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. Sat, 26 May 2018 19:33:15 +0000 <![CDATA[Scottish income tax diverges further from rest of UK to raise more from high earners]]> Today the now devolved income tax system in Scotland has diverged further from its counterpart in the rest of the UK. The threshold for the higher-rate of Scottish income tax (now 41%) was already below the threshold that applies in the rest of the UK and today increases by only 1% to £43,430, while the income tax thresholds that apply to the UK apart from Scotland have been uprated in line with inflation (3%). At the same time, there are two new rates (of 19% and 21%) within the basic rate band in Scotland, while the higher- and additional-rates have each been increased by 1% to 41% and 46% respectively (see Figure).

Figure. Marginal tax rate schedule on earned income, Scotland and the rest of the UK, 2018–19

Notes: Dashed lines include income tax, employee National Insurance contributions, and the withdrawal of the income tax personal allowance from £100,000. Assumes all earnings are from one job. The spike in the Scottish marginal tax rate at £43,430 reflects the higher-rate threshold for income tax in Scotland being lower than the upper earnings limit for National Insurance contributions (which is set for the UK by Westminster). Solid lines include only income tax and the withdrawal of the personal allowance.

In 2018–19, the Scottish income tax system will raise around £355 million more in revenue (3% of the Scottish income tax take) relative to what it would have raised had it retained the same system as the rest of the UK. Of this difference, the changes enacted this April are due to raise £225 million with a further £130 million a result of freezing the higher-rate threshold in 2017. Effectively, the Scottish government is utilising its tax-raising powers in order to reduce the squeeze on spending in Scotland. 

Scotland chooses to raise more from top half of income tax payers 

The effects of the tax changes vary across Scots. Table 1 shows the number of Scottish income tax payers in different earnings bands, and the difference in tax liability they will face in Scotland relative to the rest of the UK. (Note that technically the Scottish income tax system applies only to non-saving non-dividend income, which we henceforth refer to as earnings income.)

  • 44% of Scottish adults do not have earnings above the personal allowance and so are unaffected by these changes.
  • 31% of Scots – specifically, those earning between £11,850 and £26,000 – will pay less income tax in Scotland than the rest of the UK, although the gain is small (up to a maximum of £20 per year). The total giveaway to this group is forecast to be £23 million.
  • 25% of Scots earn above £26,000 and will now pay more in Scotland than they would in the rest of the UK. In total, these individuals are forecast to pay £380 million more in tax, of which £250 million is a result of this April’s changes. Losses are greater for those with higher earnings. For example, an individual earning £28,354 per year (the median full-time earner in Scotland) will have a £24 higher tax liability in Scotland compared to the rest of the UK, while an individual earning £170,000 per year (enough to place them in the top 1% of UK income taxpayers) will have a tax liability over £2,000 higher in Scotland.

Table 1. Number of Scottish adults and difference in annual tax liability under the Scottish income tax system relative to that in the rest of the UK, by earnings range

Annual earnings

Number (000's)

Share (%)

Difference in tax liability




No change




£0-20 lower




£0-174 higher




£174-£1,943 higher




£1,943+ higher

Note: Numbers may not sum due to rounding. Population includes all individuals in Scotland aged 16+.
Sources: Authors’ calculations using HMRC, Income tax statistics and distributions, Office for Budget Responsibility Devolved Taxes Forecast, March 2018 and Office for National Statistics 2016-based National Population Projections.

Risks with the Scottish approach 

While the changes represent a relatively broad-based tax rise, the concentration of losses among higher income taxpayers poses risks to the Scottish income tax yield. Previous work has shown that higher income individuals can be particularly responsive to tax increases. If no-one changed their behaviour in response, the Office for Budget Responsibility (OBR) estimate that this April’s changes would raise around £320 million in 2022–23 (the final year of the forecast period). After accounting for likely behavioural responses, however, the forecast revenue gain falls by a quarter to £240 million.

One risk would be that these individuals migrate south of the border into England. While genuine migration is costly, it may be easier for those with multiple properties to change their tax residence within the UK. The OBR assumes no response in terms of genuine migration, but estimates that changes in reported tax residence as a result of this year’s changes will lead to revenue losses to the Scottish Government of around £20 million – almost 10% of the predicted yield. They note that an effect of this magnitude would require only a small number of additional rate taxpayers (fewer than 20) to switch their residence. This is because when a taxpayer relocates it is not just the additional tax revenue that the Scottish Government would lose, rather all of the income tax that they pay on their earnings would now go to the Westminster Government rather than Holyrood.

A risk to the UK wide income tax base is that of individuals incorporating – setting up their own company and paying themselves in tax advantaged dividends rather than salary. Growth in company owner-managers has outstripped growth in employees since at least the 1990s, and this is projected to continue, leading to a reduction in revenues of around £2.5 billion by 2021–22. While income tax on earnings is devolved to Scotland, dividend and corporation tax are set by Westminster. If an individual in Scotland incorporates, they not only pay less tax overall, but more of the tax they do pay is paid to Westminster rather than Holyrood. This means that (i) higher income tax rates make the incentive to incorporate stronger in Scotland and (ii) that Scottish revenues are particularly exposed to the risk of incorporation. If the rate of incorporation in Scotland outstrips that in the rest of the UK, Scottish income tax revenues would be depleted as a result (with no offsetting block grant adjustment).

The OBR forecasts explicitly take both of these behavioural responses – migration and incorporation – into account, though there is considerable uncertainty about the magnitude of their effect. While the OBR and Scottish Fiscal Commission estimate of the revenue from the latest changes are similar in 2018–19, the fiscal commission estimate in 2022–23 is 10% higher than the OBR’s (£267 million vs £240 million).

These behavioural responses are likely to be relatively moderate while the UK and Scottish income tax systems differ only slightly. But the ease with which individuals may be able to avoid higher taxes north of the border suggests that policymakers may wish to pause and consider the evidence of what happens in response to this small change before implementing any more radical departures.

]]> Fri, 06 Apr 2018 00:00:00 +0000
<![CDATA[More into workplace pensions: minimum default pension contributions rise for most employees and their employers]]> From tomorrow, a large proportion of private sector employees will pay more into their pensions – and their employers will have to contribute more too. This is the first of two planned steps in the next two years that will increase the minimum contributions that most employees and their employers will, by default, make to a workplace pension. This is all part of the government’s automatic enrolment policy aimed at increasing retirement saving.

In this observation, we summarise the changes that are coming in tomorrow and discuss what effect they may have on employees and their pensions.

Automatic enrolment so far

Automatic enrolment is a flagship government pensions policy that means that all eligible employees must be enrolled into a workplace pensions scheme (a pension scheme arranged by their employers). Until tomorrow the minimum default contributions that must be made were 2% of “qualifying” earnings (between £5,876 and £45,000 in 2017–18), of which at least 1% must come from the employer. The employee can choose to “opt out” and leave the pension scheme (in which case both they and their employer will stop contributing), but unless the employee actively does this, they will remain part of the scheme.

This policy has been gradually rolled out since the Autumn of 2012, starting with the very largest employers and progressively including smaller and smaller employers. Employees are eligible if they earn the equivalent of at least £10,000 per year, are aged between 22 and the state pension age, and have worked at their employer for at least 3 months. The policy has led to an enormous increase in the proportion of private sector employees who are saving for their retirement in a workplace pension. As shown in Figure 1, it doubled between April 2012 and April 2018, rising from one-third (32%) to two-thirds (67%). This increase has reversed the decline in private pension membership seen since the 2000s, and leaves membership much higher than the levels seen in the late 1990s (it was 48% in 1997), though it is still lower than the levels seen in the public sector. Previous IFS research has shown that the increase since 2012 has been directly driven by the introduction of automatic enrolment.

Figure 1: Workplace pension membership rates among public and private sector employees, 1997 to 2017

Source: ONS ASHE Pension Tables (2011 to 2017) and Cribb and Emmerson (2016) calculations using ASHE microdata (1997 to 2010).

Although pension membership has risen enormously, the number saving large proportions of their earnings has not. Automatic enrolment has mainly increased the proportion of employees and employers making very low contributions at the minimum levels. Figure 2 shows a very large increase in the proportion of employees who are in a workplace pension but who are contributing less than 2% of their earnings since 2012 – increasing from just 2% to 32% of private sector employees. It is a similar story with employer contributions. Therefore, although many more people are contributing to a pension, it is, in most cases, only at very low levels. There is however a slight increase in the proportion of employees who are making more significant contributions: in 2012 5% of private sector employees were making an employee contribution of more than 7% to a workplace pension; by 2017 this figure had increased to 6%.

Figure 2: Distribution of employee contribution rates for all private sector employees in 2012 and 2017.

Source: Authors’ calculations using ONS ASHE Pensions Tables 2012 and 2017. 

Increases in minimum default contributions happening tomorrow

The changes happening tomorrow are aimed at increasing the amounts saved in workplace pensions. Currently, the minimum default rate that can be contributed is 2% of “qualifying earnings”, with at least 1% coming from the employer. Tomorrow, the minimum default rate will more than double to 5% of “qualifying earnings” (£6,032 to £46,350 in 2018–19), of which at least 2% will come from the employer. In April next year, the minimum contributions will rise again, to 8% of qualifying earnings, with at least 3% coming from the employer.

Figure 3 shows how much these minimum contribution rates amount to, in 2017–18 and in 2018–19, for people earning between £0 and £60,000 per year. An employee earning £10,000 per year who currently makes and receives minimum contributions will see an increase of at least £38 per year in their employer contribution (from £41 per year to £79 per year), and will make an additional £78 per year of employee contribution (from £79 per year to £119 per year). The maximum increase is for someone earning £46,400 or more, who will receive an additional £415 per year (from £391 per year to £806 per year) from their employer, and make an additional employee contribution of £818 per year (from £391 per year to £1,210 per year).

Of course, while the higher employer contributions will reduce take home pay for employees, they will not reduce it by as much as suggested here. Since contributions to pensions are exempt from up-front income tax, employees will pay lower income tax as a result of higher contributions (worth at least 20% of the contribution for those who pay income tax in 2018–19) – though of course income tax is likely to be paid when the pension is drawn). Furthermore, if they contribute through a “salary sacrifice” arrangement – they will not pay any NICs on the contributions either (either when the contribution is made or when the pension is drawn).

Figure 3: Minimum default employee and employer contributions to workplace pensions in 2017–18 and 2018–19 (£ per year)

How will people respond?

 Importantly, a significant number of private sector employees will not be directly affected by tomorrow’s changes. At least 28% of private sector workers in 2017 are already in a scheme with higher default employer contributions than the total minimum default mandated by the government.  A further 32% are not in a workplace pension scheme – either because they have not been enrolled automatically (for example if they earn below £10,000 a year then their employer does not have to enrol them automatically) or because they chose to opt out after being enrolled – , and so are not affected by the rising contributions either.

However, the change will still be a hurdle for the government’s automatic enrolment policy.  So far, the proportion of employees choosing to leave their workplace pension after having been enrolled automatically has been small. It is an open question whether employees will continue to stay in their workplace pension schemes, and whether newly enrolled employees stay in at the same rate, or whether more will choose to opt out because of higher required contributions. 

There are at least three theoretical reasons to for the government to be optimistic that the proportion of people opting out will not increase markedly. First, individuals procrastinate when making saving decisions and tend to shy away from making complex choices. This means that when they are defaulted into a workplace pension some will remain in the scheme simply due to inertia. Second, individuals would be giving up a (now larger) employer contribution if they chose to opt out, which gives them a financial incentive to remain in their pension. Third, it is possible that the default minimum rates are seen as being endorsed – perhaps by their employer, or by the government, or perhaps even by their colleagues who are remaining in the scheme – and therefore employees will choose to remain in their pension with the higher rate of contributions.  Furthermore, there is evidence from the United States, where companies have introduced automatic enrolment with higher default contributions, that suggests that coverage rates may hold up: Madrian and Shea (2001) studied a firm which had default rates of 3% from employer and 3% from the employee, and found that only 14% of workers opted-out (not that much higher than among large firms in the UK while default total contributions have been at 2%).

Pension membership rates going forwards therefore need to be carefully monitored. But a harder, and arguably more important question, is where the additional pension saving of those who stay in their pensions at higher default contribution rates is coming from. If employees are saving an extra 2% of “qualifying earnings” in their pension, are they spending less (and if so, on what) or are they saving less in other forms? That is what will matter for the government’s ultimate ambition of increasing individuals’ retirement resources – increasing workplace pension coverage and pension savings is a means to an end, rather than the end itself.

]]> Thu, 05 Apr 2018 00:00:00 +0000
<![CDATA[The gender wage gap and new employer reporting requirements]]> Tomorrow is the final deadline for organisations with 250 or more employees to fulfil the new gender pay reporting requirements. These requirements include providing information on the percentage gap between both mean and median hourly wages of men and women. Additional information has to be provided on the proportion of men and women receiving bonuses and on the gender breakdown of each quartile of the organisation’s hourly pay structure.

These data could help shed useful light and attention on an important issue. IFS researchers have recently shown the gender wage gap has not fallen very much at all in the last fifteen years: it was 23% in 2003 and is now 20%. And among graduates the gender wage gap has not fallen at all over this period. The research also shows that some, but not all, of the gender wage gap is explained by the fact that after having children women are much more likely than men to work part-time, and part-time work is associated with little or no subsequent progression in hourly wages. This is particularly important for graduates for whom full-time experience typically delivers considerable wage growth.

The greater prevalence of part-time work among women means that the government is wise to ask for reporting of hourly wages rather than annual earnings. Employers offering very family-friendly employment terms are likely to employ a greater fraction of women part-time and therefore would have large differences in average annual earnings even if there were no difference in hourly wages.

Requiring large employers to provide information on their gender wage gap could therefore be a useful step. There should be benefits from both transparency and greater awareness of the issue. Over time the data will allow the monitoring of trends in gender pay gaps in different industries.

The data can be downloaded from the Government’s gender pay gap service here. At the time of writing (7pm on Monday 2 April 2018) information was available from 7,556 organisations. The Government’s impact assessment estimated that around 8,000 private-sector and voluntary organisations would be required to provide information. And the obligation applies to public-sector employers too.

As ever, however, the statistics are limited and need to be interpreted with care. The IFS research cited above shows that part of the gender wage gap results from women on average having less experience of being in paid work full-time than men. The cause of – and therefore the appropriate response by policymakers to – this part of the gender wage gap could be very different to, for example, the part of the gender wage gap that exists prior to the arrival of a first child. And even if men and women were on average paid the same, one would expect to see some variation across employers. For some employers we should not be surprised to see considerable gender pay gaps, and to see a much bigger difference in the mean gap than the median gap. For example five organisations are currently reporting that their mean gender wage gap is 80 percentage points or more larger than their median gender wage gap and all five are English Premier League football clubs. In addition, within-organisation gender pay gaps may miss important causes of the wider gender wage gap: namely differences between men and women in the kinds of organisations that they work for. There is evidence from elsewhere that part of the gender wage gap is explained by men being more likely to work for more productive, higher paying, firms.

A careful examination of the data suggests some reporting errors. As the Financial Times has highlighted there is a greater prevalence of employers reporting a pay gap of exactly zero, at both the mean and the median, than would be expected. Currently this is true of 45 reports. There are also some other entries that look incorrect. For example one organisation reports that almost everyone in the bottom half of its pay structure (at least 99.8%) is female whereas the top half comprises both men and women – and despite this it reports that it its gap in median hourly pay is precisely zero!

The full distribution of reported gender pay gaps from the data so far is shown in Figure 1 (median) and Figure 2 (mean). Most organisations have a positive gender wage gap – that is on average men receive more per hour of work than women. This is true of 79% of organisations when considering median pay and 88% of organisations when considering mean pay. The mean gap is generally greater than the median gap indicating that the very highest paid in each organisation are disproportionately male. This is shown by the fact that in over 60% of organisations the majority of the highest-paid quarter of employees is male.

Figure 1. Distribution of reported median gender pay gaps (data as of 2/4/2018)

Source: Author’s calculations using data downloaded from at 7pm on April 2 2018 (7,556 organisations).

Figure 2. Distribution of reported mean gender pay gaps (data as of 2/4/2018)

Source: Author’s calculations using data downloaded from at 7pm on April 2 2018 (7,556 organisations).

Once the final data are available it will be interesting to see if the later reports typically have larger pay gaps than those who reported earlier (which could explain a potential greater reluctance to report). Further analysis by sector and industry – and when data from future years are available in trends over time – will also be of interest.

Postscript: The Institute for Fiscal Studies is not required to publish our gender pay statistics since we have fewer than 250 employees. Having argued that it is good that these data are being made available it would seem inappropriate for us not to provide them. In fact the mean pay of women employed by IFS is slightly higher than that of men (a gender pay gap of –4%) while median pay of women is quite a bit higher than that of men (a gender pay gap of –29%).

]]> Tue, 03 Apr 2018 00:00:00 +0000
<![CDATA[The first Spring Statement: no surprises sprung?]]> On Tuesday 13 March the Chancellor Phillip Hammond will present the first Spring Statement to Parliament.

While he has reserved the right to announce new measures in the Spring Statement “if the economic circumstances require it” these are not expected. If this is kept to – in this and future years – then Mr Hammond should be commended rather than criticised as we should not have two fiscal events every year. But the precedents here are not good: Gordon Brown’s first Pre-Budget Report in 1997 and George Osborne’s first Autumn Statement in 2010 both contained very few measures, but neither Chancellor subsequently managed to resist the temptation to have two full-blown fiscal events each year.

Mr Hammond should have some good fiscal news to announce. In November the OBR revised down its forecast of borrowing this year from £58 billion to £50 billion. Stronger than expected tax receipts since then mean that should come down further and will probably come in below the £46 billion borrowed last year. If, as is just possible, borrowing comes in below £41 billion this financial year (and we spent that amount on investment, as planned) then we would see a surplus on the current budget – i.e. the amount being borrowed will be less than the amount of spending which counts as investment, or capital, spending. Such a surplus has not been achieved in the UK since 2001–02.

While borrowing has broadly returned to pre-crisis levels, and has come down very sharply from 10% of national income peak in 2010–11, public debt, at around 86% of national income, is more than twice as high as it was pre-recession and at its highest level since 1965–66.

The improvement in public finances this year reflects strength in tax receipts. Income tax revenues have been strong, largely driven by self-assessment receipts, which based on the January numbers imply a stronger full-year out-turn than forecast. Corporation tax receipts look set to be revised up again as corporate profits continue to exceed OBR expectations. Receipts of stamp duty on property transactions have also been strong in recent months. However, the fall in the forecast borrowing in future years could be smaller than that for 2017–18 for at least three reasons:

  • First, some of the strength in receipts this year may represent a one-off boost rather than something that can be expected to recur in future years. For example self-assessment receipts in January were not as weak as implied by the OBR’s forecast, but some of this could reflect revenue that was previously expected to be received in future years.
  • Second, the market expectation is now that interest rates will rise more quickly than thought in November. The way that the Bank of England’s programme of quantitative easing is scored means that debt interest spending is highly sensitive to this, and is likely to be revised up by £1 billion to £1½ billion a year in the early 2020s as a result.
  • Third, the Government announced in January that it would not appeal a High Court decision reversing recent changes to the Personal Independence Payments rules affecting some who find it hard to make journeys due to experiencing overwhelming psychological distress. The Department for Work and Pensions has provided an early estimate that this ruling will benefit around 220,000 individuals at a cost over five years of £3.7 billion.

Even so Mr Hammond will probably be able to point to lower forecast borrowing over the next five years than expected in November. But before the champagne corks start popping it is worth comparing the latest borrowing forecasts to those made just two years ago. Then Mr Osborne, as Chancellor, was forecasting that we would have a surplus of £10 billion in 2019–20, in line with his commitment to have eliminated the deficit by that point. As shown in the Figure below, the fiscal forecasts have deteriorated quite dramatically since then, with the November forecast being for the UK government still to be running a deficit of £26 billion in 2022–23. A large part of this deterioration was ascribed by the OBR in 2016 to the likely effects of the decision to leave the European Union. A subsequent downwards revision occurred in November 2017 as the OBR downgraded forecast productivity growth in the light of the UK’s dismal performance over the last seven years.

So, the welcome modest improvement in the outlook for the public finances expected on Tuesday will still leave us looking at a much bigger deficit than expected in March 2016.

Deficit forecasts: March 2016 and November 2017 compared

Source: Office for Budget Responsibility.

New policies in the pipeline

We may not get any new announcements on Tuesday but there are plenty of new policies coming into effect this April. Motorists will benefit from yet another freeze in fuel duties, while many receiving a large inheritance will benefit from a further £25,000 increase in the main residence allowance within inheritance tax. Tax rises will hit those with more than £2,000 of dividend income, those buying sugary soft drinks, and those with mortgages on buy-to-let properties.

There are also significant public spending cuts still in the pipeline. For example spending by the Ministry of Justice is intended to fall by 16% over the next two years. This may not prove easy to deliver in the context of a stable prison population and growing problems within prisons. The NHS budget is not being cut: but after accounting for economy-wide inflation and growth in the population it is broadly flat, which will continue to be very challenging given population ageing and other pressures on the health service.

Of most immediate consequence for some households is the large package of benefit cuts announced by Mr Osborne after the 2015 general election. Additional cuts will continue to affect the lowest income half of working age households adversely. These include:

  • The third year of the four year freeze to the rates of most working age benefits. This will be much the most painful freeze so far. Inflation in September 2015 and September 2016 was very low, so inflation uprating would have implied an increase of just 1% over April 2016 and April 2017 combined. By contrast inflation has been much higher recently and benefits would have increased by 3% this April to compensate for price rises. The saving to the government, and cost to households, of the freeze this year alone is over £2 billion.
  • No longer paying a higher-rate of tax credit for new first children (a loss of up to £545 for a household), or paying additional means-tested benefits for most new third or subsequent children (a loss of up to £2,780 per extra child), which was first introduced in April 2017, will affect more families.
  • Universal credit, which on average is less generous than the system of legacy benefits that it is replacing, continues to be rolled out across the UK. Some of the biggest losers relative to the legacy system include those who own their own home, families with significant amounts of unearned income or financial assets and low-income self-employed individuals.

Finally Mr Hammond has said that Spring Statement will “consider longer-term fiscal challenges and start consultations on how they can be addressed”. There are many potential topics that he could consult on. These include: how should we confront the public finance challenges posed by an ageing population, how should the tax system be reformed in the light of trends such as the growth in the ‘gig economy’ and the shift to electric vehicles, and many others. Which issues – if any – the Chancellor chooses to start a consultation on could be the most important part of Tuesday’s announcement.


On Wednesday 14 March, the day after the Spring Statement, IFS researchers will present their analysis of the public finances, putting them in the context of recent trends, and will discuss economic challenges facing the country. They will also respond to any announcements on policy made by the Chancellor and set out the impact of measures set to come into force this April. You can register to attend here.


]]> Fri, 09 Mar 2018 00:00:00 +0000
<![CDATA[Poverty and low pay in the UK: the state of play and the challenges ahead]]> The problem of low pay is increasingly dominating the domestic policy agenda. At an event in Westminster this morning, to mark the end of a programme of research funded by the Joseph Rowntree Foundation, we are setting out and discussing the latest evidence on poverty in the UK and in particular the challenges posed by the rise of in-work poverty. The presentations from the event are available below. This observation summarises the key points.

The nature of low income in the UK has changed radically. This is due to a combination of good and bad news. In the past the great problems were the number of workless households and the prevalence of poverty in old age. Both of those problems still exist, but are now considerably smaller. In 1994-95 22% of children and 18% of all non-pensioners lived in households where no adult was in paid work, and most of these were in poverty. The latest figures are 13% and 12% respectively. As is now well known, the prevalence of poverty among pensioners has also fallen dramatically. But earnings growth for those in work has been historically weak.

All this means that today the big issue is the number of people who live in working households who are in poverty. 57% of people in poverty are children or working-age adults living in a household where someone is in paid work; up from 35% in 1994-95.[1]  This means that poverty is far more sensitive to the plight of low-earning working households than used to be the case. In-work poverty has become one of the most important challenges we face.

A closer look at trends in earnings for low-earning households

  • The major reason why paid work lifts households above the poverty line less reliably than in the past is simply that earnings growth in the UK has been so low since the early 2000s, and non-existent overall since the recession – linked to our dismal productivity performance.
  • In addition household earnings growth at the bottom end has been even weaker than the average over the past 20 years.
  • The driver of that rise in household earnings inequality has been a rise in earnings inequality among men. This in turn has occurred despite little change in inequality in hourly wages between men; it has been caused by changes in hours of work. Men with low hourly wages are now much more likely to work part-time than they were in the past – a quarter of the lowest earning men now work less than 30 hours a week. Understanding the reasons for this should be a priority.

What has been the policy response?

  • Since the mid 1990s, rising cash transfers from the state have kept inequality in net incomes between working households approximately flat, despite the rise in inequality in household earnings among the same households. Big increases in tax credits since the late 1990s were especially important in this regard: real spending on tax credits and equivalents rose from £7bn per year in the mid 1990s to a peak of £32bn in 2011, and is currently £25bn.
  • The policy direction is now very different. Cash transfers have been, and are being, cut back as part of the fiscal consolidation. Real tax credit spending is falling. The major policy lever that is now being pulled to try to prop up low earners is a large rise in the minimum wage. This is already having clear impacts - growth at the bottom of the hourly wage distribution has been far higher than for the rest over the past couple of years.
  • But the potential beneficiaries from minimum wages are different from the potential beneficiaries from cash transfers – one reason why it is unhelpful to think of them as substitutes. Higher minimum wages are paid to people with low individual hourly wages. Higher cash transfers are typically paid to people in low income households. There are reasons why one may want to do both of these things, but minimum wages are not very well targeted at households with the fewest resources.
  • Both policies share one crucial feature: they involve tradeoffs. This doesn’t mean we shouldn’t use them; but it does mean they can only be pushed so far before the tradeoffs associated with them become too big to bear. Generous tax credits are expensive and can have big effects on work incentives. Beyond some point higher minimum wages will start hurting the employment prospects of the low skilled by making it too costly to employ them. That point might still be some way off – we simply don’t know.
  • Ultimately then, if we want to prevent low-earning households falling further behind then we may have to do it the hard way: understanding better, and addressing, the underlying causes of their low earnings.

The drivers of low pay: a story of (lack of) progression

  • Taking a lifecycle view of wages tells us that low pay is highly related to lack of pay progression. The wages of the low- and high- educated, and of men and women, end up much further apart by age 40 than they were at the start of their careers.
  • Different sources of wage growth tend to go together and to be complementary to each other. For example, experience and education are both positively associated with higher wages, but the association with experience is much stronger for the high-educated than the low-educated. We also see the lower-educated getting far less job-related training than the higher-educated.
  • The fact that women’s wages fall behind their male counterparts over the lifecycle is, in part, related to a remarkable lack of wage progression in part-time work: in other words, past experience in part-time work seems to count for very little when it comes to the hourly wage that can be commanded now. Understanding the reasons for this is an important challenge.

Policy is not powerless in the face of the rising challenge of in-work poverty. Both tax credits and minimum wages have had big impacts over the past 20 years – and, for the most part, broadly the intended impacts. They are tools that governments can use perfectly sensibly. But in the long run, if low earners are to stay in touch with the rest then applying these patches may not be enough: there is probably no adequate substitute for better understanding, and addressing, the underlying problem.


End note

[1] Measuring incomes after deducting housing costs and using a poverty line of 60% of median income. All numbers in the second and third paragraphs are calculated using the 1994–95 and 2015-16 Family Resources Survey.


Incomes in low paid employment, Robert Joyce, Associate Director IFS

Poverty in the UK: past trends and future outlook, Agnes Norris Keiller, Research Economist IFS


]]> Tue, 06 Mar 2018 00:00:00 +0000
<![CDATA[Tax reliefs: look for the tax design behind the big numbers]]> The UK has over 1100 tax reliefs. According to HMRC, we forego over £400bn of revenue each year as a result of reliefs. The top 10 categories (see Figure) cost £375bn. £34bn of the cost is for reliefs from corporation tax and capital gains tax for business assets. These are big numbers well suited to catchy headlines.  You can make them particularly striking by comparing them to total government spending – around £800bn – or to corporate tax receipts – around £50bn. But these numbers aren’t particularly meaningful unless you know what tax reliefs are.

Helen Miller, Associate Director at IFS, will be part of a panel at tonight’s CIOT/IFS debate: Business tax reliefs - corporate welfare or essential elements of the tax system? Here she sets out some of the issues.

What’s a tax relief?

Much difficulty arises because there isn’t a clear definition of what counts as a tax relief.  Many people will imagine tax reliefs as bungs to special interest groups. And while many reliefs are in place to, in HMRC’s words, ‘help or encourage particular types of taxpayers, activities or products for economic or social objectives, a large number can ‘be regarded as an integral part of the tax structure’. This conceptual distinction between types of reliefs serves to highlight that they can be part of a desirable tax system. But there is room for disagreement about which reliefs should be part of our tax structure and which activities should be encouraged. No matter how we define or classify them, any debate about tax reliefs must ultimately be a debate about tax design – what is it we want to tax and how do reliefs help us achieve that?  

Economists want a broad base and low rates, right?

Economists can’t often be accused of clearly conveying their ideas. But one that seems to have been picked up widely is that a good tax system has a broad base and low rates. There’s something to this. Since almost all taxes distort (i.e. change) people’s behaviour in ways that are undesirable, it makes sense not to load all the distortions onto one tax base. A broad base also makes sense because exempting some activities or incomes creates boundaries in the tax system, which in turn create complexity and avoidance opportunities and often lead to unfairness. But, the broad base/low rate rule of thumb doesn’t imply zero reliefs are optimal. In fact, economists often advocate a narrow tax base and high rates. The treatment of North Sea oil, where there are large reliefs for investment costs but profits are taxed at 50% (and until recently at much higher rates), is a good example of this.

A better, admittedly less snappy, rule of thumb is to tax all income and deduct all costs of generating taxable income. The first part of this – tax all income – is related to the desirability of a broad base. The second part is an acknowledgement that a well designed base shouldn’t be too broad. One application of this is that it is desirable to tax companies’ profits rather than revenues.  If we tax revenue, we will tax high-revenue, high-cost activities more heavily than low-revenue, low-cost activities. It is inefficient to favour activities that have low costs.

This rule of thumb gives a starting point for rationalising some reliefs as part of a well designed tax system. To continue the example, a well designed corporate tax base will allow firms to deduct costs from turnover to arrive at taxable profits. Wage costs can be immediately deducted. But there are multiple ways to deduct the cost of investment. Using the Annual Investment Allowance, firms can immediately deduct the full cost of the first £200,000 of spending on plant and machinery. For investment above this, they can use capital allowances to deduct a proportion of their investment spending over a number of years. But, there is no deduction for investment in industrial buildings. And while a firm can deduct the interest payments associated with financing an investment through borrowing, there is no equivalent deduction for the cost of using equity finance. The upshot is that our current system does not consistently deduct the cost of investment meaning that some investments are discouraged, some are incentivised and some are unaffected by the tax system. The policy takeaway is that we could do a better job of designing the corporate tax base but this does not mean that we should move to a system with no reliefs for investment expenditure. This is a good example where evaluating a relief requires digging into the policy detail.

When to use the tax system to incentivise?

As a starting point, it is desirable to design taxes in such a way that we minimise the extent to which they lead people or firms to change their behaviour. However, there are exceptions to this. Governments may want to use the tax system to incentivise some behaviours and disincentivise others. The resulting reliefs – such as R&D tax credits, Entrepreneurs’ relief, Patent Box, film tax relief – are those that are most likely to divide opinion. In these cases, we should ensure that there is a clear policy justification, that the policy is well targeted and that the benefits outweigh the costs.

R&D tax credits are a good example of a well justified policy. The market under invests in R&D because some of the benefits from investigating a new technology ‘spill over’ to other firms. R&D tax credits are an attempt to correct this ‘market failure’. They are directly targeted at the activity that we think is under provided – R&D investment – and evaluations show that the policy is cost effective; every £1 of tax credit leads to more than £1 of R&D investment and to an increase in the number of new innovations. This doesn’t imply that the relief has no costs beyond the revenue forgone.  It introduces complexity and can lead firms to relabel activities as R&D. Even for a well justified and targeted relief, the benefits must be weighed against the costs.

It is plausible that investment by entrepreneurs also has spillovers and is underprovided. However, Entrepreneurs’ Relief – a large capital gains tax break given to most company owner-managers – is poorly targeted. Unlike R&D tax credits, it is not targeted at investment but at profits. It is available to many businesses that aren’t entrepreneurial but not available to entrepreneurs who cannot take their income in the form of capital gains. It distorts decisions over how income is received and incentivises tax motivated incorporation.

Which of the 1100 can we scrap?

Too many reliefs have weak or poorly articulated policy aims. Reliefs are often hard to rationalise as part of a tax base that reduces distortions or as a policy that is effective at achieving a well defined economic or social goal. To work out which we should scrap or modify, we should get more accustomed to digging into the details and evaluating how each relief stacks up against a clearly stated tax design.  The bar for introducing any new relief should be high.

We could scrap some reliefs and increase revenues. But £250 billion of the £400 billion of “reliefs” mentioned above results from the income tax personal allowance, NIC thresholds and VAT zero rating and exemptions. Whether these should count as reliefs at all or as central structures of the tax system is moot. It’s clear that they are not easily and painlessly available as a route to raising more revenue. Lumping all existing reliefs together encourages the idea there are huge sums of money to be raised at little or no economic cost. There aren’t.

 Top 10 categories of tax reliefs cost £375bn

Source: Notes: HMRC list 431 reliefs and cost 192. The 2017-18 cost in foregone revenue is £414bn. This should not be seen as the revenue that would be raised if we scrapped all reliefs because it does not account for any behavioural response or any interaction between reliefs. Some of the categories in the Figure are an amalgamation of reliefs (all exemptions and zero rating in VAT is in the second bar).

Other references:

The most recent Office for Tax Simplification (OTS) report list reliefs can be found here.

Data on government revenues and total spending can be accessed from the Office for Budget Responsibility’s public finance databank, available here: Figures in the text refer to 2017-18, rounded to nearest £10 bn.

Places are still available at this evening’s CIOT/IFS panel event. To join the debate at the Royal Society of Arts in London, sign up here.


]]> Mon, 05 Mar 2018 00:00:00 +0000
<![CDATA[Who benefits from benefits?]]> Debates about welfare policy often discuss benefit recipients as though they are a fixed, relatively small group of people. In reality, people’s circumstances fluctuate frequently over their lifetimes, often dramatically and in ways that matter hugely for entitlements to benefits. People’s health changes, they move in and out of work, their earnings vary, and children come and go. In new research published today in the Journal of Economic Inequality, IFS researchers use data which tracks the same individuals over long periods of time to provide a longer-run perspective on people’s interactions with the benefits system. Two key findings are that:

  • While only about a fifth of people report receiving one of the UK’s main working-age means-tested benefits at any one time, more than half do so over an 18-year period;
  • While out-of-work benefits look much more effective at directing money to the poorest in any one year, looked at over a lifetime work-contingent benefits are at least as effective at supporting the lifetime poor.

This observation discusses the key findings in more detail.

A snapshot in time can miss important changes

First, the data show that people’s circumstances are subject to so much change that summaries of who gets what at a snapshot in time can miss a huge amount of what’s going on. For example, while at any point in time only about a fifth of individuals report being in a family receiving one of the UK’s main working-age means-tested benefits, more than half of people report receiving such a benefit at some point over an 18-year period (the longest observable using this data). Because these estimates are based on survey data, which tends to under-record benefit claims, the true figure is likely to be higher still1.

This tells us that the working-age benefits system is actually providing support to a very broad group of people - the majority, in fact. It’s just that this support is typically concentrated during particular stages of life (for example, when one has dependent children, or is experiencing a spell of unemployment or low earnings), and people pass through those stages at different points in time. Considered together with the tax system, which people tend to pay into more during periods of higher income, much of what the system is doing is redistributing resources across periods of life, rather than between individuals.

Of course this redistribution across life is even greater when one considers the retirement period too – a stage in which many people start receiving cash transfers from the state after being net contributors to the system during their working lives.

Clearly, this sense of perspective is important. Having an accurate sense of what the benefits system is really doing from a lifecycle perspective is necessary if one is to have sensible debates about its (subjective) fairness.

A lifecycle view can help us better understand effects of policies

In addition, taking a lifecycle rather than a snapshot view of who gets what can lead us to very different conclusions about the relative effects of different policy options. Consider a government contemplating equally costly increases to out-of-work benefits (e.g. jobseeker’s allowance) and work-contingent benefits (e.g. Working Tax Credit). When assessed at a snapshot in time, increases to out-of-work benefits appear easily the more progressive: as Figure 1 below shows, gains from such a reform are concentrated among the lowest income tenth (decile) of individuals. This is unsurprising, as those who are currently out-of-work tend to be much lower down the income distribution at that snapshot in time than those currently in paid work.

Figure 1: Snapshot distributional impact of reforms

Source: Figure 10 in Levell, P., Roantree, B., and Shaw, J., (2017). ‘Mobility and the lifetime distributional impact of tax and transfer reforms’. IFS Working Paper (W17/17). Note: Deciles are defined on the basis of cross-sectional equivalised net household income. The height of the bars show gains as a percentage of the relevant decile’s total net (unequivalised) household income. ‘Non-employment benefits’ series shows the effect of a 16.5% increase in maximum income support, (income-based) jobseeker’s allowance and (non-contributory) employment support allowance. The ‘work-contingent benefits’ series shows the effect of an 18% increase in maximum working tax credit. In both cases, the baseline tax and benefit system is the April 2015 system. All individuals face the same system throughout life uprated in line with average earnings (AEI). To aid comparison, we have scaled gains proportionally such that the ‘All’ bars are the same across reforms.

However, many people on out-of-work benefits at one point in time are out of work temporarily and are in fact not at the bottom of the lifetime income distribution: they had higher incomes in the past or will have higher incomes in future. Conversely, many of those with the lowest lifetime incomes tend to spend much of their working-age life in paid work. As a result, they will still, on average, gain considerably from increases to work-contingent benefits. Figure 2 below shows that when gains are instead assessed over a whole lifetime (using a simulation approach), we obtain a very different impression to that obtained from a snapshot: the distributional impacts of out-of-work and work-contingent benefits now look very similar.  

Figure 2: Lifetime impact of reforms

Source: Figure 10 in Levell, P., Roantree, B., and Shaw, J., (2017). ‘Mobility and the lifetime distributional impact of tax and transfer reforms’. IFS Working Paper (W17/17). Note: ‘Non-employment benefits’ series shows the effect of a 16.5% increase in maximum income support, (income-based) jobseeker’s allowance and (non-contributory) employment support allowance. The ‘work-contingent benefits’ series shows the effect of an 18% increase in maximum working tax credit. In both cases, the baseline tax and benefit system is the April 2015 system. All individuals face the same system throughout life uprated in line with average earnings (AEI). To aid comparison, we have scaled gains proportionally such that the ‘All’ bars are the same across reforms. 

Being clear about distributional objectives is crucial for policy

Ultimately, the distributional patterns shown by both Figure 1 and 2 are of interest. One of the basic functions of the welfare system is to provide a safety net for those in short-term hardship, so we do care about snapshot distributional impacts as well as longer-term ones. But this does highlight how important it is for policymakers to be clear about whether the distributional objective behind a given policy changes is the alleviation of short-term hardship or the redistribution of lifetime resources. While there may be good reasons to pursue both, the most effective policy option, and the individuals who will be affected, can depend greatly on which is prioritised.




End Notes

1. In addition, as these estimates are based on 18 consecutive snapshots of data, they will miss short benefit claims occurring between waves

]]> Thu, 01 Mar 2018 00:00:00 +0000
<![CDATA[A 'Brexit Dividend' to spend on the NHS?]]> Regardless of Brexit, the NHS will face increasing demand and cost pressures over the coming years. The Office for Budget Responsibility (OBR) recently projected that NHS spending would need to increase by 5.3% of national income over the next fifty years in order to meet increasing demand for its services and other cost pressures —a sum equivalent to about £110 billion in today’s money. This includes meeting the pressures from a growing and ageing population, as well as the cost of new technologies and treatments, pressures that are common to public and private health systems across the world.

Such increases would be in line with historical rises in NHS spending, which rose from about 3% of GDP in the 1950s to about 8% in 2009-10, but would be far above those of the past seven years. Since 2009–10, real growth in NHS spending has averaged only 1.1% a year. Calls for increased funding have grown recently amid increasing signs of strain including lengthening waiting times and a larger numbers of cancelled operations.

