Institute for Fiscal Studies | Observations In these frequent observations, we look at aspects of topical issues related to our research programme. To sign up to receive email alerts when new observations are posted, please email Bonnie Brimstone. Sun, 29 May 2016 13:29:03 +0000 <![CDATA[Taxman to raise same proportion of national income as before crisis, but from different places]]> The Great Recession triggered the two largest annual falls in real government receipts since at least 1956. Yet, by the end of the decade, tax receipts as a share of national income are due to return to almost their pre-recession level. But, beneath this apparent stability in the overall tax take, there have been significant shifts in the composition of tax revenues.

A new IFS Briefing Note, published today, sets out ‘The changing composition of UK tax revenues’ in the decade up to 2020. Tonight, the IFS and the Chartered Institute of Taxation will hold a debate – Mind the Gaps? What are the biggest risks to the UK tax take and how might tax policy and administration respond to this? – on how taxes are changing, and what risks this poses to tax receipts. 

Under current forecasts, tax receipts in 2020–21 are due to come in at 37.2% of national income. Compared with 2007–08, the taxman looks set to raise more from VAT but less from other indirect taxes, about the same from personal income taxes but with more of that coming from the highest earners, less from the main property taxes and substantially less from corporation tax. The Treasury will be more reliant on a range of small taxes, including five entirely new taxes that, combined, are forecast to raise an additional £7.3 billion in 2020–21.

The most notable changes, highlighted below, are driven by policy choices. Whether these changes have been part of a clear and coherent overarching strategy is, to put it kindly, unclear.

Tax revenues have been boosted by an increase in the rate of VAT to 20% in 2012. Revenues from other indirect taxes have fallen, largely because fuel duty has been consistently frozen at 2011 levels. This (political) choice to deviate from increasing fuel duty in line with inflation costs £4.4 billion a year in 2015–16 terms. Should freezes persist over the next five years, fuel duty revenues will grow even more slowly than the OBR forecast and be substantially lower in the longer term.

Between 2007–08 and 2015–16, there has been a fall in the share of the adult population who pay income tax (from 65.7% to 56.2%) and, for the remaining tax payers, an increase in the proportion of income tax paid by the top 1% (from 24.4% to 27.5%). This increased reliance on a small number of income tax payers follows a longer-run trend that was driven largely by above-average increases in top incomes. Since 2008, this increased reliance has been largely driven by the policy choices to increase the personal allowance, cut the higher-rate threshold, introduce the additional rate and cut pension tax relief.

Corporation tax always moves with the economic cycle, and since 2008 receipts have been substantially hit by weak profitability in the banking sector. There have also been many reforms in this area. Overall, corporation tax policies between 2010 and Budget 2016 (including those that are due to come in before the end of the parliament) have resulted in a revenue cost of £10.8 billion a year in 2015–16 terms. Moves to broaden the base and crack down on avoidance and new taxes on banks have not been sufficient to outweigh the cost of cutting the corporation tax rate from 28% to 17%. The overall trajectory of corporation tax receipts will continue to depend on the strength of growth in corporate profits and the extent to which lower rates boost activity. A permanent decline in onshore corporation tax revenues would mark a break with the previous trend which, despite continuous predictions to the contrary, was for onshore corporation tax receipts to be quite steady over time once cyclical effects were excluded.

Two new taxes on banks – the bank levy and bank surcharge – were introduced in response to the lower revenue stream coming from banks and, in part, to the view that banks should contribute to the public finance cost of the crisis. These measures have buoyed receipts, but they were not underpinned by a clear strategy. Notably, the bank levy was ratcheted up almost constantly in an attempt to squeeze more revenue out of banks before an abrupt about-face in response to concerns that it may be having undesirable effects, including increasing the likelihood that HSBC left the UK. More thought should be given to whether, and if so how, the banking sector should be taxed differently from other sectors.

More broadly, the new taxes, which also include a diverted profits tax, an apprenticeship levy and a sugar levy, have tended to be introduced hastily and without consideration of the full set of effects.

There is always uncertainty around forecast tax receipts. The risks to revenue streams are currently larger than usual: there is still uncertainty about the strength of the recovery, it is difficult to forecast the receipts from new taxes and there is policy risk in the sense that the government may choose to deviate from the assumptions embedded in forecasts. A long-term strategy for the tax system would help to alleviate some of these risks.

]]> Tue, 26 Apr 2016 00:00:00 +0000
<![CDATA[The distribution of household wealth in the UK]]> A special issue of Fiscal Studies launched today shows how wealth is concentrated among a small number of households, and is much more concentrated than incomes. Younger generations are on course to have less wealth at each point in life than earlier generations and inheritances do little to even out wealth holdings. It also draws attention to the relative lack of data on wealth holdings, especially among the very wealthiest. Wealth is a key determinant of wellbeing. It matters to households whether they have enough savings to see themselves through retirement and it matters for how they would respond to economic shocks and to fiscal and monetary policy. So understanding the distribution of wealth matters.

In fact, as the papers published in the volume show, we have been learning a lot about the wealth distribution in recent years, especially following the introduction of the Wealth and Assets Survey. But this survey cannot tell us much about the top 1% who hold around 20% of household wealth. So it is concerning that HMRC have consulted on discontinuing their publication of statistics on top shares of wealth (derived from data on bequests). These statistics have for decades given us the only, albeit imperfect, window into the wealth of the very richest.

The rest of this observation highlights what we know about the wealth distribution in the UK.

The wealth distribution in the UK

Figure 1 (taken from ‘Household Wealth in Great Britain: Distribution, Composition and Changes 2006–12' by Crawford, Innes & O’Dea) provides an overview of the distribution of household wealth in Great Britain. The household population is divided into 100 groups– and is ordered from those with the least wealth (those on the left) to those with the most.

The least wealthy one percent of households (the 1st percentile) have negative net wealth (i.e. debts in excess of any gross wealth) of more than £12,000 per adult. Net wealth holdings are also negative for those in each of the next 8 percentiles (that is 9 per cent of households have no positive net wealth). Wealth at the median is £104,000 per adult and at the very top increases dramatically across a small number of percentiles – wealth at the 95th and 99th percentiles is £0.7 million and £1.4 million per adult respectively (and there is reason to believe that the top shares - estimated using survey data -are underestimates).


Percentile plot of total household wealth per adult

Note: Weighted sample of all households interviewed in WAS Wave 3 (2010–12). Household wealth comprises gross financial wealth, gross housing wealth, private pension wealth less mortgage and non-mortgage debt.

Wealth inequality

  • Wealth is far more unequally distributed than is income. The Gini coefficient (a summary measure for how unequal a distribution is) is 0.64 for wealth. This compares to 0.34 for net income (see Table 1 of Crawford et al.).
  • The wealthiest 1% of households hold about 20% of household wealth, the top 5% of hold approximately 40%, and the top 10% hold over 50% of wealth (see Table 1 of Alvaredo et al.). These authors also find that household wealth in the UK has become more concentrated since the turn of the century. Unfortunately, the data available do not permit more concrete statements about the extent to which this is the case.

Changes in wealth since 2006

Data covering 2006 to 2012 is now available from the Wealth and Assets Survey. Crawford et al. track changes in wealth over this period and find that:

  • Average household wealth increased in real terms between 2006 and 2012. These increases are largely driven by increases in pension wealth; average household wealth held outside pensions fell in real terms between these years, except for the youngest households.
  • The rate of increase in real wealth over the period 2006 to 2012 suggests that younger cohorts are on course to have lower real wealth on average at each age than earlier generations. 

Inheritances and the wealth distribution

Crawford and Hood, use the English Longitudinal Study of Ageing (a survey that contains a representative sample of individuals in England aged 50 or over), to investigate the effect that the receipt of inheritances has on the distribution of wealth.

  • Inheritances are smaller in absolute terms for those lower down the wealth distribution, but they are more important relative to other wealth holdings. Inheritances therefore act to make the distribution of non-pension wealth more equal. 
  • However, this inequality-reducing impact of inheritances and gifts shrinks (or even disappears) when public and private pensions are included in the measure of household wealth. The impact of transfers on the distribution of this broader measure of wealth gives a better indication of the impact of transfers on lifetime resources.


Measuring the wealth distribution is difficult and there is value in using all available data sources (survey data, data from estates, data from capital income etc.) to obtain the best possible understanding of the wealth distribution. Alvaredo, Atkinson and Morelli conclude (and we agree) that data on wealth in the UK and elsewhere are in need of substantial and continued investment if researchers are to be able to communicate firm conclusions to policymakers about trends in the wealth distribution. We hope that the papers in this Fiscal Studies issue will spur further discussions on these issues and ensure that the UK and other countries have the best possible infrastructure for the collection and analysis of household wealth.

]]> Tue, 19 Apr 2016 00:00:00 +0000
<![CDATA[The EU Budget: a guide]]> In a new report and accompanying interactive online tool out today IFS researchers provide an explanation of how the EU budget works, its size, where revenues come from and what the main areas of spending are. They also provide an estimate of the UK’s net contributions to the EU. The overall net contribution will be a little over £8 billion a year going forward, though it fluctuates from year to year and was £7.5 billion in 2012, £9.1 billion in 2013 and £5.7 billion in 2014.

This is not an estimate of how much stronger the public finances would be if we were to leave the EU. That would depend in part on the deal reached with EU – it is possible that an alternative arrangement of relations with the remaining EU countries would involve the UK continuing to make significant contributions to the EU Budget. More importantly it would depend on the economic effects of leaving. We will come back to the overall fiscal consequences of Brexit in a later publication.

The EU budget and the UK contribution

The overall EU budget is about 1% of the EU’s GNI (Gross National Income) and GDP. That compares with national budgets of between 35% and 58% of GDP. European Commission figures show that the UK’s gross contribution to the EU Budget was £11.3 billion in 2014, compared to total UK public spending of £734 billion. We received back £5.6 billion through various programmes leaving a UK net contribution of £5.7 billion.

Different numbers have been put in the public domain. Much larger numbers for the gross contribution (figures of £18.8 billion in 2014, £17.8 billion in 2015 have been quoted) don’t take account of the UK’s rebate. Ignoring the rebate in this way does not seem sensible.

As for net contributions we have taken our figures from the European Commission, which include flows from the EU which don’t go through Whitehall departments – for example grants to universities – since this is money that comes back to the UK even if it does not go via the UK government. Other figures, which exclude these flows and account for the timing of payments in a different way to the European Commission figures mentioned above, give a net contribution of £9.8 billion in 2014. The OBR has forecasted that the net contribution calculated on this basis will average £9.6 billion a year between 2015 and 2020. Since the receipts by the non-governmental sector that are excluded from these figures are around £1-£1.5 billion a year, we can expect the UK’s overall net contribution to average just over £8 billion a year in the medium term, although it’s important to note that figures can jump about rather a lot from year to year.

Funding and spending the budget

If those are the headline figures, what else can be said about the EU budget?

The EU’s budget is rather complex and opaque. While there is a rational process in place to determine its size and allocation it is, perhaps inevitably, subject to considerable political horse trading between countries. The UK’s rebate on its contributions to the EU budget is perhaps the most famous of the special deals negotiated by a member state but it is far from unique. There are numerous allowances, rebates, additional allocations and the like negotiated within it.

It raises most of its revenues from three sources. Nearly three quarters comes simply from GNI based contributions – that is countries contribute according to their Gross National Income. This is a pretty straightforward and sensible basis for funding. About 13% of the EU budget comes from so called “VAT based contributions”. This is neither straightforward nor sensible as contributions are based on a hypothetical VAT base that no EU member state actually applies. This element remains because of earlier hopes that VAT could be fully harmonised and a source of EU revenue. This method disadvantages countries in which spending on goods and services that form part of this hypothetical construct constitute a large fraction of national income. Finally tariffs on goods entering the EU provide the remainder of the budget. These tariffs sensibly belong to the EU given that the goods are entering a single EU wide market. The fact that countries which collect the tariffs get to keep 25% (falling to 20%) as costs of collection, though, seems less reasonable. The average cost of collecting taxes is, thankfully, a tiny fraction of this.

The spending side of the EU budget is dominated by two spending areas which together account for over three quarters of the budget: Structural and cohesion funds on the one hand, and agriculture and rural development on the other, each account for about 38% of total EU spending.

Cohesion funds go to the poorer EU nations – those with GNI per capita below 90% of the average. Given the way the EU is funded they, alongside the structural funds which go to poorer regions, ensure that the EU budget is redistributive from richer countries such as the UK to poorer countries, largely in Eastern and Southern Europe. Recall though that the total budget is just 1% of GNI so the scale of redistribution is inevitably limited.

Structural funds go to regions within countries according to how poor they are relative to the EU average, but also according to levels of employment and population sparsity. Relative to other rich countries the UK looks like it should do relatively well from these funds because it has some really quite poor regions such as West Wales and the Valleys and Cornwall. In fact this is offset by our high employment rates and high population density and a relatively poor track-record of getting European Commission approval for projects to be funded from our structural funds allocation. A number of other countries such as Germany have also negotiated special deals for regions that would usually not qualify under the standard rules.

The agriculture and rural development budgets remain large, though they have been shrinking as a fraction of the total budget. They have also been reformed such that they no longer directly subsidise production – we no longer create wine lakes and butter mountains. The exact basis for their allocation is obscure though – and this is deliberately so. The idea is that by avoiding explicitly spelling out the formulae used, that it is easier for agreement to be reached (otherwise much time may be spent by member states trying to tweak the formulae in ways that benefit them and can attract the support of other influential members). The cost is a lack of transparency.

The UK gets relatively little from these budgets, partly because we have relatively little farmland given the size of our population and partly because, for historical reasons, we get lower payments per hectare of farmland than many other countries. It was, in large part, because of these relatively low receipts that the UK negotiated for and obtained its rebate back in the 1980s.

There is clearly room for reform and improvement of the budget and budget processes on both the spending and revenue side. But since the overall fiscal flows between the UK and the EU are relatively small – our gross contribution is around 2% of public spending, our net contribution around 1% - the scale of the benefits to the UK from improving the budget processes should not be overestimated.


This report was produced with funding from the Economic and Social Research Council’s The UK in a Changing Europe initiative.

]]> Wed, 06 Apr 2016 00:00:00 +0000
<![CDATA[The new (not yet flat rate) state pension]]> The new state pension (also known as the ‘single-tier’ or ‘flat-rate’ pension) was legislated for by the coalition government and affects all those who reach state pension age on or after tomorrow (6 April 2016) – that is, men born on or after 6 April 1951 and women born on or after 6 April 1953. Anyone born before these dates will be unaffected.

The government has been keen to stress two selling points of the new pension arrangements. First, it will be more generous to many of those approaching retirement who have spent time self-employed or out of the labour market for reasons such as caring for children or other adults, and who accrued lower state pension entitlements under the old system as a result. Second, the new state pension will be a much simpler system. The message from the government has been relatively simple: if you ‘contribute’ for at least 35 years, you will receive a full state pension of £155.65 per week.

Short run gainers

In the short run, many will indeed gain from the new pension. Our analysis suggests that 43% of those reaching the state pension age between 6 April 2016 and 5 April 2020 are likely to receive a higher state pension under the new system than under the old system. Women and the self-employed are more likely to gain than other groups: we estimate that 61% of women and 55% of those who have been self-employed for more than 10 years will gain under the new system. We estimate that women will gain on average £5.20 per week in additional state pension income at the state pension age, and those who have been self-employed for at least 10 years will gain an average of £7.50 per week. This is, arguably, a very generous bonus to the self-employed, whose lower National Insurance contributions have historically been justified on the grounds that they accrued lower benefit entitlements. The losers will be those who have accrued fewer than ten years’ contributions – under the old system they would have received a small pension, now they may receive nothing.

Simplicity oversold?

However, the simplicity of the new system has been at best misunderstood and at worst overstated. It is certainly not the case that everyone retiring after today will receive £155.65 per week in state pension, nor even that all those with at least 35 years of 'contributions' will do so. Transitional arrangements mean that some people will receive more than this amount and others less. Our analysis suggests that only 17% of those reaching the state pension age over the next four years will receive a state pension worth exactly the single tier amount, while 23% will enjoy a higher income and 61% will receive a lower state pension income.

There are good reasons for these differences in entitlements. Those enjoying more than £155.65 per week are those who have already accrued entitlement to more than this amount under the old state pension arrangements: they will see that entitlement protected.

The main reason why so many people will receive a single tier pension below the full amount is because they have had periods in the past when they ‘contracted out’ of the additional state pensions – that is, they paid reduced rates of National Insurance contributions in exchange for reduced pension entitlement. These people will have their state pension reduced in recognition of the lower contributions that they paid. Eight-in-ten of those reaching the state pension age over the next four years will have been contracted out at some point during their lives. Although these individuals have a relatively low state pension entitlement, in return for contracting out they will have accrued rights in a private pension that are (in expectation) worth at least as much as the state pension forgone.

These arrangements for those who have been contracted out are a necessary complexity of the transition process. It would have been very generous to these people not to reduce their state pension in recognition of the lower contributions they made as a result of contracting out. In fact, the transition arrangements chosen, which allow those who have been contracted out in the past to ‘work this off’, already treat this group more generously than otherwise-identical individuals who have instead been contracted in.

There are sensible reasons why not everyone is entitled to the same state pension amount, but this is likely to still come as a surprise to many people. So much of the government and media focus has been on the 'flat rate' amount of £155.65 per week that people can easily be forgiven for expecting that amount to apply to them. In the long run anyone with 35 years of 'contributions' will get the same £155.65 per week amount, but that will not happen for many years.

The real simplicity – and long run losers

The new pension system is simpler than the old one even in the short run, but in a more subtle way. After 6th April, DWP will be informing all individuals of their 'foundation amount' – that is, how much state pension they would be entitled to now if they undertook no further activity. Then for each additional year of 'contribution' going forwards, an individual will accrue an additional state pension entitlement of £4.44 per week (i.e. 1/35th of £155.65) up to a limit of £155.65 per week. In other words, accrual to the new state pension will be 'flat rate' and it will be much easier for individuals to understand how much an additional year of activity will add to their state pension income entitlement.

What the government has been less keen to point out is that – after an initial transition period – for most individuals that rate of accrual will actually be lower than accrual under the state pension system that is being replaced. Ministers and others have frequently compared the new state pension amount of £155.65 per week to the old basic state pension amount of £119.30 per week. Understandably, many individuals have drawn from this that the new state pension is more generous. However, virtually all 'contributing' individuals (the self-employed being the main exception) would have – under the old system – accrued entitlement not only to the basic state pension but also to the 'state second pension'. Taking this into account, an extra year of activity would actually have earned you more in state pension rights under the old system than it will under the new system. Of course this has the advantage that, over the longer term, the new single-tier state pension will be cheaper than the system it replaces and therefore the reform strengthens the long-run public finances.


The long-run objectives of the single tier pension system are clear. The aim is to replace the complex mess of existing rules (which had been built up through repeated tweaks to the system) with a new, far simpler system that rewards a wide range of ‘contributions’ (whether that be paid employment or caring for children) in exactly the same way. This is a simplification and rationalisation of a complex system which has proved fearsomely difficult to reform just because of its complexity. This also represents the culmination of more than 30 years of efforts to remove the earnings-related element of the state pension that was first introduced in 1978.

But continued complexity is unavoidable in the short-run, as people are moved over from the old system to the new. There is a considerable risk of disillusionment as people start claiming pension incomes this year. Given the rhetoric around the policy, it might come as a nasty surprise to many that their state pension income is in fact less than the full ‘flat rate’ amount of £155.65 per week. It would be a shame if such disillusion was to threaten the sustainability of what is on balance a sensible reform. It is also important to be clear that in the longer-term, as well as achieving a genuinely simpler system, the new single-tier pension will be less generous – and therefore less costly to the taxpayer – than the system it replaces.

]]> Tue, 05 Apr 2016 00:00:00 +0000
<![CDATA[Is the new soft drinks levy well designed?]]> In Budget 2016 the Chancellor announced a 'soft drinks industry levy' due to take effect from April 2018. The charge will be levied on soft drinks that contain added sugar and is aimed at “help[ing] tackle childhood obesity.”

New analysis by IFS researchers, published as an IFS Briefing Note today, shows over 90% of households purchase more than the recommended share of their calories as added sugar and carbonated and non-carbonated soft drinks account for 17% of added sugar purchases. Households with children obtain around 50% more of their added sugar from soft drinks, compared with households with no children. In addition, households that purchase larger amounts of sugar overall tend to buy a higher share of their added sugar as carbonated and non-carbonated soft drinks. Comparing the top 20% and the bottom 20% of households based on their share of calories from processed added sugar, households that purchase the largest amounts of sugar get around twice as much of their sugar from carbonated and non-carbonated soft drinks as households that purchase the lowest amounts of sugar.

The case for government intervention to reduce sugar intake is that there are costs associated with consumption that are not taken into account by the individual when choosing what to eat. As is the case with alcohol consumption, these costs include publicly funded health costs of treating diet-related disease or unanticipated future health problems. In the case of sugar these costs are likely to be most severe for children and for people who consume a lot of sugar. Given that these groups get a relatively high share of their sugar from soft drinks the soft drinks industry levy looks to be reasonably well targeted.

The effectiveness of the tax in reducing sugar intake will depend on how strongly people switch from the products that are taxed, and, crucially, what products they switch to buying instead. The levy will not apply to fruit juice, milkshakes, chocolate or confectionery. If consumers respond to higher soft drink prices by switching to these alternatives then the impact of the tax on sugar intake may be partially offset. Exemptions from the levy have been promised for small operators. The justification for this is unclear and could potentially lead to an increase in small producers offering high sugar products that escape the tax. In addition, how the major producers and retailers of soft drinks respond will be important. It is unclear how much of the tax will be passed-through to consumer prices. The extent to which manufacturers “reformulate by reducing the amount of added sugar in the drinks they sell” is also uncertain and will be important.

A tax on the sugar in soft drinks targets only one source of added sugar; a broader based tax on all sources would likely lead to larger reductions in dietary sugar. However, such a tax would potentially be more complicated to implement. There may also be concerns that some foodstuffs on which the tax would fall provide other important nutrients. That is not a concern with a tax on soft drinks. It is worth noting though that, along with biscuits and confectionery, soft drinks are among the few foodstuffs that are already subject to VAT. The continued zero rating for VAT purposes of cakes and other sugary foods remains as an effective tax subsidy to their consumption.

None of this is to suggest that a tax on sugary drinks is necessarily a bad idea. The structure of the tax should be carefully designed though. A sensible starting point would be to directly tax the sugar content of targeted products. The announced soft drinks industry levy does not do this. The government have set a specific revenue target, which the OBR has computed as corresponding to a rate of 18 pence per litre for drinks with 5-8 grams of sugar per 100 millilitres, and a higher rate of 24 pence for drinks with more than 8 grams of sugar per 100 millilitres. This means that, in many cases, the tax per gram of sugar is declining in the amount of sugar per 100 millilitres in a product – see Figure 1. This creates unnecessary anomalies – for instance a 1 litre product that contains 100g of sugar will attract the same amount of tax overall, and only two thirds as much per 100 grams of sugar, as a 1 litre product containing 50% more sugar.

The new soft drinks industry levy is not scheduled to come into effect until April 2018. This gives the chancellor time to adjust the structure of the tax to make sure it better targets the sugar content of products.

Figure 1: Soft drinks industry levy by drink sugar content



Figure 1: Soft drinks industry levy by drink sugar content 

Note: We gratefully acknowledge financial support from the European Research Council (ERC) under ERC-2009-AdG grant agreement number 249529 and the Economic and Social Research Council (ESRC) under the Centre for the Microeconomic Analysis of Public Policy (CPP) grant number ES/M010147/1 and under the Open Research Area (ORA) grant number ES/I012222/1.

]]> Thu, 24 Mar 2016 00:00:00 +0000
<![CDATA[Scotland’s fiscal position: an updated assessment]]> The Scottish Government’s Government Expenditure and Revenue Scotland (GERS) estimates the overall levels of government revenues and spending in Scotland and the implicit budget deficit or surplus in the previous year. With a new version of GERS reporting figures for 2014-15 and new UK-wide forecasts in the OBR’s March 2016 Economic and Fiscal Outlook (EFO), this observation provides updated projections of Scotland’s fiscal position for the next five years.

The public finance projections

Table 1 compares our latest projections for the UK and Scotland’s overall “net fiscal balance” – that is the difference between overall revenues and overall spending, including investment spending – with similar projections we made last year. For simplicity we refer to a negative net fiscal balance as a budget deficit and a positive net fiscal balance as a budget surplus.

Table 1: Net Fiscal Balance, UK and Scotland, 2013–14 and 2014–15 (outturns), 2015-16 to 2020-21 (projections)

Net fiscal balance









Latest Projections, % of GDP




























Cash-terms difference









Previous Projections, % of GDP




























Cash-terms difference









Source: Author’s calculations using GERS 2013-14 and 2014-15, and OBR EFO March 2015 and March 2016.

Since our last projections were made the OBR has revised its forecasts, with the UK deficit now forecast to be a little higher in the period from 2016-17 to 2018-19. This reflects lower underlying revenue growth due to a weaker economic outlook, and a slower pace of spending cuts than planned back in Spring 2015. However, by the end of the forecast horizon, the budget position is expected to be a little better than previously expected: a surplus of 0.5% of national income (GDP) in 2019-20 and 2020-21. This reflects new net tax raising measures, and the extension of spending cuts into 2020-21 (as well as shuffling some revenues and spending between years as described in our Post Budget 2016 analysis).

However in the case of Scotland, our projections are for a larger budget deficit in each and every year. For instance, our latest projections imply a budget deficit of around 9.4% of national income in the coming financial year, 2016-17, compared to 6.8% in our previous projections.

There are a number of reasons for these downgraded projections, including that:

  • The estimates of Scotland’s budget deficit in 2013-14 contained in GERS 2014-15 were higher than those in the previous edition of GERS, reflecting upwards revisions to estimates of government spending in Scotland. This higher level of government spending is estimated to have persisted in 2014-15, and then feeds into our projections for 2015-16 and future years.
  • Oil and gas revenues are now forecast to be lower, following further falls in expected oil prices, and cuts to the tax rates levied on oil and gas producers. Indeed, over the next few years revenues are expected to be negative: -£0.8 billion a year, on average, between 2015-16 and 2019-20, compared to +£0.7 billion a year in last year’s forecasts. Given that the majority of these revenues would have come from operations in Scottish waters, the impact of these further declines on the Scottish deficit is proportionately much larger than that on the deficit of the UK as a whole.
  • Declines in oil and gas prices, profits and investment also mean that Scotland’s economy has grown less quickly than previously projected. This means that a given cash deficit represents are larger share of the, now smaller, economy.

The Table also quantifies the differences between the projected Scottish budget deficits and the OBR’s forecasts for the UK as a whole. In 2016-17, for instance, the “fiscal gap” is projected to be 6.5% of national income (9.4% less 2.9%), which in cash terms is equivalent to about £10.6bn, or around £2,000 per person in Scotland. That is the size of the Scottish deficit on top of its share of the overall UK deficit (which is £850 per person in the UK in the same year). The projected gap then remains at a broadly similar percentage of national income over the following 4 years.

Interpreting the figures

So what are the implications of these higher budget deficit figures for Scotland?

Scotland is largely insulated from the consequences of the substantial gap between the government revenues it generates and the government expenditure undertaken in or on behalf of Scotland. This is because the Scottish Government gets most of its funding in the form of a block grant from the UK government, and the UK government uses revenues from across the UK to pay for non-devolved items like social security benefits and defence. The devolution of tax and welfare powers under the Scotland Bill 2015-16 will transfer some fiscal risk – and fiscal incentives – if its revenues or spending grows less or more quickly than those of the rest of the UK in the years ahead. But it does not transfer any responsibility for the existing larger gap between revenues and spending in Scotland. And while the Scottish Government can vary income tax, for instance, to increase or reduce the amounts it raises from these new powers, it cannot adopt a different fiscal stance to that of the UK government (changes in revenues must be balanced by changes in spending).

Figures on the Scottish deficit would be much more important if it became fully responsible for managing its own public finances. That is if it were to be fully fiscally autonomous or an independent state.

However, it is important to realise that our projections are calculated on the same basis as GERS, which allocates to Scotland a population-based share of spending on things like defence and interest payments on the UK’s national debt. The projections also assume Scotland’s onshore revenues and spending grow in line with those in the rest of the UK. Independence could have implications for the validity of these assumptions:

  • An independent Scotland might have been able to negotiate a good deal on the share of the UK’s debt it took on. Lower debt would mean lower debt interest payments and would therefore reduce the budget deficit. However, it is worth noting that even if an independent Scotland had inherited none of the UK’s central government debt, its budget deficit would likely still be substantial: around 7.6% of national income (rather than 9.4%) in 2016-17 and 4.4% of national income (rather than 6.2%) in 2020-21 holding all other elements of our projections fixed.
  • Independence could affect Scottish economic performance. A weaker performance – which perhaps may be expected in the short term – would tend to push up Scotland’s deficit. But if, as the Scottish Government have previously claimed, independence would allow policies to grow the Scottish economy more quickly, such faster growth would tend to push up revenues and reduce Scotland’s deficit.
  • Independence would also, in principle, give the Scottish Government more freedom to tax and spend more or less, which could have implications for the Scottish budget deficit. In practice, however, if an independent Scotland faced a budget deficit anything like that in our projections, spending cuts or tax rises would be needed to put the public finances on a firmer footing.

But while the precise numbers would almost certainly differ if Scotland were independent, the recent weakening in Scotland’s public finances – driven to a significant extent by falls in oil revenues and associated economic activity – clearly would have made it more difficult for an independent Scotland to manage its public finances. The oil revenue and public finance forecasts produced by the Scottish Government in the run up to the referendum also look increasingly further away from what is now expected.

The volatility of oil revenues

Oil revenues are notoriously volatile though. For the UK as a whole, they were £6.0 billion in 2009–10, over £11.0 billion in 2011–12, and just £2.2 billion in 2014–15. This volatility also makes them difficult to forecast. The OBR, for instance, has had to revise down its forecasts in 12 out of the 13 times it has updated them.

Of course if oil prices and production had risen rather than fallen, rather than revising down earlier revenue forecasts, the OBR would be revising them up. This might mean we would be revising down projections of the Scottish budget deficit rather than revising them up as has been the case.

So it’s perhaps an unfair criticism of the Scottish Government to say it got its forecasts of oil revenues wrong – so did the OBR, and any revenue forecast for something as volatile as oil will be ‘wrong’. The right response to this is to take this uncertainty into account when setting policy. Therefore more problematic is the fact that in its analysis of the potential path of oil revenues, the Scottish Government considers scenarios where the revenues come in higher than the OBR forecasts but does not consider scenarios where revenues come in less than the OBR forecasts. In other words scenarios are skewed to the “upside” – and this can prove problematic if, as has happened, revenues keep coming in under forecast.

Having said this, it’s important to remember revenues can come in ahead of forecasts too. The OBR’s forecasts assume an oil price of $35.50 during 2016 but the recent modest rebound in prices means they have averaged around $41.00 in the last week. The additional revenues this may bring in if sustained would be far from enough to fill the “Fiscal Gap”. But it’s a timely reminder that what comes down can also go up.

Notes on methodology for projecting Scotland’s fiscal position beyond 2014–15

In order to project forward the GERS 2014–15 figures to the period covering 2015–16 to 2020–21 using figures from the OBR’s March 2016 EFO, the following method is used:

  • Spending is projected on the basis that government spending in Scotland remains the same proportion (9.2%) of UK-wide government spending as in 2014–15.
  • Onshore taxes are projected on the basis that the amount paid per person in Scotland grows in line with forecast growth in onshore revenues per person for the UK as a whole. This means onshore tax revenues per person in Scotland are projected to be 96.6% of the average for the UK as a whole, as in 2014–15.
  • Offshore (oil and gas) taxes are projected under the assumption that Scotland’s share of overall UK offshore tax revenues remains the same as in 2014–15 at 82%.

The same basic set of assumptions was used in our last projections too, although these were, of course, based on GERS 2013-14 and the figures available in the OBR’s March 2015 EFO.

We have chosen the assumptions on the basis of their simplicity. As with any economic or fiscal forecast or projection, the projections outlined in this observation are subject to a number of sources of potential error that mean actual outturns will differ. This includes errors in the OBR forecasts for the UK as a whole; and trends in spending and government revenues in Scotland relative to the UK differing from the above assumptions. There are some reasons to suggest that, if anything, the assumptions are more likely to lead us to under-estimate rather than over-estimate Scotland’s fiscal deficit relative to that of the UK as a whole. First, Scottish Government plans to borrow additional money to fund capital investment mean Scottish Government spending may fall less between 2014–15 and 2020–21 than equivalent spending in the rest of the UK. This would tend to increase Scotland’s share of overall UK government spending; in contrast, we have assumed this share would remain constant. Second, the OBR forecasts revenue growth to be particularly strong for taxes like capital gains tax, inheritance tax and stamp duties, which make up a relatively smaller share of Scottish revenues. All else equal, this would tend to suggest growth in revenues per person in Scotland would be lower than for the UK as a whole. Third, while our revenue projections account for declines in oil revenues, our projections assume that GDP from the North Sea rises in line with onshore GDP. If North Sea GDP declined, as one might actually expect, then Scotland’s cash-terms deficit would represent a larger percentage of GDP. (Of course, as noted above, if oil prices rebound, oil revenues and North Sea GDP would likely grow more quickly than we have projected).

Figures for Scotland’s deficit if it inherited a 0% share of UK central government debt are calculated by subtracting estimates for Scotland’s population share of the UK’s central government net debt interest payments from our baseline projections for Scotland’s public spending.

]]> Wed, 23 Mar 2016 00:00:00 +0000
<![CDATA[Are we 'all in this together'?]]> There has been a lot of dispute in recent days over the extent to which “we have all been in this together” or government tax and benefit reforms have been “fair”. There are obviously many different ways of assessing this. In this observation we draw on recent IFS work to provide some assessment of what has happened to living standards across the distribution and what has been the direct effect on incomes of tax and benefit policy.

In broad terms income inequality is lower than before the recession as increasing levels of employment have helped those towards the bottom of the income distribution and falling real wages have hit those in work, including higher earners. We would expect that equalisation to unwind as further benefit cuts bite and earnings start to rise, such that inequality at the end of the decade is likely to be similar to inequality at the start.

The direct effect of government tax and benefit policy, on the other hand, has been to take money from those working age benefit recipients towards the bottom of the income distribution. That reflects in part some unpicking, but by no means a complete unpicking, of the very big increases in tax credits introduced by the last Labour government. Those in the middle and upper parts of the income distribution – including most pensioners and people on average earnings and above – have been remarkably well protected from tax and benefit changes on average. Meanwhile the very highest earners - those on over £100,000 a year - have seen significant tax increases.

Chart 1 illustrates the pattern of overall changes in living standards. The solid line shows that there has been a considerable equalisation of the income distribution in the years since the recession, with incomes rising for those towards the bottom of the distribution and falling for those towards the top. This reflects a combination of rising employment, falling earnings and some increases in benefit income (between 2007–08 and 2009–10). On some measures, inequality is now at a 25 year low.

The lighter dotted line shows our projections of what will happen to incomes over the next five years. This line slopes in the other direction. The lack of real income growth at the bottom reflects further benefit cuts, while the better performance further up is dependent on real earnings rising as expected by the OBR. Finally the darker dotted line shows our projections for the period as a whole (which also of course depends on earnings rising as projected from now). It suggests that we should expect much of the recent fall in inequality to be undone over the next five years, resulting in a similar change in incomes for rich and poor over the whole period since the recession. Some evidence, perhaps, that we are all in it together.

Chart 1: Change in real household income by percentile point: 2007–08 to 2020–21 [Download the data in Excel]

 Figure 1: Change in real household income by percentile point: 2007–08 to 2020–21

The numbers in chart 1 are for the population as a whole. But in fact things look rather different for different age groups. Chart 2 shows our projections for what has happened to the median household incomes of three different age groups – those aged 22-30, 31-59 and 60+. There are big differences here. The median incomes of the over 60s are about 10% higher now than they were pre-crisis. Those of working age still have incomes below pre-crisis levels, with the youngest suffering most, albeit with something of a recent bounce back. The strong income performance among the over 60s results from the fact that pensioners were the least affected by falling real earnings, pensioner benefits were mostly protected, and some of the poorest, oldest pensioners  have died and been replaced by a generation with higher state and private pension entitlements.

Chart 2: Median income by age: 2007–08 to 2015–16 (2007–08 = 100) [Download the data in Excel]

Figure 2: Median income by age: 2007–08 to 2015–16 (2007–08 = 100)

The extent to which changes in the overall economy can be attributed to government policy is an open question. It is difficult to say whether employment levels have risen so much because of what this government has done right, or earnings levels fallen because of government mistakes.  However, changes to the tax and benefit system are clearly the result of policy decisions – albeit decisions taken in a context of an unsustainable budget deficit and falling inequality. Chart 3 shows the percentage gain or loss in income resulting from tax and benefit changes for each income decile from May 2010 to April 2015 split between pensioners, working age people with children and those without children. Chart 4 shows the same thing for the period from May 2015 to April 2019, including announcements in last week’s Budget.

Chart 3: Impact of tax and benefit reforms introduced between May 2010 and April 2015 by income decile and household type [Download the data in Excel]

Figure 3: Impact of tax and benefit reforms introduced between May 2010 and April 2015 by income decile and household type

Chart 4: Long-run impact of tax and benefit reforms introduced between May 2015 and April 2019 by income decile and household type (including universal credit) [Download the data in Excel]

Figure 4: Long-run impact of tax and benefit reforms introduced between May 2015 and April 2019 by income decile and household type (including universal credit)

Source: Hood and Elming (2016)

Focussing on the first period four strong conclusions can be drawn. First, tax and benefit changes had little effect on pensioners and much bigger effects on those of working age, especially those with children. Second, they have resulted in significant losses for those of working age in the bottom half of the income distribution. That is not surprising as a result of various cuts to working age benefits have taken effect. Third, those from the middle of the distribution most of the way up (most people on average earnings and above, certainly up to £50,000 or so a year) saw very small changes in income, on average, as a direct result of tax and benefit policy. Given the scale of the overall austerity measures implemented this group were remarkably well protected from tax and benefit changes. For this group the increase in the personal allowance and falls in petrol duty largely offset the effects of the big increase in VAT and some benefit cuts on average. Remember, though, that the protection of this group does have to be seen in the context of falling real wages which have hit their living standards. Finally the top decile, and in particular those on the very highest incomes, earning more than £100,000 a year, faced some significant tax increases. Indeed, if measures introduced in the final months of the last Labour government were included too, the average loss in the top decile would rise to 6.5% of income, making them the biggest losers over the period.

The long-run impact of planned changes over the course of this parliament follows a similar pattern, for a similar set of reasons. Again, pensioners are protected while poorer working age households are hit hard, especially those with children. This is the result of the continued protection of pensioner benefits (including maintaining the ‘triple lock’ on the basic state pension) while making further deep cuts to working-age benefit spending. Again, households in the upper half of the income distribution (but below the very top) are likely to see little direct impact of tax and benefit changes on their incomes on average, as some benefits cuts and small tax rises are offset by further increases in the income tax personal allowance, and the raising of the higher rate threshold.

So the income distribution has narrowed, but tax and benefit changes planned for this parliament will likely help take it back to something like pre-recession levels. Two final points.

First, “fairness” encompasses much more than just the shape of the income distribution or the effects of taxes and benefits on it. The sources and distribution of wealth, changes in public service spending and much else besides matters. Within the narrow remit of tax and benefit policy we have shown before how one of the most important effects of the shift to universal credit (if it ever happens) will be to strengthen work incentives for some of those who currently gain the least from moving into work or earning more. That is an important element of fairness. At the other end of the income distribution there must also be limits on the amount of tax it is “fair” to ask a higher earner to pay. People will differ over where those limits are.

Second, the pattern of effects from tax and benefit changes over this parliament results directly from three policy choices: £12 billion of working age welfare cuts; the protection of pensioner benefits; and increases in the income tax personal allowance and higher rate thresholds. Not only were these policies in the Conservative manifesto last year, they were front and centre. Neither the fact of their implementation nor their distributional consequences should be a surprise.

]]> Mon, 21 Mar 2016 00:00:00 +0000
<![CDATA[Adjusting Scotland’s block grant – the options on the table]]> The UK and Scottish Governments have so far failed to agree the new 'fiscal framework' that must accompany the transfer of tax and welfare powers recommended by the Smith Commission and set out in the Scotland Bill. Perhaps the biggest bone of contention is how to adjust Scotland’s block grant to reflect the associated transfer of tax revenues and welfare spending to the Scottish Government. With another 'deadline' for an agreement looming, this observation aims to explain the proposals put forward by each government, set out their respective rationales, and analyse recent 'compromise' proposals put forward by the UK government. An accompanying briefing note provides additional information, and a detailed report to be in Edinburgh on 22nd March will give our full assessment on the proposals or agreements as they then stand.

We find that there are clear rationales behind the positions of both the UK and Scottish governments. As our report last November showed, it seems impossible to design a system that will satisfy all the Smith Commission’s principles. Both governments claim adherence to these principles, but prioritise them differently.

The Scottish Government emphasises the principle that there should be 'no detriment as a result of the decision to devolve further powers' – i.e. that Scotland should be no worse (or better) off simply as a result of devolution. The UK Government instead focuses on the principle of 'taxpayer fairness' which holds that changes in devolved taxes in the rest of the UK should not affect the level of public spending in Scotland after the transfer of tax powers has taken place.

We also find that the recent 'compromise' proposal from the UK government represents a significant move towards the Scottish Government's position - but that the two sides remain some way apart on how differences in population growth between Scotland and the rest of the UK (rUK) should be reflected in the block grant adjustment (BGA) indexation method. There are still hundreds of millions of pounds a year in spending power at stake.

In what follows we focus on taxes – and especially income tax – but the issue of BGA indexation is also relevant to welfare devolution too.

The Scottish Government’s position and initial proposals

There is broad agreement on how to adjust the block grant immediately after devolution: Scotland’s block grant should be reduced to reflect the amount of tax revenues transferred to Scotland. The problem lies in agreeing the BGA for subsequent years. This reflects differing interpretations of the principles included in the Smith Commission report.

The Scottish Government’s position is that 'no detriment' applies on an ongoing basis. It argues that if Scotland's revenues per capita grow at the same rate as in the rest of the UK, then the Scottish budget should be neither smaller nor larger than if income tax were not devolved and funding remained determined by the Barnett Formula alone.

This leads it to favour the indexation method known as Per Capita Indexed Deduction (PCID), where each year the BGA is increased or decreased by the rate of change of devolved revenues in rUK adjusted to account for the relative change in population between Scotland and rUK. This protects the Scottish budget both from Scotland’s lower per capita tax revenues and from Scotland’s slower population growth. The Scottish Government argues that the implicit insurance against differences in population growth provided by PCID is fair due to its lack of policy levers to influence the rate of population growth in Scotland relative to rUK.

One of the UK Government’s arguments against this approach is that a world without tax devolution is not in fact the relevant counterfactual against which Smith’s ‘no detriment’ principle should be assessed. Arrangements for the partial devolution of income tax under the Scotland Act 2012 (SA2012) are already agreed: Scotland’s block grant is to be adjusted by a different mechanism which takes no account of Scotland’s relatively slow population growth. PCID would therefore be relatively more generous than the already agreed mechanisms for the partial devolution of income tax, meaning the Scottish Government would gain relative to existing plans.

But the UK Government’s main objection is that it thinks the PCID mechanism infringes the ‘taxpayer fairness’ principle, because this approach would redistribute some of the future growth in rUK income tax revenues to Scotland.

The reason for this relates to the continued use of the Barnett Formula to determine Scotland’s block grant. The Barnett Formula increases Scotland’s block grant each year by a population share of increases in comparable spending in rUK. But the PCID mechanism for updating the BGA is based on percentage changes (not population shares).

This is important because when the BGA is initially set, it will take account of the fact that Scotland raises less income tax per capita than rUK. Thereafter any percentage increase in this BGA as a result of growth in income tax in rUK would be less than the corresponding population-share based increase in Scotland’s Barnett-determined block grant. In other words, when tax revenues increase in rUK, the increase in Scotland’s block grant would be bigger than the increase in the amount taken off its block grant to account for it’s ability to raise its own income tax revenues. This would happen both for revenue increases due to economic growth and tax policy changes. And each year more and more income tax from rUK would therefore implicitly be transferred to Scotland.

This redistribution already happens under the Barnett Formula, but the UK Government’s view is that devolution implies the end of this “pooling and sharing” for the taxes that are devolved.

The UK government’s position and initial proposals

The UK Government initially argued that the so-called ‘Levels deduction’ (LD) approach should be used to index the BGA. Under this approach the change in the BGA is given by the population share of the change in comparable aggregate revenues in rUK, not the percentage change.

One implication of the LD approach is that there would be no further redistribution of income tax revenues from rUK to Scotland. This is because, whilst the Barnett formula continues to give Scotland a population share of rUK spending increases, the LD approach removes a population share of rUK tax increases. Any increase to the Scottish budget coming through Barnett as a result of changes in income tax revenues in rUK is exactly offset by the same increase in the BGA. This therefore satisfies the UK government’s interpretation of the ‘taxpayer fairness’ principle.

However, another implication of this approach is that it sets a particularly high bar for Scottish revenue growth if the Scottish budget is to keep pace with what it would be without devolution of income tax. Because Scottish revenues per capita are lower than those in rUK, Scottish revenues per capita actually have to grow at a faster percentage rate than those in rUK to keep up with the population-shared based increase in the BGA. If Scottish revenues per capita instead grow at the same rate as those in rUK, Scotland’s budget falls relative to what it would have in the absence of devolution. The Scottish Government argues that this is inconsistent with Smith’s ‘no detriment from the decision to devolve’ principle.

A recent proposed ‘compromise’

In an earlier paper published last November we showed that the difference in the amount of money available to the Scottish government under the PCID and LD approaches could easily differ by over a billion a year after just a decade or so. It is therefore perhaps unsurprising that agreement has been difficult to reach.

In an attempt to reach a compromise, the UK Government has put forward a new proposal that moves significantly towards the Scottish Government's preferred approach. The new proposal is based on the LD approach, but takes into account Scotland’s lower initial tax revenues per capita. We term this tax-capacity adjusted levels deduction, or TCA-LD. TCA-LD removes one of the Scottish Government’s major objections to LD; it would no longer require Scottish per capita revenues to grow at a faster rate than those in rUK simply to match a world without tax devolution. But in proposing TCA-LD the UK Government has effectively conceded - in practice if not principle - its initial 'taxpayer fairness' argument for favouring the LD approach. By taking account of Scotland’s lower initial tax capacity, there would continue to be some redistribution of rUK income tax revenue increases to Scotland in future.

On the other hand this proposal would not address the Scottish Government’s other main concern: relative population growth. If revenues per capita grew at the same rate but Scotland’s population grew more slowly than rUK’s, the Scottish budget would still be worse off than it would without tax devolution. So the Scottish Government argues it still does not satisfy the ‘no detriment’ principle.

The UK Government’s response to this criticism is two-fold. First, as discussed above, that the correct counterfactual against which 'detriment' should be assessed should include the Scotland Act 2012 income tax provisions, which do not adjust for differential population growth. And second, that it would be unfair for the Scottish Government to be insulated from population-based risk on the revenue side when on the spending side it gains rather than loses from Scotland’s relatively slow population growth (because the Barnett formula does not take account of this slower population growth when allocating funds). Consistency instead requires a symmetric set of risks on the revenue side as on the spending side, which TCA-LD (but not PCID) would deliver.

Simulating the options

Can we say one side is right and the other wrong? No. Both governments make arguments that are consistent with (different parts of) the Smith Commission’s principles. However, we can attempt to quantify the effects of the different proposals on Scotland’s budget using ONS population projections for Scotland and rUK, and OBR estimates of long-term revenue growth (Table 1).

Looking at income tax only, if revenues grow at the same rate per capita in Scotland and rUK, then the Scottish Government would be no better or worse off under their proposals than without any devolution of income tax alone. The UK government’s initial proposals would have seen the amount Scotland gets from its block grant and income tax (given the specific assumptions we have made) fall by 4.5% after 15 years. Given a block grant of just under £29 billion in 2015–16, this 4.5% fall would be around £1.3 billion in today’s terms.

The UK government’s compromise proposals bring this gap down to around £0.5 billion in today’s terms. The important flipside of this is that an additional £0.8 billion of income tax revenues in today’s terms would be redistributed from rUK to Scotland on top of the amount redistributed today (albeit £0.5 billion less than under the Scottish Government’s proposals).

Compared to what would happen under the Scotland Act 2012 powers, the Scottish Government’s proposals would lead to an increase in Scotland’s annual budget of around 1% after 15 years, while the UK Government’s compromise approach would lead to a decrease of around 1% (in both cases equivalent to around £0.2 – 0.3 billion in today’s terms).

Table 1: Difference in Scottish Government’s budget from block grant and devolved income tax after 15 years, relative to two “no change” counterfactuals


Scottish Government’s Proposals (PCID)

UK Government’s initial proposals (LD)

‘Compromise’ proposals (TCA-LD)

No income tax devolution




Partial devolution under Scotland Act 2012




Source: Authors’ calculations using ONS population projections, OBR’s Fiscal Sustainability Report, HMT’s Public Expenditure Statistical Analysis and Statement of Funding Policy, and HMRC’s Disaggregated Tax Revenue Statistics

Concluding thoughts

While the UK government has made significant concessions from its initial position, there are still hundreds of millions of pounds a year to play for. There are also continuing differences about the extent to which Scotland should bear risks associated with differential population change, and whether existing devolution arrangements under the Scotland Act 2012 should influence the choice of BGA indexation. There are no right or wrong answers here – just differences in opinion. Perhaps it’s not surprising that a deal is still elusive. Reaching one will take good will and further compromises by someone.



David Bell and David Eiser are at the University of Stirling and the Centre on Constitutional Change.

David Phillips is a senior research economist at the Institute for Fiscal Studies (IFS)

An accompanying research paper, Adjusting Scotland's Block Grant - the options on the table, has been published on the Centre on Constitutional Change website. 

This analysis and accompanying paper was supported by funding from the Nuffield Foundation. The Nuffield Foundation is an endowed charitable trust that aims to improve social well-being in the widest sense. It funds research and innovation in education and social policy and also works to build capacity in education, science and social science research. The Nuffield Foundation has funded this project, but the views expressed are those of the authors and not necessarily those of the Foundation. More information is available at

The work was also supported by funding by the Economic and Social Research Council (ESRC) through the Centre for the Microeconomic Analysis of Public Policy at IFS (grant reference ES/H021221/1).

]]> Tue, 23 Feb 2016 00:00:00 +0000
<![CDATA[What does the row over Google’s tax bill tell us about the corporate tax system?]]> Corporate tax has rarely excited the imagination of the public as much as in recent years. This week Google has become the latest company to attract widespread anger over the amount of tax it has paid in the UK. The sense that there are some big, profitable companies paying relatively little in corporate tax has led many to try to allocate blame. Are multinationals simply behaving badly? Is HMRC cutting sweetheart deals with favoured companies? Have politicians failed in their task of writing the tax rules?

In an attempt to shed some light on these issues, a new paper by IFS researchers published today sets out how the current tax system seeks to tax corporate profits, what problems this can lead to and how the OECD’s two year Base Erosion and Profit Shifting (BEPS) project has sought to prevent tax avoidance. This paper is a pre-released chapter from the February 2016 IFS Green Budget, produced in association with ICAEW and funded by the Nuffield Foundation and to be launched on Monday 8th February.

The most important question relates to what we’re trying to tax. The current tax rules are not designed to tax the profits from UK sales. They’re certainly not designed to tax either revenue or sales generated in the UK. They are instead designed to tax that part of a firm’s profit that arises from value created in the UK. That is the principle underlying all corporate tax regimes across the OECD. The trouble is that calculating how much profit arises from value added in any individual country can be very tricky, and is often open to honest dispute.

Multinationals operate across tax jurisdictions and create profits from activities in many countries. Working out how to allocate profits to different jurisdictions is difficult. In practice, countries have long agreed to divvy up profits according to where the underlying value was created. But there is often no single ‘correct’ answer to how much profit should be taxed in the UK. For example: if a worker in the UK and a worker in Ireland collaborate in arranging and concluding a sale, or in designing a new product, or writing a piece of software, how much of resulting income should be attributed to UK activities?

The tax rules seek to provide an answer to this. Two elements of the rules are key. Permanent establishment rules define when a firm has a taxable presence in a country. Controversy often arises when a firm has a large revenue stream in the UK, but is not deemed to have an associated presence here for tax purposes. Rules around this will change following the BEPS process and it will become more difficult for companies to claim that they do not have a permanent establishment, but this can’t change retrospectively. Transfer pricing rules dictate the prices that a firm can charge for a transaction – including payments for services or for the use of ideas – that happens between two parts of the same firm that are located in different tax jurisdictions. These are the rules that determine taxable profit allocation. Yet the rules can never be detailed enough to set out what the outcome should be in every possible case. This creates room for disagreement over what the tax rules mean. This is why HMRC is often engaged with multinationals about how much tax they pay: not because they are busy cutting special deals, but because they are trying to apply the tax rules in a consistent manner.

Multinationals are in a good position to be able to employ hordes of tax advisors that help them to conclude any uncertainty in way that leads to lower tax bills, and to take advantage of any loopholes to avoid tax. Some of those loopholes are well known and many exist in other countries’ tax regimes. For example, the well documented “Double Irish” refers to differences between Irish and US tax laws that allow US multinationals to shift profits out of Ireland to tax havens such as Bermuda. These kinds of gaps in tax systems can create opportunities for tax avoidance on a grand scale. There is literally nothing the UK government can do unilaterally about some of these loopholes.

If the outcomes produced by the current tax rules are deemed "derisory” then there are at least two options that are more helpful than complaining that firms are behaving badly.

First, governments could improve the current tax rules to prevent certain avoidance behaviours. The OECD has been seeking to foster collaboration through the BEPS project  to do exactly this. The UK has already acted to prevent some types of avoidance structures and, going forward, will join other countries in trying to prevent tax avoidance by changing the rules that determine profit allocation. The “Double Irish” structure described above will come to an end by 2020 as a result of international pressure. On Thursday, the European Commission announced new proposals that build on the BEPS project and seek further adjustments to EU tax rules to crack down on tax avoidance. Since the opportunities for avoidance arise at the boundaries between tax systems, a multilateral approach makes sense. However, governments can face a trade-off when deciding how to act: changing tax rules can help crack down on avoidance but come at the cost of reducing a country’s competitive position. Many of the actions under BEPS are merely “recommended”. Countries are under no obligation to implement them if they think they will damage their own competitiveness. It remains to be seen how the UK government, among others, will make that trade off.

Second, it is open to government to pursue a much more radical course of action: to scrap the corporate tax system as we currently know it and write a new one that better serves our objectives. The world has changed enormously since the current system was designed in the 1920s. Companies’ activities have become more global, digital and intangible. A system that allocates profits as if they were earned by separate companies will always create tensions. We could decide to live with those tensions as best we can, or we could go back to the drawing board and design a tax system based on how the world currently looks. For example, we could tax companies based on where their sales occur rather than where their profits are deemed to have arisen. We may not be ready for such radical change yet, but depending on how well the newly patched up international corporate tax system works over the next few years we may find it is worth considering whether a new set of tensions would produce a more agreeable outcome.


1. 'Corporate tax avoidance: tackling Base Erosion and Profit Shifting' by Helen Miller and Thomas Pope is a pre-released chapter from the IFS Green Budget 2016, edited by Carl Emmerson, Robert Joyce and Paul Johnson.

2. The full Green Budget 2016 publication will be launched at 10:00 on Monday 8 February 2016 at Guildhall, London. Please email if you wish to attend;

3. ICAEW is a world leading professional membership organisation that promotes, develops and supports over 144,000 chartered accountants worldwide. We provide qualifications and professional development, share our knowledge, insight and technical expertise, and protect the quality and integrity of the accountancy and finance profession.

As leaders in accountancy, finance and business our members have the knowledge, skills and commitment to maintain the highest professional standards and integrity. Together we contribute to the success of individuals, organisations, communities and economies around the world. Because of us, people can do business with confidence.

ICAEW is a founder member of Chartered Accountants Worldwide and the Global Accounting Alliance.

4. The Nuffield Foundation is an endowed charitable trust that aims to improve social well-being in the widest sense. It funds research and innovation in education and social policy and also works to build capacity in education, science and social science research. The Nuffield Foundation has funded this project, but the views expressed are those of the authors and not necessarily those of the Foundation. More information is available at

We are delighted to have produced this year’s Green Budget in association with ICAEW and with funding from ICAEW and the Nuffield Foundation.

Additional analysis will be provided by ICAEW and Oxford Economics.

We are also grateful to the Economic and Social Research Council for funding much of the day-to-day research at IFS that underpins the analysis in this report.

]]> Fri, 29 Jan 2016 00:00:00 +0000
<![CDATA[Council tax rises to ease the pace of cuts to local government budgets]]> Yesterday was a big day for local government in England. The Department for Communities and Local Government (DCLG) published the ‘Provisional Local Government Funding Settlement’ – which sets out how much in the way of core grants it plans to give each English council every year between 2016–17 and 2019–20. The settlement will result in cuts in local councils’ spending power of around 8% on average – a much smaller scale of cuts than experienced over the last parliament. The cuts will also be more evenly spread, rather than hitting poorer authorities harder as happened between 2010 and 2015. With the additional ability to increase council tax to pay for social care, the average council tax bill for a band D property could rise by £205 a year by April 2019 if these powers are used in full, with the potential for further rises to pay for police and fire authorities.

As in the last parliament, grants to councils are set to be cut substantially over the next 4 years. Taken together, the amount councils receive in Revenue Support Grant (RSG) and other grants from DCLG are set to fall by 60%.

But councils also have other sources of revenue: they retain a portion of business rates revenues and levy and retain council tax, and taken together these are a much bigger source of revenue than grants from DCLG. And these revenue sources are expected to grow over the next four years, not least because councils are expected to raise council tax fairly significantly.

Looking at how much councils will have to spend in total, including these additional sources of revenue, the cuts will be around 7% in real terms over the next four years. This is a substantially slower pace of cuts than councils had to deliver between 2009–10 and 2015–16, when councils’ spending power was cut on average by 25% in real terms. Cuts over the next four years, though, will be front loaded, with cuts of around 4% to 5% next year, on average.

These figures are national averages though, so how might the cuts affect different councils?

The impact of cuts across councils

Figure 1 shows that over the course of the last parliament, cuts in spending power were much greater for those councils which rely a lot on grants for their funding – because they have low council tax bases, or high spending needs for instance (i.e. they are poorer) – than those less dependent on grant. In the last few years, DCLG effectively cut every council’s grant by the same percentage. Of course, a given percentage cut in grant has a bigger impact on a council if it relies more on that grant for its overall spending.

Figure 1. Percentage change in real spending power 2009­–10 to 2015–16 by initial grant reliance of local authority

Figure 1. Percentage change in real spending power 2009¬–10 to 2015–16 by initial grant reliance of local authority

Source: DCLG, Local authority outturns and budgets.

Note: Greater London Authority non-police, non-fire spending power allocated to London councils in proportion to population.

But this pattern is set to change because of a change in the way DCLG allocates cuts to grants across councils. It now explicitly takes into account the differing extent to which councils rely on grants, making smaller cuts to the grants of those which rely a lot on the grant than to those councils which are able to raise more of their own revenue from council tax.

This means, looking ahead over the next 4 years, cuts to spending power will be much more evenly distributed across councils than they were over the last parliament – as shown in Figure 2. Nevertheless, cuts will still be a bit larger on average for (principally poorer) areas who are most reliant on grant (for which cuts are set to average 9.2%), than for those who are least reliant on grant (for which cuts are set to average 6.8%).

Figure 2. Percentage change in real spending power 2015–16 to 2019–20 by initial grant reliance of local authority

Figure 2. Percentage change in real spending power 2015–16 to 2019–20 by initial grant reliance of local authority

Source: DCLG, Core spending power: provisional local government finance settlement 2016 to 2017.

Note: Greater London Authority non-police, non-fire spending power allocated to London councils in proportion to population.

Why is a bigger squeeze on poorer councils still happening if DCLG is now accounting for differences in reliance on grant funding? It is because, while DCLG’s new allocation of funding accounts for the initial level of grant reliance, it does not account for the fact that things change over time as councils’ other sources of revenue – notably council tax – grow. In particular, the forecast growth in council tax rates and revenues will do less to offset cuts to grants in areas with small council tax revenues and a high degree of grant reliance than it does in areas with large council tax revenues and a low degree of grant reliance. In other words, it is the fact that the poorer, more grant-dependent areas can do less to increase their budgets by increasing their council tax that means they will still fair a little worse over the next few years than leafier, less grant-dependent places.

Spending cuts will differ not only across different local authorities but also across different areas of spending, just as it did over the last five years (an IFS briefing note published before this year’s general election provides analyses in detail the cuts to local government spending between 2009–10 and 2014–15). Looking to the next four years, spending on adult social care is likely to be particularly protected, because some of the grant to authorities is being ring-fenced for this purpose and because councils are being given the ability to raise council tax by an additional 2% a year specifically to fund adult social care. If this ring-fenced funding and council tax revenue were used to stop further cuts to adult social care and instead offer a real-terms freeze – which may not be enough given rising demands and cost pressures – real-terms cuts to other areas of spending would need to be around 12%, on average, by 2019–20. This could mean difficult choices for other services like children’s social services, refuse collection, libraries, transport, economic development, planning and housing, some of which have already seen very large cuts. Every additional 1% increase in adult social care spending would require additional cuts to other areas of spending of around 0.5%.

Council tax

The extra council tax ‘precept’ for social care not only affects councils’ budgets, it will also affect how much council tax households pay.

Over the last few years, councils wanting to raise council tax by more than 2% have had to call a referendum. None have done so (although Bedfordshire police tried and failed to win such a referendum on the local police precept). Councils with social care responsibilities will now be able to increase council tax by a further 2% – i.e. a total of 4% – without a referendum. Other councils like districts and the Greater London Authority will still only be able to raise rates by 2% a year without a referendum.

If councils make full use of these powers, the average Band D rate in England would go up by £48 a year next April (£33 in real terms), and £205 a year by April 2019 (£97 in real terms). However, it is important to note this will follow a 5-year period when most councils have been freezing their council tax, leaving the average council tax bill a little lower in real-terms in 2019 than in 2010.

Of course, councils may not make full use of the powers. DCLG assume that councils will make full use of their social care precepts but only increase council tax rates otherwise by 1.75%, rather than the full 2% ‘standard’ rise they are allowed without a referendum. This would see council tax increasing by £191 a year by April 2019 (£84 in real terms). However, if further cuts to services become too difficult to bear, some councils might hold referendums for bigger council tax rises. And police authorities and fire authorities also levy council tax precepts – increases in either of these would push council tax bills up further.

The local government financial revolution

Taken together yesterday’s local government financial settlement therefore has some big changes in it: an easing in the pace of cuts, a change in the way cuts are allocated across councils, and the return of rising council tax bills.

Bigger changes are on the horizon though. The government will soon begin consulting on how to fully devolve business rates revenues to councils. This change, planned to take place by 2020, will mean councils’ spending power in future will be more directly linked to the performance of the local economy, – meaning additional incentives to encourage growth, but greater risk when things go wrong. A key thing still to be decided is just how much divergence between the revenues of different areas should be allowed before safety net systems kick in.

The full localisation of business rates will also represent a significant transfer of additional money to councils and the government will be asking councils to take on additional responsibilities in return for this extra money. Just what these extra responsibilities will be is as yet undecided – but it may include things like police, or public health, or even the operation and funding of parts of the benefits system (e.g. attendance allowance, which is a benefit paid to disabled pensioners).

Yesterday’s announcements are, therefore, just part of some genuinely revolutionary changes that are taking place to local government finance.  

]]> Fri, 18 Dec 2015 00:00:00 +0000
<![CDATA[Fiscal responses of six European countries to the Great Recession: a crisis wasted?]]> Most European countries experienced a significant increase in government borrowing in the wake of the global financial crisis and Great Recession. Ireland and Spain are well-known for the severe difficulties they faced but France, Italy and the UK also saw borrowing rise sharply. For these countries a combination of the reliance on tax revenues related to the financial sector and asset prices (Ireland, Spain, UK) and/or stickiness in public spending levels, resulted in a large increase in borrowing in the wake of the Great Recession. Examining how these countries responded – in terms of the size, timing and composition of the fiscal measures they resorted to – highlights some interesting similarities but also some notable differences. One common theme, unfortunately, is that these fiscal responses to the crisis largely missed opportunities to improve the overall efficiency of the tax system.

A new special edition of the journal Fiscal Studies examines – for France, Germany, Ireland, Italy, Spain and the United Kingdom:

  • the evolution of GDP, employment and unemployment rates, and the public finances in the run-up to, and through, the financial crisis;
  • the scale, timing and nature of the fiscal response; and
  • the impact of the reforms on the incomes of different households and on spending on different public services.

Compared with usual cross-country macroeconomic assessments of public finances (for example, by the IMF or the OECD), this special issue relies heavily on the use of micro data and detailed microsimulation models and hence presents a deeper analysis that aims at enhanced comparability between these countries.

The impact of the financial crisis on public borrowing

In the years preceding the financial crisis, Ireland and Spain appeared to have the healthiest fiscal positions out of these six countries – having run overall budget surpluses for at least the preceding three years. However, with the onset of the crisis, it quickly became clear that this position was propped up by unsustainable revenues from – in particular – the property market. As the ‘dots’ in the figure show, the financial crisis led to a substantial deterioration in the apparent underlying strength of Ireland’s and Spain’s public finances.

Although suffering a little less severely, France, Italy and the UK also saw their underlying public finances weaken – each by just over 5 per cent of GDP. In other words, had they taken no policy action, they would have ended up borrowing over 5 per cent of GDP more every year forevermore.

The only exception was Germany, whose public finances were – in the long term – unaffected by the Great Recession. Germany was the only one of the six countries for which this looked like a ‘text book’ recession – a temporary increase in borrowing followed by a quick return to normal times.

 Size and composition of post-crisis fiscal policy response up to 2014

Notes and sources: See Figure 4 of Bozio, et al. op cit.

Fiscal response

In many ways different countries have taken very different approaches to dealing with the fiscal problems they faced. As the figure shows, France and Italy have both relied relatively heavily on raising taxes, rather than cutting spending, to reduce borrowing. By the end of 2014, two-thirds of the measures implemented in France were aimed at boosting revenue and one-third aimed at reducing spending. Both France and Italy on course to become higher tax, higher spending and high borrowing countries than they were before the crisis hit.

In contrast, the UK has relied heavily on cuts to public spending, rather than net tax rises. Up to the end of 2014–15, 82% of fiscal measures in the UK were spending cuts, rather than net tax rises. The result is that the UK is on course to have a lower level of spending, a similar level of taxation, and a lower level of borrowing than was the case pre-crisis.

Similarities in tax measures but differences on spending

Although each country has relied to rather differing degrees on tax raising measures, the nature of the tax increases implemented across the countries shows some interesting similarities. France, Ireland, Italy, Spain and the UK all chose to increase VAT rates and to increase social insurance contributions. France, Spain and the UK also all implemented income tax rises that were focussed on the highest income individuals, while France the UK both chose to reduce corporation tax rates while widening the base for this tax.

The countries have, however, made rather different choices about which areas of spending to cut. France, Italy and Spain refrained from cutting welfare benefits (and, in fact, increased them for some groups), in stark contrast to Ireland and the UK where cuts to benefit payments for working age adults played a major role in delivering fiscal consolidation. France and the UK have both chosen to afford relative protection to spending on health and education, while Italy and Spain have chosen to cut these services more deeply than other service areas.

Countries have also made different choices about which households should bear the brunt of the consolidation measures implemented so far. In Italy households with children have lost less from tax and benefit reforms than pensioner households; the reverse is true in Ireland and the UK.

“Never let a good crisis go to waste”

With France, Ireland, Italy, Spain and the UK all having planned and implemented large fiscal adjustments since the onset of the Great Recession, we might hope that policymakers would have tried to use this as an opportunity to improve the efficiency of the tax system and public spending in their countries. Or, at the very least, not to have exacerbated existing inefficiencies. Unfortunately, in many cases, the fiscal response to the crisis missed opportunities to improve the overall efficiency of the tax system. To give three examples:

  • In France, while the new corporate tax credit (which is computed on individual earnings of firms’ employees) is welcome, it would have been better to have simply reduced the relatively high level of employer social security contributions.
  • In Ireland, reforms have unnecessarily created uncertainty and distortions: there has been a succession of VAT changes (increases, reductions and then increases again), while the rate of capital gains tax has since October 2008 been increased from 20 per cent to 22 per cent, then to 25 per cent, then to 30 per cent and then to 33 per cent.
  • In the UK, given a relatively narrow VAT base, the increase in the main rate of VAT will have come at the cost of increasing distortions for both producers and consumers; the income tax schedule has also been made considerably more complicated.

With many countries still running deficits in excess of 2 per cent of national income and having a stock of government debt well above its pre-crisis level, this will not be the end of the story. Where there is a need for further spending cuts and/or tax rises, these could be an opportunity for countries to make reforms that improve the efficiency of the tax and benefit system. However, the lesson from the reforms made so far is that we perhaps ought not to be too optimistic on this front and instead may have to be content to settle for reforms that do not add to existing deficiencies.

Carl Emmerson and Gemma Tetlow, Institute for Fiscal Studies.      

This observation has been co-published with The Conversation

The Conversation

Click here for the full Fiscal Studies introduction. 

]]> Thu, 10 Dec 2015 00:00:00 +0000
<![CDATA[The options for calculating Scotland's block grant]]> The Scotland Bill, currently making its way through the Houses of Parliament, will transfer a range of tax and spending powers from Westminster to the Scottish Parliament. At the same time, an adjustment will have to be made to Scotland’s block grant funding from Westminster. Alongside things like changes to borrowing powers and fiscal institutions, these block grant adjustments (BGAs) form a key part of the new “fiscal framework” Scotland will require when these powers are transferred.

Unlike the tax and welfare powers, the fiscal framework is not part of the Scotland Bill. The report of the Smith Commission, on which the Bill is based, did not have the time to design the fiscal framework. Instead, it laid down a number of principles for its design.

The House of Lords Economic Affairs Committee, in its report “A Fracturing Union?” published today, argues that the process for determining the fiscal framework is flawed and that its design principles may not be workable and are not mutually compatible.

A new joint paper by researchers at the Institute for Fiscal Studies (IFS), the University of Stirling and the Centre for Constitutional Change, funded by the Nuffield Foundation, confirms that it is impossible to design a block grant adjustment system that satisfies the spirit of the ‘no detriment from the decision to devolve’ principle at the same time as fully achieving the ‘taxpayer fairness’ principle: at least while the Barnett Formula remains in place.

It also finds that the precise way in which the BGAs are indexed over time could mean differences of over a billion pounds a year in the Scottish Government’s budget. And it concludes by suggesting the time may now have come for a more fundamental reassessment of devolved finance – including the Barnett Formula.

This Observation article summarises some of our key findings. In what follows, we focus on tax devolution but many of the same issues arise for welfare devolution.

The options for the block grant adjustment

The Smith Commission committed to retain the Barnett Formula as the mechanism for determining Scotland’s block grant. But Scotland’s Barnett-determined block grant will clearly need to be adjusted to reflect both the new tax-raising powers and new expenditure responsibilities being devolved.

In the first year that new powers are devolved, this adjustment should be relatively straightforward, at least in principle. The initial reduction in the block grant for devolved taxes should be equivalent to the revenue forgone by the UK government.

In future years however, the process of adjusting the Barnett-derived grant becomes more complicated. The BGAs have to be indexed. If the BGAs were not indexed but fixed at the year 1 amount, then, in the face of inflation and economic growth their relative value would be eroded over time. In the case of tax BGAs, this would result in the Scottish Government gaining and the UK government losing out as time goes by. The indexation of the BGA is thus critical to achieving the spirit of the Smith Commission principle that there should be ‘no detriment [to either government] from the decision to devolve’.

Our report considers three specific options in detail.

The first approach to indexing the BGA is Indexed Deduction (ID). This indexes the change in the BGA to the percentage change in total comparable tax revenues in the rest of the UK (rUK). For example, if comparable revenues in rUK grow by 5%, the BGA also grows by 5%. Increases in income tax revenues can come about from two sources: more tax-paying workers and/or more tax per worker. Under this adjustment method, growth in either of these sources of revenue in rUK would lead to an increase in Scotland’s BGA (a fall in its block grant). This approach therefore exposes Scotland to the risk of relatively slower population growth than in rUK, assuming that increases in population will lead to higher tax revenues. On the other hand, this mechanism would allow Scotland to capture the reward of relatively faster population growth. Scotland would therefore gain from attracting and retaining more income tax payers, for instance.

The second approach is Per Capita Indexed Deduction (PCID). This indexes the BGA per capita to the percentage change in comparable rUK revenues per person. This option clearly protects the Scottish budget from the risk that its population grows relatively more slowly than the rUK’s. But equally the Scottish budget would not benefit from revenue increases that resulted from population growth. The Scottish Government would therefore lack incentives to boost growth through attracting more people to Scotland. 

The third approach is the Levels Deductions (LD). This calculates the change in the BGA as a population share of the change in comparable revenues in rUK. For example, if income tax revenues increased by £10 billion in rUK, then if Scotland’s population was 9% of rUK, Scotland’s BGA would increase by £900m. The rationale for the LD approach is that, by being based on a population share of a cash terms change in revenue, it is symmetric with the spending side of the Barnett Formula (which calculates the change to Scotland’s block grant as a population share of the cash terms change in English spending).

Comparing these options: when tax rates change in the rest of the UK

This symmetry property is useful when it comes to changes in rUK tax rates. Changes in UK tax rates for taxes that are devolved are likely to lead to a change in spending by the UK government. To the extent that this spending is likely to benefit Scottish taxpayers in some way (either because it leads to an increase in the Scottish block grant via the Barnett formula, or because it leads to an increase in ‘reserved’ spending in Scotland), the block grant to Scotland would need to be adjusted to ensure that increases in taxes in rUK do not fund higher spending in Scotland, without a corresponding increase in Scotland’s tax effort. This is the ‘taxpayer fairness’ element of the Smith Commission’s ‘no detriment’ principles.

When the additional rUK revenues are spent on services like health or education that are devolved to Scotland, Scotland gets an equivalent population share of this spending via the Barnett Formula. An increase in taxes and spending of £10 billion in rUK would, for instance, feed through into approximately £900 million more for Scotland via the Barnett Formula. Under the LD approach, Scotland’s BGA also increases by a population share of the change in UK revenues: again £900 million. So the increase in the BGA exactly offsets the increase in the underlying block grant, leaving Scotland unaffected. The ‘taxpayer fairness’ principle is satisfied.

In contrast, under the ID or PCID approaches, Scotland would gain from such a tax increase. This is because tax revenues per person are lower in Scotland than in rUK. A percentage increase in the BGA is therefore smaller than a population-share based increase. Our report shows that for a £10 billion income tax increase in rUK, equal to about 2p on each income tax rate, such gains to Scotland would amount to over £100 million a year, even though Scots were paying no more tax themselves. Or vice versa for an income tax cut. These methods therefore do not fully satisfy the ‘taxpayer fairness’ principle.

Comparing these options: underlying revenue growth

Tax revenues change not only because of policy changes though. Underlying growth in the economy and the tax base also affect revenues. Depending on the initial starting levels of revenues per person, revenue growth per person and population growth, the different options we consider can have markedly different effects on the Scottish Government’s budget.

Consider the following situation: revenues start off lower per person in Scotland, grow at the same percentage rate per person as in rUK, but the population grows less quickly than in rUK. (That is a rough approximation to what has happened in recent years).

Under the PCID approach, the BGA increases in line with the rate of growth in revenues per person, which is the same in Scotland and rUK. Hence Scottish revenues grow at the same rate as the BGA – meaning it does no better or no worse than if taxes were not devolved. If we consider the situation where revenues are growing at the same rate per person as in rUK as a benchmark against which the Scottish Government’s performance should be judged, this method therefore satisfies the spirit of the principle that there should be ‘no detriment from the decision to devolve’.

On the other hand, Scotland does lose out somewhat under the ID approach because of its lower population growth. And, under the LD approach it loses out even more, at least initially, because a given rate of growth in its revenues translates into less than a population-based share of the equivalent growth in rUK revenues (because rUK revenues started off higher per person). The LD approach therefore does not seem to satisfy the spirit of the principle that there should be ‘no detriment from the decision to devolve’: there will be detriment to Scotland under this approach, unless revenues in Scotland grow at a faster rate both per person, and in aggregate than in rUK. This might be seen as an unfair challenge for Scotland to meet.

It therefore turns out that it is impossible to design a block grant adjustment system that satisfies the spirit of the ‘no detriment from the decision to devolve’ principle at the same time as fully achieving the ‘taxpayer fairness’ principle: at least while the Barnett Formula remains in place. Some methods better satisfy the first and others the second principle.

In our paper, we estimate what would have happened if each of these approaches had been in place between 1999–00 and 2013–14 to get an idea of just how big a difference they could make.While these figures are approximate and refer to the past rather than the future, they show these differences can be substantial.

Relative to the LD method, the ID method could have resulted in the Scottish Government’s budget being around £1 billion a year higher after 14 years, with the PCID approach delivering an even bigger budget. These are quite sizeable numbers in the context of a block grant to Scotland equal to around £30 billion a year in 2013–14. 

Our analysis also shows that eventually though, if relative population decline continues, Scotland would start to do less well under the ID method than the LD method. Indeed, because the ID method never gets updated to reflect the fall in Scotland’s relative population, it can eventually imply a negative budget for Scotland if one looks far enough in the future. Clearly such an outcome would never be allowed to come to pass. But it illustrates that the ID method would not represent a sustainable long-term compromise between the PCID and LD methods that, in the short term, would be most beneficial to the Scottish Government and UK Treasury, respectively.

The difficulty of compensating for policy knock-on effects

Our paper also considers another part of the Smith Commission’s principles: the idea that one government should compensate the other if its policies have knock-on effects on the other’s revenues or spending. While this seems eminently reasonable in theory, we have significant concerns about its workability in practice. In particular, the way that individuals change their behaviour in response to tax rate changes means that the counterfactual “no change” scenario cannot be observed or easily modelled.

Calculating the knock-on effects would therefore require a series of assumptions, each subject to significant uncertainty, opening up the potential for frequent disagreement between the governments.

This therefore raises questions about the institutional arrangements for such calculations – who estimates revenue loss, and what mechanisms exist for the governments to negotiate around this – that are likely to be extremely complex to resolve. It does not seem likely that this no detriment principle can be achieved at the same time as achieving another one of the Smith Commission principles – that Scotland’s fiscal framework should be simple, transparent and as far as possible, automatic, in its operation. Indeed, if too much attention is paid to compensating for every example of knock-on effect, then the arguments and tricky negotiations that result could cause the whole system to become unworkable and unsustainable.  


Just how the block grant will be adjusted following the devolution of tax and welfare powers to Scotland is currently being negotiated by the UK Treasury and Scottish Government. These discussions are taking place behind closed doors with little information publically available about the options being considered and the effects of these options. Our paper helps fill that gap and shows that which option is chosen can have a significant effect on the Scottish Government’s spending power, and the types of risks and incentives the Scottish Government will face. It also shows that the principles set out by the Smith Commission, against which the negotiated framework will likely be judged, cannot be met in full.

For these reasons, the options available for calculating the BGAs, and other elements of the fiscal framework, should be part of the public and parliamentary debate, as much as the tax and welfare powers set out in the Scotland Bill itself have been.

Indeed it may now be time for a more fundamental reassessment of how the devolved governments are financed: including whether the Barnett Formula should be retained. Reform of Barnett may remove some of the conflicts between the Smith Commission’s principles. The Smith Commission parked these issues to one side by stating that the Barnett Formula should be retained. Making the UK’s fiscal framework sustainable for the long term may require reopening the debate. 


David Bell and David Eiser are at the University of Stirling and the Centre for Constitutional Change.

David Phillips is a senior research economist at the IFS

]]> Fri, 20 Nov 2015 00:00:00 +0000
<![CDATA[Autumn Statement 2015: the first test for the Chancellor's welfare cap]]> The July Budget painted a picture of declining borrowing over the next few years, assisted by growth in tax revenues and – in particular – further deep cuts to many areas of public spending. The latest official forecast suggested that borrowing this year would be £69.5 billion, falling to £6.4 billion by 2018–19 and then moving to a surplus of £10.0 billion in 2019–20. If achieved this would be the first overall annual surplus for the UK public finances since 2000–01 and only the ninth since Queen Elizabeth II came to the throne. However, this move into surplus is predicated on significant cuts to some areas of public spending. In preparing his Autumn Statement and Spending Review, Mr Osborne will be facing two particular challenges. The first is how to divide up the diminishing resources available for day-to-day departmental spending. The second is how to remain within his self-imposed welfare cap while also adhering to the recent House of Lords motion that requires him to reconsider the cuts to tax credits that were announced in July.

Short-term economic and fiscal outlook broadly unchanged

The good news for Mr Osborne is that the UK’s economic situation and outlook does not seem to have deteriorated since the last official forecasts were published in July. Average independent forecasts for GDP growth this year and next are identical now to what they were in July.

Despite this, figures released today by the Office for National Statistics suggest the Chancellor may be on course to slightly overshoot his forecast for borrowing this year. Borrowing over the first seven months of 2015–16 has been 10.9% lower than over the same period last year, compared to the Office for Budget Responsibility’s (OBR) forecast from July 2015 that borrowing would fall by 22.9% this year. If this trend were to continue for the rest of the year, government borrowing would overshoot the OBR’s forecast by around £11 billion.

But the picture may not be as bad as it appears at first sight. Some aspects of spending, such as that on investment, are quite lumpy and may well grow less quickly over the remainder of the year than they have done so far. However, total tax receipts do look likely to disappoint slightly this year.