In the short term, the immediate effects of the Brexit vote will have increased these pressures. The depreciation of sterling has increased the costs to the NHS of goods and supplies that are imported, and the resulting inflation has also eroded the real value of public-sector workers’ pay. In November 2016 the OBR forecast that the sterling depreciation would increase prices by 2% in the medium term. Simply compensating NHS staff for this increase in prices would cost around £1 billion in additional salaries. Indeed, in the November Budget in 2017 the Chancellor announced that the public-sector pay cap would end for most NHS staff, at an expected cost of around £700 million in 2017–18. Finding additional funds to cover such increases obviously affects the amount of healthcare that can be provided within a given budget.

Given this context, it is not surprising that NHS funding has not only been important in recent policy debates, but also played a prominent role in the 2016 referendum campaign. This famously included claims that some or all of the UK’s gross contribution (including its rebate) to the EU budget—currently about £19 billion per year, or £360 million a week—could instead be diverted to the NHS. However, higher health funding will involve more difficult trade-offs than these figures imply.

In the short term, assuming a transition is agreed, the UK will continue to contribute to the EU until the end of any transition period; there will also be some additional payments as part of the “divorce bill”. Even after the transition, the full £19 billion will not be available. In particular, the UK’s rebate—currently about £5 billion—has already been “spent” domestically on the government’s priorities and simply cannot be spent again. The remaining £14 billion could theoretically be redirected, but the government has already pledged to replace at least some EU spending (for example, farming subsidies) for some years. The size of the UK’s net contribution to the EU after taking into account both the UK’s budget rebate and existing EU spending in the UK is about £8 billion. This represents a more realistic estimate of the sum that could be redirected to domestic priorities, including the NHS.

However the impact of Brexit on the wider economy is likely to have a greater impact than the amount the UK will get back from reduced EU contributions. Brexit represents a fundamental shake-up of the UK’s relationship with our largest trade partner. If, as most economists expect, this reduces economic growth over the medium to long term, it will almost certainly outweigh the reduction in our EU contributions; a reduction in GDP of just 1% translates to a fall in tax revenue of more than £8 billion.

The OBR has already incorporated some, but not all, of the potential impacts of Brexit into its forecasts. In the November 2016 Autumn Statement, it downgraded its forecasts for investment, productivity growth and immigration, as well as raising its inflation forecast relative to what it thought would have happened without Brexit. The implied hit to the public finances is about £15 billion per year by the early 2020s, about 10% of the NHS budget, more than outweighing the UK’s net contribution to the EU. This is despite reduced investment having only a modest impact on tax receipts, as investment spending is taxed less heavily than consumption. When over time this feeds into lower company profits, the long-run increase in the deficit may be around £3.5 billion larger.

In other words, so far the implication is that Brexit has reduced rather than increased the funds available for the NHS (and other public services), both in the short and long term. Although subsequent growth after the OBR’s forecast has been slightly stronger than expected, the medium-term outlook is now gloomier than in November 2016.

All forecasts are uncertain, and the fact we do not yet know what the UK’s post-Brexit relationship with the EU will be adds additional uncertainty. However, the OBR forecast is predicated on a relatively smooth transition, putting it at the optimistic end of the spectrum; in a ‘no deal’ scenario, in which the UK reverted to World Trade Organisation (WTO) rules, the economic dislocation would be much larger. Moreover, the OBR has not as yet incorporated any longer-term negative impacts of Brexit on productivity, and some economists expect these to be significant. 

In short, Brexit is likely to mean less money for public services, including the NHS, than otherwise would have been the case. Although the NHS budget may continue to rise in real terms, it is more likely than not that lower economic growth will take away funds that would otherwise have been available for additional spending increases. Even without Brexit, addressing the pressures on NHS funding would likely require significant tax increases, extra borrowing or diverting more money away from other services. Brexit will make these hard choices even harder.

This piece was first published by UK in a Changing Europe, an ESRC funded initiative.  The UK in a Changing Europe promotes rigorous, high-quality and independent research into the complex and ever changing relationship between the UK and the European Union (EU).

]]> Wed, 24 Jan 2018 00:00:00 +0000
<![CDATA[Firms’ supply chains form an important part of UK-EU trade: what does this mean for future trade policy?]]> Summary

We often think of exports and imports as things made in one country and consumed in another – I export cars to you and import socks you’ve made. In fact, the majority of UK exports and imports are now made up of goods or services that are themselves inputs into production – I export engines to your car factory and import cotton to make into socks. This sort of trade is particularly important for understanding the UK’s trade with the EU. Over half of the UK’s imports from the EU are of such intermediate goods and services, as are nearly 70% of our exports to the EU. These shares have also been growing over time.

This pattern of interdependence is crucial for understanding trade policy. Outside the single market and customs union, UK trade would still be affected not only by the trade deals it strikes with other countries but also by the trade deals the EU has with third parties. Moreover, in such an interdependent world, multilateral trade agreements are much more valuable than bilateral ones. As a result, even if the UK decides to make its own way in striking future trade deals, it will still pay to coordinate its efforts with others – including the EU.

Full version

Globalisation, and the accompanying internalisation of firms’ supply chains that has occurred over recent decades, have changed the way we should think about trade flows and in particular the UK’s trade with the EU. It is increasingly the case that countries’ exports embody imports from abroad, whether in the form of imported raw materials, components or business services.

Supply chains play a particularly important role in UK-EU trade. Not only is the EU the UK’s largest source of imports (accounting for 54% of the total in 2016), but a majority of these imports now take the form of intermediate goods and services. These are goods or services that add to the value of a product which firms then sell either for further processing or for final consumption. Tin used to produce a can is an intermediate good. A tin of beans sold to a supermarket for sale to consumers is not.

Nearly 70% of the UK’s exports to the EU take the form of intermediate inputs to production of other goods and services. The relative importance of the EU in the UK’s trade thus largely reflects the role UK industries play in EU-wide supply chains.

Figure 1 shows the importance of intermediate goods and services in the UK’s trade with EU and non-EU countries, and how this has changed over time, using figures calculated from the World Input-Output Database. The share of goods and services imported from the EU that are used in the production of UK goods and services rose from 51% in 2000 to 56% in 2014. The remainder was made up of goods and services intended for final consumption.

Figure 1. Share of intermediate goods and services in the UK’s imports and exports, 2000–14

Figure 1. Share of intermediate goods and services in the UK’s imports and exports, 2000–14

Source: Author’s calculations from World Input-Output Tables, 2016 edition.

The share of intermediate inputs in imports from the EU is similar to (in fact slightly lower than) their share in imports from the rest of the world. However, intermediate goods and services tend to be more important in the UK’s exports to the EU than in its exports to other destinations. The share of the EU’s imports from the UK that were intermediate goods and services was 69% in 2014 (having increased from 61% in 2000).

In total, 9.3% of the UK’s inputs are sourced from the EU. However, the degree to which different UK industries make use of inputs imported from the EU varies greatly (Figure 2). They are most important for UK manufacturers, who obtain 16% of their inputs from the EU, followed by healthcare and agriculture, which obtain 13% and 11% of their inputs from the EU respectively. Inputs from the EU are, perhaps unsurprisingly, less important for service industries such as real estate. Within the manufacturing sector, the industries that make most use of EU inputs are motor vehicles, pharmaceuticals, chemicals and electronics – incidentally, industries which also tend to export relatively more of their output to the EU.

Looking instead at the importance of UK inputs for industries in the rest of the EU, the first thing to note is that the UK is a much less important source of inputs for the EU than vice versa. For example, manufacturing firms in the rest of the EU only obtain 1.5% of their inputs from the UK. One industry in which UK inputs are noticeably more important than others, however, is financial intermediation: 3.3% of inputs to the rest of the EU’s financial intermediation sector come from the UK. The total proportion of the EU’s inputs sourced from the UK is 1.4%.

Figure 2. Share of inputs imported from the EU (for UK) or the UK (for EU) by industry, 2014

Figure 2. Share of inputs imported from the EU (for UK) or the UK (for EU) by industry, 2014

Source: Author’s calculations from World Input-Output Tables, 2016 edition.

These facts have a number of immediate implications for the way the UK should think about its future trade policy, in terms of its trading agreements with the EU and with the rest of the world.

1) The increasingly interconnected nature of global trade means that a country’s imports and exports cannot be treated as independent quantities. A successful exporting country will also need to be open to imports. Exports embed imports, and so greater access to imports can boost the competitiveness and export performance of domestic firms. For example, according to the OECD TiVA database, 9.5% of the value-added embedded in the UK’s gross exports in 2011 was produced in the EU and a further 13.5% was produced in the rest of the world. These figures represented increases from values of 8.2% and 9.6% respectively in 2000. As a result, UK firms’ access to imports as well as export markets is an increasingly important consideration for future trade policy. The UK should take this into account before taking actions that would introduce or maintain tariffs or other trade barriers on its imports from the EU or third countries.

2) Demand for the exports of UK industry depends not only on the export access of UK-based firms but also on the export access of firms they supply. Since the UK is a relatively important supplier to EU firms, the trade deals the EU signs will continue to have relevance for the UK in the coming years, whether or not the UK leaves the customs union. This includes, of course, the access the EU has to the UK market.

3) The importance of international trading networks means that bilateral trade deals will tend to be of less value than multilateral deals. This is partly because of the importance of rules of origin requirements, which can create a complex ‘spaghetti bowl’ of overlapping agreements which firms involved in international supply chains must navigate in order to benefit from bilateral trade agreements.

For example, in the EU-Korea free trade agreement, a good exported from the EU is deemed to have ‘originated’ in the EU only if less than 45% of the value of inputs has been imported from outside Korea or the EU. At present, UK imports from other EU countries therefore do not count as ‘foreign’ when determining a product’s origin and so do not limit firms’ ability to sell their goods to Korea. Equally, EU firms can freely use components manufactured in the UK. Outside the EU, the UK may well be considered a third party in such trade agreements and hence EU firms may not be able to use too many UK components in exports to Korea. Since foreign components are an important input for UK manufacturers, many firms may not be able to benefit from bilateral deals the UK signs unless the UK can get its partners to agree to less restrictive rules of origin requirements (and there may be a price to pay for this). The more countries that are included in each deal, the less of a problem this will be.

A final question is whether the current importance of inputs from the EU is likely to change following Brexit. If the EU’s own tariffs exclude competition from the rest of the world, then selective post-Brexit tariff reductions on imported goods could help to improve the competitiveness of UK firms.

This effect is likely to be small though. First, tariffs on the sort of intermediate goods the UK purchases from the EU (and indeed the rest of the world) already tend to be lower on average than those levied on goods used for final consumption – about 4% on intermediate goods compared with nearly 10% on goods imported for final consumption. These figures would be even lower if we took into account services inputs that essentially attract no tariff. Second, tariff reductions will not prevent the introduction of non-tariff barriers such as customs checks that would follow if the UK left the EU’s customs union and single market. These are especially important considerations for industries that value timeliness and flexibility in their supply chains (such as the car industry), and would be difficult to negotiate away in deals with third countries.


This analysis was funded by the Economic and Social Research Council’s ‘The UK in a Changing Europe’ Initiative and also appears as a blog on their website.

]]> Mon, 08 Jan 2018 00:00:00 +0000
<![CDATA[Will the rising minimum wage lead to more low-paid jobs being automated?]]> Summary

The minimum wage is rising rapidly. In 2015 4% of employees aged 25 and over were on the minimum; this is set to reach 12% by 2020. New analysis published today by the IFS shows that those being brought within the minimum wage net are in different sorts of jobs to those who have been on the minimum previously – and in particular they are more likely to be doing jobs that appear more readily doable by machines or computers. For example, many of those on the minimum wage in 2015 were in personal service occupations, such as workers in the hospitality sector. Those jobs are not readily done by machines. But those set to be brought within the minimum wage net in 2020 are more than twice as likely to be in the ten per cent most “routine” occupations – such as retail cashiers and receptionists – as those who were directly affected by the minimum in 2015. This kind of work tends to be easier to “automate”. Ease of automation actually rises even more as one looks somewhat further up the wage distribution, reaching a peak about a quarter of the way up.

The fact that there seemed to be a negligible employment impact of a minimum at £6.70 per hour – the 2015 rate – does not mean that the same will be true of the rate of over £8.50 per hour that is set to apply in 2020. Beyond some point, a higher minimum must start affecting employment, and we do not know where that point is. The fact that the higher minimum will increasingly affect jobs that appear to be more automatable is an additional reason why extremely careful monitoring is required. Meanwhile even higher rates, as proposed for example by the Labour party, would bring even more employees in more automatable jobs into the minimum wage net. Here we document the facts on the ease of automation of work at different wage levels, discuss the implications, and discuss what more we need to know to make minimum wage policy effectively in light of this.

Full version

For employees aged 25 or over, the legal minimum hourly wage – now termed the ‘national living wage’ (NLW) – has risen from £6.70 in 2015 to £7.50 now, will rise further to £7.83 in April, and is planned to reach 60% of median wages in 2020 – which, under current forecasts of wage growth, would be £8.56. More importantly, because this is a much faster rate of growth than for wages in general, this would mean that the fraction of employees aged 25+ subject to the minimum would have trebled in just five years, from 4% in 2015 to 12% in 2020.

As we have said before, there is a case for a higher minimum wage as a tool that the government has to help those on low wages – particularly given that the rises in the minimum up to 2015 do not seem to have had significant adverse effects. However, it does need to be raised with care and according to a rigorous evidence-based process, rather than subject to a political bidding war based on nice-sounding numbers. This follows simply from two facts: first, beyond some point, a higher minimum wage will have adverse consequences for those that it is designed to help; and second, we do not know where that point is. 

The first fact should be easy to see. A £100 per hour minimum wage would make most employees too expensive to employ and, hence, lead to them being unemployed. The important question is at what point before we get to that extreme case do the significant negative effects on jobs (and/or hours of work) start to materialise. Evidence on employment effects to date suggests we had not got to that point by 2015 – but we would get there at some stage if the minimum wage continued to move up the wage distribution.

There is another reason why previous evidence is of limited value when contemplating a trebling of the numbers of employees covered by the minimum: the jobs affected by a higher minimum wage could be quite different from those already affected by it. One way in which this could be the case is if they can be more or less easily substituted with ‘capital’ (e.g. machinery or computers) – or ‘automated’.

Current technologies are relatively effective at executing ‘routine’ tasks – such as calculating supermarket bills and collecting payment – and less effective at tasks that involve forward planning or responding to unusual situations or that require human interaction. The relationship between wages and the importance of routine tasks is not as obvious as one might imagine. Some low-paid jobs actually involve a lot of non-routine work, such as waiters or nursery attendants supervising children; while routine tasks are relatively prominent within some mid-paying occupations, such as bank clerks. Without looking at the data, it is unclear whether those workers brought onto the new, higher minimum wage will be more or less easily replaced with machines than those on the very lowest wages who are already on the minimum wage.

To investigate this, we need to measure how easy it is for employers to replace employees with machines. We use a measure that has been adopted in academic research: the ‘routine task intensity’ (RTI) of each occupation. The logic is that, as discussed above, more routine tasks tend to be more readily performed by technology. All else equal, under this measure, occupations are classed as more routine to the extent that they involve ‘working to set limits and standards’ and ‘finger dexterity’. They are classed as less routine to the extent that they involve ‘undertaking direction, control and planning’, ‘hand-eye-foot coordination’ or ‘mathematical and formal reasoning’. While RTI measures are only proxies for ‘automatability’, research has found that they are both strong predictors of past technological adoption and capable of explaining labour market trends in a number of advanced economies.

The figure shows that, as the minimum wage is raised, we are increasingly affecting workers in occupations that appear easier to automate (i.e. to replace with machines). For example, the workers set to be brought within the minimum wage net in 2020 are more than twice as likely to be in the top 10 per cent most routine occupations – such as retail cashiers and receptionists – as those already paid the minimum in 2015 (11% compared with 5%). It is notable that the increases in the minimum wage that have already occurred over the past two years account for much of this – they have resulted in an especially steep rise in the ‘automatability’ of affected occupations. We still await much in the way of reliable quantitative evidence on the impacts of those recent increases, as stated by the Low Pay Commission. In any case, substitution of workers with technology would, if it occurred, be unlikely to happen immediately.

The figure also shows that ‘automatability’ does not peak until after the 25th percentile of the hourly wage distribution. This means that employees affected by further rises beyond those planned appear to be doing even more readily automatable jobs, on average, than the ones already set to be covered. This would apply, for example, to the £10 per hour minimum in 2020 promised in Labour’s manifesto, which would cut at around the 22nd percentile of the distribution.

This is not the same as saying we know there will be big adverse consequences for low-skilled workers of the planned rises in the minimum wage. The reality is more uncertain – we do not know what the employment effects will be – but also more nuanced than that. First, as with other causes of job loss, it would matter greatly whether anyone who did lose out from automation quickly found adequate re-employment. Moreover, the use of technology to replace some jobs can create new jobs that are complementary to that technology (e.g. people to service machinery). Research in the US has found some negative impacts of higher minimum wages on the employment of low-skilled people in automatable occupations whilst also finding evidence of concurrent employment gains among other groups, with mixed evidence on whether the new jobs that were created were also filled by low-skilled workers or instead were likely to be benefiting those with higher skills. Some workers may themselves be ‘upskilled’ as a result of the higher minimum wage (e.g. through training), and hence remain in employment but end up doing a somewhat different job from the one they did before (or the one that they would have gone on to do).

There is much more to learn here, and the recent and upcoming policy changes in the UK offer an opportunity to do so. As new evidence comes in, it is crucial that policymakers seeking to help low-skilled workers respond accordingly.

Percentage of employees aged 25+ in the most automatable jobs (top 10% of routine task intensity)

Note: Wage distribution based on hourly earnings including overtime in April 2015, for employees aged 25 or over, excluding those whose pay in the reporting period was affected by absence. We use 2015 data so that the position of different employees in the distribution is not itself affected by the minimum wage hikes that occurred from 2016. We use the measure of routine task intensity defined in D. Autor and D. Dorn, ‘The growth of low-skill service jobs and the polarization of the US labor market’, American Economic Review, 2013, 103(5), 1553–97.   
Source: Authors’ calculations using 2015 Annual Survey of Hours and Earnings.

]]> Thu, 04 Jan 2018 00:00:00 +0000
<![CDATA[A bigger nudge: the Government’s proposed extension to automatic enrolment]]> The Department for Work and Pensions today published the recommendations of the Government’s review into automatic enrolment. This focussed on issues around membership of, contributions to, and engagement with, workplace pensions. Over the last year I served on an expert advisory group to the review. In this piece I look at two of the recommendations – both of which are likely to boost the amounts going into workplace pensions – and the proposed trials of ways to boost pension saving among the self-employed who are not covered by automatic enrolment.

Automatic enrolment before the review

Prior to today’s review automatic enrolment worked (roughly) as follows. Employers had to enrol employees aged between 22 and their State Pension Age (SPA) who earned over £10,000 into a workplace pension. Contributions to the pension had to be a minimum of 2% of earnings between the Lower Earnings Limit (LEL, £5,876 in 2017–18) and an upper limit (£45,000), with at least 1% of this coming from the employer. These minimum contribution rates are to rise to 5% (with at least 2% coming from the employer) in April 2018 and to 8% (with at least 3% coming from the employer) in April 2019. Once enrolled employees can then choose to leave the scheme if they want. If they leave, they would not have to make any employee contribution, but they would also lose the employer contribution.

Employees who earn above the LEL but who are not automatically enrolled because they are aged 16 to 22, aged over the SPA (but under age 75), or because they earn under £10,000 can actively choose to join a workplace pension. If they do they will also benefit from an employer contribution of at least 3% of earnings above the LEL. Those earning below the LEL can also choose to join a workplace pension if they wish, but their employer is not obliged to make any contribution.

Previous IFS research, published last year, has shown that so far automatic enrolment has boosted workplace pension membership substantially, with particularly large increases in membership among lower earners and younger individuals.

Proposed reforms to automatic enrolment

The Government is proposing important changes to these parameters. First, minimum contributions will be based on all earnings up to £45,000 rather than starting at the LEL. Second, the lower age limit is to be reduced from 22 to 18. The plan is for both of these changes are to happen in the mid 2020s. This timetable is to ensure that the increase in employer obligations occur after the current period where default minimum employer contributions are rising (from 1% to 3% of band earnings) and the National Living Wage is being increased faster than average earnings.

Removing the lower earnings threshold would mean that the default minimum contributions of those earning above the LEL who are in a workplace pension rises from 8% of earnings above the LEL to 8% of all earnings (up to £45,000). In addition those earning below the LEL (who are aged between 16 and 74) would, if they choose to join a workplace pension, receive an employer contribution.

The impact of this change on default minimum contributions is shown in the Figure below. The solid lines show the total default minimum contributions (from employee and employer) before and after today’s proposal for those who earn over £10,000 and are therefore automatically enrolled. The dotted line shows the minimum contributions for those who earn below £10,000 (and therefore might not be automatically enrolled) but who actively choose to join a workplace pension.

Figure. Minimum total contributions to workplace pensions, before and after today’s proposals

For all employees earning above the LEL and only making minimum contributions the increase is worth £470 per year, of which at least £176 per year would be in employer contributions. While a relatively small amount it represents a large increase in the default minimum contribution of lower earners. For example an individual earning £10,000 would see their default minimum contribution more than double from £330 per year to £800 per year (from 3.3% of their earnings to 8.0% of their earnings). In addition the removal of the LEL would be of particular benefit to employees with multiple jobs, since they will now be able to receive employer contributions worth at least 3% of their total earnings, whereas under policy prior to today their employers might each have contributed 3% of their earnings in excess of the LEL in each of their jobs. In addition to increasing pension contributions this reform would also be a welcome simplification and reduce the risk that some employees mistakenly think that they are contributing 8% of their earnings to a workplace pension when in fact they are only contributing 8% of earnings above the LEL. (Of course those earning above £45,000 will have contributions of 8% up to £45,000 rather than 8% of all earnings.)

The reduction in the minimum age from 22 to 18 would also be likely to boost pension contributions. While those aged under 22 have seen a sizeable increase in membership of workplace pensions as a result of automatic enrolment they are still less likely to be a member of a plan than older individuals. This also feels a welcome move: in fact there is a case that the Government should have gone further and removed the age limits altogether, which would have led to all employees aged 16 to 74 earning above £10,000 a year being automatically enrolled into a workplace pension by their employer.

Taking the removal of the lower-earnings threshold and the lowering of the lower age limit from 22 to 18, the DWP estimates that this would boost contributions to workplace pensions by £3.8 billion, with £1.8 billion coming from employees, £1.4 billion from employers and £0.6 billion from up-front tax-relief. This compares to the DWP’s estimate of an additional £19.7 billion being contributed to workplace pensions as a result of automatic enrolment under existing policy in 2020. Who actually bares the cost of these contributions would be much more difficult to know: for example employers might seek to cover the cost of their contributions by slowing wage growth over coming years (as the Office for Budget Responsibility assumes in its forecasts) or by pushing up their prices.

Automatic enrolment for the self employed?

Finally the self-employed are not covered by automatic enrolment. The review does not propose a straight extension of the policy to them. This is sensible. The self-employed do not, by definition, have an employer and therefore forcing them to enrol themselves in a pension that they can subsequently choose to leave would seem odd.

Instead the Government is proposing that, in the light of the emerging lessons from automatic enrolment for employees, alternative reforms that use similar mechanisms are trialled. One leading contender would be through the tax return process. When the self-employed file their tax return they could be informed about the amount of pension saving an employee with a similar amount of income would, by default, do. This – perhaps alongside a simple process for making a pension contribution at this point – could help nudge them into a greater level of retirement saving. But we do not know how effective this would be and therefore a test and learn approach is appropriate.

One for the future

One key question remains unanswered: when minimum default contribution rates have risen to 8% of earnings should they stay there or should they be increased even further? The Government has appropriately held off making recommendations on this until evidence is available on what happens when the contribution rates rise from their current low levels. But unless the increase in minimum contributions, from 2% to 8%, that are scheduled occur over the next two years lead to a significant increase in the numbers of employees choosing to opt out of their workplace pension there will be a strong case for further increases. By the mid-2020s the nudge could be even bigger.

Carl Emmerson, Deputy Director Institute for Fiscal Studies, and member of the expert advisory group to the DWP’s review of automatic enrolment

]]> Mon, 18 Dec 2017 00:00:00 +0000
<![CDATA[Six key charts ahead of the Chancellor's Autumn Budget]]> The Government is spending around £50 billion more than it raises in revenue

Figure 1a. Government spending, 2016–17 (£771 billion)

Figure 1a. Government spending, 2016–17 (£771 billion)

Figure 1b. Government receipts, 2016–17 (£726 billion)

Figure 1b. Government receipts, 2016–17 (£726 billion)

Figures 1a and 1b show how the public sector spent and raised money last year. The largest single areas of spending are Social Security and Health, while three taxes (Income tax, National Insurance Contributions and VAT) account for 60% of government revenues.

Not unusually, the amount that the government spent last year was larger than the revenue generated. The difference between these two numbers is government borrowing, or the deficit.

Borrowing has returned to its pre-crisis share of national income and is set to fall further

Figure 2. Government borrowing since 2000–01

Figure 2. Government borrowing since 2000–01

Figure 2 shows how the deficit has evolved since the year 2000 (the last year in which the budget was in surplus), and how it was expected to evolve at the time of the last budget in March. The deficit increased dramatically between 2007–08 and 2009–10 during the great recession. After 7 years of “austerity” (spending cuts and tax rises), borrowing is now back to its pre-crisis level.

But current plans imply further spending cuts and tax rises over the next few years to reduce the deficit further, and the government aims to eliminate the deficit entirely by the mid-2020s. This is an ambitious target and is unlikely to be easy.

A likely downgrade to the outlook for productivity growth in the budget will push up forecast borrowing and mean the Chancellor has little room for budget giveaways

Figure 3. Output per hour annual growth since 1972

Figure 3. Output per hour annual growth since 1972

Figure 3 shows annual output per hour (productivity growth) since 1972, including the forecast for the next five years from March.

The most important driver of the government’s budget position is the economy – it was a reduction in the size of the economy (meaning, for example, lower wages), and a resulting reduction in tax revenues, that led to the deficit increasing dramatically in the late 2000s.

The key driver of economic growth is productivity growth, which has been historically terrible over the last 9 years and shows little sign of improvement. The forecast for productivity this time around is likely to be more pessimistic than in March – that means slower wage growth, slower economic growth and lower tax receipts which will mean higher borrowing.

Figure 4. Government borrowing under different growth scenarios

Figure 4. Government borrowing under different growth scenarios

A worse outlook for productivity is bad news for all of us, but it puts the Chancellor in a particularly tricky position. A significant productivity downgrade (represented by the ‘Weak’ and ‘Very Poor’ scenarios in Figure 4) would push up the borrowing forecast substantially. Given this it would be extremely difficult for the Chancellor to offer spending giveaways while still remaining committed to his deficit reduction targets. 

On current plans, the spending of government departments will be squeezed further over the next two years

Figure 5. Real-terms departmental budget changes, 2010–11 to 2019–20

Figure 5. Real-terms departmental budget changes, 2010–11 to 2019–20

Figure 5 shows the real terms (i.e. accounting for inflation) change in the budgets of different government departments over the next two years and for the 2010s as a whole.

The Chancellor is facing calls to loosen spending in a number of areas, and based on this picture its easy to understand why. Almost all departments have faced tight settlements since 2010. Even departments that have been relatively protected from real terms cuts, such as the Department of Health, have experienced much smaller increases than the historical norm. Yet other departments, such as DEFRA or Justice, are set to see cuts of almost 40% over 10 years, with further cuts to be delivered over the next two years.

A post-election budget is normally a time for tax rises, which could be used to finance looser spending plans, but this is likely to be difficult for a minority government

Figure 6. Long-run net tax rise from measures announced in the year following elections, 2017–18 terms

Figure 6. Long-run net tax rise from measures announced in the year following elections, 2017–18 terms

Figure 6 shows the net effect of tax rises announced in the year following every general election since 1992.

If the Chancellor wanted to ease the squeeze on spending while remaining on course to eliminate the deficit by the mid-2020s, one option might be to raise taxes. Indeed, this is what every Chancellor has chosen to do following elections since 1992. However, this may not be a feasible option this time. None of those previous governments were minority governments, and the parliamentary arithmetic this time around is likely to make it difficult to pass substantial tax-raising measures.


Figures 1a and 1b

Notes: ‘Government spending’ refers to Total Managed Expenditure. ‘Government receipts’ refers to Public Sector Current Receipts.

 ‘Other indirect taxes’ includes alcohol duties, tobacco duties, betting and gaming duties, air passenger duty, insurance premium tax, landfill tax, climate change levy, vehicle excise duties and soft drinks industry levy. ‘Corporation taxes’ includes corporation tax, petroleum revenue tax, oil royalties, bank surcharge, bank levy and diverted profits tax. ‘Property taxes’ includes council tax and business rates. ‘Capital taxes’ includes stamp duties, capital gains tax and inheritance tax. ‘Other taxes’ is a residual measure, including devolved taxes and environmental levies, which are generally part of government schemes that translate higher revenues directly into higher spending. Non-tax receipts include interest and dividend income and the operating surplus of publicly owned corporations

Sources: Office for Budget Responsibility, Public Finances Databank, October 2017 (; Office for Budget Responsibility, Economic and Fiscal Outlook, March 2017 (; HM Treasury, Public Expenditure Statistical Analyses, July 2017 (

Figure 2

Note: Figure shows Public Sector Net Borrowing

Source: Office for Budget Responsibility, Public Finances Databank, October 2017 (; Office for Budget Responsibility, Economic and Fiscal Outlook, March 2017 (

Figure 3

Note: Black lines refer to average growth rates in the relevant periods (1972–2007, 2010–2016, 2017–2021).

Source: Office for National Statistics series LZVB; Office for Budget Responsibility, Economic and Fiscal Outlook, March 2017 (

Figure 4

Note: ‘Moderate’ scenario assumes OBR downgrades growth in line with the Bank of England’s August Forecast. ‘Very poor’ scenario assumes that productivity growth is only 0.4% per year. ‘Weak’ scenario assumes OBR downgrades halfway towards the ‘very poor’ scenario (productivity growth of 1% per year on average).

Source: C. Emmerson and T. Pope Autumn 2017 Budget: Options for Easing the Squeeze, IFS Report 133 (

Figure 5

Source: HM Treasury, Public Expenditure Statistical Analyses, July 2017 (

Figure 6

Note: Colour of bar refers to party or parties in government after the election. All figures based on the cost of the measure in the final year on the scorecard.

Source: Office for Budget Responsibility, Policy Measures Database (

]]> Fri, 17 Nov 2017 00:00:00 +0000
<![CDATA[Is our tax system fair? It depends...]]> The basic question of whether our tax system is fair is at the heart of many of our public debates. Discussions of whether ‘the rich’ or companies are paying their ‘fair share’ is regularly in, or underlying, the news headlines.

These are important questions. If we want to ensure that we can raise the revenues to pay for the public goods and services that we all want, we need to be able to have sensible debates about how much tax we raise, who we raise it from and how we spend it. But tax fairness is a complex issue. It’s hard to debate because people will differ in what they think is fair. There isn’t even a single concept of tax fairness – there are many, each of which would support a different tax design.

On top of this, discussing tax fairness is tricky because people have access to different information and this matters a lot in shaping their judgements.

The Experiment

We recently ran a straw poll to demonstrate the power of information. We asked a simple question: Broadly, do you think the UK tax system is fair? In many ways this is a silly question; so broad that it’s hard to know what use the results could have. But before answering participants were randomised into one of three groups

Group 1 were just given the question.

Group 2 (the ‘rich pay a lot’ group) also received these two true statistics:

  • The point at which income tax starts to be paid has increased in recent years. 4 in 10 adults now pay no income tax.
  • The income tax system is top-heavy. The top 10% of income taxpayers pay 60% of all income tax. 

Group 3 (the ‘rich don’t pay a lot’ group) also received these two true statistics:

  • The richest 10% of income taxpayers earn more income than the entire bottom 50%.
  • Someone earning £45,000 faces the same income tax on an extra £1 of earnings as someone earning £145,000.

The Results

The results from this poll were stark. Before any information, 51% of respondents said that the tax system was unfair because the rich paid too little. In the ‘rich pay a lot group’, the proportion saying the system is unfair because the rich pay too little dropped to 33%. In the ‘rich don’t pay a lot group’, this proportion jumped to 72%. This means that about a fifth of people had their answers changed by two (true) statistics. Comparing groups two and three shows that small amount of information led to a doubling of the proportion of people who thought the system unfair.

Information also changed the proportion of people who thought the system was fair, from 16% before any information, to 25% in the ‘rich pay a lot’ group and to 8% in the ‘rich don’t pay a lot’ group. Proportionately, these are even bigger changes.

What we think about fairness depends on the information at our finger tips

This isn’t hard science – it’s a straw poll based on a small, unrepresentative sample of people who choose to click on our survey.  But the results do help to illustrate the fact that small amounts of information can radically shape people’s stated views. And the idea that information matters is backed up by other research. For example, a survey of 7,700 Vox readers conducted by the US Tax Policy Centre earlier this year also found that information - in this case provided through a quiz – changed people’s perceptions of tax fairness. A research study based in the UK found evidence that how information is framed – in particular whether statistics are given as percentages or in terms of absolute amounts of money – affects stated preferences of how progressive taxes should be.

These effects matter. People will often be exposed to small chunks of information about an issue and, as we continue to get more media tailored to our tastes, we might expect to also get exposed to more statistics.  If we are to have a successful debate about what a fair tax system looks like, or indeed about any other important issue in our society, we need to start with well-formed questions that are supported by the best available evidence.

Communicating economics

This poll was conducted for use in a public talk I gave at Manchester University. The talk addressed the question: Are the rich paying their fair share of taxes?

Organised with Manchester University and funded by the Economic and Social Research Council (ESRC), this was the first in a new series of public talks that IFS are using to ensure that as many people as possible have the tools and information needed to engage in some of society's important debates. That's part of the mission of IFS – we’re an independent, politically impartial research institute with the principal aim of better informing public debate. 

Future talks will tackle issues relating to an ageing society, income inequality and education. To keep in touch or find out more about our work, please follow us on twitter @TheIFS or on facebook at ‘The Institute for Fiscal Studies’.


129 people completed the poll. Results are available in the reference list for the related public talk (here)

IFS produces video explainers on key issues like the public finances, taxation, health spending and trends in inequality. You can see these and some recordings of our events on our YouTube channel (here)

]]> Fri, 03 Nov 2017 00:00:00 +0000
<![CDATA[Higher inflation means more pain for households from benefit freeze, less gain from £12,500 personal allowance]]> This morning the Office for National Statistics announced that inflation in the year to September was 3.0%. Normally the September inflation figure is used to uprate benefit levels and tax thresholds the following April. However, current government policy is to freeze most working-age benefits in cash terms until March 2020. Combined with the latest inflation forecasts, today’s number means that the 4-year freeze is now expected to reduce entitlements in 2019–20 by an average of £450 per year for the 10.5 million households affected (saving the Exchequer £4.6bn). When the policy was first announced, the expected average loss among the losing households was £320 per year (saving the Exchequer £3.4bn). The £130 difference is because inflation over the 4 years is now expected to come in higher than was anticipated at the time. For the same reason the Conservative commitment to a £12,500 personal allowance is less of a giveaway than expected when it was announced: £12,500 of income tax-free cash will not go as far given higher prices. This illustrates a problem with setting benefit rates and tax thresholds in cash terms several years in advance – unexpected movements in inflation will cause the generosity of the tax and benefit system to differ from what the government intended.

Usual practice is for tax thresholds and benefit rates to be uprated each April in line with the previous September’s inflation rate. However, the government’s benefits freeze means that most working-age benefit rates will be unchanged in cash terms between April 2015 and March 2020. The Conservative Party manifesto committed to a £12,500 income tax personal allowance and a £50,000 higher rate threshold by 2020, regardless of what happens to inflation. In other words, when inflation proves higher or lower than expected, the real value of benefits and income tax thresholds also changes. This observation discusses the consequences.

The government announced the benefits freeze – which affects most working-age benefits other than disability benefits – in Summer Budget 2015. The first two years of the freeze have only reduced the real value of affected benefits by 1.0% in total, as inflation has been low. However, inflation has since risen as the depreciation of sterling in the wake of the EU referendum result feeds through into higher consumer prices. Today’s inflation number, together with the Bank of England’s latest forecast for what inflation will be next September, indicate that the next two years of the freeze will represent a further 5.7% real reduction. Hence, taking the 4-years as a whole, the freeze is now expected to reduce the real value of benefits by 6.7%.

By the end of this 4-year benefits freeze in 2019–20, it will have reduced benefit entitlements by £4.6 billion per year, resulting in entitlements being on average £450 lower than they would otherwise have been for around 10.5 million households. Figure 1 shows how these losses are spread across the income distribution (which includes households unaffected by the freeze). Not surprisingly, the largest impact is toward the bottom of the distribution, where households are more dependent upon benefits for their income. Some small losses are seen even in the highest income deciles, largely because the freeze includes child benefit, which is available to many higher income households. Excluding households who only lose out from the cut to child benefit, 7.5 million households are affected, losing an average of £590 per year.

Figure 1. Distributional impact of the benefits freeze between 2015-16 and 2019-20

Figure 1. Distributional impact of the benefits freeze between 2015-16 and 2019-20

Notes: Assumes full take-up of means-tested benefits and tax credits, and that all claimants are unaffected by the changes made in April 2017 for new benefit claimants. The figure shows the average effect of the policy across all households in each decile, including those unaffected by the reform.
Source: Authors’ calculations and the IFS tax and benefit model, TAXBEN, using data from Family Resources Survey, 2015–16.

The 2015 and 2017 Conservative manifestos included commitments to increase the income tax personal allowance and higher rate threshold (HRT) – both normally uprated with inflation – to £12,500 and £50,000 respectively by 2020–21 at the latest. Because these are cash terms targets, this is a less generous (and less expensive) commitment the higher inflation turns out to be – that £12,500 of income-tax free cash will not go as far if prices have risen faster in the meantime. Had the personal allowance simply been uprated as usual in line with inflation since 2016, it would, on current inflation forecasts, have reached £11,580 by 2020–21. Relative to that, the increase to £12,500 would save a basic rate taxpayer £184 per year. However, much of the progress towards £12,500 has already been made:  in 2017–18, the personal allowance stands at £11,500, and the higher-than-expected inflation that we have seen over the past year, combined with higher expected future inflation, means that continuing to increase the personal allowance in line with inflation from now on would be expected to leave it at £12,430 in 2020–21. In other words, the commitment to £12,500 in 2020–21 would represent a further giveaway of only £14 per year for most basic rate taxpayers. (That said, because 29 million individuals would gain, even this small giveaway would still cost the government £0.5 billion a year.)

The commitment to a £50,000 higher-rate threshold (HRT) would cost the government a similar amount to the personal allowance rise, but with larger gains going to far fewer individuals. Increasing the HRT in line with inflation from now on would leave it at £48,610 in 2020–21, meaning an increase to £50,000 would save over 4 million higher rate taxpayers £139 per year. Individuals who benefit from the increase in the HRT are in the highest income 8% of adults, and 85% of the gains would go to households in the top fifth of the income distribution.

The benefits freeze and income tax policies are both examples of the practice of setting tax thresholds and benefit rates in cash terms many years in advance, regardless of what inflation in the intervening years turns out to be. The same habit can be seen in the (disintegrating) public sector pay cap, which was to limit increases in public sector pay scales to 1% for four years. This approach to policymaking means that, when what actually happens to inflation differs from what was forecast, the real changes in taxes and benefits also differ – sometimes substantially – from what the government initially intended. It is worth remembering that there are also income tax thresholds for which the policy is to freeze them permanently, implying that their real value will typically fall every year, such as the £100,000 point at which the personal allowance begins to be withdrawn and the £150,000 point at which the 45% rate becomes payable. The £50,000 point at which child benefit starts to be withdrawn is another example of this unwelcome practice.

Figure 2A plots the total fall in the real value of affected benefit rates over the four years of the benefits freeze, on the basis of the outturns and forecasts for inflation at each fiscal event since the policy was first announced. Figure 2B shows the gain for a basic rate taxpayer that the increase in the personal allowance to £12,500 in 2020–21 represents, relative to simply uprating the 2015 personal allowance in line with inflation. As the figures make clear, the expected sizes of the benefit takeaway and personal allowance giveaway have changed over time. When announced, the benefit freeze was expected to reduce the real value of affected benefits by 4.8% in total over the 4 years. At the same time, the raising of the personal allowance to £12,500 was expected to deliver a saving to basic rate taxpayers of £232 per year in 2020–21. Since sterling’s devaluation in the latter half of 2016, inflation and inflation expectations have sharply increased. This means that the benefits freeze is now expected to entail a 6.7% total cut in the real value of affected benefits, and the gain to basic rate taxpayers from a £12,500 personal allowance – relative to a 2015 allowance uprated with inflation – has been reduced to £184 in 2020–21.