As the table shows, October 2015 was a relatively bad month for receipts, with the main taxes (income tax, National Insurance Contributions, VAT and corporation tax) all performing worse than the full year forecast. Taking the first seven months of the year together, income tax, National Insurance Contributions, VAT and corporation tax have all still grown strongly relative to the full year forecast, but the slowdown in their growth over the last month means they are no longer offsetting weaker growth in other receipts. Total central government current receipts have grown by 3.0% over the last seven months, compared to the OBR’s forecast of 3.6% for the year as a whole.

Table: Growth in receipts, spending and borrowing over the year to date


% growth

£ billion


Month-on-month (outturn data)


(outturn data)

Forecast for year as a whole

(July 2015 Budget)

Forecast for 2015–16

(July 2015 Budget)

Central government receipts





Of which:





Income tax















Corporation tax










Central government spending





Of which:





Debt interest





Net social benefits















Public sector net investment





Public sector net borrowing






However, in the context of public borrowing, the size of overshoot suggested by today’s figures is relatively small. Together with the fact that the economic outlook appears little changed since July, this suggests that there are unlikely to be large revisions to the OBR’s economic and fiscal forecasts next week. Despite this, George Osborne still faces a tough challenge in making his overall spending plans add up.

Slicing the departmental cake

The July Budget suggested that George Osborne will be looking to cut departmental day-to-day spending by around 5% in real terms over the next four years. However, some large areas of spending were ‘protected’ from the start – overseas aid, defence and, in England at least, the NHS and schools. As a result, the unprotected areas (after taking account of the Barnett formula, which is usually used to determine grants to the devolved administrations) are set to see real terms cuts to their day-to-day spending averaging 27%, as we described in a recent IFS briefing note. This would bring the total cut to these budgets since 2010–11 up to 50%.

But of course the pain is unlikely to be evenly shared and from the start the Treasury asked departments to propose how they would cut 25% and 40% from their day-to-day budgets. Recent weeks have witnessed a series of announcements about early settlements for certain government departments (although without the actual settlements being published), including the Departments for Transport, Work and Pensions, Environment, Food and Rural Affairs and Energy and Climate Change. However, many of the largest departments have not yet been settled – including Education, the Home Office, Justice, and the Department for Business, Innovation and Skills – meaning that there are still important decisions to be made. Some of these departments have already seen very significant cuts to their budgets over the last five years. For example, the Ministry of Justice has already had its day-to-day budget cut by a third since 2010–11.

Fitting the welfare cap?

Whereas George Osborne always expected to have to make these tough choices about how to divide up the shrinking departmental spending pie, he probably was not expecting to have to revisit difficult decisions on welfare spending. However, the recent vote in the House of Lords (which forced George Osborne to reconsider his planned cuts to tax credits) may have resulted in just that.

At the time of the July Budget, George Osborne announced plans for £12 billion of cuts to working-age welfare spending by 2019–20 (the same total amount, albeit two years later, than committed to in the Conservative Party general election manifesto). At the same time, he lowered his so-called ‘welfare cap’ so that it exactly matched the new, lower forecast for spending on “welfare-in-scope” over the next five years. “Welfare-in-scope” covers, essentially, spending on all social security benefits and tax credits that are set by central government apart from the state pension and the most cyclical benefits.

If the OBR’s forecasts were to remain unchanged but the Chancellor had to unwind (or compensate for) some of his planned cuts to tax credits, then – without commensurate cuts elsewhere – he would breach his cap. Doing so would force him to go to Parliament for a vote to raise the cap.

There are only two ways that Mr Osborne might avoid this scenario next week. The first is for him to find some alternative cuts to other spending within the welfare cap. This comprises mainly working-age benefit spending – such as child benefit, housing benefit and disability benefits – though does also include some benefits paid to pensioners, such as pension credit and the winter fuel allowance. Such cuts would presumably be ones that were considered but rejected in favour of the tax credit cuts in July.

The second way to avoid a Commons vote is for him to hope that the OBR has – for some other reason – reduced their forecast for cash spending on welfare-in-scope. If this happens, it would give Mr Osborne a little more wiggle room under the cap. But the likely unchanged outlook for growth next year – and the fact that spending on social security benefits so far this year has run in line with the OBR’s forecast for the year as a whole (as shown by the “net social benefits” line in the table above) – suggests that any wiggle room is likely to be small. Even the very low inflation in September (which is usually used to uprate many benefits in the following April) will not have made much difference this year as zero inflation was already anticipated in the July Budget, (and the government has announced a cash freeze in working age benefits in any case).

It will be interesting to watch next week how Mr Osborne navigates these treacherous waters and avoids the obstacles he constructed for himself. As he said when he introduced the welfare cap in November 2013: “The government has a responsibility to taxpayers to control their spending on welfare; and Parliament has a responsibility to the country to hold the government to account for it.” Might this be the first time we see this principle in practice?

]]> Fri, 20 Nov 2015 00:00:00 +0000
<![CDATA[July Budget measures will strengthen work incentives overall despite tax credit cuts]]> In advance of next week’s Autumn Statement, when the Chancellor is expected to announce amendments to his planned cuts to tax credits, the Institute for Fiscal Studies has today published a comprehensive analysis of the impact of the government’s current tax and benefit plans and the National Living Wage on household incomes and financial work incentives.

Our report focuses on the effects on work incentives of changes proposed in the July Budget. It finds that both the package of tax and benefit changes, and the new ‘National Living Wage’ will, on average, strengthen incentives to move into paid work and to work more if in work.

But within this overall picture, the effect of the tax credit changes on work incentives will be different for different groups. For example, the tax credit changes will weaken incentives for lone parents to move into work but strengthen the incentive for both members of a couple to work rather than just one. And for those looking to work more, although the tax credit changes will increase by a million the number of workers who get to keep at least 40p of every extra pound earned they will also increase by half a million the numbers keeping less than 20p of each extra pound earned.

The full report is available here, a companion report funded by the Welsh Government examining the impact of the reforms in Wales is available here.

A key finding of our report is that tax and benefit changes to be introduced between now and 2019–20 will on average strengthen people’s incentives to be in paid work despite the £4.4 billion planned cuts to the tax credits received by working families. However, these cuts to the in-work support provided by tax credits do mean that the impact of the changes on work incentives is relatively modest given the size of the £12.5 billion overall cut to working age benefit spending. Generally, cuts to benefit spending would tend to strengthen work incentives as people will see bigger increases in income if they move into work or work more hours. But the devil is in the detail.

The factors driving the strengthening of work incentives include increases in the income tax personal allowance and higher rate threshold and cuts to benefits for workless families, which increase in-work incomes but reduce out-of-work incomes for many people who are not entitled to tax credits when in paid work.

The impact of the cut to in-work tax credits will be different for different groups:

  • For around 6.7 million people the cut to in-work tax credits is greater than the cut to out-of-work benefits, weakening their incentive to be in paid work. These changes will on average weaken incentives for lone parents and those in couples whose partner does not work.

  • However, reducing the amount of support given to single-earner couples through the tax credit system will strengthen the incentive for both members of a couple to work rather than just one, as they will have less support to lose if the second person starts paid work. 8.6 million people with a working partner will see their work incentives strengthened by planned changes to the (pre-universal credit) benefit system.

As well as cutting the tax credits working families receive next year, the July Budget also proposed to reduce the amount of support universal credit will give to working families. But tax credits are being cut more severely than universal credit. So relative to the tax credit system currently planned for 2019–20, universal credit will increase the amount of support given to single-earner couples, particularly those with children. This will strengthen the incentive for one member of a couple to work (rather than none) but weaken the incentive for both members of a couple to work rather than just one, as the additional support given to single-earner couples is withdrawn when the second member of the couple enters paid work.

The proposed tax and benefit changes also on average strengthen the incentive for those in paid work to increase their earnings. This predominantly arises because the cuts to tax credits mean that fewer workers will be entitled to any means-tested benefit or tax credits, so that they no longer face withdrawal of means-tested benefits if they increase their earnings. Changes to the pre-universal credit benefit and tax credit system increase:

  • the number of people who get to keep at least 60p of each additional pound they earn, by around 1.6 million,

  • the number that get to keep at least 50p, by around 1.3 million, and

  • the number that get to keep at least 40p, by around 1 million.

However, around 2.1 million workers who remain entitled to tax credits will see their incentives to increase their earnings weakened as a result of a rise in the rate at which tax credits are withdrawn as income rises. This change will increase the number of people who get to keep less than 20p of each additional pound that they earn by half a million from 1.3 million to 1.8 million. But under current plans this will only last until universal credit is introduced. By combining multiple overlapping benefit tapers into a single one, universal credit will remove the worst incentives that exist in the current system where people face withdrawal of multiple benefits over the same range of income. As a result, universal credit will reduce the number of people who get to keep less than 20p of each additional pound they earn by 1.3 million or nearly three-quarters.

Another important reform announced in the July Budget was the introduction of the ‘National Living Wage’ (NLW), a higher minimum wage for those aged 25 and over. This will also strengthen the work incentives of those who potentially benefit from its introduction. Indeed, for this group the introduction of the NLW will do more to strengthen work incentives than tax and benefit changes will. The NLW will also increase the gain to these individuals from working an additional hour. Currently, those paid below the NLW gain on average £4.22 from working an additional hour after taxes and benefit withdrawal. Tax and benefit changes will increase this to £4.49 as they mean that this group will get to keep more of each additional pound that they earn, and the NLW will increase the gain from working an additional hour further to £4.89. These figures show the extent to which taxes and benefit withdrawal reduce the gain from working an additional hour as the National Minimum Wage is £6.70 and long run level of the National Living Wage will be around £1 more than this (in today’s earnings terms).

However, the effect of the NLW is smaller for those who have the weakest incentive to increase their earnings in the first place, as much of the higher wages will be lost in lower benefit and tax credit entitlement. Furthermore, it is important to bear in mind that although the NLW may increase the amount people are willing to work, it will also make it more expensive to employ those aged 25 and over, meaning that firms may want to employ fewer of them: indeed, the Office for Budget Responsibility (OBR) estimates that the NLW will on aggregate reduce employment levels by 60,000.

]]> Thu, 19 Nov 2015 00:00:00 +0000
<![CDATA[Funding the thin blue line]]> Police forces have borne significant spending cuts arising from the government’s recent austerity programme. This article summarises the main findings of a new Briefing Note from the Institute for Fiscal Studies on government funding of the police since the turn of the millennium. This work, funded by the Economic and Social Research Council, provides important context for the ongoing debate surrounding the Home Office proposals to change the way central government funding for police services is distributed between forces. 

As an ‘unprotected’ area of public spending (unlike the NHS, schools and overseas aid), spending on police services in England and Wales fell by 14% in real terms between 2010–11 and 2014–15. This is in marked contrast to the substantial growth of police spending over the 2000s: police spending increased by 31% in real terms between 2000–01 and 2010–11. The cuts to police spending since 2010–11 have been large enough to reduce spending per person by 2014–15 roughly back to the level it was in 2002–03.

Police forces in England and Wales are financed from two main sources: grants from central government and a component added to local council tax called the police precept. The growth in spending over the 2000s was mainly due to a near-doubling of precept revenues – shown in the table below. The cuts that followed have been driven by cuts to central government grants – which fell by a fifth. There has therefore been a shift towards local financing of the police. In 2000–01, precept revenues financed a sixth of police spending. By 2014–15, this had risen to nearly a third.


There are important differences between police forces in how they fared during the period of growth in funding and the period of cuts that followed.

The forces that saw the biggest increases in police spending over the 2000s were those that increased revenues from the precept by the most. For example, North Yorkshire Police more than tripled its precept revenues between 2000–01 and 2010–11 (and saw its overall revenues increase by more than 50%), while Northumbria Police increased precept revenues by only a third (and saw overall revenues increase by just 14%).

Since 2010–11, the cuts to police spending have been driven by cuts to grants. The coalition government applied the same percentage cut to grants to all police forces each year. However, because some forces are much more reliant on grant funding than others, the cuts to spending power varied substantially across forces. Surrey Police – which received 54% of its revenues from grants in 2010–11 – saw its revenues fall by 10% between 2010–11 and 2014–15, while Northumbria Police – which received 88% of its revenues from grants in 2010–11 – saw its revenues fall by 19%.

The scatter plot below shows, for each force, how the average annual change in funding between 2010–11 and 2014–15 compared with the average annual change in funding between 2000–01 and 2010–11. This reveals that the forces that saw the biggest cuts to spending between 2010–11 and 2014–15 also typically saw the smallest increases over the 2000s. This is because the forces most reliant on grant funding in 2010–11 were also those that increased their precept revenues least over the 2000s.


Taking the period as a whole, spending on the police (excluding the counter-terrorism grant) increased on average by 7% between 2000–01 and 2014–15, but this varied from a cut of 8% for Northumbria Police to an increase of 32% in North Yorkshire.

Why some forces increased precept revenues over the 2000s by so much more than others remains an important question. It could be that the formula central government used to allocate grants to police forces did not adequately reflect their relative needs, and so forces with greater needs needed to raise more from the precept. Or it could be that people in some areas have preferences for higher spending on the police, making it easier for these forces to raise revenue locally. Understanding these different motivations is particularly crucial for the Home Office at the moment, given it is seeking to reform how central government grant funding is allocated between forces in future.

The briefing note ‘Funding the English & Welsh police service: from boom to bust?’ was funded by the Economic and Social Research Council (grant ES/L008165/1).

]]> Tue, 17 Nov 2015 00:00:00 +0000
<![CDATA[Ethnic minorities substantially more likely to go to university than their White British peers]]> All ethnic minority groups in England are now, on average, more likely to go to university than their White British peers. This is the case even amongst groups who were previously under-represented in higher education, such as those of Black Caribbean ethnic origin, a relatively recent change.

These differences also vary by socio-economic background, and in some cases are very large indeed. For example, Chinese pupils in the lowest socio-economic quintile group are, on average, more than 10 percentage points more likely to go to university than White British pupils in the highest socio-economic quintile group. By contrast, White British pupils in the lowest socio-economic quintile group have participation rates that are more than 10 percentage points lower than those observed for any other ethnic group.

These are amongst the findings of research undertaken by IFS researchers, funded by the Departments of Education and Business, Innovation and Skills (BIS), and published by BIS.

The report updates evidence on differences in higher education participation by socio-economic background, gender and ethnicity. It also explores the extent to which pupils’ performance  in national achievement tests taken at age 11, and GCSE and A-level and equivalent exams taken at ages 16 and 18, can help to explain differences in the proportion of students going on to study at university.

The research used census data linking all pupils going to school in England to all students going to university in the UK, containing over half a million pupils per cohort. It focused on those taking their GCSEs in 2007-08, who could have gone to university at age 18 in 2010-11 or age 19 in 2011-12. (It therefore predates the large increase in tuition fees which occurred in 2012.)

Differences in progression to university between individuals from different ethnic groups were particularly striking. We find that school pupils from all ethnic minority backgrounds are now, on average, significantly more likely to go to university than their White British counterparts. That is, the proportion of students from an ethnic minority background getting a place at a UK university is higher than the proportion of White British students getting a place.

This is shown in Figure 1, which plots average university participation rates of pupils from 12 ethnic groups. The light grey bars on the right show the overall participation rates by ethnic group and the black horizontal line plots the White British average. For all groups, overall participation rates exceed the White British average. Some of these differences are very large indeed. For example, Indian and Chinese pupils are, on average, more than twice as likely to go to university as their White British counterparts.

Differences in how well pupils do at school can help to explain some but not all of these gaps. For example, pupils of Black, Pakistani and Bangladeshi ethnic origin tend to perform worse, on average, in national tests and exams taken at school than their White British counterparts. Accounting for the fact that individuals from these ethnic groups have lower prior attainment than their peers therefore increases the unexplained differences in participation between ethnic minorities and White British pupils. In other words, comparing pupils with similar school attainment but from different ethnic backgrounds actually makes it more difficult to explain why ethnic minorities are so much more likely to go to university than their White British peers.

Figure 1: Percentage of pupils taking their GCSEs in 2008 who go on to university at age 18 or 19, by ethnicity and socio-economic quintile group

Figure 1: Percentage of pupils taking their GCSEs in 2008 who go on to university at age 18 or 19, by ethnicity and socio-economic quintile group

Notes: figures reproduced in Table at end of observation.

The report also considers participation at 52 of the most selective universities, or “high tariff institutions” as we call them in the report. Most ethnic minority groups are, on average, more likely to attend such institutions than their White British counterparts, but the differences are smaller than for participation among all universities, and could generally be better explained by differences in school attainment.

Even so, we still found, for example, that 34% of Chinese pupils attend one of these selective institutions, higher than the proportion of White British students who go to any university, and more than three times higher than the proportion of White British students going to a selective institution.

These results do not necessarily contradict recent evidence suggesting that ethnic minorities are less likely to receive offers from selective institutions than their equivalently qualified White British counterparts. Our research focuses on those who go to university, while evidence on offer decisions is based on UCAS applications data. If ethnic minorities are even more likely to apply to university than their White British counterparts, then it would be possible for them to be offered proportionately fewer places on average than White British students, but still go on to be relatively more likely to attend.

This research has shown that university participation rates amongst ethnic minority groups are very high on average – much higher than for their White British counterparts. Only a small proportion of this relative over-performance can be explained by differences in how well pupils perform earlier in the school system. This means that there must be other factors that are more common amongst ethnic minority families than amongst White British families which are positively associated with university participation. Moreover, we find that these other factors appear to be more important for ethnic minorities for whom English is an additional language and for those living in London. Future research could usefully explore the source of these differences.

Table 1: figures underlying Figure 1


Lowest SES quintile group


Middle SES quintile group


Highest SES quintile group


White British







Other White







Black African







Black Caribbean







Other Black



































Other Asian





















]]> Tue, 10 Nov 2015 00:00:00 +0000
<![CDATA[English schools will feel the pinch over the next five years]]> The overall funding settlement for schools will feel quite different over the next five years compared with the previous five.

Under the coalition government, school spending in England was relatively protected at a time when other areas of government saw large cuts, with many unprotected departments seeing their non-investment budgets cut by 20% or more between 2010–11 and 2014–15. Current or day-to-day spending on schools grew by 3% in real terms between 2010–11 and 2014–15. Even after allowing for the growth in pupil numbers over this period, spending per pupil still rose by 0.6% in real terms.

The picture for day-to-day spending on schools contrasts with other areas of education spending that saw larger cuts: 16–19 education spending fell by 14% in real terms over the last parliament and education capital spending fell by 34% in real-terms.

The new Conservative government has also offered schools in England considerable protection, committing to protecting day-to-day spending per pupil in cash terms over the current parliament. As shown in the figure below, this will actually mean a similar rate of growth in nominal spending as seen in the last parliament. However, increasing costs and increasing pupil numbers mean that resources per pupil are likely to fall significantly.

Our research at the Institute for Fiscal Studies (IFS), funded by the Nuffield Foundation, shows that key cost increases include the average public sector pay settlement of 1% per year announced in the Summer Budget. In addition schools will see an increased employer National Insurance Contributions bill from April 2016 as the reduced rate associated with contracting out will cease. This comes on top of the increase in employer pension contributions to the teachers’ pension scheme that came into force in April 2015.

Taking these together with pressures on other costs, we forecast that school spending per pupil is likely to fall by around 8% in real terms (based on a school specific measure of inflation) between 2014–15 and 2019–20. This is actually a less severe squeeze than looked likely at the time of the election (when we thought it would be around a 12% cut per pupil). But that only reflects the tighter 1% pay increase announced in the Budget. That will ease the pressure on schools costs, but might make recruitment and retention of teachers and other staff more difficult.

Even so, this will be the first time since the mid-1990s that school spending has fallen in real terms (when spending per pupil fell by 3.6% in real terms between 1993 and 1997).

Figure: Percentage Changes in School Spending and Cost Factors, 2010-11 to 2019-20  

Sources and Notes: Figures for nominal school spending taken from; Pupil Numbers taken from; methodology for calculating school costs for 2014–15 to 2019–20 can be found in; Growth in paybill per head figures are not available before 2014–15 and we therefore assume school costs follow the growth in the GDP deflator of 7.4% between 2010–11 and 2014–15, which is likely to slightly over-state cost increases for schools given that average weekly earnings for public sector workers only increased by 5.5% between 2010–11 and 2014–15.

Challenges ahead

A significant challenge on the teacher workforce over the next five years will be recruiting the required number of teachers, and of sufficient quality and motivation, at a time of continued public pay restraint and rising pupil numbers.

The pupil population is currently expected to rise by 450,000 from 6.45 million in 2016 to 6.9 million in 2020. If there was a desire to keep the pupil:teacher ratio constant, the number of teachers would need to increase by 30,000 (from 450,000 today to 480,000 by 2020). Although not without precedent (a similar increase occurred in the early 2000s), it could prove difficult to increase teacher numbers at a time when public sector pay seems likely to fall relative to that in the private sector. Alternatively, schools and policymakers could allow the pupil:teacher ratio to rise, implying potentially larger class sizes.

The government has also signalled its intention to reform the school funding system to ensure (sensibly) that areas with similar populations receive the same level of funding per pupil. Although it made some tweaks at the end of the last parliament, this only went some way to address the problem that similar areas can receive quite different levels of funding per pupil. However, substantial reform will be particularly difficult for policymakers to implement when overall school spending per pupil will be frozen in cash-terms. Delivering a cash-terms increase to some schools or local authorities as a part of a set of reforms would require cash-terms cuts to some other groups of schools or local authorities. Schools do have some experience of receiving overall cash-terms cuts when pupil numbers fall. However, delivering cash-terms cuts in funding per pupil could be a lot more challenging.


Schools in England experienced a relatively benign scenario under the last parliament at least relative to many other public services. But this is now set to change. We forecast they will see an overall cut of 8% per pupil in real terms over the next five years. This will be a new experience for many in the school sector – the last real-terms cuts were in the mid-1990s – and will make issues around teacher recruitment and reform of the school funding system more difficult.

However, it is important to remember that schools are still relatively protected compared with many other areas of government and other areas of education spending. Further education and sixth form spending fell by 14% in real terms under the last parliament and could experience even larger cuts under this parliament. Unprotected government departments are currently expected to see their spending fall by 27% in real-terms between 2015–16 and 2019–20.

]]> Wed, 21 Oct 2015 00:00:00 +0000
<![CDATA[The UK coalition government’s record, and challenges for the future]]> This Observation comes from an article published in a special edition of Fiscal Studies, based on IFS analysis for the 2015 general election. The articles in this special issue are currently available free of charge in the Wiley online library. With this piece of work, we conclude the analysis on our general election website.

Looking back

By any standards, the fiscal task facing the Conservative–Liberal-Democrat coalition government when it assumed office in May 2010 was a remarkable one. The UK had just experienced its deepest recession in almost a century and the deficit was at its highest point since the Second World War, at just over 10 per cent of national income. The Conservative part of the coalition had been elected on a platform of quite aggressive deficit reduction, which would require historically unprecedented spending cuts alongside at least some tax rises. In the event, the coalition agreement largely accepted the Conservative plans to reach structural current budget balance by the end of the parliament.

As things turned out, poorer-than-expected economic growth put paid to that ambition. Between 2010 and 2012, the outlook for economic growth and tax revenues deteriorated. The government did not respond to the bigger-than-originally-predicted deficit by imposing greater spending cuts or tax increases by the end of the parliament; rather, it promised further fiscal consolidation after the 2015 election.[1] The result was that, even after the tightest period for public spending in 60 years, by the end of 2014–15 the deficit still stood at about 5 per cent of national income.

The deficit reduction achieved over the parliament was still sizeable, however, and required some very substantial spending cuts as well as some net tax increases. Figure 1 shows how this deficit reduction was achieved, in terms of changes to tax and public spending as shares of national income, along with the new Conservative government’s plans for these through to 2019–20.[2] The large majority of the fiscal consolidation occurred through reducing spending as a proportion of national income, from nearly 46 per cent in 2009–10 to just over 40 per cent by 2014–15. Under the Conservatives’ plans, reductions in public spending as a share of national income will continue until 2018–19, when spending is forecast to reach around 36 per cent of national income – around the share it was in 1999−2000. In the absence of the fiscal consolidation, public spending would have remained around 45 per cent of national income.

Figure 1: Taxes and spending as a share of national income, with and without fiscal consolidation

Taxes and spending as a share of national income, with and without fiscal consolidation

Note: ‘Without consolidation’ lines ignore the impact of all fiscal policy measures that have been announced in fiscal events since March 2008. Dashed lines indicate the plans of the incoming Conservative government in 2015 based on its manifesto commitments.

Source: Authors’ calculations based on Office for Budget Responsibility (2015a), Conservatives (2015) and all fiscal events between March 2008 and March 2015.

The big fall in spending as a proportion of national income shown in Figure 1 corresponds to a cut in real, price-adjusted, terms to total public spending of ‘only’ around 3 per cent between 2010–11 and 2014–15. Of course, what happened was much more complex than simply cutting all spending by similar amounts. Departmental spending – that is, spending on public services such as health, education, defence and so on – fell by 9 per cent, while non-departmental spending – such as debt interest payments, spending on social security and spending by local authorities financed by locally-raised revenues – actually rose by 4 per cent over the period.

Nor were the cuts to departmental spending shared equally across service areas. Notably, spending on health, schools and overseas aid were protected, increasing by 5 per cent, 3 per cent and 31 per cent respectively in real terms between 2010−11 and 2014–15. Together these areas amounted to nearly 40 per cent of all departmental spending in 2010–11, and so their protection increased the average cut to other areas of departmental spending to 19 per cent rather than 9 per cent.[3]

Two papers in this Special Issue focus on what happened to specific elements of departmental spending. Luke Sibieta looks at changes to school spending. While spending on schools may not have changed dramatically over this period – the 3 per cent real increase slightly more than kept pace with rising pupil numbers over the period – there were important reforms to the school funding system. Sibieta describes these reforms and explores the implications for the distribution of school funding, showing in particular how funding became more focused on schools with the poorest intakes.

Meanwhile, David Innes and Gemma Tetlow look at changes to local government spending. They show how local authorities to some extent mirrored the behaviour of central government in protecting some areas of spending at the expense of others. In particular, spending on social care – the biggest element of spending under local control – was cut significantly less than the average. The authors also show how, by contrast with what happened to school funding, it was local authorities with the greatest needs, which had been most reliant on central government funding, that experienced the sharpest cuts in spending.

At the same time as these significant cuts in public service spending were introduced, some substantial reforms were implemented – for example, in the funding of schools and hospitals – or planned – as in the (delayed) introduction of universal credit, a plan to roll six out-of-work benefits into one single payment. When eventually implemented, the introduction of universal credit will represent one of the biggest reforms to the benefits system in many decades.[4]

Figure 1 shows that, in contrast to the substantial reductions in public spending that occurred between 2010–11 and 2014–15, taxes as a share of national income remained relatively stable over the last parliament, hovering at around 36 per cent of national income. While the direct impact of tax changes implemented by the coalition reduced borrowing only slightly, by £16.4 billion in 2015–16, this was the net effect of £64.3 billion of tax rises and £48.0 billion of tax cuts. So even if the net effect of tax changes was relatively modest, there was considerable policy activism. That policy activism, however, led to relatively little in the way of substantive reform other than in the taxation of profits through corporation tax and in the taxation of savings.

As Helen Miller and Thomas Pope in this Special Issue show in some detail, the coalition enacted a series of changes to corporation tax, with the explicit, and fulfilled, aim of increasing the competitiveness of the UK corporate tax system. The net cost of the package of onshore measures is almost £8 billion in 2015–16, equivalent to 16.5 per cent of real pre-crisis (2007–08) onshore corporation tax receipts.[5] Only increases in the income tax personal allowance – the amount of income that can be taken before an individual becomes liable for any income tax – represent a bigger giveaway in revenue terms than cutting the main corporation tax rate by 8 percentage points from 28 per cent to 20 per cent.[6]

Stuart Adam and Barra Roantree explore the coalition government’s policies on non-corporate taxes and argue that only in the taxation of pensions and savings were there important structural changes. Other changes to the tax system have been fiscally and distributionally significant, but none has resulted in real structural change or improvement, with most focusing on changing rates and thresholds.

These changes to the tax system were implemented alongside cuts to social security benefits. Hood and Phillips (2015) show that reforms implemented by the coalition will result in benefit spending in 2015–16 being £16.7 billion less than it would have been in the absence of any policy change. However, different groups have been differentially affected by these changes, with increases for pensioners more than compensated for by reductions in the generosity of working-age benefits.

Stuart Adam, James Browne and William Elming in this Special Issue provide a detailed analysis of the distributional consequences of the tax and benefit policies implemented by the coalition. Broadly speaking, they show that the overall effect of these changes was to take money from those with the highest incomes, and those dependent on benefits in the bottom half of the distribution, whilst largely protecting those in the middle and upper-middle parts of the income distribution.

Changes in the labour market were, in fact, more important determinants of the overall path of living standards for most than were tax and benefit changes. Cribb and Joyce (2015) show that real wages fell significantly after 2009, and Cribb, Hood and Joyce (2015) estimate that real median household incomes in 2014–15 were at about their 2007–08 (pre-crisis) level and still something like 2 per cent below their 2009–10 peak. This represents an extremely unusual period historically, with wages and incomes recovering much more slowly than is usual coming out of a recession.

Looking forward

The newly-elected Conservative government has entered office with a deficit of 5 per cent of national income and a commitment to substantial additional spending cuts. If implemented, they will take public spending as a fraction of national income back to the level it was at the end of the 1990s, and close to its lowest level since the Second World War.

Without doubt, implementing those additional cuts will be challenging. The Conservative government has pledged a £12 billion, or roughly 10 per cent, reduction in spending on (non-pension) social security benefits. If cut to this extent, spending on social security (excluding pensions) will fall to its lowest level as a share of national income since 1990. Even then, departmental spending would need to be cut by a further 7 per cent between 2014–15 and 2018–19 for the deficit to be eliminated as planned. Given the Conservatives’ pledges to increase health spending, protect school spending and increase overseas aid spending, other ‘unprotected’ areas of spending are looking at further cuts of around 15 per cent over this period. This would result in real-terms spending on ‘unprotected’ departments having fallen by around a third on average between 2010–11 and 2018–19.

It is this concentration of spending cuts in particular areas, rather than the overall level of cuts to total public spending, which is most important. The composition of the state is changing, with public spending ever more focused on providing health services and pensions.[7]

Looking further forward, the continued ageing of the UK population will put increased pressures on spending. On rather conservative assumptions about the likely path for health spending, the Office for Budget Responsibility (OBR)’s central projection is that spending on state pensions, health and social care will rise by 5 per cent of national income between 2018–19 and 2053–54.[8] This necessarily implies more taxes, more borrowing or less spending elsewhere to the tune of 5 per cent of national income each year. However, borrowing an additional 5 per cent of national income each year is not really an option, as the OBR illustrates that this would push public sector net debt onto an unsustainable increasing trajectory.

The future challenges are not just about reducing spending or increasing taxes though, but also about reforming public services and the tax and benefit system and making choices about the distributional consequences of any changes. For example, the National Health Service faces increasing demand from the growing and ageing population, and cost pressures from wages and high-cost drugs. The service therefore faces a considerable productivity challenge in order to ensure that service levels or quality do not fall, despite the real increase in funding that has been proposed over the next parliament.[9] Other challenges include issues ranging from the allocation of resources and use of consistent formulae, as with school and local authority spending,[10] to the extent and design of private co-payments in areas such as higher education and social care. These decisions all have distributional consequences. For example, since 2010, changes to school funding have benefited schools with poorer intakes, whilst changes to local authority funding have hit poorer areas harder. Changes in higher education financing have been rather progressive.[11]

Distributional issues are much more salient when it comes to social security spending. If the new Conservative government’s pledged £12 billion of cuts are delivered, they will almost certainly hit those on low incomes – most benefits (other than pensions) are means tested and are only received by the relatively poor. Structural reforms, such as the introduction of universal credit, might improve efficiency and will have more complex distributional effects. Even leaving aside the implementation of these planned reforms, numerous challenges still remain. There are some indexation arrangements in place that make little sense in the long run. For example, the continuation of the ‘triple lock’ for pensions – whereby the basic state pension is increased by the highest of CPI inflation, earnings growth and 2.5 per cent – will be increasingly costly and must surely be reviewed at some point, while the indexation of housing benefit means that relative benefit levels across the country in future will depend on relative rent levels back in 2012. There are some existing policies that are proving difficult to implement, such as the new (tighter) eligibility criteria for disability benefits. And there are some benefits on which spending continues to rise, despite cuts to generosity (for example, housing benefit, where increased spending is being driven by the growth of the private rental sector and the growth in private rents).

Between 2010 and 2015, cuts in welfare spending were accompanied by some very sharp tax increases for the richest, but also by a package of tax measures that ensured that those in the middle to upper-middle parts of the income distribution were protected. On the whole, though, there was little sign of coherent tax reform. The biggest reforms affected the taxation of pensions and savings, but even here many of the pension tax changes reduced the coherence of the tax system.

Going forward, the Conservative government is committed to reducing the coherence of pension taxation even further by tapering away tax relief for higher earners, thereby introducing a very high effective marginal income tax rate for some. It is also, bizarrely, committed to legislating to prevent itself from raising rates of income tax, National Insurance and VAT as well as to introducing tax breaks for owner-occupied housing into the inheritance tax system.

The need for more coherent tax reform is becoming urgent, and none of these proposals moves in an appropriate direction. A number of recent changes have also introduced anomalies into the structure of income tax[12] and, as Adam and Roantree in this Special Issue note, there are now several important thresholds in the income tax system that are fixed in nominal terms. As for the big structural problems in the tax system, no attempt has been made to integrate income tax with National Insurance contributions, with the latter acting simply as a second tax on earnings. The system of housing taxation is becoming increasingly inefficient and damaging. We continue to have one of the narrowest VAT bases in the OECD. The scope for improvement is considerable. As Adam and Roantree conclude:

Having failed to implement substantive structural reform, the coalition has left its successor with numerous challenges to address the long-standing weaknesses in the tax system identified by [the Mirrlees Review[13]] – some of which (such as increasingly unsustainable council tax valuations and ill-targeted fuel duties) are becoming ever more pressing.


[1]More detail on the evolution of the government’s public finance plans over this period is provided by Emmerson and Tetlow (2015).

[2]These plans are taken from the Conservative Party manifesto (Conservatives, 2015), which is the latest available information at the time of writing.

[3]Further detail on the extent of differences between departments in the scale of the cuts they suffered is provided by Crawford and Keynes (2015).

[4]See Hood and Oakley (2014) for a description of universal credit.

[5]‘Onshore’ revenues are essentially those not associated with the taxation of North Sea oil and gas.

[6]The personal allowance has been increased to £10,600 in 2015–16, £2,835 higher than the £7,765 it would have been if the 2010–11 allowance had simply been uprated in line with RPI inflation (used by default for uprating at the time).

[7]As noted by Crawford and Johnson (2011) and Crawford and Keynes (2015).

[8]Office for Budget Responsibility, 2015b.

[9]Crawford and Stoye, 2015.

[10]As highlighted respectively by Sibieta and by Innes and Tetlow in this Special Issue.

[11]See Crawford and Jin (2014) for a discussion.

[12]Discussed in more detail in Johnson (2014).

[13]Mirrlees et al., 2011.


Conservatives (2015), The Conservative Party Manifesto 2015.

Crawford, C. and Jin, W. (2014), Payback Time? Student Debt and Loan Repayments: What Will the 2012 Reforms Mean for Graduates?, Report no. R93, London: Institute for Fiscal Studies.

Crawford, R. and Johnson, P. (2011), The changing composition of public spending, Institute for Fiscal Studies, Briefing Note no. BN119.

— and Keynes, S. (2015), Options for further departmental spending cuts, in C. Emmerson, P. Johnson and R. Joyce (eds), The IFS Green Budget: February 2015, London: Institute for Fiscal Studies.

— and Stoye, G. (2015), ‘The outlook for public spending on the National Health Service’, The Lancet, vol. 385, pp. 1155–6.

Cribb, J., Hood, A. and Joyce, R. (2015), Living standards: recent trends and future challenges, Institute for Fiscal Studies, Briefing Note no. BN165 .

— and Joyce, R. (2015), Earnings since the recession, in C. Emmerson, P. Johnson and R. Joyce (eds), The IFS Green Budget: February 2015, London: Institute for Fiscal Studies.

Emmerson, C. and Tetlow, G. (2015), Public finances under the coalition, in C. Emmerson, P. Johnson and R. Joyce (eds), The IFS Green Budget: February 2015, London: Institute for Fiscal Studies.

Hood, A. and Oakley, L. (2014), A survey of the GB benefit system, Institute for Fiscal Studies, Briefing Note no. BN130.

— and Phillips, D. (2015), Benefit spending and reforms: the coalition government’s record, Institute for Fiscal Studies, Briefing Note no. BN160.

Johnson, P. (2014), ‘Tax without design: recent developments in UK tax policy’, Fiscal Studies, vol. 35, pp. 243–73.

Mirrlees, J., Adam, S., Besley, T., Blundell, R., Bond, S., Chote, R., Gammie, M., Johnson, P., Myles, G. and Poterba, J. (2011), Tax by Design: The Mirrlees Review, Oxford: Oxford University Press for the Institute for Fiscal Studies.

]]> Tue, 15 Sep 2015 00:00:00 +0000
<![CDATA[Getting off the rollercoaster]]> The March 2015 Budget presented a set of figures for tax and spending over the next five years that, while technically reflecting government policy at the time, were inconsistent with the intentions that each of the coalition partners presented in their manifestos. In particular, neither party was committed to the profile for spending on public services, which OBR chief Robert Chote described at the time as a ‘rollercoaster’. It was, therefore, to be expected that the first Budget of the new parliament would contain significant measures in order to bring official forecasts closer to the Conservatives' manifesto commitments. But the Budget also contained policies not mentioned in the manifesto – in particular new tax rises and less deep cuts to spending over the next three years – that further affect the timing and composition of the remaining planned fiscal consolidation.

The March Budget set out a plan to achieve an overall budget surplus by 2018–19, through a nine-year fiscal consolidation package totalling 10.0% of national income (or £185 billion). At that time, the UK was just at the end of the fifth year of the planned consolidation, with four years remaining.

The new plan set out in last week’s Budget is for a slightly longer, slightly slower, but ultimately larger fiscal consolidation – lasting for 10 years and totalling 10.3% of national income (£194 billion) from when it first started in April 2010.

This means that the government is not now expected to achieve an overall budget surplus until 2019–20 (i.e. one year later than intended just four months ago in March). (The size, timing and composition of this consolidation plan have changed gradually over the last five years – this evolution is described in Chapter 1 of this year’s IFS Green Budget.)

The Budget also somewhat changed the composition of the consolidation – shifting more towards reliance on tax rises and cuts to social security spending, and away from reliance on cuts to public service spending.

The plan set out in March entailed 11% of the consolidation coming from each of net tax rises and social security cuts, whereas the new plan has 14% of the tightening coming from net tax rises and 17% from cuts to social security spending (as shown in the Figure below). With little change to investment spending plans this leaves the share of the consolidation coming from cuts to day-to-day spending on public services falling from 56% to 45%.

By the end of this financial year, two-thirds of the consolidation is set to have been done. Of the final third, 88% is set to come from spending cuts and 12% from tax increases. This is partly accounted for by the planned abolition of contracting out for defined benefit pension schemes and partly by last week’s announced net tax rise. However, while the former is expected to boost National Insurance contributions by £5 billion in 2016–17, more than £3 billion of this comes directly from public sector employers and so really reflects an additional pressure on public service spending.

The net effect of the slower pace of the consolidation and the greater reliance on tax rises and social security spending cuts is that the cuts required to public service spending over the next three years will be smaller than implied by the March Budget – and even somewhat smaller than implied by the Conservative party manifesto. However, most of this additional potential spending was committed by the Budget to the Ministry of Defence, meaning that the prospects for other unprotected departments are little improved.


Note: This Figure updates the numbers presented in Figure 1.6 of the 2015 IFS Green Budget to include the policy announcements made in the March 2015 and July 2015 Budgets. The Green Budget contains details of the methodology and sources used to construct this figure. The data underlying this figure can be downloaded here (xls)

]]> Wed, 15 Jul 2015 00:00:00 +0000
<![CDATA[Time for tax reform]]> The 8 July Budget may prove to be George Osborne’s best chance to bring in some much-needed reforms to our creaking and increasingly incoherent tax system. This Observation suggests some important directions for reform and calls for an improvement in the way policy is made. If this is to be a Budget for productivity, then both a better, and a more predictable, tax system should be an important part of it.

Unfortunately, the Conservative manifesto does not augur well for coherent reform. We have written elsewhere about some of the problems.

Manifesto commitments

The proposal to reduce annual limits on tax-relieved pension contributions for higher earners makes little sense, and may do more to discourage earning than the 50% income tax rate Mr Osborne says he disliked so much. The commitment to offer additional inheritance tax relief for owner-occupied housing will serve further to damage and distort an already broken system of housing taxation. It will do even more to lock older home owners into possibly inappropriate properties. Costly promises to increase the income tax personal allowance even further are supposed to be paid for by unspecified and highly uncertain anti-tax-avoidance and -evasion measures.

The Chancellor is further constrained by his promise to legislate to prevent himself from raising the headline rates of income tax, National Insurance Contributions (NICs) and VAT. Between them, these taxes account for more than two thirds of total revenues. In the case of income tax though, this is probably not much change to normal practice. Other than the introduction of the 50p rate in 2010, no headline rate of income tax has been increased since 1975. And for income tax and NICs, the pledge relates only to the main rates, not thresholds or the base on which they are charged, so there remains some room for manoeuvre.

Getting beyond these self-imposed political handcuffs, though, what ought the Chancellor to be thinking about? There are two sorts of things it would be good to see. The first is a commitment to a better and more coherent policymaking process, alongside recognition of some of the challenges that the tax system will need to address. The second is a set of specific changes to improve the current system.

Needed: A long-term plan

Our process for making overall tax policy is, and has been for a very long time, broken. No strategy is laid out by government. No long-term direction is signalled. Parliamentary scrutiny is inadequate. Businesses and individuals have little idea from one year to the next what will happen to their tax liabilities. Reduced rates of income tax are introduced and abolished. Limits on tax-relieved pension saving are cut and cut and cut again. A diverted profits tax is introduced before it has been fully thought through. The bank levy is increased every six months. Inflation indexation of rates of fuel duties is postponed, postponed and cancelled time and again. Chancellors cannot restrain themselves from layering complexity upon incoherence. 

So perhaps the most important thing Mr Osborne could do next week is simply to commit to telling us broadly what he wants to do with the tax system. One route to long-term reform is set out in the Mirrlees Review published by the IFS in 2011. There are other routes to reform. The Chancellor should lay out his own strategy.

He should tell us how he means to address the growing challenges created by increased dependence on a few very well off taxpayers, by international mobility of people and of profits, and by the growth of e-commerce. And he should give us some sense of where he wants to take the taxation of pensions, the taxation of profits, the taxation of earnings, the taxation of housing and the taxation of inheritances.

These issues are too important to be left to piecemeal Budget announcements and the detailed discussion of finance bills. Choices about tax policy have big effects on economic activity and inequality. If this is to be Mr Osborne’s Budget for productivity then one of the most helpful things he could do with the tax system is reduce the amount of uncertainty about the future.

As for reforms, here are just a few of the things that both ought to happen and have not been explicitly ruled out by manifesto commitments:

  • Reintroduce a coherent system of inflation indexation into the tax system. Increasing indirect taxes in line with the retail price index and most direct tax thresholds in line with the consumer price index erodes confidence in the honesty of policymaking. To have introduced numerous elements into the income tax system which are fixed in cash terms – the £150,000 point at which the 45% tax rate is levied, the £100,000 point at which the personal allowance is tapered away, the £50,000 point at which child benefit starts to be withdrawn – is indefensible.

  • Address the absurdity of having a 60% rate of income tax on incomes between £100,000 and £121,200.

  • Set us on a course towards the real integration of income tax and NICs. The latter serves purely as an additional tax on earnings and exists only to attempt to fool the electorate into believing that they pay less tax than they do and into believing there is a meaningful link between what each individual pays in and what they get out. There is not.

  • Put in place regular and frequent revaluation of properties for council tax such that tax liabilities are based on current values, not on the relative value of properties nearly a quarter of a century ago. At the same time reform the structure of council tax such that liabilities are proportional to property values rather than much less than proportional to property value.

  • Reduce stamp duty land tax. The current system acts as a substantial barrier to mobility.

  • Remove some of the badly targeted and expensive allowances and exemptions in both inheritance tax and capital gains tax which undermine the credibility of both taxes by allowing the very wealthy and well advised to pay less tax than the merely well off.

  • Recognise that it is perfectly sensible to allow pension saving out of pre-tax income if tax is to be paid on withdrawal. But address the real subsidies to pension saving which come through the fact the employer contributions are entirely exempt from NICs and a lump sum of up to £250,000 can be withdrawn entirely free of income tax.

  • Reform fuel duties. In the short run this might just mean having them adjust monthly and automatically in response to inflation. In the long run it might well mean replacing them with another tax altogether, ideally a national system of congestion charging, because the main harm caused by motoring is congestion and because the government is committed to reducing the use of road fuels massively in order to meet climate change targets.

There are of course many more changes, both big and small, that would improve our system of taxation. With £33 of every £100 generated in the UK taken by the government in tax, making improvements can really affect the welfare of its citizens. Next week’s Budget could be the best chance this new government will have to show that it understands that. It could make a start on a process of reform which could have substantial long-term pay-offs. 

]]> Thu, 02 Jul 2015 00:00:00 +0000
<![CDATA[Weak productivity growth is not confined to a few sectors of the economy]]> Productivity is currently the most talked about topic in town, and for good reason. At the end of 2014, UK productivity remained below its pre-recession level and 16% below where it would have been had the pre-recession trend continued. Looking forward, it is only productivity growth that is likely to spur increases in real wage growth and living standards. Alongside the upcoming budget, George Osborne will set out a plan for how to boost productivity.

This Observation aims to provide some context for current discussions by setting out what the most recent data shows about the trajectory of productivity across different sectors of the economy. 

Productivity trends by industry

Figure 1 shows how total productivity – output per hour worked – fell at the start of the recession and recovered up to 2011. A similar pattern was observed in both manufacturing and service industries. Across the economy, productivity actually fell through 2011. Since then, manufacturing industries have regained some of their lost ground, but service sector productivity has been stagnant over the past three years.  A large part of what explains the trend in the production sector is the ongoing decline in North Sea oil production, which accounts for around 1.4% of the economy and is likely to see a permanently reduced rate of productivity growth. While there are differences between sectors, it is clear that none is close to where they would have been had the pre-recession trend continued.

The service sector accounts for close to 80% of output and is therefore the most important driver of economy-wide productivity growth. Within the stagnant productivity performance of the sector as a whole there is considerable variation, with some sub-sectors quickly returning to pre-recession growth levels and others remaining well below pre-recession levels.  

The relatively good news stories are administrative and support services, wholesale and retail and professional services (Panel A).  These sectors, which together account for approximately 25% of output, saw strong productivity growth before the recession and have recovered to their pre-recession levels. Administrative and support services and wholesale and retail trade are now seeing productivity growth rates above the average pre-recession trend, while professional services has seen little growth in productivity since 2011.

Accommodation and food services and ICT services were sectors with productivity growth before and immediately following the recession (Panel B). Yet in the last two years productivity has been in decline. These sectors account for around 9% of output.

Transport and storage and finance and insurance together make up around 12% of output and are notable as examples where productivity is still substantially below the pre-recession level and has shown little sign of recovery. It is often highlighted that the financial industry was important for UK productivity growth before the recession and that going forward productivity growth may be permanently reduced. However, as these figures show, weak productivity growth is a feature of much of the service sector. Even if productivity in financial services had continued on its very strong pre-recession trend, overall productivity levels would still only be 2.3% above their 2008 level, and still 11% below where it would have been had the overall pre-recession trend continued. The slowdown in productivity growth in financial services is therefore only part of the explanation for recent trends.

Finally, there are some industries that did not see growth either before or after the recession. These include government services (19% of output). As such, the weak performance in these industries does not explain why overall service sector productivity growth is so much lower than before the crisis.


Productivity growth has been weak in almost all sectors of the economy, and negative in some. The lack of productivity growth in the finance sector has been important, but cannot explain the majority of the recent weakness.

Much has been written about the various explanations for the lack of productivity growth, which include low investment in new capital, an impaired allocation of resources, higher employment as a result of weak wages and possibly some measurement error. When the Chancellor sets out his productivity plan, the focus is likely to be on those areas of public policy that we can be confident matter for productivity, such as investment in skills, science and infrastructure. Such investments could provide a welcome boost to productivity in the medium term but are unlikely to provide immediate fixes for current productivity. Finding ways to boost productivity today will be much harder. 

]]> Fri, 26 Jun 2015 00:00:00 +0000
<![CDATA[Median income inched towards pre-crisis level in 2013–14]]> The Department for Work and Pensions (DWP) has today published its annual statistics on the distribution of household income in the UK. The latest release covers the years up to and including 2013–14.

On the 16th July, IFS researchers will publish a detailed report, funded by the Joseph Rowntree Foundation, analysing what these data tell us about recent changes in living standards, inequality and poverty. In this Observation article, we briefly outline some of the key information contained in the DWP publication.

One difficulty in summarising the headline figures is that the official statistics continue to use the discredited Retail Price Index (RPI) to adjust for inflation. Since the RPI is known to significantly and systematically overstate the rate of inflation due to defects in its formula, the official figures are based on a methodology which makes changes in living standards look worse than they really are.

The use of the RPI in these figures is sensibly under review. Fortunately, the report also contains an annex with figures adjusting for inflation using the more appropriate RPIJ measure. In what follows, we discuss the RPIJ-adjusted figures rather than the more prominent (but less appropriate) headline figures.

Average incomes

After accounting for RPIJ inflation, median (middle) income grew by just under 1% in 2013­–14, following a similarly small rise in 2012–13. (The HBAI headline figures, based on RPI inflation, suggest that median income was unchanged in 2013–14.)

This represents a slow recovery in average incomes, which follows the sharp decline between 2009–10 and 2011–12 when workers’ real earnings fell rapidly. It did mean that real median income had crept back to within about 1% of its pre-recession (2007–08) level, though it was still almost 3% below its 2009–10 peak (household incomes actually continued to rise in real terms during the recession, partly because of temporary stimulus measures such as the temporary VAT cut).

Mean income grew by almost 3% in 2013–14, rather faster than the median. However, this is highly likely to reflect increases right at the top of the income distribution which were, at least in part, somewhat artificial. The cut to the top rate of income tax in April 2013 from 50% to 45% not only had the direct effect of increasing the incomes of the richest, but will also have resulted in significant amounts of income being taken in 2013–14 rather than in 2012–13 – in other words, a distortion to the timing of income. As a result the measured change in mean income will overstate the growth in underlying living standards.


In 2013–14 incomes grew at a similar rate across almost all of the income distribution, resulting in little change in income inequality. An exception seems to have been at the very top of the distribution – which as discussed above is likely to have been artificially distorted – driving a small but statistically insignificant 1ppt rise in the oft-cited Gini coefficient measure of inequality.

At 0.34, the Gini in 2013–14 remains at around the same level as in the early 1990s, but lower than before the Great Recession (the Gini was 0.36 in 2007–08). This is largely explained by the fact that while real earnings fell sharply between 2009–10 and 2011–12, benefit incomes were more stable. Since poorer households get a greater share of their income from benefits, their incomes have risen relative to higher-income households (who tend to get most of their income from earnings).

Once you take account of falling mortgage payments, however, overall inequality fell less than those numbers suggest since it is largely better off households who benefited from this reduced cost. This is illustrated in the after housing costs figures in today’s release, which show smaller falls in inequality since the recession.

Income Poverty

The latest data show little or no change in poverty rates, for the population as a whole and for the major demographic groups (pensioners, working-age adults and children).  

There was a small and statistically insignificant fall in absolute income poverty (with income measured before housing costs (BHC) and a poverty line fixed over time at 60% of 2010–11 median income), from 10.1 million individuals in 2012–13 to 9.8 million in 2013–14 (16% of the population). This leaves it at a similar level to that seen in 2009–10 and 2010–11, and below levels seen prior to the recession during which income poverty actually fell – particularly for pensioners and children – as some real benefit rates increased.

Using a relative measure of poverty, which instead uses a poverty line of 60% of median income in each year (and hence measures how low-income households fare not in absolute terms, but relative to middle-income households), the figures show 9.6 million (15% of) individuals were in poverty in 2013–14 with incomes measured BHC. This is essentially unchanged from the previous year and down from 18% in 2007–08, reflecting the fact that median income (and hence the relative poverty line) is still a little below its 2007–08 level. The rate of relative child poverty was also unchanged in 2013–14 and remained below its pre-recession level.

The lack of a measured increase in poverty in 2013–14 may come as a surprise. IFS researchers had estimated that known changes to tax and benefit policy (and in particular a number of substantial working-age benefit cuts), combined with the labour market trends recorded by the Labour Force Survey (LFS), would be likely to result in an increase in income poverty in 2013–14. The DWP document mentions employment increases (which were larger in the data underlying today’s statistics than in the LFS) and direct tax cuts as explanatory factors, and in our detailed report next month we will analyse the underlying data to assess the extent to which these do explain the changes observed, and the extent to which they are consistent with other data sources.

The most important point, though, as we emphasise every year, is that limited weight should be placed on a single year of data, which is subject to some margin of error, based as it is on a sample of the population. This margin of error means that a stable poverty rate in the data is perfectly consistent with a small increase or decrease in reality. It is broader trends measured over a number of years that tend to be the most reliable and upon which most weight should be placed.

As we argued last year, when looking at recent changes it is better to measure poverty after housing costs (AHC) due to the variation in housing cost trends across the income distribution. Unfortunately, the official document does not attempt to use a non-RPI based inflation measure to uprate the AHC absolute poverty line (i.e. the real line with which after-housing-cost incomes are compared to assess whether households are in poverty). But the change in AHC absolute poverty in 2013–14 is likely to have been very similar to the BHC change. However, looking at the period since 2007–08 (just before the recession), we know that AHC poverty trends have looked less favourable, because lower-income groups have not benefitted to the same extent from falls in mortgage interest rates.

]]> Thu, 25 Jun 2015 00:00:00 +0000
<![CDATA[How much is too much borrowing?]]> This week has seen various statements by public figures about borrowing and debt. George Osborne announced yesterday that the government will legislate to require the UK government to run a budget surplus ‘in normal times’. Meanwhile, the Scottish National Party have today tabled an amendment to the Scotland Bill to open up the possibility of full fiscal autonomy for Scotland, arguing that – even if this meant running a deficit in Scotland – this would be possible because ‘the UK has been in deficit in 43 of the last 50 years’.

So what level of borrowing can or should the UK or Scotland have in the longer-run? And what do the statements this week really mean?

Unfortunately, economic theory does not provide an answer to what the right level of borrowing for a country is. There are pros and cons of both higher and lower borrowing.

Most obviously, for a given level of taxation, lower borrowing requires lower spending. On the other hand, lower borrowing can lead to debt falling more quickly, leaving the country better placed to accommodate future pressures.

Figure 1. Borrowing


Notes: Figures for 1948–1954 are on a calendar year basis; figures for 1955–56 onwards are on a financial year basis.
Sources: Figures for the UK from Office for Budget Responsibility, Public Finances Databank,
Projections for Scotland from D. Phillips, ‘Full fiscal autonomy delayed? The SNP’s plans for further devolution to Scotland’, April 2015,


Figure 1 shows that the UK has actually run a budget surplus in only 12 years since 1948 (when comparable data begin) – and only one of the four surpluses since the mid-1970s was achieved without relying on an over-inflated economy.

This suggests that budget surpluses are not a prerequisite for a successful economy. Indeed, running budget surpluses can come at an economic cost. The government does not need to balance its books every year in order to ensure that debt is sustainable – they could simply rely on economic growth (and inflation) to erode the value of outstanding debt over time, just as UK governments did for much of the post-war period (as shown in Figure 2). On the other hand, with a surplus, debt will fall more rapidly, getting the UK more quickly to a position where it could more comfortably absorb another large shock. Figure 3 shows some illustrative projections for debt. One projection assumes a balanced budget, another assumes that the government borrows to cover investment spending. Balancing the budget in the medium term would also leave the UK public finances better placed to accommodate the longer-term fiscal pressures of an ageing population, which the Office for Budget Responsibility highlighted today in their Fiscal Sustainability Report.

Figure 2. Debt

Fig 2. Debt 

Source: Figures for National Debt are from the Debt Management Office. Figures for public sector net debt are from Office for Budget Responsibility, Public Finances Databank,


There is, however, a limit to how much the government could borrow without debt starting to rise, rather than fall.

Turning to the specific case of the fiscal arithmetic for Scotland under full fiscal autonomy, our (now much-cited) calculations suggest that Scottish borrowing in 2015–16 would be 4.6% of national income higher than the equivalent figure for the UK as a whole. This is a result of the fact that spending per person in Scotland is higher than across the UK as a whole, while revenues per person are similar. The calculations use the same methodology as the Scottish Government use for allocating tax revenues and public spending between Scotland and the rest of the UK.

This would be the result under full fiscal autonomy if – as is usually assumed – such an arrangement entailed all taxes and the vast majority of spending being devolved to Scotland, with the Scottish Government making transfers to the UK government to cover a per capita share of things like defence, foreign affairs, and the UK’s existing debt interest payments.

This 4.6% of national income figure equates to £7.6 billion, and would put Scottish borrowing at 8.6% of national income this year (or a total of £14.2 billion), compared to 4.0% of national income for the UK as a whole. £7.6 billion is essentially a measure of the annual net fiscal transfer from the rest of the UK to Scotland this year.

Our estimates suggest that, if currently planned UK-wide spending cuts are applied in Scotland, Scottish borrowing would fall from 8.6% to 4.6% of national income by 2019–20, while the UK is forecast to move to a surplus of 0.3% of national income.

While the SNP are correct to say that the UK has run deficits in most of the last 50 years, Figure 1 shows that there have been very few years when the UK has run a deficit as large as 4.6% of national income. In fact, this has only happened during the recent crisis period, the recession of the early 1990s, and the troubled times of the mid- to late 1970s. Furthermore, were the UK to run borrowing at this level, debt would be likely to continue rising over the longer term, rather than falling (the third projection shown in Figure 3).

Figure 3. Public sector net debt profiles compared


Source: The “no borrowing” scenario and the “borrowing only to invest” scenario are from Figure 2.2 of Crawford, Emmerson, Keynes and Tetlow (2015), The “borrowing only to invest” scenario assumes public sector net borrowing of 1.4% of GDP from 2019–20 onwards. The “4.6% borrowing” scenario uses the same methodology.


What level of borrowing a government chooses in the longer-term is – to some extent – a political choice. Ahead of the recent general election there was a clear distinction between the Conservative party’s stated ambition to achieve an overall budget surplus – which Mr Osborne is now planning to set in law – and the proposals by Labour, the Liberal Democrats and the SNP to allow some borrowing.

Full fiscal autonomy would give the Scottish Government control over what and how to tax and what and how to spend. But, as mentioned above, it is generally assumed that it would also lead to them losing any net transfer from the rest of the UK. If this were the case, the higher deficit they would likely face would require them to find even bigger tax increases or spending cuts than those currently planned for the UK as a whole, to avoid their debt rising rapidly. In the longer-run, full fiscal autonomy might (or might not) lead to better policies that could generate higher economic growth in Scotland, and thus allow at least some of these additional tax rises or spending cuts to be reversed. But the consequences of the short run arithmetic are not easily avoided.  

]]> Thu, 11 Jun 2015 00:00:00 +0000
<![CDATA[Public service spending: more cuts to come]]> The Conservative Party manifesto commits to cut total public spending in real terms by 1% a year in 2016–17 and 2017–18. It implies this is easy as it “means saving £1 a year in every £100 that government spends”. Spending is then to be frozen in real terms in 2018–19. If delivered this would, on the latest official forecasts, be sufficient to meet the commitment to eliminate the deficit in 2018–19, without having to announce significant further tax rises.

But as the OECD has suggested this week, meeting these spending plans will be anything but easy, as I set out on Thursday in a presentation at a pre-budget briefing organised jointly with the Institute for Government.

The reason is that keeping to these, perhaps seemingly benign, spending totals will actually require deep cuts to some areas of government. This is because underlying pressures are increasing spending in other areas. Debt interest spending is forecast to rise as both government debt and the effective interest rate on that debt rises. Spending on public service pensions is forecast to rise as the numbers receiving such pensions grows. Spending on state pensions is forecast to rise as average state pension payments continues to rise. In addition commitments to increase spending in some areas, and not to cut other areas, increases the size of the cuts required elsewhere.

One area where the Conservatives’ manifesto commits to make cuts is in social security spending (outside of protected spending on state pensions and universal pensioner benefits). But, as this IFS observation published last week points out, finding the sought after £12 billion of cuts in just two years will not be easy. Cuts of this scale amount to almost 10% of unprotected benefits. Finding such a reduction without cutting child benefit, which has been pledged this week, would mean that even more significant cuts would likely be required to spending on one or more of tax credits, housing benefit and disability & incapacity benefits.

But to meet its overall spending target, even delivering the £12 billion of cuts to social security spending would still leave the government needing a slight acceleration of cuts to departmental spending, compared to what we have seen since 2010–11. As shown in the figure, the cuts would need to increase from the 2.0% a year seen over the five years from 2010–11 to 2015–16, to 2.2% a year over the three years from 2015–16 to 2018–19. This would give a total cut of £23.8 billion across all departments between 2015–16 and 2018–19. That’s on top of the £2.2 billion of cuts taking place in 2015–16 and the £49.2 billion of cuts delivered between 2009–10 and 2014–15.

Figure. Pace of departmental spending cuts set to accelerate?

Figure. Pace of departmental spending cuts set to accelerate?

These cuts in departmental spending are to be far from evenly spread. Spending on overseas aid and the NHS is set to continue increasing in real terms, while schools’ spending per pupil is to be protected in cash terms. As we set out in our analysis of the main parties’ manifestos prior to the general election, this requires cuts elsewhere averaging 15.3% (or £30.0 billion) over the three years between 2015–16 and 2018–19. In other words two years of overall spending being cut by 1% a year, followed by a year in which overall spending is frozen, quickly becomes a 5.4% a year cut, for three years, for a swathe of public services. That would see a cumulative cut to these other unprotected departments over the eight years from 2010–11 of 32.9%.

These unprotected areas include spending by the Ministry of Defence, the Home Office, the Ministry of Justice, the Department for Communities and Local Government, the Department for Business, Innovation & Skills, and the Department for Transport. Protecting defence from any further real cuts – which would still leave its budget falling further as a share of national income – would increase the cuts elsewhere over the next three years to 18.7%, and the cuts over the eight years from 2010–11 to 36.9%.

That is not to say that such spending cuts would prove impossible to deliver. The coalition government was successful in keeping (broadly) to the spending plans set out in 2010. However it was helped by lower-than-forecast nominal growth in private sector wages, which made it easier (at least economically) to constrain the public sector pay bill.

There are further reasons to think the next phase of cuts will be harder to deliver than those achieved since 2010. Presumably efficiencies that were easy to identify and to deliver have already been made. Similarly programmes judged to be low-value have already been scrapped. And, from next April, public sector employers will find the cost of offering their staff public service pensions rises as reduced National Insurance payments for contracting out end (£3.3 billion) and new scheme valuations push up the required employer contribution rate (£1.1 billion). On top of this, new spending commitments – such as the extension of free childcare (costed by the Conservatives at £350 million), the new tax-free childcare scheme (£0.8 billion), the removal of the cap on higher education student numbers (£0.7 billion) and the Dilnot social care funding reforms (£1.0 billion) – will further increase the cuts required elsewhere.  Meanwhile demand for some public services – including social care – continues to rise as the population grows (and ages).

The cuts that the government announces later this year in next month’s Budget and the following Spending Review may turn out to be deliverable. But they certainly will not feel like is just 1% being taken out of each area of spending, nor will it require merely “£13 billion from departmental savings” as the Conservative manifesto described. While not inaccurate, these numbers give a misleading impression of what departmental spending in many areas will look like if the manifesto commitment to eliminate the deficit by 2018–19, largely through spending cuts, while not cutting spending in many areas, is to be met.

This article has also been published on The Conversation.

The Conversation

]]> Thu, 04 Jun 2015 00:00:00 +0000
<![CDATA[Benefit cuts: where might they come from?]]> The Conservative Party manifesto confirmed a long-discussed commitment to find a further £12bn of cuts to the annual social security budget, and to do this by 2017–18. That’s an £11.8bn cut in today’s prices (as all subsequent figures in this observation will be). This observation, funded by the Joseph Rowntree Foundation, briefly summarises previous IFS analysis of the context for these choices and the kinds of options that are on the table.

What is meant by a cut here is that the generosity of the system will be reduced such that annual spending is £12bn lower than it otherwise would have been. Meanwhile wider economic factors or demographic trends might also push spending up or down.  But, for example, a fall in benefit spending triggered by an increase in employment would not count as a cut in this sense. Conversely, spending on benefits remained relatively flat over the last parliament, despite net cuts of about £17bn, as underlying pressures such as falling earnings, growing rents, and growing numbers of older people approximately offset the cuts in generosity.  As always, of course, in the event of unexpected good or bad fiscal news the government would be left with a choice as to whether to allow the planned path of the deficit to change or to adjust its tax or spending plans (or some combination).

What has already been announced?

Specific benefits policies in the manifesto make a small step towards the specified total. Freezing most working-age benefits and tax credits for two years, rather than increasing them in line with inflation, would – given the current low-inflation environment – cut spending by only £1.0bn under current inflation forecasts. It implies a 1.4% real cut to the affected benefit rates by the end of the two years, which translates into an average loss among the losers of less than £100 per family per year. But it is a very broad-based cut, affecting entitlements for about 11 million families. Freezing these benefits for longer would save significantly more, particularly given that inflation is expected to be higher after 2017–18 than before it. Extending the freeze to the end of the parliament would take another £4½bn off the annual benefits bill under current forecasts. But this is irrelevant for the stated goal of finding savings by 2017–18.

Reducing the benefits cap from £26,000 to £23,000 per year would hit some families with several children and/or high rents hard: the biggest losers would be about 24,000 families who are already capped and who would lose another £3,000 per year (i.e. up to 11.5% of their income). But because in total fewer than 100,000 families would be affected and most would lose less than this, the policy reduces spending by only £0.1bn. Evidence from the current cap (discussed here) suggests that, at least in the short-term, a small minority of affected families will respond by moving into work – the cap does not apply to claimants of Working Tax Credit – and that very few indeed will respond by moving house.

Similarly, removing housing benefit from 18-21 year-old jobseekers would be a significant cut for about 20,000 young adults, but the small numbers affected limit the saving to just £0.1bn. This would increase the incentive for these individuals to move into paid work, or to claim a different out-of-work benefit (Employment and Support Allowance or Income Support) instead.

This leaves a lot not yet announced

That leaves another £10½bn of cuts to annual benefits spending that we are yet to hear about. This fiscal year, spending on benefits and tax credits is expected to total about £220 billion. The Conservative manifesto pledged to exclude more than 40% of that budget from cuts: it committed to protecting state pensions (and in fact to maintain the ‘triple lock’ on the basic state pension) and universal pensioner benefits, which account for £95bn of annual spending. This almost doubles the proportionate cut implied for the rest of the budget, to about 10%. The chart below breaks down the £125 billion of annual spending that is left unprotected.

Benefit and tax credit spending not explicitly protected by Conservatives, £ billion, 2015-16

Source: DWP benefit expenditure tables; OBR Economic and Fiscal Outlook.

The biggest items are tax credits (£30bn) and housing benefit (£26bn). Together they account for almost half of the unprotected spending. Disability and incapacity benefits between them account for almost a further third. Child benefit is the next largest. Some combination of those benefits are virtually certain to be cut if £12bn of cuts by 2017–18 are to be delivered.

What kinds of people are likely to be affected? That of course depends on the detail. But we can identify some broad likely patterns. Because of the protections outlined above, about 80% of entitlements to the benefits in the chart go to working-age families; and because the large majority of working-age benefits spending is means-tested, it would be very difficult to avoid hitting low-income households – particularly those with children – hardest: three quarters of entitlements to the benefits in the chart go to families in the bottom half of the income distribution, of which more than half go to families with children.

Specific options

In order to provide a sense of scale and of the kinds of families that could be affected, below we discuss some illustrative options for cutting working-age benefit spending. There are many more possibilities not discussed here, some of which were discussed in a previous observation looking at hypothetical options leaked to the BBC, and others of which were discussed in a chapter of February’s IFS Green Budget.

Some of the options below would of course affect overlapping groups of people. Hence, although we discuss the options in turn, the most significant impacts on family incomes will often be found where more than one cut bites at once. For example, low-income renting families could be affected by cuts to housing benefit and tax credits simultaneously, just as in the package of welfare cuts introduced by the coalition government. It is important that any package is designed coherently in view of its combined effects. Note also that some of these policies would interact with each other, in a way that means one cannot simply add up the revenue from each of them individually to arrive at a cumulative figure. For example, if you cut child benefit for large families, there is then less scope to make savings from means-testing it more aggressively.

Benefits for families with children

From January 2013, the previously-universal child benefit was tapered away from families containing an individual with a taxable income over £50,000 and removed completely for families containing an individual with a taxable income above £60,000. This meant that about 15% of families with children effectively lost some entitlement to child benefit, and about 10% lost all of it. Because (absurdly) these thresholds are fixed in cash terms over time, fiscal drag is set to double the number of families losing some or all of their child benefit over the next decade, to 2.5 million.

A way of making quick and significant cuts without affecting the entitlements of the very poorest is to take child benefit away from more families now, rather than allowing it to happen gradually and arbitrarily through fiscal drag. One could abolish child benefit per se and simply increase the child element of child tax credit (and its imminent replacement, universal credit) correspondingly, so as to compensate low-income families who claim their means-tested entitlements. This would have advantages: it would integrate the two systems of income-related support for children, rather than maintaining two systems operating in completely different ways; and it would signal the end of some undesirable features of the current child benefit charge (discussed here).It would be a big cut, meaning that the majority of families with children would lose at least £1,000 per year, with only the entitlements of approximately the lowest-income third protected. As a result it would reduce spending significantly, by around £5bn. The expansion of means-testing would come at the usual costs for work incentives and potential hassle and stigma for claimants.  One could of course design a policy to extend the means-testing of child benefit less aggressively than this (affecting fewer families and saving less money).

Alternatively, one could cut tax credits and take money away from a lower-income group of families. This would reverse some of the large increases in generosity towards this group over recent history. Returning the per-child element of child tax credit to its real (CPI-adjusted) 2003–04 level would reduce entitlements for about 3.7 million low-income families with children by an average of £1,400 per year, and would cut spending by about £5bn. For the poorest families it would mean a takeaway of £845 per child per year. Taking as an example a 2-child family where at least one parent works full-time, it would mean tax credit entitlement running out at £28,847 of gross earnings rather than £32,969.We estimate that this would increase relative child poverty by about 300,000 (or 2.5 percentage points) so, in the absence of much-needed clarity from the government on its child poverty strategy (and in particular its attitude towards the supposedly legally-binding 2020 child poverty targets) it is difficult to assess the coherence of such a policy. While about two thirds of families with children on tax credits are in work, like most cuts to means-tested benefits this policy would tend to strengthen work incentives – families would have less tax credit income to lose by increasing their earnings or to gain by reducing their earnings.

A more across-the-board cut, which featured in a document leaked to the Guardian Newspaper prior to the election, would be to reduce the amount of child benefit payable for the first child in the family by £7 per week, so that it is the same as the amount payable for subsequent children (£13.70). This would save £2.5 billion per year and would mean a flat cash-terms cut in generosity (of about £360 per year) for all families receiving child benefit.

A cut with a different distributional impact would be to abolish the amounts payable in respect of third and subsequent children. If this just applied to child benefit it would reduce spending by about £1 billion per year. Extending the same principle to means-tested support for children one would increase this sum by a further £3bn per year. However, such a policy might apply only to new births or conceptions, in which case the full saving would not be realised until the 2030s, making this of little relevance for the 2017–18 target. Losses from these cuts would be particularly concentrated towards the bottom of the income distribution: large families are more likely to be poor. Together with the fact that each losing family contains at least three children, this would again sit uneasily alongside any desire to reduce income poverty among children.

Housing benefit

The coalition government made cuts to the generosity of housing benefit totalling about £2 billion per year – though underlying trends (rising private rents relative to earnings and the growth of the private rented sector) meant that real housing benefit spending was still £1 billion higher in 2015–16 than in 2010–11. Given the scale of housing benefit, it is likely to be considered as part of any £12bn package of further cuts. As with the housing benefit cuts made in the previous parliament, further cuts may be accompanied by some additional funding for the protection of particularly vulnerable claimants, which could slightly reduce the net saving.

Cuts to housing benefit unsurprisingly tend to affect low-income families – and hence again to increase income poverty - and particularly those with high rents. They can have behavioural impacts too. There is evidence that the package of cuts to housing benefit in the previous parliament resulted in a small proportion of claimants renting cheaper types of properties; and cuts to housing benefit tend to strengthen work incentives – they mean that families have less housing benefit to lose by increasing their earnings or to gain by reducing their earnings. It is worth noting, though, that cuts to housing benefit may affect substantial numbers of in-work families. The proportion of housing benefit claimants who are in work has been rising quickly, and is now about one fifth (1.1 million).

Most of the coalition’s cuts to housing benefit were in the private rented sector, which actually accounts for only 40% of housing benefit spending. Perhaps the most obvious way to reduce generosity here is to lower the maximum amount of rent that housing benefit will cover (known as the Local Housing Allowance (LHA) rate). The coalition reduced LHA rates from the 50th percentile to the 30th percentile of rents in the local area. Early evidence suggests that this had little effect in pushing down rents, and hence resulted in most claimants paying more net rent, with a small proportion responding by moving house. We estimate that a further reduction in LHA rates to the 20th percentile of local rents would reduce spending by roughly £400 million a year, with 1.5 million low-income private renters having their entitlements reduced by an average of around £300 a year. Because LHA rates are set separately by area and family type, families whose rents are particularly low relative to other families of the same structure living in the same area would not be affected.

An alternative to reducing LHA rates is to make all claimants pay some share of the rent. This would give all tenants some incentive to shop  around for cheaper housing. Introducing a ‘co-payment’ of 10% (i.e. reducing housing benefit awards from 100% to 90% of rent up to the LHA rate) for private sector tenants would cut spending by about £0.9bn.

Any attempt to cut housing benefit spending substantially may well involve reforms affecting social tenants, as they account for the majority of such spending. Despite the so-called ‘bedroom tax’, most social housing tenants on housing benefit currently pay no net rent (i.e. their housing benefit covers all rent), subject to a means test. Introducing a new 10% co-payment for all social tenants (and reducing housing benefit by a further 10% for those already affected by the ‘bedroom tax’) would cut spending by about £1.6bn.

Whether in the private or social sectors, co-payments specified as a percentage of rent would mean that those with higher rents – including London renters and larger families - would tend to lose most in cash terms. However, even those with the lowest rents would be affected (unlike with reducing LHA rates, for example). 

As discussed above, the Conservatives have already committed to removing housing benefit entirely from 18-21 year-olds on Jobseeker’s Allowance, reducing spending by £0.1bn. One could go further: abolishing housing benefit for all aged under 25 would reduce spending by a further £1.5bn. This would be a big cut for the 300,000 affected claimants, averaging about £5,000 per year. Exempting those with dependent children – perhaps because they can less reasonably be expected to live with their parents – would roughly halve the saving. 

Disability, incapacity and carers’ benefits

Substantial amounts are spent on benefits for those with disabilities and those who care for them. With the exception of Employment and Support Allowance, these benefits are not means-tested. Hence there is more scope here than elsewhere to protect low-income families from cuts. But of course the lack of means-testing in this area of the system reflects the fact that these benefits are there to compensate for the additional costs of disability. There is evidence that, as one might expect, disability is associated with a higher level of material deprivation than income alone would suggest (though income poverty too is higher in families where someone is disabled).

If the government went down this route, then taxing universal disability benefits (and hence treating them like most other non-income-related benefits) would be one natural option (though technically a tax rise rather than a benefit cut). Those with incomes below the personal income tax allowance would, of course, not be affected. Taxing Disability Living Allowance and its replacement, Personal Independent Payments, would boost revenues by £0.9 bn. Taxing Attendance Allowance (which goes to pensioners only) as well would increase revenues by a further £0.6bn.

If the government wanted to cut benefits for higher-income recipients of Carer’s Allowance  – a benefit for full-time carers – it could simply abolish it, as apparently in a menu of options drawn up by civil servants and leaked to the BBC last month. Those with low enough incomes – most recipients of CA, in fact – would be able to claim additional means-tested benefits instead. The BBC report suggested a net saving of about £1bn per year.


The government wants to find £12bn per year of further cuts to benefits – mostly or entirely from the working-age portion of the budget – and to do this by 2017-18. This will involve difficult decisions. In many ways it provides an illustrative case study of the issues that governments always face when looking for ways to reduce spending in this area. Saving money while only affecting better-off claimants will tend to weaken work incentives. Saving money while protecting or strengthening work incentives tends to mean hitting some of the poorest in society and hence increasing poverty. We should soon find out the balance that the new government chooses to strike.

]]> Tue, 26 May 2015 00:00:00 +0000
<![CDATA[Evolving views of the taxation of saving]]> The 125th anniversary of the Economic Journal provided an opportunity to look back on several of the journal’s important papers and reflect on how these papers altered our thinking about important economic issues and laid the groundwork for further progress in economic research. One such contribution was the 1980 paper by Anthony Atkinson and Agnar Sandmo, “Welfare Implications of the Taxation of Savings,” written at a time when the most commonly held objective of tax reform was the achievement of a broad-based income tax, i.e., a comprehensive tax on all forms of income, applied to a base not eroded by deductions, exclusions, or favourable rates of tax, all of which could influence taxpayer behaviour and require higher tax rates on the remaining tax base. Apparently at odds with this objective, even then, was the basic theory of optimal taxation, which by viewing saving as a vehicle for future expenditure suggested that the taxation of income from saving – capital income – was an implicit tax on future consumption, and further implied that only the objective of taxing individuals’ future consumption, as in retirement, more heavily than their consumption earlier in life could one justify any taxation at all of the income from saving.

A recording of the session, 'The taxation of savings', at the RES conference

Confronting these opposing views, and reflecting a growing appreciation of the importance of dynamic considerations, in particular the influence of the taxation of saving on wealth accumulation and the distribution of well-being across generations, Atkinson and Sandmo (hereafter AS) reconsidered the desirable mix of different taxes, including taxes on labour income, taxes on capital income, and taxes on consumption expenditures. A key lesson from their analysis was that the influence of concerns about intergenerational equity on tax design depends on the range of instruments available to government.

In considering the appropriate combination of labour and capital income taxation, AS showed that having national debt available as a policy instrument allows the government to choose tax rates without regard to concerns for intergenerational equity. Instead, it can adjust the rate of debt accumulation to influence the relative well-being of current and future generations, borrowing more to shift burdens more onto future generations and borrowing less to shift burdens to the present. On the other hand, to the extent that debt policy lacks such flexibility (as it would if there were limits on the range of debt accumulation imposed by budget rules or the discipline of financial markets), varying the tax mix between labour income and capital income must serve two functions simultaneously: providing suitable incentives for saving by individuals within particular generations and (because labour income accrues to younger cohorts than capital income) distributing the tax burden among generations. As a consequence, capital income taxation may be called for to shift greater burdens of the cost of government activities onto older generations if reductions in national debt cannot be used to accomplish that objective.

Thus, AS provided a potential justification for capital income taxation, even through the lens of optimal tax theory, with the additional consideration of intergenerational equity taken into account. But this justification was very different from the one underlying the traditional argument for comprehensive income taxation, which might be summarized as a view that the source of income should be irrelevant to the rate at which the income is taxed. Moreover, even in the environment considered by AS, the addition of another tax instrument, in the form of a tax on consumer expenditures, largely eliminates the argument for using capital income taxes for generational equity. This is because consumption taxes, like capital income taxes, fall more heavily on older generations, while not discouraging individual saving. While this conclusion is largely implicit in the paper by AS, it follows directly from their analytical framework and helped contribute to an array of subsequent contributions, including in the two recent volumes of the Mirrlees Review, putting forward and evaluating alternative methods of implementing consumption taxation. Another motivation for relying more on consumption taxes is that they do not require one to distinguish between labour income and capital income, the difficulty of doing so being perhaps the only remaining potential justification for taxing labour and capital incomes at the same or similar rates.

While much has been learned about the taxation of saving in the years since the AS contribution, many issues remain for future research. One such issue is the integration of taxes on capital income and taxes on intergenerational wealth transfers within families. Both affect incentives for individual saving, but with different saving objectives and different implications for the intergenerational transmission of inequality through the persistence of wealth concentration. A second issue is the design of policies toward retirement saving, which for most households represents the primary reason for long-term saving. The UK, like other countries, provides substantial tax incentives specifically for retirement saving, and yet the justification for such targeted tax incentives requires further consideration, as does the design of such schemes and the consideration of alternative policies aimed at ensuring adequacy of resources among the elderly.

Alan Auerbach presented a paper at the special session, which was chaired by IFS Research Director, Richard Blundell. Richard is also director of the Centre for the Micoeconomic Analysis of Public Policy at IFS, which sponsored the session.

See all papers and sessions involving IFS staff at the RES conference 2015.

]]> Fri, 22 May 2015 00:00:00 +0000
<![CDATA[The changing characteristics of UK disability benefit recipients]]> Reform of disability benefits is high on the public policy agenda in many countries. In the UK there have been several major reforms in the last 20 years, perhaps most notably the replacement of Invalidity Benefit with Incapacity Benefit in 1995 and the replacement of Incapacity Benefit with Employment and Support Allowance from 2008. A key aim of reforms over this period has been to reduce public spending through making benefits harder to claim and through moving more recipients off these benefits and into paid work.

Over the same period there have been significant changes in spending on these benefits, the numbers receiving these benefits and their characteristics. Our new research paper, published in the Journal of Economic Perspectives, documents some of these trends.

The observed trends are dramatic. As a result, the challenges for designing appropriate public policies – both relating to the operation of disability benefits and to the consequences for employment policy – are now rather different to the past.

The percentage of men in different age groups receiving disability benefits is now more similar to the levels for women. This perhaps suggests that the issues involved with getting disability benefit recipients back into paid work may no longer be that different between men and women.

There has been systematic growth in the proportion of claimants with mental and behavioural disorders as their principal health condition, rather than physical health problems. This poses an increasingly central issue for future disability policy reform and other policies aimed at encouraging work.

Finally, the fact that an increasing proportion of younger individuals with low levels of education are receiving disability benefits might reflect the falling relative pay-off to paid work among less-educated young workers.

The new research paper finds that:

  • Spending across Great Britain on disability benefits in 2014–15 totalled £13.5 billion. At 0.8% of national income this is half the level of disability benefit spending when it was at its peak in 1995–96. A driver of this decline has been the indexation of these benefits to a measure of inflation (the Retail Prices Index up to April 2010 and the Consumer Prices Index from April 2011), while GDP has grown more quickly.

  • The overall number of individuals receiving disability benefits has fallen slightly since the mid-1990s. But this is in the presence of underlying demographic change that would have tended to push up the numbers receiving considerably – both overall population growth and the baby boomer generation reaching older working ages. The proportion of older men receiving disability benefits has actually fallen sharply since the mid-1990s, as shown in the figure below. Disability benefit receipt among men increases much less steeply with age than it used to. The distribution now looks more similar to that seen among women (right hand panel, below), for whom there has not been as sharp a decline in receipt of disability benefits at older ages.

  • Disability benefit receipt is now more related to education, and less related to age, than in the past. By 2013 25-34 year olds with a low education level were twice as likely to be on disability benefits as the highest educated 55-64 year olds. Back in 1998, the younger, less educated group were half as likely to receive disability benefits as the older, more educated group.

  • The proportion of disability benefit claims being primarily for mental and behavioural health reasons (rather than primarily for physical health problems) has steadily increased between 1999 and 2014 among men and women of all ages. For example, it almost doubled among men aged 50 to 54, increasing from 24% in August 1999 to 42% in May 2014. Among women of the same age group, the proportion increased from 28% to 42% over the same period.

  • Focussing on those aged between 50 and 59 over the period since 2002 (where we have richer data), we find that, for the same level of reported disability status, men are more likely to receive disability benefits than women and those with less education are more likely to receive benefits than the highly educated.

  • Among women aged 50 to 59 in the period since 2008, i.e. since the roll out of the new Employment and Support Allowance reforms, the reduction in numbers on benefits has been disproportionate amongst those with milder disability levels. This has resulted in a targeting of the benefit towards those with more severe disabilities. There is no evidence of a similar pattern among men of the same age group.

The paper “Disability Benefit Receipt and Reform: Reconciling Trends in the United Kingdom” was funded by the ESRC Centre for Microeconomic Analysis of Public Policy at the Institute for Fiscal Studies and is published in the Journal of Economic Perspectives.

]]> Thu, 21 May 2015 00:00:00 +0000
<![CDATA[Scotland would gain significant new powers under SNP plans for further devolution]]> The SNP manifesto sets out plans to prioritise the devolution of powers over national insurance contributions (NICs), corporation tax and welfare policy, among other areas, as a stepping stone to full fiscal responsibility.

In this Observation, we assess these plans as part of the IFS' election analysis, funded by the Nuffield Foundation. (An earlier, companion Observation looked at how the plans fit into moves towards full fiscal responsibility.)

So what can be said about the plans for devolving these specific areas? 

First, devolving NICs to Scotland could be seen as a natural next step, given that the Smith Commission has recommended devolving income tax on non-savings income.

NICs are in many ways just another income tax on earned income, and devolving the two together may allow the Scottish Government to move towards closer integration of the two taxes, if it so wished. This is something recommended by the IFS’s Mirrlees Review of the tax system. In a rational world, NICs should be treated like income tax. However, there are notional links between NICs and entitlements to contributory benefits such as pensions, and so devolving NICs could involve some tricky administrative issues. Who would pay for the pensions of people who had worked in England and retired to Scotland, for instance?

Second, the plans for the devolution of welfare in its entirety would give the Scottish Government significant new powers – not only to increase or reduce benefit rates but also to restructure the whole system.

This may result in a system better suited to Scotland’s particular needs and preferences. And devolution of the budget for welfare would allow synergies between public service spending and benefit spending to be better exploited (dealing with concerns, for instance, that the Scottish Government doesn’t benefit from the savings on benefits that result from investment in housing, education, or healthcare, say).

But as we discussed in the run up to last year’s independence referendum, any radical reform inevitably involves difficult trade-offs. Major changes benefitting some individuals would either create significant numbers of losers, many of whom would probably have low incomes, or else involve a substantial increase in overall benefit spending. The Scottish Government may also face the budget risk of benefit spending in Scotland rising more or less rapidly than in the rest of the UK – a risk currently borne by the UK government. Agreeing the block grant adjustment mechanism for welfare may be particularly tricky.

The manifesto’s proposals for changes to the benefit system might provide a guide to the type of reforms that an SNP Scottish Government would prioritise, if welfare were fully devolved. These include halting and reversing the replacement of disability living allowance (DLA) with personal independence payments (PIP), and increases to carer’s allowance and universal credit (UC) work allowances.

The first two of these are deliverable under the Smith Commission’s proposals (as disability benefits are being fully devolved), although the latter is not – and does not look like it would be possible under Labour’s proposals for more general benefit ‘top up’ powers. It is notable that each of these reforms would increase the generosity of the system, and would therefore cost money – money that would have to be found from within the Scottish budget if welfare were devolved and the plans were not adopted UK-wide.

Third, corporation tax is not a natural candidate for devolution. Relative to most other tax bases, corporate profits are particularly sensitive to differences in tax rates across jurisdictions – companies shift investment and profits between jurisdictions to take advantage of the lowest rates.

Corporation tax is therefore particularly prone to tax competition. While one might think that devolution of corporation tax would provide Scotland with a significant new lever with which to attract additional investment and profits, the lever may be less effective than hoped if the UK government responded to any reduction in tax rates in Scotland by cutting tax rates in the rest of the UK.

Instead, tax rates and revenues may be lower in both Scotland and the rest of the UK than if rates were set centrally for the whole of the UK. Furthermore, allocating profits between Scotland and the rest of the UK would entail significant administrative complexity, even if tax rates were the same and companies had no incentive to game the system.

Of course, the UK government has agreed in principle to devolve rate-setting powers (but not the tax base) for corporation tax to Northern Ireland. It says that this is justified by the particular circumstances of Northern Ireland – a land border with the Republic of Ireland, where corporation tax is just 12.5%, and a relatively weak private sector. Whether this reasoning is valid or not, it is not surprising that the SNP is asking for similar powers for Scotland.  

The proposals listed as ‘priorities’ would therefore give Scotland significant new powers which the Scottish Government may be able to use to design improved benefit and tax systems. But they would also entail additional spending and revenue risk, and involve a number of complex technical and administrative issues.  

]]> Wed, 22 Apr 2015 00:00:00 +0000
<![CDATA[Full fiscal autonomy delayed? The SNP's plans for further devolution to Scotland]]> Yesterday, the SNP published their election manifesto. Plans for full fiscal autonomy – now renamed “full fiscal responsibility” – remain as a medium term goal. But, the manifesto suggests that such plans would take “several years” to negotiate and implement. In the meantime, priority would be placed on moving beyond the recommendations of the Smith Commission with the devolution of corporation tax, National Insurance Contributions (NICs), and the welfare system.

In this Observation, part of the IFS' election analysis funded by the Nuffield Foundation, we examine what these plans would mean for the Scottish Government’s budget and powers, and the challenges involved in moving beyond them to reach full fiscal responsibility for Scotland. In doing this we look at the much cited £7.6 billion figure, originally from the IFS, which opponents of the SNP's plans have claimed would be the cost to Scotland of full fiscal responsibility. The figure is taken from recent IFS projections of Scotland’s fiscal position in 2015–16. In particular, we examine and reject a number of criticisms of the figure – that it is a snapshot for a single year and therefore irrelevant, and that it does not account for future growth in the Scottish economy.  

The Smith Commission and beyond

Following Scotland’s “No” vote in the independence referendum last September, the Smith Commission was set up with a remit of agreeing proposals for further devolution to the Scottish Parliament. It published its recommendations just two months later. On the tax side, key recommendations include the devolution of income tax rates and bands on non-savings income, air passenger duty, and the assignment of half of VAT revenues. On the spending side, around £2.5 billion of mainly disability benefits would be devolved to Scotland, as would powers over the housing elements of universal credit (UC), and powers to create discretionary payments that could provide additional support to individuals facing hardship on a case-by-case basis. Taken together, the plans would see devolved or assigned revenues making up more than half the Scottish Government’s budget, substantially higher than the 7% or so, at present.     

All of the parties represented in the Commission say that they are committed to delivering its recommendations for further devolution to Scotland in full. In addition, the Labour Party says it would go further by giving the Scottish Government not only the power to make discretionary payments, but also the power to top up any benefit – even those not being devolved, such as the state pension, child benefit or universal credit – from its own budget.

The SNP’s manifesto suggests going much further though. In the first instance, it says the SNP’s priority would be to secure the devolution of “powers over employment policy, including the minimum wage, welfare, business taxes, national insurance and equality policy”. In 2013–14, the latest year for which figures are available:

  • NICs are estimated to have raised £8.7 billion in Scotland;
  • (Onshore) corporation tax is estimated to have raised £2.8 billion in Scotland, and;
  • Spending on benefits and tax credits currently administered by DWP or HMRC is estimated to have been £17.4 billion in Scotland.

Such plans would therefore mean substantial additional revenue and spending being devolved to Scotland. The amount of taxes under the Scottish Government’s control (including local taxes) would increase by around 70% compared to the Smith Commission plans. And the Scottish Government would gain full control of welfare spending equivalent to almost half its existing budget. This would allow Scotland to not only top up existing benefits, but also engage in more significant reforms to the system. However, Scotland would remain some way short of full fiscal responsibility, especially on the tax side, where inheritance tax, capital gains tax, excise duties, and the remaining half of VAT not assigned to the Scottish Government, for instance, would continue to flow to the UK Government.  

The SNP’s assumption is that the Barnett formula would continue in operation alongside this greater devolution with the “no detriment” principle – that further devolution should not make Scotland any worse off – continuing to apply. Whether the unionist parties would agree to it applying in the face of devolution much greater than that envisaged by the Smith Commission is unclear though.

Moving to full fiscal responsibility

These plans for further devolution represent a stepping stone on the SNP’s medium term goal of full fiscal responsibility (and longer term goal of Scottish independence). The SNP have been clear that they believe that the Barnett Formula and no detriment principles should remain in place while “the Scottish Parliament’s powers remain short of full fiscal responsibility”. The implication is that once full financial responsibility is delivered, these should no longer remain in place. This would imply that the Scottish Government would become responsible for balancing its own budget. It is at this point that the unionist parties claim Scotland would face a fiscal gap necessitating spending cuts or tax rises.

These claims are based, in part, on IFS projections of Scotland’s underlying fiscal position in 2015–16. These figures show an implicit Scottish budget deficit of 8.6% of Scottish GDP compared to a budget deficit for the UK as a whole of 4.0% of GDP. A difference in borrowing of 4.6% of Scottish GDP is equivalent to £7.6 billion in cash terms. This is the figure which has been widely cited as the size of the gap Scotland would need to fill under full fiscal responsibility.

The figure has also been subject to a number of critiques. First, that it is a snap-shot relating to one year only and therefore irrelevant given that full fiscal responsibility would likely take several years to deliver. And secondly that it does not account for future growth in the Scottish economy.  How do these criticisms stack up?

It is true that the figures relate to 2015–16, and that figures for later years may differ. Projecting further into the future is also more difficult as there is more uncertainty about how Scottish and UK revenues and spending will evolve the further ahead one looks. But with these caveats in mind it is possible to undertake such projections. Table 1 shows projections for each year to 2019–20, calculated on the same basis as our figure for 2015–16. It shows that though Scotland’s deficit is projected to shrink from 8.6% of GDP in 2015–16 to 4.6% of GDP in 2019–20, the gap between Scotland’s deficit and that of the UK as a whole would, if anything, grow somewhat larger in the years ahead, reaching £9.7bn in 2019–20 (equivalent to £8.9 billion in today’s prices). The figure for 2015–16 therefore does not seem misleadingly pessimistic given current revenue and spending forecasts.  

Table 1: Net Fiscal Balance, UK and Scotland, 2013–14 (outturns), 2014-15 to 2019-20 (projections)

Net fiscal balance








% of GDP
































Cash-terms difference








Source: GERS 2013–14, OBR EFO March 2015, ONS 2012-based population projections, and author’s calculations.

The projections are calculated on the basis of Scotland’s share of UK public spending remaining at 9.2%; Scotland’s share of the UK’s oil and gas revenues remaining at 83.8%; and Scotland’s onshore revenues-per-person remaining at 97% of the UK average throughout the projection period (in each case, the same relative level as in 2013–14 as estimated in the Scottish Government’s GERS publication). This means, for instance, that the figures assume Scottish onshore revenues per person will grow at 1.9% per year in real terms between 2013–14 and 2019–20 – the same rate of growth that the OBR forecasts for the UK as a whole. In cash terms, growth in onshore revenues is projected at £14 billion, very similar to the £15 billion growth the SNP have cited, based on similar projections by Fiscal Affairs Scotland.

Of course if, with additional powers, the Scottish government could grow the Scottish economy more quickly than that of the rest of the UK then revenues would grow more rapidly, shrinking the fiscal gap. But there is no guarantee of such growth and assuming similar growth between Scotland and the rest of the UK is clearly the most useful baseline comparison.

It also provides information to assess how much faster growth would have to be to close the gap. To close the gap by 2019–20, for instance, Scottish revenues per person would need to grow by more than twice as much as forecast for the UK as a whole – 4.5% in real terms per year – between 2013–14 and 2019–20. Even closing the gap over a longer ten or fifteen year horizon would require a step-change in Scottish economic performance, and revenue generation. Such a change is not impossible, but is much easier to promise than it is to deliver. As we have highlighted before, the types of policies previously outlined by the SNP as potential ways to boost growth, such as cuts to corporation tax and air passenger duty, and increases in childcare spending, would, at least in the short to medium run, cost the government money, and widen rather than shrink the fiscal gap, even if they did boost growth.   


The SNP’s manifesto confirms a policy goal of “full fiscal responsibility”, at which point Scotland would have to fund its spending through its own tax revenues and borrowing. In the shorter term though the priority is a package of substantial further devolution of tax and spending powers that stops some way short of full responsibility. The hope seems to be that any fiscal gap opening up as a result of full fiscal responsibility would therefore be delayed by invoking the “no detriment” principle for a package that stops short of full fiscal responsibility.

Delaying a move to full responsibility for a few years would not on its own deal with the fiscal gap though. Indeed, if anything, given current spending and revenue forecasts, the gap would likely grow rather than shrink over the next few years. It would remain the case that full fiscal responsibility would likely entail substantial spending cuts or tax rises in Scotland. While a big and sustained rebound in oil revenues or significantly higher growth in Scotland could mitigate this, there can be no presumption that either would occur. 

Update 22/04/2015: A second Observation by David Phillips has now been published on the SNP's plans to prioritise the devolution of areas of tax and benefits - including national insurance contributions, welfare policy and corporation tax.

]]> Tue, 21 Apr 2015 00:00:00 +0000
<![CDATA[Education spending: what are the parties planning to protect?]]> Today, the Liberal Democrats launch their manifesto and have highlighted a commitment to protect education spending per pupil in England for 2-19 year olds in real-terms. Labour and the Conservatives have also announced protections for education spending in England. What are the implications of these different protections for how schools and education spending will evolve over the next parliament?

As we set out in our recent Election Briefing Note, the Conservatives have committed to protecting cash spending per 5–16 year-old pupil between 2015–16 and 2019–20. Since pupil numbers are forecast to grow by 7.0% between 2015–16 and 2019–20, that implies a 7% cash increase in spending on this group. Their manifesto also states that their commitments allow for a real-terms protection in schools spending over the next parliament.

Labour have committed to increasing the entire education budget in England at least in line with inflation. If both parties just meet these commitments and Labour increase all areas of education spending by equal percentage amounts, then these two pledges both imply a real-terms freeze in schools spending. However, Labour would additionally protect non-schools spending in real terms (mostly the early years budget and 16–19 education spending), which the Conservatives have not pledged to protect.

Today, the Liberal Democrats have announced that they would protect the 2–19 education budget in real terms up until 2017–18 (which would mean increasing nominal spending at least in line with inflation). After 2018–19, they would increase it in line with economic growth. Together, this makes for a 12.8% nominal increase in 2–19 education spending between 2015–16 and 2019–20, or a 4.8% real-terms increase. However, most of this increase is back loaded to after 2017–18.

How does this compare with the protections announced by the Conservatives and Labour?

Overall, Labour and the Liberal Democrats are committed to protecting a larger part of education spending than are the Conservatives. In addition, the Liberal Democrats are committed to a faster increase in spending than the Conservatives in the period after 2017–18.

Compared with Labour, the Liberal Democrats have only committed to protecting the 2–19 education budget (which they say is currently £49.2bn). Labour have instead committed to protecting the entire Department for Education resource budget (which was £54.2bn in 2014–15 in 2015–16 prices). However, Labour have committed to protecting the education budget in real-terms, whilst the Liberal Democrat commitment currently implies increasing 2–19 education spending by 4.8% in real terms.

Therefore, the Liberal Democrats are protecting a slightly smaller definition of education spending than Labour, but have committed to increasing this by more.

The Liberal Democrats say that their commitment allows them to protect 2–19 education spending per pupil in real terms. This is true as their commitment implies a 4.8% real-terms increase in 2–19 education spending between 2015–16 and 2019–20 and the number of pupils aged between 2 and 19 is expected to grow by at a similar rate. However, it is worth noting that the number of school-age pupils (ages 5–16) is expected to grow by more – by 7% between January 2016 and January 2020. If the Liberal Democrats increase all areas of education spending in equal percentage amounts, then their commitment still implies a 2.1% real-terms fall in current school spending per pupil. This though is still more generous than what is implied by the Labour and Conservative commitments, which could both imply a 6.6% real-terms fall in current school spending per pupil between 2015–16 and 2019–20.

The Liberal Democrats (and Labour) could choose to increase schools spending by more than other areas of education spending, which might be the more natural thing to do if school-age pupils are expected to grow by the most. 

]]> Wed, 15 Apr 2015 00:00:00 +0000
<![CDATA[Conservative and Labour proposals to cut pensions tax relief for those with an income above £150,000]]> The Conservatives today announced that they want to reduce the generosity of pension tax relief for those with an income of above £150,000 a year. The Labour party have already announced that they want to cut pension tax relief for this group. While only affecting a relatively small number of high income individuals both sets of proposals have the potential to be complex, damaging and counter-productive. The Labour proposals involve the introduction of a “cliff edge” such that an increase in income could leave people substantially worse off. The Conservative proposals have the odd effect of allowing anyone with an income of up to £150,000 to put £40,000 tax free into a pension, but allowing those on higher incomes to put in much less. Their proposals effectively increase the marginal tax rate faced by many of those with an income between £150,000 and £210,000 a year.

The Conservative proposals

At present those with taxable incomes over £150,000 a year – there are about 300,000 such individuals – pay income tax at 45%. Any pension contribution they make up to an annual allowance of £40,000 a year attracts income tax relief at that rate. Income tax would be paid at the point at which the pension is withdrawn – as an annuity or in some other form. The exception is that a tax free lump sum worth a quarter of the accumulated pension pot can be withdrawn.

The Conservatives propose to reduce the annual allowance to £10,000 once income reaches £210,000. In other words 50p of allowance will be lost for every additional £1 of income in a range between £150,000 and £210,000. For anyone who continued to put their income into a pension that would effectively raise their marginal income tax rate to 67.5%. Those making other choices – to consume their income immediately or to save in a somewhat less tax privileged form – will face some increase in their effective marginal tax rate, though not always to the same extent.

Whilst affecting a relatively small number of high income individuals a reform such as this would further complicate the pension tax system. It would have the curious effect of allowing those with an income of up to £150,000 to save £40,000 a year in a pension but restrict that to £10,000 a year for those with an income of more than £210,000. Given that there is a great deal to be said for a pension tax regime which allows savings to be made free of tax and then tax to be paid upon withdrawal this complication would not improve the efficiency of the system.

The biggest effect of this change would be on those with incomes between £150,000 and £210,000. It becomes proportionately less significant for those on the very highest incomes.

The amount it would raise in the long run will depend on how people respond. To the extent that those affected spend their income now rather than in the future, at least some of the apparent additional tax revenues will be brought forward rather than increased in total. Similarly if people respond by putting more money into ISAs the Exchequer will get more revenue now and less later on.

The Labour proposals

The Labour party has also proposed to reduce the value of pension tax relief for those with an income of over £150,000. Their proposals are at least as complex.

They want to reduce the rate of income tax relief for those with an income of more than £150,000 a year. This is similar to a policy proposed by the last Labour Government in its 2009 Budget. At that time, they proposed restricting tax relief on pension contributions to the basic rate (20%), but only for those with incomes above £130,000 and whose gross income plus employer pension contributions was above £150,000 (see here for a discussion). The way the policy was designed meant that some with large employer pension contributions would face a substantial increase in their income tax bill if their income rose from just under to just above the £130,000 threshold: For example an individual earning £129,000 plus an employer pension contribution of £40,000 would face an increase in their annual income tax bill of over £10,000 if their current wage were to rise to £130,000 (assuming a top rate of income tax of 50%). Effects of this kind are almost inevitable when introducing complexities such as this which result in treating people very differently once their incomes rise above a certain level.

The policy was thankfully dropped by the incoming coalition government in favour of a reduction to annual and lifetime allowances designed to raise the same amount of money.

In conclusion

We risk rushing towards something like chaos in the taxation of pensions for those on high incomes. Both Conservative and Labour plans will have substantial incentive and behavioural effects for those with incomes in the £150,000 to £200,000 range – potentially bigger effects than the 45% (or 50% under Labour) income tax rate itself.

This matters. Of course nobody on these sorts of incomes is going to be pushed into penury by such changes. But the undesirable distortions to savings behaviour and to work incentives have the potential to be significant. The two main parties seem to be competing to tie their own hands on the main tax rates whilst scooping up apparently free money from “the rich”, non-doms and tax avoiders on the other. There is a danger that the tax proposals being put forward through this general election campaign will have a long term malign influence on our tax system and economic welfare.

]]> Sun, 12 Apr 2015 00:00:00 +0000
<![CDATA[Conservatives' proposed cut to Inheritance Tax on main homes]]> Inheritance tax (IHT) applies to (some) wealth that is transferred on or shortly before death. The tax is currently charged at 40%, with each individual receiving an allowance of £325,000. Any unused proportion of this allowance is transferable to a surviving spouse or civil partner, effectively doubling the inheritance tax threshold for many couples. An IFS observation published last year considered the case for radical reform of IHT.

The Conservatives would introduce a new £175,000 per person transferable allowance for main residences when they are passed to children or grandchildren. For many couples this will give a total allowance of £1 million (£325,000 plus £175,000 each). This new allowance will be tapered away from those leaving more than £2 million with the intention that those leaving more than £2.35m will not benefit from the new allowance.

The vast majority of estates (over 90%) are not liable to IHT at the moment and therefore would not benefit. The Conservatives estimate that their policy would be a giveaway of about £1 billion. With around 50,000 estates forecast to pay IHT over the next few years this gives an average (mean) gain per IHT paying estate of around £20,000. The maximum reduction in IHT on a couple’s estate is £140,000 which will go to married couples with estates worth between £1 million and £2 million. Since the children of those with very large estates are disproportionately towards the top of the income distribution the gains from this (and in fact any) IHT cut will also go disproportionately to those towards the top of the income distribution.

Many features of the policy are similar to one analysed in a Treasury document that was leaked to, and published by, the Guardian last month. This estimates that, based on Budget 2014 forecasts, the policy would reduce the proportion of estates liable for IHT from 8% in 2015–16 to just over 6% by the end of the parliament, rather than see it rise to just over 10% under current policy.

As this HMT document argues (para 16, page 9) “there are not strong economic arguments for introducing an inheritance tax exemption specifically related to main residences”. The document lists a number of problems with the policy for example the fact that it would encourage investment in owner-occupied housing rather than other more productive investments and discourage downsizing late in life when that might otherwise be appropriate.

The Conservatives’ proposal would further complicate the IHT system. The Figure below shows the marginal rate of IHT faced by an individual with a home worth £175,000 by the size of their total estate, before and after the change (assuming their estate is to be bequeathed to a child or grandchild). The new effective IHT rate of 60% that kicks in at £2 million is due to the tapering back of the new allowance. Why the IHT rate should go 0%, 40%, 60% and then return to 40% is difficult to justify. A preferable policy would have been simply to increase the existing threshold from £325,000, whereas under current policy it is set to be frozen at this level (which is the level it was at in 2009–10) through to 2017–18.

Inheritance tax schedule for single individual with a home worth £175,000

Inheritance tax schedule for single individual with a home worth £175,000

]]> Sun, 12 Apr 2015 00:00:00 +0000
<![CDATA[Unknown quantities: Labour’s ‘non-dom’ proposal]]> Yesterday the Labour Party proposed abolishing ‘non-dom’ status except for people who only come to the UK for a short period.

Foreign domiciliaries, or ‘non-doms’, are people who live in the UK but whose permanent home for tax purposes (’domicile’) is considered to be elsewhere.

They are already taxed in full on their UK income and capital gains. But unlike other UK residents, they can opt to be taxed on the ‘remittance basis’, meaning that they are not taxed on their foreign income and capital gains unless they bring the proceeds into the UK. And non-doms get other tax privileges, notably favourable treatment of non-resident trusts, even if they do not claim the remittance basis. For example, a UK domiciliary normally faces an inheritance tax charge when they put assets into a non-resident trust, and pay capital gains tax (CGT) on any rise in value when the assets are sold; non-doms do not have to pay either of those taxes.

Although non-dom status has a history going back more than 200 years, it acquired fresh political prominence in the 2000s. Starting in 2008, the then Labour government required non-doms who had lived in the UK for seven years to pay a £30,000 annual charge, and to give up their income tax and CGT annual allowances, if they wished to be taxed on the remittance basis. Since 2010 the coalition government has tightened the eligibility rules and increased the charge for longer-term residents. The latest increase took effect on Monday and means that the charge is now £60,000 after 12 years in the UK and £90,000 after 17 years. The UK’s regime remains unusually generous, however, though international practice varies.

In 2012–13 114,800 people recorded non-dom status on their self-assessment income tax returns, though the true number of non-doms is likely to be higher than that as non-doms are not obliged to record that status on their tax return unless they claim the remittance basis. Only a minority of non-doms, 46,700 people, claimed remittance basis in 2012–13: the rest either had no significant income or gains kept abroad or else chose to pay tax on them. And among those 46,700 people taxed on the remittance basis, only 5,100 paid the charge, the other 89% presumably being people who had lived in the UK for less than seven years. Those 5,100 people between them paid £226 million in remittance basis charges. (Note that that is not the revenue yield from the existence of the charge, since the existence of the charge affects revenue from other taxes too: some people opt to pay tax on their worldwide income and capital gains to avoid the charge, while others decide not to live in the UK at all.) Of this £226 million, £43 million was paid by 3,700 non-doms who had lived in the UK for 7-12 years, and the remaining £183 million by 1,400 non-doms who had lived in the UK for more than 12 years. The government estimates that its latest changes will raise a further £90 million a year from people living in the UK for more than 12 years.

Non-doms are a diverse bunch.

A typical non-dom is a foreigner who comes to work in the UK for a few years and then returns. Such people may be rather footloose and a significant (but unknowable) fraction could be put off coming to the UK by a hefty extra tax bill. Many will also have close family members abroad, and so might be able to avoid UK tax  even under Labour’s proposals simply by having assets owned by a family member. As a matter of principle it is also not obvious how the UK should want to tax such people. Is it fair to tax them on, say, capital gains on assets abroad which they sell while in the UK but which may have risen in value before they came here? Or to levy inheritance tax on their worldwide estate if they die a year after arriving in the UK? How long should a person have to live in the UK before we start taxing them like UK domiciliaries? There is not a single ‘right answer’ to such questions, and other countries take a variety of different approaches; for example, Canada rebases all assets to market value when a foreigner arrives there so as to avoid taxing gains that accrued prior to their arrival.  Double tax treaties aim (not always successfully) to ensure that the same money is not taxed in more than one country.

But non-doms also include people born and bred in the UK. By default, a person’s domicile is simply the domicile of his or her father (or mother if they were unmarried at the time of birth), though one can change domicile by settling permanently in another country. It is quite possible for a person and both of his or her parents to have been born and lived their entire lives in the UK and yet be domiciled elsewhere, although some practitioners think that HM Revenue and Customs should be much more forceful in challenging such cases as it should surely be possible to show that after 20 years living in the UK someone really does intend to settle here. As a matter of principle it seems hard to justify taxing such people more lightly than other UK citizens. And on a pragmatic level, people who have lived in the UK for many years are presumably less likely to leave the UK in response to a tax rise – though having some ties to (and considerable income or assets in) another country might indicate that they are still more mobile than UK domiciliaries.

The Labour Party argues that abolishing non-dom status except for short-term residents would be fairer than the current system and would raise ‘hundreds of millions of pounds’.

There are three potential sources of additional revenue from Labour’s proposal:

  • Abolishing the remittance basis for those currently paying the charge. The 5,000 or so people paying the remittance basis charge are presumably doing so because the charge is less than the tax they would otherwise have to pay on their unremitted income and capital gains. This implies that there is potentially some revenue available from removing this option, but nobody knows how much because those who pay the charge do not currently have to disclose their unremitted income and gains. And the more additional tax is at stake, the more steps people will take to avoid paying it.
  • Abolishing the remittance basis for a wider group of those currently claiming it. The fact that most non-doms claiming the remittance basis have lived in the UK for less than seven years (and are therefore not liable for the existing charge) implies that the length of the grace period that Labour proposes to allow for ‘temporary’ residents before full tax becomes payable would be crucial in determining how many non-doms were affected. Labour has said that it will consult on this, though Ed Balls has suggested that the period might be two to three years and that five years would be ‘probably too long’. The shorter this period, the more non-doms will be affected and the more potential revenue at stake, but also the less tied to the UK those affected are likely to be and the more likely they are to be deterred from coming to (or staying in) the UK, with a corresponding loss of revenue. 
  • Removing some or all of the other tax advantages, such as those associated with trusts, that non-doms currently enjoy even if they do not claim the remittance basis. This may be important as in some cases the tax advantages in question can be large, though again the number of non-resident trusts and the sums involved are unknown. But it is not clear whether Labour would seek to remove these reliefs entirely – or what current EU rules on free movement of capital would allow. 

Thus the revenue impact of the policy would depend on details that have not yet been announced (such as the length of the grace period and the treatment of trusts) and on the amounts of foreign money involved, which are not known. It would also depend heavily on how non-doms responded to the reform.

 How might non-doms respond to a change?

One extreme response might be to leave the UK altogether (or, perhaps more likely, not to come here in the first place). Another would be to evade the tax by failing to declare all their foreign income etc. to HM Revenue and Customs: such behaviour would be illegal but difficult for HMRC to detect since the money is not brought into the UK, though some progress is being made on co-operating with revenue authorities in other countries to exchange information.

There could also be less extreme responses such as transferring assets to family members abroad, making more use of non-resident trusts, spending enough time overseas to avoid qualifying as UK-resident for tax purposes, or simply generating less overseas income and capital gains. Ingenious tax advisors might find more ways to avoid the tax. Much would depend on the details of the reform.

The likely scale of this type of behaviour is essentially unknowable in advance. When announcing the latest increase in the remittance basis charge in Autumn Statement 2014, the government forecast that behavioural responses would reduce the yield by around a quarter. But that estimate was necessarily somewhat speculative, and even if accurate it does not follow that the behavioural response to a larger tax rise will account for a similar proportion of revenue. Furthermore the Autumn Statement measure applied only to people who had lived in the UK for at least 12 years, who are likely to be much less mobile than those here for shorter periods. And reforming the taxation of non-resident trusts is a potentially significant aspect of the reform without recent UK precedent to suggest what its effect might be.

The overall revenue effect of such a change is equally hard to predict. To the extent that non-doms who currently pay a large amount of tax (the Daily Telegraph reports that non-doms paid £8.2 billion in income tax and National Insurance contributions alone in 2012–13) leave the country altogether, the change could cost the exchequer revenue. If there is less behavioural response and it turns out there is a lot of money currently untaxed overseas then additional revenue of £1 billion or more may be achievable.

The uncertain revenue implications of Labour’s proposed reform does not make it a bad idea: it would be wrong only ever to introduce policies whose effects can be forecast with certainty. And revenue is not the only consideration. There is also the pain caused to non-doms (who would be worse off even if they did not pay any more tax, since the behavioural response is something they would not choose to do in the absence of the reform), and the fairness (or otherwise) of taxing them more like we tax other UK residents. And there could potentially be benefits (or costs) to the rest of society from having non-doms in the UK, over and above their net contribution to the exchequer. Employing people and buying goods and services in the UK do not, in themselves, normally constitute benefits to wider society, since those resources could otherwise be put to alternative uses; but if non-doms are unusually entrepreneurial then there could be some positive spillovers to the productivity of other UK businesses. Some would argue that the presence of a super-rich and socially segregated non-dom elite detracts from the cohesion of British society; others might argue that non-doms bring welcome diversity and vibrancy to British society. The impossibility of quantifying these costs and benefits reliably does not mean they should be ignored.


Even after a series of reforms in recent years, the current non-dom rules look anachronistic. It does seem inequitable to give preferential treatment to some individuals who have lived in this country all their lives. That said, the number paying the remittance basis charge – that is, the number who have lived in the UK for at least seven years and declare substantial overseas income or capital gains – is small. And there are benefits to having highly skilled, internationally mobile individuals living and working here for periods of time. Changes which ironed out the obvious inequities whilst preserving the UK’s attractiveness to mobile workers would be welcome. At a minimum, taxation should be based on objective, observable criteria such as years of residence rather than nebulous and difficult-to-prove criteria such as where one intends to reside permanently and the array of unsatisfactory indicators that are used to proxy that.

The details of Labour’s proposal, especially with respect to the period of time for which people arriving in the UK continue to be taxed preferentially, will be important. Whether it would raise  much, if any, revenue is uncertain: we do not know full details of the policy, full details of non-doms’ overseas income and assets, or how non-doms would respond to the reform. But revenue is not the only criterion by which the proposal should be judged. It matters more broadly how far it would improve the efficiency and equity of a clearly problematic aspect of the current tax system.

]]> Thu, 09 Apr 2015 00:00:00 +0000
<![CDATA[Annuity buy-back: thoughts on the potential market and possible pitfalls for pensioners]]> In his final Budget before the General Election, George Osborne confirmed the next step on his journey to provide greater flexibility in how people can access their defined contribution (DC) pension savings: removing the tax penalty that currently applies to anyone who wants to sell their existing annuity.

On the face of it, this appears to provide welcome new flexibility for the estimated 5 million people who currently receive income from a DC annuity in the UK. But it is not clear that this reform will move us to a new utopia for those who have already purchased annuities.

First, potential purchasers may – for reasons not simply related to a desire to make a profit at the expense of the seller – not be willing to pay a price for annuities that sellers will be willing to accept. Second, there could be pitfalls if annuity holders are not able to make well-informed decisions about whether to sell or not. Finally, part of the animosity towards the old ‘compulsory annuitisation’ requirement was driven by a (not always well-evidenced) belief that annuities offered poor value for money – indeed the government’s consultation last year on removing the compulsory annuitisation requirement stated that “the annuities market is currently not working in the best interests of all consumers. It is neither competitive nor innovative and some consumers are getting a poor deal”: but, if the market for selling annuities was perceived not to work well, it is far from clear why a market for buying them back should work much better.

At present, if someone sells an annuity that they have already bought with funds from a DC pension, they face a tax charge of up to 55% (or 70% in some cases). Budget 2015 proposed that, from April 2016 onwards, people will be able to sell their annuity and pay tax at their marginal rate when they draw on the sale proceeds. The government is consulting on a number of issues, including how the secondary market for annuities should operate and what advice and guidance should be provided to potential sellers. Both are important questions.

Who will be affected by the change?

The government estimates that in 2013 there were 5 million people receiving DC annuity income totalling £13 billion – making the potential size of the secondary market for annuities significant. Using data from a representative survey of the older English household population, Figure 1 shows that 1 in 4 people aged 55 and over receive income from a DC pension annuity. This fraction varies significantly by age and sex: 1 in 7 women aged 55–64 fall into this group, compared to nearly half of men aged 65–74. The higher prevalence of DC annuities among those aged 65–74 than among older people reflects the relatively recent emergence of DC pensions, which means that fewer individuals in earlier cohorts will have saved any money in a DC pension.

Annuities pay out money until the annuity holder dies (or, in the case of joint life annuities, until they and their partner have both died). Therefore, an annuity will be ‘worth’ more if the person is expected to live for a long time, and less if they are expected to die soon.

Table 1 shows an estimate of the value of the annuities held by people aged 55 and over in England, based on the assumption that each person lives to the average life expectancy for their gender for people born in the same year. This is likely to understate the true ‘value’ of these annuities since – as Finkelstein and Poterba (2002) showed – annuitants live for longer, on average, than non-annuitants.

Table 1: Estimate value of DC annuities in payment (£)

  25th percentile Median 75th percentile
55–64 12,503 28,299 72,713


29,114 69,812
75–84 9,148 23,112 52,349
85+ 4,568 15,002 31,894
All, 55+ 10,665 25,854 61,766

Note: Estimated values are calculated as the discounted present value of the annuity stream, using a real discount rate of 3% a year and assuming that each individual lives to their age- and sex-specific cohort life expectancy, as estimated by the Office for National Statistics. These calculations also make the simplifying assumption that all annuities in payment are single life, rather than joint-life, annuities.
Source: English Longitudinal Study of Ageing, 2012–13. Weighted using cross-sectional weights.

Using this method of estimating the ‘value’ of annuities, we estimate that the median value of future income streams from annuities in payment is nearly £26,000. The value of annuities is lower among older people than younger people, which is because older individuals have fewer years left to live. There is a significant range, however, in the estimated value of annuities in payment – 1 in 4 are estimated to be worth less than £11,000, while 1 in 4 are worth more than £61,000.

Will there be a market for selling annuities?

The estimates of the value of annuities shown above assume that all annuity holders are expected to live to their age- and sex-specific life expectancy. However, not everyone will – some will live longer, some will live less long and both the individuals themselves, as well as the companies seeking to buy the annuity, may have differing views on how much this is the case.

If many of those looking to sell an annuity are those who have good reason to believe they might die soon, and if those buying the annuity cannot fully reflect this in the price (either because they do not have full information or because they are legally prevented from using some information), then prices on offer will assume that those looking to sell will be likely to die soon. These prices will be unattractive to many potential sellers. This is an issue that arises in many insurance markets and is known as ‘adverse selection’.

There are good reasons to think that annuity sellers will have more information about their potential longevity than potential purchasers are able to reflect in the prices they offer. First, women, on average, live longer than men but it may well be illegal for purchasers to offer prices that vary by sex: meaning that the resulting market will be particularly unattractive to women. Second, individuals may well have better information about their survival chances than potential purchasers can hope to elicit. For example, they know more about their own lifestyle choices (such as how much exercise they do, and how much alcohol, tobacco and unhealthy food they enjoy), detailed family histories of disease, and whether they suspect they might be developing health problems that have not yet been diagnosed. For these reasons, the secondary market for annuities might be limited or even difficult to establish at all.

Making good decisions

The argument just outlined assumes that individuals are reasonably well-informed and are capable of making complex financial decisions of the type required in deciding whether or not to sell an annuity. But valuing an annuity versus a lump sum is a complex calculation and requires people to grapple with uncertainty surrounding their own longevity, potential future investment returns and inflation. Therefore, a further concern with this policy might be that some annuity holders are not well-placed to make such decisions and may – as a result of this policy change – make a choice that is to their detriment.

Table 2: Financial decision-making capability among DC annuity holders (%)

  55–64    65–74    75–84    85+    All, 55+
Has difficulty managing money 1.6   1.2   4.3   9.7   2.5
“Let’s say you have £200 in a savings account. The account earns 10% interest each year. How much would you have in the account at the end of 2 years?”                  
£242 (correct answer) 29.2   24.9   13.4   14.0   22.6
£240 (simple interest answer) 41.8   37.3   40.9   31.8   38.9
Other answer 29.0   37.8   45.7   54.2   38.5

Source: English Longitudinal Study of Ageing, 2008–09 and 2012–13. Weighted using cross-sectional weights.

Table 2 presents some indicators of the ability of annuity holders to make a decision about whether or not to sell their annuity. The top row of the table shows what percentage of annuity holders report that they have difficulty managing their money because of a physical, mental, emotional or memory problem. The fraction that report difficulty with this is very small among younger annuity holders (less than 2% among those aged under 75) but rises to nearly 10% of annuity holders aged 85 and over.

The bottom part of the table provides an indication of higher-level financial capability – reporting answers to a question gauging individuals’ understanding of compound interest, which is a relevant concept in comparing the value of a lump sum and a future annuity income. The table shows that – even among annuity holders aged 55–64 – only 3 in 10 were correctly able to answer a question about compound interest. Among older annuity holders far fewer were able to give a correct answer: with only around 1 in 7 of those aged 75–84 or aged 85 and over giving the right answer.

This suggests that a significant minority of those who could potentially sell their annuity may not be able to work out accurately what a fair price for their annuity would be without proper advice or guidance.

There is also some evidence that individuals have difficulty valuing annuities compared to lump sums (albeit not from the UK and using experimental evidence rather than choices made in the real world). Brown et al (2013) presented individuals in the United States with an opportunity to buy a theoretical annuity and scenarios in which the same individuals were offered an opportunity to sell an annuity. Their experiments provide a number of pieces of evidence showing that individuals have difficulty valuing annuities and that the degree of difficulty with annuity choices is correlated with cognitive ability.

The researchers found a significant divergence between the prices at which individuals are willing to buy and sell annuities – with the price required to sell an annuity being typically significantly higher than the price at which people report being willing to buy an annuity. This divergence is greatest on average for those with the least cognitive ability (as measured by education, financial literacy, and numeracy).

Individuals’ valuations of annuities are also found to be sensitive to ‘anchoring’ effects. Specifically, they offered individuals various different prices at which to sell their annuity and they found that the price that individuals eventually agreed to sell at was affected by the price that they were first offered. That is, individuals who were initially offered a low price tended to end up agreeing to sell at a lower average price than individuals who were first offered a higher price. This ‘anchoring’ effect was much more likely among those with more limited cognitive abilities.


On the face of it, the Chancellor’s announcement in Budget 2015 that, from next year, people will be able to sell their annuities provides a welcome liberalisation of the market and is perhaps a natural next step to follow the changes made in Budget 2014. However, there is a significant risk that the secondary market for annuities will be limited or fail to emerge at all, due to the potentially significant problems of ‘adverse selection’ in this market. If this was the only issue, we might still not be unduly concerned about the changes. A more serious issue could arise from individuals’ potential inability to make well-informed decisions about whether or not to sell their annuity. Evidence suggests that at least a significant minority of annuity holders – in particular, older annuity holders – may struggle with the complex decisions required in valuing their annuity compared to an alternative lump sum. This suggests that, at the very least, individuals will need to have access to good quality financial advice and guidance in order to navigate this new market – if, indeed, such a market does spring into existence. 

]]> Wed, 01 Apr 2015 00:00:00 +0000
<![CDATA[New public service pensions remain relatively generous despite cuts]]> Reforms to two of the three largest public service pension schemes in England and Wales – the NHS and the Teachers’ Pension Schemes – are coming into effect tomorrow, Wednesday 1 April, changing radically how pensions for members of these schemes are calculated.

These changes are the final changes made to public service pensions by the coalition government following Lord Hutton’s Review in 2011. (Changes to the Local Government Pension Scheme – the largest public service pension scheme in England and Wales – came into effect in April 2014.)

Until now members of the NHS and Teachers’ Pension Schemes have accrued a pension that pays a fraction of their final salary from their “Normal Pension Age” (NPA). For those who joined the NHS scheme before 1 April 2008, or the Teachers’ Pension Scheme before 1 January 2007, this NPA was set at age 60; while for those who joined more recently it was, as a result of reforms implemented by the last Labour government, increased to age 65.

The latest reforms have changed the system in three key ways (further details are provided in the table below), which taken together will reduce overall pension entitlements and, therefore, reduce the amount of future taxpayer support for these schemes:

  • Pension benefits are to be calculated as a fraction of career average earnings rather than final salary.

  • The NPA has been increased to be aligned with each member’s state pension age (SPA).

  • The fraction of earnings accrued in pension for each year of service has been made more generous.

These changes only affect the pension that scheme members will accrue in future; they have no effect on the pension rights that members have already accrued: so if an individual draws a pension at their old NPA, the payment in respect of service before 2015 will be unchanged. The changes for pension accrual from 2015 onwards affect different people in different ways depending on when they joined their pension scheme, their age and their salary progression.

Those who were within 10 years of their NPA on 1 April 2012 are unaffected: they will continue to accrue pension rights under the old final salary scheme rules.

Those who were more than 10 years from the NPA on 1 April 2012 and who joined the pension schemes prior to Labour’s reforms will be made worse off on average. This is primarily because they will see an increase in their NPA from 60 to at least 66. This means that they would have to pay into their pension for at least six more years in order to receive their full pension (albeit – with extra years of service – potentially a larger annual pension) for at least six fewer years. This is a substantial fall in average generosity for this group.

Those who were more than 10 years from their NPA on 1 April 2012 and who joined the pension schemes after Labour’s reforms will, on average, only be a little worse off. Their NPA will increase slightly, their pension will be calculated based on career average earnings, rather than final salary, and they will receive more generous annual accrual. For this group our previous calculations suggest that average pension accrual – that is the increase in the value of an individual’s pension arising from one more year of service – will fall only marginally from 18.3% of salary to 17.7% of salary.

These figures are, however, just averages. Final salary schemes are more generous to those who see rapid growth in their pay, particularly when this happens towards the end of a long career. This is not the case when a career average earnings measure is used instead. So the move to a career average scheme will be particularly detrimental to “high flyers”, who will lose out relative to those who experience lower pay progression. Previous calculations by IFS researchers suggest that low educated workers, whose pay tends to rise less quickly as they age, may actually see their pensions made more generous as a result of this reform. Since they are also the group who are least likely to be covered by a relatively generous employer-provided pension in the private sector, the latest reforms actually exacerbate differences between the public and private sector in the current pattern of pension provision across education groups.

One of the other changes to public service pensions recommended under Lord Hutton’s review – an increase in employee contributions – also affected higher earners more than low earners. Employee contributions now have a “tiered” structure, with contributions substantially higher as a fraction of salary for higher earners than for lower earners. However, the majority of the contribution increases have already been implemented, rather than being introduced for the first time in 2015–16.

These reforms to public service pensions come on top of Labour’s increase in the NPA to age 65 for most new members, a previous increase in member contribution rates, and the current government’s substantial reduction in the generosity of indexation (by changing the measure of inflation used from RPI to CPI). Taken together these reforms dramatically reduce the overall generosity of public service pensions: previous estimates suggest that without any of these reforms average accrual would have been twice as large.

Despite these reductions, on average public sector employees still accrue more generous pensions than their private sector counterparts. The vast majority (87%) of public sector workers are members of an employer provided pension scheme but this was only true of 49% of private sector employees in 2014. Although automatic enrolment will continue to boost coverage in the private sector, the defined contribution schemes typically on offer in the private sector are much less generous than the defined benefit schemes available to most public sector workers. In 2014, 83% of public sector workers were a member of a defined benefit pension scheme compared to just 9% of private sector employees.

The new structure for public service pensions has considerable advantages over the previous system. The move to a career average basis is a sensible change; the previous final salary pension schemes were disproportionately beneficial to long-stayers who received pay increases towards the end of their career. For example, it is unclear why a teacher who subsequently goes on to become a headteacher should receive a more generous pension for the years in which they are working as a teacher than a teacher who does not subsequently become headteacher.

Aligning the NPA for employees of the same age will also end the practice where employees of the same age, in the same job, with the same headline pay accrue a different pension depending on whether they joined the scheme just before or just after Labour’s cut coming into effect. Finally, aligning NPAs with the SPA – which will mean that future rises in the SPA will automatically feed through into further increases in the NPA – is a coherent way for the government to help tackle the public finance challenge posed by rising longevity.

The NHS and Teachers’ Pension Scheme rules

  NHS Pension Scheme Teachers’ Pension Scheme
  Pre-Labour Labour's bequest New Pre-Labour Labour's bequest New
Normal Pension Age 60 65 SPA 60 65 SPA
Salary Final Final Career-average (revalued by CPI plus 1.5%) Final Final Career-average (revalued by CPI plus 1.6%)
Accrual 1/80th plus 3/80ths lump sum 1/60th 1/54th 1/80th plus 3/80ths lump sum 1/60th 1/57th
]]> Tue, 31 Mar 2015 00:00:00 +0000
<![CDATA[Would Labour increase taxes by over £3,000 for every working household?]]> The Conservative Party have claimed that under Labour there would be a £3,028 tax rise for every working household. This calculation assumes that Labour would increase taxes on working households by £7.5 billion in 2016–17 and £15 billion from 2017–18 onwards, with the £3,028 being the average tax rise cumulated over all years through to 2019–20.

What time period, which households?

The first point to note is that, on the basis of these figures, you get to an average £3,000 tax increase by (1) cumulating increases over four years – this is the average additional bill in total over four years, it is not an annual additional cost – and (2) dividing the total tax increase only by the number of working households not by the total number of households.

In a world in which taxes were to rise by £15 billion one would usually describe this as leaving households worse off by £560 a year – £15 billion divided by 26.7 million households.

Cumulating numbers like this over several years is, at best, unhelpful. Ignoring the existence of non-working households doesn’t help provide sensible averages either.

Which fiscal targets?

A more fundamental question to ask, though, is whether Labour would need to impose a tax rise amounting to £7.5 billion in 2016–17 and £15 billion from 2017–18 onwards to meet its commitments for reducing the deficit, assuming that the consolidation is split 50/50 between further tax rises and real spending cuts.

The Conservatives argue that this would be needed for Labour to comply with the Charter for Fiscal Responsibility which it voted for earlier this year in the House of Commons.

The Charter sets out two fiscal targets. First, that public sector net debt should be lower as a share of national income in 2016–17 than 2015–16. Second, that there should be a surplus on the structural current budget balance in the third year of the forecast horizon. This second rule means that, after adjusting for the estimated impact of the ups-and-downs of the economic cycle, total revenues should be sufficient to cover all of the government’s current spending: in other words any borrowing should be explained either by temporary weakness in the economy or spending on investment.

Let’s start with the more important, and sensible, of these targets, the target for structural current budget balance.

The latest forecasts for the structural overall deficit and the structural current budget deficit are shown in the figure below. The OBR’s forecast is that total public spending, less spending on debt interest, would be cut by £30.5 billion by 2017–18 and that this would be sufficient to deliver a current budget surplus of £16.3 billion. However, because some items of public spending – such as spending on public service pensions – is expected to rise the size of the discretionary cut to spending required to bring about this surplus is actually closer to £35 billion.

So on the face of it Labour might need a fiscal tightening of just over £18 billion by 2017–18 (the £35 billion implied by the Budget less the £16.3 billion of overachievement against the fiscal target that Labour would not actually need). Obviously, such a tightening – if half is to come from tax rises – would imply a net tax rise of around £9 billion in 2017–18 (and not the £15 billion the Conservatives suggest).

However, the target set out in the Charter for Fiscal Responsibility relates to the third year of the forecast horizon. While this is currently 2017–18, by the time of any post-election “emergency” Budget this would relate to 2018–19 (because the current financial year would be 2015–16 not 2014–15).

In that year, the Budget forecast is for a surplus on the structural current budget of £33.7 billion, brought about by a total real cut to departmental spending between 2015–16 and 2018–19 of almost £40 billion. In other words the total amount of consolidation needed beyond the cuts in 2015–16 (that Labour has signed up to) would be just £6 billion. Achieving this 50/50 through tax rises and spending cuts would imply a £3 billion tax rise from 2018–19 onwards (and not the £15 billion from 2017–18 onwards that the Conservative numbers suggest).

Latest OBR forecasts for structural borrowing


Debt target

The OBR’s latest forecasts suggest that public sector net debt will fall from 80.2% of national income in 2015–16 to 79.8% of national income in 2016–17. This assumes that there are no new net tax rises or welfare cuts but that departmental spending is cut in real terms by £18.8 billion in 2016–17. This takeaway could be reduced to just over £9 billion and debt would still be forecast to fall slightly as a share of national income. If done 50/50 through tax rises and spending cuts this would imply a £5 billion tax rise in 2016–17 (not the £7.5 billion the Conservatives suggest).

But as we have argued before this target for debt to be falling in a particular year has little to commend it.

In conclusion

It is also not entirely clear – at least to us – when Labour would want to achieve current budget balance. Their oft-stated goal is to eliminate the current budget deficit by, at the latest, the end of the parliament. If that’s all they want to achieve they may need no tax increases or real terms spending cuts – beyond those planned for 2015–16 – at all. But that is later than implied by their having signed up to the Charter for Budget Responsibility. If they take that commitment seriously then they at least need to aim to get to current budget balance by 2018–19. If that’s what they want then they will require about £6 billion of spending cuts or tax increases.

There is real uncertainty about what path the Labour party want to follow for the public finances. The Conservatives have been clearer about what they want to achieve, but they have not been clear about how they would achieve it. They would require substantially bigger spending cuts or tax increases than Labour.

There is little value in bandying around numbers which suggest either party would increases taxes by an average of £3,000 for each working household. We don’t know what they will do after the election. But neither of the two main parties has said anything to suggest that is what they are planning.

]]> Mon, 30 Mar 2015 00:00:00 +0000
<![CDATA[The search for further benefit cuts]]> The BBC has reported that civil servants and Conservative ministers have been holding discussions about possible cuts to benefits. Options discussed include cuts to child benefit, taxing certain disability benefits, and abolishing a number of non-means-tested benefits altogether, with lower-income claimants partly or wholly protected if they claim means-tested support instead.

It is not clear whether any of these reforms will become Conservative Party policy. All we know is that ministers have been discussing possible options with civil servants. Given the scale of benefit cuts the Chancellor has said he would aim for – £12 billion a year by 2017–18 – it is hardly surprising that such discussions have been taking place. The specific options reported by the BBC constitute a largely predictable list of the kind of policies that civil servants would be likely to put in front of ministers looking for benefit cuts – indeed many were discussed in a chapter of last month’s IFS Green Budget, which considered the options a future government would have if it wanted to cut benefits further.

However, the BBC’s report is a timely reminder that finding significant further savings would involve difficult decisions. The Conservatives’ figure of £12 billion represents around a 10% cut to the £125 billion spent on social security excluding the state pension and universal pensioner benefits (which the Prime Minister has pledged to protect). So far, they have outlined where less than £2 billion of these £12 billion of cuts would come from. The proposal to freeze most working age benefits in April 2016 and April 2017 would save little more than £1 billion per year (given current inflation forecasts). Other confirmed policies add little to this total: lowering the household benefits cap to £23,000 would save a further £150 million, and removing housing benefit entitlement for jobseeker’s allowance claimants aged 18 to 21 around £120 million. That leaves more than £10 billion of cuts still to find.

It is not just the Conservative Party that would face hard choices in this area – the plans set out by the Liberal Democrats last week would also require £3½ billion of cuts to social security by 2017–18. The Labour Party is not currently committed to overall cuts to social security spending.

The table below provides estimates for how much each of the potential reforms reported today might save a future government. If all of these were implemented, the total saving would be likely to fall well short of the missing £10 billion per year that the Conservatives intend to find by 2017–18 (particularly since, as discussed below, some of the savings would be unlikely to be fully realised for some years).

Table: Estimated long run annual saving from possible reforms

Limiting child benefit to two children per family

£1.0 billion (a)

Taxing disability living allowance and personal independence payment

£0.9 billion (a)

Taxing attendance allowance

£0.6 billion (a)

Abolishing contributory jobseeker’s allowance and employment and support allowance

£1.3 billion (b)

Abolishing carer’s allowance

£1.0 billion (b)

Replacing the industrial injuries compensation scheme with an employer-funded scheme

£1.0 billion (c)

Introducing regional household benefit caps

Depends on details

Incorporating council tax support in universal credit

Depends on details


a IFS Green Budget 2015.

b BBC reporting of civil service costings, which differ from those given in the IFS Green Budget. This is likely to be because the Green Budget figures do not account for non-take-up of means-tested benefits.

c DWP benefit expenditure tables.

What would each of these policies mean for the benefits system?

Limiting child benefit to two children per family

Child benefit was a universal benefit payable in respect of all children from its introduction in 1977 until the coalition government started withdrawing it from families in which at least one adult has a taxable income over £50,000. A further cut to child benefit that is reportedly being discussed is to cease payments in respect of the third and subsequent children in each family. In the long run this could save about £1.0 billion per year, with 1.2 million families losing an average of nearly £1000 per year. However, it seems likely that such a policy would apply only to new births or conceptions, so the full saving would not be realised until the 2030s – making it of little consequence in the context of the Conservatives’ ambition to find £12 billion of cuts to the annual benefits bill by 2017–18.

Taxing disability living allowance, personal independence payment and attendance allowance

These disability benefits are currently neither taxed nor means-tested, and so are worth the same cash amount to all recipients. There is an argument for keeping things this way, since the purpose of these benefits is to compensate those with disabilities for the additional costs they face. However, if a future government did want to make savings in this area, making these benefits taxable would reduce their generosity in a way that meant the individuals with the lowest incomes were protected and the highest-income recipients lost the most. We estimate that making disability living allowance and its successor personal independence payment taxable would increase revenues by around £900 million a year, and including attendance allowance (the equivalent benefit for pensioners) would boost revenues by a further £550 million a year.

Abolishing contributory jobseeker’s allowance (JSA) and employment and support allowance (ESA)

Entitlement to contributory JSA and ESA depends not on income but on one’s history of National Insurance contributions. They are remnants of Beveridge’s vision of social insurance, in a working-age social security system in which 80% of spending is now means-tested. This is an area where the coalition has made cuts, limiting the duration of contributory ESA claims (except for the most disabled claimants) to a year from April 2012. A total of £5 billion a year is now spent on contributory ESA and JSA: however, abolishing them would only save a fraction of that amount, since most recipients (those with low incomes) could claim an equivalent amount in means-tested benefits instead. The BBC report suggests the saving would be £1.3 billion a year, with over 300,000 families losing an average of £80 per week. This reform would represent another stage in the slow death of the contributory principle in the working-age benefits system.

Abolishing carer’s allowance

The government currently spends around £2.5 billion a year on carer’s allowance, a benefit for full-time carers. As with contributory JSA and ESA, abolishing the benefit would reduce overall spending by less than that amount as most recipients would be able to claim means-tested benefits instead. The BBC report suggests a saving of around £1 billion a year.

Replacing the industrial injuries compensation scheme with an employer-funded scheme

The government currently spends around £1 billion a year on benefits to compensate individuals with an illness or disability incurred through their work. The policy would be to abolish these benefits, with employers required to set up their own schemes. To the extent that employer-funded schemes replace the support currently available, the reform would shift costs from the taxpayer to businesses. To the extent that they don’t, the reform would reduce the incomes of recipients. It is not clear whether the transfer of costs from the government to employers would happen in full straightaway (e.g. whether the change would apply to existing claimants). Hence the short-term government saving is unclear.

Introducing regional household benefit caps

The current household benefit cap (of £26,000 a year) and the Conservatives’ proposed lower benefits cap (of £23,000 a year) are both much more likely to affect claimants in London than the rest of the UK, because of their higher rents (and consequently higher housing benefit entitlements). Introducing a household benefit cap that varied by region would thus allow a government to make the cap significantly more binding in other areas of the country without further large reductions in the benefit entitlements of those in London. Without details on how these regional caps would be set it is impossible to say how much (if at all) they would reduce spending.

Incorporating council tax support in universal credit

Bringing council tax support within universal credit would be a welcome change, and would in fact be a return to the original intention to incorporate it within universal credit along with the other main means-tested benefits. Keeping them separate threatens to undermine the simplification of the system and rationalisation of work incentives that universal credit promises. But there are several ways in which this could be done, and without further details it is impossible to say by how much (if at all) the move would reduce spending.


Several of the reforms mentioned in the BBC report follow a common theme. Taxing disability benefits hits higher-income claimants but not the poorest. On carer’s allowance, contributory JSA and contributory ESA, the policy being discussed is to abolish the non-means-tested benefit, with lower-income claimants again protected if they claim means-tested support instead. However, this would leave more people relying on means-testing, with the accompanying increased potential for hassle, stigma, non-take-up and a weakening of incentives to work and save. Limiting child benefit to two children per family is different: the losers from this would include many low-income families. Of course there are many other ways of reducing benefit spending: for example, the BBC report does not mention cuts to housing benefit or tax credits, which make up half of working-age social security spending.

Today’s BBC report may tell us little about how a future government would cut benefits. But it does illustrate the scale of the reforms that would be necessary to make significant savings. Even if all of the reforms discussed today were implemented, alongside confirmed Conservative party policies, the total saving would be likely to fall well short of the £12 billion per year that the Conservatives intend. But those reforms would also involve reducing entitlement to child benefit for over a million more families, taxing previously universal disability benefits, and a further erosion of the contributory principle for those of working-age. These may well not be the decisions that a future Conservative government would make. But it is likely they would have to make changes at least as radical as this to find £12 billion a year. We should be told what those changes would be.

]]> Fri, 27 Mar 2015 00:00:00 +0000
<![CDATA[There’s more to higher education funding than the RAB charge]]> Higher education funding has been a hotly-debated topic in recent times, with the RAB charge – the government subsidy inherent in the student loan system – a prominent focus of these debates. On the back of new research published today, researchers at the Institute for Fiscal Studies argue that this focus comes at the expense of a wider discussion around how much the government should subsidise the higher education system as a whole, and how best to deliver this.

The new report has re-examined the current funding system for undergraduate higher education, highlighting its uncertain public finance implications. It also provides new estimates of the financial consequences of a series of reforms to the system of undergraduate funding, which have been proposed by various interested parties, and offers a first look at the likely costs associated with several potential ways to implement the new postgraduate loan system, announced by the Chancellor in his final Autumn Statement of the parliament. 

The funding of undergraduate higher education works as follows. Universities receive tuition fees from students and teaching grants from the government. Tuition fees are capped at £9,000 per year. Students do not have to pay these fees up front; they are entitled to government-backed loans to cover the full cost of their tuition fees and a contribution towards their living expenses. Students from poorer families are also entitled to grants to help cover their living costs while at university. The loans taken out by students incur a real interest rate of RPI+3% whilst they are studying, and a real interest rate of between 0% and 3% once they have left university, depending on their income. Individuals do not start making repayments until their income rises above £21,000. They must repay 9% of their income above this threshold, and continue doing so until their loan is fully repaid or for 30 years, whichever comes first. Any outstanding debt that remains at this point is written off.

The overall cost to government of undergraduate higher education comprises two elements: the certain up-front costs associated with teaching and maintenance grants; and the uncertain long-run costs associated with issuing student loans (the RAB charge). The latter can, of course, only be estimated at this stage, as they will depend on student loan repayments for decades to come. Moreover, estimating these costs requires us to make a number of assumptions about things such as future earnings growth and the discount rate, which is a way of accounting for the fact that money paid in future is less valuable to the government than money paid now.

What is clear from our analysis is that the discount rate matters hugely when estimating the cost of higher education. In fact, it matters more than plausible changes to the rate of real earnings growth. Reducing the discount rate means valuing future repayments more highly; hence the estimated loan subsidy (and the RAB charge) falls. But note that nothing ‘real’ has changed. No additional repayments are being made; we have simply changed how highly we value these future repayments in the present.

The current debate has focused heavily on whether the RAB charge is ‘too high’. The fact that the RAB charge would fall from 43% with a discount rate of RPI+2.2% (the standard government assumption) to 30% with a discount rate of RPI+1.1% (the long-run median of real index-linked bond yields over the last decade or so) means that the assumption made about the discount rate is a key driver of this debate in a way that is perhaps not particularly illuminating given the important broader questions about the overall level of government subsidy.

Reforms that would reduce the cost to taxpayers

To the extent that there is a desire to reduce the government subsidy of higher education (and to do so by reducing the loan subsidy), the report explores the financial implications of several proposals that aim to do just that, including:

  • Freezing the income level above which repayments are made (currently £21,000) and the interest rate thresholds (currently £21,000 and £41,000) in nominal terms for seven years, and then uprating them in line with inflation (rather than earnings growth, as expected under the current system).

  • Making graduates liable for repayments based on their total income (rather than their income above £21,000) once their income passes this threshold.

  • A move to a ‘pseudo’ graduate tax, in which all graduates pay 9% of their income above £21,000 for 35 years (not just until they have covered the costs of their education).

The current system is progressive: higher-income graduates repay more than lower-income graduates. That has obvious consequences for attempts to reduce the cost to taxpayers.

Reforms that raise more money from graduates in total, whilst retaining the form of a genuine loan system in which graduates don’t pay back more than they borrowed, tend to hit lower- to middle-income graduates hardest. This is the case for the threshold freeze and total income proposals we consider.

The alternative is to try to extract repayments from the richest graduates that are bigger than the loans they originally took out, as would be the case under the ‘pseudo’ graduate tax we have modelled. Such systems would be very sensitive to the income and loan take-up behaviour of a relatively small number of individuals.

This trade-off is highlighted in Figure 1, which splits graduates into 10 equally-sized groups (deciles) on the basis of their lifetime earnings and illustrates how much we expect graduates in each decile to repay, on average, under each of these scenarios relative to the default 2012 system.


A reform that would reduce the cost to graduates

The report also explores the implications of the recently-announced Labour proposal, in which the cap on undergraduate tuition fees would fall from £9,000 per year to £6,000 per year (but university income would be replaced by teaching grants); the maximum interest rate incurred on student debt would rise from RPI+3% under the 2012 system to RPI+4%; and maintenance grants would increase.

The Labour proposal differs from the others in that it does not seek to reduce the cost of higher education borne by taxpayers but to reduce the cost borne by graduates. It would decrease the amount that is loaned out to students (and hence the long-run cost to government of issuing student loans), but substitutes this uncertain long-run cost with a known up-front contribution in the form of higher teaching grants. Because some graduates are expected to repay their loans in full under the current system, this policy would increase the overall taxpayer contribution to higher education. Moreover, because higher-income graduates are the most likely to repay their loans in full, the policy effectively benefits higher-income graduates the most.

Postgraduate loans

Finally, the report considers three proposals put forward for a postgraduate loan system of up to £10,000 per student:

  • The first is similar to the current undergraduate loan system, with repayments of 9% of income above £21,000, but with this threshold frozen in nominal terms for five years and with an interest rate of RPI+3% charged to all postgraduates (regardless of income).

  • The second assumes individuals repay 9% of their income between £15,000 and £21,000 per year.

  • The third assumes that graduates are liable for repayments based on their total income (rather than their income above £21,000) once their income passes this threshold.

Assuming postgraduate loan repayments are paid concurrently with undergraduate loan repayments, we estimate that all of these systems could operate at essentially zero cost to the government.

Moreover, these estimates are relatively impervious to what happens to real earnings growth or which graduates decided to take out a loan. This is because the vast majority of postgraduates are expected to repay their loans in full, and to do so relatively quickly, mostly because they would borrow considerably less than undergraduates. If the size of the postgraduate loan on offer were scaled up substantially, then the costs might start to escalate (and the uncertainty increase). This may be relevant when considering the design of the £25,000 loans for PhD students proposed in the Budget.

Summing up

The subsidy inherent in the student loan system now comprises a substantial proportion of the overall government subsidy of undergraduate higher education. It is thus perhaps inevitable that debates about whether the loan subsidy is too high will continue. But it is important to remember that this is only one (highly uncertain) element of government support for higher education, and having so much of the debate focused solely on how much the government is likely to subsidise student loans misses the wider issue of how much the government should be subsidising higher education full stop. There are several reasons why the government might want to subsidise higher education, including the fact that there are benefits for the economy beyond those realised by individual graduates, such as higher tax revenues, and having a more productive or healthier workforce. However, choosing the appropriate level of subsidy and targeting it efficiently is fraught with difficulty. A more open discussion about how much subsidy should be provided, for whom, and how much certainty we would like to have over the numbers would be welcome.

]]> Fri, 27 Mar 2015 00:00:00 +0000
<![CDATA[Cutting the deficit: five years down, four to go?]]> The UK is coming to the end of the fifth year of what is now planned to be an nine-year fiscal tightening. The fiscal consolidation is forecast to total £185 billion (in 2015–16 terms) and to be completed by 2018–19. This is £54 billion larger than was originally planned in the June 2010 Budget but £15 billion smaller than the plans set out in the December 2014 Autumn Statement.

In his first Budget, George Osborne outlined a plan to complete the fiscal consolidation by the end of next financial year (2015–16). However, worse-than-expected economic news over the last few years, coupled with the fact that he is now aiming to achieve a tighter fiscal position, has led to Mr Osborne increasing the size of the planned tax rises and spending cuts. As a result, even though Mr Osborne has implemented virtually all the policies he originally announced, the consolidation is not now expected to be complete until the end of 2018–19.

However, the latest plans (from Budget 2015) imply a somewhat smaller fiscal consolidation than was planned in December 2014, and one that finishes a year earlier than was then planned. This is in spite of the fact that the underlying outlook for the UK economy and public finances was little changed between December 2014 and March 2015 – in other words, this change in planned consolidation appears to reflect a change in Mr Osborne’s long-term economic thinking, rather than a response to new information.

By the end of this financial year, 60% of the total consolidation is expected to have been implemented. However, within this nearly all the tax increases and cuts to investment spending will have been implemented, while only a little over half of the cuts to non-investment, non-welfare spending will have been done. The result is that of the consolidation to date 82% has come from spending cuts and 18% has come from net tax rises, whereas for the consolidation to come the plans imply a mix of 98% spending cuts and 2% net tax rises. The figure below shows the size and timing of the currently planned fiscal consolidation as a share of national income. This shows the combined effect of tax and spending measures announced and implemented by the previous Labour government and the current coalition government since March 2008. Only a small amount of additional fiscal consolidation is planned for next year – amounting to 0.6% of national income – but the pace of consolidation is then expected to pick up again in 2016–17 and 2017–18.

Figure: Timing and composition of the fiscal consolidation

Figure: Timing and composition of the fiscal consolidation

Note: This Figure updates the numbers presented in Figure 1.6 of the 2015 IFS Green Budget to include the policy announcements made in Budget 2015. The Green Budget contains details of the methodology and sources used to construct this figure. The data underlying this figure can be downloaded here.

]]> Fri, 27 Mar 2015 00:00:00 +0000
<![CDATA[Scotland’s fiscal position: an update in light of the OBR’s March Forecasts]]> Perhaps unsurprisingly, the OBR downgraded forecasts for North Sea revenues for the next five years in its Economic and Fiscal Outlook published on the 18 March. In part, this reflects the direct impact of the fall in oil prices which have been the subject of much discussion in recent months. But other forecasting changes – like reductions in oil and gas production forecasts –, and the announcement of cuts to tax rates on profits from the North Sea, have also contributed to reductions in forecast revenues. This means, if anything, the reductions in forecast revenue are even more dramatic than was anticipated.

Table 1 shows that the OBR now forecasts North Sea revenues to average around £0.7 billion a year between 2015–16 and 2019–20, rather than the £2.6 billion a year it anticipated just a few months ago. It also shows the contribution of various factors to those forecast changes.

Table 1: Net fiscal balance (% of GDP), UK and Scotland, 2013–14 to 2015–16

Revenue forecasts and changes






December 2014 Forecast












Forecasting changes






Oil and gas prices


















Other modelling changes












Policy changes












March 2015 Forecast






Source: OBR March 2015 EFO

On average, over the next 5 years:

  • Falls in forecast oil and gas prices have reduced forecast revenues by £1bn a year;
  • Falls in forecast production have reduced forecast revenues by £1bn a year;
  • Other modelling changes have reduced forecast revenues by about £0.7 billion a year;
  • Tax cuts have reduced forecast revenues by about £0.3 billion a year;
  • And a fall in forecast (tax deductible) investment and operation expenditure offsets around £1.1 billion of these falls a year.

Last week we published projections for Scotland’s net fiscal balance based on the OBR’s December 2014 forecasts for revenues. How much of a difference do the new forecasts make to these projections? Table 2 shows the original and updated forecasts for both Scotland and the UK as a whole.

Table 2: Net fiscal balance (% of GDP), UK and Scotland, 2013–14 to 2015–16


Net fiscal balance



Projections based on various OBR forecasts







Dec 2014 forecasts




March 2015 forecasts












Dec 2014 forecasts




March 2015 forecasts




Source: GERS, 2013–14, OBR December 2014 EFO, OBR March 2015 EFO and author’s calculations.

Under our earlier projections based on the OBR’s December 2014 forecasts, Scotland’s North Sea revenues were projected to fall from around £4.0 billion in 2013–14 to around £1.8 billion in 2015–16. This would have given Scotland a deficit of around 8.0% of GDP in that year. Using the same methodology, the OBR’s March 2015 forecasts imply Scotland’s North Sea revenues will fall to around £0.6 billion in 2015–16. This would mean Scotland’s budget deficit would be 8.6% of GDP in that year.

In contrast, OBR forecasts for the UK as a whole are effectively unchanged – the budget deficit for the UK as a whole is still expected to be 4.0% of GDP in 2015–16, despite the reduction in North Sea revenue forecasts. Scotland’s projected deficit in 2015–16 is now 4.6% of GDP higher than that for the UK as a whole. In cash terms this is equivalent to a gap of £7.6 billion. Unless oil and gas revenues were to rebound, onshore revenues were to grow more quickly than in the rest of the UK, or government spending in Scotland were cut, a similar sized gap would remain in the years ahead.

Why do the forecasting changes affect Scotland and the UK as a whole so differently? There are two reasons.

First, and most importantly, because most North Sea revenues are estimated to come from the Scottish portion of the North Sea (84% in 2013–14), and because the onshore economy and tax-base of Scotland is much smaller than that of the UK as a whole, a fall in this revenue stream has a much larger impact on Scotland’s fiscal position. For instance, the reduction in forecast revenues in 2015–16 is equivalent to around 0.8% of Scottish GDP but only around 0.1% of GDP for the UK as a whole.

Second, is the fact that the fall in oil and gas prices may have a positive impact on the onshore economy (by reducing energy costs, for instance). For the UK as a whole, this positive impact on the onshore economy may be big enough to more than offset the direct impact of lower oil and gas prices on North Sea revenues (A recent report by PWC suggests this is the case). This seems much less likely to be the case for Scotland – which accounts for the majority of the UK’s North Sea output and revenues, but only a small (close to population) share of onshore output and revenues.

]]> Thu, 19 Mar 2015 00:00:00 +0000
<![CDATA[School funding increases in England targeted at most deprived and led to particularly large increases in non-teaching staff]]> School funding per pupil in England has increased substantially since the millennium. More dramatic than the average increase in funding was the increased focus on deprived schools over the 2000s. This trend started well before the pupil premium was introduced in 2010.

In new IFS research published today, we find that some of these additional resources were spent on hiring additional teachers. But a much larger amount went on higher numbers of teaching assistants, other non-teaching and non-staff expenditures.

Some of these changes were intended by policymakers at the time, but it is not clear they ever intended the scale of the change we see in terms of the numbers of teaching assistants or other non-teaching staff. Instead, we argue that the scale of the changes we observe is more likely to reflect rigidities when making staffing decisions, the flexibility of contracts and the timing of funding allocations. This work was funded as part of a grant from the Nuffield Foundation

Between 1999–00 and 2009–10, current or day-to-day spending per pupil in England increased by 5% per year, on average, in real-terms. It has continued to grow in real-terms since 2010 as the coalition has chosen to increase the current schools budget slightly in real terms, despite the deep cuts to spending on many other areas of public services (though it has chosen to make substantial real-term cuts to schools capital spending.   

Perhaps even more noteworthy is the fact that school funding in England has become much more targeted on the most deprived schools over time, as is demonstrated in the table below (here and throughout we only focus on funding for day-to-day or resource spending). At the end of the 1990s, average funding per pupil amongst the most deprived fifth of primary schools was around 17% higher than that in the least deprived fifth of primary schools. At 15% the difference between most and least deprived secondary schools was similar (based on dividing schools into quintiles in terms of the proportion of children eligible for free school meals). Between 1999–00 and 2012–13, funding per pupil rose much more strongly amongst the most deprived primary and secondary schools. As a result, funding per pupil in the most deprived primary and secondary schools was nearly 40% greater than in the least deprived ones in 2012–13, a substantial increase in the level of funds targeted at school deprivation. This increase in funding targeted at deprivation occurred both during the 2000s and after 2010 when the pupil premium was introduced. The pupil premium therefore represents a continuation of this long-run trend. 

School funding per pupil 1999–00 and 2012–13 by quintile of school deprivation (2012–13 prices)


Quintile of School Deprivation

Relative difference (most/least)


Least Deprived




Most Deprived

Primary Schools


Funding per pupil 1999–00







Funding per pupil 2012–13







Total real-terms change







Secondary Schools


Funding per pupil 1999–00







Funding per pupil 2012–13







Total real-terms change







Notes: School deprivation defined in terms of proportion of pupils eligible and registered for Free School Meals.

How has this additional money been spent? We find that a higher number of teachers per pupil and a higher real-terms cost of teachers (i.e. teachers’ salaries plus other employer costs) account for about 20-30% of the additional funding. These increases have not differed by school deprivation, except that the most deprived secondary schools have been slightly more inclined to spend the extra resource on additional teachers. A much larger proportion of the increase in funding per pupil across quintiles of social deprivation can be explained by increasing quantities of teaching-assistants and other staff per pupil (explaining about 40-44% of the increase in funding per pupil across primary schools and 31-38% across secondary schools). These factors also translated into larger amounts of increased spending for the most deprived schools. In addition to this, a substantial proportion of the increase in funding per pupil seems to have been reflected in higher expenditures on non-staffing inputs (such as Information and Communication Technology, energy, professional services and learning resources). Furthermore, some of the increasing difference in funding per pupil between the most and least deprived schools is not actually being reflected in differences in expenditure per pupil. In 2012–13, the most deprived secondary schools ran a surplus of about £260 per pupil compared with about £90 per pupil for the least deprived secondary schools (the differences for primary schools are much smaller).

What drove these changes and what lessons do they provide for the way schools make financial decisions? In the early 2000s, policymakers actively encouraged schools to make more use of non-teaching staff to release teacher time and enrich the experience of pupils. Changes in educational need may also have required increased use of teaching assistants (such as greater numbers of pupils with English as an Additional Language). A government consultation on developing the role of support staff in 2002 states schools were given sufficient funding to employ an extra 50,000 support staff over the course of the parliament. In reality, the number of non-teaching staff grew from 160,000 in 2000 to 270,000 in 2005 and to 360,000 by 2010 (all on a full-time equivalent basis). The number of teachers also grew, but by much less (growing from 400,000 in 2000 to reach 450,000 by 2010) and the ratio of teachers to teaching assistants fell from around 5:1 in 2000 to just over 2:1 by 2010. It is not clear whether policymakers ever intended a shift in the workforce on this scale. Furthermore, the recent international TALIS survey shows that schools in England are relatively unusual in their high reliance on non-teaching staff as compared with other countries.  

These shifts in the workforce are unlikely to have been driven by robust empirical evidence either, as little was available in the early 2000s. Indeed, the evidence that now exists suggests that teaching assistants have had a weak effect on pupil attainment (at best), though this could be due to poor training and deployment. We instead think that the main factors driving the scale of the change are the various rigidities schools face when making financial decisions and uncertainty over future funding allocations that encourages greater use of flexible inputs. Teachers must be employed on relatively inflexible contracts, are difficult to remove if funding was to decline and schools are not easily able to add an extra classroom. Other staff can be employed on relatively flexible and temporary contracts. Uncertainty over future funding allocations might also be driving schools' decisions to run surpluses and build up a precautionary balance. 

These findings are relevant to present policy debates in schools. First, when allocating extra funding to schools, policymakers should consider the rigidities schools face and what these might mean for resource decisions. Second, academies have more flexibility on pay and conditions of teachers than do other schools and all schools in England have been given more freedoms on teacher pay since September 2013. While academies have made limited use of these freedoms to date, where they have been used it will be important to understand whether they have led to different resource decisions being made, as well as how schools are making use of their new pay freedoms. Third, uncertainty on future funding allocations appears to sway resource decisions, both in terms of spending more on flexible inputs and encouraging precautionary savings. The current government has stated that it plans to reform the school funding system to make it simpler and rationalise allocations to schools and local authorities (and has already undertaken some reforms in this direction). However, there is significant uncertainty as to what reforms will be come in over the next few years. Such uncertainty seems likely to be encouraging schools to make greater use of flexible inputs. 

]]> Tue, 17 Mar 2015 00:00:00 +0000
<![CDATA[Taxes up, taxes down, but fundamental problems unaddressed]]> As part of its deficit reduction programme, the coalition government has made tax changes whose direct impact is to reduce borrowing by an estimated £16.4 billion in 2015–16. This net figure belies much larger changes, with £64.3 billion of tax rises being partly offset by £48.0 billion of tax cuts. But all this activity has done little to improve the structure of the tax system. As we set out in a new IFS Election Briefing Note, with funding from the Nuffield Foundation, the reforms introduced by the coalition have for the most part involved simply changing rates and thresholds with little attempt to address the fundamental structural deficiencies of the tax system. Plenty of challenges remain for whoever wins the election in May.

The biggest tax increases were implemented early in the Parliament: a rise in the main rate of VAT from 17.5% to 20%, a sharp reduction in the amount that can be saved in tax-privileged pensions, and a 1 percentage point increase in all rates of National Insurance contributions (NICs) that had been announced by the previous Labour government.

All of these exacerbate unwelcome distortions in the tax system. The first has increased the distortions created by the VAT system since no real attempt has been made to broaden the VAT base and so the difference between the treatment of different goods and services has grown.