While reasonable arguments can always be made about how high benefits or tax thresholds should be – and what path should be taken to get them there – it is difficult to see why their real values should be buffeted around by unexpected moves in inflation. But this is the inevitable result of setting policies in cash terms several years into the future: the government is allowing the future generosity of the tax and benefits system to different groups to be impacted not only by its deliberate choices, but also on the roll of the dice that is how inflation outturns differ from forecast. It would be a good habit to kick.

Figure 2A. Real change in the value of benefits from the 4-year benefit freeze implied by successive inflation forecasts

Figure 2A. Real change in the value of benefits from the 4-year benefit freeze implied by successive inflation forecasts

Figure 2B. Gain for a basic rate taxpayer from raising the personal allowance from £10,600 in 2015 to £12,500 in 2020 implied by successive inflation forecasts

Figure 2B. Gain for a basic rate taxpayer from raising the personal allowance from £10,600 in 2015 to £12,500 in 2020 implied by successive inflation forecasts

*The October 2017 values are based upon the Bank of England’s latest forecast for inflation.
Notes: With the exception of October 2017, values are based upon the OBR’s inflation forecast at the fiscal event in question.
Sources: Office for Budget Responsibility, Economic and Fiscal Outlook, various editions.
Bank of England, Inflation Report, August 2017

Tom Waters, IFS research economist and one of the authors of the report, said:

“This morning’s inflation figure, taken together with the latest inflation forecasts, means that the 4-year freeze on most working-age benefits is now expected to cut the benefits of 10 million families by £450 a year in real terms – up from £320 back when the freeze was first announced. The extra £130 loss is not the result of any deliberate decision by the government – it is the consequence of inflation being higher than was expected when the policy was set. This illustrates a problem of setting benefit rates far in advance in cash terms – it leaves the actual generosity of future benefits sensitive not only to the government’s active decisions but also to unexpected moves in inflation. It means that the risk of higher inflation – a risk that has now materialised – is being borne by working age benefit recipients.”

]]> Tue, 17 Oct 2017 00:00:00 +0000
<![CDATA[School Funding Reform in England: a smaller step towards a more sensible system, will the final leap ever be made?]]> Last week, the Secretary of State for Education announced arrangements for school funding in England in 2018–19 and 2019–20. This confirmed additional annual funding of around £900m by 2019–20 (as compared with pre-election plans) and announced the amended plans for the national funding formula. Under these new proposals, the funding local authorities receive for schools will be linked to local area characteristics; however, a new national school-level formula will now not be in place until at least 2020–21. This is a smaller step than planned prior to the election – although still one in the right direction. The slower pace of reform and additional money also mean that schools losing out under previous plans will probably see their funding situation improve slightly. This observation describes the current system, why reform is needed and the likely effects of the latest proposals.

Current school funding system

Currently, central government allocates a block of funding to English local authorities which they allocate to all state schools in their area using their own funding formulae. This results in a wide variation in funding per pupil across schools in England. A lot of this is intentional: schools in London receive more to cover higher staff costs and successive governments have deliberately allocated more funding to schools in more deprived areas. However, some of the variation is unintentional; many schools with similar characteristics in different parts of the country receive different levels of funding.

These unintentional differences arise for two reasons.

  • First, similar local authorities can receive quite different levels of funding per pupil from central government. This is because the amount going to each local authority is largely determined by their characteristics in the early 2000s and local authorities have changed a lot since then. For example, Plymouth and Bradford have similar characteristics (e.g. both have about 17% of pupils eligible for free school meals), but funding per pupil is around £500 higher in Bradford than it is in Plymouth. This is largely because back in 2004 there was a significant gap, with about 24% of pupils eligible for free school meals in Bradford and 16% in Plymouth in 2004.
  • Second, schools in different local authorities can also receive different levels of funding per pupil because local authorities make different choices over how to allocate their school funding. For example, in Norfolk, the basic amount provided for a pupil aged 11-14 is about 20% higher than for a pupil in primary school. In Darlington, it’s about 70% higher. As a result, similar secondary schools would likely receive higher funding in Darlington than Norfolk, while the opposite would be true of similar primary schools. This type of variation may be desirable; it may reflect differences in local preferences or local policymakers responding to information on the ground.

Proposals for reform

Before the 2017 general election, the government set out ambitious plans for a national funding formula for schools in England. This would replace all 152 different local authority funding formulae with one single national funding formula applying to all state-funded schools. The intention was to eradicate the differences in funding levels between apparently similar schools and, in doing so, remove the role of local authorities in allocating school funding. Sensibly, given the scale of the change, in 2018–19, local authorities would still play some role, with the national school-level formula not fully kicking in until 2019–20. There were also caps on the gains and losses schools could experience. Such protections were desirable to prevent schools from seeing large changes in funding over a single year, but doing so slows the transition to the new formula. Around 40% of schools would have still had funding levels that reflected historical factors in 2019-20, including a quarter of schools that were “overfunded” relative to the formula. The government said very little at the time about what would happen after 2019–20 and this remains a major source of uncertainty.

Under the new proposals for a national funding formula set out last week, a number of factors changed. The most important change is that the national funding formula will no longer be fully implemented until at least 2020–21. There will be a school-level formula, but it will only be used to calculate how much each local authority receives. They will then be free to allocate it (subject to certain regulations) to the schools in their area according to their own funding formulae.

There were also a number of other changes to the original proposals. First, there is more money. The average cash-terms increase in funding in pupil between 2017–18 and 2019–20 is now around 3% rather than just under 1% as under the original proposals (equivalent to a real-terms freeze). Second, there are new absolute minimum levels of funding per pupil for both primary and secondary schools. Finally, protections against losses were extended such that no school could experience a cash-terms increase  of less than 0.5% per year between 2017–18 and 2019–20 (as opposed to a cash-terms fall of 1.5% per year). The maximum any school can gain has also increased from 5.6% to 6.1% in cash-terms per pupil. However, none of these changes will affect schools directly. They will affect the amount that each local authority receives and it is the local authority (in discussion with schools themselves through ‘School Forums’) who will decide how much each school actually receives. The minimum funding levels for primary and secondary schools are not obligatory and local authorities are able to reduce individual schools’ funding per pupil by up 1.5% in cash-terms if they wish. It is sensible that this latter protection is less than the 0.5% increase in the main formula as it will allow local authorities’ funding formulae to respond to the changing circumstances of schools (e.g. if a school is becoming less deprived, its funding can go down).

What will this achieve?

Given the current state of the school funding system, the latest proposals imply school funding reform is moving in the right direction, albeit it at a slower pace than implied by policy prior to the general election. If implemented, this will get closer to a system where similar areas will receive similar levels of funding. However, the proposals will not ensure that similar schools are funded in a similar way, as local authorities will still be free to implement their own funding formulae.

We don’t know anything, however, about government plans after 2019–20, either in terms of continued transitional protections or the full introduction of a school-level national funding formula. This is a source of major uncertainty. The government still says it is their ‘intention’ to implement a ‘hard’ formula. Whether it actually happens – in particular given that this change would require primary legislation to pass through parliament – remains to be seen. 

]]> Thu, 21 Sep 2017 00:00:00 +0000
<![CDATA[How much would it really cost to write off student debt?]]> Amid continuing debate over university tuition fees there remains confusion over some important numbers. We showed before that scrapping tuition fees for new students would increase borrowing by £11 billion a year. It has more recently been suggested that debt accumulated by graduates under the £9k a year tuition fee regime should be written off. If that policy were implemented immediately it would have almost no effect on government debt in the short run, but due to reduced future repayments from graduates, would increase debt by around £20 billion by 2050. If implemented after an election in 2022 the cost would be much higher, adding around £60 billion to debt in the long run.  Suggestions that debt would rise by £100 billion are wrong. £100 billion is the outstanding value of all tuition fee and maintenance debt since 1998 – it is not the answer to the question: what would be the impact on public debt of writing off fee loans accumulated under the £9,000 tuition fee regime?

Citing concerns about access to university, Labour’s election manifesto proposed to scrap tuition fees for all future students. Our previous work outlines the impact of this on graduates and the government finances. However, following the release of the manifesto, Labour leader Jeremy Corbyn went further than this, stating in an interview with NME that he would “deal with” the debt burden of those with “the historical misfortune of being at university during the £9,000 period”1.

This sparked considerable debate, with some reports suggesting this would cost approximately £100 billion. In fact, the £100 billion figure is the total value of all outstanding tuition fee and maintenance debt right back to 1998. The outstanding fee debt of graduates who entered university after 2012 stands at £34 billion. If that were written off in its entirety it would have almost no effect on government debt in the short run, but due to reduced future repayments from graduates, would increase debt by around £20 billion by 2050 (in current day terms). The difference is made up of loans the government expects to write off anyway. Of course, if the write-off were not to occur until after a 2022 general election, the costs of writing off all tuition fee debt would be much higher – we estimate this would add roughly £60 billion to debt by 2050.

As with the policy of scrapping fees for future cohorts, it is the highest earning graduates that would benefit the most, with the lowest earning graduates benefiting very little from reduced compulsory loan repayments.  

Government finances

Figures from the Student Loan Company show that, as of March 2017, £100.5 billion was outstanding in student loans. However, as shown in the Table, this includes £11.2 billion of Welsh, Scottish and Northern Irish loans, which are administered separately, as well as £44.1 billion in pre-2012 English loans. Furthermore, only around £30 billion of the remaining £45.3 billion in post-2012 loans consists of tuition fee loans including the interest accumulated on those, with the rest made up of maintenance loans. Adding the approximately £4 billion in fee loans that will have been paid to universities since March, the current level of total outstanding fee debt of students having entered university after 2012 will be around £34 billion.

NI, Welsh and Scottish loans


English loans - pre-2012 cohorts


English loans - post-2012 cohorts


Of which:    Tuition fee loans


                    Maintenance loans


Total outstanding UK student loans


See End Note 2

Writing off the post-2012 tuition fee loans would weaken the public finances. The impact on the deficit would be rather complicated, with a substantial increase in the first year, by up to as much as the full value of the debt written off – that is there would be a one off increase in the deficit of up to £34 billion. Beyond that it would be increased only by the loss of interest that would otherwise have been accrued on the outstanding debt. Depending on how the write-off is scored it is possible that the deficit would actually be reduced in future years as less debt will be written off in those years. But of course this would all be dwarfed by the £11 billion a year cost if loans were replaced by “free” tuition going forward.  

The impact on government debt of writing off these £9k loans (and ignoring any change to the tuition fee regime going forward) is more straightforward. In future years, the Government would receive lower loan repayments and therefore debt would increase. Figure 1 summarises our estimates of the pattern of this increase. In the short run, there would be no impact on government debt as new graduates would be making repayments on their maintenance loans, meaning overall receipts are unaffected. However, the impact starts to increase as more students clear their maintenance debts, resulting in the government losing out on repayments from graduates. Government debt would continue to rise above its counterfactual level for 30 years, before levelling off at the point where outstanding student loans would have been written off anyway. Thereafter the only cost to the public finances would be the interest paid by government on the extra debt accumulated, while government debt would be permanently higher.

Figure 1. Increase in Government debt from writing off post 2012 English tuition fee loans

See End note 3

The figure highlights that with all other government policy unchanged, this policy would increase the debt by around 1% of national income by 2050; equivalent to around £20 billion in today’s terms. If instead only the amount in excess of the £3,465 charged to those going to university in 2011 were to be written off, government debt would be around £10 billion higher in 2050 as a result of the policy.

Obviously the level of outstanding post-2012 fee debt will increase under the current system as new cohorts experience the higher fees, increasing the cost of writing off these loans. For example, if a government were to come to office in 2022 set on writing off all outstanding fee debt from the post-2012 cohorts, outstanding tuition fee debt would be in the region of £100 billion4.  Writing this debt off would increase government debt by roughly 3% of national income, or £60 billion, in 2050.

Wider implications

With all else held constant, the main beneficiaries of this proposal would be high earning graduates, with low earning graduates standing to benefit very little. Under the current system, high earning graduates make the highest student loan repayments and repay the largest proportion of their debt. If a significant part of the debt were to be written off, their total repayments would therefore be reduced most. Low earning graduates, on the other hand, are forecast to repay very little of this final part of the loan; indeed around one-third would see no change at all to their student loan repayments as a result of the policy as they will never earn enough to clear even their maintenance loans.

Furthermore, not all students take out the full fee loans available to them – for example, around 7% of students starting university in 2014–15 chose to pay their fees upfront, while others did not borrow the full amount or have already made repayments on their tuition fee debt. Unless there will be some form of compensation for those that paid their fees (or part thereof) upfront, those graduates would not benefit (or not benefit as much) from any writing off of tuition fee debt. They (or their families) might reasonably feel cheated.

Writing off the tuition fee debts of those who paid the post 2012 fees in England might also place pressure for additional spending in Scotland, Wales and Northern Ireland whose populations would otherwise not benefit from this change.

There is also of the issue of those who went to university before 2012, who themselves faced tuition fees; 2011 students incurred fee debts of more than £10,000 for their degrees, for example. While these debts are considerably lower, leaving these individuals’ debts untouched while trying to address the “historical misfortune” of attending university after 2012 would seem contradictory. Adding these earlier debts to the write-off would of course add to the long run costs considerably.

Of course, our forecasts for the impact on government debt in the long run assume no policy changes designed to recoup some of the money. The government could, for example, pay for this with a modest increase in the top rate of income tax. This would do something to alleviate concerns that the policy is regressive, although high earners without student debt – people who didn’t go to university as well as those who went but do not have any outstanding debt – would lose out. 


End Notes

1. The NME interview with Jeremy Corbyn can be found at .

2. Figures from March 2017 Student Loan Company report “Student Loans in England: Financial Year 2016-17”.

3. Uses OBR central projection of GDP. The discontinuity in 2046 occurs because this is where students start to have their debts written off. The figure only includes the additional debt impact of writing off the stock of post-2012 tuition fee loans, it does not include any debt impact from scrapping tuition fees going forward.

4. This uses the estimated increase in student debt, including interest, from the OBR’s March 2017 Economic and Fiscal Outlook and assumes that tuition fee loans will stay constant as a proportion of total loans given out. It has been put in today’s prices.

]]> Thu, 14 Sep 2017 00:00:00 +0000
<![CDATA[How do the rich respond to higher income tax rates?]]> The 2017 election saw Labour propose a 45% rate of income tax on incomes above £80,000, and a 50% rate on incomes above around £125,000, with the aim of raising around £4.5 billion per year in revenues. Analysis by IFS researchers at the time concluded that Labour’s revenue estimate was probably little on the high side, but was not implausible. This reflected a high degree of uncertainty about the extent to which taxpayers would respond if their tax rate was increased, and the importance of such responses to the revenue effects of Labour’s proposals.  

Today we publish a Briefing Note that presents new analysis of how high income taxpayers respond to changes in income tax rates. The Note is based on three new Working Papers, funded by the Nuffield Foundation, the Economic and Social Research Council and the European Research Council.

The first paper updates and critiques HMRC’s 2012 analysis of responses to and the revenue effects of the UK’s short-lived 50% income tax rate on incomes above £150,000 (which was in place between April 2010 and March 2013). The second uses individual (rather than aggregate-level) data to analyse responses to the 50% tax rate, and looks in more detail at the nature of the responses. It also proposes a new approach for estimating the long-term responses to a tax rate change, when temporary re-timing effects are significant, after finding that existing approaches perform poorly for the 50% tax rate. The third paper looks at how taxpayers respond to income tax thresholds where marginal rates change (such as the higher rate threshold where the rate increases to 40%, and £100,000, where the personal allowance is tapered away, creating a 60% marginal rate).

This observation highlights some key findings of these papers. 

Would a 50% rate on incomes above £150,000 raise revenues?

In 2010 the marginal rate of income tax on incomes above £150,000 was increased from 40% to 50%. It was then cut to 45% in 2013, after HMRC had estimated that the 50% rate would probably raise no more than a 45% rate, due to the extent to which they estimate that affected individuals responded to the higher rate by reducing their taxable income. There are a number of debatable assumptions in the methodology used by HMRC to produce those estimates. However, given the range of estimates obtained in our own analysis, we conclude that the HMRC estimate is a reasonable central estimate for policy-costing purposes. 

But our analysis also shows that it is plausible that such a policy could raise or cost £1-2 billion a year in revenues. The uncertainty is due to a range of factors, not least efforts by taxpayers to bring forward income to avoid the pre-announced tax rise (termed ‘forestalling’) – this makes it difficult to disentangle these temporary changes in the timing of income from the permanent responses (which is what really matters in the long run for the public finances). In addition, revenue effects depend on the extent to which revenues from other taxes (such as VAT or capital gains tax) and revenues in future periods will be affected by these behavioural responses. Available data do not allow such impacts to be fully estimated.   

What about Labour’s recent plans?

Our research finds evidence that the overall high degree of responsiveness of high earners to the former 50% rate was driven by those with the very highest incomes: those with incomes between £150,000 and £200,000 appear to be only a third to a half as responsive to tax rates as the £150,000+ group as a whole.

This implies that Labour’s recent proposals – increasing tax rates above £80,000 – would likely raise additional revenues, although probably not as much as Labour were claiming. Furthermore, the estimates suggest that this additional revenue would likely have come predominantly from those with incomes between £80,000 and £200,000, rather than those with the very highest incomes.

What types of taxpayers and responses drive the overall results?

The new research also provides evidence on the types of response and the income sources driving overall measures of responsiveness to income tax.

As shown in Figure 1, individuals with dividend incomes, such as owner-managers of incorporated businesses, are more responsive to changes in tax rates than those with employment income. Incomes from this source jumped substantially in 2009–10 (prior to the introduction of the former 50% tax rate), fell substantially in 2010–11 (after the introduction of the 50% tax rate), and remained relatively depressed the following year. In contrast, income from employment was more stable. A similar pattern is found when examining tax thresholds: company owner-manages respond much more to these thresholds than other taxpayers, with virtually no response among employees.

This pattern is likely to reflect the fact that company owner-managers have an additional mechanism through which they can respond to income taxes: they can manipulate the timing of their income by retaining it in or taking it out of their company at a time that lowers their overall tax bills: such as bringing it forward to avoid the 50% tax rate. (This benefit comes on top of the fact that the income they take in dividends gets taxed much less, on average, than the income of employees, as IFS researchers have highlighted). 

Figure 1. Trends in different income sources and deductions for groups with incomes greater than £150,000, 2001–02 to 2011–12

Note: The blue vertical line shows when the 50% tax was announced (in March 2009) and the red line when it was implemented (in 2010–11).

Source: Authors’ calculations using SA302 data from 2001–02 to 2011–12.

Evidence from the US typically shows that increases in tax deductions play a major role in taxpayers’ responses to increases in income tax. We find no evidence of this for the UK; Figure 1 shows that the cost of deductions reported on tax returns fell when the 50% tax rate was introduced.

This may reflect restrictions to pension contributions tax relief and associated anti-forestalling rules introduced at around the same time. By reducing allowable deductions, these policies would have tended to push up taxable income, offsetting some of falls in taxable income as people responded to the 50% rate. Analysis of the change in taxable incomes would therefore lead to an underestimate of the extent to which taxpayers reduced their incomes in response to the 50% tax rate. Estimates of responsiveness based on the larger falls in pre-deductions income (which should have been affected by these deduction rule changes) imply a 50% tax rate would likely yield lower revenues than a 45% tax rate on incomes above £150,000.

However, it should be borne in mind that not all deductions relevant for the revenue effects of income tax rate changes are observed in the tax return data we use. First, as already mentioned, income can be retained in companies, so will not be picked up on personal tax returns. Second, contributions to occupational (rather than personal) pensions are deducted from salaries before incomes data is submitted to HMRC. Future work should explore the scale and potential revenue effects of responses via these margins in more detail by linking personal income tax data with corporate tax and occupational pension contributions data.

Concluding thoughts

Taken together the empirical and methodological contributions of these new papers are important. But even with careful analysis, it has not been possible to obtain precise estimates of the responsiveness of high income individuals to income tax rates in the UK. 

Part of this reflects the significant forestalling that occurred when the 50% rate was announced over a year in advance. It would be easier to evaluate the longer-term behavioural responses to and revenue effects of future changes in income tax rates if they were announced with (near) immediate effect: affected individuals would not have the time to engage in forestalling activities that hamper identification of the longer-term effects. It would also have the benefit from the government’s perspective of maximising revenues (taxpayers engage in forestalling activities to minimise their tax liabilities). However, tax changes made with (near) immediate effect may be subject to less debate and scrutiny, which may impose its own costs.

Other factors also contribute to the uncertainty: other economic or policy changes can impact on incomes, making isolation of the effects of tax rate changes tricky; and the precise nature of any response, not just its size, will impact on revenues.

Policymakers will therefore need to reconcile themselves to uncertainty about the revenue effects of changing the top rates of income tax. Of course, a central estimate will be required for the official Budget costing. But it is important to be aware of the uncertainty when designing and implementing policy, and be up front about it when ‘selling’ policies to the public.




]]> Tue, 22 Aug 2017 00:00:00 +0000
<![CDATA[How might Brexit affect food prices?]]> Brexit has the potential to have a substantial impact on the prices households pay for food. Currently around 30% of the value of food purchased by households in the UK is imported, and the major source of food imports is the EU. In comparison, only 17% of overall consumer spending is on imported goods. This means that changes in the costs of imports – for example, through changes to tariffs or movements in exchange rates – are likely to have a particularly big impact on food prices. In new analysis released today, we study how different households might be affected by changes in the costs of food imports, due to tariff changes and exchange rate movements.

So far, the clearest and most immediate impact of the UK’s decision to leave the EU has been on the exchange rate. Sterling depreciated by 13% between January 2016 and March 2017; see Figure 1. This is an important determinant of the price of imports. Imports are purchased in foreign currency: when these currencies become more expensive relative to sterling (‘sterling depreciation’), more sterling is needed to purchase the same quantity of foreign currency, and therefore the same quantity of foreign goods.

One way to gauge the impact of sterling depreciation on food prices is to look at a previous episode.

From the beginning of 2007 to the end of 2008, there was a 21% depreciation in the effective sterling exchange rate; see Figure 1 again. Over the same period, prices overall rose by 6%, but food prices rose by 15% – that is, food prices rose by 9% relative to the prices of other goods and services. This period saw sharp increases in world commodity prices for key agricultural inputs, which are also likely to have contributed to increased food prices. However, while other countries (whose currencies did not depreciate) experienced food price rises, these were neither as large nor as persistent as the increase in the UK. This suggests that exchange rates played an important role in driving higher food prices in the UK.

So far, the depreciation in sterling since the beginning of 2016 has not been associated with similarly sized effects on food prices. The consumer price for food relative to the overall consumer price level initially declined after the referendum, but it started to increase moderately towards the end of 2016. The rise in producer prices since the referendum has been somewhat faster, however, and these costs are likely to be passed on to consumers in the future.

Figure 1. Exchange rate movements and the real price of food

Note: The Producer Price Index (PPI) for food gives output (‘factory gate’) prices for producers of food products selling to the UK market. This is the price received by UK manufacturers. It covers any margin they make on the goods they sell, as well as any costs such as labour, raw materials and energy, interest on loans, site or building maintenance, and rent. Both the PPI and CPI (food) are shown relative to the all-items Consumer Prices Index (CPI). The effective exchange rate and the price indices are rebased to equal 1 in January 2005.

Source: Bank of England and the Office for National Statistics.

How food prices will change following the UK’s actual departure from the EU remains highly uncertain. Currently, the UK benefits from tariff-free trade within the EU, and the UK and other EU members levy common tariffs on products imported into the EU from other countries. As is generally the case across the world, these tariffs are, on average, higher on agricultural products than on non-agricultural products and there is a lot of variation in tariffs both across and within broad food groups.

If the UK leaves the EU customs union, it would be free to adjust the tariffs it charges on agricultural goods. Under World Trade Organisation (WTO) rules, the UK would not be able to set tariffs that discriminate between trading partners, except as part of a free trade agreement or to give developing countries special access to its market. If the UK and the post-Brexit EU fail to strike a free trade deal, it is likely tariffs would be imposed on EU imports into the UK, as the UK would be unable to impose zero tariffs on imports from the EU without also extending tariff-free access to all other WTO members. This would raise the price of food imported from the EU, which is the major source of food imports into the UK, accounting for 70% of gross food imports. Therefore if the UK did not strike a free trade deal with the EU, food prices would be likely to rise significantly. This could potentially be ameliorated if the UK reduced tariffs across the board by a substantial margin and/or decided to accept cheaper food imports that do not meet current EU regulatory standards. There would also be costs associated with non-tariff barriers, such as customs checks.

The UK could reduce tariffs from their current level, and may choose to do so, especially for imports of foods that are not domestically produced – for example, oranges and olives. It would also be free to strike free trade deals with other countries. In addition, if the UK ceased to be a full member of the EU single market, it would be able to apply different regulatory standards to food imports, leading to the possibility of importing hitherto banned produce (such as chlorinated chicken). Such changes would serve to lower the price of foods imported from non-EU countries. That said, most advanced countries seek to protect their domestic farmers through a combination of price supports, subsidies and trade barriers. The extent to which the UK would choose to act differently in this regard is not clear at this stage.

Tariff changes and movements in the exchange rate directly affect the cost of getting imported food products onto supermarket shelves. This will naturally feed into the prices faced by consumers for imported goods. The extent to which tariff changes and exchange rate movements feed through to prices is uncertain and may vary across goods. The prices of domestically produced products are also likely to change, for two reasons. First, many domestically produced products use imported inputs, and changes to firms’ costs will tend to feed through to the prices they charge for their final products. Second, changes in the price of imported goods are likely to lead to changes in the price of similar domestically produced goods because of competitive effects: for example, if the prices of imported goods rise, then domestic producers who compete in the same markets might take advantage of the opportunity to increase their prices too.

How might these changes affect different households? The first thing to consider is that lower-income households allocate a higher proportion of their spending to food than higher-income households (23% of spending for the lowest-income tenth of households versus 10% for the highest-income tenth). Poor households are therefore more exposed to rises in the general level of food prices.

It might be the case that poorer households buy less imported food and hence would be less exposed to price rises driven by exchange rate or tariff changes. In our briefing note, we show that there are big differences in import penetration between food groups – for example, 39% of fruit products are imported, while only 20% of bread and cereal products are imported. Additionally, using detailed spending data, we observe that some households spend a lot more on imported products within broad food groups than others do. For example, on average, households buy around 35% of their beef, lamb and pork from abroad, but some households buy all of it from overseas and others purchase none from abroad.

However, we find that these differences do not translate into any clear differences by income – the poor buy a similar fraction of their food from abroad to the rich. This means that poorer households typically allocate a larger share of their total spending to imported food, and therefore are more exposed to changes in the price of imported food.

There is a great deal of uncertainty over what the nature of the UK’s post-Brexit trading arrangements will be. Decisions over post-Brexit membership of the single market and participation in the customs union will have profound effects on the price and import mix of the foods on UK supermarket shelves. It is also unclear whether sterling will depreciate further – or appreciate – as Brexit proceeds. These uncertainties over tariffs and the exchange rate mean that UK households are potentially going to be affected by considerable and unpredictable changes in food prices, with the poorest households much more exposed to this risk than the richest households. 


]]> Thu, 27 Jul 2017 00:00:00 +0000
<![CDATA[Where next for tax and spend?]]> Terrible economic growth since the start of 2008 has created big problems for the finances of both households and government. The government’s deficit increased from 2.6% of national income to a post Second World War high of 9.9% of national income in 2009–10. The period of “austerity” since then has been focussed on getting that deficit down through a combination of some net tax rises and some big cuts to benefits and spending on public services. The deficit stood at 2.4% of national income in 2016–17. So the UK now has a smaller government deficit than prior to the financial crisis. Tax is higher as a share of the economy than it was in 2007–08. Perhaps more surprisingly public spending also now takes a slightly larger fraction of national income than it did in 2007–08. All that austerity has got us back to where we were, in terms of spending overall, ten years into the last Labour government’s term of office. The problem has been that the economy is far far smaller than expected.

Despite the UK’s weak economic performance over the last decade the Office for Budget Responsibility judges that there is currently no spare capacity in the economy. If correct this means we can’t rely on above normal levels of growth helping bring the deficit down further. Therefore the commitment made by Chancellor Philip Hammond in last November’s Autumn Statement to deliver a budget surplus “as early as possible in the next Parliament” was expected to require a combination of further tax rises and spending cuts.

The plans of the previous Conservative Government in the March 2017 Budget are shown in the figure below. The tax burden was forecast to continue rising – climbing to its highest level since the mid 1980s – driven by further tax raising measures such as an increase in dividend tax and council tax rises (earmarked for spending on social care). The net effect of these discretionary tax measures is to boost revenues by a further £5 billion in 2021–22.

March 2017 Budget forecast for tax and spend

March 2017 Budget Forecast: tax and spend

Public spending is forecast to continue falling as a share of national income. In part this results from planned cuts to benefits – mainly targeted at working age benefits – that are already in the pipeline. This includes a freeze to the rates of most working age benefits in April 2018 and April 2019 and also the effect of more new claimants moving onto Universal Credit (which is now less generous, on average, than the system it replaces) and more new births in large families (who no longer see a boost to means-tested benefit entitlements if they already have two children). In total these benefit cuts will reduce spending by a further £11 billion in 2021–22, and by more in the longer-term.

There is also a continued squeeze on the share of national income devoted to public service spending (defined here as total spending by government less that spent on benefits, state pensions and debt interest). Over the five years from 2016–17 to 2021–22 this is forecast to rise in real terms by £37 billion, with a £23 billion increase in day-to-day spending and a £14 billion increase in investment spending. These real increases equate to a fall in spending as a share of national income equivalent to £17 billion by 2021–22. Within this there is considerable variation: day-to-day spending is to fall as a share of national income by £27 billion while investment spending is forecast to rise by £10 billion. The spending cuts are far from evenly shared across departments: for example while the Ministry of Justice faces further budget cuts the Department for International Development’s budget continues to rise.

While the effect of spending cuts so far has been to bring public spending down to its pre-crisis level as a fraction of national income, current plans imply further cuts such that overall public spending will reach its lowest share of national income since 2003–04 in 2019–20.

So far the policy announcements by the Government do not amount to a significant change in fiscal policy. The additional spending for Northern Ireland, while sizable in the Northern Irish context, are not a big deal for the UK public finances while the abandonment of plans to means-test Winter Fuel Payments and to move, from April 2020, from “triple” to “double” lock indexation of the state pension would not have cut spending by much in the current parliament.


The Conservative Party manifesto watered down the previous commitment on the deficit to targeting a “balanced budget by the middle of the next decade”. As shown in the figure above the latest forecasts imply that this would require some combination of further tax rises and spending cuts beyond 2021–22. Continued ageing of the population, with the large generation born just after the Second World War putting an increasing strain on the NHS and social care budgets over this period, will make that all the harder to achieve. Brexit adds an additional layer of uncertainty.

On the other hand this target will not bite for a while. The new Government could therefore choose to pursue a looser fiscal policy than currently planned in the short term while claiming still to be heading for budget balance at some point in the 2020s. But given the scale of the takeaway planned for the next few years an “end to austerity” – as defined by no further net tax rises, benefit cuts or cuts to spending on public services – would require a very sharp change of direction. It would imply a £17 billion boost to planned spending on public services alongside a £5 billion net tax cut and an £11 billion increase in planned benefit spending – i.e. a giveaway of £33 billion a year. What that would do would, on current forecasts, leave us with a deficit at its current level – 2.4% of GDP – in 2021–22. That is an option in a way that it was not an option back in 2010.

We could choose to continue to run deficits of around the current level over the longer-term. If the economy were to grow as expected this would be sufficient to see debt fall as a share of national income over the longer term. It would mean that over the next few years household incomes could be better supported and a greater quality and quantity of public services could be enjoyed. But it would also involve planning to live with elevated public sector debt for longer. It could give the chancellor less room for manoeuvre if the economy were to suffer badly, for Brexit related or other reasons over the next few years; and it would almost certainly require the abandonment of the pledge to eliminate the deficit in the mid-2020s.

If instead the concern was with public spending cuts – rather than overall austerity – then an alternative option would be to raise taxes in order to reduce the planned cuts to spending on public services as a share of national income or to reduce the scale of the benefit cuts currently in the pipeline. To give one example the government could choose not to implement the planned cut to the corporation tax rate from 19% to 17%. That would raise an additional £5 billion which could be spent on public services or reducing benefit cuts, without affecting the deficit. Other tax increases are of course available.

As ever the chancellor has choices and as ever there are trade-offs. He could decide to spend and borrow more; he could decide to spend and tax more; or he could decide to stick with his current plans and see spending continue to fall. He is not going to be able to give everyone everything they want and the sooner he makes plain what his choices are the better for all.

This observation is based on a presentation by Carl Emmerson which wis given at an Institute for Fiscal Studies / Institute for Government joint event on “What next for tax and spend?” July 12 2017.

]]> Wed, 12 Jul 2017 00:00:00 +0000
<![CDATA[Proposed changes to alcohol taxation are small beer]]> The government has recently consulted on the structure of alcohol taxes. This consultation focuses on two issues: (i) the introduction of a new still cider and perry band that would increase the tax on products below 7.5% ABV, and (ii) the introduction of a new still wine band that would reduce the tax on products between 5.5% and 8.5% ABV.

In this Observation we summarise our main points from our response to the consultation (which is reproduced in full at the end of this piece).

The main rationale for subjecting alcohol to additional taxes above and beyond VAT is that there are social costs of drinking that are not taken into account by the drinker themselves. These costs include the public costs of treating alcohol-related diseases and dealing with alcohol fuelled crime. The aim of taxation should be to raise the price paid by the drinkers who generate these social costs by an amount equal to the (marginal) social cost of drinking an extra drink.

Ideally, taxes would be levied per unit of pure alcohol (ethanol), as it is the alcohol that creates social costs. However, restrictions imposed by the European Union mean that cider and wine must be taxed per litre of product. This means that for a given tax rate per litre of product, higher strength products (those that contain more pure alcohol per litre of product) face a lower per unit of alcohol tax rate. This restriction is one motivation for introducing additional tax bands for mid-strength cider and low-strength wine.

In light of this, our view on the proposed modest reforms are summarised as follows:

  • The introduction of a new cider and perry band is a small movement in the right direction. However, the proposed small tax increase would still leave the tax levied on cider and perry at a much lower level than that levied on other alcohol products.
  • The tax rates on cider should be brought in line with beer. Currently, a litre of 7.5% cider attracts less than a third of the tax of a litre of 7.5% beer. This provides an incentive for people to choose cider over beer, and to consume more alcohol, than if cider were taxed at the same rate as beer is currently taxed. This equalisation of rates does not necessarily mean increasing the tax applied to cider to beer tax levels. One possibility would be to increase the tax on cider and reduce the tax on beer to equalise them in a revenue neutral way.
  • The argument that cider is part of the rural economy is not a good reason for subjecting cider to a lower rate of tax. Alcohol taxes should be designed to deter socially costly overconsumption of alcohol. If there are reasons to subsidise the cider industry over the beer, or other industries, then appropriate policy instruments should be designed to do this. However, it is not clear why the cider industry needs government assistance, potentially at the expense of other alcohol – or rural – industries in the UK.
  • There are currently virtually no still wine products sold off-trade with an ABV less than 8.5%. This part of the market would have to increase several-fold to have any impact on total alcohol consumption.

Overall this represents a missed opportunity for a more comprehensive review of how alcohol is taxed in the UK. The annex to this Observation contains our responses to selected questions in the consultation (shown in bold).

]]> Thu, 15 Jun 2017 00:00:00 +0000
<![CDATA[Comparison of parties’ plans for education spending on 16-18 year-olds in England]]> At this election, the Conservatives, Labour and Liberal Democrats are all proposing extra spending on 16-18 education over the course of the next parliament. The Conservatives and Liberal Democrats would do enough to keep spending per 16-18 year-old pupil constant in real terms. Labour propose increasing spend per pupil by 8% in real terms. The Labour and Conservative proposals would represent larger increases than the parties’ respective commitments on schools spending per pupil for younger pupils. But all would leave spending per 16-18 year-old pupil around 10% lower than the parties’ respective proposals for secondary schools.

In this observation, we examine the main parties’ proposals for spending on 16-18 education in England, which includes students in School Sixth Forms, Sixth Form Colleges and Further Education Colleges. This area of education receives considerably less attention in public debate than other areas and seems not to have been a major spending priority for policymakers over recent decades.

Recent history and current plans

Over the last 30 years, spending on education for 16 to 18 year olds has fared substantially worse than other areas of education spending.  For example, the figure below shows that spending per 16-18 year-old pupil in 1990–91 was more than 50% higher than spending per secondary school pupil, but by 2017–18 it was 13% lower. It is not clear why students in 16-18 education should receive 13% less in resources than pupils in secondary schools, or indeed 40% less than students in higher education (see here).

The relative decline in 16-18 education spending per pupil was the result of deeper cuts in the 1990s, slower growth in the 2000s, and being one of the only areas of education spending to be cut under the coalition government. Further real-terms cuts were planned in the latest Spending Review in 2015, which would have reduced spending per 16-18 year-old pupil to £5,200 by the end of the Spending Review period in 2019-20 (in 2017–18 prices, as are all figures in this observation). This would have returned it to around its 2002 per-pupil level, and lower than the level seen at the start of the 1990s.

In Budget 2017, the Conservative government committed to following the recommendations of the Sainsbury review for post-16 education. These reforms would introduce ‘T-levels’ for technical education – slimming down the thousands of vocational qualifications currently on offer to 15 different lines of learning - and increase the number of teaching hours on some technical routes by more than 50%. The Conservative government committed £420 million of additional investment in 2021-22 for the implementation of these reforms.

If, after the Spending Review period, spending per pupil would - excluding the additional £420 million - have been frozen in cash terms between 2019-20 and 2021-22 (as was the case between 2010-11 and 2019-20), then total spending per pupil including the extra £420m would be about the same real-terms level in 2021-22 as it is today.  

The parties’ manifestos

The Conservative manifesto reaffirms their commitment to these plans, which would effectively result in a real-terms freeze in 16-18 spending per pupil over the next parliament. This has a similar effect to the Liberal Democrats’ proposal to protect per pupil funding of schools and colleges in real terms. However, these plans would still leave spending per pupil on 16 to 18 year olds in 2021-22 around 14% below its peak in 2011-12. In the case of the Conservatives, because plans for 16-18 education are slightly more generous than those for schools, the gap between 16-18 education and secondary school spending per student would shrink slightly from its current level of 13% to around 10% in 2021–22.

Spending per pupil aged 16-18: actual and planned

Notes and Sources: Spending per pupil in 16-18 education represents a weighted average of spending per student in School Sixth Forms, and spending per student in Further Education and Sixth Form Colleges. See Belfield, Crawford and Sibieta (2017); Authors’ calculations based on Conservative, Labour and Liberal Democrat General Election Manifestos; Spring Budget 2017; GDP deflator and student numbers projected forwards using ONS population projections for numbers of 16-18 year olds in England.

The Labour manifesto commits to “bring funding for 16 to 18-year-olds in line with Key Stage 4 baselines," which means that they would seek to align the basic amount provided for a pupil aged 11-16 with that provided for a pupil aged 16-18 (these basic amounts are the most important, but not the only, element of each funding formula).  The manifesto includes around £390 million in 2021-22 to achieve this, which would be additional to the funding proposed by the Conservatives to implement the Sainsbury reform. If implemented, Labour’s proposals would increase 16-18 spending per pupil to £5,800, a real increase of 8% over the parliament. However, this would still leave spending per student in 16-18 education about 11% below that for secondary schools in 2021–22, which under Labour’s manifesto plans would be about £6,500.


As with spending on schools, there is a clear difference between the Conservative and Labour plans. Labour propose an 8% real-terms increase in funding per 16-18 year-old pupil over the parliament, whereas the Conservatives and Liberal Democrats propose a real-terms freeze. Both of the Conservative and Labour proposals for spending on 16-18 year old are more generous than their respective plans for school spending, which would go a small way to reversing the relative trends in spending on pupils of different ages seen over the last thirty years. However, in each case, spending per student in 16-18 education would remain about 10% lower than it would be for secondary schools.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Thu, 01 Jun 2017 00:00:00 +0000
<![CDATA[Challenging times ahead for the NHS regardless of who wins the election]]> The Conservatives, Labour and the Liberal Democrats have all announced commitments for NHS spending in their manifestos. However, the amounts, timing and periods covered by these pledges vary by party. How do their plans for health spending in England really compare and what are the implications once the costs of demographic change are accounted for?

Under the current government’s plans set out in the 2015 Spending Review, Department of Health (DH) spending in England will increase by an average 0.8% per year in real terms between 2017–18 and 2020–21. This would increase spending from around £124 billion to around £127 billion (in today’s prices), as shown in Figure 1.

The Conservative, Labour and Liberal Democrat policies are as follows:

  • The Conservative manifesto promises “we will increase NHS spending by a minimum of £8 billion in real terms over the next five years”, and committed to “delivering an increase in real funding per head of the population for every year of the parliament”. This would suggest an increase in DH spending to £132 billion (in today’s prices) in 2022­–23, if the other (non-NHS) aspects of DH spending were frozen in real terms over this period.[1] This would be an average growth in real spending of 1.2% per year between 2016–17 and 2022–23.
  • The Labour manifesto promised a larger increase in health funding. Labour would increase spending relative to current government plans by £7.7 billion in 2017–18, rising to £8.4 billion (in nominal terms) by 2021-22. This could take DH spending to around £135 billion (in today’s prices) in 2021–22.[2] This would be an average 2.0% per year real increase in spending between 2016–17 and 2021–22.
  • The Liberal Democrats have pledged to increase spending on health and social care in England, Wales and Northern Ireland by approximately £6 billion each year, with £2 billion ring-fenced specifically for social care. This could imply DH spending of £131 billion (in today’s prices) in 2021-22, and average growth in spending of 1.4% per year in real terms between 2016–17 and 2021–22.[3]

The likely path for DH spending under each of these three parties’ plans is shown in Figure 1. All are proposing increasing health spending in real terms over the next parliament, with Labour being the most generous and the Conservatives the least generous. However, it is worth nothing that all these planned increases in health spending are well below the long run average; historically UK health spending has grown by an average 4% per year in real terms. Figure 1 shows that relative to spending having grown by 4% per year since 2009–10, the difference between the parties’ plans going forwards is modest. The next parliament will therefore continue to be an incredibly challenging period for the NHS, regardless of who wins the election.

Figure 1: Real terms Department of Health spending: actual and plan

Notes and sources: See notes 1-4 below.

To put the planned spending increases in context, it is important to remember the population is both growing and ageing. Between 2016 and 2021, the population in England is expected to grow by 3.7%, with the population aged 65 and over expected to grow by 9.2%, and the population aged 85 and over by 14.5%. This has important implications for NHS spending, since older individuals typically use more (and more expensive) health care than younger individuals. For example, average health spending on a 65 year old is estimated to be double that on a 30 year old. The ratio rises steeply at older ages, with average spending on a 70 year old three times, and spending on a 90 year old almost eight times, that on a 30 year old. Real NHS spending would therefore need to increase by 1.2% each year over the next parliament just to keep up with demographic change.

Figure 2 illustrates how age adjusted per person spending on health has changed since 2009–10 and what the parties’ plans imply going forwards. Current government plans imply age-adjusted per person spending falling over the coming parliament, while the Conservative plans would keep age-adjusted spending broadly constant across the parliament. Under Labour age-adjusted per person spending in 2021–22 would be 4% above 2009–10 (and 2016–17) levels. 

Figure 2: Real terms age-adjusted per person Department of Health spending relative to 2009-10 level: actual and plans

Notes and sources: See notes 1-4 below.

While the Conservatives', Labour's and the Liberal Democrats’ planned spending increases would be sufficient to meet demographic pressure in the short-run, it is important to note that the NHS faces cost and demand pressures over and above those that arise from demographic change. For example the increasing prevalence of chronic conditions, rising real wages and the development of new medical technologies all tend to push up health spending.

The demographic and other cost pressures facing the NHS are not new, and nor are they going to go away. The latest OBR estimate is that between 2016–17 and 2066–67 demographic pressures could increase health spending by 0.8% of national income, while demographic and other cost pressures together could push health spending up by 5.3% of national income, a sum equivalent to £109 billion in today’s terms. Funding such an increase in spending would require substantial tax increases or substantial cuts to other areas of public spending, which future governments may find difficult. Given this, the next government would be wise to consider these long run trends carefully, and start focusing on finding and implementing a long term solution to these funding pressures now. In the long run the NHS would be better served by a serious attempt to address these issues in a coherent and systematic fashion, than by the government just announcing further short term funding fixes.


[1] The Conservative manifesto only specifies how much higher NHS spending will be in 2022–23 compared to 2017–18; for intermediate years we assume spending increases at a constant rate between 2017–18 and 2022–23.  We also assume that the other (non-NHS) aspects of DH spending are frozen in real terms between 2017–18 and 2022–23 (non-NHS DH spending has fallen since 2010).

[2] Labour’s planned £8.4 billion increase in spending includes £5.8 billion for additional day-to-day spending, £2.0 billion for additional capital spending, and £0.6 billion to reinstate nurse bursaries. Labour’s spending plans are only specified relative to current government plans and interpreting this commitment is complicated by the fact that current government plans for DH spending do not extend as far as 2021–22; the 2015 Spending Review only set the department’s budget as far as 2020–21. However, if we assume that current government policy would be to increase the DH budget between 2020–21 and 2021–22 in line with the average nominal increase between 2016–17 and 2020–21 (a 2.5% cash terms increase, implying a 0.6% real terms increase), that would suggest Labour plan to spend around £135 billion (in today’s prices) in 2021–22.

[3] We assume that, under the Liberal Democrat proposals, £2 billion of the £6 billion they plan to raise from increasing income tax would be set aside for social care, and that of the remaining £4 billion 92% would be used to increase NHS spending in England (with the remaining 8% being used to increase spending in Wales and Northern Ireland). As with Labour, the Liberal Democrat plans refer to additional funding above the current government’s existing spending plans. We therefore again assume that current government policy would be to increase the nominal DH budget between 2020–21 and 2021–22 in line with the average nominal increase between 2016–17 and 2020–21.

[4] Previous spending and future spending plans (2009–10 to 2019–20) from HM Treasury Public Expenditure Statistical Analyses 2016.  2020–21 spending plans from the 2015 Spending Review.  GDP Deflator from OBR, 31 March 2017. Population estimates and projections for England are from the Office for National Statistics.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Tue, 30 May 2017 00:00:00 +0000
<![CDATA[The SNP’s fiscal plans and Scotland’s fiscal position]]> Today the SNP published their manifesto for the upcoming UK parliamentary elections. In it they set out a target of balancing the current budget by 2021–22 and plans for an increase in public spending over the next five years of around £120 billion relative to those set out in the March 2017 Budget. They also state that if they win a majority of Scottish seats, they believe they will have a strong mandate for a second referendum on Scottish independence once the ‘final terms of the [Brexit] deal [between the UK and the EU] are known’, and on a timescale chosen by the Scottish Parliament. In this context, this observation examines the state of Scotland’s underlying public finances and how they might evolve over the next few years given current economic forecasts.

The SNP’s fiscal plans for the UK

Before that though, a little more about the changes the SNP would like to see to the UK’s fiscal plans.

Like Labour, the SNP would like a target of balancing the current budget (the difference between tax receipts and day-to-day spending) by 2021–22, stating they would support borrowing only for investment purposes from that year onwards. Overall the SNP say their plans would allow £120 billion of extra spending between 2017–18 and 2021–22. This increase in spending – and borrowing – would be consistent with falls in national debt as a share of GDP, but at a slower pace than under current plans.

The additional day-to-day spending on public services the SNP plans would allow is less than that set out in Labour’s manifesto. On welfare though, the SNP plan to go further than Labour in reversing planned benefit cuts – ending the cash-terms freeze to working age benefits and revoking restrictions on tax credits and universal credit to the first two children in a family, for instance. Their plans imply a net increase in taxes relative to current plans – including an increase in the top rate of income tax, an increase in the bank levy, a bankers’ bonus tax and the cancellation of further cuts to corporation tax – although there would be tax cuts targeted at SMEs. However, unlike Labour whose plans mean increases in day-to-day spending will be largely funded from (much larger) tax increases, the SNP’s plans imply most of the increase in spending would be funded by additional borrowing.

The 2.3% of GDP target for overall borrowing set out by the SNP is consistent with investment spending increasing in line with the plans set out in March 2017 and implies lower investment than under Labour’s plans (an additional £250 billion over 10 years).   

Scotland’s public finances: the medium-term outlook

As well as proposing changes to the UK’s fiscal policy, the SNP’s manifesto set out plans that could have big implications for the management of Scotland’s public finances: a second independence referendum once the terms of the Brexit deal are known.

The latest Government Expenditure and Revenue Scotland (GERS) figures for 2015–16 estimate that the “net fiscal balance” – that is the difference between overall revenues and spending – in Scotland was in deficit by 9.5% of GDP, compared to a deficit for the whole of the UK of 3.8% of GDP. This difference reflects the relatively high levels of public spending in Scotland (around 11% higher per person than in the UK as a whole in 2015–16) which are not matched by tax revenues (which were around 5% per person below the UK average in the same year), following major falls in revenues from North Sea oil and gas.

Table 1 shows our latest projections through to 2021–22, based on the OBR’s March 2017 forecasts and the current government’s tax and spending plans. Over this period Scotland’s deficit is projected to fall to 6.7% of GDP, while the UK deficit is forecast to fall to 0.7% of GDP on the same basis. So over the next few years the gap between the two is projected to increase slightly as a percentage of GDP.

Table 1: Projected Net Fiscal Balance, UK and Scotland, 2015–16 (outturn), 2016–17 to 2021–22 (projections), % of GDP unless otherwise stated

Net fiscal balance








































Difference (£ bn)








Source: Author’s calculations using GERS 2015–16, and OBR EFO March 2017. Further details at end of observation.

Note: UK figures are taken from OBR EFO March 2017. Outturn figures for 2015–16 differ from those reported in GERS 2015–16 (at that stage the UK’s net fiscal balance was estimated to be in deficit by 4.0% of GDP in that year).

In 2017–18 the difference is projected to be 5.6% of national income (8.5% less 2.9%), which in cash terms is equivalent to about £9.5 billion, or around £1,750 per person in Scotland. That is the size of the Scottish deficit on top of its share of the overall UK deficit (which is £880 per person in the UK in the same year).

While large, this fiscal gap is actually somewhat smaller than under our last set of projections. This is because the forecasts for the UK budget deficit have been revised up by more than our projections of the Scottish budget deficit. In particular, upwards revisions to offshore revenues from North Sea oil and gas (driven by an increase in the sterling oil price) have offset more of the downward revisions to onshore revenues for Scotland than for the UK as a whole. However current projections for oil revenues of around £0.7 billion a year are still much lower than they were at the time of the last Scottish independence referendum when the OBR was projecting that revenues would be around £3 billion a year and the Scottish Government even higher (up to £8 billion a year).

It is also important to note that while changes to the overall levels of tax and spending in Westminster could lead to significant changes in headline deficit figures for the UK as a whole, they would have similar effects for Scotland: higher spending would feed through the Barnett formula into higher spending in Scotland, for instance. Therefore the fiscal gap would be little changed. What matters most for the gap is the performance of the economy and underlying tax revenues in Scotland relative to the rest of the UK.

Implications for an independent Scotland

Although under its new Fiscal Framework, the Scottish Government gains or loses if its devolved revenues grow by more or less than equivalent revenues in the rest of the UK, it does not bear responsibility for the pre-existing large fiscal gap. That is, under present constitutional arrangements, Scotland is largely insulated from the consequences of its higher implicit budget deficit. That would change though under independence.

A deficit of 6.7% of GDP, for instance, would be unsustainable as continuing to borrow that amount would lead to ballooning national debt. A newly independent country facing such a deficit would need to cut public spending or increase taxes to reduce the budget deficit to a more manageable level.

However, our projections are based on the Scottish Government’s official GERS publication, 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. With this in mind it is worth noting:

  • The Scottish Government might be able to negotiate a good deal on the share of the UK’s debt it took on at independence. Lower debt would mean lower debt interest payments and would therefore reduce the budget deficit. However, even if an independent Scotland inherited none of the UK’s net public debt, its projected budget deficit would still be substantial: around 5.2% of national income (rather than 6.7%) in 2021–22 holding all other elements of our projections fixed.
  • Without spending cuts or explicit tax rises, significantly faster growth in underlying revenues would be required to reduce the deficit to sustainable levels. For instance, to reach an overall deficit of 2.3% of GDP by 2021–22 – the SNP’s target for the UK as a whole in that year – would require Scottish revenues to grow by 4.0% a year in real terms, compared to a forecast 2.3% for the UK as a whole over the same period.

Looking further ahead, independence could also affect Scottish economic performance. A weaker performance 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. During the last referendum campaign, however, the main growth-enhancing policies proposed were tax cuts or spending increases. In the near term at least, such policies would tend to increase rather than reduce Scotland’s budget deficit.

Finally there are significant issues raised by Brexit. This may mean independence would bring benefits that were not as relevant when the expectation was that the UK would remain in the EU. An independent Scotland could seek to rejoin the EU and/or the European single market, which would bring it the benefits of preferential access to European goods and services markets. If businesses responded to the UK’s exit from the single market by shifting investment or employment out of the country, an independent Scotland might be able to exploit its geographical, cultural and institutional proximity to the UK to attract a good chunk of that. But an independent Scotland in the single market might face bigger trade barriers with the rest of the UK. Given that Scotland’s trade with the rest of the UK is worth around four times as much as its trade with the rest of the EU, the costs of these sorts of barriers could outweigh the gains from better access to European markets. The impact of Brexit on the economic and fiscal impacts of Scottish independence is therefore far from clear.

What is clear though is that the future trading relationships between Europe, the UK, and a potentially independent Scotland, and businesses’ responses to these relationships, are an important issue for debates about the economic and fiscal implications of Scottish independence. If a second independence referendum is held, in-depth analysis of the options and their impacts would be a vital contribution to public debate.

Notes on methodology for projecting Scotland’s fiscal position beyond 2015–16

In order to project forward the GERS 2015–16 figures to the period covering 2016–17 to 2021–22 using figures from the OBR’s March 2017 EFO, the following method is used:

  • Spending is projected on the basis that government spending in Scotland remains the same proportion (9.1%) of UK-wide government spending as in 2015–16.
  • 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 95.5% of the average for the UK as a whole, as in 2015–16.
  • 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 2015–16 at 78.5%.
  • 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 government net debt interest payments from our baseline projections for Scotland’s public spending.

We have chosen the assumptions on the basis of their simplicity and as a reasonable baseline case. As with any economic or fiscal forecast or projection, the projections outlined in this observation will differ from eventual outturns. This includes the OBR forecasts for the UK as a whole; and trends in spending and government revenues in Scotland relative to the UK. 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, during periods of public spending restraint, the Barnett formula tends to lead to spending levels in Scotland increasing relative to those of the UK as a whole. Second, Scottish Government plans to borrow additional money to fund capital investment mean Scottish Government spending may fall less between 2015–16 and 2021–22 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. Third, 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.


IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Tue, 30 May 2017 00:00:00 +0000
<![CDATA[A comparison of manifesto proposals on school spending in England]]>  

The Conservative, Labour and Liberal Democrat manifestos all commit to increase overall school spending in cash terms over the next parliament. Once expected inflation and student numbers growth are accounted for, Conservative plans would imply a real-terms fall in school spending per pupil of 2.8% between 2017–18 and 2021–22, which makes for a total cut of 7% between 2015–16 and 2021–22. Labour would increase school spending per pupil by 6% compared with present levels and leave spending per pupil in 2021–22 1.6% higher in real-terms than its historic high in 2015–16. Liberal Democrat plans would protect spending per pupil in real-terms at its 2017–18 level. All three major parties would continue the process of school funding reform, though with additional protections to ensure no school loses out in absolute terms. In this observation, we detail what these commitments would mean for the path of overall school spending in England and the prospects for continued reform of the school funding system.  

Increases to the overall schools budget

Under current plans for school spending in England, spending per pupil is set to fall by about 6.5% between 2015–16 and 2019–20. This fall increases to about 8% if we include some of the additional costs schools have faced in recent years such as higher pension contributions. Figure 1 shows that this would leave spending per pupil in 2019–20 at about the same real-terms level as it was a decade earlier, though still substantially higher than before the rapid increases over the course of the 2000s. Figure 1 then also shows how spending per pupil would evolve under each of the main parties’ commitments on school spending.

The Conservatives have committed to “increase the overall schools budget by £4 billion by 2022”. Once you strip out inflation, this equates to a real-terms increase in the schools budget of around £1 billion compared with the level in 2017–18. Taking account of forecast growth in pupil numbers this equates to a real-terms cut in spending per pupil of 2.8% between 2017–18 and 2021–22. Adding this to past cuts makes for a total real-terms cut to per-pupil spending of around 7% over the six years between 2015–16 and 2021–22.

Labour would reverse real-terms cuts to spending per pupil since 2015 and then protect it in real-terms over the course of the next parliament. This would require an increase in school spending of around £4.8 billion in 2017–18 prices compared with its level in 2017–18. If delivered, this would increase spending per pupil by 6% over the course of the parliament, leaving it about 1.6% higher in real-terms compared with its historic high in 2015–16 (the small real-terms increase results from compensating schools for some of the additional costs they have faced in recent years).

Under the Liberal Democrats’ commitment, spending per pupil would be frozen in real-terms over the course of the parliament, which would require a total increase in the school budget of around £2.2 billion compared with today.

These different commitments imply quite different levels of spending per pupil by the end of the next parliament. For instance, under Labour plans, we project secondary school spending will be about £6,500 per pupil, which compares with about £6,000 under Conservative spending plans.


Spending per pupil in primary and secondary schools: actual and plans

Notes and Sources: See Belfield, Crawford and Sibieta (2017).; Authors’ calculations based on Conservative, Labour and Liberal Democrat General Election Manifestos; GDP deflator and Department for Education Pupil Projections .


Uncapping public sector pay

The figures above exclude any additional money schools may receive from removing the 1% cap on public sector pay increases. Exactly how much schools might receive is, however, uncertain. Labour have committed to following the recommendations of Pay Review Bodies, whilst the Liberal Democrats have committed to increasing public sector pay in line with inflation. If Pay Review Bodies recommended increasing public sector pay in line with private sector earnings, IFS researchers have used current forecasts for average earnings growth to estimate that this would increase the cost faced by public sector employers by £9.2 billion in 2021–22, of which about 30%, or £2.8 billion, would be for schools. Increasing pay in line with inflation would increase costs faced by public sector employers, relative to the 1% cap, by £5.3 billion, or about £1.6 billion for schools.

In reality, fully compensating public sector employers would probably require less funding than these headline numbers as cash-terms growth in spending per pupil would already allow for some of the planned increases in public sector pay. Nevertheless, any compensation given to schools would probably equate to a significant further increase in the schools budget over and above what is planned. For instance, £2.8 billion equates to 7% of the current schools budget. Any additional funding could be seen as simply compensating schools for higher costs,  but one could also view such increases as providing for a higher quality of schooling than would otherwise be possible. For instance, the School Teachers Review Body has warned about the implications of the 1% pay cap on teacher recruitment and retention. Providing funding for higher pay increases could be one way to avoid such problems, though comes with a price tag.

National Funding Formula – Everyone’s a winner?

The outgoing government also proposed a national school funding formula in England to ensure similar schools receive similar levels of funding. Reforming the school funding system is long overdue, but doing so was always going to create winners and losers, at least in relative terms. Existing proposals came with transitional protections: no school would face cash-term cuts of more than 3% per pupil between 2017–18 and 2019–20 and no school would gain more than 5.6%. These protections cost £300 million and result in only 60% of schools receiving the funding level given by the formula in 2019–20.

All three major parties have decided to go further and have committed to ensuring that no school will face cash-terms cuts as a result of the national formula. This requires around an additional £350m of funding and leads to an asymmetric roll-out of the formula; the winners will still gain, but the losers will still not see their budget fall in cash terms. The above calculations ignore this as such funding is temporary and will gradually reduce as more schools move on to the formula. Though this might not happen that fast.

Starting with current policy, Figure 2 shows the proportion of schools due to receive gains and losses of given amounts as a result of the National Funding Formula between 2017–18 and 2019–20. The Figure shows three scenarios under the current policy of a cash-terms freeze in overall school spending per pupil: with no transitional protections; with the existing protections and with the additional protection that no school will face cash-terms losses. The striking result is that with a cash-terms freeze in overall spending and the additional protections that have been announced only 40% of schools are on formula in 2019–20.

As discussed above, each of the major parties has also committed to increase overall spending on schools. These increases in spending have the potential to ease the implementation by allowing some schools to move towards the formula level while their budgets remain protected in cash-terms or grow at a slower rate than the headline increase. 

Conservative proposals imply a cash-terms increase in spending per pupil of around 4.5% between 2017–18 and 2021–22. This could allow them to get around 25% more schools on to the formula by 2019–20 if some see their funding frozen in cash-terms.  However, this freeze would be on top of cash-terms freezes for some schools from 2015 onwards.


Figure 2. Cumulative distribution of changes between 2017–18 and 2019–20: primary and secondary schools

Notes and Sources: See Figures 10 and 11 in Belfield and Sibieta (2017).

In principle, Labour and the Liberal Democrats have more headroom to get more schools on to the formula faster, as their plans imply larger increases in spending per pupil. However, they could only do so by delivering below inflation increases in funding per pupil to some schools. Labour’s proposals imply school funding per pupil increasing by 9.5% in cash-terms between 2017–18 and 2019–20. Under one hypothetical scenario set out in Figure 2 above, Labour could get most schools on to the formula by 2019–20 by simply imposing a cash-terms floor of 0% and a maximum real-terms increase of 5.6%. Only about 5% of schools would actually experience a cash-terms freeze and 80% would be on formula by 2019–20. However, this policy would still result in a significant number of schools facing real-terms cuts between 2017–18 and 2019–20. If instead Labour chose to protect all school per pupil budgets in real-terms this would preserve the current distribution of school spending and only about 40% of schools would be on formula by 2019–20, with further adjustment presumably required after 2019–20.

This highlights the difficulty of introducing the NFF at a time of constrained overall funding. Such a reform necessarily creates relative winners and losers, but with no overall growth in funding the reform must create absolute losers to transfer the additional funding to those that gain. Imposing additional protections does not solve this problem; it just delays schools’ transition to the true formula level.

Luke Sibieta, an Associate Director at IFS and an author of this piece, said:

“The commitments made by each of the main parties would imply quite different paths for school spending in the next parliament. Labour would increase spending per pupil by around 6% after inflation over the course of the parliament, taking it to just above its previous historic high in 2015. Proposals from the Conservatives would lead to a near 3% real terms fall in spending per pupil over the parliament, taking it back to its level in 2010.”



]]> Fri, 26 May 2017 00:00:00 +0000
<![CDATA[Social care – a step forwards or a step backwards?]]> Yesterday, the Conservative Party proposed changes to the rules governing who is eligible for government funding for social care, and backed away from a lifetime cap on care costs. In this observation, we discuss those changes and lay out their potential effects. Taking the population of people in their 70s in England we estimate that, on becoming in need of care in their own home, 12-17% would be eligible for state support under current rules but would not be eligible under the new rules proposed by the Conservatives. Others would find they needed to use more of their own wealth to fund the costs of care in their home before the state stepped in.

What is happening to funding for social care?

Unlike health care, social care services in England are not provided free of charge for everyone. Rather, local authorities in England provide financial assistance to adults who have insufficient financial means to fund their own use of care services.

Spending on social care in England fell by 8% in real-terms between 2009­-10 and 2016-17 as a result of cuts to funding for local government. In response to widespread belief that the system is underfunded, current plans are for councils to receive enough money by 2019-20 to reverse all the cuts that have been made since 2009–10 if they choose to do so. The Labour Party have pledged to offer a little more money for social care than is currently planned, and the Conservative manifesto also suggests some additional funds would be made available.

How common is it to need social care?

In England in 2015–16, 783,000 adults were receiving local authority contributions towards the cost of long-term care – either in their own home or in a care home. 538,000 of these were over 65, 6% of the over-65 population in England. [1] Of those half a million people, most (310,000) were receiving care in their home (at an average cost of £14 to 31 an hour), with the remainder (228,000) receiving residential care (at an average cost of £553 a week) [2].  

These numbers might sounds relatively small, but it is common for people to require social care towards the end of their life. Analysis for the Dilnot Commission in 2010 found that 45% of 65 year olds could expect to spend (or have spent on them) more than £25,000 on care services, and 10% could expect to spend more than £100,000[3]. Few know in advance if they will be among those needing to spend much on long-term care.

Who is eligible for help with care costs now?

Adults are potentially eligible for public funding if they have assets worth less than £23,250. If their care needs are to be met by care provided in their own home, the value of their house is excluded from the means test. If they are in need of residential care then the value of their house may be included in the calculation of their assets.

Since 2015-16 individuals who need residential care but are not entitled to financial assistance from their local authority because the value of their house puts them over the asset threshold have had the option of making a deferred payment agreement with their local authority, which means that they do not have to sell their house during their lifetime to pay for their care costs. Their local authority would then fund their care during their life, and reclaim these costs after the individual’s death.

Even if someone is eligible for public funding on the basis of the value of their assets, local authorities are still allowed to ask them to pay some of the cost of their care on the basis of their income. When deciding how much to ask for, local authorities must leave everyone with at least a minimum amount of income (set nationally) which depends on whether they are receiving residential or home care.

What are the proposed changes?

The Conservative party have proposed two changes to the rules governing eligibility for public funding. The first is to raise the asset threshold from £23,250 to £100,000 for both residential care and care provided in the home. The second is to include housing wealth in the measure of assets considered when assessing eligibility for support for home care (and correspondingly expand deferred payment agreements to cover home care too).

The tables below look at the assets of those aged between 70 and 79 in England, and examine how these changes would potentially affect the eligibility of this group to state support for residential and home-base care should they end up needing it. The current rules governing exactly which assets count in the means-test are complicated , and quite what rules the Conservatives are proposing is unclear. We model two scenarios:

1. For those living with a partner, half of the couple’s total wealth is counted (including housing wealth where appropriate).

2. For those living with a partner, half of the couple’s non-housing wealth is counted, along with their total housing wealth (where appropriate).

The difference between these scenarios occurs when a couple has housing wealth between £100,000 and £200,000. If one member of the couple needed care, it is unclear whether the whole value of the house would be counted (rendering them ineligible) or only half of the value (leaving them eligible). In both scenarios, all the wealth of single individuals is included, and non-housing wealth is split evenly between couples.

Table 1 focuses on eligibility to public funding for residential care. It shows that 15% of individuals would see their eligibility to public funding unaffected by the proposed change, as they have assets below the old threshold (and therefore also below the new, higher, threshold). The remaining 85% could potentially gain from the Conservative proposal if they needed residential care – an additional 5-16% would be eligible for public funding immediately, and the remaining 69-80% would see state support kick in sooner if they ended up using their assets to pay for residential care.

Table 1. Assets for the purposes of existing and proposed means-test for state support with costs of residential care (% of 70-79 year olds in England)

Level of total assets

Assuming all housing wealth counts for both members of couple

Assuming housing wealth split 50-50










Source: Authors’ calculations using the English Longitudinal Study of Ageing, wave 7.

Table 2 examines eligibility to public funding for home care. It shows that:

  • around a quarter would see their potential eligibility unaffected, as they are below both the old and the new asset thresholds.
  • between a quarter and a third of 70-79 year olds would potentially lose eligibility for help with home-based care: they have less than £23,250 in non-housing wealth, but over £100,000 when housing is included.
  • A further 45% of the group could find that public funding does not kick in as quickly if they end up using their wealth to fund care costs. Under the old system, they would be eligible as soon as non-housing assets fell below £23,250, regardless of how much housing wealth they have. Under the new system, they could then also be required to spend any housing wealth in excess of £100,000 before becoming eligible for public funding.

Table 2. Assets for the purposes of existing and proposed means-test for state support with costs of home care (% of 70-79 year olds in England)


Assuming all housing wealth counts for both members of couple

Assuming housing wealth split 50-50


Total assets below £100,000

Total assets above £100,000

Total assets below £100,000

Total assets above £100,000



Non-housing assets below £23,250





Non-housing assets above £23,250





Source: Authors’ calculations using the English Longitudinal Study of Ageing, wave 7.

As explained above, it is not just wealth that affects individuals’ eligibility for state support with the costs of social care, but also income. This means that some of the group we identify as potentially losing out as a result of this reform may in practice be no worse off, as they have incomes high enough to be ineligible for publicly funded care.

To give a sense of how important that is likely to be the Figure below takes the group of people who pass the old asset test but not the new asset test (a quarter to a third of 70-79 year olds in England), and shows the number of hours of home care per week they could be expected to self-fund (based on their income) before the local authority would help cover the costs of any additional hours.

It suggests that 20% of this group have incomes sufficiently low that they would not be expected to pay for any home-based care, and a further 35% would receive public funding if they needed more than 5 hours of a care each week. In other words, the lowest-income half of the group who fail the new asset test (but passed the old one) would be likely to lose out if they needed care. Whether the other half lost out would depend on how many hours of home care they need.

To summarise, a quarter to a third of 70-79 year olds in England would pass the old asset test for home care but fail the new one proposed by the Conservatives. Of those, at least half would have got public funding for home care under the old system if they needed it (the other half have too high an income to be eligible). Hence the new policy might directly affect the eligibility of 12-17% of this group to state support with costs of care in their home.

Figure. Self-funded hours of care each week among those who pass the old asset test for home care, but fail the new asset test

Figure. Self-funded hours of care each week among those who pass the old asset test for home care, but fail the new asset test

Notes: Income split evenly between couples. Figure presented for those who pass the old asset test but fail the new asset test in our second scenario. Figures for those affected in the first scenario are extremely similar.

Source: Authors’ calculations using the English Longitudinal Study of Ageing, wave 7, national minimum income thresholds from local authority circular January 2017 (Social care - charging for care and support), and unit costs of day care (England average for external provision) from NHS Digital, Personal Social Services: Expenditures and Unit Costs England 2015-16. 

What is the impact on public spending?

64% of local authority spending on care for the over-65s is for residential care, with the remainder spent on home care. The Conservative plans would tip this balance towards a greater fraction of spending on residential care –  increasing public funding for residential care (by raising the asset threshold) but decreasing public funding for home care (by including housing wealth in the asset test).

It is not possible to be confident whether the change would increase or decrease overall spending relative to the current system on care on the basis of publically available data. What is clear is that the proposal is less generous than the version of the Dilnot commission’s recommendations that have already been put into law

What about the Dilnot Commission?

In 2011, the Dilnot Commission proposed a new system of social care funding which included a cap on the amount an individual would have to pay towards the cost of their social care over the course of their life. Once they have met the cap, all further care costs would be covered by the state.

 At the time, the government accepted the basic idea behind the proposal, though it planned to set the cap at £72,000 (higher than the £25,000 to £50,000 (in 2010 prices) proposed by the commission). Implementation of the reform was planned for 2016 and then pushed back to 2020. The Conservative manifesto has now dropped the planned reform entirely.

What has been lost by abandoning the cap?

The fundamental thing to understand about the funding of social care is that it is an insurance problem.  Individuals face a low but meaningful risk of incurring very high care costs – 10% of adults age 65 could face costs of more than £100,000. People don’t know whether it will happen to them, and lots of these care costs don’t arise until the end of life.

Ideally, you’d like to buy some insurance reasonably early on in your life that would pay out if you ended up one of the unlucky few. But insurance markets don’t exist that will let you buy social care insurance so far in advance.

A life-time cap on care costs, as proposed by the Dilnot Commission, is a solution to the insurance problem. It is effectively a form of social insurance, funded from general taxation. It may also make it easier for a private market to emerge that would offer insurance against care costs up to the cap.

By contrast, the Conservative plan makes no attempt to deal with the fundamental challenge of social care funding. That is the big problem – not how many people might win or lose.


[1] HSCIC Community Care Statistics 2015–16, table LTS001a. Figures exclude ‘commissioned support only’.

[2] HSCIC Personal Social Services Expenditure and Unit Costs 2015-16. Hourly cost of day services differs depending on whether provision is directly provided by local authority employees or commissioned externally. Day care costs are average across England for all age groups. Residential costs are specific to the over 65s.

[3] Fairer Care Funding Volume II: Analysis and evidence supporting the recommendations of the Commission on Funding of Care and Support, July 2011

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Fri, 19 May 2017 00:00:00 +0000
<![CDATA[Moving from a Triple to a Double Lock does little to long-run state pension affordability]]> Today’s Conservative manifesto announced that from 2020 onwards the state pension would be increased over time in line with average earnings or inflation whichever is highest – the so-called ‘Double Lock’. In this observation we show that this is very similar to sticking with the Triple Lock, and does little to resolve the pressures an ageing population will put on the public finances over the years to come.

The Figure shows projected state pension spending as a share of national income up to 2066–67 (based on the latest Office for Budget Responsibility Fiscal Sustainability Report, and not incorporating the recommendations of the recent Cridland Review) under three uprating rules: average earnings, the double lock and the triple lock. As we have noted before, half of the increase in state pension spending forecast over the next 50 years (0.9% of national income, or nearly £20 billion in today’s terms) is explained by the triple lock, rather than other factors.

As the figure makes clear, moving to a double lock does very little to help. State pension spending in fifty years time is only 0.2% of national income lower (less than £5 billion in today’s terms). In other words, moving to a double lock undoes only around a quarter of the damage done by the triple lock to the long-run sustainability of the public finances. So with the double lock in place spending on the state pension would still be projected to increase by 1.6% of national income (a little over £30 billion in today’s terms) over the next fifty years, with over 40% of this increase being explained by the double lock  (relative to increasing in line with average earnings) rather than other factors.

The fundamental reason for this is that it is pretty rare for both average earnings and inflation to be below 2.5%. Hence getting rid of the 2.5% element of the triple lock does little to change the projected long-run generosity of the state pension.

Figure. Projections of state pension spending

Figure. Projections of state pension spending 

The frustrating thing is that it is possible to insure pensioners against the state pension ever falling in real terms without the ‘ratchet’ effect that means both a triple-locked and double-locked state pension would rise faster than earnings or prices over the long run. The answer is the ‘smoothed earnings link’ described here which  makes sure the state pension never falls in real terms, and rises in line with earnings growth over the long run.  

We will be holding a press event on party manifestos on Tuesday 23 May 2017 at Church House Westminster, 10:00-12:00. To register, please email

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Thu, 18 May 2017 00:00:00 +0000
<![CDATA[Income tax and benefits: the Liberal Democrats and Labour compared]]> The Liberal Democrat manifesto published today includes commitments to increase all rates of income tax by 1 percentage point as well as cancelling a number of significant benefit cuts and reversing others. In this observation we provide an overview of these proposals and compare them to the income tax and benefit proposals published by Labour yesterday.

Income tax

The Liberal Democrats have proposed a rise in income tax that is very broad-based, but relatively modest per person, raising all rates by 1 percentage point. It would raise around £6bn per year and would affect the roughly 30 million adults who pay income tax. Bear in mind that nearly 45% of adults pay no income tax at all, as their incomes are below the personal allowance.

This would clearly hit people further down the income distribution than Labour’s income tax proposals announced yesterday and analysed in detail here, since Labour’s policy would affect only those individuals on at least £80,000 per year: the top 2% of adults or 4% of income taxpayers. The Figure below compares the increases in income tax bills at each level of taxable income under the two parties’ proposals. It shows that:

  • The lowest-income 24 million adults would be unaffected by either proposal;
  • The next 28 million (up to £80,000) would be affected by the Liberal Democrat proposal only;
  • The next 500,000 (up to £97,125) would be affected by both but would lose more under the Liberal Democrats;
  • The top 800,000 would be affected by both but would lose more under Labour.

It is important to bear in mind that some of the lower-income families who would lose from the Liberal Democrats’ income tax rise would gain substantially from the benefit changes the Liberal Democrats propose, discussed below.

The revenue raised by the Liberal Democrats’ income tax rises, at around £6bn per year, is considerably more certain than the revenue raised by Labour’s. Essentially the explanation is that the revenue raised by the Liberal Democrats’ income tax rise would be far less dependent on a small group at the very top, whom evidence suggests are relatively responsive but for whom there is also much uncertainty about the degree of responsiveness. As we discussed in detail yesterday (see final section here), after accounting for the fact that some taxpayers will respond to higher tax rates by reducing their taxable incomes, the middle of a plausible range of estimates for the revenue raised by Labour’s proposal is probably around £2-3 billion. But it could easily raise something like the £4½ billion that Labour expects or nothing at all.

Figure. Annual losses from Liberal Democrats’ and Labour’s proposed income tax rises

Source: Authors’ calculations using HMRC income tax statistics.


The Liberal Democrats have also proposed a number of increases in benefits levels relative to current policy plans. Most are reversals of benefit cuts implemented since 2010, or cancellations of cuts that have not yet been fully implemented but are in the pipeline: in fact, the Liberal Democrats propose to reverse nearly all of the significant cuts to working-age benefits that are currently planned over the next few years.

The Table below sets out the Liberal Democrat benefit proposals alongside those made in the Labour manifesto yesterday, along with their estimated long-run cost. For both parties, the policies listed in the Table cover around 90% of the total proposed increases in benefit spending (the other benefit changes are listed at the bottom of the observation).

The Liberal Democrats’ biggest commitments are the reversal of cuts to child tax credit , including the limiting of support to two children (at a long run cost of £5 billion a year), the ending of the freeze on most working-age benefits (£3 billion), and the reversal of work allowance cuts in universal credit (£3 billion). Note that the costings given in the  Liberal Democrat manifesto are for the short run (2019-20), but the first of these policies costs significantly more in the long run, as the cuts only affects new births (and some new claimants).

In comparison with the Liberal Democrats, Labour’s plans for benefits are modest – their manifesto allocates £2 billion to reviewing the cuts to Universal Credit, and another £2.6 billion to other benefit changes.[1]  It does not allocate any additional funds on top of that to ending the freeze on working-age benefits or reversing cuts to child tax credit.  Both parties commit to increases in benefits that come with a much lower price tag, such as scrapping the so-called “bedroom tax” (at a cost of £400 million a year) and restoring automatic housing benefit entitlement to 18-21 year olds (around £30 million).

Table. Liberal Democrat and Labour benefit proposals

Notes: The costing for 'reverse cuts to UC' refers to the cuts to work allowances announced in Summer Budget 2015. The Liberal Democrat manifesto states their intention to reverse these cuts, while the Labour manifesto allocates £2bn per year 'for review of [UC] cuts and how best to reverse them'. Labour’s manifesto commits to increasing carer’s allowance to the level of job seekers’ allowance (at a cost of £0.2bn), while the Liberal Democrats’ commits to extending eligibility of carer’s allowance (at a cost of £0.4bn). For a list of policies in the manifesto not included above, see the notes at the end of the observation.
Sources: See end of observation.


Both the Liberal Democrats and Labour propose increasing income tax. While the Liberal Democrat proposal would affect the highest-income half of adults, Labour’s proposal would only affect the highest-income 2%. But the revenue from Labour’s plans is vastly more uncertain, and highly likely to be lower than under the Liberal Democrats.

Both the Liberal Democrats and Labour propose increases to benefits. But those proposed by the Liberal Democrats are much larger – reversing nearly all of the cuts planned for the next few years.

This observation focuses just on income tax and benefits, so is only part of the bigger picture. Both parties propose other tax rises (such as higher rates of corporation tax) which would also ultimately affect the incomes of UK households. And both parties propose increases in public spending, particularly on education and the NHS.

Notes and sources

Further notes to the Table:

Benefit policies in the manifestos but not included in the table: Continuing the triple lock on the state pension (both parties), increasing parental leave (both parties), extending pension credit to certain groups (Labour), uprating the state pension for British overseas pensioners (Labour), increasing UC for some two earner couples (Lib Dem), uprating JSA & UC for those aged 18-24 in line with the minimum wage (Lib Dem), uprating local housing allowances rates in line with local rent changes (Lib Dem).

Sources for the Table:

Authors' calculations using TAXBEN and the Family Resources Survey 2015-16; The Labour Party Manifesto 2017 (; The Liberal Democrats Manifesto 2017 (, Hood, A., and Waters, T., The impact of tax and benefit reforms on household incomes, IFS Election 2017 Briefing Note BN196, April 2017, (; Budget 2013 Table 2.2 (; Budget 2016 Table 2.2 (; Department for Work & Pensions, Equality Analysis: PIP assessment criteria, February 2017 (; Department for Work & Pensions, Benefit expenditure and caseload tables 2017 (

[1] The Labour manifesto allocates “£2 billion of additional funding for Universal Credit for review of cuts and how best to reverse them” though it was suggested later that some of the funds could be used instead to ameliorate the benefit freeze.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to


]]> Wed, 17 May 2017 00:00:00 +0000
<![CDATA[More hours, more children, and more spending: early years and childcare proposals from Labour and the Liberal Democrats]]> Labour's proposals

The Labour party is proposing a major extension to government support for early childhood education and care (ECEC) in England. Done well, spending on the early years can be an excellent investment, with benefits for children’s development and the wider economy. However, the Labour party’s proposals represent a significant new cost to the public finances, while the available evidence suggests that the benefits for children’s development and mothers’ employment are likely to be modest. There is also a great deal of uncertainty around both the costs and the benefits; since Labour is proposing large extensions to both the number of children and the number of hours covered by the free entitlement, understanding how the costs, benefits, and take-up will vary by age is crucial, but there is little clear evidence available as a guide.

Policy changes

Currently, the government supports early childhood education and care in four ways.