The second has reduced the coherence of the taxation of pensions. A system which allows saving out of income before income tax, has no income tax on returns on funds in the pension as they accrue, and charges income tax on withdrawal has many attractive features. Widespread proposals(including those from the Labour party) to take the Government’s changes further, and potentially to restrict the rate of tax relief available threaten to undermine the more coherent parts of the system. There are better ways to increase tax revenues from pension taxation by reducing the excessive tax privileges associated with the tax-free lump sum and the fact that NICs are never charged on employer contributions.

The third, a rise in all rates of NICs, increases the existing incentive to shift the form in which income is taken away from earnings and towards capital income (for example, through setting up a company and taking income as dividends rather than earnings).

There have also been three big tax cuts: an increase in the income tax-free personal allowance, cuts to the main rate of corporation tax, and real-terms reductions in fuel duties.

Corporation tax has been cut, making the UK’s rate more internationally competitive, but its base continues to discourage investment and to favour using debt to finance it.

The different treatment of the income tax allowance and corresponding NICs thresholds makes little economic sense. It is hard to think of a good economic reason for wanting to take the low-paid out of income tax but not NICs, and the emphasis on income tax and neglect of NICs highlights the absurdity of continuing to have two similar but separate taxes, given that National Insurance is not a true social insurance scheme. Despite some promising rhetoric, there has been virtually no progress on integrating the two.

The cuts to fuel duties will no doubt have been welcomed by motorists. But the way in which they have been delivered, continually delaying and then finally dispensing with planned increases provides little evidence of planned reform. The unsuitability of fuel duties for tackling congestion – the biggest harm associated with driving – has not been addressed. The relationship between fuel burned and congestion caused is weak, and improving fuel efficiency and new technologies are making it ever weaker.

But the coalition’s changes to the tax system go far beyond these, with a large number of smaller measures constituting the bulk of its activity. Across the full range of taxes, the coalition’s reforms have changed rates and thresholds but have failed to tackle the underlying weaknesses in the system.

  • The additional rate of income tax has been reduced from 50% to 45%, yet nothing has been done to address the anomaly of the effective 60% income tax rate associated with the withdrawal of the personal allowance once income exceeds £100,000.

  • Council tax has been cut but allowed to get ever more out of date: we now find ourselves in the absurd position that tax bills in England and Scotland are still based on relative property prices in 1991.

  • Inheritance tax has been increased as the threshold has not kept pace with inflation, but no real attempt has been made to close the loopholes that allow many – particularly among the very wealthy – to sidestep the tax.

  • Capital gains tax has been increased but with no clear strategy for dealing with the tension between minimising disincentives to save and minimising avoidance opportunities.

  • Business rates have been cut but made more unstable and continue to discourage property-intensive production.

All in all the coalition’s changes represent a missed opportunity to improve the tax system. There have been some welcome structural reforms, but even there the job has often seemed incomplete. The jumps in stamp duty land tax (SDLT) liabilities at price thresholds have been removed for housing, but not for non-residential properties, and the more fundamental problem with SDLT has not been addressed: the effect of a transactions tax such as SDLT is to discourage mutually beneficial transactions, so that properties are not held by the people who value them most.

Meanwhile the use of the discredited retail prices index (RPI) to adjust the tax system for inflation has been ended for direct taxes, but not for indirect taxes. More problematically, an increasing number of thresholds in the tax system are not uprated at all. As income growth picks up, the number of people affected by the effective 60% income tax rate, the 45% income tax rate and the withdrawal of child benefit will increase rapidly. We estimate that this fiscal drag will lead to the number of families losing some of their child benefit doubling within a decade unless this lack of indexation is addressed.

One possible reason for optimism about the future is that the coalition has made some admirable improvements to the institutions of tax policy–making, enhancing transparency and allowing better scrutiny. That optimism should be tempered, though, by the way in which the coalition has announced tax policy itself. Some areas, such as fuel duties and business rates, have seen a stream of ad hoc, often temporary, announcements overtaking each other without a clear statement of principles or long-run intentions. Arguably, the ad hoc and inconsistent approach being taken to devolution of tax-setting powers to different parts of the UK creates similar uncertainty.

There is a better way to make tax policy. The corporate tax road map was a good start, setting out a direction of travel and providing an element of predictability. The next government would be well advised to apply this approach to more elements of the tax system, and indeed to the tax system as a whole. That would help taxpayers to plan, provide a benchmark for assessing the policies actually implemented, and facilitate debate on whether the strategy laid out is the right one. Taking the time to articulate a strategy might even lead to the adoption of better tax policy.

]]> Fri, 13 Mar 2015 00:00:00 +0000
<![CDATA[Promoting the social inclusion of the extreme poor at scale in developing countries: what do we know?]]> Poverty in Latin America has been greatly reduced over the past 20 years. However, around 15% of the population remained in extreme poverty in 2011 – defined as average daily consumption of $1.25 or less - and such households risk exclusion from the gains in living standards enjoyed by others, and from participating in society more widely1.

Policymakers in a number of countries have responded to this by implementing innovative umbrella programmes targeted at these households, which bring together a number of often pre-existing policies. An evaluation by IFS researchers of one such policy piloted in Colombia, however, suggests that it had no effect on either take up of social programmes or on labour market outcomes. Evaluation of a somewhat more intensive programme in Chile suggests that while it was more successful in promoting take up of social programmes it still had no effect on labour market outcomes. In fact the most successful programmes tend to be smaller scale and more targeted, often run by NGOs.

The difference in results looks likely to arise from the details of policy design. This provides more evidence of the importance of precise policy design, and especially of not assuming that the beneficial impacts of small resource intensive programmes will be replicated when similar programmes are implemented with less resource.

This type of umbrella programme aims to improve the living standards of the extreme poor along many dimensions, including housing, health and employment. The rationale is to tackle a range of different causes of poverty simultaneously by providing preferential access to a range of existing social services, while also improving the standard of these services. Social workers are then assigned to households to encourage the use of these programmes. In addition, these social workers help families to identify specific economic and social barriers to exiting extreme poverty, and provide guidance on how to overcome these constraints. These programmes account for an important share of social inclusion funding in Latin American countries, from 5% in Colombia in 2013 to 20% in Chile in 2014.

IFS researchers conducted an evaluation of a pilot of one such programme, Unidos, which was introduced in Colombia in 2009 and has since been rolled out nationally. This evaluation provides evidence on the impact of a large-scale pilot scheme, examining how the introduction of the programme has changed both the awareness and take-up of social programmes, and the labour market outcomes of households in extreme poverty.

The results suggest that the pilot scheme had no impact on these outcomes2. This is likely to be as a result of the fact that each social worker – the key delivery agents – was assigned as many as 150 families to work with and was thus unable to spend significant amounts of time with individual families. In other cases, social workers lacked sufficient training. Despite these negative results the programme has been subsequently rolled out nationally in a very similar manner.

Unidos in Colombia was inspired by an earlier programme introduced in Chile in 2002 under the name of Chile Solidario (CS). CS provides a greater number of home visits from social workers, and so potential benefits are likely to be greater. In addition, social worker quality is likely to be higher, the coordination of the supply side of social services seems to be stronger, and the programme also provides a small monetary grant to cover the costs of participating in CS. A recent evaluation of the programme, involving IFS researchers, suggests that CS had a positive impact on the take-up of the family allowance for poor children and the take up of employment programmes. Importantly, the take-up of programmes was mainly driven by households who were disconnected from the welfare system or were outside of the labour force before the intervention, highlighting its success in promoting social inclusion. However this was not accompanied by widespread improvements in housing or employment outcomes3.

Taken together, this evidence has policy implications that are important not only for Colombia and Chile, but for other developing countries where similar programmes are being introduced or considered. Unidos, in its current form, is unlikely to make a significant contribution to the reduction of extreme poverty in Colombia. The evidence from CS suggests that even a stronger version of Unidos is unlikely to have significant impacts in improving employment outcomes for the average member of their target population in the medium to long term.

Does this then imply that Unidos and CS should be scrapped? The answer is not so simple, and depends on the ultimate objective of the programme. CS partially achieves reductions in social exclusion, and therefore may provide a valuable impact. However, neither programme appears to succeed in alleviating extreme poverty. What then should be done to tackle this problem?

Some insights can be gained from successful smaller-scale interventions operated by non-government organisations, where existing evidence suggests that better targeted programmes are more successful at reducing poverty and promoting social inclusion. Such programmes are more focused than Unidos and CS in the sense that, rather than targeting a range of problems at once, they focus on the different causes of poverty one by one. Successful programmes include BRAC’s ‘Targeting the Ultra Poor’, started in Bangladesh in 2002 and subsequently exported to a range of other countries. A similar programme now is being implemented in Colombia through an NGO called Fundación Capital, and is using part of the Unidos infrastructure.

Despite the positive evidence associated with these more targeted interventions, incorporating any new programme into a national welfare system faces several challenges both from difficulties associated with scaling up an intervention from a small pilot scheme to the national level, and also from acquiring the approval and assistance of relevant government departments and existing institutions. This process can make it difficult to scale up programmes at a reasonable cost, even if they are well designed.

IFS researchers have designed and evaluated an intervention that uses the infrastructure of the existing Familias en Acción programme in Colombia to deliver a scalable, cost-effective and integrated early childhood programme through home visits. Experimental evidence from this evaluation suggests that this programme has been successful and this may provide some positive policy lessons in this area.

A coordinated effort among different national agencies administering a number of high-quality programmes (as opposed to an array of separate programmes) seems to be the right approach to tackling extreme poverty; but it needs to be managed and executed well. Ultimately, a further understanding of the impacts of these programmes, as well as the reforms they undergo, is crucial to decide whether programmes such as Unidos and Chile Solidario ultimately provide value for money, or whether they should instead be replaced in their entirety.

1. See, for instance, Cecchini, S. and R. Martinez (2012), Inclusive Social Protection in Latin America: a comprehensive, rights-based approach, Libros de la CEPAL No.111

2. A complementary analysis presented in a report to the Government suggested no consistent impact on a range of other outcomes such as housing, health and access to justice. See ‘Evaluación de Impacto de Juntos (hoy Unidos). Red de Protección Social para la Superación de la Pobreza Extrema’, Informe de Evaluación  Diciembre de 2011, by Fedesarrollo, Econometria, SEI and IFS.

3. Due the lack of administrative data about other social services or outcomes, Carneiro, Galasso and Ginja (2014) cannot say anything about a larger array of social programmes and services being made available.

]]> Thu, 12 Mar 2015 00:00:00 +0000
<![CDATA[How would you deal with the deficit?]]> As the Chancellor puts the finishing touches on what may be his last Budget, those outside the Treasury wait in anticipation of what new policies will be announced next week.

But what would you do in his place? Now you can be the Chancellor and set your own Budget thanks to a new online tool launched on the IFS election webpage today. The interactive webpage, complete with data visualisation and infographics, guides you through the challenging tradeoffs involved in decisions over how much to spend, how much to tax and how much to borrow. With data simulations showing you the effect on the finances of all the choices you make, now you can explore the choices facing Mr Osborne, and whoever may be the Chancellor after the election.

At the time of the Autumn Statement last December the Chancellor was planning to reduce annual borrowing over the next five years, and reach a surplus of 1.0% of national income by 2019–20. Given currently confirmed tax and benefit policies, and the official forecasts for the economy and the public finances, this would require cuts of 14.1% to real terms departmental spending between 2015–16 and 2019–20.

Next week the Chancellor may try to sweeten the electorate, giving some of that surplus away by announcing lower taxes or higher public spending. Or he may choose to stick to his borrowing plans, but reduce the squeeze on departments through further cuts to benefit spending. Or he may do something different.

While we will have to wait and see what George Osborne announces, why not explore what decisions you’d make if you were Chancellor.

Please note that this is a simple educational tool. The numbers you generate do not represent IFS analysis, and figures can differ slightly from published IFS analysis where we are able to use more sophisticated assumptions about the profile of any fiscal decisions made between 2015–16 and 2019–20.

]]> Thu, 12 Mar 2015 00:00:00 +0000
<![CDATA[Scotland’s fiscal position improves in 2013–14 but this is set to stall as oil price falls bite]]> Today, the Scottish Government published the latest version of its annual Government Expenditure and Revenues Scotland (GERS) publication covering 2013–14. In this observation we first discuss what the figures tell us about Scotland’s notional fiscal position in that year. The key finding is that Scotland’s overall budget deficit of 8.1% of GDP during 2013–14 was significantly higher than the UK wide deficit of 5.6% of GDP. This reflects the fact that disproportionately high oil revenues in Scotland are not sufficient to “pay for” higher public spending in Scotland.

We then project figures forward to 2015–16 over which period expected revenues from the North Sea have fallen further. Based on the Office for Budget Responsibility (OBR)’s December forecasts, we project Scotland’s deficit in 2014–15 and 2015–16 to be 8.6% of GDP and 8.0% of GDP, respectively, compared to 5.0% and 4.0% for the UK as a whole. The gap would likely be even larger if oil prices remained at current levels – which are significantly below those used in the OBR’s December forecasts – but could be smaller if oil prices or production rebound.

Finally, we place these figures in the context of the evolving constitutional debate – including the debate about whether Scotland should become “fully fiscally autonomous”.

Scotland’s fiscal position in 2013–14

Table 1 shows how Scotland and the UK’s net fiscal balance – which is the difference between government revenues and government spending (including investment spending) – evolved between 2009–10 and 2013–14. Figures are reported both excluding and including North Sea oil and gas revenues. 

Table 1: Net fiscal balance (% of GDP), UK and Scotland, 2009–10 to 2013–14

Net fiscal balance












   Excluding North Sea revenues






   Including geographic share












   Excluding North Sea revenues






   Including North Sea revenues






Source: GERS, 2013–14, and author’s calculations.

Excluding North Sea revenues, Scotland’s net fiscal balance was in deficit to the tune of 12.2% of GDP (or £16.4bn) in 2013–14. This represents a fairly sizeable reduction in the onshore fiscal deficit compared to the previous year, driven by government expenditure falling by 4.2% in real-terms.

Scotland’s onshore deficit is estimated to have fallen by more than that of the UK during 2013–14. This is despite the growth in onshore revenues being slower in Scotland than the UK as a whole; it instead reflects the fact the fall in government spending is estimated to have been larger in Scotland than in the rest of the UK. However, the level of the onshore deficit in Scotland remains around double that for the UK as a whole (6.0% of GDP) because government spending per person is much higher than the UK average, while onshore revenues are a little lower than the UK average.

Allocating a geographic share of North Sea revenues to Scotland unsurprisingly improves its fiscal position, although a large deficit of 8.1% of GDP remains. But these revenues did not help as much in 2013–14 as in earlier years, as declines in oil and gas production took their toll. In the two years between 2011–12 and 2013–14, Scotland’s North Sea revenues fell by more than half from £9.7bn to £4.0bn. This has driven Scotland’s overall net fiscal balance from 5.9% of GDP in deficit in 2011–12 to 8.1% of GDP in deficit in 2013–14, a period during which its onshore deficit shrank by a similar magnitude.

In contrast, the UK’s overall net fiscal deficit shrank from 6.9% to 5.6% of GDP over the same two-year period. Of course, the decline in North Sea revenues was not helpful to the UK public finances either. But, because most North Sea revenues are estimated to come from the Scottish portion of the North Sea (84% in 2013–14), and because the onshore economy and tax-base of Scotland is much smaller than that of the UK as a whole, a fall in this revenue stream has a much larger impact on Scotland’s fiscal position.

Projecting Scotland’s fiscal position for 2014–15 and beyond

The falls in North Sea revenues have continued during the current financial year. The current low price of oil, if sustained, would also lead to further declines in revenue in future years. With this in mind it is worthwhile examining the impact ongoing weakness of North Sea revenues may have on Scotland’s public finances in 2014–15 and 2015–16. We do this by projecting the figures in GERS forward using official OBR forecasts for the UK as a whole, and a number of additional assumptions (see notes below for further details).

Table 2 shows projections for Scotland and the UK’s net fiscal balance in 2014–15 and 2015–16 based on the latest OBR forecasts published in December 2014. This shows Scotland’s onshore budget deficit continuing to decline, from 12.2% of GDP to 10.3% of GDP in 2015–16, driven by growth in the economy and ongoing public spending cuts. However, under this projection, Scotland’s North Sea revenues would fall from around £4.0 billion in 2013–14 to around £1.8 billion in 2015–16. Such a decline in North Sea revenues would offset the projected improvement in the onshore fiscal balance, leaving Scotland’s overall fiscal deficit virtually the same in 2015–16 as in 2013–14 at 8.0% of GDP. In contrast, the UK’s overall fiscal deficit is forecast to decline from 5.6% of GDP to 4.0% of GDP during the same period. In cash terms, Scotland’s fiscal deficit in 2015–16 would be more than twice as high per person (around £2,600) as that in the UK as a whole (around £1,200). 

Table 2: Net fiscal balance (% of GDP), UK and Scotland, 2013–14 to 2015–16

 Net fiscal balance


Projections based on latest OBR forecasts







   Excluding North Sea revenues




   Including geographic share








   Excluding North Sea revenues




   Including North Sea revenues




Source: GERS, 2013–14, OBR December 2014 EFO and author’s calculations.

At the time the OBR made its forecasts, the oil price was $70 a barrel, and the OBR’s forecast was based on an assumed average price for 2015–16 of $83 a barrel. Since then, the oil price has fallen further, and futures markets currently have an average price of around $60 a barrel for oil to be delivered during 2015–16. Updated forecasts will be published next week alongside the UK government’s Budget but it seems likely that these oil price falls will result in lower revenues from the North Sea. For instance, a recent report by the UK parliament’s Scottish Affairs Committee suggests North Sea revenues for the UK as a whole may amount to £1.5 billion a year at an oil price of $60 a barrel. The direct effect of this would be to increase Scotland’s deficit in 2015–16 by around a further 0.3% of GDP; indirect effects on the wider economy in Scotland could be positive or negative.

The GERS figures in the context of the devolution debate

The figures described above represent Scotland’s notional fiscal position if it had to raise or borrow the money needed to pay for government spending undertaken in, or for the benefit of, Scotland. This is not the case at the moment though. Instead, most tax paid in Scotland goes to the UK government, which is responsible for defence, foreign affairs and for paying benefits and state pensions to those in Scotland. It also gives money as a block grant to the Scottish Government to pay for devolved services – like health and education. The size of this grant does not depend on how much tax revenue is raised in Scotland but is based on historic spending in Scotland, adjusted each year using the Barnett formula so that changes in spending broadly match changes in government spending in England. Scotland is therefore insulated from the fiscal implications of volatile North Sea revenues.

Under existing plans for further devolution, Scotland would be exposed to some revenue risks associated with its economy performing better or worse than that of the UK as a whole. This is because part of the Scottish Government’s block grant will be replaced with revenues from income tax raised in Scotland and a share of VAT raised in Scotland (and a number of smaller taxes). However, existing levels of funding would largely be maintained as the Barnett formula will remain in place, and as North Sea taxation is not being devolved, Scotland will remain insulated from the fiscal risk associated with these revenues.

It has been suggested that devolution could go much further, however, with the Scottish National Party calling for “full fiscal autonomy”. Under such an arrangement, all taxes and the vast majority of spending would be devolved to Scotland – with the Scottish Government making transfers to the UK government to cover things like defence, foreign affairs, and Scotland’s share of the UK’s debt interest payments. In that case the notional fiscal position set out in GERS and our projections would have direct implications. The Scottish Government would have to borrow if its spending were greater than its revenues. It would also have to bear the risk of volatile North Sea and other tax revenues.   

Our projections suggest that if Scotland were fiscally autonomous in 2015–16, its budget deficit would be around 4.0% of GDP higher than that of the UK as a whole. In cash terms, this is equivalent to a difference of around £6.6 billion. To put this in context, we project government spending in Scotland to be £68.8 billion in 2015–16, and onshore tax revenues to be £53.7 billion.

It has been suggested that the powers obtained under full fiscal autonomy would allow the Scottish Government to implement policies that would boost the growth rate of the Scottish economy, thereby improving its fiscal balance. This could be the case: full fiscal autonomy would give more freedom to pursue different, and perhaps better fiscal policy, and to undertake the radical, politically challenging reforms that could generate additional growth. There are undoubtedly areas where existing UK policy could be improved upon. But it is much easier to say things would be better if the economy grows more quickly than it is to develop and implement policies that would actually deliver that extra growth. The Scottish Government has previously suggested policies to boost growth – such as cuts to corporation tax and expanded childcare – but the immediate effect would be to weaken its finances; and it is not clear that even in the longer term the effects on growth would be enough to pay for such tax cuts and spending increases.

Today’s figures therefore illustrate that full fiscal autonomy would likely involve substantial spending cuts or tax rises in Scotland – unless oil revenues rebound and remain at consistently high levels, or credible policies to boost the growth of Scotland’s onshore economies and revenues can be developed.

Notes on methodology for projecting Scotland’s fiscal position beyond 2013–14

In order to project forward the GERS figures to 2014–15 and 2015–16 using figures from the OBR’s December 2014 Economic and Fiscal Outlook, the following method is used:

  • Spending is projected on the basis that government spending in Scotland remains the same proportion (9.2%) of UK-wide government spending as in 2013–14.
  • Onshore taxes are projected on the basis that the amount paid per person in Scotland grows in line with forecast growth in onshore revenues per person for the UK as a whole. This means onshore tax revenues per person in Scotland are projected to be 97.0% of the average for the UK as a whole, as in 2013–14.
  • Offshore (oil and gas) taxes are projected under the assumption that Scotland’s share of overall UK offshore tax revenues remains the same as in 2013–14 at 83.8%.

We have chosen the assumptions on the basis of their simplicity. As with any economic or fiscal forecast or projection, the projections outlined in this observation are subject to a number of sources of potential error that mean actual outturns may differ. This includes errors in the OBR forecasts for the UK as a whole; and trends in spending and government revenues in Scotland relative to the UK differing from the above assumptions. There are some reasons to suggest that, if anything, the assumptions are more likely to lead us to under-estimate rather than over-estimate Scotland’s fiscal deficit relative to that of the UK as a whole. First the workings of the Barnett formula and Scottish Government plans to borrow additional money to fund capital investment in 2015–16 mean that Scottish Government spending is set to fall less between 2013–14 and 2015–16 than equivalent spending in the rest of the UK. This would tend to increase Scotland’s share of overall UK government spending; in contrast, we have assumed this share would remain constant. Second, the OBR forecasts revenue growth to be particularly strong for taxes like capital gains tax and stamp duties, which make up a relatively smaller share of Scottish revenues. All else equal, this would tend to suggest growth in revenues per person in Scotland would be lower than for the UK as a whole. Third, while our revenue projections account for declines in oil revenues, our projections assume that GDP from the North Sea rises in line with onshore GDP. If North Sea GDP declined, as one might actually expect, then Scotland’s cash-terms deficit would represent a larger percentage of GDP. 

]]> Wed, 11 Mar 2015 00:00:00 +0000
<![CDATA[Labour’s proposed pensions takeaway]]> Ed Miliband and Ed Balls today announced that a Labour Government would significantly reduce the generosity of the income tax treatment of private pensions. Those with incomes above £150,000 a year would only be able to receive income tax relief at a rate of 20% (rather than the 50% marginal rate of income tax they would face under Labour), the annual pension contribution limit would be reduced by a quarter from £40,000 to £30,000 and the lifetime limit would be cut by one-fifth from £1.25m to £1m. They estimate the resulting tax increase would be sufficient to cover the around £2.7 billion increase in borrowing that Labour estimates would otherwise arise from their proposed reforms to the financing of higher education in England, including a reduction in maximum tuition fees from £9,000 a year to £6,000 a year (the reforms are discussed here).

As discussed in last year’s Green Budget a desirable benchmark for pensions taxation is a system where full tax relief is given up-front, returns in a pension are left free of personal taxation and income is taxed in full on receipt. Unfortunately Labour’s proposed tax increase would move the system further away from this benchmark.

The policy to reduce the rate of income tax relief for those with an income of more than £150,000 a year appears similar to one proposed by the last Labour Government in its 2009 Budget. At that time, they proposed restricting tax relief on pension contributions to the basic rate (20%), but only for those with incomes above £130,000 and whose gross income plus employer pension contributions was above £150,000 (see here for a discussion). The way the policy was designed meant that some with large employer pension contributions would face a substantial increase in their income tax bill if their income rose from just under to just above the £130,000 threshold: For example an individual earning £129,000 plus an employer pension contribution of £40,000 would face an increase in their annual income tax bill of over £10,000 if their current wage were to rise to £130,000 (assuming a top rate of income tax of 50%).

The policy was dropped by the incoming coalition government in favour of a reduction to annual and lifetime allowances designed to raise the same amount of money. At the time the Labour policy was expected to raise £3.6 billion (from just 300,000 individuals, implying an average (mean) tax increase for these people of £12,000 per year). However, the reductions in annual and lifetime allowances that have occurred since mean that such a policy would now be expected to raise far less than this.

Fundamentally the idea that income tax relief should be restricted to the basic rate is misguided. The error stems from looking at the tax treatment of pension contributions in isolation from the tax treatment of the pension income they finance. Pension contributions are excluded from taxable income precisely because pension income is taxed when it is received: in effect, the tax due on earnings paid into a pension is deferred until the money (plus any returns earned in the interim) is withdrawn from the fund. The tax system should treat pension contributions and pension income in a symmetric way: it is hard to see why it should be unfair for those above £150,000 to get tax relief at their marginal rate, but not for other higher-rate taxpayers to do so. Indeed, these very-highest-income individuals are less likely to be only basic-rate taxpayers in retirement, removing one of the principal (although still not well-founded) arguments for restricting relief.

The other reforms proposed today – further restricting annual and lifetime limits – are less incoherent, although still not the best way to reduce the generosity of the pensions tax system. They would be in keeping with recent reforms, repeating what was done in the June 2010 Budget and the 2012 Autumn Statement. The government estimates  that the reduction of the annual limit from £50,000 to £40,000 and the reduction of the lifetime limit from £1.5 million to £1.25 million implemented in April 2014 will together raise £1.1 billion in 2017–18 and more thereafter. Labour is proposing reductions of the same size, which should raise significantly more than £1.1 billion because far more people would be affected.

Reducing the annual allowance makes less sense than reducing the lifetime allowance. For a given level of lifetime contributions, it is not clear why we would want to penalise making occasional large contributions rather than frequent smaller contributions. In practical terms, reducing the annual allowance is more problematic, as valuing annual contributions to defined benefit pension schemes is difficult; the lower the annual limit, the more of these difficult valuations that must be done.

But the further you go down the route of cutting the lifetime limit, the more you move away from the relatively desirable system of taxation where full relief is given up-front, returns in a pension are left free of personal taxation and income is taxed in full on receipt. This would increase the risks that people will be incentivised to undersave for retirement and that more effort will be put into securing tax advantages by using more complex schemes.

If the system of pensions taxation is to be made less generous, then it would be better to tackle the two elements of the system that look generous relative to the aforementioned benchmark. First, up to one-quarter of an accumulated pension can be taken tax-free. This means that even under Labour’s proposed £1 million lifetime limit some could receive £250,000 that had escaped income tax altogether: it would be taxed neither when it was earned nor when it was withdrawn from the pension. Second, roughly three-quarters of pension contributions – those made by employers – escape National Insurance contributions (NICs) entirely. The official estimate of the cost of this NICs relief is a whopping £14 billion in 2013–14.

If Labour’s reforms are implemented, then – like the reforms to pensions taxation implemented by the current government – they would add further undesirable complexity and be a missed opportunity to rationalise those parts of the pensions tax system that are overly generous.

]]> Fri, 27 Feb 2015 00:00:00 +0000
<![CDATA[£8 billion giveaway used to boost corporate tax competitiveness]]> Corporate tax has rarely received as much attention, either from policy makers or the public, as in recent years. The coalition government has enacted a series of policy changes including reductions in the main and small profits rates, changes to capital allowances, the introduction of a preferential rate for patent income (Patent Box), changes to rules concerning the taxation of foreign income, and a raft of anti-avoidance measures. The explicit, and fulfilled, aim was to increase the competitiveness of the UK corporate tax system.

The net cost of the package of onshore measures is almost £8 billion in 2015–16, equivalent to a substantial 16.5% of real pre-crisis (2007–08) onshore corporation tax receipts. This net giveaway contrasts with an overall net takeaway from tax measures of £16.4 billion. In terms of individual tax measures, only increases in the personal allowance (at £8.0 billion) represent a bigger giveaway than the £7.6 billion cost of cutting the main corporation tax rate by 8 percentage points. We examine corporation tax policy changes and their effects in a new Election Briefing Note,which is published today as part of the IFS election 2015 analysis, funded by the Nuffield Foundation.

It has long been recognised that corporate income taxes can distort incentives in a number of harmful ways, and they are thought to have a particularly damaging effect on economic growth. The income and activities of multinationals are particularly mobile and responsive to the tax rate. A lower rate works to reduce the impact of distortions and increase the attractiveness of the UK as a location for investment. The 20% headline rate that will come into effect in April is more internationally competitive than the 28% rate inherited in 2010; it will be the joint lowest in the G20 and the joint 6th lowest in the OECD (compared to the 9th lowest and 23rd lowest respectively in 2010 ). The introduction of the Patent Box also makes the UK system more appealing for certain kinds of mobile income, while changes to controlled foreign companies rules (as well as the previous move to an exemption system for foreign source income) have made the UK more attractive as a location for multinationals to headquarter in.

However, the UK’s corporate tax base continues to embed a number of distortions – including a bias in favour of debt financing – and offers a set of capital allowances that are ungenerous by international standards, and that have been made less generous by the current government. Taken together, the changes since 2010 work to reduce the tax burden significantly for profitable companies and those that are internationally mobile, while expansions of the tax base disproportionately harm firms that invest heavily in plant and machinery.

One concern may be over the distributional consequences of the cuts to corporation tax – that is, which groups are likely to benefit most. Corporate tax cuts may be seen as a ‘tax break for big business’. However, it should be noted that the burden of taxes is ultimately on people rather than companies and the burden of corporation tax is not necessarily borne by shareholders (through lower dividends). It can also be borne by workers (in the form of lower wages or employment) or by consumers (in the form of higher prices). Indeed, the relative immobility of labour means that we might expect a substantial share of the burden of corporate tax to be on workers. As such, lower corporate taxes may feed through into higher wages in the medium term.

Corporate tax avoidance has been the subject of considerable public and media criticism in recent years. This has included a number of high-profile media case studies that ‘name and shame’ specific multinational companies. While reforming the corporate tax system to help entice multinationals (and their income) into the UK, the government has at the same time sought to crack down on avoidance activities. In 2015–16 the government expects all corporate tax anti-avoidance measures announced since 2010 to raise a little over £1 billion. This includes expected revenues of £270 million (and £350 million a year thereafter) – less than 1% of corporation tax receipts – from the new diverted profits tax (popularly dubbed the “Google tax”). The total revenue lost through corporate tax avoidance, though presumed large, is uncertain. These measures, which raise relatively little revenue going forward, reflect the government’s attempts to unilaterally reduce avoidance activity.

Multinational tax avoidance is a problem best tackled through international cooperation. The coalition government has supported the OECD’s Base Erosion and Profit Shifting (BEPS) project, which is considering what actions can be taken multilaterally to reduce avoidance opportunities and will conclude at the end of 2015. BEPS is an impressive project that is tackling a broad range of issues on a timescale designed to take advantage of political momentum. The BEPS process is effectively seeking to ‘patch up’ the current system rather than provide any fundamental reform. However, as such reform appears unlikely, especially in the near future, the BEPS process is a sensible development.

Despite a flurry of activity in the area of corporate tax policy over this parliament, challenges remain. In our briefing note we discuss: the need to maintain a competitive tax system in an international environment (where the concept of a “competitive tax system” is a moving target); the continued challenge of tackling avoidance (including any recommendations that arise from the BEPS process); the desirability of reforming the tax base to reduce economic distortions. We also discuss issues raised by the decision to devolve the setting of corporation tax to Northern Ireland from April 2017.

Corporate tax is a complex policy area in an evolving international context; any future changes to corporate tax will likely require balancing the desire to be competitive with the aim of cooperating with international efforts to make the system work better for all countries. The policies that one country views as creating a competitive system may be viewed by another as creating an opportunity for avoidance.

]]> Thu, 26 Feb 2015 00:00:00 +0000
<![CDATA[Today’s young adults are much less likely to own a home than their parents’ generation, but those owning a home well before the crisis have gained from house price increases and a sharp fall in housing costs]]> The UK has witnessed a substantial and sustained increase in house prices since the 1990s. This long-term rise in house prices, and the financial crisis which led to falls in real incomes and reduced availability of high loan-to-value mortgages, are perhaps the main causes of widespread concern about the ‘affordability’ of housing. In a new election briefing note funded by the Nuffield Foundation and published today, IFS researchers set out a range of evidence on how the affordability of housing, and – related to that – the kinds of housing that people access, have been evolving. 

The real price of purchasing a typical house, as measured by the Nationwide house price index, nearly trebled between 1995–96 and 2007–08, taking it 77% higher than its previous peak in 1989. When thinking about affordability, we of course want to account for changes in incomes too. At their peak in 2007–08, average house prices were 7.6 times average net household incomes – significantly higher than the ratio of 6.4 at the end of the previous house price boom in 1989. But real house prices then fell by nearly a quarter between 2007–08 and 2012–13. This left the price to net household income ratio at 6.2 – back to around its early 2000s level and close to its level at the end of the 1980s boom. Significant real growth in house prices resumed from 2013–14. 

When considered in combination with the lesser availability of mortgage credit since the crisis, it is not surprising that there is concern over the ease of access to homeownership. And in London, house price to income ratios have bounced back substantially more quickly since the financial crisis. Indeed, if we use pre-tax average earnings (for which, unlike net household income, we have data beyond 2012–13), we see that in 2013–14 the price to earnings ratio in London had already surpassed its pre-crisis peak. 

House prices are just one part of the story though. When thinking about the cost or affordability of housing, we should distinguish carefully between the costs of coming to own a property (which depend largely on house prices relative to families’ resources) and the ongoing regular costs associated with the consumption of housing – most importantly, mortgage interest for owner-occupiers and rents for renters. Although there are economic reasons why these costs are all related, it is possible for them to move in very different ways in the short run. Indeed, trends in these different costs have diverged emphatically since the recession.

Between 2007–08 and 2012–13, owner-occupiers with a mortgage saw a 38% real-terms fall in their real housing costs, due to large falls in interest rates. (Note that these housing costs include mortgage interest but not capital repayments, which are loan repayments and hence increase net wealth.) Their real incomes fell over this period too, but as a proportion of income their housing costs still fell from 16% to 10%. In that sense, for this particular group, the ‘affordability’ of housing has clearly been very forgiving of late. This is likely to be a key reason why the number of repossessions in the recent recession was so much lower than in the previous one. Although it has been getting harder to buy a house for the first time, it has been relatively easy for those already owning one to hold onto it. 

Meanwhile, costs for renters have not been offering anything like this scale of relief. Housing costs as a proportion of income rose for renters between 2007­–08 and 2012–13, from 25% to 27%. Even here, though, the story is perhaps a little more benign than popularly perceived. Average housing costs for private sector tenants have actually risen little faster than other (non-housing) prices since 2007–08.

Perhaps the biggest issue here is a distributional one – different groups have been affected very differently by housing trends. Changes in the regular costs of housing consumption have been very favourable for the group of people who already own a home relative to the group who don’t. Meanwhile, high house prices and tighter credit conditions make it harder to move from the latter group into the former. And of course all this follows large (and tax-free) capital gains over an extended period for a group of owner-occupiers who bought at much lower prices than it would be possible to find today.

Much of this is of particular relevance when thinking about the distribution of resources between generations. There has been a striking decline in homeownership that largely reflects differences across successive generations. The homeownership rate at age 35 of those born in the mid-1970s was 10 percentage points lower than it was for those born in the mid-1960s at the same age; the age-25 homeownership rate has halved in 20 years. There is also evidence of divergence in housing characteristics between generations. Couples aged 60 and over saw an 11% increase in floor space per person between 1996 and 2012, and single people aged 60 and over saw an 8% increase. Meanwhile, for most working-age household types, floor space per person was flat or falling. 

But there are potentially important issues here for inequality within generations as well. Young homeowners have actually done particularly well out of low interest rates – because they have the most outstanding mortgage debt, interest rates make the most difference to them – while a growing number of young adults remain as renters. And much of the housing wealth currently held by older generations could eventually come to be owned by some portion of the younger generation via inheritances passed on at death or through financial assistance from parents to children.

In summary, then, there are actually some respects in which recent trends in the affordability of housing look rather less troubling than some of the debate might suggest; but the picture varies greatly between owners and renters, with implications for the distribution of resources within and between generations.

What are the policy challenges? We argue in the briefing note that there is reason to think that historically high house prices relative to incomes partly reflect a mismatch between supply and demand, and policies that look to address this are likely to be important. There is also a strong case for reform to the taxation of housing (as discussed in the Mirrlees Review). As it happens, one of the most efficiency-enhancing tax reforms that could be made would also help credit-constrained first-time buyers: namely, abolishing stamp duty land tax – an extremely damaging tax which discourages mutually beneficial housing transactions. If the government wanted to replace the revenue, a sensible way of doing so would be through a revalued and reformed council tax. Finally, there are challenges for the public finances. A future government will need to take a view on whether it would be prepared to earmark increasing amounts of its budget to housing benefit if rents continue to rise over the long term and/or the growth of the private rented sector continues.

]]> Thu, 19 Feb 2015 00:00:00 +0000
<![CDATA[Health spending protected by more in England, but social services spending protected more in Wales]]> Health is a devolved matter: the UK government decides how to organise healthcare services in England and how much to spend on them, while healthcare in Wales is organised by the Welsh Government, which decides how much of the block grant given to it by Westminster to allocate to health, and how much to allocate to other services. Social care services are devolved to local authorities in both England and Wales and are funded by a mix of grants from central government and revenue raised locally (via, for instance, council tax).

Recent months have seen significant debate about the relative performance of – and the relative funding for – the health services of England and Wales. The Prime Minister has repeatedly raised the issue in the House of Commons, and recently claimed that while health spending had increased in England, that the health budget in Wales had been cut by 8%. The Welsh Government has disputed this figure, arguing that it has protected social services spending in a way England has not done.

In this Observation, we assess the empirical evidence on what has happened to health and social services spending in England and Wales since 2010–11. We find that the relative degree of protection offered to these services does differ between the two countries.

Health and social care spending in England and Wales: 2010–11 to 2014–15

Table 1 shows how much the UK government spent on healthcare in England and the Welsh government spent on healthcare in Wales in real-terms in 2010–11, and how much they plan to spend in the current financial year, 2014–15. It also shows how much local authorities spent (or plan to spend) on social services in England and Wales in the same years. All figures refer to resource expenditure – spending on the day-to-day provision of services – and exclude capital expenditure. (We exclude capital expenditure as figures are not available on a consistent basis for social services capital expenditure for England and Wales).

Table 1. Health and social services resource spending in England and Wales, £ billion (2015–16 prices)


Real-terms expenditure (£s billions)

% change












Social Services
















Social Services




Sources: English health spending is taken from PESA 2014, Table 1, as item “NHS (health)”. English social services spending is taken from local government revenue expenditure outturns and budgets, and is adjusted for changes in classification of social services spending, such as including in 2010–11 spending then funded via the Department of Health but now part of the regular social services budget, and transfers between the Education and Social Services components of local government spending. Welsh health spending is taken from the Welsh Government’s second supplementary budgets for 2010–11 and 2014–15, and excludes spending on the drugs misuse strategy. Welsh social services spending is taken from local government revenue expenditure outturns and budgets, and is adjusted for shifts between the education and social services portfolios. Full details are available from the author on request.

Notes: Percentage changes are calculated using unrounded numbers

The figures show that while spending on health services in England has increased by 4.3%, spending on health services in Wales has been reduced by 2.0%. Turning to social services spending, a different pattern is found. While spending in Wales has been cut by just 0.8%, spending in England has been reduced by 11.5%.

These patterns reflect the differing decisions of the Welsh and UK governments, and of local authorities in the two countries. They illustrate the trade-offs involved when trying to protect particular services from budget cuts. In the case of England, the UK government’s real-terms increases for the health budget have meant that cuts to other “unprotected” departments need to be larger. One of those unprotected departments facing larger cuts has been the Department for Communities and Local Government which provides grants to local authorities in England. Significant cuts in these grants, and the widespread freezes to council tax in England, have led to a substantial reduction in the spending power of councils and may have made it harder for local authorities to protect social services spending. In contrast, the somewhat lesser protection given to health by the Welsh Government has allowed smaller cuts to other areas, like grants to Welsh local authorities. Alongside council tax rises, this may have made it relatively easier for Welsh local authorities to protect social service spending.

It is worth noting that these differences largely reflect decisions taken in the first couple of years of the cuts: 2011–12 and 2012–13. During this period, the Welsh Government cut the resource budget for health by 3.5%, perhaps seeking to avoid the large cuts to unprotected services that protection of health spending in England had required. Starting in 2013–14, however, the Welsh Government changed tack, and has found additional money for health by cutting more deeply elsewhere. This has partially (but not fully) reversed the earlier cuts to health spending.

It is also worth asking whether there is a reason why the Prime Minister claimed health spending in Wales was being cut by 8% when we have calculated the cuts to be 2%.

The main reason seems to be that when the Prime Minister’s claim was made (on 7 January), the Welsh Government had yet to publish its second supplementary budget for 2014–15. In its first supplementary budget (then, the most up-to-date budget available), the Welsh Government had not allocated all its money for the year. It always keeps some back to be allocated in the second supplementary budget, and for several years at least, much of that money has then been allocated to health. The first supplementary budget showed a resource budget for health in 2014–15 that would have been 5.9% lower than in 2010–11, and an overall budget including capital spending 7.7% lower than in 2010–11 (a figure very close to the 8% cut claimed by the Prime Minister). Since then, the additional £250 million of resource funding allocated to health in the second supplementary budget has reduced the scale of the cut to resource spending to the 2.0% reported in Table 1 (and together with an extra £15 million of capital funding, reduced the cut to the overall budget including capital spending to 3.9%).


Devolution means that the Welsh Government has freedom to decide how to allocate the block grant it gets from Westminster across different service areas. Since 2010–11 the Welsh Government has used these powers to allocate cuts across services in a different way to the UK government in England. In particular, the Welsh Government has offered less protection to the health budget than is the case in England. But this has meant smaller cuts to other areas of government spending in Wales than in England. One such area is social services.

]]> Wed, 18 Feb 2015 00:00:00 +0000
<![CDATA[Substantial cuts made, but biggest changes to the benefit system yet to come]]> The coalition government has implemented changes to the benefit system that mean spending in 2015–16 will be £16.7 billion (7%) lower than it would otherwise have been. Real terms benefit spending, however, is forecast to be almost exactly the same in 2015–16 as it was in 2010–11, at £220 billion. This reflects the effect of underlying economic and demographic factors which are pushing up spending – most importantly an ageing population, but also weak wage growth and rising private rents. At the same time, the government has set out on a path towards radical reform of some parts of the system. But most of the major structural changes, such as universal credit, have run into problems, and are yet to be delivered. These are among the findings of a new Election Briefing Note on the coalition’s reforms to the benefit system, part of a programme of work at the IFS in the run up to the election, funded by the Nuffield Foundation.

Of course there have been some controversial benefit cuts. Cuts to housing benefit for social housing tenants (variously dubbed ‘the removal of the spare room subsidy’, or ‘the bedroom tax’), and the household benefits cap (which limits payments for most non-working families to £26,000 a year) have been particularly prominent in the public debate. Perhaps this is because they hit relatively small groups of people relatively hard – just 27,000 families are actually subject to the welfare cap, but each loses £70 a week on average. But the amounts saved from these changes (£650 million) are small in the context of the overall cuts (£16.7 billion). Proposals for a further cut in the household benefits cap to £23,000 a year – highlighted again this week - would likewise only reduce benefit spending by a further £150 million.

In fact, the biggest cuts have come from seemingly less controversial broad-based changes to benefits that affect large numbers of working-age claimants. Over half of the cuts (£9 billion worth) have come from changes in how benefits are increased each year. Increasing benefits in line with CPI rather than (the now discredited) RPI (or Rossi) since April 2011 will save £4.3 billion in 2015–16. Cash freezes to child benefit and parts of working tax credit, and the 1% increases in most working age benefits for three years, are forecast to save a further £4.7 billion in 2015–16.

There have also been other big cuts to tax credits (£3.9 billion), and child benefit; withdrawing the latter from families where someone has a taxable income over £50,000 has reduced spending by £1.9 billion. Cuts to private sector housing benefit (£1.8 billion) are also substantially larger than those affecting social sector tenants even though a majority (60%) of housing benefit expenditure goes to social housing tenants.

But despite these cuts, the reforms implemented so far largely represent an evolution of the system, rather than the revolution that was promised. All the main benefits are still in place. And although most benefits are less generous than in 2010 (also shown in recent work by CASE), the cuts only partially reverse the increases in benefits and tax credits for low- and middle-income families with children and pensioners under Labour: such families remain better off compared with an ‘unreformed’ 1997 tax and benefit system (although the same is not true for low-income working age adults without children who did not gain under Labour’s benefit reforms, and have faced subsequent cuts) – see Figure 3.5 in our recent Briefing Note on the distributional impact of tax and benefit reforms.

So what of the revolution? The introduction of universal credit - the replacement of a raft of means-tested benefits and tax credits with a new universal credit (UC) – is years behind schedule. By now, all new claims were meant to be for UC and the transfer of existing benefit claimants to UC was meant to be well under way. Instead, problems with project management and IT systems mean it is available to only some new claimants in a few parts of the country. Wider roll out is expected in the near future, but UC will still not be fully rolled out in April 2020. Implementing the coalition government’s flagship welfare reform will therefore fall largely to the next government and the government after that (if they choose to stick with it).

Changes to the disability benefit system have also been rolled out much more slowly than planned. In particular the new more stringent tests for employment and support allowance (ESA) to people already claiming support, and the replacement of disability living allowance with “personal independence payments” (PIPs), have run into problems. And fewer claimants have been found ineligible than originally expected – in part, because of successful appeals against initial decisions. Such issues and delays mean these reforms are saving much less money than hoped by now – in the case of PIPs, £1 billion less.

Beyond rolling out UC and PIP, the stage is also set for further change.

The link between housing benefit and current rents in the private sector has largely been broken. Instead the amount that can be claimed now depends, in part, on historic levels of local rents, meaning geographical relativities in housing benefit payments in 2050, for instance, will depend upon geographical differences in rent levels in 2012. This does not seem sensible and further reform looks warranted.

Child benefit, in one of the most radical structural changes, is no longer a universal benefit. But the current way in which it is withdrawn from families where the highest income individual has a taxable income over £50,000 leaves it in a strange sort of limbo, and out of step with the rest of the family-level earnings tests used elsewhere in the benefits system. That will surely require further attention.

The next government would also do well to think clearly about how benefits should be indexed over time. Moving away from the flawed RPI was sensible. But there have been various ad hoc and temporary deviations from standard indexation. For instance, most working age benefits have been subject to something close to a ’reverse double lock’. That is the government has justified below inflation increases on the grounds that earnings have been rising less quickly than prices. To continue on that route would mean that working age benefits rise less quickly than both earnings and prices over time – which does not seem a sustainable long term policy.

In contrast, in addition to having been largely protected from the cuts, pensioners have benefited from the ‘triple lock’ – the basic state pension now goes up by the highest of inflation, average earnings growth or 2.5%. When compared to the plans inherited from Labour to index to earnings from April 2012 onwards, this will increase spending on the state pension by £4.6 billion in 2015–16. This reflects the weakness of earnings growth. But perhaps it is unlikely that any government would have shifted to earnings indexation with earnings so weak. Compared to indexing to CPI inflation (now the default for most other benefits), the triple lock will cost a more modest £1.1 billion in 2015–16. Even so, the ‘triple lock’ could be very costly in the long-term as the state pension will go up more quickly than both prices and earnings, which again may not be sustainable.

Looking ahead, the next government faces the difficult decision of whether, and if so how, to make further cuts to benefits as part of continuing efforts to reduce the budget deficit. Identifying further cuts will be a challenge – especially if pensioners are again largely protected. 

]]> Wed, 28 Jan 2015 00:00:00 +0000
<![CDATA[The Smith Commission’s Proposals – big issues remain to be resolved]]> On November 27th 2014, the Smith Commission published proposals for further devolution of powers to Scotland. We now know what is to be devolved – the UK and Scottish Government now have the more prosaic task of implementing the changes. Getting the details of how the taxes and welfare are devolved will be crucial. A new Briefing Note published today, with funding from the ESRC through Centre for Microeconomic Analysis of Public Policy at IFS, analyses some of these ‘technical’ issues (and critically appraises the Smith Commission proposals more generally). In it we suggest a solution to one of the most difficult issues the Commission did not tackle – how to adjust the block grant given to Scotland when more taxes and spending are devolved. We also question some of the recommendations of the Commission – arguing that implementing them in practice might not always be feasible or fair. This observation provides a summary of these “big issues”.  

Adjusting the block grant to account for further devolution

When a tax is devolved to Scotland, and the Scottish Government gets to keep the revenues raised, a reduction should be made to the block grant Scotland currently receives. Similarly, if extra spending responsibilities are devolved, then Scotland should receive additional money to account for that.

Implementing this in the first year is relatively easy. When a tax is devolved, the block grant should be reduced by the amount of revenue that is being transferred to Scotland. When further spending powers are devolved, the block grant should be increased by how much the UK would have spent on that area in Scotland.

More difficult is determining what should happen to these adjustments in subsequent years. Keeping them fixed is typically not appropriate because inflation and economic growth mean that the amount raised from a tax or spent on a particular area tends to grow over time. The Smith Commission recognises this, by stating that these block grant reductions or additions should be “indexed appropriately”. But what does this rather cryptic phrase mean?

Unfortunately, the answer won’t be the same for every area of tax or spending. That is one reason why there is a lot of work for policy-makers and analysts still to do.

One attractive option is to index the block grant reduction to what happens to revenues from the equivalent tax in the rest of the UK. This would insulate the Scottish Government’s budget from revenue or spending shocks that hit the whole of the UK – in line with the principles agreed by the Smith Commission –, but still give the Scottish Government the incentive to grow revenues and limit expenditure in Scotland.

Revenues from devolved taxes, and spending on devolved areas, may evolve differently from comparable items in the rest of the UK for reasons completely unrelated to devolved government policy. Whether it is deemed appropriate for the Scottish Government to bear these risks depends on the importance of redistribution and risk-sharing across the United Kingdom. But trying to insulate Scotland from such risks while still providing it with the right incentives is complicated: it would involve isolating the effect of Scottish policy (which we would want the Scottish Government to bear) from other factors affecting devolved tax revenues and spending. Any modelling exercise like this is controversial – particularly when substantial sums could be at stake.

Is it possible to compensate for knock-on effects of policy decisions?

This brings us on to another one of the Smith Commission’s key principles: that the Scottish Government should bear the full revenue or spending consequences of its policy decisions, and the UK Government of its decisions. In particular it says:

“Where either the UK or the Scottish Governments makes policy decisions that affect the tax receipts or expenditure of the other, the decision-making government will either reimburse the other if there is an additional cost, or receive a transfer from the other if there is a saving. There should be a shared understanding of the evidence to support any adjustments.”

In principle this is sensible. In order to align incentives for policy making properly, each government should bear the full costs (and receive the full benefits) of its policy decisions. It also seems only fair to compensate (or penalise) the other government for ‘knock on effects’ of policy decisions. But implementing such a principle would be fraught with practical difficulties meaning that such transfers might often be infeasible.

Nearly all policy decisions could have knock on effects on the revenues or spending of the other government. But calculating what these are is inherently difficult, with much room for disagreement over the methods used – and different effects of the same policy might go in opposite directions. Therefore such compensating transfers might only be practical in a few simple cases – otherwise the system could quickly become unworkable and lack transparency.

Should the Scottish Government’s budget always be unaffected by changes to taxes in the rest of the UK which have been devolved to Scotland?

The next proposal from the Smith Commission could also be problematic. It says:

“Changes to taxes in the rest of the UK, for which responsibility in Scotland has been devolved, should only affect public spending in the rest of the UK. Changes to devolved taxes in Scotland should only affect public spending in Scotland”

It seems clear that, for instance if income tax were increased in the rest of the UK to fund additional spending on services in the rest of the UK, then there should be no knock-on effect to the Scottish budget: Scots would be paying the same taxes as before so should see neither a rise nor fall in the amount spent on them. Indexing the block grant reductions or additions in the way we suggest above would achieve this outcome.

But what if the UK government wanted to spend more money on defence or state pensions, or wanted to increase taxes to reduce borrowing. These are things that benefit the whole of the UK, including Scotland, even though they are not necessarily “spending in Scotland”. One interpretation of the Smith proposals would be that the UK government could not use an increase in income tax – one of the main taxes it levies – to fund these policies. This would be absurd and the Treasury has informed us that it was not the intention of the Smith Commission to constrain the UK government in this way. But if the UK government did use income tax (levied only outside Scotland) as part of its response to such UK-wide issues then the amount of grant transferred to the Scottish Government should be changed in response. For instance, if taxpayers in the rest of the UK were paying more in income tax to reduce the budget deficit it would be only fair that Scottish taxpayers also contribute to deficit reduction. Devolution of income tax powers to Scotland should give Scots more freedom of taxation and spending in future – but not the ability to avoid spending cuts or tax rises that are carried out to finance things that benefit the whole UK.


The Smith Commission has provided a set of proposals for further devolution of taxes and spending, agreed by the five main Scottish parties. This is a significant achievement. But many difficult issues remain to be addressed – not least, how the block grant will be adjusted to account for the additional revenues and spending areas that will come under Holyrood’s control. No system will be perfect. There is an inherent trade-off between providing incentives to the Scottish Government, and the degree of risk-sharing between Scotland and the rest of the UK. And, it will not be practical to devise a system where the UK and Scottish governments compensate each other for all the knock on effects of their policies as the Smith Commission recommends. That being said, it is possible to avoid some of the biggest pitfalls – and important too: our earlier analysis of the mistakes that were made when business rates were devolved shows that getting these details wrong can have big financial consequences.  

]]> Thu, 18 Dec 2014 00:00:00 +0000
<![CDATA[Coping with the cap?]]> In 2013 the government introduced a cap on the amount that some working-age families can receive in benefits. Today the Department for Work and Pensions (DWP) published quantitative analysis of the direct impacts of the cap and of how the affected people might have responded. We peer-reviewed this work and offered advice on the methods used, and in this observation we draw out some of the key findings. (A previous observation gave an overview of the benefits cap and commented on the potential rationale – or otherwise – for such a policy.)

The cap, which works by reducing housing benefit awards, was set at £500 per week (except for childless single people, for whom it is £350), and those receiving Working Tax Credit and some claiming disability benefits are exempt. These choices mean that the cap does not affect many people and that the overall fiscal consequences are small. About 27,000 families (less than 1% of working-age families receiving housing benefit) were being capped once the policy was fully rolled out in late 2013, with their benefit income reduced by a total of about £100 million per year. Essentially all the families who receive enough benefit income for the cap to be binding have a large number of children or high rents (and often both).

However, the relatively small number of affected families can lose substantial amounts. Half of those capped in November 2013 lost at least £46 per week as a result. Some recipients lost much more than this, so the mean loss among those capped was much higher still, at £70 per week. Given these large impacts on those affected, it is important to understand how they are responding.

Identifying causal impacts of policies on behaviour is rarely easy, but DWP’s analysis allows us to draw some conclusions about the likely effects. The Figure below illustrates perhaps the most striking evidence. It plots the amount of weekly benefit income (before any cap is applied) against the probability of claiming working tax credit (WTC) a year later, for four groups (`cohorts’) of benefit recipients: those receiving benefits in May of 2010, 2011, 2012 and 2013 (excluding those exempt from the cap, and excluding the single childless who are subject to a different cap level). The May 2012 cohort is the first that we might expect to see changing its behaviour in light of the cap: from May 2012, claimants who looked like they were set to be affected were sent a letter notifying them of this and were offered support through Jobcentre Plus. The cap was then rolled out from April 2013.

The May 2012 and 2013 cohorts were more likely to flow onto WTC than the earlier cohorts at all benefit levels shown – including for recipients below the cap level, which suggests that some of this difference is due to a wider recovery in the economy after 2011 rather than the cap. Crucially though, the divergence between cohorts begins to widen at just around the £500 point at which the cap binds. The divergence grows consistently as one moves to the very highest levels of benefit entitlement, i.e. to those who were hit hardest by the cap. (Note that there are very few claimants at these very high levels of entitlement.)

Figure: Movements onto Working Tax Credit after one year by benefit income (before capping)


Notes: £500 benefit cap introduced in 2013. Excludes single adults without dependent children, for whom a lower cap level applies (£350 rather than £500 per week). Figures are four-point moving averages.

Source: Chart 8.2 of DWP report “Benefit Cap: Analysis of outcomes of capped claimants”, available here.

DWP conducted some more formal econometric analysis of the kinds of patterns documented in the Figure. On average, claimants with benefit income exceeding the impending cap level in the May 2012 cohort – who received warning of the cap and support in dealing with it - were 1.5 percentage points more likely to flow onto WTC within a year than their counterparts just below (within £50 per week of) the cap. This is over and above any gap that would be expected simply because of observed differences in the characteristics of these two groups, such as the number of children that they have. The difference grew to 4.7 percentage points for the May 2013 cohort, after implementation of the cap. These differences did not exist for the earlier May 2010 and May 2011 cohorts (indeed prior to the cap those with higher levels of benefit income were slightly less likely subsequently to move onto WTC). This suggests that these differences provide a reasonable sense of the likely effect of the cap on movements onto WTC.

It is worth bearing in mind the caveat that starting a WTC claim is not the same thing as moving into paid work. Some people may move into work but not work enough hours to be entitled to WTC or not take up the WTC to which they become entitled. On the other hand, some might start a WTC claim when they were in work all along, perhaps because claiming this entitlement is a relatively easy way of exempting oneself from the benefit cap. To the extent that additional moves onto WTC are an accurate indicator of moves into work, these estimates suggest that around 2,000 families who were claiming benefits in May 2013 had someone move into paid work twelve months later in response to the cap. (Note: this estimate is for this cohort, rather than an overall estimate of total additional moves to WTC.)

We might also expect some claimants to move house in response, as many affected claimants are above the cap because they have high rents and hence have a large housing benefit claim. There is evidence of this, but (at least so far) only for the small number of claimants who lost particularly large amounts of benefit income as a result of the cap. For those with benefit entitlement at least £200 above the cap level, 14% of the May 2010 cohort moved within the next year; this had risen to 20% for the May 2013 cohort once the cap was in place. For benefit claimants just under the cap level, the proportion moving house within a year stayed constant at 11% for each of the May 2010, 2011, 2012 and 2013 cohorts.

This evidence helps us to learn about the behaviour of a group of benefit claimants who we previously knew little about. At its Autumn conference, the Conservative Party suggested lowering the benefits cap by approximately a further £60 per week (to about £440 per week). On the basis of the analysis published today, it would be reasonable to expect this to result in some of the affected claimants moving into work, but few moving house – the maximum possible loss of benefit income, from this additional hypothetical cut, would be £60 per week, and the current cap seems to have increased house moves only among those who lost substantially more than that.

Nevertheless, there is still much we do not know. There are various other possible responses to this reform, including cutting back on spending, running down savings (or building up debts) or getting help from family or friends. Analysis of benefits data cannot tell us about this (although DWP have also conducted surveys of, and in-depth interviews with, those subject to the cap and these provide some information on other potential responses). What the quantitative analysis does tell us is that the large majority of affected claimants responded neither by moving into work nor by moving house. For this majority, it remains an open question as to how they adjusted to what were, in many cases, very large reductions in their income.

]]> Mon, 15 Dec 2014 00:00:00 +0000
<![CDATA[The government’s proposed new postgraduate loan scheme: will the RAB charge really be zero?]]> In the Autumn Statement last week, the Chancellor announced a new government-backed loan scheme for postgraduates. Loans of up to £10,000 are to be made available for under-30s studying full-time or part-time for taught masters courses from 2016/17. The government expects 92,000 people to be eligible for the loans in the first year they are available, and intends the long-run cost of the scheme to be zero, meaning that the full value of loans will be repaid, on average. While full details of the loan system have yet to be announced – and will be subject to consultation – the documentation accompanying the Autumn Statement highlighted one way in which the scheme could operate. This observation assesses the example loan scheme put forward and raises some issues that should be considered during the consultation.

The loan system announced by the Chancellor should help individuals who may otherwise be prevented from continuing into postgraduate study because they are credit constrained. If the value of the loans issued were not repaid in full, then it would also represent a government subsidy for postgraduate education, just as there is in the undergraduate student loan system. (Previous IFS research estimated that the government can only expect to recoup 57% of the value of undergraduate student loans, representing a considerable government subsidy.) However, the Autumn Statement made clear that the government intends the loans issued to postgraduates to be repaid in full, on average, suggesting that it does not wish to subsidise loans for postgraduate study.

In the illustrative scheme put forward in the Autumn Statement loans of up to £10,000 will be made available to those under the age of 30 studying for taught masters courses. English students studying in the UK and EU students studying in England would be eligible. The loans will be subject to a real interest rate of 3% (i.e. interest will be charged at RPI plus 3%) and will be repaid at a rate of 9% of income above a lower-income threshold of £21,000, which would be frozen in nominal terms for 5 years. We assume that all other features of the repayment system follow that of the undergraduate loan system, i.e. that individuals incur an interest rate of RPI plus 3% while they are studying; that the lower earnings threshold is uprated by average earnings growth after the five year period; and that postgraduate debt will be written off after 30 years.

Our estimates of the future repayments that would be made under such a scheme are outlined in Figure 1. These estimates are calculated for English-domiciled students using repayments made on the basis of lifetime earnings rather than income and assume that repayments are made in line with the repayment schedule (with no under- or over-payments). We estimate that the government could expect to recoup 100% of the value of these loans in the long-run (i.e. that the so-called RAB charge – the long-run cost to the government of issuing student loans – is effectively zero).

This differs from estimates of the long-run cost of the undergraduate loan system – which we estimate will cost the government 43p for every £1 loaned out – for three reasons. First, and most importantly, postgraduates will borrow considerably less: the maximum postgraduate loan will be £10,000, while the average undergraduate loan is around £44,000. Second, according to recent estimates, postgraduates earn around £200,000 more over their lifetimes, meaning they repay more quickly. Third, the illustrative loan scheme suggests that they would be charged higher interest rates on their loans than most undergraduates. 


Figure 1: Net Present Value (NPV) of postgraduate loan repayments made and RAB charge, by distribution of postgraduate lifetime earnings distribution (2016 prices)

Figure 1 also shows how repayments vary across the distribution of lifetime earnings amongst postgraduates. It shows that in all but the first decile of lifetime earnings, loans are close to being repaid in full, on average. Indeed, less than 15% of postgraduates do not repay the full value of their loan, and repayments are, on average, highest in the third decile of lifetime earnings. This relative lack of progressivity is a consequence of the positive real interest rate charged and the fact that repayments are lower than the interest accrued at relatively low earnings. Individuals higher up the earnings distribution also repay more quickly, on average, with those with the highest lifetime earnings repaying in full within seven years of graduating.

There are a number of assumptions that underlie these estimated figures, all of which introduce risk to the government that the long-run cost of issuing loans to postgraduates may turn out to be different from zero. Two important assumptions for the total cost of undergraduate loans were the long-run real earnings growth rate and the population of borrowers. However, for postgraduates these assumptions are less important, as the loan is relatively small in comparison.

For example, while changing our assumption about long-run real earnings growth from 1.1% to -1% increased the RAB charge from 43% to 52% for the undergraduate loan scheme, for postgraduates it increases only slightly, from -0.9% to 1.3%. Similarly, even if we were to assume that only the lowest earning 56% of postgraduates take out loans (this is the take-up rate assumed by the Autumn Statement), the RAB charge would still be very small: around 3%. This lack of sensitivity arises mostly because the value of the loans issued is relatively small compared to the undergraduate scheme, but is also partly driven by the features of the illustrative loan scheme. If, for example, the interest rate charged were to increase with income, then the burden of repayment would fall more heavily on higher earners, but the RAB charge would also become more sensitive to changes in the population of borrowers. This represents an important trade-off that should be considered when designing the repayments system.

One assumption to which our findings are sensitive is that of repayment compliance. In 2012/13, 13.5% of new postgraduate students in the UK were EU domiciled. If the same proportion of EU students took out loans, but the government were unable to collect any repayments from these students, we estimate that the RAB charge would increase to 12.6%. This is another important consideration for the policy consultation.

An additional concern with the introduction of the government’s illustrative example is the high marginal tax rate that individuals would face under the new system. Since repayments on the postgraduate loan would be made “concurrently” with undergraduate repayments, individuals earning between the lower loan repayment threshold (of £21,000 in 2016 prices) and the higher income tax-rate threshold would face marginal tax and employee NICs rates of 50%, while those earning above the higher rate tax threshold would face marginal rates of 60%. This could potentially affect the labour supply decisions of young postgraduates and hence may have wider consequences for growth and productivity.  

It would also introduce a further complication to the tax schedule faced by postgraduates: while the repayment threshold for undergraduate loans looks set to be uprated in line with average earnings growth, the Autumn Statement documentation suggests that the repayment threshold for postgraduate loans would be frozen in nominal terms for five years (we have assumed that it would rise at the same rate as the undergraduate threshold thereafter). This would create an increasingly large range over which repayments would be due on any outstanding postgraduate loan but not on any outstanding undergraduate loan. Of course, this is just an illustrative example, and the government will consult before outlining the specifics of the policy, but it will be important to consider such features.

Ultimately, the success of the policy will be determined by the responses of both students and universities. Are credit constraints a primary driver of individuals’ decisions not to stay on for postgraduate study, and if so is the new scheme sufficient to alleviate these constraints? The response of universities is also uncertain. While the proposed postgraduate loan scheme does not link loans to fees in the same way as it does at undergraduate level, institutions with high market power might still respond to the increased availability of credit by raising prices, which would reduce the effectiveness of the policy in making the upfront costs of postgraduate study cheaper. Thus, while the introduction of a loan scheme is broadly welcome, the devil will, as always, be in the detail.

]]> Tue, 09 Dec 2014 00:00:00 +0000
<![CDATA[Household incomes set to start growing again, but slowly and unequally]]> In work funded by the Office of the First Minister and Deputy First Minister in Northern Ireland, the IFS today published an update of our previous income projections for Northern Ireland and for the UK as a whole. This observation draws out what this research tells us about the likely path of household incomes in the UK in the next few years.

Throughout this observation, we focus on changes in incomes relative to inflation as measured by the CPI. Official government statistics still used the flawed RPI measure of inflation; figures using this measure are shown in the briefing note.

Unsurprisingly, the recent recession had a significant impact on household incomes. Median (middle) income fell by 5% relative to CPI between 2008–09 and 2011–12, before growing slightly in 2012–13, the latest year for which official statistics are available. Our projections show that median income is likely to have fallen again relative to CPI in 2013–14, but might now be growing (slowly) in real terms. This means that we expect median household income in 2015–16 to be roughly where it was in 2010–11 (after adjusting for CPI inflation), though still about 3% below its 2007–08 level.

These trends in median income tell us nothing about what is going on at higher or lower income levels. During the recession, income inequality fell as state benefits broadly retained their real value due to price indexation, while workers’ earnings failed to keep pace with CPI inflation. Tax rises affecting better-off households also played a role in reducing inequality. However, this does assume that all households face the average inflation rate - previous IFS research has shown that lower-income households were experiencing higher rates of inflation than richer households over this period. In addition, our projections show that this pattern of income changes is likely to be reversing now: cuts to working-age benefits were accelerated in 2013–14 and will continue to accumulate until at least 2015–16, while earnings growth starts to recover and middle- and higher-income households benefit most from large increases in the income tax personal allowance. As we show in our report, the fact that the incomes of low-income households are projected to fall between 2012–13 and 2015–16, both relative to CPI inflation and median income, means that we would expect both absolute and relative poverty measures to increase over this period.

It is important to note that our projections rely on forecasts for the macroeconomy produced by the Office for Budget Responsibility. A new set of these forecasts will be produced alongside the Autumn Statement next week. Changes to these forecasts – for example, if earnings growth is once again downgraded, or inflation is lower – will have an impact on our projections. We will therefore be updating our projections as part of our election analysis next year.

Figure 1. Real income growth by percentile point: 2007–08 to 2015–16

Figure 1. income growth

Note: Figures adjusted for CPI inflation.

]]> Mon, 24 Nov 2014 00:00:00 +0000
<![CDATA[Does GP Practice Size Matter? The relationship between GP practice size and the quality of health care]]> This week the Care Quality Commission (CQC) published comparable information on all GP surgeries in England for the first time, with the aim of identifying which practices should be prioritised for newly introduced inspections.  Understanding whether and why some GP practices fail to provide quality care is important given the vital role that GPs play within the NHS. In recent years, there have been substantial changes to the organisation of GP practices, with a shift away from single-handed practices towards practices with six or more full time equivalent GPs. This observation summarises the findings of an IFS report published today, which describes these trends in detail and examines the relationship between practice size and a variety of practice outcomes and behaviours.

Trends in the organisation of GP practices

Between 2004 and 2010, the total number of full-time equivalent (FTE) GPs in England grew by 15%, rising from 32,263 to 37,173 FTE GPs. Over the same period, the registered population grew by just 5%.  As a result, the number of FTE GPs per individual increased by 10%.

Meanwhile, the number of GP practices fell.  In 2004, 9,016 practices were registered.  By 2010, this had decreased to 8,832.  These trends together reflect in a shift towards a smaller number of GP practices which on average employ a larger number of GPs.  In 2004, there were 3.6 FTE GPs per practice.  This rose to 4.2 in 2010. In particular, there has been a marked decrease in the number of single-handed GPs, while practices with six or more FTE GPs have become more common. 

In the figure below, we divide practices into four categories:

  • single-handed (one or fewer FTE GPs per practice)
  • small-medium (more than one and up to three FTE GPs)
  • medium-large (more than three and up to six FTE GPs)
  • large (more than six FTE GPs)

The share of small-medium practices and medium-large practices stayed flat between 2004 and 2010, while the percentage of single-handed GPs fell (from 22% to 15%) and the share of large practices grew (from 16% to 23%).

Figure: Share of GP practices, by FTE GPs (all practices)

Note: Includes all GP practices with at least one registered GP in the year in question.

Source: Kelly and Stoye (2014), Figure 2.2

Practice size, behaviour and patient outcomes

Overall there is a small, positive relationship between the size of GP practices and the score that they achieve under the Quality and Outcomes Framework (QOF). QOF scores (introduced in 2004), partially determine payments to practices and are used to provide an overall measure of the clinical quality of a GP practice. We find that single-handed practices achieved the lowest average scores, while large practices achieved the highest.  These findings remain even after controlling for differences in the characteristics of the populations and areas that small and large practices serve.  However, there is variation across the different domains captured by QOF.  Interestingly, single-handed practices achieved the highest average scores for patient experience but lower scores in the other domains.  This means that patients appear to be more satisfied with the care received in small practices, despite objective measures suggesting that they receive worse clinical standards of care.

We also find some evidence that single-handed GPs are more likely than GPs in larger practices to have high emergency admission rates among their patients for conditions that would potentially have been avoidable. There are a number of conditions (known as ambulatory care sensitive or ACS, conditions) for which emergency hospital admissions are potentially preventable if the patient receives timely and effective primary care. These include acute conditions (such as ear, nose and throat and urinary tract infections), chronic conditions (such as diabetes and asthma) and vaccine preventable conditions (such as influenza).  Smaller practices are found to be significantly more likely to be among the 20 per cent of practices with the highest emergency admissions rates for ACS conditions, and this relationship holds even when taking account of other patient and practice characteristics.  This relationship is particularly strong in the case of chronic conditions, with single-handed GP practices 4% more likely to appear in this category relative to large practices.  This perhaps suggests that practices with fewer GPs are less able to spend significant periods of time managing long-term conditions.

We also find that smaller practices behave differently from larger ones in how they refer patients on for secondary care. Single-handed and small-medium practices are, other things being equal, less likely to refer their patients for secondary care than GPs in larger practices are.  For example, single-handed practices are 12% more likely to appear in the 20 per cent of practices with the lowest referral rates (once adjusting for the demographic composition of their patients).These practices are also less likely to refer patients to independent sector providers. This may indicate a lower willingness or ability of GPs in smaller practices to offer a full range of treatment options to patients.

It is important to note however that there is significant variation across practices of the same size, and some smaller practices provide extremely good quality care. There are also many factors that could affect the outcomes examined in our report, and which may differ between small and large practices.  As a result, the relationships between GP practice size and GP behaviour are not necessarily causal.  For example, if small practices typically treat patients with worse health on average, the findings related to higher than expected emergency admissions in small practices could simply reflect underlying health differences rather than any impacts of practice size.  Other GP characteristics may also vary across different practice sizes in some unobservable way. 

These findings overall suggest that smaller practice sizes typically deliver worse health outcomes for patients, particularly in single-handed practices.  Although we cannot say for sure that the small number of GPs is the sole cause of this relationship, it is likely to be an important factor.  It is important for policymakers to be aware of this, and understand it better so as to address it.


]]> Thu, 20 Nov 2014 00:00:00 +0000
<![CDATA[What is happening to spending on social security?]]> Over the course of this parliament, the government have made changes to benefits and tax credits that, at the time they were announced, were expected to reduce spending in 2014–15 by £19 billion relative to a world of no policy change. In fact, real spending (after adjusting for CPI inflation) will be only £2.5 billion lower in 2014–15 than it was in 2010–11. This observation explains why.

If we adjust for prices using the CPI, the move from RPI to CPI for uprating benefits (expected to save over £4 billion in 2014–15) did not bring about a cut in spending over time – rather it meant benefits were no longer expected to increase in real terms. So we’re left with around £15 billion of cuts to explain, compared to an actual fall in real terms spending of around £2.5 billion.

Over a third of the remaining £12.5 billion gap is down to higher spending on pensioner benefits. Combined spending on state pensions, pension credit, and universal pensioner benefits such as the winter fuel payment will be £5 billion higher in 2014–15 than in 2010–11. All of that increase is explained by the rising cost of state pensions. To some extent this reflects an ageing population, with the number getting the state pension up 400,000 over the period. But it is mostly the result of higher spending per pensioner, with each recipient getting nearly £500 a year more on average. This was, for the most part, planned – as each new cohort of pensioners retire, they benefit from past decisions giving them more generous entitlements than previous generations.

Once we strip out pensioner benefits, we’re left with around £7.5 billion of extra spending to explain. In order to help with that task, Table 1 shows real terms spending on major working-age benefits in 2010–11 and 2014–15, alongside announced cuts (excluding the move from RPI to CPI indexation).

Table 1. Changes in spending and announced cuts for some major working-age benefits

Notes: Spending figures adjusted for CPI inflation. Spending on employment and support allowance not shown as cuts announced by previous government. Disability living allowance figures include personal independence payment. Cut to JSA is from 1% uprating and is approximate.

Source: Department for Work and Pensions benefit expenditure tables, Office for Budget Responsibility policy measures database.

The first thing to draw from the table is that there is actually more than £7.5 billion of extra spending to explain, once you take account of the fact that spending on JSA has fallen by around £1.3 billion as a result of lower unemployment.

Housing benefit explains £3 billion of the extra spending. Despite announced cuts of over £2 billion, real terms housing benefit spending will be nearly £1 billion higher in 2014­–15 than 2010–11. This was unanticipated – the OBR’s welfare trends report shows expected spending in 2014–15 has risen by nearly £3 billion since their June 2010 forecast. As they explain, the private rented sector has grown faster than expected, private rents have grown faster than expected, and earnings have grown more slowly than expected – all of which increase housing benefit spending. That slower-than-expected earnings growth also increases tax credit spending. Government cuts to tax credits total £4.6 billion, but spending in 2014–15 is expected to be down less than £3 billion on its 2010–11 level.

It’s not all about macroeconomic conditions though. The difference between the £1.2 billion cut to DLA spending that was announced and the £1.6 billion increase in spending that is now expected reflects the significant delays to the government’s replacement of DLA with the less generous personal independence payment. And although the introduction of employment and support allowance is not included in the £19 billion of cuts (since it was announced by the previous government), it too has saved much less than expected (for reasons discussed here).

All this has important fiscal consequences. Working-age benefit spending has always been sensitive to the unemployment rate. But the rapid growth of housing benefit and tax credits over the couple of decades (documented in this briefing note published today) means that slow earnings growth now has the potential to push up spending too. Much of the hoped-for savings from the introduction of ESA have failed to materialise, and it is an open question whether the personal independence payment will be any different. Mr Osborne wants further cuts to social security spending to help reduce the deficit. He may end up having to make cuts just to stay on track.

Note: The DWP benefit expenditure tables adjust spending figures for economy-wide inflation (as measured by the GDP deflator) – in those terms spending is expected to be £5 billion higher in 2014–15 than 2010–11.

]]> Mon, 17 Nov 2014 00:00:00 +0000
<![CDATA[What is welfare spending?]]> The government has started to send out information on how tax revenue is spent to individuals who pay income tax or National Insurance contributions. It has broken down spending into a number of categories. The biggest of these is "welfare", which represents a quarter of total spending. State pensions also appear as a separate category, accounting for 12% of spending. In this observation we look at what counts as pension and welfare spending, and offer some alternative breakdowns.

Spending on state pensions is straightforward. This is essentially just the annual spend on the basic state pension and earnings-related state pensions from various different schemes.

"Welfare" spending, at 25% of the total, is taken directly from the government's public expenditure statistical analyses. It is the total spending defined as "social protection" in these analyses, less spending on state pensions. Total spending on social protection comes in at £251 billion in 2013-14, which is about 37% of total public spending of £686 billion (before accounting adjustments). Take off £83 billion of spending on state pensions and you get to £168 billion on "welfare" - very nearly a quarter of total spending.

What is included in that "welfare" total?

It includes £28.5 billion on "personal social services". This is a number that in many analyses one would want to report separately from other welfare spending. It includes spending on a range of things, such as looked-after children and long term care for the elderly, the sick and disabled. Unlike other elements of "social protection" it is not a cash transfer payment and in many ways has more in common with spending on health than spending on social security benefits.

Another £20 billion of the spending counted under welfare is pensions to older people other than state pensions. That includes spending on public service pensions – to retired nurses, soldiers and so on[1]. This is not spending that would normally be classed as "welfare". The rest of the pay package of a public sector worker is included as departmental spending within the department of that worker. One could either report such pension payments separately or, like pay, as part of the relevant spending function. The pay of nurses counts as health spending. One could count their pensions in the same way.

That leaves around £120 billion of other welfare spending, which can be broken down in a number of different ways.

Since the government has chosen to report state pension spending separately, one obvious division would be to separate spending on those of working age and those of pension age. In addition to state pensions a further £28 billion is spent on pensioners, of which £15 billion goes on benefits specifically for that group, such as pension credit, attendance allowance and winter fuel payment, while the remaining £13 billion is largely spent on housing benefit and disability living allowance. So of the £205 billion or so spent on tax credits and social security benefits about £111 billion is spent on those over pension age and £94 billion on those of working age.

Figure 1 and Table 1 show this breakdown of the 25% of total spending described as "welfare" by the government, alongside the 12% spent on state pensions. 4% goes on "personal social services", 3% on public service pensions, 4% on other benefits for pensioners, and the remaining 14% on benefits for those of working age.

Figure 1. Breakdown of "welfare" as a share of total spending

Sources: Public Expenditure Statistical Analyses, Department for Work and Pensions benefit expenditure tables.

Table 1. Breakdown of "welfare" as a share of total spending


Sources: see Figure 1.

It could of course be argued that this would provide too much detail. But there are five categories reported in the government's breakdown of spending, each representing less than 2% of total spending. If it is worth reporting contributions to the EU budget which represent less than 1% of total spending then there might be a case for providing this additional breakdown of "welfare spending".

There are of course other ways of breaking down spending on social security benefits. The OBR, for example, show spending on the elderly, sick and disabled, families with children, the unemployed and help with housing costs (table 2.1 here) . In our survey of the benefits system we provide a similar split (table 3.1 here), separating out spending according to its function as best we can. There are clearly different judgments one can make here (we include attendance allowance in spending aimed at the sick and disabled, rather than the elderly for example), but it does give a good sense of the overall spending priorities within the system. Table 2 shows this split, including the 12% of total spending going on state pensions, and excluding the 4% spent on personal social services and the 3% reported as going on pensions other than the state pension (both of which the government includes in "welfare").

Table 2. Welfare spending by function


Note: categories correspond to the primary recipient of a given benefit, rather than capturing all of the expenditure on each group.

Source: Department for Work and Pensions benefit expenditure tables.

There are different ways of reporting how our taxes are spent, and there is a balance to be struck between the amount of detail presented and clarity of message. Lumping a quarter of total spending into one bucket labelled "welfare" may not strike the most helpful balance, especially when it includes such diverse items as spending on social care, public service pensions, disability benefits, child benefit and unemployment benefits.

Note: The IFS provides detailed breakdowns of tax revenue here, benefit spending here, and other public service spending here.

[1]It is also not the number reported by the OBR when looking at spending on public service pensions (see table 4.24 here). The OBR reports gross spending of £36 billion and "net" spending – i.e. gross spending less contributions received – of £10.5 billion. But for the purposes of this note we stick with the numbers published in PESA.

]]> Tue, 04 Nov 2014 00:00:00 +0000
<![CDATA[How far through the consolidation are we?]]> The Prime Minister, in an article in today's Times, said "In this parliament we will have made £100 billion of savings while cutting income tax by £10.5 billion. In the next parliament we plan to make £25 billion of savings while making £7.2 billion of income tax cuts". This repeats a statement made at this year’s Conservative Party conference. The implication is that most of the planned cuts in public spending have been made and that the promised tax cuts in the next parliament are relatively modest by comparison with what has been managed in this parliament.

The problem is that the two numbers for “savings” are inconsistent and should not be compared with one another. Public spending cuts planned for the next parliament are about as severe as those which have been implemented over the past five years. Depending exactly how you make the comparison you could say we will be a bit less or a bit more than half way there by next April. (See for example and from this year’s IFS Green Budget).