  • There is a ‘free entitlement’ to 15 hours of childcare for all three- and four-year-olds as well as two-year-olds in the 40% most disadvantaged families (from September 2017, this is due to rise to 30 hours a week for the approximately 390,000 families of three- and four-year-olds living in two-earner or single working parent homes).
  • The tax credit system supports disadvantaged families with childcare costs through the childcare component of Working Tax Credit and Universal Credit.
  • Working parents are eligible for childcare vouchers through their employer that are not subject to income tax or national insurance contributions.
  • These vouchers are being phased out in favour of the tax-free childcare accounts that working parents have been able to use from April, which offer a 25% subsidy on care spending.

All told, the government currently spends about £6.5 billion on childcare through these programmes.

The Labour party is proposing to vastly expand the first of these supports – entitlement to free childcare – and to introduce its own subsidy programme for additional hours of childcare. In exchange, they would eliminate spending on demand-side subsidies – through the tax credit system, childcare vouchers, and tax-free childcare – for children under five (it’s unclear what would happen to the childcare support these programmes offer for older children). In particular, the party is proposing to offer:

  • An additional 15 hours per week of free childcare for all three- and four-year-olds, not just those in working families;
  • 15 hours of free care for all two-year-olds (not just the most disadvantaged) from 2020-21, rising to 30 hours a week the following year;
  • Subsidised additional care over and above this extended free entitlement; and
  • Some free childcare for one-year-olds and extending maternity pay to 12 months in the longer term.

Labour has promised to ensure that these childcare policies would be delivered in high-quality settings with a graduate-led workforce. Since more qualified staff require higher wages, for these quality improvements to be sustainable the rate at which childcare providers are reimbursed would have to rise. Although the party has not committed to a specific hourly rate, they reference a report by the Family and Childcare Trust that suggests that a graduate-led workforce would require reimbursement rates of £7.18 for three- and four-year-old places and £8.62 for one- and two-year-olds. Under current plans, these rates would be £4.88 and £5.39 respectively from September this year, so this would represent a more than 50% increase for providers.

Programme cost

These are substantial changes which would sweep away many elements of what is currently a very complex system of subsidies, effectively replacing them with freely available childcare for two-, three- and four- year olds. All together, the changes to the free entitlement would see at least 400,000 families with two-year-olds eligible for free childcare for the first time and would offer more free hours to another 1.6 million children. If the proposed subsidised hours are offered to all two- to four-year-old children, 2 million families would be eligible for cheaper childcare.

The costs are correspondingly large. By 2021-22, Labour estimates that these programmes would require around £5.3 billion in funding (including additional spending on supporting new childcare places). This would increase spending on childcare by around 70% over current projections.

However, the costs might rise further in later years. First, these figures exclude the costs of offering free or subsidised care to one-year-olds and of extending maternity pay to 12 months. Even the cost of policies that are included – 30 free hours of childcare for all two- to four-year-olds and additional subsidised care – might not be fully captured by spending in 2021-22. The full 30-hour entitlement for two-year-olds will only become available in that year. Since take-up of these free hours is expected to rise for the first few years of the programme, the full running costs might not be felt until the middle of the next parliament. For example, take-up of the current offer for disadvantaged two-year-olds rose from 58% in 2015 to 68% the following year; a similar 10 percentage-point increase in take-up of Labour’s two-year-old offer would imply around £650 million in additional spending each year.

The case for subsidy

The cost of these policies is a good indication of their ambition. Labour’s programme represents a complete overhaul of state support for childcare in England. It would replace a system that largely channels resources through parents with a programme to directly fund childcare providers. It would replace the current patchwork of subsidy programmes with a single system of free and subsidised care. And, most importantly, it would significantly increase the scope of universal benefits for children from the age of 1 to the age of 4.

Evaluating whether these expensive changes would be in the public interest requires understanding the goals that such a major overhaul of the childcare system is intended to achieve.  That might be to improve children’s outcomes, help parents to move into work, or both.

Based on its manifesto, the Labour party appears to be more interested in the first of these goals. It promises that a better-educated early years workforce would ‘improve child development’, while party leader Jeremy Corbyn promised to ‘invest in our young people’ during his manifesto launch speech.

Impacts on children

Existing evidence suggests that the original 15-hour free entitlement had small academic benefits for children, but these didn’t last in the longer term. Labour’s proposals would mean that free childcare would be available from an earlier age (two) for many children and would be available for more hours for three- and four-year-olds. The extent to which this would improve these children’s later outcomes remains uncertain, and will depend on the extent to which families take up the additional hours and the quality of care that will be on offer.

Surveys estimate that only about 63% of two-year-olds use formal childcare – this means that there is scope for Labour’s proposed universal free entitlement to move more children into early education, which could benefit their academic attainment. However, existing research is largely unable to tell us about the “optimal” age to start childcare. While the EPPE study suggests that starting childcare before the age of three is linked to better intellectual development, other studies suggest that starting before age two could actually harm children’s development.

The impact of Labour’s proposals for 3- and 4-year-olds is quite different. Today, 98% of this age group is already taking up their free entitlement place. As a result, if anything, Labour’s policies would increase the number of hours of early education used (rather than the number of children using it). The impact this will have on children is unclear: evidence on the “right” amount of time for children to attend centre-based childcare is limited, and some studies even find that children who have longer days in childcare tend to have worse behaviour.

International evidence suggests that the benefits of childcare are often stronger for the most disadvantaged children, meaning that early education programmes can help to reduce inequalities in children’s development. Unusually, however, the researchers investigating the original three-year-old free entitlement in England did not find larger effects among disadvantaged children than among their more affluent peers. They suggest that one possible explanation for the small impacts of the policy was that the expansion of free entitlement was delivered through childcare providers in the private, voluntary, and independent sector, where quality is on average lower than in the maintained sector.

High-quality care

In this respect, Labour’s willingness to invest in order to bring up the quality of childcare settings is welcome. If done well, it means that the reforms they propose could have larger impacts on children. Labour claims that it would ‘transition to a qualified graduate-led workforce, by increasing staff wages and enhancing training opportunities.’ On the one hand, this would presumably help the sector to attract and retain higher-quality workers. And by moving towards a system of funding providers directly, the government might have more leverage to ensure that the quality provided by all settings meets minimum standards. 

On the other hand, it’s unclear which qualifications and training Labour hopes that early years’ staff will take. While the EPPE study found that childcare settings with more qualified staff also provided better-quality care, the qualification framework has undergone numerous changes since it was published in 2003. More recent research suggests that the link between staff qualifications and children’s academic achievement is weaker now than it used to be. 

Parental employment

Although the Labour party seems to be focused on how early years provision can improve child outcomes, it’s also important to take into account the potential for additional benefits for parental labour supply. Evidence on this is also patchy.

Previous work by IFS and ISER researchers suggests that offering free full-time childcare can help mothers to work. However, these effects are only found when offering full-time childcare (between 30 and 35 hours per week) to a mother’s youngest child. Even then, the gains are small – after a year of full-time provision for their youngest child, mothers were only 3.5 percentage points more likely to be in work relative to mothers whose youngest child was at the end of their first year of part-time entitlement. This means that extending free childcare from 15 to 30 hours for one cohort of 690,000 four-year-olds moved 12,000 mothers into work.

The expected effect that offering full-time care to a two-year-old would have on her mother’s chances of being in work is unclear. On the one hand, parents of younger children are more likely to prefer to look after their child themselves or use help from family and friends. Almost half of parents of two-year-olds who are not using formal childcare say that this is because of their preferences (compared to around 27% who cite cost as the biggest obstacle). We might expect that making care freely available for to these families would, at least in the short run, do relatively little to change their preferences and move these mothers back into work. 

On the other hand, extending the free entitlement to two-year-olds might have larger effects on mothers’ employment than at older ages. By providing childcare support earlier on, it would help mothers to return to work sooner, which could make it easier for them to find a job. A shorter time away from work could also help mothers continue to develop skills and experience, raising their earnings in the longer term. And the flexibility from subsidised additional hours would help mothers whose jobs don’t fit neatly into the current restrictions on when free entitlement hours can be taken.


The Labour party has proposed an ambitious new early years agenda that would fundamentally change the English government’s approach to childcare. Labour would replace the current system, targeted at working and disadvantaged families, with a more generous free entitlement available to all children aged two, three, and four. At the same time, they would dramatically increase the rates received by childcare providers, in the hopes that higher payments and more opportunities for training would improve the quality of the early years workforce.

This reform seems to be aimed primarily at benefiting children’s development; however, existing research cautions that the gains are likely to be modest, and the evidence on extending childcare for more hours and to younger ages is mixed at best. The long run effects of such a major change in policy on parents’ employment are also genuinely uncertain.

There are arguments in favour of universal benefits on this scale, many parents would no doubt welcome such a change, and the current complex patchwork of demand- and supply-side subsidies could certainly do with an overhaul. But the costs of this policy are both very large and uncertain. It is incumbent on those proposing this scale of policy change to do a better job of setting out who exactly they hopes to benefit, and how.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

Liberal Democrat proposals

The Liberal Democrats have also made major promises on extending the childcare system in England. Their manifesto sets out two goals for these promises: more generous childcare support will ‘help parents afford work’ and ‘impact on children’s attainment as they enter school.’

Like the Labour party, the Liberal Democrats are proposing a system that would eventually become much more generous to parents. In particular, the party is proposing:

  • 15 hours of free entitlement for all two-year-olds;
  • 15 hours of free childcare for children aged nine to 24 months in working families;
  • A long-term goal of 30 hours’ free childcare for two- to four-year-olds and for younger children in working families; and
  • An increase in the Early Years Pupil Premium to £1,000 per pupil per year.

There are obvious similarities between some of these policies and the proposals put forward by the Labour party to increase the availability of free childcare. However, the Liberal Democrats’ proposals don’t represent quite the overhaul that Labour is suggesting. They will start by extending part-time entitlements to new children, while Labour proposes both hours and coverage extensions at the same time. They have stopped short of promising a fully universal system, instead targeting additional care for the youngest children to those in working families and increasing the funding rate for places taken up by disadvantaged children. And it seems that the Liberal Democrats would keep the current system of subsidies through tax-free childcare, childcare vouchers, and tax credits.

As we have discussed in this observation, the extent to which these policies will improve children’s outcomes is unclear; certainly, the available evidence suggests it is unlikely that the changes they have proposed will have the ‘huge impact on children’s attainment’ that is promised in the manifesto unless the quality of care offered is very much improved.

While the Liberal Democrats rightly note the importance of high-quality provision, the measures they propose to increase the quality of provision are less comprehensive than those put forward by Labour. For example, the support they are offering through the Early Years Pupil Premium (EYPP) is less generous than the higher reimbursement rates that Labour seems to be proposing. The EYPP provides additional funding to disadvantaged three- and four-year-olds in early education, and is currently worth up to £300 per eligible child. Raising it to £1,000 per eligible child works out to £1.23 per hour in additional funding for disadvantaged children, a rise of about 25% over current average funding levels for three- and four-year-olds. This is about half as generous as Labour’s proposed 50% increase in the reimbursement rate, and would apply to fewer pupils.

The labour supply effects of the Liberal Democrats’ proposals are similarly uncertain. Research finding that only full-time care for the youngest child raised a mother’s chances of being in work (and even then by a relatively small 3.5 percentage points relative to the last term in part-time care) is not encouraging. However, as with the Labour party’s proposals, it is possible that extending free childcare to younger ages will reduce the amount of time that mothers spend out of work after the end of their maternity leave. Since longer absences from the labour market make it harder to find a new job, it’s possible that policies aimed at keeping mothers attached to the labour market might have larger benefits for labour supply than have been found at older ages.

As with the Labour party’s proposals, costing the Liberal Democrats’ proposals is challenging without making strong assumptions about take-up rates and future increases in funding. The party has allocated £5.5 billion for education and the early years spending in 2019-20, but this funding envelope includes other policy commitments such as extending universal free school meals to all primary pupils, protections for school and further education budgets, and support for professional development for teachers. Based on estimates of the cost of these other policies, it seems that the Liberal Democrats are expecting to spend just over £2 billion in 2019-20 on their commitments to extend the free entitlement and increase the Early Years Pupil premium. Of this, we would expect around £900 million to be spent on the extension of 15 free hours to all two-year-olds (assuming current funding rates, similar levels of take-up to what is seen among the 40% most disadvantaged two-year-olds who are already eligible for the offer, and not accounting for savings elsewhere in the childcare system).

This observation will be updated in due course if child care policies are announced by other parties

]]> Wed, 17 May 2017 00:00:00 +0000
<![CDATA[Labour Party manifesto: background information]]> The Labour party is due to publish its June 2017 election manifesto later today. Below you will find detail on recent IFS analysis, which may help to contextualise and assess some of their policies. All analysis has been produced with funding from the Nuffield Foundation and can be found on our dedicated 2017 election micro site:

1. Policy: increase the minimum wage to reach £10 per hour for all employees (excluding those eligible for the apprentice rate) aged at least 18 in 2020.

In Minimum wages in the next parliament, by Jonathan Cribb, Robert Joyce and Agnes Norris Keiller, we set out some of the potential effects of introducing this policy.

Jonathan Cribb, a Senior Research Economist at IFS, and an author of the report, said, “Carefully set minimum wages can be a useful policy tool for governments seeking to help those on low wages. But at some point, higher minimum wages will reduce the employment of lower skilled workers. Since we do not know where that point is, sudden large increases are risky. They endanger the jobs of those they seek to help.  Labour's proposal to rapidly extend a £10 per hour minimum wage to 18-24 year olds, regardless of the state of the economy in 2020, is a particular gamble. Given how harmful periods of unemployment can be for young people, the long term costs could be considerable.”

2. Policy: scrap tuition fees

We have a detailed IFS ‘observation’ on this policy: Labour’s Higher Education proposals will cost £8bn per year, although increase the deficit by more. Graduates who earn most in future would benefit most. In it, we highlight the expected short-run impact on the public finances, the impact on graduates and the wider policy implications.

Jack Britton, a Senior Research Economist at IFS, said: “This commitment to scrap tuition fees follows previously announced plans to bring back maintenance grants for the poorest students. Both would represent a major reversal of the last 20 years of Higher Education (HE) policy. They would increase government borrowing by nearly £13 billion in the short run and by around £8 billion a year in the longer-term. Those who go on to earn the most would benefit the most from this policy, as they are the ones most likely to have repaid their loans in full under the old system.”

3. Policy: increase schools spending

In an IFS ‘observation, Labour’s proposed boost to education spending, Jack Britton and Luke Sibieta, compare Labour’s education spending commitments to existing government plans.

In Election battleground in the playground?, Luke Sibieta outlines the potential implications of freezing school spending in real-terms after the election. 

Luke Sibieta an Associate Director at IFS said: “School spending was protected by the coalition government but is now due to fall in real terms per pupil for the first time in decades. Labour’s plans would, at substantial additional cost, not only reverse these planned cuts but provide substantial additional funding to schools.”

4. Policy: increase corporation tax back up to 26%

IFS analysis by Associate Director Helen Miller shows that Labour’s reversal of corporate tax cuts would raise substantial sums but comes with important trade-offs.

Helen Miller, an Associate Director at IFS, said: “A rate of 26% would, just, leave the UK with the lowest headline corporation tax rate in the G7. But we would move down the competitiveness ranking relative to some other EU countries. A higher rate would also reduce the incentive for both domestic and multinational companies to invest in the UK, which might cause concern as we prepare to leave the European Union. It is also likely that the initial revenue raised would reduce over time as companies invest less and change behaviour in other ways.”

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to




]]> Tue, 16 May 2017 00:00:00 +0000
<![CDATA[Labour’s Higher Education proposals will cost £8bn per year, although increase the deficit by more. Graduates who earn most in future would benefit most]]> Yesterday’s leaked Labour manifesto included a commitment to scrap tuition fees. This follows their previously announced plans to bring back maintenance grants for the poorest students. Both would represent a major reversal of the last 20 years of Higher Education (HE) policy. These proposals increase the upfront government contribution to HE by £1 billion compared to the current system. However, these policies would increase the government deficit by around £12.7 billion, of which £11 billion is from scrapping fees, because loans made to students don’t add to the deficit. The forecast long-run cost is lower, at around £8 billion, because a significant proportion of student loans are not expected to be repaid.

In this observation, we highlight the expected short-run impact on the public finances, the impact on graduates and the wider policy implications. We focus on the impact in England because Higher Education is devolved; however, the Barnett formula would add to the cost of this policy.

Government finances

Table 1 summarises our modelling of the impact of the reforms. Replacing fees with teaching grants would increase the up-front government contribution to HE by £1 billion compared to the current system. This is driven by the additional spending on the fees of students who do not take out student loans and paid some or all of their fees up front. Otherwise, the up-front cash outlay – and hence contribution to government debt – is unchanged.

The big difference, however, is in the impact on the measure of the deficit that we typically focus on. This is entirely because of the way these things are accounted for by the government; teaching grants count towards the deficit in the short run, while tuition fee loans do not. Consequently, scrapping fees adds around £11 billion to the deficit. This is £10 billion for current borrowers and an extra £1 billion for the current self-financers.

The long-run impact on government finances is smaller than this, because some – though not all – of the tuition fee loans would have been repaid. We estimate that the present value (to government) of long-run student repayments is £6.5 billion. This reflects the real long-run cost of removing tuition fees and is therefore a better estimate of the true cost of the policy to the government. This is still a substantial amount.

Adding free maintenance grants to this has a similar short-run effect; up front government outlays are unchanged (assuming student support in the form of “cash in pockets” is held constant), but about £2 billion is added to the deficit. However, without tuition loans, a lot of this maintenance debt would have been repaid. Consequently the long run cost of reintroducing grants will now be much larger; around £1.5 billion. Combined, this results in a long run cost of £8 billion (see End Note 1).  

Money flows under various student finance systems (2017 prices)


£9,250 fees

No maintenance grants

Zero fees

No maintenance grants

Zero fees

With maintenance grants

Cost per borrower




Total up-front government spend




Of which, loans




Long-run graduate contribution




Long-run taxpayer subsidy




Total costs (including non-borrowers)




Total up-front government outlay




Of which, direct grants




Total long-run government contribution




See End Note 2

Student repayments

We now turn to the impact for students. These reforms would dramatically reduce the level of debt students hold on graduation. Average debt amongst borrowers would fall from £50,800 to £22,400 with the removal of fees and again to £16,600 with the reinstatement of maintenance grants for poorer students. The remainder is the maintenance loans that all students can access.

Clearly, average student repayments would decline as a result. However, because student loans are repaid as a proportion of income, this average figure has very different impacts across the earnings distribution (shown in Figure 1). As high-earning graduates repay the largest share of their student loans, they benefit the most from the removal of tuition fees. The repayments from the highest-earning graduates (those earning more than around £100,000 a year on average, over their lifetime) would fall by 67% from £93,000 to £30,000, while the lowest-earning would benefit very little.

Reintroducing maintenance grants also benefits high and middle-earning graduates. Although maintenance grants are paid to students from the lowest income families, these students only pay off the value of the additional maintenance loan if they subsequently have high earnings. The additional impact of reintroducing maintenance grants is to reduce repayments of the middle and high -earning graduates with little impact on lowest 30% of earners.

Expected average lifetime repayments by decile of graduate lifetime earnings for 2017–18 cohort (2017 prices, not discounted)

See End Note 3

A striking figure from the current system is the fact that more than three-quarters of individuals expected to have some debt written off at the end of the 30 year repayment periods. For this majority of individuals, student loans are almost indistinguishable from an additional 9% “graduate tax” on their earnings. With fees removed, this share would decline considerably to around 37%, and further still – to 29% – with the removal of maintenance loans.

These Labour proposals would represent a radical shift in the funding of HE in England, with new cohorts of students facing a drastically different system to their recent predecessors. This has a significant impact on the amount they can expect to pay for HE over their lifetime. Students who entered HE in 2011 are expected to contribute around £23,000 in current prices, whereas students entering HE in 2017–18 will contribute more than twice that (111% more). Under these Labour proposals, future entrants would only pay £17,800 on average.

There might be concerns about the equity of this; students who paid £9,000 fees would be doubly hit by large student debts, and the tighter public finances resulting from the subsequent introduction of free tuition fees. One option would be to compensate these students by clearing or reducing their tuition fee debts. This would be extremely costly, however, as the outstanding stock of loans for these graduates is around £30 billion.

Wider implications

Up to this point we have assumed that university funding is unchanged as lost fee income is replaced by increased grants. However, a potential disadvantage of this reform is the concern that it may eventually lead to university resources being squeezed due to pressure on public finances – this was a key justification for the introduction and subsequent increase of tuition fees. This is particularly pertinent when considering the fact that the cap in student numbers has now been removed. Although we have yet to see increases in student numbers, this would be more likely to occur when there are no tuition fees. Large student number increases would increase government cost, and potentially reduce funding-per-student at universities as it did through the 1980s and 1990s.

On the other hand, one disadvantage of the current system is that the government has limited flexibility to target government funding at high priority subjects or students without increasing cost; as so much of the long-run taxpayer contribution to HE is through loans that are not repaid, this affords little control to government as to where the money goes – it is concentrated on students who do not have the highest future earnings, rather than those who provide the highest social return. Under the Labour party proposals, the vast majority of government funding is through grants which could – for better or worse – be targeted at specific institutions, subjects or students in a cost-neutral way.


End Notes

1. The order with which we remove fees and add grants affects the estimates of their individual contribution to the cost. If fees are removed first, scrapping them costs around £6.5 billion, while reintroducing maintenance grants costs £1.5 billion. If maintenance grants are added first, their reintroduction costs £500 million and scrapping fees costs £7.5 billion. This is driven by debts not being fully repaid by graduates. Note that the £500 million estimate here is larger than the £270 million estimated by IFS in 2015. This is driven by the reduction in the discount rate that means future repayments are now valued more highly.

2. Note: All figures are given in 2017 prices, in net present value terms using the government discount rate of RPI + 0.7%. These figures apply to young full-time English-domiciled students studying at the 90 largest universities in England starting in 2017–18. Cohort of students is held constant across systems. We assume that all students taking out loans do so for the full amount to which they are entitled, that there is no dropout from university, that graduates repay according to the repayment schedule and that they have low unearned income. This assumes cohort size of 365,700 based on 2015–16 Higher Education Statistics Agency (HESA) estimates of English-domiciled first-year full-time undergraduates. We assume 10% non-take-up of loans, approximately in line with Student Loans Company (SLC) data on loan uptake. We assume full take-up of grants. Source: Authors’ calculations using IFS’s graduate repayments model.

3. Note: Figures in 2017 prices, deflated using CPI inflation, not discounted. These figures apply to young full-time English-domiciled students studying at the 90 largest universities in England starting in 2017–18. Cohort of students is held constant across systems. We assume that all students take out the full loans to which they are entitled, that there is no dropout from university, that graduates repay according to the repayment schedule and that they have low unearned income. Source: Authors’ calculations using IFS’s graduate repayments model.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Thu, 11 May 2017 00:00:00 +0000
<![CDATA[Labour’s reversal of corporate tax cuts would raise substantial sums but comes with important trade-offs]]> Today, the Labour party will announce that they would not implement planned corporation tax cuts and would reverse most of the cuts introduced since 2010. This would be the first time the main rate of the modern corporation tax in the UK had been increased. The policy could raise around £19 billion in the near term, but substantially less in the medium to long run because companies would respond by investing less in the UK. Part of the revenue raised would be used to increase spending on education by £9 billion a year (£8.4 billion in 2017–18 terms).

Alongside this observation, we are today publishing a new analysis of what’s happened to corporation tax and receipts since 2010. Rate cuts announced since 2010 are forecast to cost at least £16.5 billion a year in the near term, with tax raising measures bringing the net cost to £12.4 billion a year. Under current plans, corporation tax receipts are forecast to be 2.3% of national income by 2021–22, substantially below the pre-recession high of 3.2%.

The Labour policy

Labour would increase the headline rate of corporation tax from 19% in 2017–18 to 21% in 2018–19, 24% in 2019–20 and 26% in 2020–21, as shown in Figure 1. This would return the rate to its 2011 level. For companies with annual profits below £300,000, they would reintroduce a small profits rate at 20% in 2018–19, rising to 21% in 2020–21. So under Labour, from April 2020 most profits would be subject to corporation tax at a rate of 26%, substantially higher (9 percentage points) than the 17% rate that would apply to most profits under the government’s current plan.

Figure 1: Rates of UK corporation tax under current and Labour’s proposed plans  

Figure 1: Rates of UK corporation tax under current and Labour’s proposed plans

Source: Author’s calculations.

The revenue raised

This substantial rise in the rate of corporation tax would raise a substantial sum. HMRC estimate that a 1 percentage point increase in the rate of corporation tax this year would raise £2.7 billion a year in 2021–22 (£2.3 billion a year in 2017–18 terms). If this figure is used, a simple ‘back of the envelope’ calculation suggests that Labour’s plans (relative to a 17% rate) could raise around £19 billion in 2021–22 (£16.8 billion in 2017–18 terms). That compares to forecast onshore corporation tax revenues of £53 billion in 2021–22. However, this should be seen as an overestimate of how much revenue would be raised because at higher tax rates an increase in the rate would lead to a larger reduction in UK investment and therefore bring in less revenue.

Labour’s increase in the rate of corporation tax, if it did raise revenues by £19 billion, would add 0.8% of national income to government receipts, which are already forecast to rise to the highest share of national income since 1986–87.

The trade-offs

Increasing tax on corporations may appear to be an attractive way to raise revenue. But, as ever, there are trade-offs:

  • Increasing rates will raise less revenue in the medium to long run because firms would respond by investing less in the UK. This in turn would depress economic activity and lead to fewer jobs and lower wages. There is a very high degree of uncertainty about how large these effects are but estimates suggest that they may be substantial. The potential size of these effects is an indication of why the OECD and others judge corporation tax to have a particularly damaging effect on economic growth.
  • Of course, when considering longer run effects it is also important to consider how a government would spend revenue raised, since these decisions (such as higher spending on education) will also have effects on the size of the economy.
  • All taxes are paid by people and corporation tax is no different. Higher rates can reduce the returns to company owners (shareholders), but there is also evidence that a significant share of the burden is passed to workers in the form of lower wages.

Where a 26% corporation tax rate would leave the UK

A rate of 26% would still leave the UK with the lowest rate in the G7. We would, however, move down the competitiveness ranking relative to some other EU countries (Figure 2). Under both the current Conservative plans and the Labour plans, the UK would continue to have a less competitive tax base than other countries because we allow a smaller share of capital expenditure to be deducted from revenues each year. A higher rate will reduce the incentive for both domestic and multinational companies to invest in the UK.

An unfortunate side effect of the Labour plan is that it will entail reintroducing the so-called Small Profits Rate – a preferential rate for companies with low profits. The oft-cited justification for having a separate and lower small profits rate is to encourage new business formation and, in particular, entrepreneurship. However, there is a lack of compelling evidence that levying a lower rate of corporate tax on the basis of companies’ profits achieves this aim. In addition, the redistributional motivation for a progressive personal income tax system does not apply to firms. Reintroducing the small profits rate would reintroduce unnecessary and unwelcome complexity into the corporation tax system. 

Figure 2: International comparison under Labour’s plan

Figure 2: International comparison under Labour’s plan

Note: Measure refers to the 2017 combined corporation income tax rate, including local taxes where relevant.

Source: OECD Tax Database,

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Wed, 10 May 2017 00:00:00 +0000
<![CDATA[Labour’s proposed boost to education spending]]> Today, Labour will announce a series of education spending commitments as part of an overall plan for a National Education Service. Exactly how a National Education Service would differ from the current institutional arrangements is not wholly clear. What is clear is that Labour plan to spend significant amounts of additional money on education as compared with current government plans, about £8.4 billion a year in today’s prices by 2021–22 (or £9 billion in cash-terms). This observation details the education spending commitments and compares them to existing government plans. An accompanying observation looks at the rise in corporation tax that Labour would use, in part, to pay for these commitments. Unless otherwise stated, all figures are quoted in 2017–18 prices and will thus be slightly lower than those quoted elsewhere.


The current government is part way through making the first real-terms cut in school spending per pupil in England since the mid-1990s, which will total about 6.5% between 2015–16 and 2019–20, or about 8% if we include some of the additional costs schools have faced in recent years such as higher pension contributions. If delivered this would leave spending per pupil in 2019–20 at about the same level it was a decade earlier, though still substantially higher than before the rapid increases over the course of the 2000s. Alongside this the current government plans to implement a national funding formula for schools in England. Such a reform was always going to produce relative winners and losers. However, with the current funding situation so tight, it means that schools facing the biggest losses from a national funding formula (a 3% cash-terms cut in funding per pupil between 2017–18 and 2019–20) would likely receive a real-terms cut of 10% between 2015–16 and 2019-20 after accounting for average cost pressures.

Labour have proposed to reverse all real-terms cuts to date and then to maintain school spending per pupil in real-terms at this new higher level. This would amount to an extra £4.5 billion a year in today’s prices (about £4.8 billion in cash-terms), or a more than 10% real-terms increase in the schools budget in England, as compared with its current level of about £41 billion. This is higher than previous estimates by IFS researchers as it relates to a more comprehensive measure of school spending (including the core schools budget, the pupil premium and all types of state-funded schools). In addition, Labour would ensure no school loses out from a national funding formula in real-terms, at a cost of £325 million in 2019–20 in today’s prices.   

Meeting these commitments would represent a significant increase in school spending. It would also continue a pattern of schools being relatively insulated from public spending cuts. Schools were one of the few areas of public services (along with the NHS and aid spending) to be protected from cuts to day-to-day spending under the coalition between 2010 and 2015. Protecting losers from the national funding formula comes at a relatively low cost, but only implementing the National Funding Formula for winners is the easy part of school funding reform. At some point, policymakers should find a way to get the losers on to the formula too. 

This comes in addition to Labour’s earlier announcement to extend free school meals to all pupils in primary schools. They estimate this will cost about £700-900 million a year and is the subject of separate IFS observation looking at the potential costs and benefits of universal free school meals.

Further and Higher Education

Labour have also proposed to bring back the Education Maintenance Allowance (at an estimated cost of £540 million a year in today’s prices) and restore maintenance grants for higher education (at an estimated cost of £1.7 billion a year in today’s prices).

Bringing back the Education Maintenance Allowance (EMA) would reverse a 2011 reform that replaced it with the 16-19 Bursary in England (the EMA was preserved in the rest of the UK). The EMA was a conditional cash transfer that paid 16-19 year olds from lower income households £30 per week to stay in school. The 16-19 Bursary was similar in spirit, but had a significantly lower annual budget of £180 million compared to the £560 million a year that was spent on the English part of the EMA. This suggests bringing back the EMA would cost around £380 million a year; Labour’s £540 million figure suggests they might not abolish the 16-19 Bursary. Previous IFS work found that the removal of the EMA had a small, negative impact on participation and Level 2 attainment, and that short run savings were outweighed by the estimated long run costs, through reduced earnings and higher welfare payments. 

Re-issuing maintenance grants for the poorest students would reverse a change that came into force in September 2016. Labour estimates this would cost £1.7 billion a year by the end of the parliament in today’s prices. In the long-run, however, it is likely to cost less than this as a good chunk of the maintenance loans they would replace would not be repaid.

Adult Education

Finally, Labour have promised to abolish upfront fees for adult learners in further education colleges. They cost this at about £1.4 billion a year in today’s prices by the end of the parliament. The actual cost will depend on the precise details of the policy which are not yet fully clear. However, there are two things we do know. First, providing something for free generally leads to high demand. Second, empirical evidence ( suggests that most existing adult skill qualification offer poor economic returns in the labour market. Therefore, without clear rationing and a framework that directs people to high-quality qualifications, this policy could end up being poor value to the taxpayer.


Labour have promised significant increases in education spending. If the additional £8.4 billion, of which £4.8 billion is for schools, is spent well then it will make a positive difference. And the latest economic evidence from the US suggests boosts to school spending can improve pupil attainment and their earnings. There is, however, much we still don’t know about Labour’s plans. What are their plans for early years, tuition fees, or spending in sixth forms and further education colleges? We will hopefully find out more over the next few weeks as Labour publishes its manifesto.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Wed, 10 May 2017 00:00:00 +0000
<![CDATA[Free school meals for all primary pupils: Projections from a pilot]]> Hunger in the classroom is linked to lower attainment, poor behaviour, and worse health outcomes. In the last few years, UK policymakers have tried to reduce the number of children with poor nutrition by expanding school food initiatives, including providing free school meals to all children in Reception through Year 2 in English schools. The Labour Party has promised to extend this free meal entitlement to children in Years 3 to 6 (ages 7-11) in England in order to ‘benefit the educational attainment and health of all children’.

Universal free school meals can improve attainment in some circumstances. A 2012 pilot study by IFS researchers and others found that Year 6 students in Newham and Durham, where all primary children were offered free school lunches, made around two months’ additional progress over a two-year period compared to similar children in other areas. In this observation we argue that extending the policy nationwide would come at a significant cost and might not lead to similar gains. Other policies, such as offering free breakfast clubs (as is the case in Wales and as trialled in England) might be a cheaper and more effective way to improve both education and health outcomes.

Counting the costs

Labour is proposing to take a benefit that is currently means-tested – available only to pupils whose families meet certain criteria of disadvantage – and extend it to all students. Universal programmes can bring benefits, such as removing the stigma that might otherwise prevent eligible students from taking up free meals. However, universal benefits are also costly: rather than targeting funds at the most disadvantaged, they spread the money across all children, including those whose families are currently paying for school meals.

This ‘deadweight loss’ can be substantial. The pilot study estimated that around 40% of the total cost was spent on providing free meals to students whose parents would otherwise have paid for a school lunch. Since these students are consuming the same meal that they would have otherwise eaten, there aren’t likely to be very many benefits for their health or attainment from having government pick up the tab. On the other hand, this represents a significant giveaway to their families. For example, the policy would cost £11.50 per week for a family with one child newly eligible for free meals, around a sixth of the £70 that similar families spend on food each week.

Currently, the government pays £2.30 for each meal taken under the universal infant free school meals programme introduced in 2014. Taking into account the number of pupils in Years 3-6 who are not already eligible for free school meals and the pilot’s estimated take-up rate of 90%, extending free meals to all primary children would cost the government around £800 million per year.

In addition, there are likely to be other upfront costs from one-time investments such as renovating school kitchens and cafeterias to provide additional meals. The government provided £170 million over two years to meet these costs during the recent infant free school meals roll-out. Since Labour’s proposal would extend free meals to four new year groups rather than three, the additional one-time funding needed to upgrade facilities could be as much as £225 million.

Although the extension of free school meals would only apply to primary school children in England, there are public finance implications for the devolved nations as well: because Labour is proposing additional spending on this policy (paid for by levying VAT on private school fees), the block grants to the three devolved nations will increase under the Barnett formula. The new spending in England would result in £150 million in additional funding for the three devolved nations each year, plus one-time additional funding of around £45 million related to the upfront costs. This brings the total cost of extending free school meals to all primary pupils to around £950 million each year, with upfront costs of as much as £270 million.

Broadening the benefits?

The pilot found that Year 6 students in areas of universal provision made an additional two months’ progress over the course of two years relative to similar students in other areas. These are significant effects, roughly the same size as the benefits from national programmes such as the “literacy hour”. However, it’s far from certain that universal free school meals would be as effective if rolled out nationally.

One reason for caution is the difference between the pilot areas and the average English local authority. Both Newham and Durham are relatively disadvantaged. If pupils in better-off areas are more likely to pay for school meals or to have healthier packed lunches, the gains from making school lunches free to all students are likely to be smaller.

Further, while the pilot study found that universal free school meals improved test scores, it wasn’t able to pinpoint how these improvements came about. Evidence from other countries suggests that more nutritious meals can reduce hunger and keep students healthier, which can in turn improve behaviour in the classroom or reduce the number of illness-related absences. However, the pilot study found little support for any of these mechanisms: offering universal free lunches did not improve parents’ impressions of their children’s focus and behaviour; absence rates in the pilot areas remained about the same as in the similar comparison areas; and although children in the pilot areas had healthier foods at lunch, they still ate roughly the same amount of junk food over the course of the day and were no more likely to be at a healthy weight. Without understanding what’s driving the headline academic gains, it’s difficult to know whether the relationship between free school meals and test scores would be the same in other areas with a different local context.

Rather than providing free school lunches for all children, policymakers wanting to tackle student hunger could support school breakfasts instead. IFS research has found that support for a one-year breakfast programme in disadvantaged schools delivered similar academic benefits to universal free school meal provision (though the gains were higher in Year 2 than Year 6). The breakfast clubs also significantly improved behaviour and concentration, and reduced absences – and did so at around one-tenth of the cost per pupil of universal free school meals.

Extending free school meals to all primary school children would cost around £950 million each year. It would not directly benefit the poorest children, who are already entitled to free lunches. While there is some evidence it might raise attainment overall, we don't understand how or why, and so the effect of extending this nationwide is uncertain. In the context of constrained public resources it is important to be much clearer about effectiveness before spending a large amount of money on a new universal entitlement.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Tue, 09 May 2017 00:00:00 +0000
<![CDATA[If politicians talk about the rich, always ask who they mean]]>

This observation has been revised to correct an error in the paragraph preceding Figure 2, shown by the crossed-out text, and the labels shown in Figures 1 & 2. 

Most of us would like somebody else to pay for the public goods and services we consume. Politicians tend to give the impression that this is their intention too – often through somewhat vague references to "the rich", “the very best-off” or “those with the broadest shoulders”. However, it is rarely clear from such statements exactly who is being referred to. In this observation we provide some figures on how wealth and income are distributed in the UK to help us assess who these people might be.

The distinction between wealth and income is an important one, often glossed over. High-income people are those who receive a large flow of money over a period of time (from things like salaries, bonuses, dividends or profits for the self-employed) whereas high-wealth people are those who have accumulated a valuable stock of assets by a point in time (typically in property, pension pots or other financial assets). While there are many people with both high wealth and high incomes, the overlap is far from perfect. There are wealthy individuals with low current incomes, such as older people who own a valuable house but have little income; and there are high-income people who have accumulated little wealth, such as young renters with high-paying jobs. These groups would be among those who might want somewhat more precision from politicians who claim to have plans to increase taxes on “the richest”.

It is also often unclear from politicians’ statements how much wealth or income one should have to qualify as being rich. What might they have in mind?

Take wealth to begin with. While official statistics on the very wealthiest in society are, unfortunately, limited, unofficial data highlight that there are some extremely wealthy people living in the UK. For example, each year the Sunday Times publish a ‘Rich List’ of 1,000 individuals and families, each with estimated wealth of more than £100 million. These are clearly extremely wealthy people. However, they only represent less than 0.00002% 0.002% of the UK adult population. Most people’s idea of “the richest” is probably broader than that.

The Office for National Statistics carries out a household survey that gives a broader view of the distribution of wealth, although it does not do a good job at covering the very wealthiest, who tend not to respond to household surveys. Figure 1 plots the distribution of reported total net wealth from this survey (including the value of houses, private pension pots and financial assets, net of mortgages and other debts). It shows that being in the top 1% (around 260,000) of households in Great Britain required more than £2.9 million in total net wealth. This may well be an underestimate for the reason given above, but it is still almost 13 times as much as the median household, which sits in the middle of the wealth distribution. To be in the top 10%, a household needs to possess a more modest £1 million in net wealth: still more than 4 times the median, but only a third of the household wealth at the 99th percentile.

Figure 1. Distribution of total net wealth among households in Great Britain, 2012-14

Source: ONS Wealth and Assets Survey.


Much – but by no means all – of the variation in wealth between households is a result of their being at different stages of their lives. Because people typically save during their working life and then run down their assets in retirement, such ‘lifecycle effects’ would result in substantial inequality in wealth even if everyone had the same income and spending trajectories over their lives. It is therefore not surprising that ONS figures show that 25% of individuals aged 55-64 live in a household with more than £1 million of wealth, placing them in the top tenth of households, compared to just 4% of those aged 25-34. However, there is also significant inequality in wealth between individuals of the same age (e.g. another 23% of those aged 55-64 live in a household with wealth of less than £200,000), and there is good evidence that the greater wealth of older individuals today partly represents genuine inequality in lifetime resources across generations.

What about income? Who are the highest-income people? There are different ways of answering this question depending on factors such as whether we mean the personal income of individuals (which is what the income tax system deals with) or the total income of their household, whether we measure incomes before or after tax, and which sources of income are included. Below we briefly highlight two key official data sources on different income measures.

Data collected by HMRC for income tax purposes tell us about the distribution of individual (taxable) incomes among income taxpayers. In all that follows, then, bear in mind that about 40% of adults do not pay income tax at all (i.e. their taxable incomes are too low to be included in these statistics). In 2014-15 almost 50,000 income tax payers (0.1% of UK adults) had more than £500,000 in annual pre-tax income, amounting to £55.5 billion between them. 39% of this – £21.8 billion – accrued to just 5,000 individuals; an average of £4.4 million per person among that group. This highlights the staggering degree of income inequality found at the top of the distribution.

Given these extraordinarily high incomes, it is perhaps less surprising that those with merely very high incomes do not always feel especially well off, even though they clearly are relative to the vast majority of the population. Figure 2 plots the distribution of income among individual income tax payers in 2014-15. The person at the 95th percentile of taxpayers (i.e. with an income greater than 95% of income tax payers) had pre-tax income of about £72,000 per year. There can be little question that they have a relatively high income: £50,000 per year more than that of the median income tax payer. However, it is also £90,000 per year less than someone at the 99th percentile. They might therefore be forgiven for feeling more similar to the median taxpayer, who is 23.6 13.7 million people (45 percentiles) below them in the distribution, than to the taxpayer just 2.1 1.2 million people (4 percentiles) above them.

Figure 2. Distribution of income among individual income tax payers in the UK, 2014-15

Source: HMRC Table 3.1a

Of course these figures capture income only over one year. Because income tends to vary systematically over the lifecycle (though to a lesser extent than wealth), and can be volatile from year to year, a much higher proportion of people will have relatively high incomes at some point during their lives than in any single year.  

The income measure available from HMRC data is useful but limited. It does not account for the fact that people live in households of varying sizes with varying needs; or that some individuals will be the only source of income in their household, while others will be one of two or more; or that some households will have non-taxable income like certain state benefits. To get a better sense of who the best-off are in terms of living standards, we can use an income measure produced by the Department for Work and Pensions (DWP) based on an annual survey of more than 20,000 households in the UK. This measures each individual’s total household income from all sources after direct tax, and rescales (‘equivalises’) it to take into account the fact that households of different sizes and compositions have different needs (for households containing very high-income people it also applies an adjustment based on tax data, to try to deal with the problem of these people not responding well to surveys).

Figure 3 shows the household income after taxes and benefits needed to enter the top 10% and 1% on this basis for individuals in three different types of household in 2015-16. A single adult without children can be considered in the top tenth of the income distribution if they have a total annual net income of at least £33,000, whereas a couple with two children aged 14 or younger would need more than twice this (£68,800) to be in the top 10%. Similarly, to be in the top 1% on this basis, a single adult would need annual income of more than £86,000, compared to £179,800 for a couple with two children, and £128,400 for a couple without children.

Figure 3. Annual net income required to be in top 10%/1% of equivalised household income


Note: Incomes are measured after taxes and benefits but before housing costs have been deducted. Children are all assumed to be under the age of 14. Figures rounded to nearest £100.
Source: Authors’ calculations for the UK using the Family Resources Survey, 2015-16.

The acceptable amount of income or wealth inequality, and the appropriate degree of redistribution carried out by the state, is ultimately a matter of political choice for governments and electorates. But such choices should be underpinned by an informed understanding of how income and wealth is actually distributed. So the next time you hear a politician or journalist talk about ‘the rich’, ask who they mean.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Fri, 05 May 2017 00:00:00 +0000
<![CDATA[A ‘double lock’ on the state pension would still be a bad idea]]> On the basis of Theresa May’s answers at yesterday’s Prime Minister’s Questions, there has been significant speculation about whether the Conservative manifesto will pledge to maintain the ‘triple lock’ on the basic state pension (and the new single-tier pension) – guaranteeing that the annual rise will be the highest of inflation, average earnings growth and 2.5%. In this observation we discuss the problems with the triple lock, along with alternatives.

The most obvious problem with the triple lock is that 2.5% figure is totally arbitrary – it is entirely unclear why in a year where inflation and earnings growth are less than 2.5% the state pension should rise faster than both. But the biggest problem with the triple lock is the ‘ratchet effect’. Because the state pension grows in line with the highest of earnings, prices and 2.5% in the long run it will increase faster than all of them. This is the reason why the triple lock has such a significant effect on state pension spending in the Office for Budget Responsibility’s long-run projections. Relative to increasing in line with earnings each year the OBR estimates that the triple lock could increase pension spending by 0.8% of GDP in 2060–61, which is equivalent to £15 billion in today’s terms. The key thing to note is that getting rid of the 2.5% guarantee and moving to a so-called ‘double lock’ would not eliminate this issue. The state pension would still rise faster than both earnings and prices in the long run, and would still eventually become unaffordable.

The solution to the ‘ratchet effect’ is to go with an idea we discussed in our 2015 election analysis (page 15 here), which has since been recommended by the Work and Pensions Select Committee who describe it as a ‘smoothed earnings link’:

“The state pension would be uprated with earnings, but with temporary price-indexation when inflation exceeded wage growth. Price indexation would continue once earnings growth again exceeded inflation, but only for as long as the value of the state pension remained above [an] original fixed minimum proportion of average earnings. Indexation would then revert to earnings.”

As the Committee noted, this policy would ensure that the state pension rose in line with earnings over the long term (rather than rising faster than earnings), but also that pensioners were protected from real cuts to their income in occasional periods of falling real earnings.

Of course, you might have a different set of objectives when choosing how to uprate the state pension, and different objectives could lead to different choices. But whatever your objectives, neither the triple lock nor the double lock is the right answer.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the General Election is informed by independent and rigorous evidence. For more information go to

]]> Thu, 27 Apr 2017 00:00:00 +0000
<![CDATA[Election battleground in the playground?]]> School funding in England has shot up the political agenda in recent months and could well be a major battleground at the upcoming general election. In this observation, we set out what current government plans imply for spending on schools in England, what each of the main parties has said to date and the potential implications of freezing school spending in real-terms after the election. 

In 2017–18, day-to-day school spending in England is expected to represent about £38.5 billion, which translates into about £4,800 per pupil in primary schools and £6,200 per pupil in secondary schools (excluding 6th forms). The outgoing government committed to freezing school spending per pupil in cash-terms up to 2019–20. This implies a real-terms cut in spending per pupil  of about 6.5% between 2015-16 and 2019–20, which represents the first real-terms cut to school spending per pupil since the mid-1990s and the largest fall over a four year period since at least the late 1970s. If implemented, this would leave school spending per pupil at about the same level in 2020 as it was in 2010, undoing the increases which were actually afforded during the 2010-2015 parliament. As shown by the graph below, this would still leave spending per pupil much higher than that seen before the rapid growth over the 2000s.

Spending per pupil in primary and secondary schools: actual and plans

Notes and Sources: See Crawford, Belfield and Sibieta (2017).

As well as changing the size of the school spending cake, the current government has proposed a different way of dividing this up amongst schools. In particular, a recent consultation proposed introducing a national funding formula for schools in England (, which would replace the 152 different funding formulae used by local authorities in England (for more details, see our recent briefing note The current system of funding schools is out-of-date and in need of reform. We haven’t had a proper funding formula since the early 2000s, which has allowed various inequities across areas to develop, which will only grow if left unaddressed. It is to their credit that the outgoing government came forward with detailed proposals for a funding formula, and to their discredit that previous governments kicked this particular can down the road.

Such a reform would always produce relative winners and losers. With overall funding so tight it will also result in absolute losers. Transitional arrangements mean that no school will see a cash-terms fall in funding per pupil of more than 3% between 2017–18 and 2019–20, and no school can receive a gain of more than 5.6%. However, combined with real-terms cuts to overall school spending, this means that schools facing the biggest losses will see a real-terms fall in funding per pupil of close to 10% between 2015-16 and 2019-20. Even the biggest winners will get only a real-terms freeze over the same time frame.

We don’t yet know what each of the main parties will propose at the upcoming general election. A re-elected Conservative government could choose to implement existing plans up to 2019–20, or could decide to top them up. The government seems committed to the principle of a national funding formula, but is yet to publish updated proposals following the consultation.  

How things would change under a Labour government is not yet clear. What we do know is that Labour seems to support the principle of a national funding formula, with the Shadow Education Secretary Angela Rayner recently saying “Now the principle behind a fair funding formula, I absolutely, whole-heartedly agree with it.” ( Labour also opposes real-terms cuts to school spending per pupil. For instance, the Shadow Chancellor John McDonnell, recently told the NUT conference that “Our priorities are to oppose the cut in funding per pupil...”  (

What would the implications be of freezing school spending per pupil in real-terms?

As the table below shows, this would equate to additional spending on schools of around £1 billion in 2019-20 as compared with existing government plans, or £2 billion of extra spending by 2021-22 compared with 2017–18.  This extra money would make it a bit easier to implement school funding reform as the baseline would be a real-terms freeze rather than a real-terms cut. But note that it would inevitably still imply real-terms cuts for relative losers from a national funding formula if this policy was pursued. Both implementing a national formula and ensuring no real terms cut for any school would require more money than that.

Of course any additional money spent on schools is money that could not be spent elsewhere.

Day-to-day school spending in England (£billion, 2017-18 prices) under different scenarios


Existing plans

Real-terms freeze in spending per pupil
















Notes and Sources: Existing plans based on baseline funding for schools, central services and high needs funding in national funding formula consultation and an assumed freeze in spending per pupil up to 2017-18 ( Existing plans beyond 2017-18 includes additional funding for transitional arrangements over and above cash-terms freeze in spending per pupil. Real-terms freeze scenario calculated based on DfE projections for pupil numbers up to January 2022 ( and OBR projections for the GDP deflator (

Over the course of the campaign we will likely find out more about commitments made by the major political parties on school funding and other aspects of schools policy. We will therefore update this analysis as more information becomes available.

IFS Election 2017 analysis is being produced with funding from the Nuffield Foundation as part of its work to ensure public debate in the run-up to the general election is informed by independent and rigorous evidence. For more information, go to

]]> Wed, 26 Apr 2017 00:00:00 +0000
<![CDATA[Scotland’s income tax schedule to differ from rest of the UK for first time]]> From tomorrow, Scottish residents will for the first time be subject to a different income tax schedule from those resident elsewhere in the UK. This is because of the Scottish parliament’s decision to use recently devolved powers over income tax bands and rates for non-savings and non-dividend income to freeze the higher-rate threshold (the point at which the rate of income tax rises from 20% to 40%) for the new financial year.

By default, the higher-rate threshold would rise from its current level of £43,000 to £43,900 in 2017–18 both in and outside Scotland (given standard inflation uprating and UK-wide changes to the personal allowance). Instead, the UK parliament has decided to increase this to £45,000, reducing taxes for 4.4 million people at a cost of around £600 million per year, while the Scottish parliament has decided to hold this fixed at £43,000, increasing taxes for around 350,000 of Scotland’s 2.5 million income taxpayers, raising around £60 million per year. While initially relatively small – the maximum tax rise in 2017-18 is £180 per year –, this difference looks set to at least double over the coming years; a result of the Conservative Party manifesto promise to raise the higher-rate threshold to £50,000 by the end of the parliament, and the Scottish government’s stated intention to increase it by at most inflation over the same period (implying a Scottish higher-rate threshold of at most £46,000 by 2020–21 given current forecasts).

The Scottish parliament now has significant power to set the rates and thresholds of income tax that apply to the earnings of Scottish residents, but it has no such control over the taxation of savings and dividend income or National Insurance contributions (NICs). One consequence of this is that the earnings of Scottish residents will be subject to a peculiar tax schedule, shown in the Figure below, with a combined rate of income tax and employee NICs that rises from 32% to 52% at £43,000 per year, before falling back to 42% at £45,000.

Combined income tax and employee NICs schedule, 2017–18

Note: assumes standard personal allowance and all income from earnings which are constant through the year.

This is because employee NICs are levied at a rate of 12% up to a threshold known as the Upper Earnings Limit (UEL), and 2% above. The UK government has decided to keep this aligned with the income tax higher-rate threshold that applies outside of Scotland: as the Figure shows this means that employees with straightforward tax affairs outside Scotland will pay income tax and employee NICs at a combined rate of either 12, 32, or 42%, compared to the more complicated schedule facing those in Scotland.

Scotland’s newly craggy rate schedule should serve to remind that retaining separate systems of income tax and NICs creates substantial complexity in the overall system of earnings taxation. In fact, these parallel systems can already create incoherent rate schedules of the kind depicted in the Figure above for those whose earnings fluctuate through the financial year, simply because income tax is assessed against annual income whereas NICs is assessed within each pay period. Complexity of this kind looks especially needless given that NICs are now essentially just another income tax levied on earnings, with virtually no link remaining between the NICs someone pays and the benefit entitlement they accrue. There is therefore a strong case for merging the two taxes, as set out in the IFS-led Mirrlees Review.

In the absence of such moves towards integration on a UK-wide basis, devolving NICs would allow the Scottish parliament to limit the additional cragginess in the tax schedule that arises from not following Conservative Party plans to raise the UEL and higher-rate threshold in the rest of the UK. As previously argued by IFS researchers, the ability to set NICs rates and thresholds would be a logical counterpart to those powers already granted over non-savings and non-dividend income tax, if devolution of further tax revenues to Scotland were seen as desirable.

Decisions by the Scottish parliament to set income tax rates or thresholds that differ from those elsewhere in the UK will also have implications for the incentive to carry out similar work through different legal forms. This is because Scottish company owner-managers can take income in the form of dividends which are taxed less heavily than the incomes of either the self-employed or employees, at rates and bands determined by the UK parliament. By setting the higher-rate threshold for earned income below that set in Westminster for dividend income, the Scottish parliament exacerbates this existing tax incentive to carry out economic activity in Scotland through corporate form.

As well as distorting the decisions individuals make about how to work, deviating from the UK government’s income tax policy can have implications for the Scottish budget beyond the revenue directly raised or foregone. For example, under the current fiscal framework, were incorporation to rise faster in Scotland than elsewhere in the UK – as it might in response to a larger tax differential – Scottish revenues would fall without any offsetting block grant adjustment.

These issues highlight just some of the complexities that will arise as the Scottish parliament makes use of its substantial new powers of taxation. But they are also a useful illustration of some of the more fundamental deficiencies that continue to afflict the tax system throughout the UK, and remain in need of reform.

]]> Wed, 05 Apr 2017 00:00:00 +0000
<![CDATA[Significant cuts to two parts of the benefit system to be phased in from next week]]> Next week will see the introduction of significant cuts to the generosity of two parts of the working-age benefits system. From Monday (3rd April), new recipients of employment and support allowance deemed healthy enough to carry out 'work related activities' will get up to £1,500 less each year than existing recipients. From Thursday (6th April) families with new-born third and subsequent children will no longer receive any additional support in recognition of that child through tax credits and universal credit, and support for new born first children will be around £500 a year lower. Because of the restriction to new claims or new births, existing claimants will not see their benefit income fall. But in combination, the changes are expected to reduce government spending by over £5 billion a year in the long run. Below, we consider each reform in turn.

Cut to employment and support allowance

Employment and support allowance (ESA) is the main out of work benefit for working age individuals who are judged not to be ‘fit for work’ due to a health condition. There are currently around 2.2 million individuals claiming ESA, of whom 250,000 are waiting for a health assessment, 1.6 million are in the ‘support group’, and 400,000 are the ‘work-related activity group’ (WRAG). The latter group are those deemed healthy enough to carry out ‘work related activities’, such as CV preparation or skills training.

From next Monday (3rd April) new WRAG claimants will receive £73.10 a week – the same as jobseekers’ allowance (JSA) claimants – rather than £102.15 a week as is currently the case (those in the support group are unaffected). This change will not create immediate losses of benefit income, because only new recipients are affected. Ultimately though, of course, all claims will be assessed under the new less generous rules. To give a sense of how quickly this will cut the generosity of benefits in practice, in the recent past around 60,000 people a year have started an ESA claim and ended up in WRAG - so we would expect approximately that number to get less money over the coming year than they would otherwise have got. In the long run this is expected to save the government about £650 million per year, with around 500,000 recipients getting £1,400 a year less than they would otherwise have got, on average.

What do we know about the sorts of people who are placed in the ESA WRAG? First, around half of them are entitled to ESA because of mental or behavioural disorders. Second, they tend to be somewhat older than JSA claimants, with about half being between 50 and the state pension age compared to about a quarter for JSA. Third, they tend to be on ESA for a relatively long time, as shown in Figure 1. Hence, while this change will align the weekly entitlements of ESA WRAG and JSA claimants, it is worth bearing in mind that the ESA claimants will tend to live on these amounts for substantially longer – around four in five WRAG claimants have been claiming for over two years, compared to less than one in five for JSA.

Figure 1. Proportion of claimants by length of claim, various benefits


Note: Before being placed in WRAG or support group or being declared ‘fit to work’, claimants must go through an assessment, during which they are entitled only to the basic JSA rate. The above data include time spent in the assessment phase, which typically takes at least 13 weeks. This is part of the reason why the ‘up to 6 months’ bars are small for the WRAG and support group. However, since this policy only affects claimants post-assessment, the left-hand stacked bar does give an accurate picture of claim durations for the group affected by the policy change. Source: DWP Tabulation Tool: Employment and Support Allowance, May 2016, Office for National Statistics, UK Labour Market: March 2017, Table BEN02

People might respond to this change in several ways. First, they may not choose to claim ESA in the first place: since JSA will afford the same financial support as WRAG, the financial incentives to go through the medical assessment rather than accept the additional work conditions of JSA are reduced. Second, those placed in the WRAG might challenge the decision to try to get into the support group and receive the now much higher entitlement. At the moment around 20% of those placed in WRAG challenge the decision at least once, so there is considerable scope for this to become more prevalent. Third, as the government’s policy costing document points out, they may try to claim other benefits. The main option available here is personal independence payment, a non-means tested disability benefit. Not only does this provide income directly (between £22 and £141.10 per week), but receipt can also be an automatic passport to higher ESA entitlements. Fourth, they could move into work. Claimants may be constrained in the extent to which they can respond in this way – WRAG claimants have after all been declared by the government to have ‘limited capability for work’. On the other hand, a DWP survey found that 30% of WRAG claimants are already looking for work, and some research suggests that employment decisions among the disabled can be sensitive to the level of disability payments. However, many – perhaps the majority – will not respond in any of these ways and will therefore have to make do with an average of £1,400 a year less than they would otherwise have got.

Cut to tax credits and universal credit for families with children

The generosity of tax credits (and universal credit, which is replacing tax credits and three other working-age means-tested benefits) currently depends on the number of children in a family. A first child increases maximum entitlement by £3,325 per year, and each subsequent child increases maximum entitlement by a further £2,780 per year. This will not, however, be the case for families with children born on or after the 6th April. First children born on or after that date will no longer receive the extra £545 of entitlement (known as the ‘family-element’) while third and subsequent children born on or after that date will no longer receive any additional tax credits or universal credit. The scope of the policy is then planned to widen in November 2018 – the new rules on third and subsequent children will also apply to anyone who makes a new claim to universal credit from then on, regardless of when their children were born. It is worth noting that these families will continue to receive child benefit alongside tax credits or universal credit, at around £1,100 per year for the first child and £700 per year for each subsequent child.

Since the changes only affect families with new births and then (towards the end of next year) new claims, no one will see their benefit income fall between one period and the next as a result of this policy. But these are substantial changes to the long run generosity of the system, expected to reduce government spending by a significant £5 billion a year in the long run. About £3 billion of that £5 billion is expected to come through limiting support to two children, which means substantial cuts in support for larger families. Eventually the reform will mean about 600,000 three child families getting around £2,500 a year on average less than they would have got, with a further 300,000 families with four or more children getting £7,000 a year less on average. The remaining £2 billion annual long-run saving comes from removing the £545 'family element', which will ultimately affect around 4 million families (on top of any losses from the limiting of support to two children).

What kinds of families will lose from these changes in the long run? About two thirds of those families affected will have at least one adult in paid work. Looking specifically at the 900,000 families with three or more children currently on tax credits – the group who, if treated under the new system, would stand to lose particularly large amounts - around half (52%) are couples with at least one adult in work and a further 14% are working lone parents, while 14% are workless couples and 20% are workless lone parents.

Figure 2 puts these cuts to tax credits and universal credit in their historical context, showing how maximum child-related benefit entitlements for families with different numbers of children have evolved (by ‘maximum’, we mean the amount payable if income and assets are low enough for entitlement not to be reduced by the means test). Support in 2020 for new claimants with three or more children (the far right-hand bar) will be much less generous in real terms than when tax credits in their current form were introduced in 2003. However, real-terms entitlement will still be higher for such families than it was when Labour came to office in 1997, such were the increases in generosity in the late 1990s and early 2000s.

Figure 2. Maximum child-related benefit entitlement (inflation-adjusted)

Note: Figures include entitlement to Child Benefit, Child Tax Credit and child-related components of Income Support and Universal Credit. Housing-related benefits are not included. All amounts are calculated for a couple with no private income.

How might people respond to these changes? The cut to the generosity of means-tested benefits for families with children will tend to strengthen the financial work incentives of the parents – who will have less support to potentially lose by increasing their earnings - and this could lead to increased employment rates. On the other hand, the decision to reduce support for new (but not existing) claimants from November 2018 could actually create strong disincentives to work (or to earn more) in some cases. For example, a parent on tax credits offered a one-year fixed term contract after November 2017 would want to take into account that, if they lost tax credit eligibility for that year and then became eligible again after that, they would be treated under the much less generous rules for new claimants.

As well as potentially affecting work decisions, the government has explicitly stated that it hopes to affect the decisions low-income families make about how many children to have. There is some evidence that at least the timing of births in the UK has been affected by past changes to the generosity of benefits, but no conclusive evidence on whether such changes affect the number of children families choose to have.

The government also argues that removing means-tested support for third and subsequent children improves the fairness of the system, by ensuring low-income families face more of the costs of choosing to have additional children. But whether the aim of the policy is to make benefit recipients face more of the financial costs of choosing to have more children, or to affect their actual decisions around how many children to have, it is unclear why the new rules will also apply to new claimants with dependent children of any age from the end of next year. This hits families who made the decision to have a child before the new regime was in place. Of course, widening the scope of the policy in this way will bring some of the long-run savings forward into the end of this parliament.

]]> Thu, 30 Mar 2017 00:00:00 +0000
<![CDATA[Disability benefit spending and the recent change to regulations]]> Personal Independence Payment (PIP) is the main working age benefit paid to disabled people who may face additional costs of living, and has been gradually replacing Disability Living Allowance (DLA) since 2013. It is not means-tested and is paid regardless of whether an individual can, or indeed does, work. On average, recipients in 2016–17 got £99 a week.

In November 2016 two legal judgements (see here and also here) on the eligibility criteria for PIP went against the Department for Work and Pensions (DWP). In response, last month the Government laid regulations to ‘clarify’ the PIP criteria and, in effect, reverse the effects of the legal judgements. These came into force on March 16th. On Monday, the House of Lords will debate one motion to annul the regulations and a second to call for a review.

To understand these judgements, and the changes in the regulations, it is important to understand how PIP entitlements are determined. PIP payments are composed of either or both of a “daily living” component, paid if the individual needs assistance with any of a range of activities such as eating and washing, and a “mobility” component, paid if the individual needs help to travel. Both components may be paid at either a standard or an enhanced rate. When an individual applies for PIP, they are awarded points on the basis of the extent to which they cannot carry out 12 activities which are deemed to be important to everyday life (10 “daily living” activities and 2 “mobility activities”). The total number of points they get for daily living and mobility activities determines whether they get nothing for that component, the standard rate for that component or the enhanced rate.

The two Upper Tribunal judgments in November 2016 made decisions about how many points individuals with certain conditions should receive with respect to certain activities. The much more significant of these two decisions, in terms of the number of individuals affected and the implications for government spending, was in relation to Mobility Activity 1 – “Planning and following a journey”. It was decided that individuals who had to be accompanied on journeys in order that they didn’t suffer overwhelming psychological distress qualified for certain higher categories of points. The DWP equalities analysis estimated that the precedent set down by this judgement would see around 143,000 additional PIP awards (at an average weekly rate of £40) and around 21,000 individuals receive the enhanced rather than the standard PIP mobility rate (a gain of £36 per week). The affected individuals would predominantly be those with a range of psychiatric disorders.

In the absence of any action, DWP has estimated that spending following the tribunal judgements would be £910 million per year higher by 2021–22 and £3.7 billion higher over the period 2017–18 to 2021–22, than otherwise. These estimates do not take account of the fact that any of those brought onto PIP who were also receiving Employment and Support Allowance would also qualify for an additional premium in that benefit.

The Government has said that the tribunal judgments interpreted the PIP regulations in a way which it did not originally intend and on 23 February, laid regulations to amend the PIP regulations. The Office for Budget Responsibility forecasts that the new regulations will limit the fiscal impact of the tribunal judgments to £110 million in 2017–18 with no cost in any subsequent years.

A movement in the forecast level of annual spending on PIP of the magnitude of just under £1 billion is significant. But the bigger picture is that spending on disability benefits has consistently exceeded forecasts in recent years by billions of pounds. The figure shows real-terms spending on disability benefits, had spending grown at the rates forecast in recent Budgets. It also shows what the Budget 2017 forecast would have been in the absence of the Government’s new PIP regulations.

Note: Disability benefits spending includes Personal Independence Payment, Disability Living Allowance and Attendance Allowance.

Source: Authors’ calculations using DWP Benefit Expenditure and Caseload Tables and Equality Analysis: PIP assessment criteria.

When it was introduced, PIP was expected to lead to a 20% fall in both caseload and spending by 2015–16 compared to DLA. But this fall in caseload has not materialised. While Budget 2015 forecast that the caseload for disability benefits would fall by 6% between 2014-15 and 2017–18, it is now expected to rise by 3% over this period. This, along with a higher number of individuals receiving the enhanced PIP rates than originally expected, means that spending on disability benefits continues to grow strongly in real terms, contrary to previous forecasts. Total growth in real-terms spending on disability benefits over the 10 years to 2017–18 is now expected to be £5.9 billion, which is almost twice the £3.2 billion growth forecast as recently as March 2015.

Judgements about who should qualify for disability benefits and how much they should receive will depend on a large number of factors other than their fiscal cost. But it is worth keeping sight of the fact that we are spending more on disability benefits than we used to spend, and more than we expected to be spending.

]]> Fri, 24 Mar 2017 00:00:00 +0000
<![CDATA[Would you rather? Further increases in the state pension age v abandoning the triple lock]]> On Tuesday afternoon MPs will debate the recent report by the Work and Pensions Select Committee on “intergenerational fairness”. Drawing on evidence the IFS, the committee argued that triple lock indexation of the state pension should not continue beyond 2020.

For a given amount of spending on state pensions there is a trade-off between the level of the state pension and the state pension age from which it is paid. With the triple lock a full single-tier pension in 2060 is projected to be worth 27.5% of average earnings and will be available from age 69. If we were to abandon the triple lock and instead index in line with average earnings beyond 2020, then a full single-tier pension in 2060 would be projected to be worth 24.2% of average earnings. For the same cost as the triple lock this could be paid from roughly age 67½. Alternatively if we wanted a higher pension at, say, 30.7% of average earnings then keeping within the same cost envelope could be achieved by increasing the pension age further to roughly age 70½.

The triple lock states that each year the state pension will increase by the highest of the increase in earnings, the increase in prices (as measured by the Consumer Price Index) and 2½%. Over the longer term this would eventually prove to be a financially unsustainable method of indexation. This is because every time earnings grow by less than either 2½% or prices then the value of the state pension would ratchet up as a share of average earnings.

Of course it is not unreasonable to argue that the state pension should be made more generous (though it is should be remembered that – as IFS research has shown – pensioners are no longer a particularly poor group in society). But if the government wants to increase the level of the state pension relative to earnings, it should choose the level it wants (and potentially a path to get there) rather than allowing the somewhat haphazard increases relative to earnings that result from the triple lock. The last few years – in which earnings growth has been extremely weak – have seen triple lock indexation boost the value of the state pension dramatically, relative to both average earnings and prices. Between April 2010 and April 2016 the value of the state pension has been increased by 22.2%, compared to growth in earnings of 7.6% and growth in prices of 12.3% over the same period. This has pushed the value of the basic state pension up to its highest share of average earnings since April 1988. This increased benefit to pensioners came at the cost of an increase in spending of roughly £6 billion a year in 2015–16 compared with earnings indexation, and roughly £4 billion relative to CPI indexation, over the period since April 2011.

Projecting the cost of the triple lock (relative to linking the level of the state pension to earnings growth) over the longer-term is difficult. If the economy delivers strong growth in earnings, without periods of boom and bust, then the lock would seldom apply and the state pension would mostly grow in line with earnings. But if the economy delivers weak and/or volatile earnings growth then the value of the state pension would, over time, ratchet up relative to average earnings. In its January 2017 Financial Sustainability Report the OBR projected state pension spending over the next fifty years under both the triple lock and an alternative scenario of earnings-indexation beyond the end of this parliament, with the cost of the triple lock being based on movements in earnings and prices over the thirty years from 1991 to the end of the their current forecast period in 2021. These projections are shown in the Figure below. Without the triple lock – the series shown in light green at the bottom of the figure – spending is projected to increase by 1.1% of national income between 2020–21 and 2060–61 (from 5.0% to 6.0%). This is equivalent to £21 billion in today’s terms. Under the triple lock spending (as shown by the dark green line) is projected to rise by 1.8% of national income over the same period (from 5.0% to 6.8%), which is equivalent to £35 billion in today’s terms, or some £15 billion more than under earnings indexation.

Figure. OBR projections of state pension spending

Source: Table 3.2, page 35 and Chart 3.10, page 58, of Office for Budget Responsibility, Financial Sustainability Report, January 2017, (

Both these projections assume that the Government raises the state pension over time in line with the intention, announced by the then Chancellor George Osborne in the 2013 Autumn Statement, “that people should expect to spend, on average, up to one third of their adult life in receipt of the State Pension”. Recommendations on the level of the state pension age beyond 2028 (by when it is due to have reached age 67) are to come in the next couple of months from the independent review led by John Cridland. But on the basis of the “one third of adult life” statement from the Government, and the latest central longevity forecasts from the Office for National Statistics, the OBR calculates this would see the state pension age for men and women rise to 68 by 2041 and to 69 by 2055. This is in contrast to current legislation which has it rising to age 68 by 2046 and then not rising any further.

The OBR also projects state pension spending under the scenario where the triple lock is left in place, but that the state pension age rises only in line with the already legislated increases. This is shown in the red line at the top of this figure. Under this scenario projected spending in 2060–61 is 0.5% of national income – or £10 billion in today’s terms – higher than under the central scenario.

This implies that a one year increase in the state pension age (from age 68 to 69) by the mid 2050s reduces projected state pension spending in 2060–61 by 0.5% of national income, or £10 billion in today’s terms. But in the same year the triple lock is projected to cost 0.8% of national income, or £15 billion, more than under earnings indexation alone. Therefore these projections imply that, over the next forty years, the additional cost of triple lock – rather than earnings – indexation is roughly equivalent to the cost of increasing the state pension age by 1½ fewer years. So, in other words, keeping the triple-lock and increasing the state pension age to 69 in the mid-2050s might have similar public finance implications to moving to earnings indexation and instead only increasing the state pension age to 67½.

That is not to say that if we were to abandon the triple lock then the state pension age should increase less quickly. After all, as the Figure above shows, earnings-indexation of the state pension along with increases in the state pension age to 69 by the mid 2050s still leads to projected spending increasing as a share of national income over the next fifty years. If instead we scrapped the triple lock the scale of the increase in spending on the state pension – and therefore the cuts to spending elsewhere, or tax rises, needed to fund it – would be substantially reduced.

]]> Mon, 27 Feb 2017 00:00:00 +0000
<![CDATA[How will the receipt of social care change in future?]]> How we fund and provide social care for the older population is an important policy issue. A National Audit Office report published this week suggests that the recent implementation of the Better Care Fund has so far struggled to reduce social care pressures imposing additional costs on NHS hospitals. The IFS Green Budget, also published this week showed that per-capita adult social care spending by government has fallen since 2009–10.

An important component of adult social care – which helps individuals to carry out everyday activities that they have difficulty with – is provided domestically in private residences.  This is distinct from the care provided to individuals in residential nursing homes (3.4% of the over 65 population in England lived in a residential nursing home at the time of the 2011 census), and can be informal, or either publicly or privately funded. Increases in the size of the older population – the most recent forecasts from the Office for National Statistics (ONS) estimate an increase of 84% in the 85+ population between 2010 and 2030 – will push up demand for these services. However, the extent to which demands rise in line with the population depends on how the needs of older people change in future, in addition to the ability of individuals to make alternative arrangements for care (such as informal care from a relative). A new IFS report, published today and funded by the Health Foundation, therefore examines the extent to which changes in the receipt of care across successive birth cohorts might offset (or add to) increasing demand for care from population growth.

The report uses information from the English Longitudinal Study of Ageing (ELSA). This provides a representative sample of the English population over the age of 50. Individuals are interviewed every two years, providing information on their economic circumstances, health and wellbeing. This includes information on the difficulties individuals face in doing everyday tasks, and whether (and from who) they receive assistance with these tasks. The report studies how the use of care has changed among those aged 65+ across four different birth cohorts (1915-24, 1925-34, 1935-44 and 1945-54).

One factor which may reduce the amount of care required going forward is that in the future each person will need less care at any given age. If individuals become healthier across birth cohorts then those of a given age will require less social care in future. For example, the typical 85 year old today may have very different needs than an 85 year old ten years ago.  A reduction in the per-person need for care over time may therefore help to offset some of the pressures from population growth.

The report indicates that the proportion of men aged 65 years and above reporting any difficulties with daily activities has fallen across birth cohorts. Men born between 1925 and 1934 were 4.7 percentage points less likely to report any difficulties at a given age compared to a man born between 1915 and 1924.  This difference was even larger when looking at men born between 1935 and 1944, who were 7.3 percentage points less likely to report any difficulties compared to the 1915-24 cohort at a given age. This compares with the 61% (81%) of men who were aged 65 (85) and above who reported any difficulties in 2010.  However, no corresponding differences were seen for women.  These results suggest that rates of need for formal care may therefore be falling for men, but seem less likely to do so for women.

A second reason why we might expect to see reduced rates of formal care provision in future is due to the increased availability of informal care provided by partners. One consequence of increasing life expectancy, particularly for men, is that individuals more commonly remain in couples at older ages – they are less likely to be widowed at any age. Unlike health care, where the majority of care is provided formally by the government through the NHS, the majority of social care is actually provided informally. In 2010, more than 50% of ELSA respondents who reported receiving any assistance with activities received some form of informal care. The most common source of informal care is from a spouse. With more couples surviving into old age more will be able to rely on informal care from a spouse and hence, perhaps, need less formal care. This is especially likely to be the case for women, as increases in male life expectancy (which remains lower than female life expectancy) have sharply increased the number of women with partners at older ages. The results indicate that this is indeed the case for women, with later cohorts of women receiving more help from a spouse. For example, women born 1935-44 are 5.8 percentage points more likely to receive care from a spouse at a given age than are women born 1915-24. No statistically significant results are found for men.

So, by comparison with earlier generations, men from later birth cohorts are healthier, and women from later cohorts are more likely to receive informal care. Does this lead to a reduction in the use of formal care across cohorts?  Results show that men born in later birth cohorts report receiving less formal care at a given age than their counterparts in earlier birth cohorts. The central estimates imply that the proportion of men aged 86-95 receiving formal care will be 2.4 percentage points lower in 2020 than in 2010.  This compares with the 13.2% of men in this age group who reporting receiving formal care in their own home in 2010.  However, the results do not show a corresponding change for women, among whom 29.9% reported receiving formal care in 2010.

What does this mean for likely levels of formal care required in future to meet the needs of the population? Taken together, the evidence suggests that a reduction in needs and the increase in informal care across birth cohorts will do little to offset the large increases in need for formal social care driven by demographic changes:  the estimates imply that demand for formal care for men aged 86-95 will still be 41% higher in 2020 than in 2010. For women, where there is no evidence of reduced per-capita use of formal care, demand for formal care will be 15% higher in 2020 than in 2010. These pressures will be even greater by 2030, with a population which will have aged further. Our estimates also suggest that there will be rather less in the way of an offsetting reduction in needs at this time.  Policymakers therefore should not rely on reductions in future needs to offset the additional cost of providing care in future, and should confront the tough choices over how to fund this care going forward.

]]> Fri, 10 Feb 2017 00:00:00 +0000
<![CDATA[Sweetening the sugar tax?]]> In Budget 2016 the Chancellor announced a ‘soft drinks industry levy’ that aims to reduce consumption of sugar sweetened soft drinks. The levy is due to take effect from April 2018 with two rates, one applying to mid-sugar drinks (with 5-8 grams of sugar per 100 millilitres) and a higher rate applying to high-sugar drinks (with more than 8 grams of sugar per 100 millilitres). A recent article in The Lancet: Public Health considers the possible consequences of the levy for a series of health outcomes, such as obesity, type 2 diabetes and dental care. In this Observation we propose a simple change to the soft drinks levy which would increase the likelihood of it having a beneficial effect on these outcomes.

The Lancet study is an important one, but it is also important to interpret the results accurately. The study carefully maps hypothetical changes in manufacturer behaviour through to health outcomes – the outcomes that the policy ultimately targets. However, in order to assess the impact of the levy (or any fiscal policy) there is an important first step, which is that we need to estimate the various ways in which manufacturers and consumers will respond to the policy. The paper is clear that “for each of these responses, the magnitude is uncertain” and it considers three hypothetical scenarios:

1) manufacturers lower the sugar content of mid-sugar drinks by 15% and high-sugar drinks by 30%,

2) manufacturers  increase prices by an amount equal to 50% of the tax on the product, with a maximum price increase of 20% and this feeds through to purchases based on estimates of consumer price responsiveness,

3) manufacturers change the marketing of mid-sugar and high-sugar drinks, change portion sizes and introduce new products such that there is a 12% reduction in the market share of high-sugar drinks and a 6% increase in the share of mid- and low-sugar drinks.

The article concludes that hypothetical scenario (1) would result in a larger reduction in sugar consumption from sugar sweetened drinks, and therefore would result in greater improvements in health outcomes, than either hypothetical scenarios (2) or (3).

What the analysis does not speak to is whether the soft drinks industry levy will have the hypothesised effects; will it lead manufacturers to reformulate drinks, as in hypothetical scenario (1), or to any of the effects hypothesised in scenarios (2) and (3)?

How manufacturers and consumers respond to the levy is complicated and uncertain. However, careful economic analysis can help shed light on the possible effects of the levy.

The government have set a revenue target, which the Office of Budget Responsibility has stated implies a rate of 18 pence per litre for mid-sugar drinks, and a higher rate of 24 pence per litre for high-sugar drinks. The dashed line in Figure 1 shows how the proposed tax per gram of sugar (shown on the right hand vertical axis) varies with the sugar intensity of drinks (the amount of sugar per 100 millilitres). The bars show the distribution of the sugar contents across all soft drinks purchased in 2010-11 (with fraction of purchases shown on the left hand vertical axis).

An obvious and important feature of the tax is that there is not a straight line relationship between sugar content and tax charged. There are two ‘kink points’ and between these points the tax per gram of sugar is declining as the amount of sugar increases. This has implications for the strength of incentives that manufacturers will have to reformulate products.

Figure 1

Notes: Fraction of purchases is based on the authors’ calculations using the Kantar Worldpanel for 2010-2011.

Consider, for example, a manufacturer of a 1 litre bottle of drink that contains 11 grams of sugar per 100 millilitres (this is the highest bar in the figure, meaning this was the most commonly bought type of sugary drink). Under the proposed soft drinks levy they face a tax of 24 pence per bottle or 22 pence per 100 gram of sugar in that product. If they reduce the sugar content to 10 gram per 100 millilitre then the tax paid is still 24 pence, and the tax rate increases to 24 pence per 100 gram of sugar, while if they increase the sugar content to 12 gram per 100 millilitre the tax paid remains at 24 pence and the tax rate decreases to 20 pence per 100 gram of sugar. Therefore, the soft drinks industry levy creates no incentives, within each of the mid- and high-sugar product bands, for manufacturers to reduce the sugar in their products. It is only if a manufacturer moves its product from the high- to mid-sugar or mid- to low-sugar band that it will lower the tax it faces. That means that few manufacturers will have any incentive to lower sugar content a little. They will effectively have a choice between a big change in product formulation to lower the tax they face, or no change at all. Big changes may be difficult to implement if consumers have a strong enough taste for sugar.

An alternative, and simpler, tax design would be to tax products at a constant rate per sugar content; an example of such a tax is shown by the dotted line in the figure at a rate of 20 pence per gram of sugar. This would lead to clear incentives for all manufacturers to reduce the sugar content of their products. Consider again a manufacturer of a 1 litre bottle of drink that contains 11 grams of sugar per 100 millilitres. Now the product will face a tax of 22 pence. Lowering the sugar content of the product by 1 gram per 100 millilitres reduces this by 2 pence and increasing the sugar content by 1 gram per 100 millilitres raises tax due by 2 pence. This alternative system could also incorporate bands, like the soft drinks industry levy, to create additional incentives for manufacturers to lower the sugar content of their products sufficiently to move their product into lower tax bands.