So what is the £25 billion number? That is quite straightforward. It is the real terms cut in total public spending (excluding debt interest) expected for the years 2016-17 and 2017-18. This is a measure of the reduction in public spending (relative to economy wide inflation) which the government expects will occur over that period.

Note it is not a measure of cuts over the whole parliament: it excludes 2015-16 and 2018-19. It is also not a measure of discretionary changes in spending resulting from policy decisions over, for example, spending on public services or social security. It is a measure of the change in total non-interest spending. It is because some elements of spending, driven for example by an ageing population, have been rising, and will continue to rise, that such tough decisions have had to be taken over spending on public services, social security and taxation.

All that said this is a sensible measure of spending. It measures what is happening to the total amount spent by the government on things other than debt interest. On that basis we think total spending will fall by about £28 billion over the next parliament compared with about £23 billion of spending cuts which will have been implemented by the end of this parliament. We are just under half way there.

Indeed if you look at the period of the consolidation, the scale of the total deficit, the scale of the cyclically adjusted deficit, or the scale of spending reductions relative to pre-crisis expectations you get to broadly the same answer. On some bases we will be a bit more than half way through the cuts come the next election, on others a little bit less. There is no sense in which we will be 80% of the way through (which is what is implied by the statement that £100 billion of savings have been made with £25 billion to go).

So what is the £100 billion number which Prime Minister described as savings made over this parliament?

For a start it includes cuts planned for 2015-16 – i.e. in the next parliament, not this one.

More importantly it is calculated on an entirely different basis. It is a measure of policy activism, of the discretionary cuts implemented over the period from June 2010 to the end of 2015-16. It also takes account of the reduction in debt interest spending arising from a tighter fiscal stance. It is not a measure of the change in total spending. So, for example, it includes the effect of policy changes which have reduced spending on social security and public sector pensions to significantly below what they would have been in the absence of the cuts. It does not take account of the fact that other pressures have increased spending in these areas.

Again this is an entirely sensible way of looking at things. It is gives a good sense of the squeeze that is being applied to public spending. But it is not comparable with a number based on total spending.

The £25 billion number simply measures the change in spending relative to a baseline period, asking "how much less in total are we spending now relative to the baseline period". The £100 billion number measures the impact of discretionary policy change, asking "how much less are we spending compared with a world in which we had made no policy choices (relative to a particular counterfactual)".

There are different ways of looking at what is happening to public spending. They have different merits and are useful in their own right. To say that there have been "savings" of £100 billion over this parliament is a useful and interesting statement if one is clear about what it means (and takes account of the fact that it includes savings planned for 2015-16). It is also interesting and useful to say that £25 billion of "savings" are planned over the next parliament (or at least for the years 2016-17 and 2017-18). But it is not useful to use those two numbers together to suggest that the vast majority of the planned cuts have happened. They haven't.

]]> Thu, 30 Oct 2014 00:00:00 +0000
<![CDATA[Employment and retirement – explaining recent trends]]> Employment rates of men and women in their 50s and 60s have been growing for nearly two decades – a trend that was only briefly halted during the recent recession. Employment rates for men age 65-69 are at their highest levels since 1974 and continue to grow while employment rates for older women are at record levels. There are reasons to expect that this growth will continue as changes in private pensions and state benefits increase incentives to remain in work longer. For women this upward trend reflects the continuation of a long-term trend since at least the late 1960s. In contrast, recent growth in employment among older men follows a period of dramatic decline between the late 1960s and mid-1990s. In two new publications released today, researchers from the IFS set out levels of employment among older men and women in England in 2012–13, and survey explanations for why these have increased so substantially since the mid-1990s.

Figure 1: Employment rates of older men, 1968–2013

Source: Family Expenditure Survey (1968–1982), Labour Force Survey (1983–2013). Figure 2.1 in Retirement in the 21st Century

Figure 2: Employment rates of older women, 1968–2013

Source: Family Expenditure Survey (1968–1982), Labour Force Survey (1983–2013). Figure 2.6 in Retirement in the 21st Century

Employment in 2012-13

In 2013, 87% of men and 77% of women aged 50 in England were in paid employment or self-employment. Labour force participation declines sharply with age, with the largest declines occurring around the State Pension Age, which in 2013 was between the ages of 61 and 62 for women, and 65 for men. After age 61, employment rates of women were below 50%; for men, employment rates were below 50% from age 64 onwards. A small minority continue working into their 70s, and even at the age of 74 nearly one-in-ten men and one-in-twenty women were doing some form of paid work.

Older people are much more likely to be working part time. This pattern applies to both men and women: more than half of men still in work in their late 60s are working part time, while for women this rises to more than 80%. Self-employment also increases significantly with age, accounting for 44% of men and 20% of women who are working ages 65-69. Recent ONS research has shown that the number of over-65s who are self-employed has doubled in the past five years to nearly half a million.

Health is an important determinant of work status and older people on average have worse health than younger people. For those in the worst health, employment is already low at age 50, while those who are in poor health but still in work in their early 50s tend to leave work more quickly than their healthier counterparts. However, even among those aged 70–74, six-in-ten men and seven-in-ten women report that their health does not limit the kind or amount of work they could do – suggesting that, while health may be a very important factor for some people, it is not a limiting factor for labour force participation for most people, even in their early seventies.

Employment since the mid-1990s

As Figures 1 and 2 demonstrate, women have seen rising employment since the late 1960s, while male employment plummeted between the 1960s and 1990s before increasing since then. Although our research has focused on explaining the more recent rise in employment (particularly among men), this is best understood in a longer term context: some explanations for more recent trends point to a reversal or amelioration of downward pressures seen in the 1970s and 1980s, while others point to genuinely new forces acting on labour force participation of older people.

Between the late 1960s and mid-1990s a number of factors affecting both the demand for and supply of older workers put downward pressures on the labour force participation of older people. On the demand side, the recession of the early 1980s, the decline of traditional manufacturing industries and a general shift in employer demand against low skilled employees all had a disproportionately large effect on older men. On the supply side, increases in the coverage and generosity of private pensions and the widespread availability of early retirement packages made retirement a more attractive option for many., A large increase in the fraction of older people claiming disability benefits, which is more readily explained by economic conditions than changes in health, suggests that for some these benefits acted as an alternative route to retirement. Although some of these pressures also affected women, these downwards pressures were more than compensated for by rising labour market attachment of successive cohorts.

Employment rates among older men started rising in the mid-1990s, while employment among women over 50 increased at a faster rate than in the previous decades. On the demand side, strong economic growth from 1995 to 2007 will have contributed to rising demand for all workers (including older workers), while the continued trend towards services may have increased employment opportunities appropriate to older workers. On the supply side, this period saw a drying up of the generous early retirement packages previously offered by many employers, while reforms to the disability benefits system in 1995 and the mid-2000s reduced the on-flow and increased the off-flow from these benefits, respectively. On the other hand, pension coverage was broadly stable among those approaching the State Pension Age, while overall wealth increased (driven in large part by rapidly rising house prices), factors which (if anything) we would have expected to reduce labour supply. Continued improvements in health are likely to have increased employment further, though this cannot explain the reversal in male employment in the mid-1990s as health was also improving in the earlier period.

Some new factors may also have been at play. A range of policy changes are likely to have increased both demand for and supply of older workers, including legislation allowing people to draw their pension while continuing to work for the same employer, anti-age discrimination legislation, and increases in the female SPA.

Looking forward, many of the forces that have contributed to rising employment among older people over the last two decades are likely to persist. On the demand side, employment among older workers has fared remarkably well in the recent recession compared to previous recessions, and the trend towards more flexible, service sector jobs is likely to continue. On the supply side, there is unlikely to be any relaxation of access to disability benefits, while the full impact of the shift from defined benefit (that is, salary-related) pensions to (typically less generous) defined contribution pensions, which have less sharp financial incentives to leave work, is yet to be felt and could well lead to later retirement.


]]> Thu, 23 Oct 2014 00:00:00 +0000
<![CDATA[The rise and demise of the National Scholarship Programme: implications for university students]]> There has been heated debate over the increase in tuition fees to £9,000 a year for many students that occurred in 2012. But another major change to the support for disadvantaged students was introduced at the same time: not only were universities required to provide details of their proposed financial support schemes and access programmes before they were allowed to charge fees above £6,000, but also the government introduced a National Scholarship Programme (NSP), designed to offer additional financial support to students via their universities. In new work funded by the Nuffield Foundation and the ESRC, IFS researchers provide an in-depth analysis of the financial support that universities have been offering their students since 2012, what role the NSP played in this provision and what are the likely consequences now that the government has announced that the NSP will, from 2015, no longer provide support for undergraduate students.

The National Scholarship Programme: introduction and demise

Before 2012, universities had to offer a bursary of at least 10% of the (then £3,375) tuition fee to every student from a household with income of less than £25,000 per year. Students received this money in addition to any grant or maintenance loan they were entitled to from central government.

In 2012, a National Scholarship Programme was introduced with the promise of £50m government funding in 2012, £100m in 2013 and £150m in 2014. Universities had to at least match the government allocated funding in order to receive it. While NSP funding can only go to students from households earning below £25,000, students from this group are no longer guaranteed an award, as they were under the previous system of bursaries.

In late 2013, the government unexpectedly cut NSP funding in 2014 from £150m to £50m and from 2015 will provide no further funding. For the 2014 intake, universities are obliged to maintain their previously-planned match funding, thus providing at least three times as much funding as provided by government. This observation looks at what financial support has been available from universities for low-income students each year since 2012, how this varies across different groups of institutions, and how levels of support are likely to change, and for whom, now that the NSP has been abolished.

Financial support for students: 2012 to 2014

Overall, we estimate that students starting university in 2014 will receive, on average, £635 per year in financial support direct from universities. Of this, £100 was in the form of fee waivers and £535 in the form of cash support (such as discounts for accommodation). Those with lower parental incomes (of less than £25,000 per year) will get rather more than this overall average – about £1,466 per year (of which 16% – around £240 – is in the form of fee waivers). This represents a small drop from the £1,604 that was available, on average, to poorer students in the 2013 intake – most likely as a result of the unexpected reduction in NSP funding from the government. This is shown in the figure below.

Figure: Average financial support per year by cohort


Research-intensive universities (such as members of the Russell Group) tend to have more generous financial support schemes than less research-intensive ones (such as members of the Million+ or University Alliance groups). For example, we estimate that students going to Russell Group universities in 2014 will, on average, receive around £1,250 per year from their university, of which around £130 is a fee waiver, while those in Million+ universities will receive around £320 per year, of which £100 is in the form of a fee waiver. Amongst those whose parental household income is no more than £25,000 per year, the respective average amounts of support are around £2,860 in the Russell Group and £720 in Million+ institutions.

As well as supporting students from poorer backgrounds, universities are also increasingly using financial support as a means of attracting high-achieving students, particularly since the relaxation of student number controls: for example, when student number controls were removed for students scoring at least ABB in 2013 (down from AAB the previous year), 18 out of the 23 universities that had AAB scholarships in 2012 changed their eligibility threshold to ABB. On average, we estimate that students with ABB or above who enrol in 2014 will receive around £1,060 financial support per year, compared with £386 for those without at least ABB.

The current system of university bursaries and fee waivers is unnecessarily complex and opaque. In around one-third of the 90 large universities we looked at, there is no transparent scheme that guarantees low-income students any financial support. The schemes also vary hugely across universities, with the most generous packages available at Russell Group institutions. It is also worth noting that a not insignificant proportion of the support available comes in the form of fee waivers, which – for many students – are of little benefit, since they are unlikely to pay off their student loans within the 30-year repayment period, with or without the fee waiver.

What is likely to happen to financial support for students in future?

With the abolition of the National Scholarship Programme from 2015, the level and distribution of direct university financial support are likely to change again. NSP spending constitutes a small part of total financial support in research-intensive universities, but a very large part for other universities. For example, we estimate that the government funding of NSP constitutes more than 20% of all financial support for low-income students at Million+ universities for the 2014 intake, while nearly 70% is the match funding that universities are obliged to provide under the NSP (according to their commitments before the unexpected cut to NSP government funding). By contrast, the respective proportions for the Russell Group are around 5% and 13%. As the NSP is to be abolished from 2015 onwards, the ability and obligation of less research-intensive universities to provide financial support will be diminished. The most likely outcome is therefore that financial support will become even more focused on highly-qualified entrants who study at research-intensive universities.

Early indications from OFFA access agreements for 2015/16 suggest that highly-selective institutions are planning to redirect money from bursaries/fee-waivers to more long-term outreach work such as summer schools and mentoring, while universities with more diverse student bodies are planning to redirect money towards helping disadvantaged students engage with their studies and settle into university life, rather than on old-style bursaries.

It therefore looks likely that bursaries for disadvantaged students will fall across all types of institutions. This may mean that disadvantaged students are worse off in the short run. However, the overall impact of these changes will depend on whether focusing on outreach activities and/or engagement activities once in university is more effective at improving access and retention for disadvantaged students in the long run.


This briefing note and accompanying observation are based on an earlier paper on the National Scholarship Programme and Bursaries, part of a project funded by the Nuffield Foundation. Our most recent analysis, taking into account the 2013 and 2014 Office for Fair Access (OFFA) agreements was funded by the ESRC.

 The Nuffield Foundation is an endowed charitable trust that aims to improve social well-being in the widest sense. It funds research and innovation in education and social policy and also works to build capacity in education, science and social science research. The Nuffield Foundation has part-funded this project, but the views expressed are those of the authors and not necessarily those of the Foundation. More information is available at

]]> Wed, 22 Oct 2014 00:00:00 +0000
<![CDATA[Will the welfare cap fit?]]> This morning the Office for Budget Responsibility (OBR) published its first Welfare Trends Report. At the request of the Chancellor George Osborne this is to be a new annual report looking at the trends in, and drivers of, spending on benefits. This important frequent review of spending on the individual components of the social security and tax credit budget is a welcome addition to the OBR’s remit. Sensibly, rather than just covering spending covered by the new welfare cap, the OBR has chosen to cover the total cost of social security and tax credits, including state pensions.

Today’s welfare trends report sets out revisions made to forecast spending on different benefits over different forecast vintages. Between March 2011 and March 2014 forecast spending in 2015–16 on housing benefit has been revised up by £2.5 billion (11%, as shown in Chart 4.4), as both the expected number of renters and the level of rents relative to earnings exceeded the OBR’s forecast. Over the same period forecast spending on incapacity benefits in 2015–16 has been revised up by £3.5 billion (34%, as shown in Chart 4.5) as the numbers expected to be receiving these benefits has increased, claimants of incapacity benefit have been transferred to employment and support allowance (ESA) slightly more slowly than anticipated, and a greater proportion of those transferred have been deemed to eligible for ESA (and for the more generous rates of ESA).

In future revisions of this type might lead to policy action being required if a breach in the Chancellor’s new welfare cap is to be avoided. The cap applies to “welfare in scope” spending: that is the total cost of social security and tax credits, excluding spending on the state pension, jobseeker’s allowance (JSA) and housing benefit for those receiving JSA. In the March 2014 Budget the Chancellor set this cap at the same level as the OBR’s forecast for spending on welfare in scope for the years from 2015–16 through to 2018–19. This means that in 2015–16 the cap is set at £119.5 billion (out of total social security and tax credit spending of £218.8 billion) rising to £126.7 billion in 2018–19.

Further details of the possible rationale behind the cap, and the way it operates, can be found in Chapter 2 of the February 2014 IFS Green Budget. Broadly, the way that the cap operates is that in every Autumn Statement the OBR will assess compliance with the cap. It will be deemed to be breached if policy action has caused forecast spending to exceed the cap, or if forecasting changes have caused forecast spending to exceed the cap by more than 2%. If breached, the intention is that the Chancellor will either take policy action to bring spending back within the cap or will obtain Parliamentary approval to increase the cap. Evidence from recent years (specifically Budget 2012) shows that forecast spending on welfare in scope can be revised up by more than 2% and therefore this 2% margin for error is not so large as to make the cap irrelevant (see slide 26 here)

The logic behind the 2% buffer is that it allows a tighter cap to apply to increases in welfare spending brought about by policy changes without the risk of relatively small fluctuations in forecast spending frequently leading to breaches of the cap. In some cases defining what counts as a policy change might not be straightforward. A newly announced increase in a benefit rate is a policy change, while a higher than expected birth rate that pushes up child benefit spending is a forecasting change. But what about an operational decision to delay or slow the rollout of a reform expected to reduce spending? The OBR’s judgements over such factors in future could make the difference between the cap being met and the cap being breached: but at least these judgments will be made by the OBR and not by the Chancellor.

The OBR also identify the likely causes of any breach of the welfare cap in future years. One such risk is higher-than-expected September CPI inflation, which would lead to higher spending on most benefits within the cap. Roughly speaking, a permanent 1% increase in the price level from September 2015 would boost forecast welfare in scope spending by 1% and at a stroke use up half of the 2% buffer. However if the Conservatives form the next Government this risk to spending could be significantly reduced. The Chancellor pledged in his recent conference speech to freeze most working age benefits in 2016–17 and 2017–18, which would transfer the risk of higher-than-expected inflation from the welfare budget to the recipients of those benefits in those years. The other main risk described in the OBR’s report is operational risk. For example, the OBR’s forecasts are now based on the (already delayed) transition from disability living allowance (DLA) to personal independence payments (PIP) reducing claimant numbers by one quarter and spending by £2.8 billion in 2017–18 (or by 17% of what it would have been). Finally, there will inevitably be other forecasting errors – largely due to things that could not be anticipated. Such errors could also lead to forecast welfare spending being increased (or reduced), potentially making the difference between the cap being met and the cap being breached.

Today’s OBR report is welcome. And the welfare cap could, in principle, help governments avoid undesirable increases in spending. But neither the welfare trends report nor the welfare cap are a guarantee of better policy making. If cuts are deemed to be the right response to higher projected welfare spending then those cuts should be well argued and well designed, rather than simply those that are the quickest or politically easiest to achieve.

IFS public finance observations are generously supported by the Economic and Social Research Council (ESRC).

]]> Thu, 16 Oct 2014 00:00:00 +0000
<![CDATA[How do the parties' fiscal targets compare?]]> With the general election now just eight months away the political parties are starting to set out what they would aim to achieve in government. In this observation, we describe what each of the three main UK parties have said about their fiscal targets and discuss how these differ from current coalition government policy and from each other. Meeting the Conservatives’ target could result in debt falling more quickly than would be the case under the rules proposed by Labour or the Liberal Democrats. But doing so would require tax rises or significantly greater cuts in public spending than Labour and the Liberal Democrats would require to meet their rules – on top of those that are already planned up to 2015–16.

What are the government’s current plans?

The current government has committed to meeting two fiscal targets:

  • The fiscal mandate states that the cyclically-adjusted current budget should be forecast to be in balance or surplus at the end of a rolling five-year forecast horizon.
  • The supplementary target requires that public sector net debt should fall as a share of national income between 2014–15 and 2015–16.

While borrowing over this parliament is on course to be substantially greater than the Office for Budget Responsibility (OBR) forecast back in June 2010 the additional spending cuts pencilled in for the next parliament are sufficient for the latest official forecasts to suggest that the fiscal mandate is being met. Indeed, it is being met by some margin, with a surplus of 1.5% of national income (or £26 billion in today’s terms) forecast for 2018–19, which is currently the last year of the forecast horizon.

However, the latest OBR forecast suggests that the supplementary target is on course to be missed. Although borrowing is forecast to fall over the coming years, the OBR still expects public spending to exceed public revenues by a significant margin in the next few years and so the stock of public debt will continue to grow more quickly than the economy. As a result, the OBR forecasts that debt will rise from 77.3% of national income in 2014–15 to 78.7% in 2015–16, before falling slightly to 78.3% in 2016–17.

Table: Latest OBR forecasts for borrowing and debt













As % GDP












Current budget deficit












Cyclically-adjusted current budget deficit












Public sector net borrowing












Public sector net debt












Source: Outturns for the current budget deficit, borrowing and debt in 2013–14 are taken from the ONS’ statistical bulletin Public Sector Finances, July 2014. Forecasts for borrowing and debt as a % of GDP are taken from Tables 4.39 and 4.43 of the OBR’s March 2014 Economic and Fiscal Outlook.

What are the Conservatives’ objectives?

What have they said?

At the Conservative party conference in September 2013, George Osborne set out his party’s objective of achieving an overall budget surplus. Specifically, he committed that “provided the recovery is sustained, our goal is to achieve that surplus in the next Parliament.”

So far, the Conservative party have made no specific commitments on the path for debt, aside from those signed up to (but currently on course to be missed) by the current coalition government. However, since future debt depends on future levels of borrowing, the commitment to achieve an overall budget surplus will imply something about the future path of debt – exactly what it implies will depend on how quickly they reach a surplus and how large a surplus they achieve.

How does this compare to current government policy?

The major difference between what the Conservative party would aim to achieve and the coalition’s fiscal mandate is that the Conservative party want to achieve an overall budget surplus – rather than just a surplus on the current budget. In other words, the Conservatives want to ensure that government revenues are sufficient to pay not only for current spending but also for investment spending.

The latest OBR forecast suggests that current government policy is consistent with the Conservatives’ objective of achieving an overall surplus in 2018–19, albeit with a small margin of error. Specifically, the March 2014 OBR forecast was for public sector net investment in 2018–19 to be 1.4% of national income, and therefore for public sector net borrowing to be –0.2% of national income (i.e. in surplus).

However, the latest forecasts for the public finances imply further deep cuts to public service spending, which have not yet been set out in any detail. To achieve the currently forecast levels of borrowing, without any further tax increases or cuts to welfare spending, the government would need to cut spending by government departments by a further 10.6% in real terms (or £37.6 billion) between 2015–16 and 2018–19. This is on top of the £8.7 billion cut that has already been set out for 2015–16.

What are Labour’s objectives?

What have they said?

In January 2014, Ed Balls set out Labour’s objectives for the fiscal position. As he said then, Labour’s objective is to “deliver a surplus on the current budget and falling national debt in the next Parliament”.

How does this compare to current government policy?

Labour’s first target relates to the current budget balance. The main difference between Labour’s target and the fiscal mandate is that Labour have committed to delivering a balance or surplus on the current budget at some point in the next parliament, whereas the fiscal mandate instead requires that the cyclically-adjusted current budget must always be forecast to be in balance or surplus after five years. A second difference between Labour’s target and the fiscal mandate is that the former targets the headline current budget balance, whereas the latter relates to the cyclically-adjusted measure.

Labour have not been precise about exactly when they would seek to deliver a surplus, or how large a surplus they would aim for. Ed Balls said that “How fast we can go will depend on the state of the economy and the public finances we inherit.”

The OBR’s latest forecasts suggest that, on current plans, the current budget will reach a surplus (of 0.5% of national income) in 2017–18, increasing to 1.5% of national income in 2018–19. Unless Labour seek to deliver as large a surplus as this (or larger), their target potentially implies a looser fiscal position than current policy – that is, they may be able to cut taxes or increase spending relative to current policy by 1.5% of national income in 2018–19. For example, if Labour chose to achieve exactly a current budget balance in 2018–19 and allocated all the extra borrowing to easing the planned cuts to departmental spending (rather than to tax cuts or increases in benefit spending), then the cumulative cut required to departmental spending would be reduced from £37.6 billion to £9.3 billion (or 2.6%) between 2015–16 and 2018–19.

The OBR’s latest forecasts also suggest that Labour could operate somewhat looser fiscal policy than current government policy and still achieve their target of debt falling within the next parliament, since the latest forecasts are for debt to fall from 2016–17 onwards.

What would the Liberal Democrats do?

What have they said?

The Liberal Democrats have stated that they will seek to balance the cyclically-adjusted current budget from 2017–18 onwards and to “significantly reduce national debt as a percentage of GDP, year on year, when growth is positive, so that it reaches sustainable levels around the middle of the next decade.”

How does this compare to current government policy?

The Liberal Democrats’ first target (to achieve a cyclically-adjusted current budget balance from 2017–18) relates to the same measure of borrowing as the fiscal mandate. It differs from the targets discussed by Labour and the Conservatives in that they plan to target a measure of borrowing that explicitly attempts to abstract from temporary strength or weakness in the economy. On the other hand, both the Conservatives and Labour have alluded to similar allowances for the economic cycle when discussing their targets, as they have suggested that whether and exactly when their targets would be achieved would be dependent on the state of the economy. The latest OBR forecasts suggest that the output gap will close in 2018–19, meaning that cyclically-adjusted measures of borrowing are almost exactly the same as headline measures of borrowing in that year (as shown in the Figure). As a result, achieving the Liberal Democrat target for a cyclically adjusted current budget balance implies almost exactly the same thing in that year as achieving Labour’s target of a headline current budget balance.

Current government policy is, according to the OBR’s latest forecast, consistent with achieving a cyclically-adjusted current budget surplus of 0.7% of national income in 2017–18 and 1.5% in 2018–19. Therefore, similar to Labour, the Liberal Democrats’ commitment could be consistent with running looser fiscal policy than the government is currently planning. For example, if the Liberal Democrats chose to achieve exactly a cyclically-adjusted current budget balance in 2018–19 and allocated all the extra borrowing to easing the planned cuts to departmental spending (rather than to tax cuts or increases in benefit spending), then the cumulative cut required to departmental spending would be reduced from £37.6 billion to £8.6 billion (or 2.4%) between 2015–16 and 2018–19.

The Liberal Democrats’ target for debt is consistent with the latest OBR forecast on the basis of current government policy, as outlined above.


The parties’ fiscal targets differ from one another and from the current government’s objectives and plans in two main ways. First, the exact measure of “borrowing” that is targeted differs. Second, the timescale over which a surplus must be achieved varies – for example, the current government’s fiscal mandate is explicitly forward looking with a rolling five-year target, while each of the three parties’ proposed new targets for borrowing seem to relate to a fixed date – that is, all of them refer to achieving the objective during the next Parliament. There are strong arguments in favour of forward-looking, rather than fixed date, targets and all the parties would be well advised to consider rephrasing their objectives in this way.

The Figure below illustrates how the different measures of borrowing being targeted differ from one another. The Conservatives and Labour have not yet been specific about when exactly they would seek to achieve budget balance. The latest OBR forecasts only run to 2018–19 but there would be one further full financial year in the next parliament, assuming it begins in 2015–16 and runs its full 5-year course.

The targets chosen by Labour and the Liberal Democrats would, in principle, allow them to have a higher level of spending and/or lower level of taxation than the Conservatives would require to meet their target for borrowing. However, for both Labour and the Liberal Democrats this would, of course, come at the cost of debt declining less quickly as a share of national income than implied by either current government projections or the Conservatives’ fiscal target.

For all the main UK parties, based on the latest official forecasts for the economy and public finances, achieving their fiscal targets will require further tax increases, or cuts to welfare spending or public services in the next parliament. None of the parties have yet provided the electorate with full details of these tough choices.

Figure: Comparing the parties’ fiscal targets

Comparing the parties' fiscal targets


Current budget deficit

The difference between current spending and total (i.e. current) receipts.

Current receipts

Government revenues, including from taxes, National Insurance contributions, surpluses made by public corporations, and interest and dividends received from the private sector.

Current spending

Government spending excluding investment spending.

Cyclically-adjusted current budget deficit

The current budget deficit adjusted for temporary rises/falls that are associated with a temporarily weak/strong economy. As with cyclically adjusted public sector net borrowing, this measure gives a better indication of the underlying health of the public finances than the headline measure.

Cyclically-adjusted public sector net borrowing

PSNB adjusted to take account of the current position in the economic cycle. This measure provides a more accurate picture of the underlying health of the public finances, as it strips out borrowing associated with temporary economic fluctuations. For example, in a downturn borrowing is likely to increase (e.g. as unemployment benefit spending rises), but should fall again as the economy recovers. The cyclically-adjusted measure of borrowing attempts to remove the temporary effects of a downturn/boom, and therefore indicates how much of borrowing is ‘permanent’ and reflects long-run weakness or strength in the public finances. During downturns, cyclically-adjusted PSNB will be lower than headline PSNB, while during booms the reverse will be true.

Fiscal mandate

One of the government’s two targets for the fiscal position, as set out in the Charter for Budget Responsibility. The current fiscal mandate states that the cyclically-adjusted current budget must be forecast to be in balance or surplus by the end of a rolling, five-year forecast horizon.

Primary balance

The difference between total receipts and spending, excluding spending on debt interest payments.

Public sector net borrowing (PSNB)

The difference between total public spending and total revenues. For this measure the ‘public sector’ is defined as including all public corporations, including the banks that were taken into public ownership during the financial crisis.

Public sector net borrowing excluding banks (PSNB ex)

This measure of borrowing is the difference between total public spending and total revenues, defining the ‘public sector’ as excluding publicly owned banks. This measure provides a better indication of the underlying health of the public finances than PSNB and is the measure most commonly referred to in public debate.

Public sector net debt (PSND)

The public sector’s financial liabilities, less its liquid financial assets; in the words, the stock of what the government owes to the private sector, less the value of assets that it could easily and quickly sell if it wanted to. This measure covers all debt and liquid assets held by the public sector, where the ‘public sector’ is defined as including all public corporations, including the banks that were taken into public ownership during the financial crisis.

Public sector net debt excluding banks (PSND ex)

This measure of debt is similar to PSND but is calculated based on a definition of the ‘public sector’ that excludes publicly owned banks: in other words, the liabilities of these banks (and their liquid assets) are ignored for the purposes of calculating PSND ex.

Public sector net investment (PSNI)

Government spending on investment, less depreciation of the capital stock. This can be thought of as the amount the government is adding to its capital stock – that is, after taking account of the wear and tear on existing assets.

Supplementary target

One of the government’s two targets for the fiscal position set out in the Charter for Budget Responsibility. The supplementary target requires that public sector net debt should fall as a share of national income between 2014–15 and 2015–16.



]]> Fri, 19 Sep 2014 00:00:00 +0000
<![CDATA[The Scottish NHS - more financially secure outside the UK?]]> The future of the welfare state, and particularly of the NHS, has taken centre stage in the Scottish independence debate in recent days. Scotland’s Deputy First Minister Nicola Sturgeon has said that a 'Yes' to independence would free the Scottish NHS from an agenda of privatisation and public spending cuts. On the contrary, say the 'No' campaign: an independent Scotland’s precarious government finances would mean a need for 'austerity plus' that would put NHS funding at risk. Gordon Brown has said he wants to 'Nail the NHS lie' of the 'Yes' campaign.

Given this political rhetoric, there is a clear need for some impartial analysis. So, in this observation we try to set out some of the facts on both recent changes in NHS spending in England and Scotland, and the prospects for the future whether Scotland is in or out of the Union.

First some facts about what we know has happened in recent years:

1. Health is already a devolved matter – that is health policy and funding is the responsibility of the Scottish Government. This means the Scottish NHS does not have to make more use of private sector providers just because the English NHS is (and indeed, it hasn’t been). And the Scottish government decides how much money to allocate to the Scottish NHS from the overall block grant it receives from Westminster. To a significant extent then, the future of the Scottish NHS is already in Scottish hands.

2. Between 2009-10 and 2015-16 spending on the NHS in England will, on currently announced plans, have risen by about 4% in real terms despite an overall fall of 13% in English departmental spending.

3. Over the same period the vagaries of the Barnett formula mean that Scotland will have had to cut overall public service spending by less – by about 8% rather than 13%. But the Scottish government has chosen to protect the NHS in Scotland slightly less than it has been protected in England. Spending on the NHS in Scotland has fallen by 1%.

4. Analysis we published last year shows this is not a new pattern. Between 2002–03 and 2009–10 – years of plenty for public services rather than cuts – real-terms health spending per person grew by 29% in Scotland compared with a 43% increase across the UK as a whole. This was despite overall public service spending per person growing by a very similar amount in Scotland (26%) and the UK as a whole (28%).

So it seems that historically, at least, Scottish Governments in Holyrood have placed less priority on funding the NHS in Scotland (and more on funding other services) than governments in Westminster have for England.

 Table 1: Overall and health budgets (£s billions, 2013–14 prices), England and Scotland


2009–10 spend

Planned 2015–16 spend

% change

UK Government




Overall “English” departments
















Overall Scottish government








Notes: 'English' Departments are defined as those where the bulk of that department’s work is devolved to the Scottish Government. English departmental spend includes Departmental Expenditure Limits (DEL) only, with the exception of the Department for Communities and Local Government where an adjustment has been made to account for the localisation of council tax benefit and business rates. Scotland spend includes the Scotland DEL (i.e. the block grant) and business rates revenues. Business rates revenues are added to the Scottish DEL to ensure greater comparability with figures for “English” departments (business rates revenues are counted as part of English DEL but as a separate item for Scotland). Ignoring Scottish Business rates revenues, Scotland’s block grant on its own is set to fall by about 10.5% in real terms.  

That is the past though. What about the future?

1. If Scotland remains part of the UK, the Scottish Government is likely to face continued cuts to its block grant until 2018–19 as the UK government continues with its efforts to eliminate the budget deficit (although again, the Barnett formula will mean those cuts are smaller for Scotland than for England). Further cuts in overall budgets will make it harder for both the Scottish and UK governments to continue to protect NHS spending – the cuts to other unprotected areas would need to be very large. But without such protection, managing the rising demand for and costs of health care would likely be difficult indeed.

2. A devolved Scottish Government could use the powers it will acquire under the Scotland Act to increase taxes to give it more money to spend. If it were to increase the Scottish rate of income tax by 1%, for instance, it would raise around £400 - £450 million a year. This would be enough to boost health spending by around 4%.

3. Independence would give the Scottish government more freedom to set spending and tax policies. It would also, in principle, have more freedom to borrow. That freedom would be constrained by the size of the debt it would likely inherit and the willingness of markets to lend. On most plausible scenarios it is hard to see how an independent Scotland could “end austerity” in the short run. In work published this summer we showed how, on the basis of the independent OBR’s oil forecasts, an independent Scotland would likely still have a deficit of 2.9% of GDP (borrowing of about £800 per person in today’s terms) by 2018-19 even if it followed current UK government tax and spending plans – plans that are forecast to lead to the UK as a whole actually having a small budget surplus by the same year. In this case an independent Scotland would need to implement bigger spending cuts (or more tax rises) than the UK as a whole or try to borrow more. This means it would likely be harder rather than easier to protect the NHS.

The Scottish Government is more optimistic about oil revenues than the OBR – hoping for £7 billion in 2016–17, instead of the £3 billion forecast by the OBR. It says this will allow it avoid further cuts to public spending after 2016–17, which could mean more money for the NHS, among other things. But in fact even with oil revenues at £7 billion a year, the Scottish government’s budget position would still be if anything slightly weaker than that forecast for the UK as a whole. If these more optimistic forecasts prove correct an independent Scotland would still need to borrow more or tax more than the rest of the UK if it wanted to increase spending on the NHS while protecting other services. It already has the power to tax more or divert resources from other spending areas. A new country inheriting a substantial debt might not find relying on additional borrowing either easy or conducive to longer run fiscal stability.

And of course, while oil revenues could turn out to be higher than forecast (as the Scottish government hope) they could also turn out to be lower than forecast.

In the short term, then, it is hard to see how independence could allow Scotland to spend more on the NHS than would be possible within a Union where it will have significant tax raising powers and considerable say over spending priorities. Previous IFS work on the longer-term outlook for an independent Scotland’s finances suggests that under a wide range of scenarios, a combination of the eventual fall in oil revenues and an ageing population could make for a tougher fiscal outlook for Scotland than the rest of the UK and hence less room for additional spending on things like the NHS. Faster economic growth in an independent Scotland would help square all these circles and allow more spending. It is possible that might happen and the Scottish Government says independence would give it the tools to do that. But such faster growth is not certain, by any means.

]]> Thu, 11 Sep 2014 00:00:00 +0000
<![CDATA[Product reformulation effective in reducing dietary salt intake]]> In 2003 the UK Government set a target of reducing the average salt intake of adults to 6g per day. To help achieve this it adopted a two pronged salt reduction strategy, encouraging voluntary product reformulation by the food industry and simultaneously running a consumer awareness campaign that highlighted the negative health risks associated with high salt intake.

Recent analysis by IFS researchers, published as an IFS Working Paper, finds that between 2005 and 2011 there was a 5.1% reduction in the average salt content (grams per 100g) of British households’ grocery purchases. It also shows that the actions of firms were crucial in driving the decline in the salt content of grocery purchases. The decline was entirely due to the reformulation of food products by manufacturers to reduce their salt content; households actually switched slightly towards saltier food products.

These findings point towards the potential effectiveness of encouraging product reformulation in reducing the intake of other nutrients that might be consumed in excess, such as sugar and saturated fat. This strategy has the potential to change the diets of individuals who might be resistant to changing their behaviour. Despite the recent marked decline in salt intake, in 2010, according to the Health Survey for England, 60% of people were still consuming more salt than the government’s targeted population salt intake of 6g per adult per day, and the average salt intake was 7.9g per person per day. Therefore the government still has a long way to go to reach its goal.

The research uses data from the Kantar Worldpanel on the food grocery purchases made for consumption inside the home by a representative sample of British households from 2005 to 2011. The sample includes 15,000 to 25,000 households at any point in time. We decompose the 5.1% reduction in the average salt content of households’ grocery baskets into the proportion that was due to:

1. Households switching between food products, for example, switching from salted to unsalted butter, or from crisps to apples

2. Product reformulation by firms to reduce the salt content of food products

We find that the decline in average salt content of grocery purchases was entirely due to product reformulation by firms. The effect of consumer switching was to slightly increase the average salt content of the shopping basket between 2005 and 2011. If consumers had not switched to saltier products, then the average salt content of grocery purchases would have fallen by 6.5% (compared to the 5.1% that actually occurred).

We split the overall effect of product reformulation on the salt content of grocery purchases into the contribution made by six broad food groups. Figure 1 shows that reformulation of products in the processed food and grains groups were principally responsible for the decline in the average salt content of households’ shopping baskets. This was primarily due to reformulation of bread, condiments, breakfast cereals, biscuits, pastry and pies.

Figure 1. Effect of product reformulation, by food group

Source: Calculated from Kantar Worldpanel 2005-2011 data.

The impact of reformulation was larger for households in lower socioeconomic groups, in part because these households, on average, buy a larger share of their groceries as processed food. However, households in the lowest two socioeconomic groups (groups D&E) switched more strongly, on average, towards saltier food products compared with households from higher socioeconomic groups, partly offsetting the greater benefit they received from product reformulation.

These results highlight one of the advantages of targeting product reformulation compared with a public information campaign; reformulation has the potential to affect all consumers, while the provision of information is likely to affect different groups of households to varying degrees (if at all). In particular, reformulation is an effective way of changing the diets of individuals who may be unable or unwilling to process the information provided to them. It may also be that these are the individuals that policymakers are most interested in targeting - households from lower socioeconomic groups bought more salt per person per day, on average, over 2005-2011.

Our results suggest that the UK government’s salt reduction programme (combining an information campaign with voluntary product reformulation) has been at least partially successful in reducing the average salt content of households’ grocery baskets. Product reformulation has been entirely responsible for this decline. There is little evidence that the information campaign had any impact, although it is possible that in the absence of the information campaign individuals may have switched more strongly to salty foods. It is also possible that product reformulation itself could have driven some households to switch towards saltier products, because they like the taste of salt. Nevertheless, these findings point to the potential for product reformulation to help the government to achieve its aim of reducing average intake of other nutrients such as sugar.

]]> Thu, 14 Aug 2014 00:00:00 +0000
<![CDATA[Household incomes stabilised in 2012–13]]> The Department for Work and Pensions (DWP) has today published its annual statistics on the distribution of household income in the UK. The latest data cover years up to and including 2012–13. On 15th July, IFS researchers will publish a detailed report, funded by the Joseph Rowntree Foundation, on what these data tell us about living standards, poverty and inequality in the UK, putting this in the context of the recent recession and post-recession period as a whole, and the longer-term. This observation briefly highlights some of the key findings from DWP’s report.

Average incomes

Average incomes were relatively stable in 2012–13, in comparison with the sharp falls recorded between 2009–10 and 2011–12. Official statistics recorded no change at the median (middle) and a small and statistically insignificant fall at the mean, after accounting for inflation.

The official figures still show median and mean incomes in 2012-13 6% and 9% below their 2009–10 peaks respectively. The large post-recession falls in incomes followed a period of slow income growth that began in the early 2000s, far pre-dating the recession. The net result is that the official measure shows both measures of average income no higher in 2012–13 than in 2001–02.

However, these headline statistics compare real-terms living standards over time using an inflation measure based on the retail prices index (RPI). This is generally agreed to overstate inflation, and has been stripped of its “National Statistics” status. Because it overstates inflation it will overstate real income falls and understate real income rises. An Annex in the DWP report shows that figures using the Office for National Statistics’ improved RPIJ measure of inflation paint a slightly less sobering picture. According to those figures, real median income actually grew slightly, by 1%, between 2011–12 and 2012–13 (though this is not statistically significant). On this basis, the level of real median income in 2012–13 was around the same level as it was in 2005–06 (rather than 2001–02 according to RPI) and 4% lower than its 2009–10 peak (rather than 6% according to the RPI).

Income inequality

Income inequality barely changed between 2011–12 and 2012–13. However, inequality was substantially lower in 2012–13 than it was before the recession (measuring incomes before deducting housing costs). This is largely because between 2007–08 and 2011–12 the pay of workers grew much less quickly than prices, while benefit entitlements tended to grow roughly in line with prices. Benefits account for a relatively large share of household income towards the bottom, whereas earnings account for a relatively large share further up. After almost two decades in which inequality had changed little, this was enough to return inequality to its lowest level since 1996–97. We would, however, expect to see recent falls in inequality start to unwind in future releases of data for 2013–14 and beyond, as real earnings stop falling while cuts to working-age social security entitlements accelerate.

The DWP also present an alternative set of figures which measure incomes after housing costs have been deducted. Recorded falls in inequality were smaller on this measure. This is to be expected, because mortgage interest rates have fallen substantially and it tends to be higher-income households which have large mortgages.

Income poverty

In 2012–13, 10.6 million individuals (17%) had a household income below the official absolute poverty line (e.g. below £272 per week for a childless couple, net of taxes and inclusive of benefits), measuring incomes before deducting housing costs. This was about 200,000 individuals fewer than the previous year, and a poverty rate no higher than before the recession. However, when incomes are measured after deducting housing costs, the number below the poverty line (e.g. below £235 per week for a childless couple) rose by about 600,000 in 2012–13 to 14.6 million (23%). This is about 2.0 million (2ppts) higher than in 2007–08. The difference is explained by the fact that housing costs have fallen by less for those around the poverty line than for the population as a whole – only the after-housing-costs measure accounts for this.

Using a relative measure of poverty, which instead uses a poverty line of 60% of median income, the figures show significant falls between 2007–08 and 2012–13. This reflects the fact that the incomes of low-income households were hit less hard (proportionately, as well as in cash terms) over that period than the incomes of middle-income households, and again this is explained by rapidly falling earnings and relatively stable benefits. Measuring incomes before deducting housing costs, relative poverty fell by 1.3 million (3ppts) between 2007–08 and 2012–13, to 9.7 million (15%). Measuring incomes after deducting housing costs, relative poverty fell by 400,000 (1ppt) to 13.2 million (21%).

]]> Tue, 01 Jul 2014 00:00:00 +0000
<![CDATA[Students’ GCSE performance plays a key role in explaining HE participation choices and outcomes]]> With exam season in full swing and discussions about the future of GCSEs and A-levels continuing, new research from the Institute for Fiscal Studies, published today by the Department for Education, investigates the role played by secondary schools – and the subjects and qualifications for which pupils study while they are there – in explaining whether students go to university and how well they do once there.

The report uses rich administrative data linking all secondary school pupils in England to the university records of those who attend somewhere in the UK. This data tells us which schools pupils attend and provides us with very detailed measures of attainment at the end of secondary school, including the qualifications and subjects each pupil studies and the grades they receive. This is linked to information on whether students go to university and if so where. It also follows them through university, allowing us to see whether they drop-out within two years, complete their degree within five years and graduate with a first or a 2.1 conditional on completing their degree.

We find substantial differences in the likelihood of going to university on the basis of secondary school characteristics. For example, 62% of pupils in the top quintile of school performance (those with the highest proportions of pupils achieving 5 A*-C grades at GCSE) go to university at age 18 or 19, compared to just 17% of those in the bottom quintile of school performance. The differences are even starker if we focus on those attending “high status” institutions (which we define as Russell Group institutions, plus those with similarly high research quality; 41 institutions in total). Pupils in the top quintile of school performance are 11 times more likely to attend a high status institution than pupils in the bottom quintile group.

Of course it is no surprise that pupils are more likely to go to university if they attend schools where both they and their classmates get good GCSEs and A-level results. Indeed once we account for pupil characteristics and performance at GCSE and A-level, we can explain virtually all of the difference in participation rates between pupils attending different types of schools. And by no means all of this is due to the characteristics of pupils before they reach secondary school, highlighting the importance of secondary school as a time to intervene to increase HE participation rates. Indeed, our work highlights the particular importance of ensuring that pupils from all schools make the right choices over the subjects and qualifications they take at GCSE, and that they maximise their chances of getting good grades at this level.

There is also an important association between school performance and HE outcomes. Pupils who went to the highest performing secondary schools are nearly 7 percentage points less likely to drop out within two years, just over 10 percentage points more likely to complete their degree within five years and nearly 18 percentage points more likely to graduate with a first or a 2.1 than pupils who went to the lowest performing secondary schools.

But once we account for differences in background characteristics and a rich set of measures of attainment at the end of secondary school, the raw relationships reverse: pupils from high performing schools are now more likely to drop out, less likely to complete their degree and less likely to be awarded a first or a 2:1 than similar pupils with similar attainment from low performing schools. This remains true if we compare pupils from different schools who attend the same universities and study the same subjects.

These differences are even starker if we compare pupils from selective and non-selective state and independent schools. For example, when comparing pupils with the same background characteristics and prior attainment, studying at the same universities in the same subjects, those from selective independent schools are 2.6 percentage points more likely to drop out, 6.4 percentage points less likely to complete their degree and 10.3 percentage points less likely to graduate with a first or a 2:1 than pupils from non-selective community schools.

These results provide some evidence that, amongst apparently similar students with similar GCSE and A level results, those from less effective state schools may on average have higher ‘potential’ than those from private, selective  or more effective state schools. Whilst recognising that this is a result that holds on average and of course not for every student, this is something which universities may want to be aware of in setting entry requirements.

]]> Fri, 06 Jun 2014 00:00:00 +0000
<![CDATA[Oil: the continuing source of disagreement]]> The debate on Scottish independence has been enlivened today by the publication of two separate reports on the outlook for the public finances of an independent Scotland, one from HM Treasury and one from the Scottish Government. These reports appear to come to quite different conclusions about the likely strength of Scotland’s public finances as an independent country compared to continuing as part of the UK.

The publication from the Scottish Government presents a much more optimistic medium term picture for Scotland than our own previous analysis and that of some other organisations. In 2016–17 the Scottish Government is projecting that total spending in Scotland would exceed revenues by (that is, they would have a net fiscal deficit of) 2.8% of national income – slightly greater than the 2.4% deficit currently forecast by the OBR for the UK as a whole in that year. This is significantly more optimistic than the 5.2% deficit forecast for Scotland that we estimated in our analysis published in early March. It is also more optimistic than the medium-term forecast for Scottish fiscal balance underlying the Treasury’s publication today.

What explains this difference?

There is broad agreement over two key facts. Public spending per capita is significantly higher in Scotland than in the UK as a whole. At the same time onshore tax revenues per capita are similar – in fact a little lower in Scotland.

The main point of disagreement is the different forecasts for revenues from North Sea oil and gas used. Our figures and the Treasury’s figures are based on the Office for Budget Responsibility’s projections. The Scottish Government report instead uses their own – higher – forecasts for North Sea revenues. Their figures assume that Scotland will receive £6.9 billion (or 4.1% of Scottish GDP) in tax revenues from offshore oil and gas production in 2016–17, rather than the £2.9 billion (1.7% of GDP) forecast by the OBR.

The figure below shows a comparison of the estimated net fiscal balance for Scotland and the UK under the Scottish Government’s and the OBR’s projections for North Sea revenues. (The dark bars are the figures contained in the Scottish Government’s report today: somewhat oddly perhaps, the Scottish Government report incorporated the higher oil and gas revenues in their projections for Scotland but not in the comparator figures for the UK fiscal position.)

Notes: The first set of bars indicate the figures contained in Scottish Government (2014), Outlook for Scotland’s Public Finances and the Opportunities for Independence. The names assigned to the three Scottish scenarios also correspond to those used in that publication.

Sources: Authors’ calculations based on: Scottish Government (2014), Oil and Gas Analytical Bulletin: May 2014; Scottish Government (2014), Outlook for Scotland’s Public Finances and the Opportunities for Independence; Office for Budget Responsibility (2014), Economic and Fiscal Outlook: March 2014.

It is difficult to know exactly what offshore oil and gas revenues will be in future. There is genuine uncertainty about exactly how much will be raised from this source over the coming years, as is demonstrated by a comparison of recent official forecasts and subsequent outturns. The chart, therefore, serves to highlight how much more sensitive Scotland’s fiscal position would be to fluctuations in this source of revenues than is the UK as a whole. This would pose its own challenges to the management of the public finances in an independent Scotland – as we and others have discussed elsewhere.

Does this mean that, if oil revenues come in as forecast by the Scottish Government up to 2016–17, the public finances of an independent Scotland would be in a sustainable position beyond that?

Unfortunately the longer-run picture is not that simple. There are two important fiscal challenges that an independent Scotland would have to face in the longer run. First, an ageing population, which would tend to increase public spending (particularly on health care and pensions) and reduce some tax revenues. Second, as reserves of oil and gas are depleted over the long run, offshore revenues are likely to decline. These pressures will tend to increase public spending and reduce revenues, and therefore increase borrowing and debt, in the absence of other offsetting policies.

In other words, even if the Scottish public finances were to look favourable in 2016–17, the long run picture may look very different. These challenges face not only Scotland but the UK as a whole. However, our previous work suggests that the issues (particularly the decline in revenues from offshore production) would be more acute for Scotland. For example, one relatively optimistic scenario considered in our analysis published last November (which implied that Scotland would achieve a net fiscal surplus in 2016–17) suggested that – once demographic pressures and the decline of North Sea oil revenues are factored in – an independent Scotland could require a fiscal tightening around 1% of GDP larger than required by the UK to get debt to a similarly “sustainable” position. The HM Treasury analysis published today sets out a similar type of analysis of the long run position for Scotland’s public finances to that which we conducted last November, with similar conclusions.

So why does the Scottish government’s analysis today suggest a different picture?

Today’s analysis from the Scottish Government illustrates how the long run outlook for Scotland’s debt might look more favourable were productivity growth, employment and population growth to be higher than has been the case for the UK in recent times. The scenarios they illustrate suggest that debt could be on a decreasing, rather than increasing, path every year from 2018–19 onwards. There are two concerns with this analysis.

First, whilst it may turn out to be the case that the Scottish economy does a great deal better after independence than it would as part of the UK, planning on this as a central assumption seems less than cautious.

Second, this analysis makes the important assumption that real public service spending per person does not grow faster if productivity growth turns out to be higher. This implies that public spending as a share of national income declines over time – something that has not been borne out in the UK over the long term. If instead public service spending per person were assumed to grow in line with productivity growth (the assumption made in our analysis and by the OBR when modelling the long run outlook for the UK public finances), the higher productivity growth, employment and population growth simulated by the Scottish Government would have little beneficial impact on the long run Scottish public finance position. The quantity and quality of public services provided would, however, be higher.

This analysis forms part of 'The Future of the UK and Scotland', a research programme funded by the Economic and Social Research Council (ESRC). It brings the best of UK social science to the debate about Scotland’s constitutional future and its implications for the rest of the UK. For further information please visit the programme website.

]]> Wed, 28 May 2014 00:00:00 +0000
<![CDATA[Budget 2014 pension reforms: increased flexibility, but for whom?]]> In Budget 2014, George Osborne announced significant changes to the way many will be able to access funds in defined contribution (DC) pensions. These changes will extend to a wider group the flexibilities already available to some. The reforms could have wide-ranging implications for the level and profile of individuals’ retirement incomes, their welfare, demand for different financial products, and demands on publicly funded means-tested support for pensioners.

This observation aims to shed light on how many people might be affected in the next few years and how these people might respond by describing the proportion and characteristics of people currently aged 55 to 74 who are likely to have been affected by the Budget announcements. We draw on data from the English Longitudinal Study of Ageing (ELSA) collected in 2010–11. This observation summarises some of the key messages that are described in more detail in tables and figures published today, which will form part of a presentation by Carl Emmerson at an event hosted by the Strategic Society Centre.

Prior to April 2014, most people with a DC pension faced a strong incentive to buy an annuity with their pension fund by the age of 75 because otherwise they faced a tax charge of 55% on assets withdrawn from the fund. They could alternatively take up to one-quarter of their fund tax-free and purchase an annuity with the remainder of their pot, with the annuity income then being taxed at their marginal income tax rate. Two groups of people had more flexibility than this, however. First, those with only small amounts of money in DC pensions were able to withdraw the whole amount, receiving one-quarter tax-free and facing tax at their marginal rate on the other three-quarters. This applied to up to three pension pots worth no more than £2,000 or to all pension funds if an individual’s total private pension wealth was below £18,000 (provided the pension provider(s) concerned allowed such ‘trivial commutation’). Second, those relatively well-off individuals who could demonstrate that they had at least £20,000 a year of other secure annuitised income could withdraw as much as they wanted from their remaining pension funds and face tax at only their marginal rate (known as flexible drawdown).

What the Budget announced was essentially that the flexibilities available to those with small pots or a high level of annuitised income would be extended to all. From April 2015 onwards, all individuals aged 55 and over will be able to withdraw as much as they want from their DC pension funds and face income tax at their marginal rate, rather than 55%.

How people, and markets, will respond to the changes announced in the Budget will depend, at least in part, on who it is that has been affected: How much DC pension wealth do they have? What other forms of retirement resources do they hold? What are their other characteristics?

A significant fraction of individuals will be entirely unaffected by George Osborne’s announcements – either because they have no money in DC pensions or because they were already eligible for trivial commutation or flexible drawdown. The tables and figures published in our briefing note today describe the number and characteristics of those aged 55 to 74 who are likely to be affected. By focussing on those already aged 55 to 74, our figures give a picture of the potential near-term impact of the reforms. (The number and characteristics of people affected in later cohorts – among whom DC pension membership will be much more widespread – are likely to be different. We do not attempt to characterise them here.) For much of the analysis we focus on those aged 55 to 59 since these individuals are more likely not to have already purchased an annuity.

We estimate that about half of men and two-thirds of women (or six-in-ten of all) aged between 55 and 59 have no money currently held in DC pensions. There are two other smaller groups of individuals who are also likely to be unaffected by the Budget announcements – those whose DC pots are below the trivial commutation thresholds (8%) and those who we estimate will have at least £20,000 of secure income from state pensions, defined benefit (DB) pensions and already annuitised DC pensions (2%). This leaves just under four-in-ten men and just over two-in-ten women (or three-in-ten of all those) aged between 55 and 59 who will experience greater flexibility as a result of the Budget changes.

As the table shows, nearly 60% of those likely to get additional flexibility are men. Those affected are generally in better health, better educated and more likely to own their own house than the average. They also expect to live longer.

The amount of money held in DC pension funds by this group of people varies widely. Half of them have less than £53,000, but one-in-four of them hold more than £116,000 in DC pension funds. However, DC pension wealth is only one component of the household wealth portfolio for this group. There is considerable variation in the importance of DC wealth within the portfolios of those likely to get greater flexibility: for around half of them DC wealth accounts for less than 10% of their total household wealth. For just over half of those likely to be affected, DC wealth is their only form of private pension wealth, but just over a quarter actually hold more wealth in DB pensions than in DC funds. The figure below shows that private pension wealth as a whole makes up on average a third of total household wealth for those who might get greater flexibility.

Overall the group who might get more flexibility has higher average wealth than those unlikely to be affected by the reforms: we estimate that those getting more flexibility have median wealth of £680,000, compared to £550,000 among those unaffected.

These characteristics suggest that those who will get greater flexibility might be relatively well-placed to receive, and to act appropriately upon, information, guidance and advice that they are given over how to manage their own finances. However, the fact that they are in relatively good health and expect to have a greater than average chance of living to older ages may complicate their retirement planning and potentially increase the costs of making an inappropriate decision. One concern that has been raised about this policy is that the greater flexibility will allow people to spend all their funds quickly and then fall back onto taxpayer-funded means-tested benefits. However, the figures presented here suggest that – in the near-term at least – this may not be a significant concern, as the vast majority of those most likely to get the greater flexibility are home owners (and thus unlikely to qualify for housing benefit) and on average have significant other assets in addition to their DC funds. The picture is likely to look different in the longer-term, however, as DC pension coverage becomes much more widespread.

Notes: Sample size = 2,015.
Source: Authors’ calculations using English Longitudinal Study of Ageing, wave 5 (2010–11).

Notes: Sample size = 2,015.
Source: Authors’ calculations using English Longitudinal Study of Ageing, wave 5 (2010–11).

]]> Thu, 15 May 2014 00:00:00 +0000
<![CDATA[Tax without design]]> Giving the annual CTA (Chartered Tax Adviser) Address today, Paul Johnson, IFS director, sets out some of the numerous ways in which tax policy in recent years has unnecessarily made the tax system more complicated, less efficient and less transparent.

The coalition government have pursued two tax policies with considerable consistency and at remarkable cost – the increase in the income tax personal allowance to £10,500 and the cut in the headline rate of corporation tax to 20%. Between them these are set to cost getting on for £20 billion a year – a remarkable cost in any parliament, let alone under current fiscal circumstances. But consistency in these two particular policies stands out among the complexity and uncertainty created elsewhere.

Responsibility for much, perhaps most, recent poor tax policy rests with both this government and its predecessor (and often earlier governments too). Gordon Brown, Alistair Darling and George Osborne all made some improvements to the tax system. But poor tax policies where this government and the last are both culpable include:

  • There is a basic rate of income tax of 20%, a higher rate of 40% and a top rate now of 45%. What is less well known is that the last government introduced a rate of 60% on a band of income starting at £100,000. This government has maintained it and effectively increased its range considerably. There is now a 60% rate of income tax on income between £100,000 and £121,000 (where it drops back to 40%). It’s hard to make much sense of that.
  • Several elements of the income tax system no longer adjust with inflation. The point at which the 45p rate becomes payable, and indeed the point at which the 60p rate becomes payable, is fixed in cash terms and has already fallen by more than 12% relative to the Consumer Prices Index since its introduction. More people will gradually be pulled into these higher rates. There is apparently no plan to stop this.
  • This government has accelerated a trend overseen by recent governments which has fundamentally altered the nature of our system of income tax, namely a continued increase in the number of higher rate taxpayers. Numbers have risen from less than 2 million in 1990 to nearly 4 million in 2007 and well over 5 million by 2015. The problem is not necessarily so much the fact of the change – there is a case for, and a case against, such a system – but the fact that this fundamental change to our tax system, which appears to have the support of the three main political parties, has never been announced or properly debated.
  • Governments of all stripes have continually cut income tax whilst increasing National Insurance Contributions (NICs) – a tax on earned income. The only reason for this is that income tax seems to be more salient and therefore increases to NIC rates are politically easier.
  • This government has followed in the pusillanimous steps of its predecessors in failing to carry out any revaluation of properties for council tax in England. Council tax, our main property tax, is assessed on relative property values which will soon be a quarter of a century out of date. This situation is becoming increasingly absurd.
  • The last government and this one raised rates of Stamp Duty Land Tax time and time again. This is one of the worst designed and most damaging of all taxes, yet revenues from it are due to hit £15 billion within just a few years. At the extreme a £1 increase in sale price can now trigger an additional £40,000 tax bill. The tax helps to gum up the entire property market.
  • Having set out a clear direction for the taxation of pensions the last government set in train an absurd complication of the regime. This government has moved in a slightly better direction but has dramatically reduced the generosity of income tax allowances for pension contributions. It has not made any long term direction clear. Once again it, like its predecessor, has ignored the role of NICs. The Labour Party appears set on further complicating the system for those with incomes above £150,000 while the Liberal Democrats want to do this for all higher-rate taxpayers too. Long term stability in pension taxation is crucial. What we have is costly short term meddling and complication.
  • Both this government and the last have continually planned for increases in fuel duties and then decided to delay, and then abandon them. The continued failure even to raise duties in line with inflation adds to a long term loss of revenue and reduced tax on the external costs created by motorists.

But the current government hasn’t just carried on with mistakes made by its predecessor. Among problems of its own creation are:

  • The failure to increase the threshold for paying National Insurance Contributions at the same time as increasing the income tax personal allowance. It is hard to think of a rationale, and none has been offered, for continuing to increase the latter while not increasing the former. More than one million low paid workers now pay NICs but not income tax;
  • The way in which the transferable allowance for married couples and civil partners is being introduced effectively creates an infinite marginal tax rate. Since the full value of the allowance is withdrawn as soon as one partner becomes a higher rate tax payer, it is possible to become more than £200 a year worse off as a result of a £1 pay rise;
  • Huge policy uncertainty has been introduced through continual (at least annual) changes to elements of business rates, corporation tax allowances and carbon taxes.

And of course the last government could fill its own hall of shame with damaging tax policy including:

  • The introduction and then abolition of a 10p starting rate of income tax. The chaos around the abolition is what is remembered. The introduction was the mistake. A new 10p band of income tax achieves nothing that could not be better and more simply achieved by an increase in the personal allowance. Yet despite the lessons one would hope they had learned, the Labour Party now promises the reintroduction of this starting rate.
  • The introduction and then abolition of low rates of Capital Gains Tax (CGT). Again it is the furore around the abolition which is perhaps better remembered but it was the introduction of a 10p rate of CGT which allowed hedge fund managers to claim they were paying lower rates of tax than their cleaners. There is still some way to go to sort out CGT.
  • The introduction and subsequent abolition of a 0% rate of corporation tax for the lowest profit companies. It seemed to surprise the government that this led to a wholly predictable upturn in the number of self employed people deciding to incorporate and hence a loss of revenue to no good purpose.

Sadly the list of examples of poor policy could be greatly extended. And if one were to include the failure to sort out existing problems then, despite some improvements made by both the current government and its predecessor, it could be extended over many pages.

And this all matters. Complexity, uncertainty and inefficiency in a tax system which takes £4 in every £10 generated in the economy costs a huge amount in lower welfare, less economic output and straightforward inequity. Even if we can’t have perfection something close to coherence and consistency would be nice. In the words of former US Treasury Secretary William Simon we should have a tax system which looks “like someone designed it on purpose”. That feels like a low benchmark, but it is one we are a long way from meeting.


]]> Tue, 13 May 2014 00:00:00 +0000
<![CDATA[House prices: what do the statistics really mean?]]> Different house price indices are giving conflicting messages

House price indices are some of the most closely watched economic indicators in the UK and it can feel as if barely a week goes by without a new set of widely-reported statistics. But what do these numbers really mean and why do they sometimes look very different? In a new IFS briefing note we explore some of the reasons for the differences between these indices and some of the methodological difficulties in constructing house price indices in general.

There are now several well-established house price indices that are published on a monthly basis. All the indices show house prices on a general upward trend, but exactly how fast they are increasing is less clear. The various indices can and do differ significantly from month to month, leading to a confusing picture of what is happening. This matters because commentary on whether the UK, and especially London and the South East, is experiencing a ‘boom’ or even ‘bubble’ in house prices, often rests on a particular house price statistic. Two sets of data illustrate this conundrum.

The first graph shows average house prices portrayed by several well-known indices in early 2014, relative to their previous peak in 2007-08. All but one (Halifax) suggest that London house prices, in nominal terms, had exceeded their previous peak. But the differences between the indices are huge. According to an index produced by LSL Property Services / Acadata (LSL Acad), based on Land Registry data, London house prices are now more than 30% above their 2007-08 peak, whilst the Halifax index suggests they have not quite reached that peak. And when it comes to house prices for the UK as a whole three indices suggest prices are still around 5% or more below their previous peak while two show them 5% above.

Note: Land Registry and LSL Acad cover England and Wales only. All figures compare February 2014 to previous index-specific monthly peak, except London figures for Nationwide and Halifax which compare 2014Q1 to previous quarterly peak.
Source: see Figure 4 in IFS Briefing Note 146.

Over longer periods one might expect the various indices to converge on a common trend. But this turns out not to be the case either. In the graph below, we look at the cumulative increase in house prices according to the various indices since 2003. All the indices show a broadly similar pattern. But they disagree on the extent of growth over the decade: the LSL Acad index shows an increase of over 55%, while the Land Registry index suggests an increase of around 35%. This is a large discrepancy.

Source: see Figure 3 in IFS Briefing Note 146

Why house price indices differ

House price indices are based on sales data, which can be misleading

House price indices are based on the prices of properties actually sold in a given period. However, because not all houses are sold in every period, house prices indices are based on a small sample of the total housing stock.

The major challenge with using sales data to measure house prices is that the mix of houses being sold from month to month can vary significantly. In particular, a simple average of sales prices might vary even if house prices stayed the same. For example, suppose sales volumes in London increase relative to the rest of the country, but the price of all houses stay the same. Because houses in London are typically more expensive than elsewhere, the average value of a house sold will increase, even though prices have in fact remained the same.

Most indices adjust sales data

Most house price indices adjust sales data to account for changes in the mix of houses being sold. The most popular method, used by the ONS, Nationwide and Halifax, is to use sales data to put a price on the various characteristics of a house, such as an additional bedroom, a garden, or the location. The values attached to each characteristic are then combined into the price of a synthetic house (e.g. a two-bed, detached house, with garden, in Reading), whose characteristics are representative of the houses being bought and sold. The LSL Acad index follows a broadly similar methodology.

In part, the indices differ because they make different choices about which characteristics matter, and how to measure their value. In addition, the indices differ in terms of how they combine individual characteristics in order to construct an ‘average’ house price. At one extreme, the Halifax index tracks the price of a house that was typical of transactions in 1983, while at the other end, Nationwide and ONS revise their definition of the average house to reflect transactions over the past two or three years, respectively (the ONS index is ‘chain-linked’ in order to provide a consistent series despite these annual revision to what constitutes the ‘average’ house). While the Halifax index has the advantage of consistency (it measures the price of the same property over time), the Nationwide and ONS may be more relevant, because they measure the price of a property which is typical of recent transactions. This may help explain the relatively fast growth in the ONS and Nationwide indices since 2009, as they reflect the shift in sales volumes towards the more expensive London market.

The Land Registry take a different approach

The Land Registry (LR) index is based on another method, known as repeat sales. The LR measures the increase in the price of each property sold in a given month compared to the price of that same property the last time it was sold. The average increase in prices is essentially the average of these individual increases in price. This approach ensures the index only measures genuine increases in price, without requiring the level of detail described in the methods above (assuming the quality of the property has not changed substantially between sales, in which case an increase in ‘price’ might just reflect an increase in quality). However, it is still affected by changes in the mix of sales: for example, a shift in sale volumes towards properties whose price is increasing rapidly will show up as an increase in the overall rate of price growth, even if in fact each property is growing at the same rate as before. More generally, properties which sell more frequently will be over-represented in this index.

Indices also use different data

Different indices are also based on different data. The LSL Acad index is based on Land Registry data which covers every registered property transaction in England and Wales, while the LR index itself uses just those properties which have been sold at least twice since 1995. The ONS, Nationwide and Halifax indices are based on mortgage data, which by definition excludes cash sales, with the ONS drawing on the largest sample, while Nationwide and Halifax use their own sales data which might be influenced by their lending policies. It is not simply the case that the index with the largest sample is ‘best’, because the indices also differ in their methodologies.

What price is right?

Given the underlying differences in both methodology and data, it is not surprising that the various house price indices provide different estimates of house price changes. It is not clear that any single method is “right” or clearly superior to the others. It probably makes sense to look at what is happening to all the indices in thinking about trends in house prices. The tendency of commentators to pick on each new piece of information in isolation as it is reported is not helpful. Ideally we would have a composite index based on all the various methods of construction. But that would be a huge undertaking. And it would raise considerable methodological difficulties of its own.

Source: see Table 1 in IFS Briefing Note 146

]]> Fri, 09 May 2014 00:00:00 +0000
<![CDATA[Cutting the deficit: four years down, five to go?]]> The UK is in the fifth year of what is now planned to be a nine-year fiscal tightening. This fiscal consolidation is forecast to total £178 billion in today’s terms by 2018–19. This is £62 billion larger than was originally planned in the June 2010 Budget. In his first Budget, George Osborne outlined a plan to complete the fiscal consolidation by the end of next financial year (2015–16). However, worse-than-expected economic news over the last few years (coupled with the fact that he is now aiming to achieve a tighter fiscal position) has led to Mr Osborne increasing the size of the planned tax rises and spending cuts. As a result, even though Mr Osborne has implemented virtually all the policies he originally announced, the consolidation is not now expected to be complete until the end of 2018–19.

By the end of this financial year, 55% of the total consolidation is expected to have been implemented. However, within this nearly all the tax increases and cuts to investment spending will have been implemented, while only just under half of the cuts to non-investment, non-welfare spending will have been done. The figure below shows the size and timing of the currently planned fiscal consolidation as a share of national income. This shows the combined effect of tax and spending measures announced and implemented by the previous Labour government and the current coalition government since March 2008.

Figure: Timing and composition of the fiscal consolidation

Timing and composition of the fiscal consolidation

Note: This Figure updates the numbers presented in Figure 1.2 of the 2014 IFS Green Budget to include the policy announcements made in Budget 2014 and some other changes to our methodology. The Green Budget contains details of the methodology and sources used to construct this figure. The data underlying this figure can be downloaded here.

]]> Tue, 29 Apr 2014 00:00:00 +0000
<![CDATA[Death to the death tax?]]> Last week the Prime Minister, David Cameron, stated that he would like to increase the inheritance tax (IHT) threshold, reviving memories of the 2010 Conservative Party manifesto pledge to increase the threshold to £1 million. Such a reform would leave IHT affecting few estates and bringing in little revenue, perhaps raising the question of whether it was worth continuing with the tax at all. But equality of opportunity concerns might point to an alternative direction for reform with the replacement of IHT with a tax on lifetime receipts. At the very least IHT is in need of reform.

Currently IHT at death is charged at a rate of 40% on estates’ values in excess of a threshold, while transfers made up to seven years before death are taxed at reduced rates and transfers made seven years or more before death are not normally taxed. Bequests to a spouse or civil partner are exempt from IHT and, since October 2007, any unused allowance is transferred to the surviving spouse. The threshold is currently set at £325,000 which means that married couples can collectively bequeath £650,000 tax-free. Agricultural land, certain business assets and bequests to UK registered charities are also exempt from IHT.

As illustrated in the figure below, receipts from IHT fell sharply following the 2007 reform and the onset of the financial crisis and associated falls in asset prices, and at less than 0.2% of national income (0.6% of all tax revenue) they remain at low to moderate levels by historical standards.

The Office for Budget Responsibility (OBR) forecasts that IHT will have raised £3.5 billion in 2013–14 and that this will increase to £5.8 billion in 2018–19. If the increase materialises, this would bring receipts as a share of national income to a level slightly above the previous peaks in in 2007–08 and 1986–87, as shown in the figure below, and would be the highest level of receipts since at least 1973–74 (not shown). The numbers paying IHT are also forecast to increase sharply from just 2.6% of deaths in 2009–10 to 9.9% of deaths in 2018–19, as shown in the figure.

Inheritance tax receipts and taxpayers

Inheritance tax receipts and taxpayers

Note: Figure shows total receipts and number of estates liable for IHT and (for years before 1986–87) capital transfer tax at death.

Source: Authors’ calculations based on data from HM Revenue and Customs, the OBR and the Office for National Statistics.

In part this predicted growth in receipts and taxpayers is due to the economic recovery and, in particular, a forecast bounce back in house prices. It is also in part caused by government policy. The last Labour government announced that the IHT threshold would be frozen at its 2009–10 level of £325,000 through to 2014–15 and the coalition government has chosen first to continue with this policy and then, in Budget 2013, to extend the planned freeze through to 2017–18. This eight-year freeze represents a cut of 22% or £70,700 relative to inflation as measured by the Consumer Prices Index. Thus while Mr Cameron wants to increase the threshold, the policy of his government is for the threshold to continue falling in real terms even after then next general election.

The last Conservative Party manifesto pledged an increase in the threshold to £1 million. Our calculations using data from HMRC suggest that in 2010–11 an increase in the threshold to £1 million would have cost the exchequer at least £1.8 billion. A £1.8 billion giveaway would have reduced IHT receipts by 70% from £2.6 billion to just £0.8 billion. It would have reduced the numbers paying IHT by three-quarters from 2.8% of deaths to just 0.7% of deaths. Of the total £1.8 billion giveaway £0.8 billion would go to those taken out of IHT as a result of the reform while £1.0 billion would go to those still liable for IHT. Much of the gain would go to those in London and the South-East since, in 2010–11, they paid half of all IHT. Given that IHT receipts are projected to increase sharply over the next few years an increase in the threshold to £1 million in, say, 2018–19 would cost considerably more than £1.8 billion.

Increasing the IHT allowance to £1 million would abolish IHT for all except a very small number of very rich families who do not plan their affairs in a tax-efficient manner (by giving away all except £1 million or £2 million of their assets at least seven years before their death, for example). And those individuals who would still be paying IHT after the reform was introduced would be paying considerably less: a couple whose joint estate was worth more than £2 million at death would attract £540,000 less tax as a result of this reform. This policy would leave IHT as a tax that very few estates were liable to pay and that raised little revenue.

The IFS-led Mirrlees Review, funded by the Nuffield Foundation and the ESRC, noted that IHT is “a somewhat half-hearted tax, with many loopholes and opportunities for avoidance through careful organization of affairs. This leads to charges of unfairness and makes a principled defence of inheritance tax difficult”. (See ch.15 Mirrlees Review) Inheritance tax is not very effective at achieving wealth redistribution. Were the threshold raised to £1 million it would also be much less effective in terms of raising money.

In these circumstances more radical change should surely be considered.

One option would be simply to abolish the tax. That would cost little more than increasing the threshold to £1 million, though the total cost would be £5.8 billion in 2018–19 relative to current policy. It would simplify the tax system and get rid of an ineffective and unpopular tax which can be criticised in any case as a source of double taxation in cases where bequests are financed from earnings that have already been taxed.

There is, however, also a case for moving in a different direction. The main justification for an inheritance tax is on grounds of equality of opportunity. Individuals do not control whether or not they are born to wealthy parents and therefore it might be seen as unfair that some are able to inherit much more than others. This argument points to taxing what each recipient gets, rather than what each donor passes on. It is also hard to see in principle why gifts made during someone’s lifetime should be treated differently from bequests at death, The simplest way to avoid IHT is simply to pass on assets above the threshold well before death – easier for the very wealthy than for the merely quite wealthy whose main bequeathable asset is their home.

The Mirrlees Review therefore argued that, if concern for equality of opportunity justifies taxing transfers to the next generation, a more logical approach than the current one – albeit with practical challenges of its own – would be to tax individuals at progressive rates on the total amount of gifts and inheritances that they receive over their lifetime. Systems broadly along these lines were proposed by the Conservative government in 1972 and by the Liberal Democrats as recently as 2006, and one has been in place in Ireland since 1976. (See ch. 8 Mirrlees Review)

Less radical reforms to IHT could also improve it by reducing the scope to avoid the tax. Such a package of reforms could involve removing or reducing the reliefs given for agricultural land and business assets and extending the reach of the tax to gifts made more than seven years before death (an increase to 15 years is current Liberal Democrat policy (See Liberal Democrat Conference)). But whatever happens IHT should not live on in its current form.

We are grateful to the Nuffield Foundation and to the Economic and Social Research Council (ESRC) for supporting much of the research that underpins this analysis.

]]> Fri, 04 Apr 2014 00:00:00 +0000
<![CDATA[No new money, yet more generous support for childcare]]> Last year’s Budget announced several major changes to the way in which the government proposed to support childcare for working families. After considering the responses to the public consultation on these proposals, the Government has today announced more details on its new Tax Free Childcare scheme and the way in which childcare will be supported in Universal Credit.

The Tax-Free Childcare scheme, due to begin in autumn 2015, involves the Government topping up payments for childcare made by parents, in a way that is equivalent to parents’ spending on childcare being free of basic-rate income tax. At the same time, the existing scheme that allows employers to give childcare vouchers free from tax will be stopped for new claimants. The new Tax Free Childcare scheme will not be available to parents who receive tax credits or Universal Credit: they are instead able to receive childcare subsidies through the tax credit or UC system. Today’s announcement means that the planned system will be significantly more generous than initially envisaged, providing support to children up to 12 straight away, will provide a higher level of support, and will provide more generous support for childcare in Universal Credit. Yet the Treasury has not increased its estimate of the total cost, as it has revised down considerably its estimate of how many families will benefit.

Specifically, we learned several things today:

  • Rather than being available initially only to children under 5 and then rolled out gradually to children under 12, the new Tax Free Childcare scheme will be available to children under 12 within its first year. The Government says an additional 0.7 million families will therefore benefit during the first year.
  • Although the rate of subsidy remains at 20%, the maximum government subsidy payable for any one child will now be £2,000 a year, rather than £1,200, which will help those (currently relatively few) parents who spend more than £6,000 a year (just under £120 a week) per child on childcare.
  • To be eligible for Tax Free Childcare, all parents in the family will need to earn at least £50 a week and less than £150,000 a year (with some exceptions where one adult is sick or disabled or on parental leave).
  • All parents on Universal Credit will be entitled to an 85% subsidy on childcare spending. This is more generous compared to the announcement in Budget 2013, when the Government said that those paying income tax would receive an 85% subsidy and other claimants would receive a 70% subsidy. (This will be tapered away from better-off families in the same way as other elements of Universal Credit.)
  • A new Early Years Pupil Premium will provide nurseries, schools and other providers of funded early education with extra money for disadvantaged three- or four-year olds

We also got a bit more information about the way that the Tax-Free Childcare scheme will work in practice and how it will interact with Universal Credit. On balance, today’s announcements suggest that the Government’s priority is to devise a scheme that is simple for parents to understand and access: they will not need to report changes in personal circumstances in real time, but only once a quarter, and they will not need to prove that they are in work, but instead only need to self-certify that they are earning more than £50 a week (with different rules applying to the self-employed). Broadly, this simplification is to be welcomed. But there are risks. In particular, policing the requirement to earn £50 a week will be difficult, as no income tax or NI is payable at (or anywhere near) that level of earnings. There will also be a very large incentive for some second earners to claim that they are earning that much: it could be worth thousands of pounds in childcare subsidy.

In what will clearly be an unwelcome decision for several companies which make money by administering the current employer childcare vouchers system, the Government has decided that the new Tax Free Childcare scheme will be delivered by HMRC in partnership with National Savings and Investments (NS&I), an Executive Agency of HM Treasury. Here, the Government has decided that simplicity to parents and the economies of scale that arise from having just one provider are more important than any benefits that would have arisen from having several companies compete to administer the accounts. It has also announced that parents will not need to pay fees to access the scheme.

Surprisingly, today’s announcements come with no new money. Extending the new Tax Free Childcare scheme to all children under 12 within its first year will cost money compared with a world where it was limited to children under 5, but the Treasury can make this announcement without altering its public spending plans because it has significantly revised down its estimate of how many families are likely to be eligible for the scheme (from 2.5 million to 1.9 million). It is not clear what has led to this dramatic change, and so we cannot judge whether these new estimates are any more plausible than the initial ones, but the fact that the change is so large suggests that the Treasury would benefit from being more open about the way it costs new policies. The increased generosity of support for childcare under Universal Credit will be matched by savings elsewhere in Universal Credit that will be confirmed in the autumn.

These announcements confirm that childcare is likely to be an important political battlefield in the months leading up to the general election. In this context, it is crucial that the Government and the opposition parties are clear about the objectives and evidence underlying their proposals. Today’s announcements indicate that the Government’s main motive is to help parents move into work. As we pointed out in the IFS 2014 Green Budget, we know remarkably little about the impact of the policies to support childcare that have been introduced in England in recent years. And there is no consistent evidence from other countries that childcare support has large effects on parental labour supply. While today’s announcements bring welcome simplifications to the new Tax-Free Childcare scheme, and an increase in generosity that will certainly be welcomed by families on Universal Credit using childcare, and better-off families who spend more than £6,000 a year on childcare, the extent to which it will deliver its intended goals is essentially unknown.

]]> Tue, 18 Mar 2014 00:00:00 +0000
<![CDATA[Scotland's fiscal position worsened in 2012–13 as North Sea revenues fell]]> Today, the Scottish Government published the latest version of its annual Government Expenditure and Revenues Scotland (GERS) publication covering 2012–13. The IFS will be publishing a full report on these figures and the last pre-referendum update of our assessment of Scotland’s fiscal position next month. But what are the key findings that jump out of the latest year of data?

For the first time in 5 years GERS suggests that Scotland's net fiscal balance, or budget deficit, was worse than that of the UK as a whole even when allocating North Sea revenues to Scotland on an illustrative geographic basis. Until now these revenues have been enough to more than outweigh the fact that public spending per head is significantly higher in Scotland than in the rest of the UK whilst non oil revenues have been much the same. But not in 2012–13.

Two things appear to have happened that explain a large part of this. First, total UK-wide North Sea revenues were substantially lower than in the recent past (£6.6 billion compared to £11.3 billion in 2011–12, for instance). Second, a smaller proportion of North Sea revenues have been allocated to Scotland in 2012–13 than 2011–12, although there have also been downward revisions to the estimated share in 2011–12’s compared to last year’s publication.

As we have said before it is important to think of oil revenues differently to other revenues because they are so volatile (and likely to run down in the longer term). These figures reaffirm the fact that an independent Scotland, like the rest of the UK, would have a substantial fiscal deficit. But higher spending in Scotland makes that deficit bigger than in the UK as a whole if oil revenues are not high enough to compensate.

The fiscal position

GERS includes two measures of the overall fiscal position. That is, the current budget balance, which is the difference between non-investment spending and revenues, and the net fiscal balance, which is the difference between total spending (including investment) and revenues. Figures are presented excluding North Sea revenues, as well as including population-based and illustrative geographic shares of North Sea revenues. Table 1 shows how each of these measures evolved between 2008-09 and 2012–13.