A tax per gram of sugar is unlikely to have any higher administrative burden than the proposed levy.  It is already necessary to know the sugar content of the product in order to determine into which band it falls. This is the only information that is necessary to levy a tax per gram of sugar. We already operate taxes of this form, both beer and spirits excise taxes are applied per gram of ethanol, rather than per litre of product.

The soft drinks levy may be a reasonable starting point in attempting to reduce excess sugar consumption: there is evidence that households that buy a lot of sugar and households with children purchase a disproportionate amount in the form of soft drinks. However, we believe the effectiveness of the levy could be substantially improved, and the chances of the reformulation hypothesised in the favourable scenario (1) of the recent Lancet paper improved, by a simple and feasible change to its structure.

The new soft drinks industry levy is not scheduled to come into effect until April 2018. The chancellor has time to adjust the structure of the tax to make sure it better targets the sugar content of products and thereby increase the likelihood that it will be effective at encouraging product reformulation.

Note: We gratefully acknowledge financial support from the European Research Council (ERC) under ERC-2009-AdG-249529 and ERC-2015-AdG-694822, from the Economic and Social Research Council (ESRC) under the Centre for the Microeconomic Analysis of Public Policy (CPP), grant number RES-544-28-0001 and under the Open Research Area (ORA) grant number ES/I012222/1

]]> Fri, 16 Dec 2016 00:00:00 +0000
<![CDATA[How far do today’s social care announcements address social care funding concerns?]]> In yesterday’s English Provisional Local Government Finance Settlement, the government announced councils will be able to set a ‘Social Care Precept’ of 3% a year over the next two years, rather than the 2% a year previously planned, to raise additional funds for adult social care. Coming on top of the 2% increase in bills councils are allowed to levy to generate funding for other services (such as libraries, children’s services and refuse collection), if these new powers were used in full, Band D council tax could increase by an average of £67 (£27 in real-terms) in April 2017 and £136 (£56 in real terms) by April 2018.

However, the flip side is that if councils make full use of the 3% precept in each of the next two years, they will not be able to use the precept in April 2019: a cap of 6% in total over the next three years applies. In other words, council tax increases can be brought forward to raise money in the short term, but this will do nothing to plug the longer-term funding issues adult social services (and councils more generally) face. Furthermore, the new ‘Adult Social Care Grant’ also announced yesterday is only available in 2017–18, and is largely a relabeling of money councils were going to receive anyway.    

We calculate that yesterday’s announcements could increase the amount available to spend on adult social care by a maximum by £700 million over the next two years relative to previous plans. But they provide no boost to spending beyond that.

Adult social care spending and the social care precept

In the last financial year, 2015–16, councils’ net expenditure on adult social services was £14.6 billion, a little over one third of their overall spending on services (excluding education). Measured on a consistent basis, this has fallen in real terms by almost 20% since 2009–10 despite a growing elderly population.

In the 2015 Spending Review, the Chancellor announced two policies to provide additional funding for this area:

  • The ‘Social Care Precept’ allowing councils with social care responsibilities to increase council tax by an extra 2% a year (on top of the usual 2% limit) if that funding were allocated to adult social services. This raised around £380 million in 2016–17, which would have risen to £1.7 billion per year in real-terms by 2019–20, if used in full (See row 1, Table 1).
  • The ‘Improved Better Care Fund’ was additional grant funding, amounting to £100 million a year in 2017–18 and £1.4 billion a year in real-terms by 2019–20 (see row 2, Table 1). This funding was to be allocated to councils in such a way as to offset differences in their ability to raise their own revenues for social services via council tax.

Taken together, these plans meant £3.1 billion of ‘extra’ funding ring-fenced for social care by 2019–20. If this money were added to existing budgets, real-terms spending on adult social care would have grown 1.3% in real terms this year and 14.7% in real terms by 2019–20.

Table 1. Potential additional resources for adult social services, £ millions (2016–17 prices)

Policy measure





Original plans





(1)      Social Care Precept





(2)      Improved Better Care Fund





(3)      Total ((1) plus (2))










New plans





(4)      Social Care Precept





(5)      Improved Better Care Fund





(6)      Other extra grant funding





(7)      Total ((4) plus (5) plus (6))










Total extra funding under new plans ((7) minus (3))






Note: Figures for 2016–17 reflect revenue forecasts based on actual usage of Social Care Precept in that year, as reported in Figures for later years assume full use of Social Care precept by relevant authorities (on top of 2% increases in council tax rates each year for general purposes). The ‘Other extra grant funding’ refers to the net effect of the shift of funding from the New Homes Bonus to the Adult Social Care grant for councils with social care responsibilities. It reflects a net shift of £75 million from District Councils to councils with responsibility for social care.

Councils’ budgets for this year show them planning to spend £14.4 billion on adult social services – actually less than the £14.6 billion spent last year, despite the £380 million forecast to be raised from the ring-fenced ‘Social Care Precept’.  This has happened because councils’ overall budgets fell between the two years and the way the ring-fence works is that councils simply have to confirm that they are using the precept and Better Care Fund to spend more than they otherwise would have on such services. Thus a fall in spending is perfectly consistent with satisfying the ring fence if councils state that they have cut spending by less than they otherwise would have done.

Looking ahead, the Government yesterday announced that in April 2017 and April 2018, councils will be able to set the precept at 3% rather than 2% as previously planned. If used in full, this would raise a little over £200 million more than previous plans next year, and £425 million more in 2018–19.

The government also announced a new £240 million Adult Social Care Grant for 2017–18. This comes from reductions to the generosity of the New Homes Bonus, paid to councils based on the number of new homes that are built in their areas. Thus, councils with social care responsibilities see a net increase in their grant funding of only £75 million in real-terms (paid for by a net cut to grants to District Councils, who will also see smaller New Homes Bonus payments, and are not responsible for social care provision). Added to the revenue from the higher ‘Social Care Precept’, the new plans therefore provide an additional £280 million in ring-fenced funding next year on top of the extra already in the pipeline. This is not an insignificant sum, but is small relative to the scale of the estimated funding gap in the social care sector – see for example this joint paper by the Nuffield Trust, Health Foundation and Kings Fund which estimated a funding ‘gap’ of £1.9 billion in 2017–18, for instance).  

Importantly the amount spent on social care may well not go up in line with the ring-fenced grants: councils could still cut spending provided that they state that they have used this money to cut by less than they otherwise would.

Councils’ overall spending power, including the ring-fenced funds for social care funding, is set to fall 2.6% next year. If social care spending fell in line with this overall cut, that would mean a cut of around £375 million in real-terms. If, on the other hand, the ring-fenced funds are used in full to boost spending on social services, spending would rise by around £425 million (3%) in real-terms compared to this year. But to balance their budgets councils would need to make cuts of around £1.5 billion (5.3%) to other service areas.  

Looking further ahead, if councils make full use of the 3% precepts in 2017–18 and 2018–19, they will not be able to levy an additional precept in 2019–20 (the usual 2% limit on council tax would apply). In other words, the main impact of yesterday’s announcements was to bring forward funding increases for adult social care. How the government plans to address longer-term spending pressures arising from an ageing population, increases in labour costs as a result of the National Living Wage, and the costs of a more generous social care regime planned for 2020 and beyond, remains unclear.

Impact across councils

Because of differences in council tax revenues and social care needs around the country, the relative amount councils can raise from the ‘Social Care Precept’ varies substantially between councils: areas with high valued homes and low needs can raise relatively more. The ‘Better Care Fund’ is being allocated to compensate for these differences, so that by 2019–20, the overall additional funding available is much more equally distributed around the country.

However, the backloading of the Better Care Fund (it is set to be just £100 million next year), and the front-loading of the increases in council tax via the ‘Social Care Precepts’ mean that in 2017–18 and 2018–19, areas with lots of high valued properties and/or low needs will see a relatively bigger increase in the resources available for social care. Overall, yesterday’s announcements brought forward into the next two years relatively more funding in areas with higher valued properties and/or lower needs.

Areas with lots of low valued properties and/or higher needs will then see a bigger boost than more affluent areas in 2019–20 as the ability to levy additional Social Care Precepts is exhausted and the Better Care Fund – targeted at those low valued / high needs areas – continues to increase.

Households’ council tax bills

We now turn to what yesterday’s announcements mean for households’ council tax bills.

In 2015–16, the average Band D council tax bill in England was £1,484. Adjusting for consumer price inflation, this was 7% lower than in 2010–11, reflecting a widespread council tax freeze. The ending of the council tax freeze and the introduction of the ‘Social Care Precept’ led to a 3.1% cash-terms (2.1% real-terms) increase in council tax in 2016–17, taking the average Band D rate to £1,530.

All councils outside London able to use the precept report using the full 2% allowed in 2016–17. However, at the same time, a substantial minority of councils decided not to make use of the full 2% increase to council tax for general purposes. This, of course, kept households’ bills down, but means that these councils did not raise as much as they could via council tax to support local services (including adult social services).

Looking to the future, the amount of council tax households in England will pay will depend on the extent to which councils make use of the new higher 3% social care levy over the next two years, and the increases in their ‘general’ council tax rates. If all councils make full use of the precept, and of the 2% increases to standard bills that are allowed, average Band D council tax would increase £67 next April (£27 in real terms), and would be £136 higher by April 2018 (£56 in real terms). Increases in different parts of the country will vary with the biggest increases allowed in those areas where council tax rates are already highest (because the standard 2% increases and 3% precepts are bigger when initial council tax rates are higher).

However, council tax bills would still be lower in real-terms in 2018–19 and 2019–20 than they were in 2010–11: the increases in bills this year and over the next few years will not fully undo the coalition government’s council tax freeze.


Yesterday’s announcements therefore represent a modest increase in funding for social care in the short-term, largely paid for by above-inflation increases in council tax, and a small reduction in grants to District Councils next year.

]]> Fri, 16 Dec 2016 00:00:00 +0000
<![CDATA[Raising GCSE attainment crucial to get more young people from disadvantaged backgrounds into university, but work to promote social mobility cannot end when they arrive on campus]]> Reducing socio-economic gaps in education outcomes has been at the heart of government strategy to raise social mobility for many years. Achieving higher educational qualifications enables individuals to earn more, on average, so if those from poorer backgrounds are less likely to attain these qualifications than those from richer backgrounds, then the socio-economic circumstances of parents and children will continue to be inextricably linked.

Because the benefits of going to university have been found to be large in the UK, access to higher education has become a focus of efforts to raise social mobility – although it is far from the only route. Despite decades of policy action, however, young people from richer families in England are still around three times more likely to go to university than their peers from poorer backgrounds, with nearly 60% of state school students from the least deprived fifth of families going to university at age 18 or 19 compared to less than 20% of those from the most deprived fifth of families.

Stark statistics like this are at the heart of why recent governments have pledged to increase the number of students from disadvantaged families, schools and neighbourhoods who go on to university – and some (marginal) progress has been apparent in recent years. But the slightly faster growth in university participation rates amongst those from poor backgrounds over the last decade or so pales in comparison to the overall size of the gap: a reduction of 1-2 percentage points in a gap of nearly 40 percentage points does not signify much in the way of substantial progress.

A recently published book, Family Background and University Success, written by researchers at the Institute for Fiscal Studies, the UCL Institute of Education, and the Universities of Cambridge and Warwick, the key findings of which are due to be presented at a conference at the Nuffield Foundation today, draws together the latest quantitative evidence for England, to provide new insight into what drives these gaps – and hence what policymakers might need to do to reduce them. It also considers socio-economic differences in access to different degree courses, as well as dropout, degree class and labour market outcomes. Efforts to get more young people from poor backgrounds into university will be ineffective if those students do not attend the best courses for them; if they do not complete their degree or get a high degree class; or if their labour market prospects are still not as good as those of their richer peers.

One of the key messages highlighted by the book is the crucial role played by attainment earlier in the school system in explaining socio-economic gaps in university access. This point is illustrated by the figure below, which takes as its starting point differences in the likelihood of going to university between state school students from the 20% richest and poorest families in England. When looking at attendance at any university in the UK (the bright green bars) this equates to a gap of around 37 percentage points.

Differences in the % of state school students from the richest and poorest 20% of families who go to university, controlling for attainment at different ages

Differences in the % of state school students from the richest and poorest 20% of families who go to university, controlling for attainment at different ages

Source: Figure 5.1 of Crawford et al. (2017), Family Background and University Success, Oxford University Press.

The remaining bars show how this gap changes when differences in attainment at various ages are taken into account. In other words, they show how much more likely a rich student with a given level of attainment at a particular age is to go to university than a student from a poor background with the same level of attainment at the same age. The smaller the remaining gap, the more important is that measure of attainment in explaining why pupils from rich backgrounds are more likely to go to university than those from poor backgrounds, and hence the more important it is to raise attainment at that age amongst poorer students in order to reduce the socio-economic gap in university access.

The second and third bright green bars show that the remaining unexplained gap falls substantially when we account for measures of attainment during primary school (at age 7 or age 11), suggesting that how well children do during this early phase of education is already highly predictive of whether they are likely to go to university. But it is the measures of attainment at the end of secondary school that are most important. The (missing) fourth bright green bar shows that accounting for which qualifications young people take, in which subjects, and which grades they receive at age 16, can explain all – not nearly all, but all – of the gap in university access between the richest and poorest students.

The importance of prior attainment in explaining socio-economic gaps in university access is not a new finding. But this figure uses more detailed measures of attainment than many previous studies, and is the first to find that the gap can be entirely explained by a rich set of measures of attainment at age 16. This highlights the importance of intervening early: if the government wants to substantially reduce or even close the socio-economic gaps in university access, then it must do more to increase the attainment of students from poorer backgrounds by the end of secondary school.

The dark green bars in the figure focus on the selected group of students who make it to university and attend a ‘high status’ institution. There are, of course, many potential ways to define ‘high status’; here we use information on the university’s average research quality, taking the 24 Russell Group institutions and adding any others whose average research quality is above the lowest in that group, giving around 40 institutions in total. These bars show how much more likely a state school pupil who goes on to university from one of the 20% richest families is to attend a ‘high status’ institution than a state school pupil who goes on to university from one of the 20% poorest families.

Looking across all state school students, around 28% of university entrants attend one of these ‘high status’ institutions. But there are large socio-economic differences here too. Nearly two fifths of university entrants from the 20% richest families attend a ‘high status’ institution compared to just 15% of university entrants from the poorest fifth of families – a gap of 23 percentage points.

Comparing this raw gap (the first dark green bar in the figure) to the other dark green bars shows that the relative importance of attainment at different ages that we saw across all institutions is repeated here too – with one exception: while we were able to explain all of the socio-economic gap in university attendance using attainment at age 16, we are not quite able to do this when looking at our group of ‘high status’ institutions. There remains a small – but statistically and economically significant – gap of around 4 percentage points in the likelihood of attending one of these institutions between state school students from different socio-economic backgrounds.

It is not clear from the evidence we have how much of this gap arises from differences in application rates to these institutions by students from different socio-economic backgrounds, differences in offers made by universities to students from different backgrounds, or differences in acceptance rates of these offers – but it is clear that more needs to be done to ensure students from lower socio-economic backgrounds are fully informed about the potential benefits of attending one of these institutions before making their university application decisions. (The book also shows that graduates from these institutions tend to go on to earn more in the labour market.)

Just as when looking at overall university access, however, the key to reducing these gaps remains increasing the academic performance of students from lower socio-economic backgrounds earlier in the school system. Of course, there are no easy answers on how to do this. But universities can and do play a role in trying to achieve these aims, with many institutions now working with disadvantaged young people or schools to help increase attainment at secondary or even primary school. The evidence on how effective such programmes are at raising attainment and increasing progression to university is weak though. The Education Endowment Foundation has been blazing a trail in this respect by trialling programmes to raise attainment amongst disadvantaged students in English schools over the last 5 years. But there is a clear need for more rigorous evidence on ‘what works’ to emerge from the university sector too.

Just as vital, though, is that universities continue to support students – especially those from non-traditional backgrounds – once they arrive on campus. Even amongst the selected group of students who make it to university, the book highlights further differences in the likelihood that those from different socio-economic backgrounds will complete their degrees and achieve a 1st or a 2:1 at the end of their course. For example, comparing students who arrive at university with the same school qualifications, and who study on the same degree course at the same institution, those from the 20% richest backgrounds are around 5 percentage points more likely to complete their degree within 5 years and, conditional on having completed their degree, around 4 percentage points more likely to achieve a first or a 2:1 than their counterparts from the 20% poorest backgrounds.

The book also shows that students from lower socio-economic backgrounds are less likely to work in professional occupations after they graduate and are likely to earn less over their working lives than graduates from richer socio-economic backgrounds. For example, even comparing graduates from the same courses at the same universities, those from higher income families earn around 10% more, on average, than those from lower income backgrounds 10 years after graduation. Similar differences also persist at older ages and are only slightly reduced (to around 6%, on average) by accounting for all manner of other ways in which these graduates differ from each other – including prior school attainment and degree class, as well as ‘non-cognitive’ (social) skills.

Does this mean that those from lower socio-economic backgrounds should not bother going to university? Definitely not. In fact, some evidence suggests that the ‘returns’ to university – the average difference in earnings between those who do and do not obtain a degree – are higher for those from poorer backgrounds. Thus, while graduates from poorer backgrounds tend to earn less than those from richer backgrounds, average earnings for those who don’t go to university are also lower amongst those from poorer backgrounds – and the socio-economic gap is even larger amongst the group of non-graduates.

These findings all suggest that socio-economic gaps in educational attainment open at an early age and grow over time. All of the gap in university entry by social background can be explained by attainment at age 16. But gaps remain in terms of the type of institutions students access, and open up again at university and beyond, with differences in degree completion and degree class, as well as labour market performance. Closing the gap in both attainment and longer term outcomes is clearly going to require action throughout the school years and into, and beyond, university as well.

End note:

Copies of the book can be purchased from Oxford University Press. Free access to the book can be arranged for members of the press: please contact Bonnie Brimstone for further information. 

Some, but not all, of the material covered by the book can be found in earlier freely available publications. For example, previous IFS research has highlighted the importance of prior attainment in explaining socio-economic gaps in university access. It has also considered the determinants of socio-economic differences in dropout, degree completion and degree class and labour market outcomes 10 years after graduation, as well as mid career.

]]> Mon, 05 Dec 2016 00:00:00 +0000
<![CDATA[Council-level figures on spending cuts and business rates income]]> Council-level figures on spending cuts and business rates income

Last month, researchers at the Institute for Fiscal Studies launched the first paper from a new programme on local government finance. This paper looked at a range of issues including, changes in councils’ spending and revenues over the last seven years, and issues related to the evolving English business rates retention scheme (BRRS). Today, we publish two spreadsheets with information for individual council areas:  a spreadsheet showing changes to councils’ spending on services between 2009–10 and 2016–17; and a spreadsheet showing relative gains and losses from the BRRS since it was introduced in 2013–14.

Changes in councils’ service spending

The first spreadsheet looks at how spending on services by local government has changed in different areas of the country over the period 2009-10 to 2016-17. This is complicated for two reasons. Firstly, in England one area can have services provided by multiple authorities (e.g. by a district and a county council). Secondly, the responsibilities of local government have changed over this period as services have been devolved.

We correct for the former by treating all areas of England as though they are a unitary authority – adding the spending of county councils to that of their constituent district authorities, and splitting the spending of combined authorities between their members. We correct for changes over time by excluding spending on areas of new or changing responsibility, in particular education, public health and some aspects of social care.

On this basis, we find that real-terms cuts to local authority service spending in England have varied from 4% in East Riding of Yorkshire and 5% in Hampshire* to over 40% in several authorities including Westminster (46%) and Salford (45%). Almost 1 in 3 councils** have faced cuts of 30% or more. As a result of the way grants have been cut local authorities that received the largest share of their funding from government grants in 2009–10 experienced the largest cuts to their service spending. The 10% of authorities most dependent on grants in 2009-10 received an average cut of 33%, compared to 12%*** for the 10% of authorities least dependent on grants.

Cuts to local government grants in Scotland and Wales were smaller over this period than in England, and as a result service spending has fallen by less on average.  In Wales the size of cuts also varied less across local authorities, ranging from 6.5% in Wrexham to 22.6% in Denbighshire. However in Scotland, the variation across authorities was almost twice that of Wales, ranging from a small increase in East Renfrewshire (1.6%) to a cut of 29.0% in Glasgow City.

Relative gains and losses from the Business Rates Retention Scheme

The second spreadsheet looks at relative gains and losses from the BRRS between 2013–14 and 2016–17, updating earlier analysis presented in our first paper.

To do this, we first calculate each council’s income from the rates retention scheme using outturns data (NNDR3 returns) for 2013–14 to 2015–16, and estimates as of January 30th 2016 for 2016–17 (from NNDR1 returns). As figures are not readily available on how income has been distributed within “business rates pools” made up of multiple councils, we have used a stylised pooling scheme based on the features of a number of pooling agreements (see methodological notes below for full details). Any safety net payments (to prevent very large losses) or levies (on revenue growth for councils with high business rates revenues) are also taken into account.

We compare this income to what a council would have received if its baseline funding level – the amount of revenue from the scheme each authority was judged to need at the start of the scheme in 2013 – had instead been increased in line with growth in national business rates revenues. Gains or losses relative to this baseline are presented in cash-terms, and as a percentage of councils overall funding (excluding grants specifically for education, housing benefit, or public health).

It is important to note that the figures presented are gains/losses relative to this counterfactual sharing of business rates revenues: we do not know how much councils would have received in the form of grants from central government in the absence of the scheme.

Results shows a wide range of gains and losses from the scheme. 24 councils see income gains between 2013–14 and 2016–17 equivalent to more than 5% of their overall funding, with Suffolk Council’s 24% gain equivalent to £16.5 million over four years. At the other end of the spectrum, 127 councils have experienced lower income than they would have received if revenues were pooled nationally. Westminster is recorded as suffering the greatest fall in both cash and percentage terms – 2.8% or £31.2 million over the same four years. This large loss reflects the fact that Westminster has been making large provisions in case of successful appeals against rateable values. If the appeals are ultimately unsuccessful, some of these losses will not actually materialise and their position will ultimately look less bad.    

Gains and losses reflect not only relative business rates performance in different parts of the country but the design of the scheme. District councils, for instance, retain 40% of the growth of business rates income in their areas. We estimate that as a result, as a group they will have gained almost £300 million over the last four years, equivalent to 2.4% of their overall budgets. The counties that cover the same areas receive only 9% or 10% of the growth in business rates and rely instead on top-ups to their rates income that increase online in line with inflation. As a result they have gained just £10 million (or 0.02% of their much larger overall budgets) during the same period (and would actually have lost were it not for the fact that they  have gained as a result of pooling with district councils). Fire authorities, London boroughs and metropolitan boroughs have all lost a little relative to what they would have received if rates income had been pooled nationally.

The second worksheet in this spreadsheet shows what would have happened if there were no safety net payments or levies on growth. Unsurprisingly gains, and especially losses, would have been larger. Westminster, for instance, would have lost the equivalent of 20% of its budget – £224 million over four years – as a result of its large provisions for appeals, and the fact it would not have received the £193 million of safety net payments that in reality it is set to.

Looking ahead, IFS researchers will build on this analysis, to consider how different councils may fare under different scenarios for revenue growth and different policy options for the 100% business rates retention scheme now under development.

This observation is part of a new programme of work on local government finance and devolution funded by a consortium of funders including Capita, CIPFA, the ESRC, PwC, the Municipal Journal and a range of councils across England.

Methodological note on Pools

This note explains how we treat Business Rates Pools in our figures.

First, all councils within the pool receive what they would have received from the BRRS if outside the pool. Second, if there is a surplus after this has been done, 50% of this surplus is shared proportionate to each council’s baseline funding level and 50% is shared proportionate to each council’s growth in business rates income above its baseline funding level for that year (among those that have experienced positive growth). If there is instead a deficit, the burden of this is shared in entirety in proportion to baseline funding level.

In two-tier areas, typically upper tier authorities (counties) receive most of the gains allocated according to baseline funding, and lower tier authorities (districts) receive most of the gains allocated according to business rates revenue growth.



Please note, figures for a minority of councils have been corrected. In most of these cases the cuts are now estimated to be larger, but the changes are small in magnitude relative to the overall variation in cuts and does not affect the overall picture. Specific changes to the text include the following.

*This used to state that the largest cuts had been 0.6% in Surrey and 1.3% in Hampshire

**This used to say '1 in 3' councils.

***This used to say that the least grant dependent decile had face a cut of 9% on average.

]]> Mon, 28 Nov 2016 00:00:00 +0000
<![CDATA[The distribution of healthcare spending: an international comparison]]> A special issue of Fiscal Studies published today looks at patterns of individual-level health spending across a range of countries and finds some important similarities. It shows how health spending is concentrated in the last years of life, how significantly more is spent on the poor than on the rich and how health spending tends to be concentrated on a relatively small number of people with high needs.

The countries considered – Canada (Quebec), Denmark, England, France, Germany, Japan, the Netherlands, Taiwan and the United States – have health care systems that differ in important ways, such as the role of government in the funding and provision of care and the extent to which patients face direct costs for treatment they receive. However, all have experienced rising health care costs as a share of national income over the past few decades and face the pressures from ageing populations and new effective but expensive treatments.

National accounts data from 2012 show that the US, which operates a market-based system, spends a much higher share of national income on health care (16%) than European countries and Japan (where shares are typically between 8% and 11%). Much less is known about the extent to which these health care systems also lead to differences in how health care spending is distributed across the population. The papers published today use individual-level data on health care use and costs from national health care systems and private insurers to provide new evidence on the similarities and differences in patterns of medical spending across countries. The work focuses on three main aspects of spending, important for designing health care policy and understanding the principal drivers of rising demands on health care budgets: the relative cost of end-of-life care, where the high costs associated with dying have been a focus of academic and policy interest; the distribution of health care spending by income; and the share of health care spending accounted for by the most expensive patients.

Comparing the patterns of spending across all the countries provided five main findings:

  • First, the care of those in the last year of life is costly, but represents a relatively modest fraction of total health care spending. In all of our countries, those who die account for less than 1% of the population each year, but at least 4% of total medical spending. In England, where information is restricted to hospital care, £1 in every £10 spent on hospital care is spent in the last year of life. Among the population aged 65 and over, £2 in every £10 is spent in the last year of life. These figures for England are towards the top of the range found for other countries on similar measures of spending. For all the countries we study, these costs are sizable, but do not appear large enough to be the key driver of differences in medical spending either over time or across countries. To place these figures in context, while total spending on the NHS in England doubled in real terms during the 2000s, reducing spending on end-of-life care by half would reduce total spending on hospital care by 5%.
  • Second, we present new evidence on the income gradient of medical spending. Total medical spending typically decreases with income in any given year, with poorer people consuming more medical resources than richer people, both for the population as a whole and within any given age groups. In England, those living in the poorest fifth of local areas have hospital costs that are 14% higher than the richest fifth, for the population as a whole. Those aged 65 and over in the poorest areas have 35% more spent on them than those in the same age group in the richest fifth of local areas. In Denmark, where information on income is available at the individual level, average medical expenditure for the poorest fifth of the population is around two–and–a-half times that of the richest fifth.

We do not attempt to adjust for medical needs, and therefore make no claims about whether or not the distribution of medical spending is fair. However, in England, Denmark and elsewhere, health expenditures are financed through progressive taxation. The patterns of spending by income therefore indicate that such health care systems typically act to redistribute resources from rich to poor. Furthermore, we show that in countries that use patient cost-sharing (where patients must pay directly for a certain share of the cost of medical care), such as the US and Taiwan, the share of spending accounted for by the poorest in society is lowest. Thus, while patient cost-sharing may help control spending on medical care goods and services, it also means that more of those goods and services will flow to richer patients who can better afford them. Countries that ration health care using non-price mechanisms, such as the guidelines of the National Institute for Health and Care Excellence in England and Wales, have greater health care use by low-income people relative to those with high incomes.

  • Third, health care spending in any given year is highly concentrated in all countries. For hospital spending, a tenth of the population in England account for £8 in every £10 spent on hospital care. Similar levels of concentration are found in all other countries. For countries that have more information on broader measures of health care costs, which include services such as primary care and prescriptions, concentrations are a little lower at between a half and two-thirds of spending accounted for by a tenth of the population.
  • Fourth, the US is a clear outlier in terms of total medical spending as a share of national income and average total spending per person, but looks much more similar to other countries in terms of how that spending is concentrated. Although the levels are higher in the US, health care spending is somewhat less concentrated in the US than in most countries, in terms of both the share accounted for by the top medical spenders and the share going to those in the last year of life. This picture is consistent with much higher health care spending in the US generated by higher prices or greater use of medical goods and services for the average patient, rather than with very high levels of expensive care received by a small number of people at the end of their lives.
  • Finally, the data available on medical spending in England are far less comprehensive than those for comparable countries in Europe, despite operating a national public health system. Information is restricted to hospital use, which constitutes less than half of NHS spending. The scrapping of the programme, which sought to bring together information from different health and social care settings, means that unfortunately this situation is unlikely to change in the near future. Moreover, maximising the use of hospital data that are available in England is hampered by restrictions on access and delays in linking to survey data, even when there is consent from individuals to do so. This places limitations on the extent to which patterns of individual-level spending and the links between costs and service provision in different parts of the health and social care system can be fully understood. The fact that data coverage and access are far more restricted in England than in many other countries reduces the amount of research carried out on the efficacy of health care delivery across the NHS, thereby making it harder to uncover possibilities for service improvement. Ultimately, patients may pay the price for this.

Percentage of total medical spending on those in the last calendar year of life as a share of aggregate spending


All medical care,
excluding long-term care

Hospital care

Aged 65 and over




















United States































United States



Note: Spending is as a share of spending in that category and for that age group. For example, hospital care spending for those aged 65 and over is hospital spending of those aged 65 and over in the last calendar year of life divided by total hospital spending of everyone aged 65 and over in that country. Germany and Japan use private insurance data and thus estimates from these countries may not be fully representative of the populations of these countries. Data for England are restricted to hospitals only.  All entries of – indicate that the information or breakdown is not available.


]]> Thu, 17 Nov 2016 00:00:00 +0000
<![CDATA[The fall in sterling: who is hit by the rise in inflation?]]> This morning the Office for National Statistics announced that CPI inflation rose to 0.9% in the year to October, down from an inflation rate of 1.0% in the year to September but still substantially up from 0.6% in the year to August. The Bank of England expect inflation to rise further, to 2.4% in 2017 and 2.8% in 2018 – considerably higher than the 1.5% and 2.1% expected back in May. Most of this forecast increase is driven by the recent devaluation of the pound, which pushes up the price of imports. In this Observation we look at how the overall 2.5% increase in the price level which is likely to result from sterling’s decline since the June referendum will affect the prices of different goods. We then look at whether this is likely to have a bigger effect on poorer or richer households.

Since the eve of the EU referendum in June, the value of sterling has fallen by around 12%. The Bank of England estimate that 60% of a fall in sterling is passed through into higher prices faced by UK firms and consumers (this is less than 100% as foreign firms selling into the UK may cut their prices in response to reduced UK demand). UK firms can deal with this increase in costs by switching to (now comparatively cheaper) domestic producers, allowing their profits to fall, or raising their prices. The Bank estimate that in the longer run (around three years) the latter effect dominates, with all of an increase in import costs being passed through to UK consumers in the form of higher prices.

This affects not just goods bought directly from abroad, but also goods which are produced using imported inputs. The stories of British brands raising prices in the wake of the depreciation due to the increased cost of imported ingredients are possible examples of this indirect effect.

Those goods and services that have higher import contents – either through direct imports or inputs into products made in the UK – are likely to see larger price rises following a devaluation. Table 1 shows the expected long run effects on the prices of several broad categories of goods, covering around a third of average household spending. Bear in mind that the table simply isolates the expected impact of the fall in sterling, and is not an overall forecast of inflation for these goods which will also be affected by factors other than exchange rate movements. The Table also shows how richer and poorer households tend to spend different proportions of their budgets on different items.

Table 1: Budget shares by expenditure quintile and expected price increases as a result of the 12% fall in sterling since June 23rd, selected items


Expected price increase (%)

Budget share by expenditure quintile (%)







Clothing and footwear
























Household goods & domestic services








Petrol and other vehicle running costs
















All (CPI 2016 weights) a




All (Budget share weights) a




Source: Authors’ calculations using Living Costs and Food Survey (LCFS) 2014, ONS, Import intensity for each COICOP class, Indirect import content of domestic final demand by product and component 2010, and Input-output supply and use tables 2010.

(a) For ”budget share” weights we use the average LCFS budget share for each item. CPI weights, which are appropriate for calculating the overall price level, are calculated using a different source and different methodology.

Note: See methodology section below for assumptions. The expected price increase of the ‘clothing and footwear’ item is calculated using CPI 2016 weights. Budget shares are calculated at the two digit Classification of Individual Consumption according to Purpose (COICOP) level. Items in the table correspond to COICOP groups of the same name, except “Household goods and domestic services”, which refers to the COICOP group “Goods and services for routine maintenance”, and “Petrol and other vehicle running costs”, which refers to the COICOP group “Operation of personal transport equipment”. Budget shares are calculated for out of spending on goods covered in the CPI (so excluding mortgage interest). Figures are for Great Britain.

Almost 40% of food is imported, meaning that it is expected to see a relatively large price increase of 2.9% as a result of sterling’s devaluation. Clothing and footwear – with import contents of 40% and 50% respectively – are also affected significantly. Petrol and other vehicle running costs see a substantial price rise, but for a slightly different reason. Although only a modest share of petrol is imported, oil is traded on a world market denominated in dollars, which makes even domestically-produced oil highly sensitive to the sterling-dollar exchange rate. Conversely, the prices of more service-oriented items such as education or household goods and domestic services, and more heavily taxed items such as tobacco, tend to be less affected by changes in import prices.

Because different households spend their money on different things, the pattern of price increases summarised above can mean that some households are hit harder than others. For example, households which spend especially high fractions of their budgets on petrol, food and clothing would tend to be hit harder than average.

Clearly the big increases in food prices will hit poorer households particularly hard – while on average 16% of household spending is on food, almost a quarter of the spending by poorer households goes on food. On the other hand increased petrol and clothing prices hit richer households harder. Put these effects together and we find that essentially all major demographic and income groups are affected similarly. Figure 1 shows what happens when we use our budget shares to weight the price increases of different goods. Each expenditure decile is expected to see costs rise by roughly 2.7% on average, though with a slightly bigger effect for richer households. Broadly, this is because those at the upper end of the expenditure distribution are more affected by rises in vehicle, holiday, and furniture costs, while those at the lower end are more affected by rises in food, telephone, and utility costs (because these account for a relatively large share of their budgets). In combination this gives the fairly flat profile seen in Figure 1.

Figure 1: Average expected increase in costs as a proportion of expenditure as a result of the devaluation since June 23rd, by expenditure decile


Source: As listed below Table 1. Note: See methodology section below for assumptions. Note: The ‘all’ figure here uses average LCFS budget shares as weights (see footnote to Table 1).

Looking at the effects by region, or by whether or not the household has children, or by whether the household includes a pensioner, all tell a similar story: these groups all see a similar rise in costs as a result of the devaluation. That said, there is a wide distribution of effects according to household spending patterns. The 5% of households most exposed to exchange rate changes are projected to experience a hit of at least 4.1%, while the 5% least exposed see their prices rise by less than 1.5%.

Of course how much individual households are affected by sterling's depreciation does not just depend on the goods they buy but also the degree to which their real incomes adjust as inflation rises. Some households are better protected than others, for example those whose incomes are indexed to inflation (or better), such as those receiving an occupational pension or state pension. Normally working age benefit recipients would also be at least partly protected as benefits usually rise in line with prices, but, as we have discussed before, their benefits have been largely frozen in cash terms, meaning that their income from this source is fully exposed to future inflation. Those in work will, unless they are able to negotiate a bigger pay rise, find that their earnings will stretch less far than they otherwise would have done.

Notes on methodology for estimating price impacts from exchange rate changes

Estimates for price impacts of each item are calculated on the following assumptions:

  • Non-energy imports increase by 60% of the fall in sterling.
  • Energy imports increase by 100% of the fall in sterling.
  • Increases in import prices are fully reflected in consumer prices.
  • UK package holidays have a 0% import intensity and overseas package holidays have a 100% import intensity
  • The relationship between indirect import content by Classification of Products by Activity (CPA) component and Classification of Individual Consumption according to Purpose (COICOP) group is the same as the relationship between household consumption by CPA component and COICOP group.
  • The Office for National Statistics provides the data on import intensity ratios at the two-digit COICOP level. These give the imports valued at pre-tax prices divided by the final gross value of goods sold in each category. To account for the fact that increasing import costs will lead to an increase in the value of taxes such as VAT which are charged on each good, we scale import intensities up according to the average rates of proportional taxes charged within each COICOP group. To calculate average VAT rates at the level of two-digit COICOP codes, we first sort goods at the three-digit level into VAT categories (standard rate, zero rate etc.), then take an expenditure weighted average of the applicable VAT rates within each of the two-digit spending groups.

 As a result of the level of detail in the data we use, one factor that we cannot account for here is whether richer or poorer households tend to buy more imported goods within a particular category (such as “food”). This is because we use the average import intensity for each category, and apply the associated price rise to all households. But if, for example, households further up the expenditure distribution tend to buy more Italian shoes or Wagu beef, they might have a higher than average import intensity in the clothes and food categories, and consequently would be more affected by the devaluation than shown in Figure 1. Conversely, if they tend to buy more British made suits or cheese, they will be less affected.

]]> Tue, 15 Nov 2016 00:00:00 +0000
<![CDATA[A tighter benefit cap]]> Some of the numbers in the Table on the percentage of affected households in each region after the reform were corrected on the 14th November. The original table is at the bottom of the observation.

A lower cap on the total amount of benefits that households can receive comes into force tomorrow, affecting four times as many households as the previous benefit cap. Like the previous cap it will apply to out-of-work households of working age (with some exemptions, mainly due to disability). The cap will now be £23,000 a year in London and £20,000 elsewhere (there are lower caps for single adults without children set at £15,410 in London and £13,400 elsewhere). This compares to £26,000 nationwide under the previous cap, which has been in place since 2013. In this observation we look at the implications of a lower cap for government spending, the impact on the households affected, and how they might respond.

The previous cap

In 2015–16, the £26,000 cap directly reduced benefit spending by £65 million. 20,000 households are currently affected (as of this August), while 79,000 have been capped at some point since its introduction in Summer 2013. This cap almost exclusively affects families with large numbers of children or very high rents (who are receiving lots of housing benefit), or both: there is almost no other way to be getting so much in benefits. More than half have at least four children. Two-fifths live in London – where rents and Housing Benefit levels are high. Perhaps more surprisingly, half of those households currently affected rent from their local council or a housing association.

The cap does not apply to claimants of Working Tax Credit, providing a strong financial incentive for potentially affected families to do enough hours of paid work to qualify for it: 16 hours per week for lone parents and 24 for couples with children.

The direct effects of lowering the cap

The government’s impact assessment states that ignoring potential behavioural responses, the reduction in the cap will increase the number of families affected to 88,000, and will deliver an additional £100m a year of savings to the exchequer in the long run (with a slightly larger annual saving in the short run). This is less than 1% of the £12 billion of cuts to annual benefit spending planned by the current government during this parliament.

However, for those families who are affected, the impact can be large. Households who are already capped will lose a further £3,000 per year (in London) or £6,000 per year (elsewhere). The government expects those households newly affected by the cap to lose an average of £2,000 a year. In addition, unlike many of the cuts to benefits being implemented during this parliament, the lowering of the cap will result in affected families seeing cash drops in benefit income between one housing benefit payment and the next. This will feel different to real benefit levels falling as cash payments fail to keep pace with inflation, and changes that are being phased in by only applying to new claimants.

The reduction in the cap will also change the profile of the households affected. Most significantly, in large part because the cap will now be lower in the rest of the country than in London, those affected will be more evenly distributed geographically than was previously the case, as shown in Table 1. Under the old cap, 42% of affected households were in London. That figure is expected to fall to 22% under the new cap. The other side of that coin is that the policy will become significantly more important in many regions outside of London: in the North East for example, the number affected will rise from 600 to 4,000 households.

Table 1. Number of households affected by the benefit cap in each region and nation of Great Britain









North East





North West





Yorkshire and the Humber





West Midlands





East Midlands















South East





South West




















Source: Pre-reform numbers – authors’ calculations based on August 2016 data from DWP Stat-Xplore. Post-reform numbers – Impact Assessment for the benefit cap

The reduction in the cap also means that its impact will be somewhat less concentrated on households with lots of children. Smaller households will no longer need to have extremely high housing benefit (HB) entitlements to be affected. Take, for example, an out-of-work couple with 2 young children renting in the private sector, where HB awards are themselves subject to caps which vary by local area. It was almost impossible for the previous benefit cap to affect them unless they lived in London, as the total of jobseeker’s allowance, tax credits, child benefit and the applicable HB cap in their area was less than the overall benefit cap everywhere outside of London. The same household could (if their rent is high enough) be affected by the new cap in more than half of local areas across England. For a lone parent with 2 young children the overall cap can now bind in a number of local areas outside of London, including Reading, Basingstoke and Bath – again it was almost impossible to be affected outside of London under the old cap.