Table 1: Current and net fiscal balance (% of GDP), UK and Scotland, 2008–09 to 2012–13

 Current and net fiscal balance (% of GDP), UK and Scotland, 2008–09 to 2012–13

Source: Government Expenditure and Revenues Scotland, 2012–13

Excluding North Sea revenues, Scotland’s public finances suffered a very substantial deficit in 2012–13: an 11.2% of GDP current budget deficit; and a 14.0% of GDP net fiscal deficit. However, this represents a slight improvement on 2011–12, when the deficits were 11.6% and 14.7% of GDP, respectively. This improvement in the onshore budget balance is a result of onshore revenues growing more quickly (by 2.7%) than government expenditure (0.5%). Indeed, the modest improvement in the onshore budget in 2012–13 was, if anything, slightly larger than that in the UK as a whole, although the level of the onshore deficits remains much higher in Scotland (for the UK as a whole, the onshore deficits are close to the overall deficits reported in Table 1).

The present Scottish Government has traditionally focused on fiscal measures including an illustrative geographic share of North Sea revenues. This seems reasonable: if Scotland were to become independent then it seems likely that apportionment of the North Sea would be on a geographic basis. Allocating a geographic share of North Sea revenues to Scotland improves its fiscal position substantially, although a large deficit remains: a 5.9% of GDP current budget deficit (versus 11.2% without North Sea revenues), and an 8.3% of GDP net fiscal deficit (versus 14.0%).

However, this represents a deterioration since 2011–12, when the current budget deficit including a geographic share of North Sea revenues was just 3.1% and the net fiscal deficit was 5.8% of GDP. This deteriorating fiscal position reflects an estimated fall in Scotland’s share of North Sea revenues from £10.0 billion in 2011–12 to £5.6 billion in 2012–13, following falls in oil and gas production and increases in production and investment costs (which are deductible from taxable profits). Previously, the Scottish Government had been forecasting North Sea revenues of between £6.7 billion and £7.2 billion in 2012–13.

These falls in North Sea revenues also mean that, whereas including a geographic share of North Sea revenues Scotland’s fiscal position was stronger than that of the UK in every year between 2008–09 and 2011–12 (and substantially better in 2008–09 and 2011–12), in 2012–13 the fiscal position was a little weaker in Scotland than in the UK as a whole. For instance, the net fiscal balance (including investment spending) is estimated to have been 8.3% of GDP in deficit, compared to 7.3% for the UK as a whole.

Of course, the decline in North Sea revenues was not helpful to the UK public finances either. But, because most North Sea revenues are estimated to come from the Scottish portion of the North Sea (84% in 2012–13), and because the Scottish economy and tax-base is much smaller than that of the UK as a whole, a fall in this revenue stream has a much larger proportional impact on Scotland. For instance, the fall in North Sea revenues in 2012–13 was equivalent to around 0.3% of GDP for the UK as a whole, but 3.1% of GDP for Scotland. The latest figures therefore illustrate how sensitive an independent Scotland’s public finances would be to volatility in North Sea revenues. As we have argued before, and the Scottish Government’s independent fiscal commission has recognised, this would require a Scottish government to design fiscal rules that reflect this sensitivity. And, if North Sea revenues continue to fall and remain at low levels, as forecast by the OBR, further fiscal tightening in addition to that planned by the UK government would be required to return Scotland’s public finances to health. Even if North Sea revenues do rebound in the medium-term, as the Scottish Government forecast in its independence White Paper, it might be wiser to bank that money to improve the public finances, and to prepare for the longer-term fiscal challenges of an ageing population and the eventual diminishing of North Sea revenues.

Comparing GERS 2012–13 with the IFS’s previous projections

These latest official figures follow publication, last week, of updated medium-term projections by the IFS. We anticipated that Scotland’s budgetary position would weaken in 2012–13. However, our projection for the net fiscal deficit of 6.8% of GDP was somewhat smaller than the outturn of 8.3% of GDP. What explains these differences?

First, part of the difference can be explained by the fact that the decline in North Sea revenues was accompanied by a decline in the total amount of economic output produced in the North Sea which led to a fall in Scottish GDP. A lower GDP means a given cash deficit represents a larger fraction of GDP. The GERS figures take account of this fall in GDP, but our projections did not (as the focus was on the outlook to 2016–17 and 2018–19 and official GDP forecasts are not available for Scotland).

There are also differences between our spending and tax projections for 2012–13 and the outturns reported in GERS.

Our projections for spending were around £0.3 billion lower and our projections for tax revenues around £1 billion higher than the actual outturns reported in GERS. Around half of the gap for taxes is due to Scotland’s North Sea revenues being lower than we projected. Our projections were based on 94% of UK North Sea revenues being allocated to Scotland on a geographic basis, which was the share reported for 2011–12 in last year’s GERS publication. Today’s publication instead allocates just 84% of North Sea revenues to Scotland (it also has revised the share for 2011–12 down to 88%). These revisions reflect the fact that it is difficult to divide up tax revenues between the different parts of the UK when they are collected on a national basis.

If these higher levels of spending, lower levels of taxes and GDP, and the resulting bigger fiscal deficit were to persist, Scotland’s medium-term fiscal position would be somewhat weaker than our projections last week suggested. This is likely to be unwelcome news. If the government of an independent Scotland wanted to make up this gap, the additional tightening of 1.5% of GDP (8.3% - 6.8%) that would be required equates to around £2.2 billion in today’s terms. This would come on top of the 2.9% of GDP tightening (£4.2 billion) we estimated would be required to match the fiscal tightening planned by the UK government for between 2016–17 and 2018–19. And, if the OBR’s forecasts for overall North Sea revenues turn out to be correct, even then, Scotland would have a net fiscal deficit of around 2.5% of GDP in 2018–19, compared to a forecast of budget balance for the UK as a whole in the same year.


How healthy an independent Scotland’s public finances would be is obviously a key issue for the forthcoming referendum: whether an independent Scotland would need further tax rises and spending cuts, or could ease the austerity, is one factor in whether independence would make Scottish families better or worse off. The GERS figures released today, because they are backwards rather than forwards looking do not provide the answer to this question. But they do remind us of three important things:

  • That the health of an independent Scotland’s public finances would be affected by the strength of North-Sea revenues to a much greater extent than the UK as a whole;
  • That these revenues can be extremely volatile from year-to-year;
  • And that forecasting the public finances is a tricky business.

This means that in planning for independence, the Scottish Government should be cautious in its fiscal assessment, and avoid building its budget on the back of optimistic forecasts for North Sea revenues.

]]> Wed, 12 Mar 2014 00:00:00 +0000
<![CDATA[Does offering higher teacher salaries improve pupil attainment?]]> Research from the US and England shows the central importance of teacher effectiveness in determining pupil attainment. Understanding how to attract high quality teachers to the profession and to particular schools is therefore clearly crucial. In new research released today, we examine salary scales and pupil attainment in primary schools in and around London. For these schools, and for the salary differences of just under 5% that we observe, we do not find evidence that higher salary scales for teachers have much impact on pupil attainment. This suggests that if individual schools offered salary differentials on this scale, they would not necessarily attract more effective teachers.

Estimating the impact of teachers’ pay on pupil attainment is complicated as teachers' pay is largely set by national agreements; any variation in salaries tends to reflect the experience of the teacher so it is difficult to separate the impact of teacher pay from teacher experience. In innovative new research, we compare teachers in the same geographical area, with the same levels of experience (on average), who are paid different amounts because of London weighting: teachers in the London area receive higher salaries to compensate them for a higher cost of living (and schools receive extra funding in order to pay for these higher salaries). There are four pay zones in total (inner London, outer London, fringe London and the rest of England and Wales). The current starting salaries for teachers across these zones are shown below, together with the top of the upper pay scale.

Table of current starting salaries

We compare pupil attainment in primary schools close to the fringe boundary, where schools are broadly comparable in pupil composition either side of the boundary. As shown above, the salary scales for teachers just inside the London fringe area are just under 5% or around £1,000 higher than they are for teachers at schools just outside.

We find little evidence that these higher teacher salary scales increase pupil attainment in English and maths at the end of primary school. The difference in pupil attainment between schools on either side of the pay boundary is very close to zero for both English and maths. Furthermore, due to the relative precision of our estimates, we are able to rule out even quantitatively small effects (the largest possible effects covered by a 95% confidence interval would be 0.07 and 0.02 standard deviations in English and maths, respectively).

Our work is of course just one piece of evidence. It is based just on primary schools, and the results could be different for secondary schools. But our results are consistent with other research; work from the US (see here also) has suggested that teachers’ decisions about which school to work in are more sensitive to non-pecuniary aspects of the job (such as the mix of pupils at the school and the school environment) than pecuniary ones. Other work for England exploits collective bargaining for teachers to show that a 10% increase in local ‘outside’ wages for teachers (what teachers could earn outside the profession) reduces student test scores at age 16 by about 2%.This suggests that highly effective teachers may be sensitive to pay differentials when they make their initial career choices or when they choose which area of the country to teach in, but the effect is relatively small. Taken together, our respective estimates suggest that the effect of relatively small changes in pay on pupil attainment is either zero or relatively small, and that this seems to apply both to the effect of local wages and salary differentials across schools.

One further important qualification to emphasise is that we are looking at relatively small changes in pay. This doesn’t tell us much about the effect of bigger changes in pay and in particular doesn’t speak to the question of what would happen over the long run if the relative pay of teachers were to change significantly. For instance, previous research has suggested that long-run changes in relative pay are linked to the average cognitive ability of individuals who choose to become teachers.

Nevertheless, the apparent weakness of the relationship between pay and outcomes may reflect a number of issues.

It may be that, when making appointments, schools find it difficult to distinguish between more and less effective teachers. That might imply a need for better information. Indeed there is a strong case for improving the availability of information about the relationship between particular teachers and the outcomes for the pupils they teach.

Pay differentials may be effective when linked more directly to performance. There is some evidence of a modest positive effect of performance-related pay on pupil attainment, though it appears to matter a lot exactly how such systems are designed.

More generally there is a case for focussing on effective advertising and promotion of teaching. The availability of (cost) effective teacher training routes (which IFS is currently researching) is likely to be important. And there may be scope to increase the quality of applicants to teacher training.

Overall there is a remarkable lack of clear evidence about which combination of measures is likely to be most effective in attracting more high quality teachers into the profession or in attracting the best teachers to particular schools. Our research suggests that modest across-the-board pay rises are not likely to be the main answer, at least in the short run.

It is normal to highlight the need for more research. It is urgently needed in this area. In particular, it is vitally important to put together data that would make such research possible. The Department for Education collects data on pupil performance. It also collects data on teachers. It has consistently, and for a long period, refused to collect the necessary information that would allow these two datasets to be linked. This means that we remain largely ignorant about which are the most and least effective teachers and which policies are most effective in helping attract and retain them.

]]> Thu, 06 Mar 2014 00:00:00 +0000
<![CDATA[The next five years look better but tough fiscal choices remain for Scotland]]> The Office for Budget Responsibility’s (OBR) latest forecast for the UK’s public finances – published in December 2013 – presented a significantly better picture for the UK’s fiscal position over the next five years than had been suggested by their previous forecast in March. This was the result both of higher expected economic growth up to 2018–19 and the government’s decision to continue to freeze public spending in real terms for one more year in 2018–19. In a speech this evening to the David Hume Institute, IFS Director Paul Johnson will describe how this new information affects our estimate of the outlook for Scotland’s public finances over the next five years. This observation provides more detail behind that analysis: we set out here how the new forecasts from the OBR affect the medium-term fiscal outlook for Scotland that we and the Scottish government, among others, previously presented.

The OBR now thinks that the economy will perform better over the next few years, meaning that most of the major tax revenues will rise more quickly and public spending (many components of which have already been fixed in cash or real terms for the next five years) will decline more rapidly as a share of gross domestic product (GDP). As a result, the official forecast for public sector net borrowing in the UK up to 2017–18 was revised down significantly. This is shown by the red lines in Figure 1. In the 2013 Autumn Statement, the government also pencilled in one more year of freezing public spending. As Figure 1 shows, this (if it in fact occurs) will strengthen the public finances even further in 2018–19.

Between March and December last year, the OBR’s forecast for public sector net borrowing in the UK in 2018–19 improved from 2.0% of GDP to –0.1% of GDP (i.e. the OBR is now forecasting a surplus). Our estimates suggest that the borrowing projection implied for Scotland has fallen from 4.3% of GDP to 2.5% of GDP; these figures assume that all the spending cuts currently pencilled in by the UK government are also implemented in Scotland. In other words, both the UK and Scotland are now expected to have a stronger fiscal position in 2018–19 than was expected in March last year as a result of an improved economic outlook and the pencilling in of an additional year of spending cuts. However, it is worth noting that the OBR expects the better economic performance over the next five years to have reduced scope for further economic recovery after 2018–19. This means that the long-term fiscal improvement implied by the OBR’s new economic forecasts is somewhat overstated by the differences shown in Figure 1.

Figure 1: Public sector net borrowing projections for the UK and Scotland

Public sector net borrowing projections for the UK and Scotland

Notes: All figures assume that the tax increases and spending cuts currently announced and pencilled in by the coalition government are implemented, including those planned in Scotland for after the date of potential independence. Sources: Figures for ‘Mar 2013’ are from Amior, Crawford and Tetlow (2013). Figures for ‘Dec 2013’ are authors’ calculations using Office for Budget Responsibility (2013), Economic and Fiscal Outlook: December 2013.

Our estimate of Scottish borrowing of 2.5% of GDP in 2018–19 assumes that all the tax increases and spending cuts currently announced and pencilled in by the UK’s coalition government are implemented, including those planned in Scotland for after the date of potential independence. If, instead, the Scottish government did not want to cut non-debt interest public spending as a share of national income after 2016−17, then borrowing in 2018−19 would be 5.4% of GDP (rather than 2.5% of GDP) – this is shown by the dotted blue line in Figure 1. In other words, to reach a position where borrowing was at 2.5% of GDP in 2018-19 the government of a newly independent Scotland would need a fiscal tightening of 2.9% of GDP between 2016-17 and 2018-19.

Oil revenues

There was one source of revenues that the OBR took a more pessimistic view on in December than they had in March: that was revenues from oil and gas production. As the vast majority of these revenues are generated in Scotland, a downgrade to this revenue stream has a much more adverse effect on Scotland’s fiscal balance than it does for the UK as a whole. (Oil and gas revenues for Scotland implied by the two OBR forecasts are shown by the black lines in Figure 2.) As a result, the improvement in the fiscal position for Scotland suggested by the OBR’s December forecast (shown by the blue lines in Figure 1) is not nearly as large as for the UK – indeed, in 2013–14 and 2014–15 our calculations suggest that Scotland would have a weaker fiscal position than was implied by the OBR’s March forecast.

Figure 2: Comparing forecasts for Scottish oil and gas revenues

Comparing forecasts for Scottish oil and gas revenues

Notes: Scottish oil and gas revenues based on the OBR forecasts are calculated assuming that Scotland is allocated 94% of UK revenues. Sources: Office for Budget Responsibility (2013), Economic and Fiscal Outlook: March 2013; Office for Budget Responsibility (2013), Economic and Fiscal Outlook: December 2013. Scottish Government (2013), Oil and Gas Analytical Bulletin: March 2013 (

The outlook for Scotland is, however, very sensitive to the tax revenues received from oil and gas production and some forecasters take a very different view from the OBR on the prospects for these revenues over the next few years. In particular, the White Paper published by the Scottish government in November presented figures for Scotland’s fiscal position based on a forecast that Scottish oil and gas revenues would be somewhere between £6.8 billion (described as ‘Scenario 2’ in Figure 2) and £7.9 billion (described as ‘Scenario 4’ in Figure 2) in 2016–17. We estimate that the OBR’s forecast at the time implied revenues for Scotland (allocated on a geographic basis) of just £4.5 billion and that their latest forecast implies revenues of just £3.3 billion in that year.

We estimate that the OBR’s forecasts imply that Scottish borrowing in 2017–18 will be 3.6% of GDP. However, if Scottish oil and gas revenues were to turn out as projected under the Scottish government’s ‘Scenario 2’ or ‘Scenario 4’, borrowing would instead be 1.8% or 1.0% of GDP respectively. These latter figures suggest a better position for Scotland’s public finances, similar to the position for the UK as a whole, which the OBR projects will have borrowing of 1.2% of GDP in 2017–18.

It is worth noting in this context that it looks like the Scottish government’s forecasts for revenues under these scenarios have been too optimistic in 2012–13 and, with the vast majority of payments already having been made for 2013–14, that their forecasts for this year also look to be too optimistic. It remains to be seen whose forecasts might be more accurate going forwards.


The latest OBR forecasts for the UK’s public finances imply a slightly stronger fiscal position for a potentially independent Scotland in the medium-term than their previous forecasts suggested, although with higher borrowing over the next couple of years. This is good news but would require the government of a newly independent Scotland to continue with the spending squeeze currently planned by the UK’s coalition government. The latest OBR forecasts also illustrate the sensitivity of Scottish public finances to oil revenues. Neither the OBR nor the Scottish government can know for sure what will happen to these revenues. What is clear is that fiscal decisions in an independent Scotland would need to be taken in the context of considerable uncertainty over this very important part of the budget – and in the context of long term pressures both on these revenues and arising from an ageing population (as we have highlighted).

This analysis has been produced as part of our ESRC funded work on The Future of the UK and Scotland. This research programme aims to clarify some of the fiscal choices that might face Scotland were it to become independent.

]]> Tue, 04 Mar 2014 00:00:00 +0000
<![CDATA[50p tax – strolling across the summit of the Laffer curve?]]> On Satuday 25th January, Ed Balls, the shadow chancellor announced that if elected, a Labour government would return the rate of income tax payable on incomes above £150,000 to 50%. What would the effect of this be? The contention of Ed Balls and Ed Miliband is that it would raise a meaningful sum of money to help reduce the budget deficit and make for a fairer tax system. The contention of their opponents is that it could result in an exodus of talent from the UK, a reduction in entrepreneurial drive, and an increase in tax avoidance and evasion, and may actually reduce the amount of tax paid.

Who is right is important not just for the lucky few who have incomes high enough that they would be directly affected. It matters for everyone because the Exchequer is, perhaps worryingly, reliant on this very small group of individuals for a very large fraction of revenue: the 1% of income tax payers with incomes in excess of £150,000 pay somewhere between 25 and 30% of all income tax. How these people would respond to a change in tax rates can therefore have big implications for overall tax revenues.

Perhaps the best evidence we have at present is that produced by HMRC, and signed off by the Office for Budget Responsibility, in 2012. This suggested that cutting the 50p rate to 45p could reduce revenues by about £3.5 billion in 2015–16 if there was no change in behaviour by affected individuals. However, once one allows for behavioural response, their central estimate was a cost of just £100 million – a very small amount of money. The best available estimate of what reversing the cut would raise is therefore about £100 million too.

However, it is important to bear in mind that there is substantial uncertainty around this central estimate. Calculating the revenue effects allowing for behavioural response requires one to estimate the “taxable income elasticity” – the extent to which taxable income changes when the tax rate changes. HMRC’s central estimate is that this elasticity was 0.45, which is broadly in line with estimates by IFS researchers based on the last time the top rate of income tax changed – in the 1980s – and with estimates from a number of other countries. If instead the true elasticity was 0.35 (which is well within the range of uncertainty), reducing the top rate of tax from 50p to 45p will have cost the exchequer about £700 million, whilst if the true elasticity was 0.55 (again, within the range of uncertainty), it will have actually raised about £600 million. In other words cutting the top rate of tax may have cost the government a bit more than it thought or actually raised a bit. This evidence led the OBR chairman Robert Chote to conclude that, whatever the precise answer, we were “strolling across the summit of the Laffer curve”.

Has anything changed since then? Ed Balls and Ed Miliband have suggested that the most recent HMRC statistics show those paying 50% income tax have paid some £10 billion more over the three years 2010–11 to 2012–13 than was thought back in 2012 when HMRC did their analysis. The statistics in question are estimates and projections of tax liabilities based on the Survey of Personal Income, which is itself based on a sample of tax records. The versions of these tables from 2012 and 2013 do show a difference of around £10 billion in the total amount of tax paid by those paying the 50p rate in the years between 2010¬–11 and 2012–13. Is that an indication that the 50p rate was more successful in raising revenue than HMRC assumed in their analysis?

Looking carefully at the notes that accompany the 2012 tables cited by the Labour Party shows that the figures for 2010–11 to 2012–13 were projected tax payments based on how much tax was paid in 2009–10, before the 50p tax rate was introduced. Had the HMRC’s analysis of taxable income elasticity been based on these original, lower, projections for tax revenues then we might have good reason for questioning their analysis. In fact it was not. In their analysis, HMRC made use of the actual 2010–11 tax returns of people paying by self-assessment (who make up the vast majority of people who were paying the 50p rate of tax). This is pretty much the same data (bar a few late filers and those few 50p rate tax payers not required to fill in a self-assessment form) that then goes into producing the Survey of Personal Income data used in the most recent 2013 tables. These tables show 50p tax payers paid £34.5 billion in income tax in 2010–11 on an income of £86.5 billion (and HMRC’s analysis assumed their income was £87 billion). So in fact there is little additional evidence to suggest that a 50p rate would raise more than was estimated by HMRC back in 2012.

Of course, even if increasing the top rate of income tax raised little or nothing one might still consider it worthwhile if one had a very strong preference for reducing inequality. Some effect has of course already been felt since the recession hit through the combination of the 45p rate itself, a major reduction in pension tax reliefs, and the withdrawal of the personal income tax allowance for those on high incomes. Our analysis of the effect of policy changes since the recession suggests people with incomes over £100,000 have on average seen a bigger percentage hit to their incomes from tax and benefit policies than people in any other part of the income distribution. But the group with incomes over £150,000 a year remains extremely well off relative to the majority.

The uncertainty around HMRC’s estimates mean it is also possible that the 50p rate would be somewhat more effective at raising revenue than their initial analysis suggests. HMRC made their calculations at great speed on the basis of one year’s data that had only just become available. Indeed only around 95% of the data was available at the time they made the calculation. By now they have data for 2011–12 too, and soon they will have data for 2012–13 as well. Given this there is certainly a case for HMRC looking again. IFS researchers also now have access to much of the relevant data and we will also be looking at this issue over the course of the year.

But at the moment, the best evidence we have still suggests that raising the top rate of tax would raise little revenue and make, at best, a marginal contribution to reducing the budget deficit an incoming government would face after the next election.

]]> Mon, 27 Jan 2014 00:00:00 +0000
<![CDATA[Hard choices ahead for government cutting public sector employment and pay]]> After the 2013 Budget, the Office for Budget Responsibility (OBR) forecast that general government employment (that is employment by central and local government) would fall by 1 million between 2010–11 and 2017–18. However, with the Chancellor choosing to pencil in further cuts to departmental spending in 2018–19, last week the OBR suggested that a further 130,000 general government jobs would go in 2018–19, bringing the total fall to 1.1 million by 2018–19.

Despite the scale of these cuts, it was not the reductions in government employment that really stood out in the OBR’s latest forecasts. Instead it was the reductions to forecast growth in public sector pay. Public sector pay is now forecast to be 3.6% lower in 2017–18 than expected in June, when our analysis used the OBR’s March forecasts and incorporated an estimate of the effect of the 1% pay award in 2015–16. This change of view in part reflects the fact that earnings growth in the public sector has been weak so far this year, with no growth in pay in the 3rd quarter of 2013 compared with the same quarter a year before. This has led the OBR to reduce its forecast for public sector pay growth in 2013–14 to only 0.5% compared with a forecast of 2.2% in March. It has also reduced its forecast for each year up to 2016–17 compared to the March forecast.

At the Spending Round in June, based on the forecasts then available, we showed that the headline public-private pay gap was likely to return to its pre-recession level by 2015¬¬–16. As can be seen in Figure 1, this was after public pay had grown faster than private pay during the recession. The latest forecasts now suggest an acceleration of the fall in public pay relative to private: the pay differential is predicted to return to its 2007–08 level (and the level seen in 2008–09) this year – two years earlier than previously forecast. Furthermore, OBR projections imply that public sector pay is set to grow less quickly than the private sector in each of the years after 2013–14 too. This implies that by 2018–19, public sector pay is predicted to be 6.4 percentage points lower relative to private sector pay than it was before the crisis in 2007–08.

Notes: Data to 2012–13 estimated using Labour Force Survey data. Forecasts based on authors’ calculations using OBR Economic Fiscal Outlook March 2013 and December 2013. “Forecast June 2013” incorporated an estimate of the 1% pay award in 2015–16 announced in Budget 2013.

The recent trends in average public sector pay relative to private sector pay do not appear to be driven by a change in the composition of the workforces. The green line on Figure 1 shows that after we control for observed differences in individuals’ age, sex, experience and education, following a similar methodology to our previous analysis, the trends in the differential in recent years are little altered. Of course, controlling for the observed differences, the ‘conditional’ differential is lower than the average raw premium, as public sector workers are more likely to be highly educated and in professional or similar roles. More details on this latest analysis are available here.

If, as these projections suggest, public sector pay is set to fall relative to private pay by 8 percentage points between 2012–13 and 2018–19, it seems likely that public pay will fall lower relative to private pay than its level in the early 2000s when parts of the public sector had difficulties recruiting and retaining staff. This has some important implications. First, if the current forecasts are correct and private sector earnings growth continues to be higher than public sector earnings growth, some public sector employers may well find it increasingly difficult to retain and recruit high quality workers. Second, if that leads the government to want to mitigate the squeeze in public sector pay but to keep workforce costs as planned it would have to absorb even more cuts to the size of the workforce, beyond the cuts of more than 1.1 million already planned between 2010–11 and 2018–19.

In making such decisions, both the government and pay-review bodies need to pay great attention to indicators of whether the public sector is facing any difficulties in recruiting and retaining high-quality staff, and decide on settlements in light of any such evidence. While public sector pay relative to private sector pay was forecast to return to its pre crisis level by 2015–16, squeezing public sector pay may have been a relatively easy way to cut departmental spending. Given the current OBR forecasts, the choices ahead for the government now look rather harder.

This analysis has been developed during an ongoing project funded by the Joseph Rowntree Foundation, to whom we are grateful for support. Over the next year, we will seek to publish further analysis of changes to the public workforce and public sector pay.

]]> Thu, 12 Dec 2013 00:00:00 +0000
<![CDATA[A give and take Autumn Statement?]]> Now scheduled for December 5, the Autumn Statement is – for the third year in a row – set to be a “fiscal event” as the Chancellor George Osborne provides more detail on the giveaways announced during the party conference season. The data so far this year – including today’s public finance numbers – suggest that we can expect to see a welcome upwards revision to the forecast for economic growth and a downward revision to the forecast for the headline deficit. But any improvement will be small relative to the level of the deficit forecast in the Budget, and the deficit this year will still be very high by historical standards and relative to what was projected at the start of this Parliament and compared to what the Chancellor is ultimately hoping to achieve. So as he prepares for the Autumn Statement, if the Chancellor is planning to make good on the promises of giveaways made during the party conference season he should also be considering ways of raising revenue to pay for them.

The March Budget forecast – produced by the Office for Budget Responsibility (OBR) – was that the UK economy would grow by 0.6% in 2013, which was similar to the 0.9% average of independent forecasts surveyed by the Treasury in the previous month. The latest survey of independent forecasters – published on November 18 – suggests that the average growth rate now forecast for this year is 1.4%, an increase of 0.5 percentage points since the Budget. The Bank of England has also revised up its expectation of growth in 2013 by a similar amount: from 0.9% in February 2013 to 1.4% in November 2013. The outlook for growth in future years has also, on average, been revised upwards too.

The monthly data on the public finances since the Budget suggest that the headline deficit is also set to be better than the OBR’s Budget forecast. Central government receipts would need to grow by 2.8% in 2013–14 for the OBR’s forecast to be correct. The data from the first seven months of this financial year suggest that they have in fact grown by 4.6% compared to the same period last year. Receipts from National Insurance contributions, self-assessment income tax, and stamp duty land tax have grown particularly strongly over the last seven months compared to the OBR’s forecast for the year as a whole and easily outweighs a more than £2 billion shortfall in receipts from the Swiss capital tax arrangement.

Should central government receipts continue to grow at this rate for the remaining five months of this financial year (which of course they might not), and if the OBR’s forecasts for public spending and for other receipts were to prove accurate (which again they might not), then the deficit this year would be £7 billion lower than the Budget forecast. An error of this magnitude would be impressively small – at least compared to the historical record of the Treasury who, on average, made an absolute error in projecting borrowing one year out of around 1% of national income or £16 billion in today’s terms.

Of course even if the deficit forecast for this year is revised down by £7 billion, it would still be £113 billion, which – at 6.9% of national income – would be the fourth largest UK deficit seen between the end of the Second World War and 2008–09. Furthermore, a deficit of £113 billion in 2013–14, while lower than projected in the March Budget, would still be substantially higher than the £60 billion that Mr Osborne projected for this year at the time of his first Budget in June 2010.

When considering the scope for permanent net giveaways in the Autumn Statement, what matters is the extent to which any reduction in borrowing this year feeds into lower expected borrowing in future. The Swiss tax was largely one-off so the shortfall here will not persist, suggesting that underlying revenues could be £9 billion higher than previously forecast. One reason why they might not is that the OBR could judge that faster economic growth this year simply means less spare capacity remains – that is there is now less scope for the economy to grow before inflation pressures return. In this case, its growth forecast for future years ought to be revised down, and any additional strength in revenues this year compared to what was previously forecast might not persist indefinitely. The faster growth in 2013 can explain just under half of the faster growth in revenues this year, which could suggest that only around £4 billion of the £9 billion strengthening in revenues might persist into future years. Of course the OBR could judge that the faster growth this year has not been associated with much, or indeed any, reduction in spare capacity this year. In that case, more of the lower projected deficit this year might be considered to be permanent rather than temporary – giving more scope for permanent giveaways.

No doubt Mr Osborne would like to be able to use any fall in borrowing to justify a significant net giveaway in the Autumn Statement: not least because of the cost of the measures announced at the Liberal Democrat and Conservative Party conferences. In total three significant policies outlined at these party conferences imply a giveaway of around £2 billion a year, including: universal free school meals for those in their first three years at primary school from September 2014 (£600 million a year); a transferable income tax allowance for some married couples from April 2015 (£700 million); and an aspiration to cancel the increase in fuel duties currently planned for September 2014 (£700 million).

So, should Mr Osborne implement all these measures – and possibly more – without any offsetting tax rises or spending cuts? In both Autumn 2011 and Autumn 2012 Mr Osborne reacted to a worsening outlook for the public finances by allowing projected borrowing to rise considerably over this parliament and pencilling in further spending cuts for the next parliament. A symmetric approach to good news would be to plan to borrow less in this parliament and to reduce the amount of austerity planned for the next.

Mr Osborne also announced in conference season a desire to eliminate the total deficit by the end of the next parliament, so as to hasten the reduction in public sector net debt. The OBR’s March 2013 forecast was that the deficit would still be at £42 billion in 2017–18. That is to say, even after the implementation of all the austerity measures already announced up to 2015–16 plus those spending cuts pencilled in for 2016–17 and 2017–18, considerable further austerity (requiring either a further two year freeze in total public spending or a further net tax rise) would still be needed to meet Mr Osborne’s proposed new deficit target. So perhaps any talk of pre-election giveaways should wait.

IFS public finance observations are generously supported by the Economic and Social Research Council (ESRC).

]]> Thu, 21 Nov 2013 00:00:00 +0000
<![CDATA[Entry to grammar schools in England for disadvantaged children]]> There are 164 grammar schools in England, which select their pupils on the basis of performance in entry tests in Year 6. These schools educate 4% of the Year 7 pupils in England. They are concentrated in selective local authorities such as Kent, Buckinghamshire, Slough and Trafford, although a third of grammar schools are in non-selective local authorities (including London). Despite their small number, grammar schools attract a significant amount of attention. New research published today by IFS researchers, which was funded by the Sutton Trust, investigates the extent to which children from disadvantaged backgrounds are disproportionately unlikely to attend grammar schools.

Using administrative data on all children in state schools in England, the research investigates whether pupils who are eligible for Free School Meals (FSM) are particularly unlikely to enter a grammar school. FSM eligibility is a good measure of coming from a poor household, as it indicates that a parent is in receipt of a means tested out of work benefit.

Comparing the proportion of children eligible for FSM in grammar schools to the proportion eligible for FSM in the surrounding areas shows that relatively few disadvantaged children go to a grammar school. In selective local authorities, 3% of grammar school entrants were eligible for FSM, compared with 17.5% at other state schools.

Given the selective nature of grammar schools, if pupils eligible for FSM have poor academic performance, they are unlikely to pass a selective entrance test. Unsurprisingly our research found that higher academic performance in Key Stage 2 tests in Year 6 dramatically increases the probability of getting a place at a grammar school. But, as has been shown before, FSM eligible children are far less likely to have high academic performance. According to the latest statistics, about 14% of disadvantaged pupils (mostly those who are or who have ever been eligible for FSM) achieved Level 5 in both English and maths at age 11, compared with around a third of other pupils. As well as having lower attainment at age 11, pupils from poorer families are also more likely to experience other potential disadvantages, such as having English as an additional language, or going to a more challenging primary school.

In areas operating the grammar system, we find that two thirds of children who achieve level 5 in both English and Maths at Key Stage 2 and who are not eligible for free school meals go to a grammar school. This compares with only 40% of similarly high achieving children who are eligible for free school meals. Furthermore, even after allowing for a wider range of factors that may depress pupils’ academic achievement, as well as pupils’ level of prior achievement in Year 6, this difference remains at just over 12 percentage points . Therefore it is not purely low attainment that prevents FSM children from attending grammar school. We found similar patterns in London local authorities and in grammar schools located in local authorities that do not operate a full grammar system.

Eligibility for free school meals only allows us to look at the poorest group of families. To get a handle on whether these socio-economic differences extend further up the income distribution, we compare pupils’ chances of getting into a grammar school according to the level of deprivation in their local area. This shows that 4% of pupils in grammar schools live in the poorest fifth of neighbourhoods, around 21% come from the middle quintile and 34% live in the richest fifth of neighbourhoods. Children living in the richest areas are disproportionately likely to attend a grammar school compared with those living in areas close to the national average, who are in turn more likely to attend grammar schools compared with those living in the poorest areas. The differences do thus seem to extend further up the income distribution.

While pupils from more disadvantaged backgrounds are less likely to attend grammar schools, a further striking result from the research is that large numbers of pupils who enter grammar schools come from primary schools outside the English state education system. Most of these students coming in from outside the state system are likely to have attended private schools. Children attending private schools, who tend to come from more affluent families, make up 13% of Year 7 entrants to grammar schools. In comparison, on average 6% of 10 year olds are enrolled in a private school nationally. Unfortunately we do not have any measures of prior achievement for pupils in private schools, so it is impossible to compare the likelihood of enrolling in a grammar school of state and private school pupils with similar levels of prior achievement.

This evidence indicates that some parents may use private primary schooling to try to increase the probability of their child attending a selective state secondary school. Of course others may purchase individual tutoring for their children to help them prepare for the entrance exams. Surveys of individual grammar schools have suggested that a high proportion of the students in those particular schools had been tutored for the entrance exam, though the extent of tutoring nationally for this purpose is not known. Clearly these are advantages which children from poorer backgrounds are unlikely to benefit from.

What drives the differential in attendance at a grammar school between pupils from different socioeconomic backgrounds? Unfortunately, this research is unable to pin down the exact mechanism for the under-representation of poorer pupils in grammar schools. It may be that children from poorer backgrounds are less likely to apply to a grammar school in the first place, or that those who do sit the exams are less likely to pass the entrance exam, perhaps due to lack of preparation for that kind of test, or a combination of both.

Future research is needed to determine exactly what is driving this grammar school attendance gap between richer and poorer pupils. This is an important research question as if it is the result of low applications then the policy response should focus on encouraging greater numbers of applications. If applications are similar, then the focus should be on the nature of the test and levels of preparation.


This paper, Entry into Grammar Schools in England is authored by Jonathan Cribb (IFS), Luke Sibieta (IFS), and Anna Vignoles (Cambridge and IFS).

It was published as part of a larger report, Poor Grammar: Entry into Grammar Schools for disadvantaged pupils in England by Jonathan Cribb, Luke Sibieta, Anna Vignoles, Amy Skipp, Fay Sadro and David Jesson. The report has been funded by the Sutton Trust.

IFS receives significant funding from the ESRC through the ESRC Centre for the Microeconomic Analysis of Public Policy at the Institute for Fiscal Studies.

]]> Fri, 08 Nov 2013 00:00:00 +0000
<![CDATA[The crucial role of good evidence in evidence-based policymaking]]> In a time of continuing fiscal austerity, policymakers increasingly want to know ‘what works’ and for whom, in order to target scarce resources on those who will benefit most and to ensure that policy has the desired impact upon those it is designed for. Basing policy decisions on evidence is undoubtedly a good thing – but only if the evidence used is robust, unbiased and methodologically sound.

Of particular concern is whether the introduction of a new policy or a change in an existing policy would cause (i.e. directly lead to) a change in the relevant outcome of interest, or whether the two are simply correlated. What do we mean by this? Two factors are causally related if movement in one factor causes a change in the other. For example, heating water causes water to boil. Two factors are correlated if they move together, but one factor doesn’t cause the other. This occurs when another factor (or multiple factors) influences both. For example, higher ice-cream sales are correlated with lower water levels in reservoirs, but greater ice-cream consumption does not cause water levels to fall; both factors are driven by external weather conditions.

This distinction is crucial when thinking about how to use evidence to inform policymaking. It would be very worrying indeed if a government thought it could increase water levels in reservoirs by taxing ice-cream sales. But it is easy to think of many rather less silly examples, particularly in the social sciences, where the relationships between factors are more complicated, and it is harder to say with any degree of certainty whether one causes the other, or whether the two are simply correlated, both driven by one or more other factors.

New IFS research funded by the Nuffield Foundation on the link between parents’ marital status and relationship stability and child outcomes illustrates these challenges. It shows that married parents are, on average, less likely to separate while their child is young than cohabiting parents. It also shows that children born to married parents have slightly higher cognitive and socio-emotional development, on average, than children born to cohabiting couples; they are also less likely to engage in risky and antisocial behaviours, such as underage smoking, drinking and cannabis use. Does this evidence prove that marriage causes improvements in relationship stability and child development?

The short answer is no.

We know that married parents differ from cohabiting parents in many ways other than their marital status (and in ways which we can be confident were determined before they decided whether or not to get married). For example, they are from different ethnic backgrounds, have typically grown up in more advantaged homes themselves, and are more highly educated. To the extent that these factors influence both the likelihood that a couple marries, and the likelihood that they stay together (or raise a child with higher cognitive or socio-emotional development), we can think of these as the ‘external weather conditions’ in the example above: ignoring them gives a misleading impression of the effect of marital status on relationship stability or child development.

Our research shows unequivocally that accounting for these other characteristics is important: once we compare cohabiting and married parents who ‘look the same’ in all respects other than their marital status (i.e. are from the same ethnic background, have grown up in similar home environments, have similar education levels, and so on), we substantially reduce or even eliminate the link between marital status and relationship stability or child development. This suggests that much of the raw relationship is due to the fact that different types of people choose to get married, rather than that marriage has a large causal effect on relationship stability or child outcomes.

There is an important caveat to our findings, however: we are not always able to eliminate the difference in outcomes between married and cohabiting parents using only characteristics that are fixed (such as ethnicity) or observed in childhood, long before marriage decisions were taken. In some cases, we must rely on the inclusion of characteristics (such as household income and housing tenure) that are observed after a couple’s decision to marry or cohabit, and so have the potential to be affected by marriage. Our judgement – and the view of most other academic research on this topic – is that these effects are likely to be small. But as a result, it is possible that we could be understating any potential positive impact of marriage if characteristics beneficial to child development or relationship stability are strongly influenced by the decision to marry.

Having said this, however, even if a statistically significant association does remain between parents’ marital status and relationship stability or child development, this estimate does not necessarily represent the causal effect of marriage on these outcomes. This is because there are likely to be many factors that influence both the likelihood of getting married and a couple’s relationship stability or their children’s development that are difficult to observe and account for in our analysis, but that were apparent before their decision to marry. Good examples might be a couple’s level of commitment to one another, or their attitudes to childrearing. Without access to richer data including these characteristics (or a source of exogenous variation – an external factor that affects the likelihood of getting married), it is impossible to know whether any significant association would remain after accounting for all of the ways in which married and cohabiting couples differ.

Understanding whether and to what extent factors are causally related or simply correlated is crucial from a policy perspective. Policy decisions should ideally be based on evidence of a causal relationship. This is not always possible, however, in which case it is important to be clear about exactly what we do, and don’t, know.

The work we are publishing today on the link between parents’ marital status and relationship stability and children’s outcomes is a good example of where one has to be very careful about the conclusions one is drawing. There is a strong correlation between parents being married and children who are more successful academically and in other ways. But there is not good evidence of a causal link – though we can’t say for sure that such a link does not exist. Researchers must be careful not to interpret or present statistically significant associations as evidence of causation. In turn, policymakers must be cautious about using such associations as a basis for policymaking.

The Nuffield Foundation has funded this project but the views expressed are those of the authors and not necessarily those of the Foundation –

]]> Fri, 25 Oct 2013 00:00:00 +0000
<![CDATA[EU Commission labels UK Patent Box harmful tax competition]]> The EU Commission has opined that the UK’s Patent Box breaches the code of conduct for business taxation, which aims to prevent countries from operating policies that result in harmful tax competition. The Patent Box provides a 10% rate of corporation tax for the income deemed to be derived from patents. The policy, originally announced by the previous Labour Government, came into effect in April this year. Once fully in place it is estimated by the Office for Budget Responsibility to cost £1.1 billion a year. One aim of the policy is to encourage innovative activities in the UK. The government highlighted that the policy focuses on patents, as opposed to other forms of intellectual property, because they have "a particularly strong link to ongoing high-tech R&D".

In previous work IFS researchers have highlighted some of the concerns surrounding the Patent Box. Notably, the policy is poorly targeted at incentivising research. This is largely because the policy targets the income from a successful idea and not the underlying research activities. Much of the benefit will accrue to large profitable firms that earn the majority of income from patents. The Patent Box may help to ensure that such firms continue to locate activity in the UK. However, the policy does not require that the innovative activity underlying a patent take place in the UK. In addition, the extent to which the policy succeeds in attracting new activity has likely been eroded by the introduction of similar policies in 10 other European countries.

The Code of Conduct group also raises concerns that the Patent Box may not be sufficiently related to the real activities that the policy aims to promote. They have challenged the specific design of the UK regime based on their assessment that it breaches two provisions. First, the Patent Box may grant tax advantages without requiring any real economic activity in the UK. This may arise if, for example, a firm owns the intellectual property in the UK but conducts research and commercialisation in other countries. Second, the rules for determining eligible profits are deemed to depart from internationally accepted principles. In contrast to other countries, the UK Patent Box does not require profits to be associated with individual patents. This potentially allows for a broad scope of income to be included in the provision. In both cases the Code of Conduct group have raised concerns that the UK rules are not sufficiently transparent and may lead to multiple interpretations, some of which allow abuse of the regime.

The code of conduct is not binding, although it has tended to be followed by governments. The Commission's assessment will be discussed by Member States at a meeting of the Code of Conduct Group on 22 October. The UK government will likely argue that the policy does not breach the provisions. If the Commission upholds the claim the UK government will come under pressure to amend the policy.

However, the broader issue is whether Patent Boxes in general result in tax competition that lowers revenues for governments without providing offsetting benefits. The regimes in Belgium, France, Spain and the Netherlands have previously been reviewed and not deemed to represent harmful tax competition. Recent research at IFS describes the range of policies available in the UK and 10 other European countries, including France, Spain and the Benelux countries. Many offer much lower rates than the UK (including 0% in Malta and 2% in Cyprus) and include a broader range of intellectual property. In some respects elements of their design may be preferable to the UK Patent Box. But there are also regimes that are more generous than the UK and more likely to attract income independently of economic activities. While the precise designs of regimes are different, in practice many of their economic effects are likely to be similar. All raise concerns over tax competition. There is now a strong case for the EU Commission to undertake a full review of the policies.

]]> Fri, 18 Oct 2013 00:00:00 +0000
<![CDATA[Tax revenue in England, Scotland, Wales and Northern Ireland]]> HM Revenue and Customs (HMRC) yesterday published estimates of the revenue raised by different HMRC taxes in England, Scotland, Wales and Northern Ireland. These taxes cover 80% of total revenue, since they do not include revenue not collected by HMRC, such as council tax, business rates and vehicle excise duty. The results for 2012–13 are summarised in the figure below. This is the first time that most sources of revenue have been broken down to show the amount raised in Wales and Northern Ireland separately and what is particularly striking about these estimates is how much lower taxes per person in these areas are than in England or Scotland.

Government revenue per person across the UK, 2012-13

Government revenue per person across the UK, 2012–13

Note: ‘Capital taxes’ are capital gains tax, inheritance tax and stamp duties on shares and property. The HMRC estimates run from 1999–00 to 2012–13, although for all the main taxes the 2012–13 (and often 2011–12) ‘estimates’ simply assume that the share of revenue is the same as in previous years.

Source: Authors’ calculations using HMRC statistics and 2012 population totals from the Office for National Statistics.

Wales and Northern Ireland

HMRC tax revenue per person in 2012-13, excluding revenue from North Sea oil and gas, is 26% lower in Wales (at £5,400) and 23% lower in Northern Ireland (£5,700) than in UK as a whole (£7,300). This largely arises because Wales and Northern Ireland have less income and wealth than the rest of the UK and correspondingly raise less revenue per person from all the main taxes on earnings, savings and profits: income tax, National Insurance contributions (NICs), corporation tax, capital gains tax, inheritance tax and stamp duties all yield at least 25% less revenue per person in both Wales and Northern Ireland than in the UK as a whole.

Wales also brings in significantly less VAT revenue per person than the UK average. But in Northern Ireland VAT revenues are only slightly lower than in the UK as a whole – because Northern Irish households (or non-household actors that cannot reclaim VAT on their purchases such as VAT-exempt traders and some public sector bodies) spend more of their income on goods and services subject to VAT. And Northern Ireland actually provides 18% (£76 per year) more revenue from fuel duties per person than UK as a whole and a remarkable 79% (£120 per year) more tobacco duty.


North Sea oil and gas aside, tax revenue in Scotland (£7,100 per person in 2012–13) looks much more like that in the UK as a whole (£7,300). Scots do pay £290 per year less in income tax on average, partly because incomes in Scotland are more equally distributed, with fewer of the very high-income individuals who provide such a large share of income tax revenue in the UK as a whole. But Scots contribute slightly more in VAT and in alcohol and tobacco taxes.

Unlike for Wales and Northern Ireland, these patterns were largely known already, since the Scottish government already produces estimates of Scottish revenues (the latest being for 2011–12) in Government Expenditure and Revenue Scotland (GERS) . In most respects HMRC have used a similar methodology and produced similar estimates to those in GERS.

In three areas, however, HMRC attributes somewhat lower revenues to Scotland than GERS does: onshore corporation tax (15%, or £0.4 billion, lower), taxes on North Sea production (12%, or £1.3 billion, lower) and stamp duty on shares (40%, or £0.1 billion, lower).

Onshore corporation tax

UK corporation tax applies to profits created from economic activities located in the UK. Where firms operate in more than one part of the UK, it is difficult to know how much of their profits are attributable to activities in each location, and companies are not required to divide up their profits in this way.

Both GERS and HMRC approximate the division of onshore corporation tax revenue based on estimates of the location of profits. GERS uses a regional measure of profits contained in the ONS regional accounts that is predominantly based on the share of wages earned in Scotland. HMRC allocates individual companies’ profits to regions, largely based on the location of their employment, and aggregates this to get the share of profits in each location.

These different methods yield somewhat – albeit not radically – different estimates. Onshore corporation tax revenue in Scotland in 2011–12 is estimated to be £3.0bn (9.0% of UK total) in GERS and £2.5bn (7.7% of UK total) by HMRC. This £400m difference is equivalent to about 1% of all Scottish revenue.

Neither of these estimates is clearly superior to the other, and both may be some way off. Profits are not necessarily generated in proportion to the number of employees, or their wages. Some employees may be more instrumental in generating profits than others; and profits also arise from capital assets – both physical (such as buildings and equipment) and intangible (such as intellectual property and brand value) – the location and contribution of which may differ from the location and wages of employees. Calculating how much of a company’s profits are attributable to economic activity in different locations is conceptually and practically difficult and is the source of many problems in international corporate taxation.

North Sea oil and gas

How North Sea oil and gas revenues would be divided in the event of Scottish independence is uncertain. Both HMRC and GERS show two illustrative possibilities. One divides North Sea revenues according to Scotland’s share of the UK population. Unsurprisingly, on this basis GERS and HMRC both report that Scotland’s share of North Sea revenues in 2011–12 was 8.4%.

The other approach taken is to divide revenues on a geographical basis according to the location of individual oil and gas fields. On this basis GERS estimates Scotland’s share in 2011–12 at 94% (£10.6 billion), while HMRC’s estimate is 83% (£9.3 billion). This disparity is not because they divide fields between England and Scotland differently: both use a boundary that was established in the Scottish Adjacent Waters Boundaries Order 1999. Rather, it derives from differences in their models of how much taxable profit arises from different fields.

Stamp duty on shares

GERS allocates revenue from stamp duty on share transactions to Scotland based on the proportion of share owning UK adults that are resident in Scotland. That is a poor guide to what Scotland’s share of stamp duty revenue would be under independence. The number of people owning shares does not tell us the value of shares traded each year (the tax base). More fundamentally, the location of the shares’ owners is not the relevant consideration for stamp duty, which is paid on transactions of shares in UK registered companies.

HMRC's approach recognises this, allocating stamp duty revenue according to companies’ registered addresses (weighted by share turnover). HMRC’s lower allocation of stamp duty revenue to Scotland – 5.0% (£139 million) rather than the 8.3% (£231 million) recorded in GERS for 2011–12 – is therefore the better guide.

The new HMRC statistics are a welcome additional source of information. In some cases the estimates are only rough approximations to tax receipts in different locations, and further devolution or Scottish independence might change the location of revenues. However, these figures reveal some interesting differences across different parts of the UK.

This work has been funded by the Economic and Social Research Council through its ‘Future of Scotland’ programme.

]]> Thu, 03 Oct 2013 00:00:00 +0000
<![CDATA[The new tax break for some married couples]]> Note: The fourth paragraph of this observation was revised on 1st October 2013 to correct errors in the estimates of the numbers of families eligible for this tax cut. These corrections are indicated in the text.

The Prime Minister David Cameron has announced how the Government proposes to recognise some marriages and civil partnerships in the income tax system. From April 2015 it plans to make up to £1,000 of the income tax personal allowance transferable between adults who are married or in a civil partnership, so long as the higher-income adult is a basic-rate taxpayer. We estimate that this would cost the exchequer around £700 million per year. The precise cost will depend on the rate of take-up, as people will presumably have to make an active claim to HMRC to benefit, and the extent to which individuals change their behaviour in order to qualify.

The proposal would work as follows. If an individual were not using all of their income tax personal allowance – because their income was less than the allowance, which is set to be £10,230 per year when the policy is introduced – then they would be able to transfer up to £1,000 of any unused allowance to their spouse. This transferred allowance would lower the spouse’s tax bill by up to £200 a year: the amount of basic rate (20%) income tax that would be paid on £1,000. However, the transferred allowance will not be available to higher rate or additional rate taxpayers – those with taxable incomes exceeding £42,285 in 2015–16.

As the maximum gain is less than £4 per couple per week, effects on incomes and incentives (which we set out below) would be small. The social message sent by the tax break looks more significant than its financial consequences for families, as the Prime Minister has stated himself. But as a structural change to the tax system it may ultimately turn out to be more important. First, it would re-introduce an incentive to marry in the income tax system, just when Married Couples’ Allowance - now available only where one spouse was born before 6 April 1935 - had been almost entirely phased out. The benefits system affects many people’s financial incentives to cohabit (either married or unmarried) rather than live apart, mostly in the form of implicit ‘couple penalties’ (IFS researchers have quantified these here). But only the inheritance and capital gains tax systems currently affect financial incentives to marry, albeit providing big incentives to do so for the relatively wealthy. Second, the transferral of allowances between spouses would re-introduce an element of joint income taxation. This is inescapable if the aim is to assess people based on family characteristics (in this case, to target single-income families) as our benefits system already does, rather than based only on individual characteristics. There is a much wider, principled debate behind all this about the role of joint versus individual assessment.

Who would gain and who would not?

The families that would qualify for the tax break are couples in marriages or civil partnerships where one individual is a basic rate taxpayer and the other does not pay any income tax - that is, 3.4 3.9 million of the 12.4 million (28% 31%) couples in a marriage or civil partnership. This includes 1.2 1.4 million of the 7.8 million families with children. 2.3 2.7 million of the 3.4 3.9 million eligible families have someone in work. The rest are mostly married pensioners. The eligible families tend to be around the middle or lower-middle of the income distribution: most are in the third to sixth income decile groups. As neither two-taxpayer couples nor higher-rate taxpayers would benefit, the cash would go to families with somewhat less high incomes, on average, than under a straightforward increase in the personal allowance that cost the same amount.

Three sorts of couples in a marriage or civil partnership will not be eligible for the tax cut:

  • Couples whose members both have incomes above the personal allowance.
  • Couples containing a higher-rate (40%) or additional-rate (45%) income tax payer.
  • Couples whose members both have incomes below the personal allowance: they would pay no income tax in the absence of the policy, so cannot benefit from any income tax cut.

Clearly then, the policy is not a general recognition of marriage in the income tax system. It is also worth noting the likelihood that take-up will be less than 100% among those eligible, as an active claim to HMRC will presumably be required.

Work incentives

The policy would benefit one-earner married couples but not zero- or two-earner couples, so it would increase the financial incentive to be a one-earner married couple. In particular, it would strengthen the incentive for married couples to have someone in paid work (provided that they would be a basic-rate income taxpayer), since they would pay less tax on their earnings. But married couples containing a basic-rate taxpayer would face a weaker incentive to have the second person in work (assuming that their potential earnings would take them above the personal allowance), because this would result in a reduced allowance for their working partner.

Since the transferred personal allowance would not be available to higher rate taxpayers, workers benefiting from a transferred allowance would have a weaker incentive to increase their taxable income above £42,285 (or a stronger incentive to make more pension contributions or charitable donations, which can be deductible from taxable income, to remain in the basic rate band). Indeed some could be worse off after a pay rise, or better off after a pay cut, because the transferred allowance will be withdrawn in ‘cliff-edge’ fashion – that is, withdrawn in its entirety once income rises above a threshold rather than tapered away gradually over a range of income. Income tax liability would jump by £200 per year when taxable income crossed £42,285, and hence post-tax income would be higher if taxable income were just below this threshold than if it were just above it.

One striking feature of the policy is that it complicates the income tax system. A transferable personal allowance for married couples capped at £1,000 and then withdrawn using a cliff-edge at the higher-rate threshold is not the simplest to understand. It is three years since another cliff-edge at the higher rate threshold was announced at the 2010 Conservative Party conference as a way of effectively means-testing Child Benefit, only to be removed and replaced with a less egregious taper at Budget 2012. The amounts involved here are less than in that case, which perhaps explains the willingness to cliff-edge again rather than implement a taper. Nevertheless, it is difficult to escape the conclusion that an income tax system which makes some people worse off after a pay rise has something wrong with it. Simpler ways to provide more support to working low- and middle- income married couples would include a higher work allowance for married couples in Universal Credit (with, if the government wished, a further restriction to one-earner couples).

]]> Mon, 30 Sep 2013 00:00:00 +0000
<![CDATA[When should summer born children start school?]]> Research from the IFS and others consistently finds that, on average, pupils born later in the academic year perform significantly worse in school than those born at the start of the academic year. As well as achieving lower test scores and assessments from teachers, on average, children’s confidence in their academic ability is also affected. Many parents of summer born children are therefore concerned about how well their child performs at school, and some have called for more flexibility in the timing of entry to primary school as an appropriate policy response. (Of course, such a policy may have other goals, aside from reducing differences in educational attainment, but it is this objective that has been cited in the latest debates.)

These concerns seem to have been heeded by the government: the childcare minister, Elizabeth Truss, has recently expressed concern that local authorities are not doing enough to offer parents flexibility over when children can start school, and guidance issued by the Department for Education in July highlights that there is no statutory barrier against delayed entry (where a child starts school a year later with a lower age group, becoming the oldest in the class), as well as the more common deferred entry (where children start school later but join the year group that is correct for their age) .

It is, of course, understandable that parents are concerned about the welfare of their own children. But the government has wider objectives; they must think about how the education system affects all pupils, as well as particular subgroups, and here the decision is not always so clear cut.

We know that, on average, deferred entry is not in the interests of summer-born children. IFS research has compared the performance of children who start school across areas of England which operate different admissions systems. We find that deferred entry to school does not close the gap in educational attainment between those born at the start and end of the academic year. In fact, on average, those born later in the year benefit slightly from starting school at the beginning of the academic year with their older peers, rather than joining them up to a year later. This is because the benefit of additional time in school more than outweighs the disadvantage of starting school slightly younger.

Assessing the impact of delayed entry to school is less straightforward, as this practice is not commonly observed in England. It is possible that it would benefit some of those who become one of the oldest in their class, rather than one of the youngest. This flexibility may have adverse consequences for other children, however: those that do not delay entry will become the youngest in the class and may have peers that are over a year older. This increase in the age-range of pupils may make teaching harder, and may simply shift the problem of lower attainment and self-belief to other children. Moreover, there are still differences in attainment between children born at the start and end of the year in countries in which it is common for children to delay or defer entry to school, and some that follow such policies (e.g. Scotland) are actually thinking about reducing or removing these flexibilities.

More importantly, IFS research has concluded that it is the age at which a child sits a test, rather than the age at which they start school, that is the main determinant of differences in attainment between those born at the start and end of the academic year. This suggests that a policy of providing age-adjusted test scores, so that a child receives feedback about their attainment relative to others their age, rather than others in their class, would help address the educational inequalities we observe. Children’s self-belief may also improve with timely and regular feedback of this kind, although some differences, such as the higher proportion of summer-born pupils identified as having special educational needs, may require additional policy responses. Moreover, unlike other policies designed to reduce these attainment gaps, providing age adjusted test scores benefits all children, as everyone is given appropriate feedback; older children that should receive more support will be identified, and younger children that are doing well for their age will be reassured.

It is important to note that IFS research cannot be used to conclude whether increasing the statutory school starting age from 5 to 6 or 7 (as suggested by a group of educationalists today) would be beneficial for children or help address the inequality between those born at the start or end of the academic year: a policy of this kind has not been trialled in England. Our research suggests, however, that within the confines of the current system, allowing deferred or delayed entry to school is highly unlikely to eliminate the disadvantages faced by summer-born children – as we find that it is age at test, rather than age of starting school, that matters most – and indeed may be detrimental to the school experiences of others.

]]> Thu, 12 Sep 2013 00:00:00 +0000
<![CDATA[Further falls in income across the distribution in 2011–12]]> The Department for Work and Pensions (DWP) has published its annual statistics on the distribution of income in the UK. The latest data cover years up to and including 2011–12. IFS researchers published a detailed report, funded by the Joseph Rowntree Foundation (JRF), on what these data tell us about living standards, poverty and inequality in the UK. In this observation, we briefly highlight some of the key findings from DWP’s report.

Average incomes

Average incomes fell for the second successive year in 2011–12. Official statistics recorded a reduction of 3% at the median (middle) and 2% at the mean, after accounting for inflation (measuring incomes before deducting housing costs, BHC). This comes on top of large falls in 2010–11, leaving median and mean income 6% and 7% below their 2009–10 peaks respectively.

These large post-recession falls in average incomes follow a period of slow growth that began in the early 2000s, far pre-dating the recession. The net result is that the official measure shows both mean and median income in 2011–12 no higher than in 2001–02, after adjusting for inflation.

However, it is important to bear in mind that these statistics make real-terms comparisons of incomes over time using an inflation measure based on the retail price index (RPI). The Office for National Statistics (ONS) now acknowledges that RPI inflation overstates the true rate of inflation facing households, and that this problem has got worse since 2010 (see here). The DWP has (sensibly) indicated that it is therefore reviewing the measure of inflation used in its income statistics. This is potentially important. If one were to adjust incomes using the new RPIJ index – which does not suffer from the same technical problems as the standard RPI – then recent trends in real incomes would still look bad, but slightly less so.

According to the RPIJ, real median income fell by 4% between 2009–10 and 2011–12 (compared with 6% according to the RPI) and real mean income by 6% (compared with 7% according to the RPI). On this basis, real median income in 2011–12 had fallen back to the level last seen in 2004–05 and real mean income to the level last seen in 2005–06 (rather than 2001–02 in both cases using the RPI).

Income inequality

The falls in incomes in 2011–12 were of similar (proportionate) magnitude across the income distribution. Income inequality was therefore essentially the same as in the previous year.

But inequality was substantially lower in 2011–12 than it was before the recession. Between 2007–08 and 2010–11, the incomes of lower-income households had held up better than the incomes of higher-income households. This is largely because, over that period, real earnings fell whereas benefit entitlements grew roughly in line with prices.

In work published recently, projections by IFS researchers suggested that the reduction in inequality between 2007–08 and 2011–12 will be temporary, and will have been almost unwound by 2015–16. This is because, if the Office for Budget Responsibility’s (OBR’s) forecasts are correct, most of the falls in real earnings had already occurred by 2011–12, whereas a number of cuts to the working-age welfare budget are being implemented over the current parliament.

Income poverty

In 2011–12, the numbers in absolute poverty (defined as having a household income below 60% of the 2010–11 median, with incomes measured BHC) rose from 9.8 million to 10.8 million – its highest level since 2002–03 – but relative poverty (defined as having a household income below 60% of the contemporary median, with incomes measured BHC) was broadly flat. This is because incomes fell by similar proportionate amounts for all income groups in 2011–12, so low-income households became worse off in absolute terms, but maintained roughly the same position relative to middle-income households.

However, relative poverty was substantially lower in 2011–12 than it was before the recession in 2007–08, having fallen from 11.0 million to 9.8 million to reach its lowest rate (16%) since 1986 (again measured BHC). This reflects the more general reduction in income inequality between 2007–08 and 2010–11 highlighted above. The reductions in relative poverty have been particularly large among pensioners and children. Relative child poverty in 2011–12 is at its lowest level since the mid-1980s, having fallen by about one-quarter since 2007–08; the rate of relative pensioner poverty has fallen by more than one-quarter over the same period, from 23% to 16%.

]]> Thu, 13 Jun 2013 00:00:00 +0000
<![CDATA[Labour's record on poverty and inequality]]> Today, the Oxford Review of Economic Policy publishes a special issue on Labour's economic record when in government between 1997 and 2010. As part of this, IFS researchers assess Labour’s record on income inequality and poverty. Here, we show how income inequality changed little but child and pensioner poverty fell significantly. We suggest, though, that these falls in poverty might prove fragile given that they were mostly based on very large increases in spending on benefits and tax credits. We also reflect on the main lessons for today’s policymakers. One such lesson is that how you spend money is more important than how much you spend. Governments need effective means to establish what works and what doesn't, and patience to see whether policies bear fruit in the long-run.

Labour had very clear objectives to reduce poverty amongst families with children and pensioners, and accorded these objectives high priority. Tony Blair made a famous commitment to end child poverty within a generation, and Gordon Brown promised to ‘to end pensioner poverty in our country”. However, it is much less clear that Labour took a strong view on the appropriate level of inequality within the top half of the income distribution, as indicated for example by Peter Mandelson’s famous statement that he was “intensely relaxed about people getting filthy rich as long as they pay their taxes.'

What happened to poverty and inequality under Labour?

If our summary of Labour's distributional objectives is accurate, then outcomes reflected those objectives quite closely. Turning first to poverty, both absolute and relative measures of income poverty fell markedly among children and pensioners - although the scale of the changes did not always match the considerable ambition, as set out explicitly in the case of the government’s child poverty targets.

By contrast, the incomes of poorer working-age adults without dependent children - the major demographic group not emphasised by Labour as a priority - changed very little over the period. As a result they fell behind the rest of the population and relative poverty levels rose. Since childless working-age people started the period with low levels of poverty compared with other demographic groups, one consequence of these trends was that the risks of poverty across the major demographic groups converged under Labour. This is illustrated in the Figure below.

Figure: Relative poverty rates since 1996-97

Figure: Relative poverty rates since 1996-97

Notes: Years refer to financial years. Poverty line is 60% of median income. Incomes measured before deducting housing costs.
Source: Family Resources Survey.

With falls in income poverty, one might expect to have seen a fall in income inequality. Indeed inequality did fall across much of the distribution. Those on relatively low incomes did a little better than those with incomes just above the average. However, those right at the top saw their incomes increase very substantially with the result that, on most measures, overall inequality nudged up slightly.

What drove these changes?

The substantial falls in pensioner and child poverty were largely driven by very significant additional spending on benefits and tax credits. Reforms since 1997-98 resulted in an £18 billion annual increase in spending on benefits for families with children and an £11 billion annual increase on benefits for pensioners by 2010-11 (see here). Our modelling suggests that child and pensioner poverty would either have stayed the same or risen, rather than fall substantially, had there not been these big spending increases. Meanwhile, Labour’s tax and benefit changes had relatively little net impact on the top half of the income distribution, or on low-income adults without dependent children – the group whose poverty rate did not fall. However, there is evidence to suggest that these reforms prevented a larger rise in inequality than actually occurred under Labour.

There were also increases in employment which had a small but detectable effect on income poverty. For example, reductions in the proportion of children living in workless families acted to reduce relative child poverty by about 2 percentage points.

There were many other Labour initiatives that could be considered anti-poverty policies. These include the introduction of the National Minimum Wage, Sure Start, increased financial support for childcare, significant increases in education spending and an expansion of the number of young people going on to higher education. Any payoffs from most of these measures will be long run, rather than immediate. It is of course very difficult to predict precisely what effects they will ultimately have on overall levels of poverty and inequality - and we will never know for sure, as their effects will inevitably happen alongside many other factors which continue to affect the income distribution. The verdict on the effects of Labour’s period in power on poverty and inequality is necessarily incomplete.

Reflections for the future

The fiscal climate is now very different and welfare spending can be cut back just as quickly as it is increased. As most of the currently detectable effects of Labour policy on poverty rates came through benefit increases, Labour’s legacy may prove fragile. Its durability will thus depend crucially on whether wider policies on things like education and childcare do have long-run impacts.

We would draw a few key lessons.

First, a sensible strategy for achieving distributional objectives will probably involve a lot of patience. Tax and benefit policy has quick impacts, but most other plausible policy levers do not. Many policies that might achieve sustainable impacts are likely to bear fruit only over the medium to long term, and certainly beyond the timeframe of a normal electoral cycle.

Second, the uncertainty about the long run effects of Labour’s policies on the income distribution highlights a wider point, which governments of all complexions should take seriously: major changes to things like childcare and education policy should, wherever possible, be implemented in ways which allow their effects to be robustly evaluated. We know much less than we should about policy effectiveness because so much major policy is implemented without proper evaluation (see here).

Finally, there are large gains to be had from designing the tax and benefit system skilfully: a government’s chosen level of redistribution can be achieved in more and less efficient ways. Stronger work incentives should be focused on those who are more responsive to such incentives (such as lone parents and parents of school-age children, as argued in the IFS’ Mirrlees Review). Labour’s reforms made some moves in this direction (see here). Another important goal should be simplicity and transparency. The current government’s Universal Credit offers a major opportunity to further this goal by replacing a jumble of overlapping means tests with a single integrated one, although reforms such as that to Council Tax Benefit risk undermining some of its potential advantages.

]]> Thu, 06 Jun 2013 00:00:00 +0000
<![CDATA[Cutting the deficit: three years down, five to go?]]> Following the financial crisis in 2008, the UK – like many other advanced economies around the world – has been undertaking a significant fiscal consolidation. The current government plan is for eight successive years of tax increases and spending cuts intended to offset the permanent rise in public borrowing that became apparent after 2008. The UK’s fiscal consolidation started at the beginning of this parliament (in 2010–11) and is expected to continue until 2017–18, which will be well into the next parliament.

According to recent analysis from the International Monetary Fund (IMF), in 2010, the UK is estimated to have had the fourth highest level of structural government borrowing among the 29 advanced economies for which comparable data are available. However, at least according to plans published so far, the UK is intending the fourth largest fiscal consolidation among this same group of countries and so by 2017 the IMF forecasts that the UK will have a lower level of structural borrowing than many other advanced economies – being below the average in the Euro area and below the average among the G7 and G20 countries.

The UK’s fiscal consolidation began in 2010–11, when the Labour government’s fiscal stimulus package was withdrawn and the incoming coalition government implemented some immediate spending cuts and some in-year increases in indirect taxes (including increasing the main rate of VAT from 17.5% to 20% from January 2011). Additional austerity measures are due to be implemented each year through to 2017–18. Taken together, the measures announced since the March 2008 Budget are expected to reduce public borrowing by 9.1% of national income (or £143 billion in today’s terms) by 2017–18, as shown in the Figure. This is larger than the 6.5% of national income (£101 billion) fiscal tightening originally planned by the coalition government in 2010. Since then permanent downgrades to the outlook for the UK economy and public finances have led the government to pencil in further austerity measures, but these are not scheduled to be implemented until after the next general election. (Box 5.1 in the 2013 IFS Green Budget provides more detail on how these figures are calculated.)

The current financial year is forecast to see a somewhat smaller fiscal tightening from discretionary policy changes (1.2% of national income) than seen in recent years (1.5% in 2012–13 and 1.8% in 2011–12). While discretionary spending cuts and tax increases were implemented in 2012–13, the level of cash borrowing last year was almost exactly the same as in the previous year. This is due to underlying upward pressure on the cash level of public spending, and weak economic performance leading to tax revenues growing only marginally; thus cash borrowing fell very little between 2011–12 and 2012–13. A similar story is forecast for this year: a combination of a weak outlook for economic performance, underlying upwards pressures on spending and further austerity measures being implemented means that cash borrowing is forecast to be largely unchanged once again.

By the end of this year, the government plans to have implemented just over half (52%) of the total planned consolidation. However, within this, the tax rises and investment spending cuts have been relatively front-loaded, while the benefit cuts and the cuts to day-to-day spending on public services have been relatively back-loaded. By the end of this financial year, virtually all of the planned tax increases (95%) and cuts to investment spending (90%) are planned to have been implemented, while only 58% of the total cuts to benefit spending and under one-third (31%) of the overall cuts to other current (i.e. non-investment) spending will have been achieved. The allocation of some of the remaining cuts to non-investment spending will be announced in the Spending Round on 26 June, which will allocate departmental budgets for the 2015–16 financial year.

It is worth noting, however, that the composition of the additional austerity measures planned for 2015–16 and beyond could change. First, it is possible that the Spending Round this summer could announce further cuts to, for example, spending on benefits or tax credits from 2015–16, which would ease the required squeeze on other current spending: for example, if the government wanted to continue cutting departmental spending in 2015–16 at only the same rate as is now planned on average over the four years from 2011–12 to 2014–15, they would need to announce a further £1 billion of cuts to other areas of non-investment spending. Second, history suggests that net tax rises are often announced in the 12 months following a general election (see Figure 5.13 of the 2013 IFS Green Budget); any new tax increases could be used to reduce the spending cuts required after 2015, something which the Budget document (paragraph 1.59) makes clear is a possibility

Figure: Timing and composition of the fiscal consolidation

Timing and composition of the fiscal consolidation

Note: This figure updates the numbers presented in Figure 5.8 of the 2013 Green Budget to include the policy announcements made in Budget 2013 and some minor changes to our methodology. The Green Budget contains details of the methodology and sources used to construct this figure.

The data underlying this figure can be found here

]]> Wed, 08 May 2013 00:00:00 +0000
<![CDATA[Deficit unchanged]]> The March Budget forecast by the Office for Budget Responsibility (OBR) was for the headline deficit, excluding the impact of transfers related to the Royal Mail Pension Scheme and the Asset Purchase Facility, to be £120.9 billion in 2012–13, compared to £121.0 billion in 2011–12. This £0.1 billion projected fall in the deficit meant that the Chancellor’s Autumn Statement assertion that “[the deficit] is falling and it will continue to fall each and every year” remained on course to be met. This was achieved by Government departments’ pledging to underspend their budgets by a greater degree than normal, with some payments deliberately pushed from 2012–13 into 2013–14.

The evolution of borrowing through the last four financial years is shown in the Figure below. Between its peak in 2009–10 and 2011–12 overall borrowing is estimated to have fallen by almost a quarter in cash terms (from £158.9 billion to £120.9 billion). By contrast, for most of 2012–13 borrowing is estimated to have been running above the level seen last year. It was not until the first estimate of borrowing in February 2013 was released – the day after the Budget – that cumulative borrowing over the year to date was estimated to be lower than over the same period in 2011–12. Borrowing over the first eleven months of 2012–13 is currently estimated to have been 3% below the amount borrowed over the first eleven months of 2011–12.

Figure: Reducing the deficit?

Public sector net borrowing

Note: Public sector net borrowing excluding the transfer of assets of the Royal Mail Pension Scheme (£28 billion in April 2012) and the transfer of Asset Purchase Facility assets from the Bank of England to HM Treasury (£3.8 billion in January 2013 and £2.7 billion in February 2013).

On Tuesday, the Office for National Statistics is due to release its first estimate of public sector net borrowing in March 2013 and, therefore, for the whole of 2012–13. Whatever these data indicate – whether they suggest borrowing was slightly higher or slightly lower in 2012–13 than in the previous year – will not be the last word on the matter as revisions to past data are inevitable. We will not know for some time whether the deficit in 2012–13 was actually higher or lower than the deficit in 2011–12.

Whatever Tuesday’s figures suggest, it is important to bear in mind that there are no direct economic consequences either way. In economic terms it is irrelevant whether the deficit is slightly higher or slightly lower in 2012–13 than in 2011–12. Either way the bigger picture is the same: the Government has implemented a combination of tax rises, welfare spending cuts and cuts to spending on public services and brought about a reduction in the deficit between 2009–10 and 2011–12. However, while 2012–13 also saw further austerity measures being implemented, weak economic performance has meant that the deficit was largely unchanged from its 2011–12 level. The same is forecast to be true in the current financial year: the OBR's forecast is that borrowing will fall by just £0.9 billion to £120 billion in 2013–14. This would leave the deficit largely unchanged for three years.

]]> Fri, 19 Apr 2013 00:00:00 +0000
<![CDATA[Women working in their sixties: why have employment rates been rising?]]> Today the Office for National Statistics released its monthly report on employment in the UK. Employment was essentially unchanged at 29.7 million in January 2013, higher than the pre-recession peak of 29.6 million in April 2008. Previous IFS research has attributed this strong employment performance in part to increased labour supply as a response to wealth losses during the financial crisis and changes to pensions and benefits implemented since the last recession, which encourage individuals to be in work.

The remarkably rosy picture for total employment belies contrasting trends between different groups. The employment rate amongst 25–29 year olds was 77.5% in the last quarter of 2012, the most recent period for which data on employment rates disaggregated by age are available. This is well below the peak of 80.4% seen in 2008. On the other hand, employment rates for older people have continued to increase over this same period. Indeed, the employment rate of 60–64 year old women has grown at the same rate since early 2010 as it did during the decade up to 2010. The proportion of 60–64 year old women in work rose from 25.6% in 2000 to 34.1% in 2010 while from 2010 to late 2012 it rose by a further 2.2 percentage points (to 36.3%) in just under three years.

What explains such contrasting experience between the young and the old? There are long term trends which have seen employment rates for older women rising for many years. But since 2010 increases in the state pension age for women have played a central role. Our estimates suggest that this has been responsible for the greater part of increased employment among 60–64 year old women since 2010.

Since April 2010, the state pension age for women has increased from 60 to 61½, and it is legislated to reach parity with that for men (at age 65) in 2018. In recently published IFS research we estimate that the increase in the state pension age for women from 60 to 61 that occurred between April 2010 and April 2012 increased employment of 60 year old women by 7.3 percentage points.

Using this estimate, we can calculate how employment rates would have evolved if the female state pension age were still 60. These new results, presented in the Figure below, show that – in the absence of the increase in the state pension age – employment rates of women aged 60 to 64 would have risen by only an estimated 0.3 percentage points between 2010 and 2012 (to 34.4%), instead of the large growth to 36.3% that was actually seen. In other words, the increase in the female state pension age explains an estimated 85% of the growth in the employment rate of this group that has occurred since early 2010.

Figure: Employment rates for 60–64 year old women since the crisis began with and without state pension age increase

 Employment rates for 60–64 year old women since the crisis began with and without state pension age increase

Source: Authors’ calculations using Labour Force Survey.

Our research also finds that the increase in the female state pension age has increased employment rates of affected women’s husbands – albeit to a lesser extent than the women’s own employment rates – explaining some of the resilience in employment rates of older men that has also been observed since 2010.

Is it possible that the increase in employment rates for older workers is a cause of the lower employment rates for younger people? In the short term, especially if demand for workers is weak, it is possible that firms might substitute towards keeping on older workers rather than hiring younger workers. However, having more older individuals in work will also boost overall demand in the UK economy and therefore increase demand for workers of all types. Previous IFS research finds no evidence that encouraging early retirement for older workers led to an increase in the employment rates of younger people. In the longer run, therefore, it seems particularly unlikely that higher rates of employment for older people would lead to lower employment rates of younger people.

In summary the vast majority of the increases in employment rates among older women seen since 2010 can be explained by the increase in the state pension age for women, which began to rise from 60 in April 2010. This has already led to the majority of 60 year old women being in paid work for the first time ever. As the state pension rises further throughout this decade, this trend of increasingly long working lives for women is likely to continue.

This observation is part of a project funded by the Nuffield Foundation.

]]> Wed, 17 Apr 2013 00:00:00 +0000
<![CDATA[The rapidly changing state]]> On Wednesday the government announced that the 2013 Spending Review will be published on 26 June. The Review is expected to cover spending decisions for 2015–16, though in the Autumn Statement the Chancellor pencilled in cuts stretching through to 2017–18. By that time, departmental spending will have been cut by nearly 19% in real terms since 2010–11, if current plans are followed. This is an unprecedented period both in terms of the scale of the consolidation and in terms of the extended period of year-on-year spending cuts. In this Observation, we look at a different aspect – what is happening to the composition of public spending. On current forecasts, public spending will take the same proportion of national income in 2017–18 as it did in 2003–04. But it will be being spent on quite different things.

Given forecasts from the Office for Budget Responsibility (OBR) for the size of the economy, total public spending is planned to fall from its peak of 47.4% of national income in 2009–10 to 39.5% by 2017–18. This is a dramatic fall, but historically it is the figure for 2009–10 which is unusual. Spending as a share of national income increased by 6.7 percentage points in just two years between 2007–08 and 2009–10, largely as a result of the loss to national income associated with the financial crisis and recession.

In fact, after eight years of austerity, spending as a share of national income is forecast to be back close to its long-run average, and at almost exactly the same level it was in 2003–04, about halfway through the last government’s period in office. However, while spending as a proportion of national income is forecast to be essentially the same in 2017–18 as it was in 2003–04, the composition of that spending will be very different. This is illustrated in Figure 1, which shows the proportion of total public spending accounted for by a number of large components of spending. Figure 1 also shows the level of spending in real terms. It is important to note that while spending as a share of national income is the same in these years, since the economy is forecast to be larger in 2017–18 than it was in 2003–04, real spending (in 2012–13 prices) in 2017–18 is planned to be £125 billion more than in 2003–04 – an increase of 22%.

The biggest difference between the two years is in spending on debt interest payments, which is forecast by the OBR to increase from £28.4 billion in 2003–04 to £62.0 billion in 2017–18. This is an increase in real spending of 118%, and would leave debt interest payments accounting for 4% more of total spending in 2017–18 than they did in 2003–04.

In addition, expenditures on health and on pensioner benefits are forecast to account for greater proportions of total spending in 2017–18 than they did in 2003–04, with real growth in spending of 36% and 37% respectively. These trends are driven by the policies of both the last government and the current one, which are in fact remarkably similar in many respects in terms of the relative priority given to different areas of spending. Whilst the last government increased spending across the board, it raised spending on health faster than spending on other public services. Whilst this government is responding to the very big deficit it inherited by cutting public service spending rather dramatically, it is protecting health spending. In both periods, health spending continued to rise as a proportion of the total. Both governments have also been relatively generous to pensioners. The last government raised means-tested benefits for pensioners rather rapidly. This one has largely protected pensioner benefits from the cuts inflicted on the rest of the social security budget.

Spending on non-pensioner benefits is also forecast to increase in real terms, by 14% from £79.4 billion to £90.5 billion, but to account for a slightly smaller proportion of total spending in 2017–18 than in 2003–04. This is actually one area where the spending priorities of the current government appear to differ from those of the previous government: spending on working-age benefits rose rapidly under the last government as a result of discretionary policy choices, whilst this government is making relatively large real cuts.

Spending on debt interest, social security benefits and health accounted for just over half of total public spending in 2003–04, but all bar £14 billion of the £125 billion increase in public spending is forecast to be accounted for by these components. Real spending on all other areas, including education, defence, public order and safety, and all other non-health public services, is forecast to increase on average by a fairly meagre 5% between 2003–04 and 2017–18, and to take up an ever smaller proportion of total public spending.

The government is currently making big choices about the shape of the state as well as about its size. On current plans, we are moving ever more rapidly towards a state focused on welfare and particularly on health and on pensions. As the population ages, this focus on health and pensions will become still more evident. However, whether spending a diminishing fraction of national income on other public services is a sustainable choice is an open question.

Figure 1: Proportion of total public spending accounted for by various components

Proportion of total public spending accounted for by various components

Note: Total public spending and debt interest spending in 2017–18 are forecasts from the OBR. Pensioner benefits forecast is the Department for Work and Pensions (DWP) forecast of DWP benefit expenditure directed at pensioners. Other social security forecast is the OBR forecast for total social security and tax credits less DWP benefit expenditure directed at pensioners. Health spending forecast assumes that real health spending in 2017–18 is equal to the 2010–11 level.

IFS public finance observations are generously supported by the Economic and Social Research Council (ESRC).

]]> Mon, 18 Mar 2013 00:00:00 +0000
<![CDATA[Reforms to alcohol taxes would be more effective than minimum unit pricing]]> Last week the Prime Minister David Cameron told MPs that he was determined to “deal with” the “problem of deeply discounted alcohol in supermarkets and other stores” and said that the Government was considering the results of a consultation into a proposed minimum unit price for alcohol.

New analysis by IFS researchers, published as an IFS Briefing Note today, argues that a reformed system of alcohol excise taxes focused on strong drinks would be more effective at targeting those drinking above recommended levels than would a minimum unit price for alcohol.

A tax-based reform would also raise additional tax revenues for the government. In contrast, minimum pricing would generate significant windfall revenues for alcohol retailers and manufacturers and reduce revenue from alcohol taxes.

Together, these findings point to a tax-based approach as preferable to minimum unit pricing in dealing with problem drinking. Since tax reform would require EU-wide agreement, we argue that the government should work now to win the necessary support to enable such reforms to be implemented, even as the debate about the compatibility of minimum unit pricing with EU legislation continues to rage.

The research draws on data recording the off-trade alcohol (purchased in supermarkets and off-licences) bought by more than 21,000 British households during 2010. Two main policy reforms are compared:

1. A minimum price for alcohol of a 45p per unit;
2. A reform of alcohol excise taxes so that the tax rate depends explicitly on alcohol content. Rates vary across alcohol types and increase directly in line with alcohol strength.

The total amount of off-trade alcohol purchased following each reform falls by the same amount: around 2.4 billion units (where a ‘unit’ is 10ml of pure alcohol) from an initial level of 37 billion. However the effects of the policies differ in important ways. The research compares how these policies impact on households according to the average amount of alcohol they buy. This gives an important measure of whether or not they are well-targeted on heavy drinkers. The impact on government tax revenue and industry revenues is also assessed.

Figure 1 shows the average percentage change in alcohol units purchased across households grouped by initial purchase levels. For heavy drinkers (those purchasing more than 35 units per adult per week on average), a 45p per unit minimum price reduces alcohol purchases by around 8%. For moderate drinkers (fewer than 7 units), the fall is 4%. Reform to the system of excise taxes would reduce purchases by heavy drinkers by 9.5%. Moderate drinkers would see a fall of just 2.6%. This tax reform is better targeted on heavy drinkers than the minimum price.

Figure 1. Average change in units of alcohol purchased, by policy and purchase group

Average change in units of alcohol purchased, by policy and purchase group

Sources: Calculated from Kantar Worldpanel 2010 data.

Both policies raise total consumer spending on off-trade alcohol by a similar amount, around £550 million for the minimum price, and £640 million for the tax reform. The policies differ enormously in how that extra spending is distributed between government tax revenue and industry revenue. The minimum unit price reduces tax revenue by £290 million (drinkers reduce their purchases without any compensating increase in tax rates), but increases industry revenue by £840 million. In contrast, tax reform would increase tax revenues by £980 million.

Both reforms would probably be slightly regressive. Low-income drinkers tend to buy cheaper and stronger alcohol than high-income drinkers purchasing similar amounts of alcohol units. The government should be more concerned with the distributional effects of policy as a whole. Nonetheless, if policy makers were concerned about the distributional consequences of alcohol pricing policies in particular then some of the revenue raised through tax reform could be used to help ameliorate this concern. This is an option not available for minimum pricing.

The analysis looks only at off-trade alcohol which accounts for about three in every four alcohol units purchased. Minimum unit pricing and tax reform are likely to have different implications for the on-trade (alcohol in pubs and restaurants). On-trade prices are much higher, so a minimum unit price would have little direct impact there. Tax reforms would affect on-trade alcohol as well. To the extent that problem drinking takes place in the on-trade it is not clear that we would only want price reforms to affect off-trade consumption. If those who drink to excess on-trade purchase stronger products, then an excise tax reform would still look well-targeted. Furthermore, a tax reform targeted on alcohol strength may well still raise off-trade prices relative to on-trade since, on average, alcohol purchased off-trade is stronger than alcohol purchased on-trade.

It is somewhat curious that the recent debate around alcohol pricing has focused so much on a minimum unit price. Much less has been made of the fact that the current, complex set of alcohol taxes does a poor job of targeting harmful drinking. Reforming the way in which we tax alcohol could provide a better route to reducing harmful drinking than the introduction of a minimum price. It would reduce alcohol purchases among heavy drinkers, have less impact on moderate drinkers, and raise additional useful revenue for the government. Heavy drinkers also buy cheap alcohol, but a minimum price would be less well targeted than the tax reform, reduce tax revenues and increase the revenues of alcohol retailers and manufacturers.

]]> Mon, 18 Mar 2013 00:00:00 +0000
<![CDATA[Making tax policy]]> Taxes, like death, are unavoidable. But we can design our taxes. We are not bound to have a tax system as inefficient, complex, and unfair as our current one. To improve things, we need to see the system as a whole, we need to design the system with a clear understanding of the population and economy on which it operates, and we need to apply economic insights and evidence to the design. We also need a much more informed public debate and a much better set of political processes than the ones we currently have.’

So my co-authors and I said in the introduction to the Mirrlees Review published in 2011. The question is: what is it that prevents better decision-making and the creation of a better system? That is an issue discussed in an IFS Briefing Note published today.

It’s not the lack of importance. Four pounds in every ten generated in the economy is collected in tax. The way the tax system is designed inevitably has a huge effect on the operation of the economy. That is inescapable. By designing the system better, there are really substantial economic benefits to be reaped. As we showed in the Mirrlees Review, the costs of a poorly designed tax system run into many billions of pounds in reduced output and welfare. In that review, we also set out one possible strategy for unlocking some of these benefits, a strategy based on treating the tax system as just that, a system, and recognising that there are often substantial costs associated with treating similar activities differently.

And it’s not that there haven’t been some attempts to improve the process of tax policymaking. The current government has attempted to produce a more coherent structure for developing tax proposals, from policy formulation through to legislation and implementation. This new approach to tax policymaking is backed up by the Tax Professionals Forum. There is also the innovation of the Office of Tax Simplification.

Nor is it a want of consultation. The amount of consultation in the current policymaking process is almost overwhelming.

But there do seem to be three very substantial barriers to better policy.

First, tax policy is made annually (or increasingly biannually) in Budgets (and Autumn Statements). No government or opposition lays down a strategy for the tax system as a whole. There is every indication that policies are made in isolation, and with little sense of how they fit together over the long run. While there is much consultation, most of the effort is put into tinkering with the existing system to try to make the bits that don’t work well work a bit better without necessarily thinking more fundamentally about why the bits in question are not working well, or continue to give rise to ‘avoidance’ issues. There is little attention paid to any form of long-term strategic framework for the development of the tax system as a whole.

Second, there is a remarkable lack of challenge within the executive and effective scrutiny and challenge from the legislature. The Treasury challenges spending departments. Who is to challenge the Treasury when it makes tax policy? The enhanced role of HM Treasury, in the wake of the O’Donnell Review, may at times have limited the challenge that comes from HM Revenue and Customs. While the Treasury Select Committee and the Public Accounts Committee have increased their oversight of tax policy, they remain poorly resourced for focusing on such a large and complex issue.

Third, the general quality of political and public debate is limited, allowing poor policy to be passed off as good all too easily, resulting in an unhelpful focus on specific parts of the tax system, and often making change politically difficult. Very specific elements of the system – notably rates of income tax – achieve totemic significance far beyond their true importance, while changes to other parts of the system are often given wholly inadequate attention.

Tax policy is inevitably a difficult and highly political area. But there are things that can be done to improve the process for making policy.

There should be a tax strategy and objectives set out by each Chancellor against which policy proposals can be measured. This would increase transparency and accountability and, if followed, would also increase stability and certainty.

More challenge and scrutiny should be introduced into the policymaking process both within the executive and by the legislature. If necessary, parliament should be given greater resources, the House of Lords should play a greater role, and consideration should be given to creating a specific select committee devoted to tax policy.

We need an explicit process for reviewing the effectiveness of tax policy post implementation.

There is a case for at least a public review of the working of the relationship between HM Treasury and HMRC.

The Autumn Statement should be more explicitly focused on consultation on fiscal policy and the presentation of the latest economic and fiscal forecasts. It should not be an additional opportunity for Chancellors to pull rabbits from hats.

There is an urgent need for an improvement in the public debate about tax policy.

And of course we need politicians willing to be honest in their communication, open and long term in their thinking, analytical in their approach and courageous in their decisions. National welfare really is reduced by billions of pounds because we have a tax system that is not well designed. Finding ways to improve the policy development process should be a matter of priority for all political parties.

We need to see taxes as part of a system rather than individually, we need a strategy against which to judge proposals, and we need more transparency and more challenge and accountability in the tax policymaking process.

Note: Paul Johnson will be presenting this material at a joint IFS and Institute for Government event Better Budgets: Making Tax Policy Work on Wednesday 20 February 2013.

]]> Wed, 20 Feb 2013 00:00:00 +0000
<![CDATA[Better options exist to help low earners than 10p tax rate]]> In a speech in Bedford today, the Leader of the Opposition, Ed Miliband, has proposed reintroducing a 10% income tax rate on a narrow band of income, to be paid for by a new ‘mansion tax’ on residential properties worth more than £2 million. This observation discusses the merits of both of these policies and suggests more economically sensible policies that would have broadly the same effects.

‘Mansion tax’

Mr Miliband did not give precise details of the design of his ‘mansion tax’ in his speech, but a similar proposal from the Liberal Democrats in their 2010 General Election Manifesto would have involved a tax based on 1% of a property’s value above £2 million. Thus, a property worth £3 million would face a charge of £10,000 a year. No firm costings are available for such a tax, but in 2010 the Liberal Democrats estimated that 70,000 properties would be affected and the total yield from this tax would be £1.7 billion a year, implying that the average charge would be over £24,000 per year. Of course, a higher or lower charge could raise more or less revenue.

The ‘mansion tax’ has a sensible logic underpinning it: if residential property is to be taxed, it makes sense to levy such a tax in proportion to property value and base it on current valuations. By contrast, Council Tax, the existing tax on residential property in England and Scotland, has neither of these features as it is based on 1991 property values and is set far from proportional to those values, with higher-value properties significantly under-taxed: for example, in a local authority setting the current English average band D rate, someone with a property at the midpoint of band D will pay 1.85% of its 1991 valuation in Council Tax whereas someone with a property at the midpoint of band G will pay 1.00% of its 1991 valuation. Therefore, rather than adding a mansion tax on top of an unreformed and deficient council tax, it would be better to reform council tax itself to make it proportional to current property values. As with the mansion tax, it would be the most valuable properties that saw the largest increases in tax liability from such a reform.

10p tax rate

A 10p tax rate previously existed between 1999 and 2007 before it was (in)famously abolished for non-savings income by Gordon Brown in his last Budget as Chancellor, and still exists for savings income that falls into the first £2,710 of taxable income. Reintroducing Labour’s 10p rate in full would reduce tax revenues by around £7 billion a year and benefit basic rate taxpayers by up to £271 a year, though given the numbers above, it is unlikely that anything close to this amount could be raised through a mansion tax alone. To introduce a £1,000 10p tax band, which would benefit all 23.5 million basic-rate taxpayers, one would need to raise £2 billion, or around £30,000 on average from each of the 70,000 owners of ‘mansions’. (Higher-rate taxpayers would not benefit from the Labour party’s proposals as they would reduce the point at which the higher 40p income tax rate starts to be applied. The proposal would therefore slightly increase the number of higher rate taxpayers, which is something the current Government has done to limit the gains to higher rate taxpayers from its increases in the income tax personal allowance.)

A 10p tax rate would reduce taxes for those on low incomes and strengthen their work incentives. A far simpler and more sensible way of achieving these aims would be to spend the same amount of money on increasing the personal allowance – a policy on which the current government has already spent £9 billion a year. This would have virtually the same impact on individuals’ tax payments (see figure below), be slightly more progressive, take some people out of income tax altogether and avoid the complexity involved in introducing a new income tax rate. An even better alternative, which would help those who already pay no income tax because their incomes are below the personal allowance but do pay employee National Insurance Contributions (from April, there will be 1 million such people earning between £7,748 and £9,440), would be to increase the point at which individuals start paying employee National Insurance Contributions. This would also bring the income tax and National Insurance systems more in line and would take some people out of direct tax altogether. And if one wanted to focus the gains from the policy on low-income working families rather than basic-rate taxpayers generally, increasing Working Tax Credits would be another sensible alternative to look at.

Figure: Effect of a £1 billion giveaway through either a 10p tax rate or an increase in the personal allowance on combined income tax and employee NICs schedule

 Effect of a £1 billion giveaway through either a 10p tax rate or an increase in the personal allowance on combined income tax and employee NICs schedule

Note: Reforms considered are £1,000 10p tax band and £500 increase in the personal allowance. In both cases, there is an offsetting reduction in the higher-rate threshold so that higher-rate taxpayers are unaffected.


Both the proposals made today by the Labour party will be familiar to tax policy watchers. One, the mansion tax, reflects a problem with the current tax system – council tax is regressive as high-value properties are under-taxed relative to low-value properties – but does not take the idea forward to its logical conclusion. The other, the proposal for a new 10p starting rate of income tax, has no plausible economic justification. It would complicate the income tax system and achieve nothing that could not be better achieved in other ways. It appears to repeat the same error perpetrated by Denis Healey in 1978 (undone by Geoffrey Howe in 1980), Norman Lamont in 1992 and Gordon Brown in 1999 (which he himself undid at considerable political cost in 2007). To have observed lower starting rates of tax being introduced and abolished by governments of both complexions over the last three decades and then to propose the same thing again suggests a remarkable failure to learn from history.

]]> Thu, 14 Feb 2013 00:00:00 +0000
<![CDATA[Welcome simplification of state pensions but younger generations lose]]> The Government’s proposed new state pension system will involve everyone accruing the same level of state pension each year whilst working (so long as they earn above the lower earnings limit), or claiming benefits for being unemployed, looking after children aged 12 or under, or caring for sick or disabled adults. Anyone with 35 years of qualifying activity of this kind will receive a fixed pension of £144 a week.

A lot is being claimed for these reforms. The reforms have some desirable features but – as is often the case with pension reforms – the overall effects will be more complex. The gains from the proposed reforms are as follows:

  • The proposed system looks straightforward and – in terms of its clarity –appears to be a clear improvement on the complex miscellany of rules which govern the current system.
  • There will be some winners from the proposals. In particular, those who expect to have significant periods of self-employment are likely to gain. Currently they accrue rights to the basic state pension (BSP) but not S2P. Therefore, they would be likely to receive more state pension under the proposed system than they would get under the current rules. However, the case for the self-employed paying a lower rate of NI than employees would then be weakened, and the Government may decide to increase their NI contributions in return for this increase in benefits.
  • A second potential group of winners, at least in terms of state pension rights, is members of contracted out defined benefit (DB) occupational schemes. The proposed reforms would end contracting out and thus increase the state pension rights that this group will accrue. Whether or not they ultimately receive a higher pension overall will depend on whether the DB schemes in question make use of the ability that the Government plans to grant them to adjust the terms of the scheme accrual to deal with the extra costs imposed by ending contracting out. The government has said it does not intend to reduce accrual rates within public sector schemes – although any increase in pension rights for public sector workers will be at least partially offset by their higher employee National Insurance (NI) contributions.
  • There will also be a larger group of winners in the short-run – specifically, those who reach State Pension Age after 2017 and who have 30 years of BSP accrual but who have not yet earned a state pension, including S2P, of £144 a week. This group includes some low earners and some of those who spent time out of the labour market caring for children or disabled adults prior to 2002. However, this group will also be made up of those who have spent some significant period contracted out of the second tier state pension – that is, those who chose to give up their rights to additional state pension in return for either paying a lower rate of NI or receiving a refund of NI contributions which they will have invested in a private pension scheme. This latter group are, overall, higher-than-average earners. Of course, those who made a conscious decision to remain contracted-in may feel aggrieved that they took the wrong bet on future government policy change; among this group, those who have already accrued £144 or more of state pension rights will stop accruing any more from 2017 and hence would retire with a lower state pension than they might have expected.

But there are also some groups that will be worse off financially as a result:

  • These proposals imply a cut in pension entitlements for most people in the long run. This is because, since 2002, coverage under the current pension system is almost as broad as under the proposed system (activities including employment, unemployment and looking after children aged under 12 are all credited as contributions), and the annual pension accrual from such activities is higher under the current pension system. For example as set out in the table below, in 2017–18, most low earners and non-workers will under the current system accrue £5.05 (£3.59 plus £1.46) of additional weekly state pension rights for ‘contributing’ for one extra year, provided they have not already accrued 30 years of contributions at that stage (those who already have 30 years, would accrue £1.46 of additional weekly pension). Higher earners would accrue £5.81 a week of state pension (or £2.22 if they already had 30 years of contributions). In the proposed new system, these same people would accrue £4.11 of additional weekly state pension (or nothing if they had already accrued 35 years of contributions). The key point is that £4.11 is less than £5.05 (what most lower earners would accrue under the current system) and the gap for high earners is, of course, even greater.
  • Therefore, in the long run, the reform will not increase pension accrual for part time workers and women who take time out to care for children. In fact, in common with almost everyone else, these groups would end up with a lower pension at the state pension age under the new system than they would do under the current system.
  • In common with all pension regimes, one needs to take account of how both rights and pensions in payment are indexed. At present S2P in payment is indexed to CPI, while BSP is triple-locked (at least for this parliament) and, presumably, the flat-rate pension would be indexed to at least earnings. Simply comparing the value of BSP plus S2P with the FRP at the state pension age will therefore overstate the extent to which rights have been reduced, since overall the state pension under the proposed system would be indexed more generously.
  • The term ‘long term’ applies to anyone who will have spent 30 years or more in creditable activities (including employment, looking after children and other caring) since 2002. This, therefore, certainly includes all those who entered the labour market from 2002 onwards (that is, those born in 1986 or later) and arguably also applies to those born from about 1970 onwards.
  • Those who expect to have fewer than 10 years of contributions or credits will be particular losers from the reforms, since they will receive no state pension under the proposed scheme – though the Department for Work and Pensions (plausibly) claims that the majority of people in that position are those who have spent only a short part of their working life in the UK.

We might also be wary of claims by ministers that this will be the final radical reform of state pensions. Similar claims were also made for the reforms announced in 1998, 2002 and 2006. Such frequency of reform makes it harder for individuals to plan for retirement appropriately. That said, it is perhaps more plausible than usual that today’s proposals will prove to be the last radical overhaul for a while, not least because it is moving in the same direction as the reforms implemented in the last parliament and also because a simpler system may prove to be more robust. We have spent more than 25 years gradually unpicking the State Earnings Related Pension Scheme (SERPS), which was first introduced in 1978 without sufficient consideration of the long-term consequences for the public finances.

Overall, the proposals expected in today’s White Paper look like they will bring about a welcome simplification. However, it is important to be clear that – while there will be a fairly complex pattern of winners and losers from the reform in the short-term – the main effect in the long run will be to reduce pensions for the vast majority of people, while increasing rights for some particular groups (most notably the self-employed).

Table: Summary of current and proposed state pension systems

 Summary of current and proposed state pension systems

* Caring for a sick or disabled person for more than 20 hours a week, or a registered Foster Carer, and claiming Carer's Credit.
** Primary legislation requires that the level of the BSP should be increased each year in line with average earnings growth. The current government has committed to a more generous indexation arrangement (the triple lock) for the duration of the current parliament.
*** Individuals can accrue additional State Second Pension entitlement in each year between age 16 and the year before reaching State Pension Age. Therefore, the maximum years of contributions varies depending on an individual’s SPA. For those with an SPA of 65, the maximum years of contributions is 49. For those born on or after 6th April 1978, the SPA is currently set to be 68 and therefore the maximum years of contributions will be 52.

]]> Mon, 14 Jan 2013 00:00:00 +0000
<![CDATA[The Effects of the Welfare Benefits Up-rating Bill]]> The Welfare Benefits Up-rating Bill, which gets its Second Reading in the House of Commons tomorrow, proposes to cap the annual increases in most working-age benefits at 1% in cash terms in 2014–15 and 2015–16, in addition to the 1% cap on increases already confirmed for 2013–14.

By default, working-age benefits and tax credits are up-rated each April in line with the Consumer Prices Index (CPI) in the preceding September. CPI inflation was 2.2% in September 2012 and is forecast to be 2.6% and 2.2% in the Septembers of 2013 and 2014 respectively. The government plans instead to uprate the majority of these benefits by 1% for the next three years (though only the proposed April 2014 and April 2015 cash increases are included in the bill). Given current forecasts of inflation, this implies a cumulative 4% real cut in the benefits affected. The actual real-terms cut, though, will depend on future levels of inflation. If inflation is much lower than forecast then the real cuts will be small. If it is higher than forecast then the real cuts will be bigger than is currently expected. Similarly the effect of the measures on benefit levels relative to earnings levels will depend upon future rates of earnings growth. There is an arbitrariness to this eventual outcome. That is the result of specifying future benefit changes in nominal terms rather than, for example, relative to actual price or earnings changes.

When announcing the policy, the Chancellor drew attention to the fact that out-of-work benefit rates have grown substantially faster than earnings since 2007, as shown in Figure 1. This essentially reflects the fact that earnings have not kept pace with inflation, whereas rates of out-of-work benefits (Income Support, Jobseeker’s Allowance and Employment and Support Allowance) have risen broadly in line with prices (though there have been discretionary cuts to the rates of other benefits and tax credits).

Figure 1. Earnings and out-of-work benefits (Jan 2007 = 100)

 Earnings and out-of-work benefits

Sources: Past earnings from ONS series DTWM, ROYK, MGRZ, MGRQ, past benefits from DWP (JSA personal allowance for a single person over 25), forecasts of inflation and earnings growth from OBR

If earnings fall relative to benefit levels, then being in work becomes less financially attractive, all else equal. At present, out-of-work benefits for a home-owning couple without children stand at 27.7% of what the family’s net income would be if one of the couple earned £560 per week. If instead earnings had grown as fast as benefit rates since 2007, that fraction would have been 25.3%. However, higher wages would not have significantly affected the incentive to work for all groups. For a lone parent with two children and rent of £100 a week working 16 hours a week at the minimum wage, out-of-work income is 80% of in-work income in both scenarios, because almost all of the additional earnings are lost in withdrawn benefits.

There is, of course, nothing special about relativities in 2007. Historically, earnings have tended to rise in real terms and hence outstrip price-indexed benefit rates (although, particularly for families with children, above-indexation increases have seen some entitlements grow considerably faster than prices). Relative to that “normal” scenario, the recent fall in earnings compared to out-of-work benefits looks even more striking. On the other hand this is a reminder that when real earnings growth returns, current policy implies that benefit rates will fall relative to earnings.

Figure 2 shows the distributional effect of the benefit reductions contained in the Welfare Benefits Up-rating Bill in the context of all the Autumn Statement tax and benefit announcements and the consolidation package as a whole. This shows that the majority of the impact of the Autumn Statement announcements will come through the changes to up-rating policy set out in the Bill, but that this is only a small fraction of the net ‘takeaway’ from households resulting from the overall fiscal consolidation.

As far as the effects of benefit up-rating measures are concerned, reductions in entitlement are unsurprisingly concentrated in the bottom half of the income distribution. The lowest-income decile group see the largest fall in entitlements as a percentage of income (1.5%) as a result of measures in the Bill, and the second decile see the largest decrease in cash terms, losing about £150 per year on average.

Because the proposed uprating changes apply to almost all benefits and tax credits, both in-work and out-of-work households are affected. Of 2.8 million workless households of working age, 2.5 million will see their entitlements reduced, by an average of about £215 per year in 2015 –16. Of 14.1 million working-age households with someone in work, 7.0 million will see their entitlements reduced, by an average of about £165 per year. Note that this figure includes 3.0 million families who lose only from the cuts to Child Benefit, at an average of about £75 per year (monetary amounts are in current prices).

As the chart shows, other elements of the consolidation package have been much more significant, especially for those with the very highest incomes. It is important to see the effects of this Bill alongside the other tax and benefit changes, as well as falling real earnings which have hit those in the middle and upper parts of the overall distribution the hardest.

Figure 2: Distributional impact of tax and benefit reforms

January 2010 - April 2015 inclusive, as if Universal Credit fully in place

 Distributional impact of tax and benefit reforms

Note: Income decile groups are derived by dividing all households into 10 equal-sized groups according to income adjusted for household size using the McClements equivalence scale. Source: Authors’ calculations using 2010–11 Family Resources Survey and TAXBEN, the IFS’ tax and benefit micro-simulation model.

When cutting public spending dramatically to help reduce an unsustainable budget deficit it is almost inevitable that spending on benefits and tax credits – which account for 30% of the government’s total budget – will be targeted. This policy achieves savings through an across-the-board reduction in the real value of benefits for people of working age at a time when real earnings have been falling.

But we don’t know two things. First, the actual effects of the bill on real benefit rates are unknown, because they depend on future price levels. This exposes the poorest in society to inflation risk. Second, we don’t know the government’s view on how benefit rates should be indexed in the longer run. We ought to.

]]> Mon, 07 Jan 2013 00:00:00 +0000
<![CDATA[Withdrawal symptoms: the new 'High Income Child Benefit charge']]> On Monday, Child Benefit will effectively become an income-related benefit for the first time, reducing public spending by an estimated £1.5 billion in 2013–14. We estimate that about 820,000 families, in which at least one adult has a taxable income exceeding £60,000 per year, stand to lose all their Child Benefit via a new income tax charge unless they change their behaviour in response. And about 320,000 families in which the highest-income adult is on between £50,000 and £60,000 would have some, but not all, of it clawed back. The affected families stand to lose an average of about £1,300 per year. The remaining 85% of families currently receiving Child Benefit will be unaffected for now, although more will be affected in time because the £50,000 threshold is planned to be frozen in cash terms.

In the context of a large fiscal consolidation, one can understand why the government is looking at universal benefits going to those on higher incomes as one candidate for cutbacks. But various features of the design of this particular policy look problematic.

First, much attention has understandably focused on inequities of treatment between 1-earner and 2-earner families. The means test uses information only on the taxable income of the highest-income family member so, for example, a 2-earner couple with taxable income of £100,000 split equally between them would retain all Child Benefit, but a 1-earner couple or lone parent with taxable income of £60,000 would lose all of it. This kind of situation does not arise with existing means-tested welfare payments, which are based upon family income. (But note that, because a family’s Child Benefit entitlement will depend upon the income of the highest-income individual, this is not simply an argument about family versus individual assessment: this policy is an unusual hybrid of the two.)

A second unusual feature of the reform relates to its impact on incentives for individuals to reduce their taxable income by, for example, working less or contributing more to a private pension. All families’ Child Benefit entitlements will be exhausted once the taxable income of the highest-income adult reaches a fixed level – £60,000 – regardless of how much Child Benefit has to be withdrawn. To achieve this, the rate at which it is withdrawn as income rises above £50,000 has to vary with the number of children in the family, i.e. with the amount of Child Benefit the family receives. Specifically, affected taxpayers will pay back one per cent of their family’s Child Benefit for every £100 by which taxable income exceeds £50,000. One per cent of Child Benefit is £10.56 per year for a 1-child family, and an additional £6.97 per child for larger families. Hence the marginal tax rate between £50,000 and £60,000 is increased by about 11 percentage points for the first child and by an additional 7 percentage points for each subsequent one. So, for example, while about 320,000 people will find that their marginal income tax rate increases to more than 50%, about 40,000 of them - those with three or more children - will find that it jumps to at least 65%. This is illustrated by the Figure, which shows marginal income tax rates for taxpayers with different numbers of children.

Figure: Marginal income tax rates in 2013-14 for the highest-income member of families with children

Source: Author’s calculations.

This unusual feature of the policy has another consequence in the long run. Since higher Child Benefit entitlements imply higher withdrawal rates, marginal tax rates between £50,000 and £60,000 will also have to rise over time as cash-terms Child Benefit rates rise in line with prices (or indeed in line with anything else). This may sound obscure and technical, but in time it would be important. For example, indexation of Child Benefit rates at 2% per year would increase these marginal rates by a further 5 percentage points for a 3-child family over one decade. To avoid this, some families with an adult on more than £60,000 would have to be allowed to retain some of their Child Benefit. But there are currently no provisions to do this: the £60,000 withdrawal end-point is not a parameter that is planned to be uprated over time.

That is an important detail, but the government could and should address it without any structural upheaval. Other problems are more fundamental to the design of the policy. It creates a series of administrative complexities, including the need for up to 500,000 more individuals to fill in income tax self-assessment forms according to HMRC. But perhaps the biggest concern is the incoherence it creates in the welfare system. We already have the Child Tax Credit, and soon its imminent replacement, namely the child additions within Universal Credit. The reform to Child Benefit will mean that we have two systems of income-related support for children. But the relationship to income will be completely different in each case: based on family income in one case and the income of the highest-income family member in the other; withdrawn at different rates as income rises; and with the withdrawal starting at very different income levels.

With the introduction of Universal Credit, the government will integrate six of the seven existing means-tested benefits and tax credits into one. IFS researchers have argued that the basic principles behind this move have much to commend them. But almost simultaneously, the government is introducing a new and separate means test for Child Benefit which will work in a completely different way, as well as making the seventh existing means-tested benefit – Council Tax Benefit – much more complicated by asking every Local Authority in England to design its own scheme. It is unclear whether the net effect of all this will be to improve the welfare system.

]]> Fri, 04 Jan 2013 00:00:00 +0000
<![CDATA[A defining issue? The government’s pledge to raise the share of revenue from green taxes]]> Amidst all the discussion of the Chancellor’s fiscal rules following the Autumn Statement, rather less attention has been paid to another tax-related target. The Coalition Agreement set out a commitment to “increase the proportion of tax revenue accounted for by environmental taxes.” The Environmental Audit Committee reported last month that whether or not this pledge is met depends on how ‘environmental taxes’ are defined. This observation looks at the status of this pledge after the Autumn Statement, and asks whether such a target is sensible in the first place.

The Treasury finally clarified its interpretation of the pledge in July:

  • environmental taxes should make up at least as big a part of total revenue in 2015/16 as in 2010/11;
  • only taxes whose primary objective is to encourage pro-environmental behaviour change (rather than, say, revenue-raising) count as ‘environmental’.

International bodies such as the OECD and Eurostat, however, define environmental taxes not according to their intent, but on whether the tax encourages pro-environmental outcomes. On this basis, the ONS classifies taxes such as fuel duty and air passenger duty as environmental, which the Treasury does not.

The table summarises Treasury and ONS definitions of environmental taxes. We also suggest a third definition, which perhaps best reflects all those taxes which are environmental either in terms of intent or outcome, and for which we have revenue forecasts to 2015/16. This includes company car taxes which, like vehicle excise duty, depend on the fuel efficiency of the car. However, in line with international practice, it excludes VAT on fuel duty since VAT is a general consumption tax rather than a particular environmental tax. As shown at the foot of the table, we find that the definition of green taxes does indeed affect whether or not the target is met.

Three definitions of ‘environmental taxes’, and compliance with target

 Three definitions of ‘environmental taxes’, and compliance with target

Note: Figures for 2010/11 are out-turns and 2015/16 are forecasts. Revenues are taken from the OBR Economic and Fiscal Outlook, consistent with the December 2012 Autumn Statement. The exception is company car taxes, where forecasts are taken from a written ministerial statement to the House of Commons on 16 July by Economic Secretary to the Treasury, Chloe Smith.

On the Treasury definition, the government would easily meet its pledge: the green tax share is set to more than double, from 0.4% to 0.9% of revenues. Green taxes in 2015/16 could fall by £3.3 billion (56% of forecast environmental receipts that year) before the target is missed.

On the ONS definition, however, the green tax share will fall from 7.8% to 7.1% of revenues, breaching the target. Revenues from this set of taxes would have to rise by £5.3 billion (11%) in 2015/16 to meet the target.

On our definition, the pledge is also missed, with the green tax share falling from 7.3% to 7.0%. Green taxes would need to rise by £2.3 billion (5%) to hit the target.

Why do the results differ? Total green taxes are substantially lower on the Treasury definition, mainly because it excludes fuel duties which are estimated to raise £27.8 billion in 2015/16. Including fuel duties makes the pledge much harder to meet: their share of total revenues is set to fall by 0.8 percentage points by 2015/16.

Excluding fuel duties also allows the Chancellor to make further concessions on duties in future years without jeopardising the target. Indeed, if fuel duty is not considered an environmental tax, it actually makes the target easier to meet by reducing total ‘non-environmental’ revenues. The Autumn Statement announced that the planned duty rise for January 2013 was to be cancelled, and future planned inflation adjustments in duties were pushed back by five months from April to September each year. Together these reforms cost around £1.7 billion in 2015/16, the year in which compliance with the green tax target is to be judged. This is almost the entire amount by which, on our definition, the pledge would be breached.

It may well be reasonable to take different views on the precise definition of what constitutes a green tax. More fundamentally, we should ask whether the pledge to raise their importance in total revenues has any particular merit, setting aside the definitional issues. Ideally, taxes (including green taxes) should be raised in the most effective way, rather than to hit some essentially arbitrary target for receipts from one part of the system. Whether environmental taxes make up a bit more or a bit less of total revenues in 2015/16 than they did in 2010/11 is not of much real consequence. Moreover, the green tax share of revenues is not really a good indicator of a government’s environmental credentials. The tax system can be ‘greened’ without raising more money: reforms to vehicle excise duty to base payments on fuel efficiency are one example. Other environmental policies such as regulation and subsidies operate outside the tax system but could still have important environmental benefits. Some green tax revenues may erode away as people change their behaviour in response to the tax. The green tax share also depends on total revenues which can be very sensitive to overall macroeconomic conditions.

It took the government more than two years to lay out the rules against which it wished its green tax target to be judged. In the end, its chosen definition makes the pledge easy to accomplish but implies a very limited role for green taxes. If they really do make up less than 1% of total receipts, then any ambition to move towards a significant role for such taxes in the future (such as that adopted by the Liberal Democrats at their 2010 conference) appears to be a very long way away.

More fundamentally, pledging to meet certain targets which are then defined in such a way as to make them trivial both to meet and in their apparent importance does not look like an effective way of gaining credibility. A green tax definition more in line with international convention would be much more constraining and require the government to take policy action. As a result, there will almost certainly be an argument at the next election as to whether or not the pledge was met. A better argument would be that it wasn’t really worth making in the first place.

]]> Mon, 10 Dec 2012 00:00:00 +0000
<![CDATA[70th anniversary of the Beveridge report: where now for welfare?]]> ‘Benefit in return for contributions, rather than free allowances from the State, is what the people of Britain desire’

Sir William Beveridge, 1942

Published by parliament on 1 December 1942, the Beveridge report proposed a system of national insurance in which flat rate benefits, paid for by flat rate contributions, would provide support in the event of unemployment, ill health or old age.

70 years later these foundations remain visible in some elements of the welfare state. Entitlement to contributory Jobseeker’s Allowance, contributory Employment and Support Allowance and the state pension is still determined by National Insurance (NI) records. However, even this continuity is something of an illusion. The Basic State Pension (BSP), by far the largest of the remaining “contributory” benefits, is no longer contributory in any real sense – it is virtually universal. The unemployed, the sick and those caring for children all build up BSP entitlements as though they were making NI contributions.

In truth very little of today’s welfare system bears even a passing resemblance to the system envisaged in the Beveridge Report. To some extent this is the product of the substantial differences between the context in which the report was written and the world we live in today. His proposals were for a country in which, for the most part, men worked and married women didn’t, the only lone parents were widows, and life expectancy was lower than the pension age. The UK today is a very different place. From 96% in 1949 the share of working age men in work has fallen to 76%, while the share of working age women in work has risen from around 40% to 66%. At the time of the report less than 1 in 20 births was outside marriage; today more than 1 in 5 children grow up in a lone parent household. Male life expectancy has climbed from around 63 in 1940 to 78 in 2010, but the male state pension age remains the same as in Beveridge’s day.

These changes in employment, family structure and longevity exposed the limitations of the report. A system based on the contributory principle could not accommodate groups in obvious need of support but without a history of contributions, such as lone parents and the long-term unemployed. And as early as the 1950s, the pension demands of an ageing society led to the abandonment of any actuarial link between contributions and benefits. National Insurance Contributions stopped being flat rate in 1961.

As the role of the contributory principle in the welfare system has declined over time, successive governments have relied increasingly on means-testing to determine eligibility. This shift has been particularly dramatic for those of working age, as illustrated by the figure. Between 1978-79 and 2012-13, the share of expenditure on working-age welfare payments accounted for by means-tested benefits has increased from 26% to 80%. Of the £95 billion budget for working age benefits and tax credits this financial year, more than £75 billion will be means tested, compared with spending of less than £10 billion on contributory benefits. (Note that we have included Child Benefit in the set of means-tested benefits for 2012-13). Much of the increase in the share of means-tested expenditure is the result of extending the reach of the social security system beyond tightly defined groups (the unemployed, carers etc.) to all those with low incomes. In 2011-12, the government spent £22 billion boosting the incomes of families in work through tax credits, an idea that would have seemed bizarre to Beveridge who took it for granted that those in work would not need support.

Expenditure on working age welfare

Note:All social security spending on pensioners is excluded, not just that on specific pensioner benefits. Child benefit expenditure in 2012-13 is counted as means-tested.

Source: Authors’ calculations from Department for Work and Pensions Benefit Expenditure Tables and HMRC accounts.

Why does this history matter? For three reasons. First, because the result of appending a means-tested system to a social insurance framework is more complexity and less transparency than necessary. Second, because we maintain the fiction of an insurance system through National Insurance Contributions, which in fact act just like an additional income tax for most people most of the time. As we have argued previously, this is complex, costly and distorting and seems to be used primarily to make tax rises less transparent. And third because this history should lead us to recognise the need for our welfare system to adapt in a planned way as the world changes.

70 years after the Beveridge Report, the government faces some big choices as it seeks to make substantial cuts to the social security budget. These decisions need to be taken with the long term in mind and with a view to creating a coherent system that is both affordable and effective in meeting the needs of a changing population.

]]> Mon, 03 Dec 2012 00:00:00 +0000
<![CDATA[What is the impact of a 45p minimum unit price for alcohol?]]> Today the Home Office has published its consultation on policies to reduce the social costs associated with alcohol consumption. The most high-profile measure is a planned minimum unit price (MUP) for alcohol for England and Wales aimed at ensuring that “alcohol can only be sold at a sensible and responsible price”. The Government proposes that the minimum price should be 45p per unit (a unit is 10ml of pure alcohol); the Scottish government is already planning to introduce a MUP of 50p from next April.

As we have argued before, a preferable way of establishing a price floor for alcohol would be to reform the existing alcohol tax system in two ways. First, move towards a more equal tax treatment of different types and strengths of alcohol in which the tax is based on alcohol content. Second, restrict alcohol from being sold for less than the total tax due on each product to prevent tax increases from being absorbed by retailers. Whilst this reform would affect all alcohol, not just cheaper products, it would at least ensure any additional revenues flow to the Exchequer. By contrast, a minimum price would act as a transfer to the alcohol industry. The size of any windfall from minimum pricing and precisely who benefits is unclear: it would depend on the supply- and demand-side responses to the policy and the contractual arrangements between producers and manufacturers. Nevertheless the sums involved could be large: if there were no behavioural response at all, we estimate that a 45p MUP could transfer almost £1.4 billion per year from alcohol consumers to the off-licence alcohol industry.

To look at the effect of a 45p MUP we use detailed data recording the off-licence alcohol purchased by more than 21,000 households in 2010. The data do not include alcohol bought in pubs and restaurants. However, it is unlikely that a minimum price would have much direct effect for on-trade prices.

The consultation claims that minimum pricing is intended to “reduce the availability of alcohol sold at very low or heavily discounted prices”. However, we find that very nearly 60% of off-licence units cost less than 45p (Table 1), suggesting that the impact of minimum pricing is not just on the very cheapest alcohol. There are big differences across alcohol types: 85% of cider units would be directly affected by a 45p MUP, whilst alcopops and sparkling wines are relatively unaffected.

Table 1: Average off-licence retail price per unit and proportion of units under 45p, by alcohol type

Average off-licence retail price per unit and proportion of units under 45p, by alcohol type

Source: calculated using data from Kantar Worldpanel, 2010. Note: prices uprated to October 2012 values using the all-items RPI.

The effects also vary across different types of household. Low income households would see the largest increase in prices. Households with incomes below £10,000 per year pay on average 43p per unit, and 69% of their units cost less than 45p. This compares to an average cost of 53p per unit for those with incomes above £60,000, for whom 44% of units cost less than 45p.

Heavy drinkers would also be more directly affected. Households consuming fewer than 7 units of off-licence alcohol per adult per week pay just over 50p per unit on average, compared to almost 42p for those consuming more than 35 units per adult per week. Despite this, the impact of minimum pricing would certainly be felt by moderate drinkers as well. More than 48% of units purchased by those drinking less than 7 units per adult per week cost less than 45p, compared to 73% of units for the heavy drinkers.

Table 2 estimates the increase in total food and drink spending that would result from a 45p MUP, assuming no changes in consumer or retailer behaviour except that alcohol sold below the minimum price rises to 45p per unit. The average effect is about 0.9% of grocery budgets, or £32 per year in cash terms. The impact is quite similar for different income groups on average. The largest effect, 8.3% of grocery budgets on average, is felt by heavy drinkers with household incomes below £10,000 per year. However, only 3% of households in this income group fall into this consumption range (a similar share to other income groups). By contrast, almost a quarter of households in this income group do not consume off-licence alcohol at all, and so would not be directly affected by minimum pricing.

Table 2: Proportional increase in food and drink spending following 45p minimum price, by income group and alcohol consumption level (units per adult per week)

 Proportional increase in food and drink spending following 45p minimum price, by income group and alcohol consumption level (units per adult per week)

Source: calculated using data from Kantar Worldpanel, 2010. Note:

The most important issue in determining the impact of minimum pricing will be whether those who generate the greatest social harms from their consumption drink less as a result of the policy. That will depend on how well targeted on those drinkers a minimum price is and how their consumption responds to price rises – something we know relatively little about. Even less is known about how producers and manufacturers might respond and what this will mean. For example, the Office of Fair Trading recently noted that a MUP could give retailers an incentive to market low-cost alcohol more aggressively as it will raise the profit margin on each unit sold. Other important questions are what happens to the price of alcohol which is already selling above the minimum price? What happens to the range of alcohol offered for sale? What happens to the prices of other products besides alcohol? All of these issues will be crucial in judging the efficacy of the policy and should form part of any robust evaluation of its impact. It would be preferable, though, to pursue reforms which would allow something close to a minimum price to be implemented through the alcohol tax system.

]]> Wed, 28 Nov 2012 00:00:00 +0000
<![CDATA[Possible changes to the Retail Prices Index: what they are and why they matter]]> In October the Office for National Statistics announced a consultation on possible reforms to the Retail Prices Index (RPI). These reforms could have far-reaching consequences. Today, we are publishing a working paper with our thoughts on technical aspects of the proposed changes.

The UK has two main measures of consumer price inflation, the Retail Prices Index (RPI) and the Consumer Prices Index (CPI), which can give quite different impressions of how prices are changing from year to year. Some of the differences between these two measures are quite easy to explain. For instance, the CPI does not include mortgage interest costs while the RPI does. But there is one difference that has been a source of confusion for those using the two indices: the RPI and CPI use different mathematical formulae to work out how prices are changing, meaning that even if they were fed the same raw price data, the two would report different inflation rates.

This impact of this specific difference between the RPI and CPI is known as the ‘formula effect’. Historically, it has consistently pushed up RPI inflation relative to CPI inflation (see Chart 1). Users of these indices have long been entitled to ask what reason is there for this difference – why is a given method preferable in one index but not the other? These questions only became more pertinent when the size of the formula effect nearly doubled in size following a seemingly minor change to the way clothing prices were sampled in 2010. That’s why in October the ONS started seeking views on whether to change the formulae used in the RPI to bring them in line with those used in the CPI (other options in the consultation would serve to reduce the formula effect without eliminating it entirely).

Chart 1: RPI and CPI inflation rates and the formula effect, 2005-2012

 RPI and CPI inflation rates and the formula effect, 2005-2012

Source: Office for National Statistics

So what are these different formulae and why should the ONS want to change them in the RPI? Indices like the RPI and CPI have to combine many different prices into a single inflation figure. They both start by working out the price changes of individual ‘items’ (such as ‘white unsliced bread’) from a detailed set of price quotes (such as prices of different brands of white unsliced bread). In the RPI, in some cases, a simple average is taken of the price changes of the individual brands. This is the RPI’s ‘formula’, which the ONS is considering replacing. In the CPI, on the other hand, a geometric average is usually taken. This method was first proposed in an essay on gold prices written in 1863 by the British economist William Stanley Jevons, and is now called the ‘Jevons index’. This difference may sound extremely arcane, but it has a big effect. According to estimates of the size of the formula effect shown in Chart 1, using the same averaging methods in the RPI as in the CPI would reduce RPI inflation by just under one percentage point on average each year.

The case against the current RPI formula essentially boils down to the fact that it can give quite odd results. For example, if prices for some item go up one year and then fall back to their original level the next, the RPI will show that item as being more expensive at the end of the period. This does not seem like a very desirable property for a price index to have, and few other countries still use the RPI method partly for this reason. The CPI’s geometric average does not lead to the same problem. However, as we point out in our working paper, moving to the CPI method would not prevent the RPI as a whole (rather than the RPI for particular items) giving these sorts of results, because of the way inflation rates for the individual items are later combined together into an overall figure. Indeed, the same problem occurs with the overall CPI as well. So this issue needn’t be fatal for the RPI’s formula, though taken together with other problems it does add to a cumulative case for change. The working paper goes into more detail about this and other issues.

In any case, even if we could all agree that the current RPI method is inappropriate, reforming it will inevitably create losers. Plenty of contracts are signed guaranteeing one party or the other a return based on RPI inflation. For instance, owners of inflation-indexed bonds get a return equal to RPI inflation plus some yield each year (protecting the value of investments from being eroded by unexpected increases in prices). If RPI inflation were to be reduced, then they would see their investments fall in value, with holders of long term bonds suffering the greatest losses. Importantly, most benefit payments and tax rates have already switched to being linked to the CPI and so any change to the RPI would not affect them. One exception might be increases in excise duties such as fuel, alcohol and tobacco taxes, which are still linked to RPI, though one would expect the government to adjust the way these are uprated if the RPI were changed.

The ONS consultation will close by the end of November, and the intention is that any change would be implemented from March 2013. The eventual decision would have to be considered by the Bank of England, who would assess whether or not the change would be “materially detrimental” to bondholders. If so, responsibility for approval would then pass to the Chancellor. Of course, from his perspective, a reduction in the amount of interest he needs to pay to the holders of government debt would be a welcome windfall - helping him cut the deficit at a stroke. Jevons himself used his own index to assess the consequences of new gold discoveries on prices and noted that "The most remarkable effect of the depreciation of gold is a considerable reduction in the National Debt". Using his index in the RPI might have rather similar consequences today.

]]> Tue, 27 Nov 2012 00:00:00 +0000
<![CDATA[Late changes to Council Tax Benefit reforms]]> In a ministerial statement last week the government announced a significant change to its policy to localise Council Tax Benefit (CTB) from next April. In this observation we ask why such a significant change has been announced to a policy two years after it was first announced, less than six months before councils will have to implement it and after many have already consulted on the structure of proposed schemes.

CTB provides support to 5.9 million low-income families, more than any other means-tested benefit or tax credit in the UK. The government is proposing to localise support for council tax from 2013–14, abolishing CTB across Britain and giving grants to local authorities in England and to the Scottish and Welsh governments to design their own systems for providing support for council tax to low-income families. On top of this, the government planned to cut by 10 per cent the funding it provides for council tax support. This would save around £500 million a year.

We have analysed the effects of these proposals in some detail, concluding that localisation would create considerable complexity just as Universal Credit is being rolled out with the intention of simplifying things. It also has the potential to undermine many of the improvements to work incentives that Universal Credit is intended to deliver. For councils to save the full 10 per cent by which funding was being cut by making the system less generous, either the means test would have to be so severe that some people would be worse off after a pay rise – or else councils would have to collect some local tax from the very poorest for the first time since the poll tax. Many councils are consulting on schemes which would have these sorts of consequences.

Just last week - two years after the policy was originally announced, less than four months before local authorities have to finalise their new schemes, and only a week before the third reading of the bill in the House of Lords - new proposals have been forthcoming. In a ministerial statement a £100 million package was announced. This money – which amounts to a fifth of the total planned savings – will be available to councils whose schemes meet a particular set of criteria that the government considers “best practice”. It will, apparently, be available for one year only.

Councils will be eligible for the money if nobody currently on full CTB ends up paying more than 8.5% of their council tax liability (in practice, the costs of collecting such small amounts from very low income households who are not used to paying council tax mean that councils may well prefer to give a full rebate to such households); if the rate at which the benefit is withdrawn as income rises is no higher than 25% (compared with 20% at the moment); and if there are no “cliff edges” in the system.

Even with an extra £100 million to soften the blow, it is hard to see how most councils could design schemes that meet these criteria within the reduced funding intended for council tax support. So it looks as if the government is aiming to pay councils not only to design schemes that the government likes, but to design schemes that don’t cut support as much as councils’ funding is being cut, leaving them to make up the shortfall from elsewhere in their budgets.

It is hard to square this development with a policy whose stated aim was to devolve responsibility. And why the additional money should be appropriate in the first year of the policy and not later is unclear. But perhaps most worrying is what this says about the policy-making process. The potential downsides that the government seems to be trying to ameliorate – losses for the poorest households and weakening of work incentives – have been obvious to many observers for a long time. Yet this announcement has come very late in the process. The bill had already completed its passage through the House of Commons and scrutiny by a committee of the Lords had finished: last week’s announcement came on the eve of a key Lords debate on amendments to the bill, and just a week before the third reading in the Lords. Many councils have already been running public consultations on draft proposals (as the bill requires them to do) yet are now being incentivised to change their proposals at the last moment – perhaps only to revert to their original plans when this extra funding is withdrawn a year later.

The case for well thought through reform of the welfare system is overwhelming. The dangers of less fully considered reform - as this one appears to be - are considerable.

]]> Thu, 25 Oct 2012 00:00:00 +0000
<![CDATA[The £10 billion question: where could the Chancellor find welfare cuts?]]> In his speech to the Conservative party conference today, the Chancellor of the Exchequer again stated an intention to reduce welfare expenditure by a further £10 billion per year by 2016–17.

Given the £18 billion per year of welfare cuts already planned by 2014–15, total welfare spending in that year is currently forecast to be £214 billion. Of that, £94 billion will be spent on state pensions and Pension Credit. If these are protected from cuts, the additional £10 billion of savings would have to be found from the remaining £120 billion of welfare spending: a cut of 8.3% in that part of the budget. The figure below illustrates who is currently entitled to these benefits and tax credits, showing average entitlements for families in each decile group of the working-age income distribution. It shows that approximately ¾ of the entitlements go to the lowest-income half of working-age families, and ¼ to the top half (note that Child Benefit is already set to be withdrawn from families containing an individual with annual taxable income exceeding £50,000 from January 2013, which will reduce the fraction of welfare entitlements going to higher-income families).

Figure: Average weekly entitlements to benefits and tax credits among working-age families in 2012–13

 Lifetime repayments under old and new systems

Note: Income decile groups are derived by dividing all families into 10 equal-sized groups according to income adjusted for household size using the McClements equivalence scale. Decile group 1 contains the poorest tenth of the population, decile group 2 the second poorest, and so on up to decile group 10, which contains the richest tenth.

Source: Authors’ calculations using 2009–10 Family Resources Survey and TAXBEN, the IFS’ tax and benefit micro-simulation model.

One simple way of reducing welfare expenditure would be not to go ahead with the usual indexation of benefits in line with prices next April. The Chancellor asked “how can we justify the incomes of those out of work rising faster than the incomes of those in work?” A freeze for all working-age benefits and tax credits would save around £2 billion per year. But the Chancellor’s comments might suggest that this would only apply to out-of-work benefits (Jobseeker’s Allowance, Employment and Support Allowance and Income Support), which would – as the figure makes clear – substantially reduce the potential savings.

The Prime Minister and the Chancellor have mentioned two other specific areas where they believe the current benefit system is too generous. The first is Housing Benefit for the under 25s, who comprise almost 400,000 (8%) of the 5 million Housing Benefit claimants. Abolition of Housing Benefit for this group would save nearly £2 billion per year. Realistically though, it seems highly likely that some of the group will be made exempt from this cut. A crucial issue here is how the government would distinguish between those who can and cannot reasonably be expected to live with their parents. For example, more than half of Housing Benefit spending on under-25s goes to individuals who themselves have dependent children: might the government include them in the group who could reasonably be expected to live with parents? The answers to these kinds of questions will determine how much less than £2 billion per year the government would save from this policy, and how workable it would be in practice.

The second area that has been highlighted is state support for large out-of-work families. We have argued previously that this principle seemed implicit in an existing government policy: the overall household benefits cap, to be introduced in April 2013, will in practice only affect families with large numbers of children and/or high housing costs. If the government believes that these families are currently entitled to too much, it is better to be explicit about that when designing welfare policy so that the perceived problem is targeted as precisely as possible, rather than addressing it indirectly via an overall cap on benefits. In terms of the revenue consequences, though, this does not get us to £10 billion per year of welfare cuts. About 330,000 out-of-work families have at least three children, meaning that to save £1 billion a year would require entitlements to be cut by an average of £3,000 per family for this group.

As always, then, there are all sorts of tradeoffs and issues to think about when designing welfare policy. It is clear, however, that there is more we have yet to hear about if the government is to cut the welfare budget by an additional £10 billion per year.

]]> Mon, 08 Oct 2012 00:00:00 +0000
<![CDATA[Reforms to higher education finance: who wins and who loses?]]> The Government’s reforms to higher education funding - involving an increase in the cap on tuition fees to £9,000 per year and the removal of most direct funding for teaching - have now been implemented. This has led to greater variation in fees across universities, and even across different subjects within the same university, although average headline fees are, at £8,660 per year, close to the cap.

Accompanying this, however, are changes designed to make the student finance system more 'progressive'. Its basic principle remains unchanged: loans are available to undergraduates for tuition fees and living costs, which they pay back once in employment. Compared to the old system, however, the earnings threshold above which graduates make repayments has been increased from £15,795 p.a. to £21,000 p.a. (in 2016 prices), the maximum period over which those repayments are made has risen from 25 to 30 years, and many graduates will face above-inflation interest rates for the first time ever. Up-front cash support for most students has also risen, including a £50-million government contribution to the National Scholarship Programme to support students from the poorest backgrounds. This will triple in value to £150 million by 2014.

Recent IFS research, supported by the Nuffield Foundation, provides the first detailed analysis of the financial implications of these reforms for students, graduates, taxpayers and universities. The new system eventually saves the taxpayer around £1800 per graduate, driven by a dramatic cut in direct public funding to universities. But for universities, this cut is more than offset by almost £15,000 in additional fee income per graduate - a 140 per cent rise over the old system. Thus the total amount spent - from both private and public sources - on higher education is expected to increase as a result of these reforms. On average, universities will be better off financially as a consequence.

The average student will also be better off while at university, enjoying an increase in cash support of some 12 per cent. But the main 'winner' from the reforms is the taxpayer while the main 'loser' is the average graduate, marking a significant shift in the burden of higher education funding away from the public sector and towards private individuals.

However, these headline changes for the average graduate mask some important variation. To consider in more detail the impact of the reforms on graduates, we estimate how much each graduate in a cohort would be expected to repay over their working life. Figure 1 shows how this total repayment varies with total lifetime earnings, under the old and new systems. The two lines cross at around the 27th percentile of the earnings distribution; in other words, the poorest 27 per cent of graduates will actually be better off under the new system.

Figure 1. Lifetime repayments under old and new systems

 Lifetime repayments under old and new systems

Low-earning graduates benefit from the increase in the earnings threshold, which (combined with the debt write-off after 30 years) ensures that the majority of their loan is never repaid. This makes the new system substantially more progressive than its predecessor: the richest graduates are likely to repay ten times as much as the poorest, and would even pay back more than the value of what they borrowed.

What does this imply for university attendance amongst disadvantaged students? The progressive features of the repayment system should provide some grounds for optimism: as long as students are well informed and not averse to the kind of debt involved - repayments of which only depend on one’s ability to pay - participation rates should not suffer. But there are grounds for concern if students have difficulty understanding the complexities of the new system - which are substantial - or if they are deterred by the prospect of higher borrowing regardless. Efforts to increase participation amongst students from disadvantaged backgrounds will require clear, precise information to be provided about the costs and benefits of going to university in both the short and long run. Only time will tell if that goal has been achieved.


]]> Tue, 02 Oct 2012 00:00:00 +0000
<![CDATA[Pensions for the masses]]> On Monday, a very radical reform of workplace pensions in the UK will start to be rolled out across the country. The majority of employees who work for large private sector companies will - for the first time - find themselves enrolled automatically into an employer-sponsored pension scheme; for some of these people it will be the first time that their employer has offered to make a contribution to any pension scheme - a minimum of 3% of gross salary (within a band), provided the employee contributes at least 4%. Employees have the choice to opt out subsequently, but those doing so will risk losing the contribution from their employer. Ultimately, most employees aged between 22 and the State Pension Age will come within the scope of this policy, but it will take some time - these new rules will not cover all employers until February 2018. A new, state-owned pension provider - the National Employment Savings Trust (NEST) - has been established to ensure that all employers can provide their employees with access to a relatively low-cost pension scheme.

Automatically enrolling employees into a pension scheme is likely to have a significant impact on the number of private sector employees who contribute to a pension. In 2011, just one-in-three private sector employees were members of a workplace provided pension scheme (compared to over four-in-five public sector employees). An increase in pension coverage is likely for two reasons. First, evidence suggests that when employees are defaulted into a pension more will remain a member of that scheme than would have made an active decision to join a scheme. Second, many private sector employees will find themselves able to receive a pension contribution from their employer for the first time, which will make pension saving financially more attractive. The low level of coverage among private sector employees pre-reform gives plenty of scope for a significant increase in pension coverage.

While there are good reasons to expect pension coverage to increase, the impact on overall saving is less clear. This is for a number of reasons. First, evidence suggests that defaulting employees into pensions leads to more individuals choosing to contribute the default amount, perhaps because individuals decide not to bother reviewing this level of contribution in the belief that the default amount has, in some sense, been recommended. While this would lead to an increase in pension saving among those who would not otherwise have saved in a private pension, it would lead to a reduction in saving among those who would otherwise have chosen to save more than the default amount. Second, if individuals do increase their pension saving as a result of this reform, this does not necessarily mean that overall saving will rise. An increase in pension saving could be funded by a reduction in the amounts households save in other forms or, arguably even worse, by some running down their debts less quickly than they would have done. Third, the cost of increased employer contributions (and any additional costs to businesses of administering new systems) will need to be financed from a combination of lower wages, higher prices or lower profits; all of these could depress saving.

Although the impact on total saving in the economy is ambiguous, automatic enrolment should nonetheless lead to a significant boost in pension coverage among private sector employees, and those who have good reasons for not saving in a pension at the present time remain free to opt out if they wish. Current economic conditions may lead to opt out rates being higher than was anticipated when this policy was first suggested by Adair, now Lord, Turner in 2006. Evidence on the actual impact of automatic enrolment - on pension coverage, pension saving, overall saving, and levels of earnings - should be gathered in order to help inform future decisions over the precise policy design. However, a challenge for anyone hoping to answer these questions is that the reform is being implemented in a way that will make robust estimation of its impact difficult; careful consideration of data will be needed to shed light on precisely what effect automatic enrolment is having. Simply getting more people to save in a pension will not achieve the Government's overall objectives if it is also accompanied by a reduction in the amounts saved into pensions or saved in other forms.

There are a number of features of both automatic enrolment and the establishment of NEST that will need to be monitored closely and perhaps reformed once the initial impact of the policy changes has been seen. On automatic enrolment, two key decisions will be whether the default minimum contribution rates should be changed, and whether administrative burdens on employers (which have been of particular concern to small businesses) could be further eased without inducing other negative consequences. As far as NEST is concerned, a key decision to be made is whether any of the additional regulatory restrictions on NEST (such as annual contribution limits and restrictions on transferring existing funds in), which do not apply to other pension providers, should be lifted. Ideally, in the longer-term there would be open competition between pension providers, with NEST being neither advantaged nor disadvantaged relative to other market players.

]]> Fri, 28 Sep 2012 00:00:00 +0000
<![CDATA[GCSEs - in need of reform? ]]>

Tomorrow, thousands of 15 and 16 year olds will receive their GCSE results. In the future, these exams may be considerably different. It has been reported that the education secretary, Michael Gove, is planning to make radical changes to the structure of GCSEs in England. One potential change is to ensure there is only one exam board such that there is no competition between providers encouraging schools to choose the easiest option. The other is to replace the current system of GCSEs with a dual system of exams, one harder and one easier. Recent analysis by IFS researchers throws light on the desirability of these potential reforms and whether they would actually represent a substantial change compared with the current system. The evidence also raises a third and more fundamental question: what are GCSEs for?

Since the first group of 15 and 16 year olds sat GCSEs in 1988, grades have trended up year-on-year. Between 1996 and 2011, the proportion getting five GCSEs at A*-C (including English and Maths) increased from 35% to 58%. These increases could reflect improvements in student's knowledge. However, the sheer pace of this increase may reflect 'grade inflation', with exams getting easier. It is hard for outsiders to judge - itself a problem to the extent that it reflects a lack of confidence in the system. And the regularity of the year on year increases can reinforce that concern.

However, there are reforms that could reduce perverse incentives for schools in the current system, which is one motivation behind reforms to GCSEs proposed by the education secretary.

Choosing exam boards and qualifications

The qualifications system in England and Wales currently involves a number of exam boards competing for shares in a regulated market. One danger is that they might compete by offering qualifications that are easier to pass. A school choosing between two qualifications which will count the same in league tables, and perhaps to the outside world more generally, has a clear incentive to choose the one it thinks more of its students will pass - the easier one. It is hard to judge the degree to which this has actually contributed to rising GCSE results and multiple exam boards may be able to offer more diversity than just one. However, the incentives are clear. The theoretical case for eliminating such perverse incentives seems strong.

Rising GCSE results in the late 2000s were also buoyed by a proliferation of "GCSE equivalent" qualifications, which meant that by 2009-10 42% of pupils took 'Vocational-Related Qualifications', 30% took BTECs and 10% took GCSEs in vocational subjects. These 'GCSE-equivalent' qualifications tend to be taken by more disadvantaged pupils. Work by IFS researchers showed that schools that had moved most aggressively into these GCSE-equivalent qualifications were also the ones that improved their league table position by the most. This is suggestive evidence that schools use variation in qualification difficulty to game the league-tables. Correlation is not causation, but evidence also led the Wolf Review to similar conclusions. The government is currently implementing the recommendations of this review, including slimming down the number of vocational qualifications that count towards league tables.

A dual system

At the same time, Mr Gove is reported to be considering replacing GCSEs with a dual system of harder exams like the old O levels and easier ones like the old CSEs. His stated aim is to challenge more able students and to ensure that higher standards are maintained. Given that there are already different tiers of GCSEs and many other vocational qualifications, it is not yet clear how much difference any such change would make in practice. Nor is it clear how a new system of exams would affect the behaviour of schools and parents.

Of course there is much room for improvement in school standards. According to international rankings, the UK performs around the OECD average for reading and maths, slightly above average for science, but well behind high-performers such as Canada, Finland and South Korea. Perhaps Mr Gove's proposals could help deal with this?

The trouble is that if you dig under the average results from international comparisons what you find is that the performance of pupils in England is rather unequal with some strong associations between social class and performance. A worrying finding from the research cited above is that it has tended to be poorer pupils who have been directed away from traditional GCSEs in recent years. Recent research published in Fiscal Studies also suggests that by international standards there is a particularly big gap between social groups in the performance of the most able - with the most able pupils from lower social groups doing much worse than the most able from more privileged backgrounds. Seeking to stretch the most able pupils from all backgrounds may well help to deal with some of these problems. However, there may well be other factors holding back able pupils from disadvantaged backgrounds, such as differences in pupil attitudes or aspirations. Such a policy also seems unlikely to aid less able pupils in general.

The role of GCSEs

There is perhaps a third more fundamental question that we should be asking, and which certainly needs to be stated clearly before any reforms are introduced. And that is, what exactly are GCSEs for? England is actually rather unusual in having a high stakes school leaving exam at 16. Most countries focus on exams when most young people in fact leave school at 17 or 18. The system in England looks rather like a left over from a time when the majority of young people did expect to leave school at 16. Now that the vast majority stay on past 16 to do further qualifications there must be some question over the role of a set of exams which may signal to some that leaving at 16 is expected, particularly in the context of government policy to raise the "education participation age" to 18.

GCSEs do perform other roles as well. They are often used to hold schools to account for their performance and are one of the only external measures of attainment universities can see when making offers of places. However, other measures of school success could be used in league tables (such as a core of subjects or post-16 results) and may well be more desirable if schools have been using the present system to boost their league table position. It would also be odd to justify retaining GCSEs on the basis they are used for university admissions. Currently, the majority of children don't go to university and other reforms could improve the flow of information to admissions tutors, such as entrance exams or running the application process after A-Level results have been published. England is also extremely unusual in allowing those who do stay on past 16 to drop study of maths and English, as was pointed out in the Wolf Review.

Perhaps an even more radical rethink of the role of GCSEs and the structure of the public examination system is called for if we are to ensure that these exams serve a valid purpose and young people are best served for the future.

]]> Wed, 22 Aug 2012 00:00:00 +0000
<![CDATA[Behavioural economics and tax reform]]> Recent years have seen a rapid expansion of interest in behavioural economics, which draws on ideas from psychology to shape economic theory. A burgeoning body of evidence – both from experimental laboratory studies and out in the field – suggests that behavioural insights can help to better explain observed behaviours. This observation considers three ways in which behavioural economics might be important for the design, implementation and reform of tax and benefit policy.

First, as we discuss in much more depth in new research funded by the Nuffield Foundation,* behavioural insights help in thinking about how people respond to policy. The way taxes and benefits are presented (even in seemingly trivial ways like their name) can have an effect on their impact: previous IFS research has demonstrated that Winter Fuel Payments are much more likely to be spent on gas and electricity than unlabelled income. People may also respond differently as policies become more complicated, perhaps because they resort to simpler rules of thumb to guide decisions when things become hard to process. For example, there is evidence that people confuse marginal tax rates (how much of the next pound earned will be taken in tax) with average tax rates (how much of everything earned is taken in tax) when deciding how much to work, perhaps because average rates are easier to understand if the structure of marginal rates is complicated.

Second, behavioural insights could help explain why tax reform is often hard to implement, despite the evidence set out in the recent Mirrlees Review of the likely benefits of sensible reforms. The Review made the case that policymakers should think about reform in the context of the whole tax and benefit system, when thinking about issues like progressivity or work incentives for example. However, experimental evidence has found that people suffer from a phenomenon known as ‘disaggregation bias’: a tendency to focus on individual components of something rather than the whole. Since taxes and benefits interact in a number of ways, reforms to the system inevitably involve changes to different component parts amongst which the big picture effects can be lost. For example, expanding the base of VAT and using other taxes and benefits to compensate low income households could help improve the overall efficiency of the tax system and leave it as progressive as before on average. But the regressive nature of the VAT reform in isolation would almost certainly be the focus of attention of such a change and make the package of reform less acceptable. Governments may respond to disaggregation bias on the part of voters, but policymakers may well suffer the same bias themselves since their attention is often focused on a single aspect of the wider system.

Behavioural factors can also mean that even small tax changes are sometimes hard to implement. The model of prospect theory suggests that people feel a loss much more acutely than a gain of equivalent size, and that the additional pain caused by larger losses gets smaller and smaller. As a result, lots of small tax increases could be perceived as more costly than a single big rise. This could help explain why tax escalators, which increase tax rates by a small amount year after year, have often not been implemented as planned. Similar issues might even have some resonance for why recent attempts to impose the full rate of VAT on (amongst other things) takeaway pasties and fixed holiday caravans (raising probably less than £100 million) failed, whereas a £12 billion increase in the VAT rate went ahead.

Finally, as the IFS has long argued, tax and benefit policy should be based as far as possible on good evidence. This is equally true in thinking about what behavioural insights mean for policy, and our research looks in detail at the evidence base. Some recent evidence has come from policy trials looking at the impact of ‘nudges’ which relate to aspects of tax policy. For example, HMRC and the Behavioural Insights Team at the Cabinet Office conducted an experiment with the wording of letters sent to people asking them to pay outstanding taxes. People receiving a letter telling them about compliance rates in their local area were much more likely to respond quickly than those sent a generic letter, suggesting a role for ‘social norms’ affecting behaviour. Such findings help us to understand the usefulness of behavioural insights in making aspects of tax policy more effective, and we applaud the growing use of controlled experiments in this way.

However, behavioural economics is not just about developing ways to nudge people, and our research makes it clear that these wider insights are important for tax and benefit policy more generally. The most pressing evidence gap relates to measuring the welfare costs of getting policy ‘wrong’ by not taking behavioural insights into account. For example, how much higher are optimal cigarette taxes if people cannot commit to giving up? Are back-to-work bonuses or benefit sanctions more effective to stop people procrastinating when searching for work and how large should they be? What will the new system of Universal Credit mean for people’s perception of stigma in the benefits system and take-up rates?

The challenge going forward therefore seems to be developing the evidence base to answer these sorts of questions. This will involve three elements: 1) carefully considering how behavioural issues might impact on policy at the design stage 2) ensuring that policy evaluation (as far as possible) allows us to disentangle potentially conflicting behavioural factors that might drive people to respond to interventions in different ways and 3) using the results to develop and refine theories of behaviour.

*The Nuffield Foundation is an endowed charitable trust that aims to improve social well-being in the widest sense. It funds research and innovation in education and social policy and also works to build capacity in education, science and social science research. The Nuffield Foundation has funded the project on which this observation is based, but the views expressed are those of the authors and not necessarily those of the Foundation or the Institute for Fiscal Studies. More information is available at

]]> Thu, 19 Jul 2012 00:00:00 +0000
<![CDATA[Are you really "in the middle"?]]> How rich are you? When asked this question, most people think they are in the middle. From people in the poorest third of the income distribution, up to people in the richest five or ten per cent, many will happily report that they think they're in the middle. Obviously, not everyone can be in the middle and IFS has used the latest data on household incomes to give you an accurate picture of where you actually are in the distribution.“Where do you fit in?” is a simple tool that lets you know your position in the income distribution and can be accessed online here or downloaded as an iPhone App

So how can you compare your income to that of other people in the UK? One way is to compare your salary to that of other people working in the UK. In 2011, the median gross salary for people in full-time work was around £26,000 per year. But this does not tell the full story. Of the 48 million adults in the UK, about 11.7 million are above state pension age and mostly live off income from pensions. About 5 million working-age adults live in households where no-one works and are thus largely dependent on out-of-work benefits. There are many people in work receiving benefits as well. Some people have income from self employment or investments and some pay higher taxes than others. All of these factors affect the total income that households can use to spend on the items of their choice, and hence their material living standards.. Moreover, a single person on the median salary of £26,000 will tend to have a significantly higher standard of living than a person supporting a spouse and child on the same salary.

While there are many ways to measure income, let alone broader measures of living standards, here we use the same definition as the government in its official income statistics. We measure total household income after direct taxes (including council tax) are deducted and after all benefits are received. We also make an adjustment for family size so that it is possible to compare the living standards of households with different numbers of people living in them.

In using “Where do you fit in?” all you have to do is enter the number of adults and children in your household, add up the take-home income of each person (don’t forget to include any benefits you receive) and enter how much council tax you pay. Our tool will show you where you and your household fit into the UK’s income distribution. Who can we call average? That depends on family type. Imagine a family with two adults and two young children. If both parents go to work and each of them earns a gross salary of £19,000 per year, then they have a net income of around £600 per week. Adjusting for household size, this leaves them with income near the middle. Interestingly, this is only just above the level of gross earnings that a recent report by the Joseph Rowntree Foundation deemed necessary to achieve an adequate standard of living for this family type, according to society’s view of acceptable living standards. Thinking about other family types in the middle of the distribution, a couple without children who both have gross annual salaries of £14,000 would also be right in the middle.

Those towards the bottom of the income distribution are more likely to be on benefits. For example, a single unemployed person living alone receiving Jobseeker’s Allowance plus Housing Benefit of £80 per week to cover their rent would be at the 10th percentile. A single pensioner with similar housing costs and dependent on the Basic State Pension and Pension Credit would be at the 32nd percentile.

At the other end of the spectrum, many might be surprised to find that a childless couple each earning £30,000 (before tax) would find themselves in the top 10% of the income distribution and would be in the top 1% if earning £85,000 each. Of course, those with children need rather more to achieve the same living standards. A single earner with two children would need to be on close to £100,000 to be in the top 10% and £285,000 to be in the top 1%.

Overall those in the middle of the distribution have income about twice the level of those at the 10th percentile, while income at the 90th percentile is twice that at the median. Within the top 10%, incomes shoot up as you get closer to the top of the distribution, and have been ‘racing away’ more and more quickly over time in recent history.

We hope that you find “Where do you fit in?” informative. Punch in your numbers and you might be surprised where you end up.

]]> Tue, 17 Jul 2012 00:00:00 +0000
<![CDATA[The fiscal challenge ahead]]> The Office for Budget Responsibility's (OBR's) second fiscal sustainability report, published today, reminds us of the daunting task facing governments even after the current economic and fiscal crisis has abated. Demographic change and the potential loss of revenue from some important taxes will open up a further hole in the public finances of the same sort of magnitude as that which the current fiscal consolidation is designed to deal with. In other words we are likely to need to find many tens of billions of pounds, and probably another £100 billion or more, of spending cuts and tax increases to deal with these pressures.

There is some good news. We do have some time to prepare. Whilst the current £123 billion of fiscal consolidation is planned to take place over a seven year period, the OBR is looking 50 years ahead. In addition there are signs that the government is paying some heed to growing demographic pressures - announcements, for example, to increase state pension ages to 66 by 2020 and 67 in 2026 with the possibility of then linking it to changes in longevity.

But the scale of the change is such that much more will be required. On the spending side there are two main budgets which, under current policies, will be dramatically increased by demographic change - that for state pensions and that for health care. As we showed, based on last year's OBR report, even on conservative assumptions, between them health and pensions are projected to account for half of everything that government does by 2060. As today's OBR report illustrates in stark terms, there are really big uncertainties here about what will happen to health spending. If it increases only as a result of ageing, it will take up an additional 2.3% of national income by 2060. If, on the other hand, spending on health rises faster than this - as it has typically done in the past - as productivity rises less fast than in the rest of the economy, then spending on health could more than double to over 16% of national income. If taxes aren't increased to accommodate these costs then other budgets will have to be squeezed quite dramatically or the quality and quantity of services provided free at the point of use by the NHS scaled back considerably.

As we also pointed out, to some significant extent the state has already adjusted to shifting pressures. Demographic change and spending pressures are not new. In 1979, health accounted for one pound in every ten spent by government. That proportion has already nearly doubled. It is set to go on rising. Social security spending has also taken an increased share of spending over the last 30 years rising from 23% of the total to 29%. That means of course that other elements of spending have been squeezed - defence, housing and support for industry have all lost out. Are there now budgets which can cope with a similar squeeze over the coming decades? The housing budget has all but disappeared as has government spending on industry and subsidies for teaching at universities. There is a limit to what can be found from defence. Budgets for the Ministry of Justice, the Home Office and grants to local government are all being cut by more than a fifth over this spending review period with more pressure likely to come in the next period. Low hanging fruit is becoming harder to find.

So what about the other side of the government balance sheet - tax revenues? Here, unfortunately, the OBR also paints a rather gloomy picture with respect to some significant sources of revenue which are in imminent danger of falling or disappearing entirely. We have seen in the last month how difficult this government, in common with its predecessor, is finding raising fuel duties even in line with inflation. Put this alongside increasing efficiency of cars and the result is a long term decline in revenues from a tax which currently raises around £30 billion a year. This decline is likely to continue and, if we are to meet our climate change targets, these revenues should eventually disappear entirely. As we have argued the long standing economic case for moving to a system of road user charging is becoming ever stronger.

Other pressures on the revenue side include those on taxes from North Sea oil and gas, which the OBR forecasts will fall to very low levels by 2040. In addition there are likely to be continuing pressures on other tax bases, notably corporation tax, as a result of globalisation and potentially increasingly fierce international tax competition. Up until now industrialised countries, including the UK, have been remarkably successful at maintaining corporate tax revenues over long periods (despite high profile cases of large companies managing to minimise tax payments). But we are beginning to see international constraints impact on tax system design, not least in the introduction of the so called 'patent box' designed to impose a much lower rate of corporation tax on profits derived from patents - profits which are considered more internationally mobile.

Of course the future is uncertain. But the OBR's report highlights an inescapable fact - we will need to make a whole series of very big and serious choices over the coming decades about the type of state we want to live in. In the end we have only three choices - increase taxes, rein in and reform spending on health and pensions, or cut other areas of spending. There really is no fourth way.

As we have said before this requires a serious debate. We need to think about how we want to run and fund our health service. We urgently need a serious debate about the structure of our tax system. At the very least we will need to replace the revenues lost from motoring, oil and, probably, corporate taxes. If we decide to accommodate some of the spending pressures by taxing more, then some combination of income tax, National Insurance Contributions and VAT are likely to have to take the strain. As we have argued forcefully in the Mirrlees Review there are many ways of changing the tax system to make it more efficient. If we are to do more with the system then it becomes all the more important that it be as efficient as possible. And that will mean tackling some long standing problems that politicians of all persuasions have consistently shied away from. As well as needing to grasp the nettle of road pricing we need, among much else, to sort out a VAT system which applies zero rates to more spending than almost any other comparable country, a council tax system which remains based on relative house prices in 1991 and is regressive with respect to its tax base, and our income tax and national insurance systems that run pointlessly and confusingly in parallel with one another.

The OBR's analysis is welcome and it is powerful. It should be sparking a national debate about how to respond to the challenges it is warning us about.

IFS public finance observations are generously supported by the Economic and Social Research Council (ESRC).

]]> Thu, 12 Jul 2012 00:00:00 +0000
<![CDATA[Whitehall departments overachieve budget cuts in 2011–12]]> Today HM Treasury published Public Spending Statistics which contain, for the first time, estimated spending outturns by Whitehall departments for 2011-12. The figures show that, in total, departmental spending was £6.7 billion lower than was planned this time a year ago. Had departments spent all their planned budgets in 2011–12, departmental spending would have fallen in cash terms by £4.4 billion between 2010-11 and 2011-12 - a real terms fall of 3.5%. As a result of the underspends, cash departmental spending actually fell by £11.0 billion, which equates to a 5.2% real terms fall.

The £6.7 billion underspend is comprised of a £5.3 billion underspend by Whitehall departments, and £1.4 billion of funding held in reserve by HM Treasury this time last year that subsequently did not need to be allocated to departments. The £5.3 billion underspend is equivalent to 1.4% of departments’ budgets. In other words, on average departments spent 98.6% of the money they were allocated in 2011–12; although – as we shall show – there was considerable variation in the relative size of underspends across departments.

One obvious reason why departments might underspend is that they might have been given permission by the Treasury to move funds into the next financial year under the new ‘Budget Exchange System’. Under this system departments who surrender any underspend (up to a 'reasonable limit') in advance of the end of the financial year can get an equivalent increase in their budget in the following year. However, of the £5.3 billion underspent by departments in 2011-12, only £0.9 billion was successfully surrendered by departments through the Budget Exchange, and will now be available to be spent in 2012-13. This means that over 2011–12 and 2012–13 the total government deficit is on course to be £4.4 billion lower (i.e. £5.3 billion less £0.9 billion) as a direct result of these underspends.

The majority of individual Whitehall departments underspent on their 2011-12 budgets. The figure below shows, for departments who underspent by more than £0.1 billion, the proportion of their budgets that were surrendered through Budget Exchange (and will therefore be available to spend in 2012-13) and the proportion of their budgets that were underspent and will not be transferred into future years. The department that underspent the largest proportion of their budget was the Department for Energy and Climate Change whose £0.4 billion underspend was 13.9% of their 2011–12 budget. In absolute terms, the largest underspender was the NHS – the biggest Whitehall department – who underspent by £1.7 billion (of which only £0.3 billion was surrendered through Budget Exchange), which is equivalent to 1.6% of its 2011–12 budget.

The use of Budget Exchange varied across departments, even amongst those who ended up underspending on their budgets. For example, the Department for Transport underspent by £0.7 billion (or 5.1% of its budget), of which none was transferred through Budget Exchange, while the Department for Culture, Media and Sport underspent by £0.1 billion (or 5% of its budget), of which around half was transferred to 2012-13 through Budget Exchange.

Figure: Underspends by departments (departments who underspent by more than £0.1 billion)

Figure: Underspends by departments

Note: Chancellor’s Departments includes HM Treasury, National Savings and Investments, Government Actuary’s Department, HM Revenue and Customs, National Investment and Loans Office, Royal Mint and Crown Estate Office.

Small underspends would not be surprising – indeed it would be amazing if every department managed to spend its allocation exactly. But relatively large underspends that are not qualifying for Budget Exchange might seem strange, and particularly so in an era where most departments are facing cuts. One possible explanation is that departments have been trying hard to ensure that they do not end up with an overspend, which might be particularly unfavourably looked on by the Treasury in the current era of austerity. Another is that Whitehall departments have looked ahead to the cuts they face in 2012–13, 2013–14 and 2014–15 and decided that over-delivering on the cuts to date would leave them better placed to keep within these tight budgets going forwards. Departmental spending plans for 2012-13 now currently imply an average real cut of 0.8% in real terms from their 2011-12 level, which would have been a 2.6% average real cut if departments had spent their planned budgets in 2011-12. Whatever the motivation behind the underspends, to the extent that departments are underspending while still maintaining the quality and quantity of public services being provided, this is good news.

IFS public finance observations are generously supported by the Economic and Social Research Council (ESRC).

]]> Fri, 06 Jul 2012 00:00:00 +0000
<![CDATA[Measuring and addressing child poverty]]> Today’s figures show that the previous government’s relative child poverty target for 2010-11 was missed, despite large reductions in that measure of child poverty since 1998-99. And in the latest year of data, relative child poverty fell, but not because low-income households with children got any better off in absolute terms.

The government has re-stated its commitment to the 2020-21 income-based child poverty targets that it inherited (and had voted for). As we argue below, this leaves the government in the position of having a target looming in just eight years, without policies which are likely to transform the distribution of income anywhere near radically enough over that kind of timescale. The government has also re-iterated its view that income-based child poverty targets have important limitations, and that relative poverty measures are problematic. It has announced it will be consulting on defining broader measures in the Autumn.

There are bound to be reasonable disagreements over what child poverty means and how to measure it (as well as how to address it). All sorts of things - income, spending, health, education, access to services - matter to the well-being of the poorest children, and there is obviously scope for debate about the relative importance of each. And the argument over whether poverty should be thought of in “relative” or “absolute” terms has been going on for centuries.

So what are the issues, and what can we say about them?

Absolute versus relative poverty

The fact that relative poverty can fall without absolute living standards increasing (as happened in 2010-11) is not a new discovery: it follows straightforwardly from defining poverty in relative terms. And in the short run we probably do care most about how the poorest children are faring in absolute terms on the chosen measures - are their incomes, educational attainment and so on higher or lower this year than last? Year-to-year movements in relative income poverty which simply reflect volatility in middle incomes - and hence in the relative poverty line - would lead one to perverse conclusions if such measures were the only ones being considered.

But what about the long run? Society’s intuitions about what constitutes a minimum acceptable standard of living do seem to be rooted in time and place to at least some extent. The real income levels that society is willing to tolerate at the bottom of the income distribution are rather higher now than they were in the Victorian era. So we probably do want to think about the poor in the context of the living standards of their contemporaries and the resources available to the society in which they live.

Absolute living standards matter and, over longer periods, so do relative living standards. Each measure provides different information. Each is relevant, and each is clearly incomplete. A single number cannot be expected to capture all of the concerns that someone might have about the distribution of living standards. Reasonable people will disagree about the relative importance of each, and it is probably sensible to keep an eye on both (as the previous and current governments have done, at least in terms of the poverty indicators that they officially track(ed)).

Poverty with respect to what?

Whether absolute or relative, a poverty measure needs to specify the object of concern. The previous government clearly believed that material living standards are at the core of what poverty is about: it chose to track three income-based indicators of child poverty, and included them (plus one more) in its 2010 Child Poverty Act which set ambitious targets for 2020-21.

But income does not perfectly capture even material living standards. For example, it does not account for the availability and quality of public services, or for the ability of people to spend more than their income by drawing on savings or in anticipation of higher income later. Many students have low incomes, but we might want to think about them quite differently from the long term sick and disabled who have little prospect of higher incomes in future.

And one might see poverty as about much more than material living standards. Broader indicators of children’s well-being, such as their physical and/or mental health, matter. In the context of child poverty one could go even further, viewing it as about not only current outcomes but about lack of opportunity or ‘life chances’. The current government seems to have both of these in mind when talking about child poverty, and included health-based indicators and measures of educational participation and outcomes in its Child Poverty Strategy. One can certainly quibble about which things are taken to be constitutive of poverty, rather than causes or symptoms of it, or indeed which measures simply represent other things that we care about that have nothing to do with poverty per se. But there are without doubt a variety of dimensions of well-being that we should care about, each with multiple causes and multiple consequences, and it is sensible to be interested in all of them (whilst thinking carefully about what causes what, and how patterns of deprivation can be broken).

The 2020-21 income-based targets

Whilst supplementing them with a broad range of other indicators - which is sensible - the government is retaining the 2020-21 income-based child poverty targets that it inherited (and which both governing parties voted for in 2010). But the role of those targets is increasingly unclear, given other aspects of government policy.

IFS researchers have shown that the large reductions in income poverty among children under Labour relied very heavily on fiscal redistribution towards low-income households with children. For example, in 2010-11 about £18 billion more was being spent on benefits and tax credits aimed at families with children than if Labour had just applied default indexation rules since coming to power (see here). This is not surprising, given that the government was pursuing quite short run income-based targets, first for 2004-05 and then for 2010-11: tax and benefit changes tend to have rapid and relatively predictable impacts on incomes.

But fiscal redistribution is not costless. There is an inescapable trade-off between increasing redistribution and strengthening financial work incentives. And a pound spent on benefits is a pound not spent on other things which might improve children’s lives more cost-effectively in the long run, such as education, health or social services (or indeed a pound spent on completely different objectives).

The government has made clear that it does not view further large increases in fiscal redistribution as appropriate, at least in the current fiscal climate. Some people will agree with that judgement, and some will not: the extent to which the costs of increased fiscal redistribution are worth incurring is of course a subjective question. But there does seem to be a tension between this judgement and the retention of the income-based 2020-21 child poverty targets. Why? Because the targets loom in just eight years. Over short periods - and anything less than a generation is probably short term when it comes to making radical and politically feasible changes to the income distribution - it is difficult to think of tools beyond fiscal redistribution which would achieve a transformation of the large scale required.

]]> Thu, 14 Jun 2012 00:00:00 +0000
<![CDATA[The road ahead for motoring taxes?]]> The case for a radical overhaul of the way that motoring is taxed is set out in a report, funded by the RAC Foundation, that we are publishing today. We recommend a move to a widespread system of road pricing. The revenues raised could be used to reduce other motoring taxes. Such a move would generate substantial economic efficiency gains from reduced congestion, reduce the tax levied on the majority of miles driven, leave many (particularly rural) motorists better off, and provide a stable long-term footing for motoring taxes without necessarily raising net additional revenue from drivers.

The economic rationale for road pricing is compelling. Road use generates costs which are borne by wider society instead of the motorist. These 'externalities' mean that in the absence of taxation or pricing, there is an inefficiently high level of road use. Taxes can help bring private demands into line with the socially desirable level. Several different externalities are associated with motoring. Some, like carbon emissions from burning petrol and diesel, are easily addressed through fuel duties as the costs depend entirely on fuel use. Others, notably congestion but also the costs of noise and accidents, vary enormously according to where and when someone drives. Driving in rural areas late at night imposes no congestion cost upon other motorists. Driving in conurbations at rush hour generates large congestion costs. Taxes on fuel cannot vary according to time and location, and so are fundamentally unable to account for this variation. Taxes on road use, however, would be able to do so. The potential efficiency gains from better-targeted taxes are large: the 2005 Eddington Review estimated benefits of up to £25 billion per year.

The figure below illustrates the point. Drawing on figures from the Department for Transport, it shows an estimate of the marginal external cost associated with each kilometre of road use from least costly on the left to most costly on the right. They vary enormously from just 0.9p per kilometre to 245p per kilometre (and, if anything, these figures probably understate the variation in the marginal externality). For a car of average efficiency, driving a kilometre leads to a fuel duty payment of around 5.5p irrespective of location and time. The bluntness of fuel duties is clear. Even ignoring other taxes on motoring, perhaps half of kilometres are taxed too much, one-quarter taxed at roughly the appropriate level, and one-quarter taxed too little, often substantially so.

Figure 1. The distribution of the marginal external costs of motoring

Source: Authors' calculations based on Department for Transport (DfT) data.
Notes: The marginal external cost distribution is derived using estimates of the total motoring externality for all major types of road (conurbation, urban and rural) across different congestion bands. These have been weighted to construct the distribution by using estimated 2010 values of the proportion of total distance driven on each combination of road type / congestion band.

There is a second, fiscal, argument for road pricing. Fuel duties and Vehicle Excise Duty (VED) raise around £38 billion per year. But these revenues do not appear to be sustainable in the medium-term. Forecasts from the Office for Budget Responsibility suggest that by 2029/30, revenues from these taxes will be some 0.9% of national income lower than today. This equates to more than £13 billion per year in current terms. This decline is partly down to improved vehicle efficiency and the growth of electric vehicles. In the extreme case in which all vehicles were electric, revenues from fuel taxes and VED would disappear altogether - but the problems of congestion would clearly remain.

How could this revenue be replaced? £13 billion is approximately equivalent to a 3½p increase in the basic rate of income tax, an increase in the main rate of VAT to almost 23%, or a 50% rise in rates of fuel duties. None of these options are particularly palatable, and there is little sense in ever higher fuel duty rates for the shrinking base of motorists relying on conventional fuel. Indeed, the difficulty of sustaining revenues through further duty rises has already been demonstrated by the consistency with which both this government and its predecessor have announced, and then failed to implement, duty increases. From their peak in March 1999, real (inflation adjusted) fuel duty rates were 16% lower by December 2010. Real fuel duty per kilometre driven fell by 13% between 2000 and 2010, from 6.3p to 5.5p.

Road use, however, appears to be a more sustainable tax base. The Committee on Climate Change estimates that total distance driven in the UK will rise by almost one-quarter from 516 billion vehicle kilometres in 2010 to 637 billion by 2030. Revenues from road pricing would not erode as vehicles become more efficient, and the charge - assuming it varied by where and when the driving occurred - would provide the right signals to deal better with variation in congestion costs.

As we have argued elsewhere, there are multiple long term pressures on taxes and spending. Just as we need prepare for, and debate how to pay for, demographic change, so we should be taking the time to debate and design a system of road pricing. The groundwork needs to be laid now. In this context the potential reforms to ownership of the strategic road network recently outlined by the Prime Minister simply enhance the need for a sensible, joined-up strategy. Before signing long-term leases to let private firms manage key roads, the government should consider carefully whether this would compromise the ability to implement national road pricing if it - or a future government - were finally to accept the overwhelming logic for it in the years ahead.

]]> Tue, 15 May 2012 00:00:00 +0000
<![CDATA[Labour market changes unlikely to reduce inequalities]]> Politicians often say they want to see income inequality or poverty fall. The most explicit current example of such an ambition is the set of targets in the 2010 Child Poverty Act. Meeting these would require large increases in the real incomes of low-income households with children, and a very substantial narrowing of the gap between these households and those on middle-incomes, by 2020-21.

Political and economic constraints mean that achieving these targets solely by using benefits and tax credits to redistribute to low-income households looks virtually impossible, particularly in the context of the Government’s plan to eliminate the budget deficit mostly through spending cuts (rather than tax rises). It is natural, therefore, to ask whether the targets could be hit by transforming the distribution of private income over the next eight years. With this in mind, research co-authored by us and published today considers the likely impact of changes in the mix of jobs over the rest of this decade on typical measures of income poverty and income inequality.

The research combined long-run projections of the state of the labour market produced by the Institute for Employment Research at the University of Warwick with a model of household incomes produced by the IFS. Our previous work had already suggested that relative poverty would rise between now and 2020, partly because, over the period as a whole, wages are likely to grow in real terms while income from benefits and tax credits will be pegged to inflation, and partly because of the cuts to welfare spending announced by this government. In our new research, we found that projected changes in the mix of job types (where “job types” are defined by industry, occupation, skill level, gender, region, and full-time or part-time status) over this decade are likely to widen inequalities in household incomes further, and will consequently cause measures of relative poverty to rise further too. This is mostly due to the anticipated continuation of long-running labour market trends: for example, high-wage occupations are expected to continue to replace middle-wage occupations, but with little change in the prevalence of low-wage occupations in a so-called “hollowing out” of the labour market.

We also looked at plausible alternative scenarios for the labour market. The scenarios held the total number of jobs in the economy fixed, but changed the sorts of jobs available, or the people working in these jobs, or the pay differentials between different sorts of jobs. Even scenarios which one might reasonably expect to reduce inequality and relative poverty often have surprisingly little overall impact on the income distribution. One reason for this is that many individuals on low pay do not live in households with a low income, usually because they live with other adults who are also earning, and this means that labour market trends which may help the low paid are not necessarily well-targeted on households with low incomes. For example, reducing the gender pay gap would benefit women in low-income households but would also benefit women in high-income households, with almost no net impact on income inequality or relative poverty. Of course, this is not to suggest that one should not be interested in narrowing the gender pay gap for reasons other than reducing household income inequalities, but it highlights that governments should be clear about which objectives are achieved by which policies. Another reason is that a relatively large fraction of households with low income do not have anyone in work. This constrains the extent to which labour market changes can reduce poverty or inequality, unless the total level of employment can be substantially increased.

Of course, it is impossible to account fully for everything which may affect the distribution of income, and there remain important unanswered questions. For example, the research did not consider the scope for training and skills policies to increase the total level of employment in the economy. We did allow for the direct effects of Universal Credit on household incomes, but did not consider how the associated changes to financial incentives and the requirements imposed upon recipients to look for better-paid work would affect incomes. But with relative child poverty projected to reach almost 26% in 2020-21 in our baseline scenario, there is no way that Universal Credit alone will get the Government back on track to hit its 2020 target of reducing this measure of poverty to just 10%.

This research adds to the evidence that policy-makers who want to reduce income inequalities in the coming years will have to swim against a rising tide of inequality-increasing changes in the mix of job types, and a less redistributive tax and benefit system. The targets for 2020-21 specified in the Child Poverty Act were always ambitious (requiring a lower rate of relative child poverty than the UK has experienced in the past 60 years), and it looks increasingly doubtful that anything could bring about the transformation required in the time available, given constraints such as the fact that many of the parents of 2020-21 have already left education. If the Government disagrees with this assessment, it needs to explain urgently and credibly how such a rapid and spectacular transformation is going to take place. The only other constructive way forward is surely for the Government to set itself different objectives which do reflect its view of what is both achievable and desirable. It is difficult to see the value of extremely ambitious targets without a credible plan for meeting them.

]]> Wed, 09 May 2012 00:00:00 +0000
<![CDATA[How significant is a minimum unit price for alcohol of 40p?]]> The Home Office has today published its alcohol strategy which contains a number of measures designed to curb excess alcohol consumption and the associated social harms. The most high-profile is the plan to introduce a minimum price per unit of alcohol in England and Wales (a similar policy is currently before the Scottish Parliament), with 40p per unit set as an illustrative rate.

As in our previous analyses of minimum pricing, we continue to argue that a preferable approach would be to introduce a floor price for alcohol through the duty system, moving towards a more equal tax treatment of alcohol by type and strength combined with a restriction on selling alcohol below the total tax levied on it. Such a system could be designed to achieve an increase in the price of low cost alcohol similar to that that from a minimum price. But it would have the advantage of raising money for the Exchequer, whereas a minimum price would transfer revenues to the alcohol industry instead. If set at 40p, we estimate that these transfers could be as much as £850 million per year. The alcohol strategy document suggests that the government will "... work with industry to use any additional revenue to provide better value to customers in other areas ..." though quite how this could be enforced or monitored is hard to see.

How significant would a 40p minimum price be? We examine the policy using detailed data recording the off-licence alcohol purchased by more than 19,000 households in 2010. The data do not include alcohol bought in pubs and restaurants. However, it is unlikely that a minimum price would have much direct impact for on-licence prices.

Table 1 shows that the average off-licence alcohol unit sells for around 44.8p. Overall, 47% of units would be directly affected by the proposed minimum price. The policy would therefore have a significant impact for off-licence alcohol retailing, and would not simply affect the very bottom of the price distribution. There are also substantial differences across alcohol types. Over 80% of cider units are directly affected. By contrast, alcopops and sparkling wine would be relatively unaffected.

Table 1: Average retail price per unit and proportion cheaper than 40p, by alcohol type


Average retail price per unit


Source: calculated using data from Kantar Worldpanel, 2010. Note: prices uprated to February 2012 values using the all-items Retail Prices Index from the Office for National Statistics.

Off-licence purchase prices vary by household income and by overall alcohol consumption level. The average price for those with incomes below £10,000 per year is around 42p per unit, compared to 51p for those above £60,000. Households consuming fewer than 7 units of alcohol per adult per week pay almost 49p per unit, compared to 41p for those consuming more than 35 units. The policy would therefore affect low income and heavy drinkers more directly. Moderate consumers, though, are quite substantially affected as well: 37% of units purchased by those drinking fewer than 7 units per week cost less than 40p, compared to 59% for those drinking more than 35 units.

Table 2 estimates the average increase in total food and drink expenditure by income and consumption group following a 40p minimum price. It assumes that there is no change in consumer or retailer behaviour other than an increase in the price of alcohol below the threshold. The average cost is about 0.6% of grocery budgets, or £21 per year in cash terms. The largest impact is felt by low income, heavy drinkers who lose 5.9% of their grocery budget on average. The measure is slightly regressive overall, though the differences across income groups are not particularly large and it would not seem sensible to judge the policy simply on its distributional effect.

Table 2: Increase in food and drink grocery budget following 40p minimum price, by income group and average weekly off-licence consumption level


Average retail price per unit

Source: calculated using data from Kantar Worldpanel, 2010. Note:

This analysis does not attempt to model the behavioural response to minimum pricing. However, many of the potential effects are hard to predict. In particular, how would retailers and manufacturers change the price of more expensive alcohol? What would happen to the range of products on the market? What would happen to non-alcohol prices? These indirect effects could be very significant. The coalition may benefit from a Scottish laboratory if the policy is implemented by Holyrood first. The impact of minimum pricing in Scotland should therefore be robustly evaluated with the policy in England and Wales being responsive to any lessons learned.

]]> Mon, 26 Mar 2012 00:00:00 +0000
<![CDATA[A £10,000 personal allowance: who would benefit, and would it boost the economy?]]> Despite the planned net fiscal tightening of £123bn per year by 2016-17, the coalition government has sought to cut income tax for low-income individuals by increasing the income tax personal allowance to £10,000 by 2015-16. Significant increases to this allowance - the amount of income that is not subject to income tax - have already been implemented or announced, at a combined cost to the government of nearly £5 billion per year: it increased by £1,000 in 2011-12 to £7,475, and it will increase by a further £630 in 2012-13 to £8,105 (£700 and £210 more than it would have gone up by under default indexation rules in those years respectively).

How much more would it cost the government to meet its target of reaching a £10,000 personal allowance by 2015-16? Normal price indexation of the personal allowance (the assumption underlying the Office for Budget Responsibility's forecasts for the public finances) would mean that it would reach £8,885 in 2015-16 without any discretionary policy changes given current inflation forecasts. Compared to this baseline of £8,885, a personal allowance of £10,000 in 2015-16 would mean that tax revenues were £5.3 billion lower in that year. To reach that ambition earlier, as several Liberal Democrat and Conservative politicians have recently advocated, would reduce income tax revenues by more than this, as shown in the table below:

Year £10,000 personal allowance introduced   Annual cost to the Exchequer
2012-13   £8.9bn
2013-14   £7.7bn
2014-15   £6.5bn
2015-16   £5.3bn


Note: Costs are for the year in question: if the personal allowance were subsequently increased in line with inflation (the default indexation policy), the cost would remain roughly the same. Alternatively, if the personal allowance were frozen at £10,000, the cost would fall as shown in the table.

Source: Author's calculations using the IFS tax and benefit microsimulation model, TAXBEN, run on uprated data from the 2009-10 Family Resources Survey.

These figures all assume that the threshold at which the higher 40p rate of income tax starts to be applied is unaffected. This means that higher-rate taxpayers get the same benefit from the higher personal allowance as those who remain basic-rate taxpayers. If the government wanted to lower the cost of achieving a £10,000 personal allowance, it could prevent higher-rate taxpayers from benefiting at all by reducing the threshold at which the 40p rate starts to be applied: this would reduce the cost to the Exchequer from £5.3 billion to £3.3 billion in 2015-16. This approach would lead to a significant increase, of around 600,000 in the number of higher-rate taxpayers. Under current policy proposals this would also mean that about 200,000 more families with children would lose their Child Benefit because an adult in that family would become a higher rate taxpayer after the lowering of the higher rate threshold (these Child Benefit savings would account for about £300 million of the reduced cost to government of meeting its £10,000 target in this way).

The gain from increasing the personal allowance to £10,000 in 2012-13 (without changing the higher rate income tax threshold) would be £379 a year to each individual taxpayer aged under 65 with an annual income between £10,000 and £116,210. There are three groups of individuals who would not benefit: those aged 65 or over, who already have tax allowances exceeding £10,000 (other than those with incomes above around £29,000 who see their allowance reduced to the level of the allowance for those aged under 65); those who do not have incomes high enough to pay income tax anyway (more than a third of the adult population); and those who have the personal allowance fully withdrawn as their income exceeds £120,000. Those with incomes between £8,105 and £10,000 would see their income tax liabilities fall from less than £379 to zero. A gain of £379 is of course larger as a percentage of income to a low-income individual than someone with a higher income, although it is important to remember that the poorest third of adults do not benefit at all because their incomes are already below the personal allowance.

But if we examine the distributional impact at the family level (which is normal for this sort of analysis, since we would expect at least some degree of income sharing within families) we get a different pattern to the one we might expect. This arises because two-earner couples, who tend to have higher family incomes, can benefit twice over from the increase in the personal allowance because both members of the couple would see their income tax liabilities fall by £379, meaning that they would gain £758 in total. Thus, the highest average cash gain occurs in the second-richest tenth of the income distribution (some of the richest tenth would not benefit because of the withdrawal of the personal allowance above £100,000, lowering the average gain for this group). As a percentage of income, the gain is roughly the same from just below the middle to just below the top of the income distribution, with the bottom and the very top gaining by less than this.

To summarise, the common assertion that increasing the personal allowance is progressive is true if one considers the gains across individual income taxpayers. It is not true if one considers the gains across all families as relatively few of the poorest families contain a taxpayer and two-earner couples gain twice as much in cash terms as one-earner families.

Distributional impact of increasing personal allowance to £10,000 in 2012-13, by income decile group

Income deciles

Notes: Income decile groups are derived by dividing all families into 10 equally-sized groups according to income adjusted for household size using the McClements equivalence scale. Decile group 1 contains the poorest tenth of the population, decile group 2 the second poorest, and so on up to decile group 10, which contains the richest tenth.

Source: Author's calculations using the IFS tax and benefit microsimulation model, TAXBEN, run on uprated data from the 2009-10 Family Resources Survey.

Some have argued that a higher income tax allowance would be a good way of introducing a short-term fiscal stimulus for the economy. In our annual Green Budget, we argued that the case for a fiscal stimulus was not clear cut, but said that if the government did choose to introduce one, it should be timely, targeted and temporary. Increasing the personal allowance does not seem to meet any of these criteria. An increase in the allowance would not be especially timely: individuals would not see the effect in their pay packets until the Autumn. It would not be well targeted: the Office for Budget Responsibility suggests that increased investment spending or spending on benefits (or indeed cuts to the main rate of VAT) would deliver a larger direct boost to the economy in the near-term. And being a long term government ambition it would not, of course, be temporary.

Increasing the income tax allowance takes low income people out of income tax. Therefore, it is the best way of focusing income tax cuts on those with lower incomes. And it will strengthen work incentives, especially for low earners. But it is important not to claim too much for a policy which, especially in the current fiscal climate, is expensive. By definition it will not help those on the lowest incomes, who do not pay income tax anyway. And in the current context it is clearly not the best way of delivering a short term fiscal stimulus - and it should not be pursued for that reason. Any stimulus needs to be timely, targeted and temporary.

]]> Fri, 09 Mar 2012 00:00:00 +0000
<![CDATA[How can policymakers raise household saving?]]> Concern that too little saving is done by a significant number of UK households has long been a motivation for government policy. This is particularly the case for retirement saving: the Pensions Commission estimated in 2004 that perhaps nine million people were under-saving for their retirement. The issue is potentially wider than retirement saving alone: recent IFS research found that the median family had little more than a thousand pounds in liquid financial wealth in 2005. However, any intervention by policymakers designed to increase household saving rates should ideally be based on high quality evidence on the efficacy of different policies.

Today, a new British Academy policy centre report, authored by IFS researchers, is being launched. The report examines the evidence base for four different types of policy to promote household savings:

  • Using financial incentives such as tax-favouring and matching;
  • Financial education, training and information provision;
  • Choice architecture and 'nudges';
  • Social marketing, using techniques from commercial marketing to promote social goals.


The report critically assesses the literature on whether policies in these areas, both in the UK and abroad, have been successful in encouraging households to save more. Many of these policies have been actively pursued in the UK. Financial incentives include generous tax treatment for pension saving, and matching policies such as the previously piloted Saving Gateway. The government has recently consulted on a review of Personal, Social, Health and Economic education in schools which could recommend making financial education a statutory component. Under choice architecture, 'auto-enrolment' into employer provided pensions is set to be rolled out from October. This compels employers to default most workers into a pension scheme from which employees can subsequently choose to opt-out, rather than the current system into which people typically need to opt-in.

The report concludes that, despite the obvious importance attached to the issue in the UK and internationally, significant gaps in the evidence base for whether policymakers can raise overall household saving remain. There are of course some high quality exceptions, but relatively few studies are able to combine credible counterfactuals (what would happen to saving and wealth in the absence of the policy?) with good measures of total savings. For example, there is convincing evidence that financial incentives can raise the amount saved in the favoured savings vehicle. There is much less compelling evidence on whether these incentives increase total savings or just represent assets being shifted from one form of saving to another.

Similar issues arise in assessing 'nudge'-inspired policies like changing default options for pensions saving. There is plenty of evidence that auto-enrolment raises participation in pension schemes (often substantially) but much less sense of the extent to which it raises total saving, if at all. Auto-enrolment can lead some people to save less than they otherwise would have done: most people who auto-enrol stick to the default contribution levels and investment funds which are often quite conservative.

Many studies assessing financial education or training look at the impact of policies on financial knowledge or on intended changes to savings behaviour, but there is not much good evidence that these outcomes translate into genuinely higher household saving rates.

Some other key conclusions from the study are:

  • Not much evidence can be found on the long-term impacts of these policies, and whether they help entrench saving habits;
  • Interventions are sometimes packaged together (such as combining matching with financial education). It is hard to disentangle which aspects of the package are effective in isolation, and whether the impact of the policy as a whole exceeds the sum of its parts;
  • Relatively little evidence, except for that around financial incentives, is specific to the UK;
  • The possible use of 'social marketing' to promote savings has been little-researched.

While the evidence base is in general limited, this is not a reason for inaction if there is a genuine need for intervention. Indeed, it presents something of an opportunity for the government. The move to auto-enrolment in the UK, for example, ought to be robustly evaluated, to provide evidence not just on how the UK scheme should be improved but also for other countries thinking of similar reforms. New pro-savings policies should also be introduced in ways which allow their effects, short- and long-term, to be assessed properly, ideally through randomised trials. This requires proper measurement of total saving and wealth, considering possible spillovers to those not directly affected (for example, some studies of workplace financial education show an effect on the saving behaviour of non-treated workers or family members outside work), and perhaps allowing for random variation in the way policies are 'framed' which also appears to affect behaviour.

Of course while trials are important they may not always be appropriate or easy to implement. Thus evidence in this area needs to feed into the continued development of theoretical models of saving behaviour. Together with good household-level data on saving and assets (which has been much improved in the UK by the introduction of the Wealth and Assets Survey in 2006), this will allow researchers to assess the likely impact of policy compared to a modelled counterfactual even when trials are not feasible. Using multiple, complementary approaches in this way will hopefully mean a significant improvement in the evidence base in the future.

]]> Wed, 22 Feb 2012 00:00:00 +0000
<![CDATA[Thoughts on a benefits cap]]> This week, debate over the Government's Welfare Reform Bill has returned to the House of Commons. An element that has grabbed a lot of attention is the proposed benefit cap for working-age households (excluding those claiming Disability Living Allowance or Working Tax Credit), which will be set at £350 per week for childless single people and £500 per week for other households. This is now expected to affect about 67,000 households in Great Britain when implemented in 2013-14, reducing their benefit entitlement by an average of £83 per week and cutting the benefits bill by about £290 million in that year. To put this in context, other planned cuts to welfare spending amount to about £18 billion per year by the end of this parliament, and will affect millions of working-age benefit recipients.

How could households be in receipt of more than £500 per week in benefits? Put simply, they must have either a large number of children or high housing costs (or both). A couple with four children and no private income would be entitled to about £373 per week in Jobseeker's Allowance, Child Benefit and Child Tax Credit. If they paid rent of £127 per week or more (plausible rent levels for those who rent privately or are in social housing in London), a Housing Benefit claim to cover this would result in total benefit income of at least £500 per week. A smaller family could also be affected by the cap if they live in a particularly high-rent area such as London and consequently claim a large amount of Housing Benefit (for an example, a 3-bedroom household who rent privately can claim up to £340 per week in Housing Benefit to cover their rent). The Government's Impact Assessment estimates that 69% of households that will be affected have at least three children, and 54% live in Greater London (where rents are high).

So what will this policy achieve, apart from reducing state benefit payments to about 67,000 households with lots of children and/or high housing costs? The Government has said that it hopes there will be two forms of behavioural response: families may move to cheaper accommodation to reduce their housing costs, and/or take up paid work because their out-of-work benefit entitlement will have been reduced. A third possible form of behavioural response is in fertility rates, since the cap will effectively reduce state financial support for some large families (see here for previous IFS research on fertility and financial incentives). If this were the main intended impact, though, one would expect to see the policy affecting only new claimants of child-contingent benefits. A fourth possible behavioural impact is for fewer people to cohabit, since the benefits cap is to apply at the household level, and hence living apart could split benefits across households and mean that neither is subject to a cap. This 'couple penalty' is presumably something the Government would not be keen on, as it has said that it wishes to reduce couple penalties in the tax and benefit system.

Crucially, is a benefits cap the best approach to take to deal with benefit payments that the Government deems excessive? If it thinks that the benefit system is giving some families a level of entitlement that is too high, it must believe that some benefit rates are inappropriately high. The best-targeted response would surely be to change those benefit rates. In this particular case, the logic underlying the Government's belief that no family should receive more than £500 per week in benefits would point towards cutting the amount families receive for having large numbers of children and/or reducing the value of housing costs against which people can claim Housing Benefit.

The apparent simplicity of instead just placing a cap on total benefit receipt might look appealing, and may well be politically expedient. But it seems incoherent for a Government to set a system of benefits which it evidently thinks gives some families excessive entitlements, and to then attempt to 'right this wrong' with a cap. If starting from scratch, this is surely not the approach one should want to take. And very shortly the Government will be starting from scratch - its planned Universal Credit is to replace almost all of the existing system of means-tested benefits and tax credits for those of working age. If it has a view on the maximum reasonable level of benefit entitlement for these people, then it should design Universal Credit (and in particular, the child and housing cost additions within it) to reflect that view. It is not clear what is gained from instead layering a cap on top of a system that is designed to allow higher payments.

The approach of tweaking particular benefit rates, rather than imposing a post hoc cap on total benefit receipt, would also force the Government to think carefully about (and be explicit about) the features of the current benefits system that it considers inappropriate. Apart from improving the quality of its solution to the perceived problem, this may also improve the quality of wider debate about the issue. After all, it would make it crystal clear what precisely the debate is about.



]]> Thu, 02 Feb 2012 00:00:00 +0000
<![CDATA[Reforming VAT: a promising proposal, but much more still to be done]]> On December 6th, the European Commission published a communication outlining proposals for reform of the EU VAT system with the aim of making it simpler, more efficient and more robust. As part of the evidence gathering for this communication, a study on the workings (and failings) of the existing VAT system by a consortium of economic institutes, including the IFS, was commissioned. Here we briefly discuss the plans set out by the European Commission and some of the findings of our study. To summarise, if the Commission's proposals are fully realised, complying with VAT procedures, particularly when trading across borders, would be considerably simpler, and the system should be more robust to fraud: hence, the plans represent a genuine improvement over the existing system. However, although the plan envisages some reduction in the scope of VAT exemptions and reduced rates, here significant problems look set to remain. This reflects the fact that responsibility for the use of reduced rates of VAT and exemptions largely lies with the Member States themselves, rather than the Commission.

One focus of the Commission's proposal is on improving VAT administration, with the hope of reducing compliance costs for firms, and increasing the capacity of revenue authorities to detect and prevent fraud. A central website which would provide information on VAT rules and rates for each EU member state is proposed, as is the provision of both more information about policy changes and improved facilities for consultation prior to policy changes. More radically, the Commission advocates a standardised VAT declaration by 2013, and potentially, further standardisation of VAT procedures and forms, and the adoption of a One Stop Shop approach where many traders would need to only deal with the tax authority of one member state (generally the country in which they are based). Such moves could reduce compliance burdens for businesses engaged in cross-border trade, especially smaller businesses. Our study shows that reductions in the costs of complying with different VAT rules in different countries could boost cross-border trade, GDP and household consumption so such efforts should be welcomed. Plans to encourage the sharing of information and best practise between national revenue authorities and to investigate the potential of a trans-national anti-fraud team also seem sensible.

Another promising proposal is to abandon the long-standing objective of moving towards a VAT based upon taxing cross-border trade at the rate of VAT applicable in the exporting country (the origin principle). Doing this will allow the Commission to focus on improving the operation of the existing EU VAT system largely based on the destination principle: taxation in the importing country at the importing country's VAT rate. Such a system has real benefits. For instance, cross-country differences in rates of VAT should not distort where firms choose to source their inputs from: they pay the domestic rate of VAT whether they buy things from a domestic supplier or a supplier based in another EU country.

Perhaps more controversially, the Commission envisages broadening the VAT base with efforts to reduce the extent of exemptions in the public interest (including in the public sector), and in the provision of passenger transport services. This will initially focus on areas where the distortions to the European Internal Market and competition with non-exempt firms are greatest. The report also recognises that the plethora of reduced rates of VAT increases the complexity of the system for relatively little gain. A review of the current rate structure of VAT is proposed, with the aim of abolishing reduced rates that are harmful to the Internal Market, or that apply to goods for which consumption is discouraged by other EU policies, whilst ensuring similar goods and services face the same VAT rate.

This represents a step in the right direction, but will leave many exemptions and reduced rates in place. However, as stated by the Commission "the Member States are primarily responsible for limiting as far as possible the scope of such [reduced] rates", and they also have some discretion about when to apply exemptions. Implementing a reduction in the use of exemptions and reduced rates is likely to be political difficult: those sectors benefiting from them will lobby against their abolition, and it is unlikely to be popular with national electorates either. But doing so would bring real economic gains.

Exemptions, in particular, are anathema to the whole principle of VAT. Where they exist, there are significant distortions to decisions by firms of whether to self-supply or purchase goods and services from the market, and to competition between exempt and non-exempt firms and firms in different EU countries. There is also an increase in compliance and administrative costs for those firms that have to allocate input tax between exempt and taxed transactions. Through these mechanisms, exemptions reduce productivity and output, impede the Internal Market and reduce the international competitiveness of European industries. They also mean forgoing significant revenues: for instance, it is estimated that the VAT exemption of financial services costs the UK Treasury around £10 billion a year.

The case for most reduced and zero rates of VAT is little better. While VAT rate differentiation can be progressive, other taxes and transfers can target the rich and the poor more directly, achieving more redistribution for a lower cost. Whilst it is true that poorer households typically spend a larger fraction of their budgets on items such as food or domestic energy (subject to reduced rates of 0% and 5%, respectively, in the UK) the rich spend more in absolute cash terms. This means that it is in fact to rich households that most of the cash the government is forgoing goes to, making the case for such reduced rates rather weak. Similarly, the particular features of VAT mean that it is rarely well targeted for encouraging the use of 'socially beneficial' goods and services. Reduced rates of VAT can only encourage purchases by final consumers, when often business use of the goods in question can be equally beneficial (such as for environmental products); and the encouragement provided is proportional to price, when often the benefit from consumption is no greater for more expensive varieties of the good in question.


Furthermore, reduced rates of VAT distort households' spending patterns and thus tend to reduce welfare. As part of our study for the EU Commission we examined the welfare consequences of abolishing zero and reduced rates of VAT. We found that because of less distortion to household's spending decisions, removing all zero and reduced rates and using the revenue to reduce the main rate would lead to welfare gains amongst the winners that exceed the welfare losses of the losers to the tune of £1.1 billion in the UK. This is a clear measure of the loss of economic efficiency resulting from the current system. Whilst such a reform would, on its own, be regressive, changes to direct taxes, tax credits, and benefits could be used to address this concern, redistributing the gains from winning households to the poorer losing households. We have previously shown how this could be done.

Hence, whilst the European Commission's proposals are a good step towards improving the functioning and structure of VAT in the EU, it would be a shame if they were seen as all that was required to improve the system. As also emphasised in the recent Mirrlees Review, a more fundamental broadening of VAT to cover more goods and services at the standard rate could bring real economic gains, not least a significant reduction in complexity, and on average need not leave poorer households worse off.

]]> Mon, 19 Dec 2011 00:00:00 +0000
<![CDATA[Inconsistent and inefficient UK carbon prices]]> The United Nations Climate Change Conference will take place in Durban, South Africa, between 28th November and 9th December 2011. Representatives from governments and international organisations will meet to try and agree on ways to reduce global greenhouse gas emissions.

Emissions reduction is like a form of investment, with upfront costs for an expected long-run benefit. Agreeing on who should bear the costs, in what form, and by when, is one of the most complex issues for global politics today. One principle, though, is clear: minimising the total cost of achieving lower emissions requires that it is done in the most efficient way possible. Failing to do this risks resulting in greater costs than need be or even a consequent loss of confidence in carrying through policies to combat climate change. As we describe the UK's current policies leave plenty of scope for improvement.

The damage caused by an individual's emissions affects everyone through climate change effects, but this 'externality' is not taken into account in the private decision of how much to emit. One way for policy to address this is to set an appropriate carbon price and so shift the cost back to the polluter.

A single, consistent carbon price across different sources of carbon emissions (and ideally across countries as well), is a necessary condition for minimising the cost of emissions reduction. To see why, suppose there were a carbon price of £20 per tonne applied to all emissions. This would give firms and households an incentive to carry out measures to reduce their emissions that cost less than this for each tonne of carbon saved. Suppose instead that firms faced a price of £30 while households faced a price of £10. Firms would now engage in more costly measures to reduce emissions, and households would do less. The total cost of reducing emissions by the same amount as before would rise.

Reducing the cost of carbon for households relative to firms might be done with distributional concerns in mind. But it is not at all clear that distributional objectives are best met through variable carbon prices rather than other parts of the tax and benefit system. Furthermore, higher carbon prices for firms may well ultimately fall on households through higher prices or lower wages, but in ways which are much less transparent (and thus harder to compensate for) than direct carbon charges on the household sector.

Existing policies impose both explicit and implicit prices on carbon emissions, which vary greatly across energy sources and energy users. In the UK all electricity produced by burning fossil fuels is subject to the EU Emissions Trading System. Layered on top of this, businesses face the Climate Change Levy (CCL) and the Carbon Reduction Commitment (CRC). The CCL charges firms on the basis of their use of gas or non-renewable electricity. Companies in some industries also get discounts of up to 80% in exchange for making a Climate Change Agreement. The CRC is a trading scheme, which requires firms that exceed a particular threshold of electricity use to purchase allowances for each unit. It has some interaction with the EU ETS, and this layering of trading schemes creates particular complexities. Households are not subject to the CCL or CRC. Indeed, household use of gas for cooking and central heating attracts no carbon tax at all. The figure shows the total carbon prices faced by households and businesses across a number of emissions sources in 2011-12 and 2013-14.


There are some things we exclude from the chart below. First, domestic energy is also subject to the reduced 5% VAT rate, so it is effectively subsidised by the tax system. Second, there are additional effects on energy bills of various cross-subsidies for efficiency measures and green energy production, such as the Carbon Emissions Reduction Target and the Feed-In Tariff.

It is not really possible to determine what the effective carbon price is for vehicle fuels. If we assume that all the tax levied on these fuels were there to capture the cost of carbon, then the effective carbon price would be a massive £251.74 per tonne for petrol, and £219.43 for diesel. In reality, most of the social costs imposed by driving take the form of congestion costs, so we should think of much of the tax as capturing these costs rather than emissions.


Implicit Carbon Prices

Carbon prices

Note: Only the CCL, Renewables Obligation, CRC, Carbon Price Support Rate, and EU Emissions Trading Scheme are included here. The business rate assumes participation in the CRC.

As can be seen from the figure, taking the current set of policies, for which we can straightforwardly estimate the carbon price, it is clear that households are charged much lower carbon prices than firms, and that gas used for heating is charged a lower price than electricity.

In addition, all users face lower implicit carbon prices on coal-fired electricity relative to gas-fired electricity. Businesses pay £22.20 more, and households £6.90 more, for a tonne of carbon emitted from gas-fired rather than coal-fired electricity. This is because neither the CCL nor the Renewables Obligation (RO) discriminate between non-renewable fuels on the basis of their carbon content, raising the implicit carbon price for the relatively less polluting gas compared to coal.

This inconsistency in carbon prices will continue in the coming years, with the gas-coal differential for business electricity rising slightly to £22.70 in 2013-14. The introduction of the Carbon Price Support Rate (due to be implemented in April 2013) will help ameliorate the dispersion of carbon prices because it depends explicitly on carbon content and applies to domestic and non-domestic sectors alike. But this effect is small, and is offset by the expansion of the Renewables Obligation, which does not depend on the carbon content of non-renewable fuels.

The current system is not the uniform carbon price necessary for emissions reduction to be done in the least costly way, but rather a complicated tangle of policies. Other interventions, such as targeted subsidies and product regulations are likely to be needed to achieve emissions reductions. But there is no reason these should substitute for, rather than complement, a uniform, transparent carbon price.

]]> Fri, 25 Nov 2011 00:00:00 +0000
<![CDATA[Small beer? assessing the Government's alcohol policies]]> The March 2011 Budget raised the duty on strong beers (above 7.5% ABV) by 25% and cut the duty on weak beers (2.8% or less) by half. The Home Office has also announced plans to ban 'below-cost' alcohol sales in England and Wales, with 'cost' defined as the total tax (duty and VAT) due on each purchase, though precisely when the ban is to be implemented is unclear. In Scotland, the SNP Government has introduced a Bill to implement a minimum price per alcohol unit from 2012, following an unsuccessful attempt to do so last year. A new Briefing Note published today looks at these three policies in depth, using detailed data recording the off-licence alcohol purchases of more than 25,000 British households.

The changes to beer duty came into force in October. For the products affected, the effect is significant: the total tax (including VAT) due on a 4 x 440ml purchase of 2% ABV lager fell from 78p to 39p, whilst at 8% ABV the tax rose from £3.33 to £4.17. However, strong beers made up just 0.8% of all off-licence alcohol units purchased in 2010 and weak beers just 0.2%, so the policy has a big effect on a very small part of the market. The biggest impact is likely to be on households who consume large amounts of alcohol (more than 35 units per adult per week), who buy 23% of all off-licence units but 53% of strong beer units. Notably, the reform did not affect cider. This means that, on a per-unit basis, the duty on a strong 7.6% ABV beer (23.2p) is now more than three times higher than the duty on a 7.6% ABV cider (7.1p). Rather than switching to low-strength beer, the reform might encourage strong beer drinkers to consume strong cider instead.

A ban on 'below-cost' sales introduces a de facto minimum price for alcohol. For example, a litre bottle of vodka at 40% ABV could sell for no less than £12.25 and a 75cl bottle of wine at 12.5% ABV for no less than £2.17. The impact of the ban is likely to be small: only 1% of off-licence units retailed at less than the tax due in 2010 (see Table). Spirits sell below cost more often whereas cider and alcopops virtually never do. We find no evidence that large supermarkets were more likely to sell below cost; rather, discount stores, off-licences and corner shops were more likely to do so. Nor is there strong evidence that households who drink more bought below cost products more often.

Average off-licence prices per alcohol unit and proportion of units below 45p and 'cost', 2010


Average off-licence prices

Note: prices in December 2010 values. Source: Calculated from Kantar Worldpanel data.


These two reforms would have only marginal effects on the market for alcohol. By contrast, a minimum price could be a much bigger policy. Introduced across Britain at 45p/unit - the level proposed in Scotland last year - minimum pricing would directly affect 71% of off-licence alcohol units (see Table), and nearly 9 out of 10 households who buy off-licence alcohol. Unlike the below-cost ban, a minimum price would apply to all alcohols at the same rate. Lager and cider (which have low per-unit prices) would be particularly heavily impacted, whereas alcopops would not.

The effects would be largest for those who drink the most: about four-fifths of units bought by those consuming 30 or more units per week cost less than 45p. This lends some weight to the idea that minimum pricing targets heavier drinkers, but even amongst those consuming more moderately (6 to 10 units per week), two-thirds of units purchased would be affected.

A previous Observation and Briefing Note discussed some of the economic issues around minimum pricing. What would be most useful for the debate around minimum pricing would be convincing evidence not just of the direct effects, but the indirect impacts as well - how would retailers change the price of alcohol currently selling above the minimum price, or the price of other products? How would manufacturers change the range of alcohol products available once cheap goods are no longer competitive on price? As yet we do not have clear answers to these questions, but these secondary effects are extremely important in assessing the overall impact of minimum pricing.

In the absence of any behavioural responses or wider price changes, our estimates suggest that a 45p/unit minimum price would transfer £1.4 billion a year from drinkers to alcohol producers and retailers. If the intention is to raise the overall price of alcohol, it would seem preferable that these revenues flow to the Government instead. Annual real-terms increases in alcohol duties to 2014/15 are planned, but more fundamental change to the structure of alcohol taxes should be pursued as well. At the moment, different types and strengths of alcohol are taxed at very different effective rates per unit (see Chart), without a clear rationale. Whilst the tax system could not easily replicate a minimum price directly, a sensible starting point would be to tax all alcohol equally on the basis of alcohol content, perhaps in conjunction with a 'below-cost' ban, which would then introduce a floor price directly through the tax system. For example, taxing all alcohol at 25p/unit (the current rate for spirits) would increase the duty rate on typical beer and wine products by around 30% (and more than double the duty on typical ciders), and raise around £1.5 billion per year. Such a reform would require a change in EU Directives which restrict the basis on which alcohol is taxed, and the Government should seek to make this happen.

Effective rate of duty per alcohol unit by alcohol type, as at October 2011

Duty per alcohol unit by alcohol type

Source: Calculated from HMRC figures. Notes: Cider and perry rates are for 'still' ciders; different rates apply to 'sparkling' ciders which are a very small part of the cider market. Ciders above 8.5% ABV are treated as wine or spirits. Wine above 22% treated as spirits. Spirits duty rates apply to alcopops.

]]> Thu, 24 Nov 2011 00:00:00 +0000
<![CDATA[Corporate tax revenues]]> In October most large companies (those with previous annual profits of over £1.5m) made the third of four instalments due in 2011-12 for their corporation tax. Oil and gas firms paid one of their three instalments. In March this year the Office for Budget Responsibility (OBR) forecast that corporate tax receipts for the current financial year would increase by 14.2%. Figures released today show that over the first seven months of this financial year corporate tax receipts were running 0.7% below the same period last year. In October receipts actually fell by 6.9%, compared to the same month last year, with the OBR attributing this to financial sector firms being adversely "affected by weaker investment banking activity, writedowns on euro area sovereign debt and provisions for the mis-selling of payment protection insurance" and a steep fall in gas and oil production, partly as a result of increased maintenance and temporary shutdown of fields.

Corporate tax raises significant revenues - around £43 billion, or 8% of total revenue in 2010-11. Corporate tax revenues (in cash terms) peaked in 2007-8, and fell in each of the two following years. Corporate tax receipts from North Sea oil companies fell particularly sharply. This trend was reversed in 2010-11 when corporate tax receipts increased by 18%. The strongest growth came from the financial services and North Sea sectors.


Corporation Tax net receipts, by sector, 2000-01 to 2010-11 (£ billions)


 Corporation Tax net receipts, by sector


Source: HMRC, Table 11.1


The financial services sector has tended to be a large contributor to UK corporate tax receipts. Revenues from this sector, as well as from North Sea companies, are among the more volatile. Going forward there are a number of uncertainties related to amount of revenue the government will raise from the financial sector. There is uncertainty about the future level of financial sector profits and whether they will remain in the UK. There is also some uncertainty over how the sector will be taxed going forward (the government introduced a Bank Levy in January this year and there are calls from, among others, the European Commission to introduce a so called 'Tobin Tax' on financial transactions).In its latest medium term forecast the OBR set out that the share of total receipts from the financial sector (including PAYE, NICs and corporation tax) is not expected to return to the 2007-08 peak by the end of the forecast period in 2015-16 (see Box 4.1)

Who pays tax?

First, it is important to remember that ultimately companies don't pay tax, people do. The actual incidence of corporate tax is borne either by the owners of capital (in the form of lower investment returns), by workers (in the form of lower wages) or by consumers (in the form of higher prices). Of these groups, capital tends to be the most mobile. As a result corporate tax tends to get shifted to workers or consumers - and especially workers - but with a larger distortionary effect than if these groups had been taxed directly. This is especially important in discussions about companies paying their 'fair share'. With this in mind, we discuss how tax payments vary across companies.

As you would expect, the large companies that earn the majority of profit also pay the largest amount of tax. In 2007-08 the largest 1% of companies (by taxable profit) accounted for around 80% of corporate tax receipts. Many small and medium companies pay no tax at all, either because the make no profits or because they have losses that can be carried forward and offset against tax liabilities. We also expect that as profit increases, the proportion of profit paid in tax will increase. This is because allowances, which act to reduce tax liability, become much less important at higher income levels.

Recent work by the Oxford Centre for Business Taxation provides a detailed analysis of the amount of tax UK firms pay as a percentage of their profits. This varies across companies depending on a large number of factors. They find that most large companies, and those within multinational groups, pay on average a higher proportion of their income in corporate tax. However, there are a small group of the very largest companies that appear to pay little or no tax in the UK.

Tackling tax avoiders?

Tax avoidance has attracted increasing attention in recent years, with a number of groups calling for government action to increase the amount of tax we collect from companies.

There are a range of ways in which firms seek to minimise their tax bill. We expect most firms to take full advantage of allowances and provisions in the tax code - no one expects companies voluntarily to pay more tax than is necessary. Some firms may be more aggressive in their tax planning and seek to exploit loopholes or favourable interpretations of uncertainty in tax legislation. While this activity is legal, many would argue that it is not in the spirit of the law. However, these activities are distinct from tax evasion, in which firms illegally manipulate their tax liability.

The precise characterisation of what counts as tax avoidance is subject to much debate and has many grey areas. The OECD defines tax avoidance as 'the arrangement of a taxpayer's affairs that is intended to reduce his tax liability and (...) is usually in contradiction with the intent of the law it purports to follow'. Of course, not all parties will interpret the 'intent of the law' in the same way. These issues and the possible ways in which the UK government could seek to address avoidance are reviewed in a recent Tax Law Review Committee paper.

We do not have precise measures of which firms are avoiding tax or by how much. The tax gap - the difference between the amount of tax that firms 'should' and actually do pay - is not only hard to define conceptually but is extremely difficult to estimate accurately. Clearly, this difficulty arises largely from the need to determine the correct amount of tax a company owes.

HM Revenue and Customs have produced an analysis on the UK tax gap, noting the many impediments to the exercise. The estimated corporation tax gap was £6.9 billion in 2008-09, which represented 14% of the overall gap - i.e. 14% of difference between total theoretical liabilities and all collected tax receipts.

The government are taking a number of steps to reduce tax avoidance, including considering the possibility of introducing a General Anti-Avoidance Rule (GAAR) - a broad set of principle-based rules designed to prevent tax avoidance. A report commissioned by the Government to consider the merits of a GAAR, and published yesterday concluded that "a moderate rule which does not apply to responsible tax planning, and is instead targeted at abusive arrangements, would be beneficial for the UK tax system" and that the focus should be on "abnormal" practices. But a grey area between what is normal and what is abnormal might be difficult to avoid.


]]> Tue, 22 Nov 2011 00:00:00 +0000
<![CDATA[Rate cutting, base broadening: a reduced incentive for investment?]]> As part of a large package of corporate tax reforms, the coalition government announced a series of cuts in the statutory rate, alongside broadening of the tax base. The main rate of corporation tax was reduced from 28% in 2010 to 26% from April this year and will fall to 23% by April 2014. From April 2012, the main rate of capital allowances will fall from 20% to 18%, the special rate from 10% to 8% and the Annual Investment Allowance from £100,000 to £25,000. These changes operate to reduce the proportion of the previous year's expenditure on certain types of capital that can be deducted from revenue to calculate taxable profits.

Ed Miliband's accusation that the coalition's policy will be to the detriment of manufacturing firms appears to be based on the fact that the package of tax changes will be relatively more beneficial to high-profit firms that invest less in plant, machinery and buildings. This assertion is correct. However, the conclusion that this leads to a tax system rigged against investment fails to acknowledge that firms invest in many types of capital, importantly including intangible capital. Some of the relative winners will be those firms that make important long term investments in skills and ideas, which benefit relatively less from current allowances.

In addition, the OBR predicts that the cost of capital will be lower for new investment by non-financial companies. As ever there is a trade-off here. Corporate taxes are known to be distortionary. Rate cutting and base broadening have opposing effects on firm incentives to invest. Overall the package of measures will give the UK a slightly more competitive tax system. In the longer term making the system significantly more competitive may mean raising less revenue from corporation tax.

On the specific issue of the appropriate level of capital allowances an important principle to consider is that of neutrality - the tax system should aim not to distort firms' decisions over how to organise their activities, how much activity to undertake and where that activity is located. To do so creates inefficiencies and is therefore costly. Since we don't know the actual economic rate for depreciation for each asset, there is debate over the relevant rates. Broadening the tax base alongside cuts to the rate is a policy mix that that has been favoured by policy makers across the developed world for the last 30 years. Indeed, this was a trend followed by the last Labour government.

Changing the tax burden

The Treasury estimates that the package of tax rate and allowances cuts will be broadly revenue neutral - the 2014-15 revenue gain from reducing allowances, allowing for some changes in behaviour but not accounting for any change in the level of investment, will almost exactly offset the estimated cost of reducing the main rate. Within this there will be relative winners and losers.

Taken together the cuts to the statutory tax rates and capital allowances will benefit high-profit firms with lower investment in plant, machinery and buildings (excluding those subject to the Bank Levy) relatively more since they gain more from the rate cuts than they lose from the base broadening. The base broadening will have the largest impact on those firms that invest intensively in the types of capital-long-lasting equipment and machinery - that are subject to capital allowances.

However, this does not mean that the policy necessarily favours low investment firms. Some firms in the manufacturing sector, as well as some in capital-intensive service sectors such as transport, will be relative losers. But there are high investment firms that benefit relatively less from current capital allowances. Importantly, this includes firms that invest in intangible assets that are not subject to allowances (even though investments in skills and ideas can depreciate in conceptually the same way as a machine). Intangible assets represent an important input into production for most firms; in the UK knowledge investment overtook fixed capital investment in the mid-1990s and is now about 50% higher. Many firms, including those in high tech manufacturing, that make long term investments in skills and ideas, are likely to be relative winners of the package of reforms. Therefore, while the package of tax changes does represent a redistribution of the tax burden, the outcome cannot be characterised as simply as saying that high investment firms are losers and low investment firms winners.

The crucial point about a lower rate is that it will help attract (and avoid deterring) internationally mobile capital, which is often highly related to investment in intangible capital.

The OBR forecast included in the June 2010 Budget sets out its judgement that overall the cuts in the corporation tax rate will more than offset the reduction in investment allowances such that the 'cost of capital for new investment is lower for all non-financial companies, and the rate of return from the existing capital stock is higher'.

This is wholly consistent with the figures reported in the Commons library briefing note, showing that the reduction in the Annual Investment Allowance (from £100,000 to £25,000) is estimated to affect between 100,000 and 200,000 businesses. The report also highlights that the package is expected to boost investment and that more than 95% of businesses in the UK will be unaffected (because their qualifying capital expenditure will continue to be completely covered by the annual investment allowance).

How competitive is the UK corporate tax system?

The key aim of the government's package of corporate tax measures - which includes reforms to the 'controlled foreign company' (CFC) rules, a cut to the small profits rate and the introduction of a Patent Box - was to "create the most competitive corporate tax regime in the G20".

A recent report by the Oxford Centre for Business Taxation set out that while the UK has a relatively low statutory tax rate - it ranks 7th out of the 19 independent G20 countries (excluding the European Union) - accounting for less generous capital allowances means that the improvement in UK competitiveness is lower for some measures.


  • The UK's effective average tax rate (EATR) - the relevant measure for considering where firms locate discrete investment projects - is just over 26% in 2011, ranking it 9th. This represents a fall from 4th in 2002, and places the UK above the G20 average. The coalition's package of reforms, conditional on other countries not changing their tax systems, would place the UK 5th in the ranking.


  • The effective marginal tax rate (EMTR) - the measure relevant for considering the level of investment - is just under 23% in 2011, giving it a rank of 15th, which will fall to 14th by 2014.


Research suggests that part of the fall in statutory rates that has been seen across countries in recent decades can be attributed to governments lowering tax rates in response to lower rates elsewhere, in an attempt to attract and retain increasingly mobile capital. Over time, in the face of even more mobile capital and potentially greater tax competition, governments should expect to raise less revenue from corporate tax. Indeed, it may be difficult to substantially increase the competitiveness of the UK tax system with revenue neutral tax changes.

Cuts to corporate tax - which raises significant revenue; around £43 billion, or 8% of total revenue in 2010-11 - are often politically unpopular. However, it is worth noting that the ultimate incidence of corporate tax always lies with households and is borne either by the owners of capital (in the form of lower dividends), by workers (in the form of lower wages) or by consumers (in the form of higher prices). Capital tends to be much more mobile than workers or consumers, and so corporate tax tends to get shifted to domestic factors - and specifically labour - but with a higher associated deadweight cost than if those factors had been taxed directly.

]]> Tue, 01 Nov 2011 00:00:00 +0000
<![CDATA[A fat tax in the UK?]]> Yesterday, the Government published a "call to action" which lays out its strategy for tackling obesity in England. The document sets out plans for government to work with the food and drinks industry to achieve a reduction in obesity. One measure for tackling diet related health problems is to introduce some form of 'fat tax'. Although not mentioned in the report, following the introduction of such a policy in Denmark last Monday, the Prime Minister David Cameron stated that introducing a similar tax in the UK "is something we should look at". The Danish tax places a surcharge on foods that contain more that 2.3% saturated fat, including products such as butter, milk, cheese, pizza, meat and processed food. Mr Cameron suggests that introducing a similar measure could be one way of tackling rising health costs and falling life expectancy among some groups of people. Many other countries are considering policies aimed at improving diet through increasing the prices of certain products (for example, Hungary recently imposed a tax on foods with high fat, salt and sugar content, various US States have adopted or contemplated soft drink taxes and the Scottish Government intends to introduce a minimum price for alcohol).

The rationale behind levying a tax on foods that are rich in saturated fat is that, by increasing their price, the tax will encourage people to eat less of these foods. The hope is that this will help curb consumption of saturated fat which, according to the Food Standards Agency, is on average 20% above the recommended level. Excessive saturated fat consumption is a risk factor for heart disease, so lowering consumption should lead to a reduction in the incidence of heart disease, all else equal.

IFS has recently received long-term funding from the European Research Council and the Economic and Social Research Council to investigate the impact of these types of policies.

It is important to be clear about what, precisely, the objective of such a policy is - what role is there for government in influencing what individuals decide to eat? Are people not in a better position than government to trade off their likes and dislikes and to choose to eat the products they favour most? The two main reasons why government intervention might be justifiable are:

  1. Some people may not be fully informed about the amount of saturated fat they can eat as part of a healthy diet and which products contain large quantities of fat. Providing information may seem like a natural response, but it may be that some people are difficult to reach or are not capable of processing this information accurately. Young children are the most obvious example. Raising the price of foods rich in saturated fat may help reduce consumption levels in a similar way to conveying information.
  2. While most of the costs of ill health arising from a poor diet are borne by the individual themselves, some costs are incurred by other members of society, for example, through increased health care costs and increased sick leave. When individuals decide what to eat they have no incentive to take these external costs into account; this results in a socially excessive level of consumption. By increasing the price of consuming fatty foods, a tax is one way of tackling this over-consumption.

Whether a 'fat tax' is the best policy response to high levels of diet related illness will depend on what of these failures in the market the government is targeting. Griffith and O'Connell (2010) discuss these failures and how well suited various policy options are to addressing them (see Griffith and O'Connell Public Policy towards Food Consumption, Fiscal Studies, vol. 31 no. 4 pp. 481-507) .

Before deciding that introducing a 'fat tax' is a good way to tackle poor diet, the government must also be clear about what element of diet it seeks to influence: is the aim to reduce consumption of all sources of saturated fat, only certain sources (for example, confectionary but not milk), or a set of nutrients often excessively consumed (like salt and sugar, in addition to saturated fat). This will influence the choice over the appropriate set of products to tax.

A second question is what impact any particular policy reform would have. This is not a simple question to answer. It depends on the specific structure of the tax, the market environment, and most importantly on how sensitive consumers are to changes in price. Both the structure of the tax (for instance, whether it is specific or ad valorem) and the nature of competition in the market will determine how taxes are fed through to prices. Griffith, Nesheim and O'Connell (2010) explore these issues by estimating the impact of a fat tax levied on butter and margarine and show that these considerations have a very substantial effect on how a tax would impact consumers, firms and government tax revenue (see Griffith, Nesheim and O'Connell Sin taxes in differentiated product oligopoly: an application to the butter and margarine market, CEMMAP Working Paper CWP37/10). Although the relationship between tax and price changes is a complicated and partially understood one, there are some things the government can do to maximise the potential rise in prices, given the magnitude of the tax; for instance, Griffith, Nesheim and O'Connell (2010) provide empirical evidence that a specific tax is likely to result in higher prices than a comparable ad valorem tax designed to raise the same amount of revenue.

The response of consumers to any price changes will be crucial. The more sensitive a consumer is to price the more effective the tax will be at reducing consumption of the taxed goods. But whether the tax is successful in improving diet will also depend on what individuals would choose to consume instead. A tax levied on butter, but not margarine, may be effective in curbing saturated fat consumption. But since margarines tend to contain more salt than butter, the policy may have the perverse consequence of exacerbating other health problems.

A final important point concerns whether a 'fat tax' would be regressive or progressive. As lower income households tend to spend a higher fraction of their total budgets on food, a disproportionate proportion of any such tax is likely to fall on their shoulders. But the point of such a policy is to encourage people to change their behaviour. Often lower income households are the most price sensitive, which would mean they are likely to change their behaviour most as a consequence of a price rise. To assess the overall effect of the tax and whether low or high income consumers would be affected more, we would need to set the costs imposed by the tax on consumers through higher prices against the potential health benefits arising through their changed behaviour.

Governments around the world are considering how policy can address the increase in diet related disease. One policy response that is increasingly being proposed and adopted is to introduce a 'fat tax'. For this to be an effective response to high and growing levels of diet related health disease, a number of circumstances must be met. Firstly, the government must be clear about why intervention is justified and how exactly it wants to improve diet. Secondly, to be effective, any tax must feed through into higher prices. Although the relationship between prices and tax is complicated, there is some evidence that by structuring the tax carefully, the government would be able to influence how much prices increase by. Thirdly, the price increases must bring about the desired changes in consumer behaviour. There is substantial debate about the role for government in fighting diet related health disease, and how government can best help. There is less evidence about what the impact of introducing policies like fat taxes would be: how exactly would the tax feed through to prices and how would consumers respond? Researchers at IFS are working on improving the evidence base to help policymakers understand the likely impact of various specific reforms.

]]> Fri, 14 Oct 2011 00:00:00 +0000
<![CDATA[Can the UK learn from developing countries? The case for proper policy evaluation]]> On Monday the respected economist Professor Esther Duflo delivered the IFS annual lecture discussing approaches to helping the world's poor (see slides). Professor Duflo drew on several, novel examples of randomised controlled trials (RCTs) - a method which she has put to good use in much of her own research. Randomised trials have done much to improve social policies in the developing world, and a key lesson is that we should make greater use of this important tool for evidence-based policymaking in the UK.

In an RCT, a policy is evaluated by randomly assigning it to some individuals or regions (a "treatment" group) but not to others (a "control" group). The effectiveness of the policy can then be assessed by looking at differences in outcomes between these treatment and control groups. Common in medical research, RCTs are increasingly favoured by economists as well because, by removing contaminating factors correlated with both take-up of a policy and the outcomes of interest, they allow us to estimate a policy's direct, causal impact on outcomes of interest more easily. There are of course some limitations to the use of RCTs, and they will not be appropriate on all occasions, but they can often be extremely useful as a means to assess objectively how well particular programmes are working.

Because of this, RCTs are playing an increasingly important role in the evaluation of policy and are now especially common in developing countries where Professor Duflo's work has been particularly influential. IFS's own Centre for the Evaluation of Development Policy (EDePo) has been involved in evaluating the results of RCTs from India to Colombia in areas such as microfinance, reproductive health, early childhood development and education. The evidence gathered in such trials has had an important influence on policy for two reasons. First, estimates of program impacts estimated via a well designed and implemented RCT are clean and easy to communicate. Second, the availability of a RCT gives researchers a better understanding of individual responses to different initiatives - allowing them to estimate richer and more credible behavioural models which can be used to give valuable insights into possible reforms. In other words, such trials need not just give a thumbs up or thumbs down for a particular policy but can also tell us how they can be improved. For example, data from the RCT of Mexico's PROGRESA (a programme that among other things paid grants to families who sent their children to school), allowed IFS researchers to simulate responses to various changes in the grants. This led to recommendations to refocus the grants toward older pupils who were found to be more sensitive to the financial incentive - a reform which is now itself being piloted in two northern cities of Mexico.

By comparison, RCTs have been quite rare in the UK. While there are many examples of small scale trials, there are only a few examples of full policy evaluations of the kind seen in other countries (such as the Employment Retention and Advancement demonstration, and the Skills Conditionality Pilot) and several examples of policies that could have been trialled (such as the "synthetic phonics programme", and the Work Programme), which have instead been rolled out nationwide without any prior robust evaluation.

Are there perhaps good reasons for the lack of RCTs in the UK? One objection to their use might be their cost. However, the costs of conducting trials need to be set against the costs of implementing large and costly programmes before we have a good understanding of their effectiveness, and therefore a lack of evidence of whether the programme could be improved, should be expanded or whether it should be scrapped. Indeed, the fact that external donors to developing countries often demand RCTs of the policies they are funding, suggests that they can help to ensure value for money. A second objection is that it is unfair to "make guinea pigs" out of certain regions or people. Arguably, however, rolling out untested policies nationally makes guinea pigs of us all! There have also been occasions where policies have been piloted non-randomly in certain areas before being extended nationwide - and for which full randomisation would have been only a small, beneficial, and one would have thought relatively uncontroversial, extra step.

Evidence from RCTs now plays an increasing role in formulating policies in developing countries. The UK could gain by following their example. As Esther Duflo and her co-author Abhijit Banerjee write in their book Poor Economics we should "accept the possibility of error and subject every idea, including the most apparently commonsensical ones, to rigorous empirical testing".


]]> Wed, 28 Sep 2011 00:00:00 +0000
<![CDATA[How does the UK's planned fiscal consolidation compare?]]> Yesterday the International Monetary Fund (IMF) published its latest Fiscal Monitor, which surveys public finance developments across the world and updates their fiscal projections. The UK launch of the report was held at the IFS in London on 5 October (here). The data in the report can be used to provide a useful comparison of the effects of the financial crisis on the public finances in the UK with the effects felt abroad, and some indication of how the coalition government's planned fiscal consolidation currently stacks up against those planned in other countries.

Back in 2007, out of the 28 advanced economies for which the IMF compiles data, the UK is judged to have had the fourth highest level of cyclically-adjusted borrowing (that is, underlying borrowing, after adjusting to take account of the estimated impact of the ups and downs of the economic cycle). In other words, the UK entered the financial crisis and recession in a relatively poor position, with a level of cyclically-adjusted borrowing that is thought by the IMF only to be exceeded by that seen in Greece, Ireland and Portugal. As we have previously shown this is because while public sector borrowing was cut by Labour over the period from coming to office in 1997 and the start of the financial crisis in 2007 it was not reduced by as much as in the vast majority of other major industrialised countries over this period.

The crisis also affected the UK economy and public finances relatively badly. The UK experienced the 9th largest deterioration in its fiscal position during the crisis, resulting in us reaching the fifth highest level of cyclically-adjusted borrowing across the 28 countries (with the UK's peak level of cyclically-adjusted borrowing reached in 2009).

Table: UK's level of and change in borrowing between 2007 and 2015, compared to 27 other advanced economies



 UK rank Notes

Cyclically adjusted borrowing  


2007 (pre-crisis)4th highestGreece, Ireland and Portugal higher

Peak year5th highestGreece, Iceland, Ireland, Spain higher

201516th highest 


Increase, 2007 to peak9th largest 

Reduction, peak to 20155th largestGreece, Iceland, Ireland and Portugal larger

Reduction, 2007 to 20154th largestGreece, Ireland and Portugal larger


The 28 advanced economies on which data comparable to the UK are available are: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hong Kong, Iceland, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Singapore, Slovakia, Slovenia, South Korea, Spain, Sweden, Switzerland and the United States.


The coalition government has set out a fiscal consolidation plan which is projected by the IMF to see cyclically-adjusted borrowing reduced year-on-year until 2015. This is currently forecast to be the 5th largest reduction in cyclically-adjusted borrowing among the advanced countries considered by the IMF - only Greece, Iceland, Ireland and Portugal are currently on course to see a larger reduction.

The Figure below compares the size, timing and ultimate effect of the fiscal consolidation planned by the UK's coalition government between 2010 and 2015 to the fiscal plans of various other countries who are judged to have had similar levels of cyclically-adjusted borrowing to the UK in 2010. These can be broadly divided into two groups: those with a similar profile of fiscal tightening to the UK (shown in panel a) who are forecast to end up in a much stronger fiscal position by 2015 than they were in 2010, and those with a much slower profile for tightening (if any) than the UK (shown in panel b), who on current plans are not expected to reduce their borrowing position as substantially by 2015.

Panel (a) shows that Greece and Ireland had a slightly worse borrowing position than the UK in 2010, but are forecast to be on course to cut borrowing similarly rapidly through to 2014. Portugal meanwhile also had a similar level of borrowing in 2010 but is forecast to be on course to complete its planned tightening in 2013. The UK is currently unique amongst these countries in that the coalition's planned tightening already continues through to 2015, and therefore the UK is currently forecast to end in a better borrowing position than both Greece and Ireland. All four of these countries (including the UK) are forecast to be in a much better borrowing position by 2015 than they were in 2010. The other country that is tightening faster than the UK between 2010 and 2015 is Iceland, but the profile for the consolidation over this period only tells part of the story for Iceland; their cyclically-adjusted borrowing has already been brought down by nearly 10% of national income between 2008 and 2010.

Panel (b) shows that Spain, Japan and the USA all have slightly better cyclically-adjusted borrowing positions than the UK in 2010. The USA and Spain are forecast to experience a similar magnitude of fiscal tightening to the UK between 2010 and 2013 but with little further tightening planned thereafter. This would leave them with a relatively weak cyclically-adjusted borrowing position in 2015, forecast at over 4% of national income, suggesting that further measures to reduce borrowing are more likely to be needed in those countries. Japan, whose fiscal position has been affected not only by the global recession but also by a series of natural disasters, is forecast by the IMF to have continued loose fiscal policy over each of the next five years and cyclically-adjusted borrowing still in excess of 7% of national income in 2015.

Figure 1: How do the fiscal consolidations planned in other industrialised countries compare to the UK's plans?

Graph A: cyclically adjusted borrowing

Graph B: cyclically adjusted borrowing

There are of course many other advanced countries, which have had lower levels of cyclically-adjusted borrowing over the last few years than the UK. Generally speaking these countries are currently planning only small fiscal consolidations over this period, if they are planning reductions at all.

This comparison of the coalition government's fiscal consolidation plans with those in other countries is a picture that may change over time. Some other countries - particularly those highlighted in panel (b) - may eventually extend their plans for tightening. But there is no escaping the fact that the UK was affected relatively badly by the crisis and the government's plans will need to be harsher than those of most other advanced economies if we do not want the UK to remain with relatively high levels of public sector borrowing, which is where it stood before the financial crisis and where it currently stands.

The UK launch of the latest IMF Fiscal Monitor was held at the IFS on 5 October - click here for more details.


]]> Wed, 21 Sep 2011 00:00:00 +0000
<![CDATA[China is investing rapidly in skills and science: UK should do the same]]> China has experienced unprecedented investment in skills and science, which has resulted in rapid growth in innovative outputs. New evidence suggests that Chinese inventors have the capacity to engage in research at the technology frontier. Such trends have fuelled widespread concerns over Western economies' ability to maintain their dominance in knowledge creation and high skill employment. However, innovation is not a zero-sum game; the success of China need not be at the expense of the West. The key challenge for the UK and other knowledge economies is to invest now in order to foster a highly skilled workforce that is able to both compete for and engage collaboratively in tomorrow's breakthroughs and that is flexible enough to adjust to changing conditions.

Recent years have seen increasing attention on China's technological performance, which is unsurprising given the barrage of statistics showing that innovative activities in China are growing at an astounding rate. Over the last decade, there has been rapid growth in investment in Research and Development (R&D) such that the proportion of Chinese national income invested in R&D (1.1%) is now comparable to that in the UK (1.15%). At the same time, large investments in education have produced a proliferation of Chinese graduates, almost half of which study for science and engineering degrees.

This investment in research capacity has been translated into equally impressive growth in innovative outputs. For example, in 2010 China was the fourth largest filer of patent applications to the World Intellectual Property Organization. Under a simple extrapolation of current trends, China could overtake the US to be the world's largest filer of patent applications by 2015. Contrary to much of the academic literature to date but in line with public perception, the proportion of Chinese innovation that stems from fundamental research is at least as high as in the West. That is, Chinese inventors display the capacity to innovate alongside US and European inventors at the technology frontier. An important part of these trends have been driven by the investment of Western multinationals who account for 25-30% of the R&D investment in China and who increasingly create new technologies using Chinese inventors.

Given these trends, concerns over Western economies' ability to maintain their dominance in knowledge creation and high skill employment are not surprising - innovation has been the engine of economic growth and lies at the heart of increased living standards. Nor are they wholly unfounded. Recent work prepared for the European Commission concluded that if the recent trends in R&D continue then "in 2025, the United States and Europe will have lost their scientific and technological supremacy for the benefit of Asia". Others have articulated the potential for shifts in the global job market for science and engineering workers towards China to erode US dominance by diminishing the current comparative advantage in high tech production. However, there are many reasons why these trends are not necessarily bad news for the West. Firms locating activity in China, whether to adapt products to local markets or gain access to specific skilled workers at a lower cost, can lead to standard gains from trade - firms experience improved performance and may transmit knowledge back to the home country. There is evidence that knowledge flows across national borders and is less restricted by distance than was the case 20 years ago. Importantly, innovation is not zero sum game - that more research is being carried out in China does not necessarily imply that less will be undertaken in the West. Indeed, it might be expected that there are more synergies in the creation of new technologies than new goods or services such that an increase in knowledge output in China compliments, rather than substitutes for, knowledge created in the West. In addition, China represents a new market for technologies developed in the West.

The Policy challenge

The challenges for Western governments relate not to devising policies to deter investment in China or other emerging economies, but to ensuring that they make sufficient investments in their own economies such that they remain leaders in innovation. In large part this means ensuring that we have a high skilled workforce - that can engage, both competitively and collaboratively, in creating new knowledge and that is flexible enough to adjust to changing economic conditions - and that we invest sufficiently in science and research.

In a 2006 speech, China's President, Hu Jintao, launched a plan to make China an innovation-oriented economy and leading science and technology power, proclaiming that "by the end of 2020... China will achieve more science and technological breakthroughs of great world influence, qualifying it to join the ranks of the world's most innovative countries". In support of this ambition, the Chinese government has made unprecedented investments in research capacity and increased incentives for firms to invest in innovation. Juxtapose this with recent policy moves in the UK, and in the West more generally, where austerity packages to reduce borrowing have commonly involved real cuts to the budgets for science, direct research funding and higher education. Even where science budgets have been relatively protected from spending cuts, as in the UK, investment as a proportion of national income is due to fall in coming years. Failing to invest sufficiently in science and skills can be short sighted. The impact of such spending occurs in the long run, in the form of higher productivity and economic growth. Being able to compete with China in 10 years time requires investment in skills and research today. This was recognised in 2009 by US President Obama's American Recovery and Reinvestment Act that increased in spending on science, stating, "We'll provide new technology and new training for teachers so that students in Chicago and Boston can compete with kids in Beijing for the high-tech, high-wage jobs of the future."

The current economic climate should not prevent investment in our capacity for economic growth in the future. The impact of China's rise will depend largely on whether we are with them at the technology frontier or onlookers from the sidelines. We should choose the former.



]]> Mon, 05 Sep 2011 00:00:00 +0000
<![CDATA[How rich are you?]]> How rich do you think you are, compared to everyone else in the UK? Do you feel pretty well-off, rather poor, or just 'in the middle'? It's a simple enough question, but you may be surprised by the answer. Today, the IFS launches its first ever iPhone app, which estimates where in the income distribution you and your household fit in.

Of course income isn't everything, and there are lots of ways of measuring income. We use the same measure as the government in its official poverty statistics - that is household income, after direct taxes (including council tax) and benefits, and adjusted for family size. This means that we don't count a millionaire's spouse as being 'poor' simply because they have no personal income. It also means that larger families need a higher income than smaller ones in order to attain the same standard of living.

Imagine we were to line up everyone in the UK according to this measure of household income, forming a 60 million-strong conga line. Ignoring the fact that such a conga line would stretch more than half way around the world, lets further imagine that it passes you by at supersonic speed over a single day starting at 00.00 am. Over the first minute, you might catch a glimpse of the very small number of mega rich individuals. For example you might see CEOs and premier league footballers dancing past accompanied by their partners, such individuals could be earning over £2.5 million a year and have a weekly net income of around £23,000, over fifty times greater than the person in the middle of the conga line (the median, who you will see in 12 hours time). By 0.15 am you would have seen the richest 1% of individuals, such as bankers, surgeons or top-flight lawyers accompanied by their partners. Assuming they had no children, they would need a weekly net income of at least £2,400 to be in this top 1% (requiring income of well over £225,000 per year before tax for a single-earner). That's an income six times greater than median income. By 2.24 am, you would have seen the richest 10% of individuals, and would now be seeing some surprisingly "ordinary" families. A couple with no children where each earns £30,000 a year before tax, giving them a weekly net income of £840, would just make it in to the top 10% of the distribution.

Fast forwarding to midday, we are now exactly half way through the conga and you would now be seeing the median individual. Many different families could occupy this position. A couple with no children would need a weekly net income of £420 (e.g. a single-earner couple with an annual salary of just over £30,000 before tax). A couple with two children would, however, need a weekly net income of about £640 (e.g. a single earner with an annual salary of £45,000 before tax). By 7.55 pm you would be getting to the 17% of individuals with household incomes below the poverty line (60% of the median income). These would include couples with two children with net incomes of less than £380 per week, and lone parents with two children whose net incomes are less than £300 per week. It would also include pensioner couples with weekly net incomes of less than about £250 per week.

At the very end of the line at 11.45 pm, we might expect to see the poorest of the poor. We would almost certainly see some people struggling to make ends meet, but this is not going to be true for everyone with a low income. We would also see a number of people with temporarily low incomes - students and some self-employed who have made losses in the most recent period.

Now let's ask again: whereabouts do you think you'd fit into this richest-to-poorest conga line? Odds are that you think you'd be somewhere near the middle - not at the front, naturally (unless your surname is Abramovich), not even particularly near the front, but somewhere towards the centre of the line. Odds are, however, that you're wrong. Research conducted for the Joseph Rowntree Foundation has shown that the majority of people in the UK appear to believe that they are 'in the middle' of the income distribution. Yet in reality (or at least in our imaginary conga line) many of these people are miles away from the middle. From poorer people in the bottom third of the income distribution, up to people in the top five or ten per cent, many will happily report that they think they're 'in the middle.'

If you're intrigued to know exactly where you fit in to our imaginary conga line - you're in luck. The IFS website has long offered a calculator called (imaginatively enough) 'Where Do You Fit In?' Punch in the number of people in your household, their income (after tax) and the amount of council tax you pay, and it'll tell you exactly where you fit in to the UK income distribution. Better still, today we've launched IFS's first iPhone app, which does the same - just follow this link or search for 'IFS' in the App Store. (Of course, if you own an iPhone, odds are you're already some way up the income distribution.) If you don't, you can carry out the same exercise on our website

]]> Thu, 14 Jul 2011 00:00:00 +0000
<![CDATA[Higher debt for decades?]]> Much attention has focused on the damage done to the UK's public finances by the financial crisis and associated recession, and the painful tax rises and spending cuts required if borrowing is to be brought back to pre-crisis levels. A longer-term perspective is provided by the publication today of the Office for Budget Responsibility's (OBR's) first Fiscal Sustainability Report. This suggests that keeping the Government's finances in a sustainable position in the longer-term will require further uncomfortable decisions to be implemented over the medium-term, on top of delivering the fiscal retrenchment already planned for the next few years.

We know that (Figure 2.6) the increase in Government borrowing since the recent crisis began could not have been left unchecked. It would have left the UK's public finances in an unsustainable position with debt forecast to rise remorselessly throughout this century. The fiscal tightening outlined in the March 2010 Budget by then Chancellor Alistair Darling would, if it had been delivered and maintained, have been sufficient to avoid this situation. The swifter and larger fiscal tightening planned by Chancellor George Osborne will (again if delivered and maintained) see debt peaking sooner and at a lower level, and then falling faster (and therefore leading to lower debt interest payments) than under Mr Darling's plan.

However as Chart 1, taken from today's OBR report shows, even if the government manages successfully to implement the fiscal tightening it has planned for this parliament, in the longer run demographic pressures (particularly from the ageing of the population) will place upward pressure on public spending. If left unchecked these would lead to higher borrowing and, as shown in the top line in Chart 1, a rising stock of public sector net debt from the late 2020s. These are not new issues, but current and future governments will need to decide how to address these pressures if UK public finances are to become sustainable over the long run. If future Governments were able to maintain borrowing forevermore at the level currently forecast for 2