Those examples illustrate a broader point. It is possible for the benefit cap to quickly affect many more out-of-work families in an area, once its level falls below the sum of the HB cap in that area for the family type in question and the other (nationally-set) benefit entitlements. Given the existence of these ‘tipping points’, further changes to the level of the cap could again have big effects not just on the number of households capped, but also on the types of households that are capped, in terms of number of children, geography and so on. This highlights one consequence of the approach of simply layering an overall cap on top of the benefits system, rather than addressing the underlying benefit rates (those for HB and child support) which cause the perceived problem: there is a risk of arbitrariness in its effects. It would be sensible for the government to set out a clear vision of which families it thinks receive excessive amounts of benefits and why.

How might people respond?

As discussed in a previous observation, we have fairly robust evidence that about 5% of those affected by the previous cap responded to that cap by moving into work. An even smaller fraction – and only those who lost particularly large amounts of income – moved house in response.

Hence it would be reasonable to expect the further lowering of the cap to result in some increases in employment as people respond to the strengthened financial incentive to be in paid work, and in some of those affected moving to a cheaper home. Another possible response is for households to claim disability benefits in order to become exempt from the cap. Of the 79,000 households who have been capped since its introduction in Summer 2013, 12,000 became exempt due to a disability benefit claim (we do not know how many of those 12,000 would have started a disability claim if the cap did not exist, so this does not reveal the effect of the policy).

However, all this suggests that the majority of those affected will not respond by moving into work, moving house or claiming a disability benefit. For that majority it is an open question how they will adjust to the loss of income. One mitigating factor that is likely to be significant is Discretionary Housing Payments (DHPs): money paid at the discretion of local authorities to help tenants deemed to be struggling to pay their rent. In 2015–16, £25 million of DHPs were allocated specifically to help tenants affected by the benefit cap (offsetting almost 40% of the direct £65 million saving from the cap). About 40% of those affected by the benefits cap so far have successfully applied for DHPs. These DHPs look likely to play a key role in mitigating the impact of the cuts on some of the families affected - whilst rendering the net fiscal savings from the cap all the more trivial.


Original Table 1.






North East





North West





Yorkshire and the Humber





West Midlands





East Midlands















South East





South West




















]]> Sun, 06 Nov 2016 00:00:00 +0000
<![CDATA[Breakfast clubs work their magic in disadvantaged English schools]]> Children who come to school hungry are less attentive, more disruptive and less likely to understand and remember the day’s lessons. UK policymakers are trying to address these problems by implementing school nutrition programmes, including new school food standards in England, a universal breakfast programme in Wales and a universal entitlement to free school lunches for children aged 4–7 in England.

New research by IFS researchers in collaboration with the National Children’s Bureau finds that offering relatively disadvantaged primary schools in England support to establish a universal, free, before-school breakfast club can improve pupils’ academic attainment.

In this study, funded by the Education Endowment Foundation, the charity Magic Breakfast offered support to 53 schools to establish breakfast clubs. The package of support lasted for one year (the 2014/15 academic year) and included as much food as required (free of cost), a £300 grant to each school to offset start-up costs such as buying a freezer, and advice and guidance from a dedicated ‘School Change Leader’.

To establish the effect of breakfast clubs on academic achievement, we compared the attainment of children aged 6/7 and 10/11 in the schools that were randomly chosen to receive this support with that of children in a group of 53 similar schools that did not receive the support that year (the ‘control’ group). Although the breakfast clubs were available to children of all ages in each school that received support, we focused on pupils in Years 2 and 6 because the assessments in these years are comparable across all schools in England.

The effect of breakfast clubs on attainment

Year 2 children (aged 6/7) whose schools were offered support to open a breakfast club made the equivalent of two months’ additional progress in reading, writing and maths over the course of a year compared with students in the control group of schools. Year 6 children (aged 10/11) had similar gains in English, though the effects on maths and science were smaller. The gains in attainment for younger children are a similar size to those found in previous research, which led to the expansion of free school meals to all infant pupils in England. This means that there is consistent evidence that school nutrition programmes can improve academic attainment.

How does breakfast club provision affect attainment?

To understand why setting up a breakfast club led to higher pupil attainment, we used surveys and administrative data to analyse how the breakfast club support affected pupil hunger, absences from school, late arrivals to school, teachers’ perceptions of student behaviour and concentration, and pupil health (as measured by Body Mass Index). We found that:

  • Gains are likely to be the result of the content or context of the school breakfasts, rather than of increasing overall breakfast consumption. Offering schools support to hold breakfast clubs markedly increased the number of students eating breakfast at school. Compared with students in the control group, more than three times as many students in schools that received the intervention ate breakfast at school (22% versus 7% in control schools). However, the impact of the breakfast clubs on the number of students eating breakfast at all was modest (91% in breakfast club schools versus 89% in control schools). This means that a large number of pupils switched from eating breakfast at home to eating breakfast at school. This suggests that the breakfast clubs’ positive impact on attainment came from the content and context of school breakfasts, such as eating more nutritious food or building stronger relationships with other pupils and staff in a relaxed environment.
  • Pupil absences declined as a result of breakfast club provision, falling by almost one half-day per year. The effect was particularly strong for authorised absences, which are primarily due to ill health. This suggests that the breakfast club might have improved pupil health, although we did not find strong evidence to support this when looking at the average Body Mass Index of students in Year 6. Late arrivals were not significantly affected by the offer of a before-school breakfast club.
  • Behaviour and concentration in the classroom improved substantially as a result of the breakfast club provision, suggesting that a better classroom learning environment is an important mechanism through which the intervention might improve attainment. The improvement in teachers’ assessments of their classroom learning environment was equivalent to moving a classroom from average ratings of behaviour and concentration to ratings in the top quarter of the schools in our sample.

Is breakfast club provision cost-effective?

These gains in pupil achievement were delivered at relatively low cost. Dividing the costs by all pupils in the school, the intervention cost just £11.86 per eligible pupil over the course of the academic year. It also required 2.6 hours of staff time per eligible pupil per year. It should be noted, however, that the breakfast club take-up rates were relatively low – the average school’s take-up rate was between 13% and 52%. An increase in take-up would lead to higher costs, but also potentially higher impact on attainment.

It is also worth noting that, while relatively disadvantaged students (those eligible for free school meals) were more likely to attend the breakfast clubs, the intervention was more effective at raising the attainment of pupils from less disadvantaged backgrounds (those not eligible for free school meals). This suggests that support for school breakfast clubs might not reduce socio-economic gaps in pupil attainment.


The 2013 School Food Plan recommended that schools with relatively more disadvantaged pupils should establish breakfast clubs to help address the problem of pupil hunger. Resulting from this, the Department for Education committed to provide funding for breakfast clubs in schools where more than 35% of pupils are eligible for free school meals and there is no existing breakfast club provision. The government’s Budget in March 2016 also included a pledge for a further £10 million a year to expand breakfast club provision from September 2017.

Our results indicate that additional funding of this kind can boost attainment, improve the classroom learning environment and reduce absences in disadvantaged schools – and all at relatively low cost per pupil. Universal breakfast club provision in disadvantaged schools should therefore be considered by schools allocating their pupil premium budget (and rightly by government) as a way to enhance pupils’ experience of school, and ultimately their educational attainment. As breakfast clubs are set to expand across the country, further research is needed to determine the most effective model of provision – for example, whether before school or as part of a soft start to the school day. Future academic research should be targeted at better understanding how health and education policies can interact to improve both children’s health and education outcomes. For example, is adequate nutrition a prerequisite for any educational improvements from traditional academic interventions?

More generally, this work shows that health- and nutrition-based policies can have real impacts on educational outcomes. In fact, providing a breakfast club in disadvantaged schools looks more cost-effective than both the universal provision of free school meals for infant pupils and many other interventions targeted directly on educational outcomes. The improvement in classroom behaviour and concentration in schools randomly selected for Magic Breakfast support is exceptional. In the policymaking world, the effect of Magic Breakfast provision is as close to magic as an intervention can get.


]]> Fri, 04 Nov 2016 00:00:00 +0000
<![CDATA[Falling sterling, rising prices and the benefits freeze]]> This morning the Office for National Statistics announced that inflation, as measured by the CPI, was 1.0% in the year to September. This is somewhat higher than the 0.6% the Office for Budget Responsibility forecast in the March Budget. But since then many forecasters – including the Bank of England – have revised up their forecasts for future inflation as the sharp drop in the value of the pound since the referendum is expected to push up prices. This observation focuses on one consequence that higher inflation would have: the fact that it would reduce the real incomes of working age families receiving benefits that the Government has frozen in cash terms through to March 2020.

Since the UK referendum vote to leave the EU the value of the pound has declined sharply. From almost $1.50 on the eve of the result, the pound has fallen to around $1.20; against the euro, the pound has fallen from around €1.30 to €1.10. This sharp decline will increase the price of imported goods. Just after the Budget, in its April World Economic Outlook, the International Monetary Fund (IMF) forecast CPI inflation to rise to 1.9% in 2017 before settling at 2.0% from 2018. By its October forecast the IMF had revised its inflation expectations to 2.5% in 2017 followed by 2.6% in 2018. Those forecasts were based on the exchange rate as it stood in mid-September; since then the pound has fallen a further 7% against the dollar and so, if anything, these numbers may in fact underestimate future price rises.

Normally many of those on the lowest incomes would be at least partially protected from the impact of higher prices by the rules that govern the annual uprating of benefits and tax credits. By default, benefit and tax credit rates are (with some exceptions, most notably the state pension) increased each April in line with the annual CPI inflation rate of the previous September – higher prices lead to higher benefit rates (albeit with a lag). However, in the July 2015 Budget the Government announced that, as part of its attempt to cut annual social security spending by £12 billion, most working-age benefit and tax credit rates would be frozen in cash terms until March 2020. This policy represented a significant takeaway from a large number of working age households. But it also represented a shifting of risk from the Government to benefit recipients. Previously, higher inflation was a risk to the public finances, increasing cash spending on benefits. Now the risk is borne by low-income households: unless policy changes higher inflation will reduce their real incomes.

Figure 1 shows how the size of the expected cut in generosity resulting from the four-year cash freeze has increased in the light of upwards revisions to forecast inflation. As of March 2016 the freeze represented a 4% cut in the value of those benefits affected relative to previous plans (given OBR inflation forecasts). As a result, 11.5 million families were expected to lose an average of £260 a year, saving the government £3.0 billion in 2019–20. Given the latest inflation forecasts from the IMF, the policy now represents a 6% cut to affected benefits. The same 11.5 million families are now expected to lose an average of £360 a year (£100 a year more than expected in March), saving the government £4.2 billion in 2019–20 (i.e. an additional £1.2 billion over what was expected back in March). Greater losses are found among families – typically those on lower incomes – who receive more in benefits: for example, ignoring the 3.2 million families who only receive child benefit, the average loss from higher inflation rises to £140 per year (with the other 8.3 million families affected now expected to lose an average of £470 a year).

Figure 1: Average expected income change per family affected as a result of the benefit freeze, March and October 2016 based forecasts, 2016–17 prices

Average expected income change per family affected as a result of the benefit freeze, March and October 2016 based forecasts, 2016–17 prices

Source: Authors’ calculations using the IFS tax and benefit model, TAXBEN, run on the Family Resources Survey, 2013–14.

Setting benefit rates in cash terms rather than relative to prices is becoming something of a habit in the UK: prior to the four-year freeze from April 2015, the rates of many working age benefits were capped at 1% in cash terms in April 2013 and April 2014. There are at least two reasons why cutting benefit rates through limiting or eliminating cash-terms increases is a habit that should be kicked, regardless of the desired generosity of the system. First, from the government’s perspective, this way of cutting benefits means the size of the cut and the saving to government depends on (unknown) future inflation. In the last few years, lower-than-expected inflation has led to smaller cuts in the generosity of the system than the government intended – this was highlighted by the OBR in Chapter 3 of their latest Welfare Trends report, published last week. Second, from the benefit recipient’s perspective, there is a reason that benefits are uprated in line with prices by default – since one purpose of benefits is to provide a minimum standard of living, their level should reflect the cost of purchasing the goods and services required to provide that minimum standard. While it is perfectly reasonable to argue – as the 2015 Conservative Party manifesto did – that the working age benefit system should be made less generous over this parliament, it is hard to see why the appropriate size of cut should be arbitrarily determined by the impact of movements in sterling on prices.

]]> Tue, 18 Oct 2016 00:00:00 +0000
<![CDATA[Business rates revaluation reveals growing gap between London and the North]]> The results of the latest business rates revaluation reveal a growing divergence in property prices between London and the rest of the country. Increases in the value of non-residential property in the capital are set to raise rates bills by 11%, on average, increasing the tax take by over £700 million. This will be offset by reductions in bills and revenues in most of the rest of England, and especially the North, as property values fall behind. Growing differences in property prices reflects broader evidence of a growing divergence in economic performance over the last few years. And it will contribute to the ongoing trend of the UK government becoming more and more dependent on revenue from London to fund services across the whole – which may pose difficulties if more revenue sources are devolved to the local level. This observation discusses this and other issues related to today’s revaluation figures.


Business rates – a tax paid by businesses and other occupiers of non-residential property and based on the value of that property – are a big deal: across the country as a whole they are forecast to raise £28.4 billion in 2016–17. This is almost as much as is forecast to be raised from council tax (£31.4 billion)– a tax levied on residential property, which in aggregate is worth around 4-5 times as much as non-residential property.

At the moment, business rates are assessed on 2008 property values. Today, the Valuation Office Agency has announced the updated 2015 property values that will be used to calculate business rates from April 2017. This is also a big deal: between 2008 and 2015 property values have changed very differently in different parts of the country, with very large increases in some areas (notably ‘gentrifying’ parts of London), and big decreases in others (such as ‘struggling’ town centres in much of the rest of the country). These relative changes will have a huge effect on business rates bills.

The business rates system works in a slightly odd way. The tax bill is calculated as the value of a property multiplied by a tax rate (called ‘the multiplier’).  The multiplier generally increases in line with inflation in the years between valuations such that revenues also rise roughly in line with inflation (although in 2014 and 2015 the government capped the increase at 2%).

Perhaps surprisingly, across England as a whole, the value of non-domestic property is estimated to have risen by around 11%, on average, between 2008 and 2015. If the same multiplier (i.e. tax rate) were kept, revaluation would therefore lead to overall business rates revenues rising.  To stop this happening, the multiplier would have to be reduced by 11% to ensure that average bills do not increase simply as a result of the revaluation.

The impact of revaluation on rates bills across England

Rateable values – and hence rates bills – are changing differently for different properties and different parts of England. The figure below show the average estimated change in rateable value and the implied change in business rates bills as a result of the revaluation for the different regions in England, as well as the so-called ‘central rating list’ (which includes things like transport, energy, water, and telecoms infrastructure that are often national in scope, and therefore are not allocated to any particular council or indeed regions).

Figure 1. Average change in rates bills due to revaluation, by region, before accounting for adjustment to multiplier to account for expected appeals

Average change in rates bills due to revaluation, by region, before accounting for adjustment to multiplier to account for expected appeals

Source: CLG Business Rates Revaluation Consultation and VOA rateable value data.

It shows bills falling by 10% or more, on average, in the northern regions of England, but increasing by 11% in London, on average. Such changes would mean that in today’s terms, more than £700 million more business rates revenues would eventually be raised in London and around £1.2 billion less in the other regions of England (with an increase in revenues from the central list making up the missing £500 million or so). 

The fact that London’s rates bills will be rising reflects its non-residential property market outperforming that in the rest of England – presumably because London is an increasingly relatively more attractive place to set up shop (or office). The same pattern of London economically outperforming the rest of the country, and especially the north, can be seen in many other indicators. Gross value added increased around three times as fast in real terms in London as in the rest of England between 2008 and 2014, for instance. The numbers of people in employment has grown by 15% in London since 2008, compared to 4% in the rest of England. And of course, residential property prices have risen far beyond their pre-Credit Crunch peak in London, but still lag in much of the North.

This trend means the UK government is becoming more and more dependent on revenues – from many other taxes like income tax, as well as business rates – from London to fund services across the country as a whole. At the same time there is growing pressure for devolution of more of London’s revenues to the Greater London Authority – a difficult square to circle if these trends continue.

The impact on individual businesses

The impact of revaluation on individual properties’ business rates bills can be much larger than the impact on average regional bills. For instance, 242,000 small properties – those with a value less than £28,000 in London and £20,000 outside London – are estimated to see their rateable value rise to such an extent as to imply increases in their bills of at least 24%. On the other hand, 141,000 would be due cuts of more than 20%.

To ease the pain of business rates increases for those seeing the biggest increases, it has become customary to offer ‘transitional protection’ – funded by slowing the ‘gain’ from those benefitting from cuts to their business rates.  The government is currently consulting on the details of the scheme this time round but its preferred scheme would apply to almost half of properties in 2017–18 and almost 5% even after five years in 2021–22: a long transition. There is clearly a trade-off between offering such protection and ensuring business rates respond fully to the changing reality of local property markets and economies. If there are worries about big changes in bills at revaluation, a better solution would probably be to revalue more frequently – e.g. every 2 or 3 years –, because the changes in relative values should be smaller over these shorter time periods.

The way the transitional protection system is set up also, in effect, redistributes from large properties (where increases in bills are capped at relatively high levels, and cuts at relatively low levels) to small properties (where the opposite is true). For instance, the government estimates only 8,800 large properties will be benefitting from transitional protection in 2017–18; but 9,700 will still be paying for transitional protection even in 2021–22. 

Will I pay more next year?  

The rates bills people will actually pay from next April will be affected by two other factors though. First, if history is a guide, many ratepayers will successfully appeal against their new values and get cuts in their rateable values and bills. To recoup these expected losses, the government will make an upwards adjustment to the multiplier, pushing up all rates bills. Second, as is usually the case, the multiplier is set to increase in line with September 2016 RPI inflation, forecast to be 1.7%.

After accounting for these factors as well, the government estimates the standard multiplier will be 0.48 next year, around 3% lower than the current 0.497.  This means that properties that have seen their rateable value increase by around 3% or less can expect a cut in their rates bill next year, while those whose value has increased by more than 3% can expect an increase in their bill next year.   

The impact on councils

Finally, it’s worth considering the impact on those who collect business rates – councils. Initially, the revenues of individual councils won’t change as the result of the revaluation (the government will redistribute funding between different councils to ensure no council wins or loses overnight). But, there may still be winners and losers in subsequent years. First, because councils in England notionally keep 50% of any change in their business rates revenues, any subsequent growth (or decline) in the business rates tax base will be worth more in those areas where rateable values have increased by more than average – and hence rates bills will increase –, and vice versa. Second, councils will have to bear their share of any successful appeals against the new values – and while estimates of the impact of such appeals can be made when the new valuations are introduced, such estimates are unlikely to be completely accurate.

The revaluation proceeds even bigger changes to the way business rates fit into the local government finance system. April 2017 will also see Liverpool and Manchester become pilots for a 100% business rates retention system that is set to roll out across England by 2019–20. This will have a big impact on the kinds of financial risks and incentives councils face. The IFS is launching a major research programme to look at these and other issues related to local government and devolution. Our first report, setting out the context and key issues, will be published on 26th October.


This analysis is part of a project funded by the Local Government Finance and Devolution Consortium. Members of the consortium include Capita, CIPFA, PwC, the Municipal Journal and the Society of County Treasurers.


Notes on sources

Information on rateable values is available from the Valuation Office Agency (see for regional summaries). Information on the implied changes in bills and transitional protection is available from the Department for Communities and Local Government (

Statistics on GVA ( and employment ( available from the Office for National Statistics. House Prices are available from the Land Registry ( 


]]> Fri, 30 Sep 2016 00:00:00 +0000
<![CDATA[Can grammar schools improve social mobility?]]> Today, the Secretary of State for Education is due to outline proposals that would allow an expansion of grammar schools across England. This could represent a significant shift in the education system in England. As ever there would be costs and benefits to such a change. It does appear that those who attend grammar schools do, on average, somewhat better than similar children in the comprehensive system. On the other hand, those in selective areas who don’t get into grammar schools do worse than they would in a comprehensive system. The real question for education is whether we can have the benefits without the costs. Do London schools point the way forward?

Entrants to current grammar schools are four times as likely to have been educated outside of the state system than to be entitled to free school meals despite the fact that across the population at least six times as many 11-12 year olds are entitled to free school meals than were previously educated outside the state system.

There is robust evidence that attending a grammar school is good for the attainment and later earnings of those who get in. But there is equally good evidence that those in selective areas who don’t pass the eleven plus do worse than they would have done in a comprehensive system.

There are benefits from a selective system for those who make it into selective schools. Expanding grammar schools may thus be a way of improving the educational achievement of the brightest pupils and there is clear evidence that this is an area where England lags behind other countries. However, those who don’t get into grammar schools do worse than in a comprehensive system. Is there a way of getting the benefits without the costs?

  • It seems likely that the only way of ensuring that a socially representative group of children attend grammar schools would be through a quota system. This would have obvious disadvantages.
  • To ensure that those not getting into selective schools do not suffer as a result requires us to understand more about why their outcomes are currently so poor and address those issues – whether it be lower quality teaching, fewer resources, negative peer group effects, or unduly low expectations.
  • A more productive route might be to look at those areas, like London, where overall standards and results have improved dramatically in recent years. Around half of pupils eligible for free school meals in inner London achieve 5 or more GCSEs at A*-C, double the proportion outside London. Furthermore, inner London has been particularly effective for high levels of attainment, with around 15% of pupils eligible for free school meals achieving 8 or more GCSEs at grade B or above in inner London, compared with 6% outside of London.
  • This high level of school performance has been put down to a variety of factors, including improved past primary school performance, greater numbers of high-achieving ethnic minorities and improved practices within and across schools (e.g. greater collaboration, better leadership and extensive use of data).

Grammar schools therefore seem to offer an opportunity to improve and stretch the brightest pupils, but seem likely to come at the cost of increasing inequality. Inner London, by contrast, has been able to improve results amongst the brightest pupils and reduce inequality. This suggests that London schools probably offer more lessons on ways to improve social mobility than do grammar schools.

]]> Mon, 12 Sep 2016 00:00:00 +0000
<![CDATA[Fall in oil price pushes up Scottish deficit to 9.5% of national income in 2015-16]]> Today the Scottish Government released the latest version of Government Expenditure and Revenue Scotland (GERS) covering 2015–16. In this observation we discuss what we can learn about Scotland’s fiscal position from these figures. The main finding is that further declines in revenues and output from the North Sea oil and gas sector pushed up Scotland’s budget deficit a little – at the same time the deficit continued to shrink in the UK as a whole.

Today’s estimates

Table 1 shows figures for revenues, spending, and the net fiscal balance (i.e. the budget deficit or surplus) for Scotland for each year between 2011–12 and 2015–16, measured as a percentage of national income. It also shows the net fiscal balance for the UK as a whole, and the ‘fiscal gap’ between Scotland and the UK. The top panel shows estimates from the most recent GERS publication. The bottom panel shows estimates from the previous edition of GERS and projections we made for 2015–16 earlier this year.

Table 1: Fiscal aggregates, UK and Scotland, 2011–12 to 2015–16

Looking at the most recent estimates from the top panel, the first thing to note is that the Scottish budget deficit widened from 9.1% of national income in 2014–15 to 9.5% in 2015–16. This is in contrast to the picture for the UK as a whole where the deficit fell from 5.0% to 4.0%. The fiscal gap – i.e. how much larger Scotland’s deficit is than the deficit of the UK as a whole – therefore increased from 4.1% of national income (£6.5 billion) to 5.5% of national income (£8.6 billion).

What explains these differences?

It is not what happened on land. ‘Onshore revenues’ (i.e. revenues other than from taxes on North Sea oil and gas) grew at a very similar rate in cash terms in Scotland (3.7%) as in the UK as a whole (3.8%). Similarly, public expenditure rose at a very similar rate (1.0% as opposed to 0.9%). The Scottish ‘onshore deficit’ therefore actually shrank by a similar magnitude (1 percentage point of national income) to that of the UK as a whole.

The divergence in trends is instead explained by what happened out in the North Sea. The fall in the oil price depressed the total value of oil and gas production in the North Sea, which had two effects:

  • First, there was a further fall in tax revenues from oil and gas production: estimated as just £60 million in 2015–16 (0.0% of national income) compared with £1.8 billion in 2014–15 (1.1% of national income) and £9.6 billion back in 2011–12 (6.3% of national income). This fall in offshore tax revenues pushed up the cash-terms budget deficit.
  • Second, the lower value of oil and gas production put downward pressure on Scottish national income. Indeed, Scottish national income is estimated to have fallen 0.5% in cash terms in 2015–16, compared to an increase of 2.4% for the UK as a whole. Lower national income means that any given cash-terms budget deficit represents a larger proportion of national income.

Taken together these two effects of lower oil prices explain the growing gap between the relative size of Scotland’s budget deficit and that of the UK as a whole.

Comparing the most recent estimates with previous estimates and projections

Comparing the top (most recent GERS) and bottom (previous GERS and projection) panel of the Table we can see that the Scottish budget deficit was slightly lower than we had projected earlier this year: 9.5% of national income as opposed to 9.8%. The fiscal gap between Scotland the UK was also a little lower: 5.5% of national income (£8.6 billion in cash terms) as opposed to 6.0% (£9.4 billion).

These (small) differences reflect revisions to data and methodology in the most recent edition of GERS that also affect estimates for previous years. In particular, estimated public expenditure in Scotland has been revised down by several hundred million pounds in each of the previous four years (2011–12 to 2014–15). Key changes include stripping out expenditure on Crossrail in London (which was previously unidentifiable and therefore allocated to Scotland in proportion to population), and other changes to the “accounting adjustment”. Moves from budget estimates to outturns for Scottish local authority spending also led to a downward revision to spending in 2014–15.

Perhaps an even more important factor has been the upward revision to estimates of Scottish national income since the last edition of GERS was published. For instance, national income is now estimated to have been £156.8 billion in 2014–15 as opposed to £153.3 billion in the last edition of GERS. This means any given cash-terms deficit would be a lower share of (the higher) national income.    

This reminds us that the allocation of spending and revenues between Scotland and the rest of the UK is a somewhat imprecise science, and is subject to both data revision and methodological change over time. But these revisions do not detract from the fact that GERS is an important publication that provides a detailed and very valuable picture of Scotland’s underlying public finances. The bringing forward of GERS (the most recent edition was published with around a 5 month as opposed to a 11 month lag) would mean, all else equal, more potential for revisions in future years (a little less data is now available when the figures are first produced). However, the data are now much more timely, so this is probably a (small) price worth paying.

Looking to the future

Looking ahead to 2016–17 and beyond, what can we say? Low oil prices – which are forecasted to persist for some time – are likely to mean continuing weakness in North Sea revenues, although the recent fall in the value of the pound might help a little (as oil is priced in dollars, which are now worth more pounds). This means that Scotland’s implicit deficit is likely to remain substantially greater than that for the UK as a whole – as we projected back in March.

However, there is more uncertainty than usual about just what the path for Scotland’s public finances will look like in the years ahead, given the recent vote to leave the EU. The consensus among economists is that leaving the EU will lead to a weaker economy than would otherwise have been the case, depressing tax revenues and pushing up some areas of spending (like welfare spending). All else equal and without further policy action, this would likely increase budget deficits for both Scotland and the UK in the years ahead above those projected or forecast back in March. Whether this would widen or narrow the gap between Scotland and the UK as a whole would depend on whether Scotland were more or less affected by the impact of the EU vote than the rest of the UK.

Given this uncertainty – and given that it is not yet clear how the UK government will change its fiscal plans in response to any economic slowdown – any projections we made for future years would be subject to wider-than-normal margins of error. For this reason we will await more information on the potential state of the economy, public finances and future fiscal policy before updating our longer-term projections for the Scottish public finances.    

The authors would like to thank the ESRC who funded this obervation as part of the public finance observations series.

]]> Wed, 24 Aug 2016 00:00:00 +0000
<![CDATA[Who will ‘Help to Save’ help to save?]]> The 2016 Budget included a policy designed to encourage around 3.5 million individuals in low-income working families to save more, known as ‘Help to Save’. From April 2018, those in families claiming working tax credit and in working families in receipt of universal credit (except those with very low earnings) will be able to place up to £50 a month in a new savings vehicle. The government will add 50% to those contributions after two years, and then do the same for another two years’ contributions, meaning a maximum government top-up of £1,200 for those who build up the full £2,400 of contributions to an account. Both members of a couple are able to have an account, so a couple that placed £4,800 in an account over two years would receive a total match of £2,400. However, the government expects the policy to cost only £70m in 2020–21, implying an average government top-up of only £20 in that year per eligible individual.

The previous Prime Minister, David Cameron, explained that the rationale for the scheme was to help low-income families build up a ‘rainy day fund’ – a buffer stock of savings that enables them to deal better with unexpected changes in income or expenditure. But the Figure shows that, of those individuals who would be eligible for ‘Help to Save’ were it in place now:

  • 20% live in a household that already has savings (excluding pension and housing wealth) of over £2,000;
  • 44% live in a household that already saves at least £10 a month;
  • 30% live in a household that already reports being able to afford an unexpected expense of £750.

In fact, just over half (53%) of the eligible group meet at least one of these three criteria. For them, the introduction of ‘Help to Save’ is arguably unnecessary to achieve the stated aim of the policy (to build up an adequate ‘buffer stock’ of savings), and for many will represent an opportunity to receive a government subsidy for savings that already exist or that would have been put aside in any case.

One could restrict eligibility to renters, on the basis that owner-occupiers (who make up nearly half of the eligible group) are much more likely to already have a ‘rainy day fund’. Of eligible individuals in rented accommodation, 10% are in households with at least £2,000 of savings, 35% are in households already saving at least £10 a month, and 16% are in households that report they could afford an unexpected expense of £750. 60% of eligible renters meet none of those three criteria (compared to 31% of eligible owner-occupiers).

Another way of targeting the policy more precisely at those without a ‘rainy day fund’ would be to make a lack of existing savings a condition of eligibility, by introducing an asset test. The (recently closed) government consultation on Help to Save states that the government is considering restricting eligibility to those with less than £2,000 of savings. If one excludes those in households with more than £2,000 of savings in financial assets, the proportion of eligible individuals whose household already saves £10 a month falls to 36%, and the proportion in households who could already afford an unexpected expense of £750 falls to 19%. 60% meet neither criterion, compared to 10% of those eligible with more than £2000 of savings.

If one combines the two restrictions (making only renters with less than £2,000 of savings eligible) the proportion of eligible individuals in households who already save £10 a month falls to 30%, the proportion in households who could already afford an unexpected expense falls to 11%, and the share of individuals meeting neither criterion rises to two-thirds.

Figure. Options for the targeting of ‘Help to Save’

Note: Sample is eligible households as defined in the text, excluding households containing more than one ‘benefit unit’ and those for which the information on saving each month is missing.

Source: Authors’ calculations using 2013–14 Family Resources Survey and TAXBEN, the IFS’s tax and benefit microsimulation model.

There are potential downsides of introducing additional eligibility criteria such as these: as in so much of welfare policy, there is a trade-off between targeting the policy precisely and avoiding unwanted distortions to incentives. If those who already have some savings were excluded, that would create an incentive not to save now (or to spend down savings) in order to receive a subsidy on savings made later on. Of course, by targeting help on those receiving in-work benefits, the existing Help to Save policy will have incentive effects too: it will strengthen the incentive to claim those benefits (which would encourage some to do more paid work but would discourage others).

There is also a deeper and critical question about which groups are really ‘under-saving’. The key justification for giving a household extra money only if it places funds in a savings account, rather than giving it extra money regardless and letting the household decide what to do with it, is that we have reason to believe that the household is saving less than is ‘appropriate’ given its circumstances. It would be helpful for future research to shed more light on which groups are really under-saving in this sense. If a household has appropriately judged that, given its income, putting more money aside would mean forgoing too much in the way of spending now, then a subsidy that is available only if the household saves more is not the best way to help it. For example, the lower share of renters with a ‘rainy day fund’ might partly reflect the fact that they are not responsible for expenses associated with home maintenance and so have sensibly chosen to accumulate a lower stock of savings than owner-occupiers. Distinguishing better between households who have low saving for those kinds of reasons, and households who have genuinely saved ‘too little’ given their circumstances, is crucial for the design of policies such as these.

Note: Funding from the Joseph Rowntree Foundation is gratefully acknowledged.

]]> Mon, 15 Aug 2016 00:00:00 +0000
<![CDATA[Determining “what works” in the delivery of local services: getting citizens more involved]]> Finding new ways of engaging citizens in the delivery of local public services may become increasingly important in the years ahead. At a launch event today, we presented the initial findings from a small scale, but robustly evaluated scheme within the London borough of Lambeth. Through this scheme we have shown that carefully designed interventions can have positive effects, engaging local communities in improving the environments in which they live. We believe this provides a strong signal for what might be possible more generally and would encourage local authorities to work with us and other researchers to design and evaluate other interventions aimed to engaging local residents.

In collaboration with Lambeth Council , we trialled a scheme that asked residents to volunteer to become a ‘Street Champion’ and coordinate efforts to improve the cleanliness and attractiveness of their local environment, with the council offering providing advice and support along the way. 170 residential streets were split into five different groups. One group were the control group, where nothing changed. In a second group of streets, residents received an invitation to become a ‘Street Champion.’ In the remaining three groups, residents received different forms of incentives to test their impact on people’s willingness to get involved. The different incentives were individual rewards (e.g. free garden waste collection), incentives that would benefit the community as a whole (e.g. graffiti removal services) and rewards that emphasised individuals’ identity as Street Champions (e.g. hi-vis vests, polo shirts). We then collected data about the actual activities of Street Champions, the cleanliness of streets and residents’ overall perceptions of their local area.

There were four main findings from this evaluation.

First, citizens are keen to get involved in the delivery of local services, but they’re more likely to act as a complement rather than replacement for existing services. In streets where people received an invite and/or incentive to become a Street Champion, around 3 people per street expressed an interest in becoming a Street Champion. In around half of these cases actual activities followed (e.g. clean-up events, street meetings, Facebook groups). There was, however, no evidence of any impact on levels of litter or other aspects of street cleanliness. Where there was a significant impact was in terms of ‘beautification’ (e.g. installing planters). Streets in the ‘identity incentives’ group were 17 percentage points more likely to show evidence of beautification than the control streets. This is a big impact considering that only 11% of all streets in the experiment showed evidence of beautification. This suggests citizens are very keen to provide services and amenities that are complementary to existing service provision. The big unanswered question, however, is the extent to which citizens are willing to step in were there to be falls in service provision.

Second, (some types of) incentives can increase the number of people coming forward and their level activity. The identity incentives had the biggest impact on actual activity levels. They doubled the number of people willing to come forward and the average number of activities per street as compared with a simple invite. For example, providing identity incentives increased the number of clean-up events by 15% percentage points, a very substantial change considering that only 6% of streets in the experiment did one of these. The community-wide incentives also increased the number of people willing to come forward and the activity relative to a simple invite, just not by as much as the identity incentives. The individual level rewards increased expressions of interest, but had a negligible impact on activity levels compared with a simple invitation.

Third, getting citizens involved in the delivery of local services can have valuable spill-over effects. We found good evidence that beyond immediate effects, these schemes increased satisfaction and levels of social interaction. In the identity incentives streets, residents were more likely to express satisfaction with their local area. In the community-wide incentives group residents reported increased levels of social interaction on their street. These types of impacts are very valuable to observe. Academic research has recognised the value of social capital in communities, but there have been few examples of methods that can shift it. Getting citizens more involved in the delivery of local services may be one way of stimulating social interactions and connectedness.

Finally, we think we have shown the value of partnerships between academic researchers and local councils to determine “what works.” We have worked closely with Lambeth Council over the last two years to develop the design of the experiment and to analyse the policy implications of the data coming out of it. The scheme eventually rolled-out was very much informed by the results of the experiment. We as academic researchers have also gained a much greater understanding of how local government works and the challenges facing policymakers. We hope that more councils and academics will follow this example.

Our launch event and some of the foundation work required for the pilot with Lambeth has been funded by the Economic and Social Research Council under the IFS Impact Acceleration Account. We are grateful for their support.

]]> Mon, 11 Jul 2016 00:00:00 +0000
<![CDATA[Strongest employment growth in twenty five years in 2014-15 pushed average incomes above previous peak]]> Median household income in the UK rose by 3% in 2014–15 after adjusting for inflation. This was the fastest rise in average incomes since the early 2000s, finally taking median income 1% above its previous peak, and was accompanied by income growth right across the distribution. These are the most striking findings from today’s release of the latest official statistics on the distribution of household income by the Department for Work and Pensions (DWP). This short observation sets out some of the reasons for the strong income growth, and highlights some of the other things we learn from the new statistics. On 19th July, IFS researchers will launch a detailed report, funded by the Joseph Rowntree Foundation, which will provide a much more comprehensive analysis using the data underlying the statistics.

Strong income growth was driven by a recovering labour market

Median income – the net household income of the person in the middle of the distribution - fell following the recession by a total of 4% between 2009–10 and 2011–12. Since then, median income has grown by 5%, with weak growth in 2012–13 and 2013–14, and robust growth of 3% in 2014–15. This leaves median income 1% above its 2009–10 peak.

The robust average income growth in 2014–15 was driven by a strong recovery in the labour market, and in particular by rising numbers of people in work. According to the Labour Force Survey, the employment rate of working-age people (aged 16-64) rose by 1.3ppt to reach 73.2% in 2014–15, the highest employment rate of any financial year since data began in 1971. This was the largest increase in the employment rate for over 25 years; the last time employment grew faster was in 1988–89. It was driven by the private sector, and was disproportionately growth in full-time work: according to the ONS, full time work accounted for 73% of employment in 2013-14, but for 81% of the employment growth between 2013–14 and 2014–15. Low inflation also meant that the average real earnings of employees grew slightly in 2014–15 - by between 0.4% and 0.8% depending on the measure used.

Income growth for rich and poor leaves inequality little changed

Incomes grew right across the income distribution in 2014–15: by 1% at the 10th percentile, 3% at the median (middle) and 3% at the 90th percentile. The differences between these growth rates are not statistically significant. Income inequality has been broadly stable since the recovery from recession began in 2011–12, with the widely-cited measure of inequality known as the Gini coefficient unchanged (at 0.34) since then. This also means that income inequality remains lower than before the recession: the large falls in earnings between 2009–10 and 2011–12 tended to hit higher-income households most, because they get a bigger share of their total income from earnings than lower-income households. Looking further back inequality remains no higher than in 1990 – though far higher than before the rapid rise in inequality during the 1980s.

 Absolute income poverty falls significantly in 2014–15

The government’s statistics include a measure of ‘absolute low income’ poverty, which counts the number of people in a household with an income below a poverty line which is fixed in real terms over time. Hence, falling absolute poverty means rising real incomes for low-income households.

Measuring incomes after deducting housing costs (as we believe is preferable when measuring poverty), the statistics reveal a statistically significant decrease in absolute poverty in the UK in 2014–15, from 22% to 20%. There were also falls in absolute poverty for all major demographic groups: children, working-age adults, and pensioners. The bigger picture, though, is that overall absolute poverty is only slightly below its level a decade ago in 2004–05. That is an unusually long period over which to see such little income growth for low-income households, though it does follow a period in the late 1990s and early 2000s when incomes grew particularly rapidly for low-income households. There are differences between groups though. In particular, absolute poverty rates for pensioners have continued to fall significantly over the past decade - and we have shown previously that they now have the lowest poverty rate of all the major demographic groups.

The other headline measure of income poverty is known as “relative poverty”: an individual is said to be in relative poverty if their household income is less than 60% of contemporaneous median income. This means that relative poverty rises (or falls) if the gap between low-income and middle-income households rises (or falls), regardless of changes in absolute income levels. Similar growth in income across the income distribution means that relative poverty was broadly unchanged in 2014–15. As with changes in absolute poverty, there have been only quite small fluctuations in relative poverty rates over the last 10 years, with relative poverty in 2014–15 the same as its level in 2004–05.

Note: On 19 July, IFS researchers will present the key findings from the latest in the series of flagship IFS annual reports on living standards, poverty and inequality in the UK. Funded by the Joseph Rowntree Foundation, the report will analyse the official data on the distribution of household income in the UK, up to and including the latest year of data for 2014-15. Register here.



]]> Tue, 28 Jun 2016 00:00:00 +0000
<![CDATA[Leaving the EU would almost certainly damage our economic prospects]]> Jagjit Chadha is director of the National Institute of Economic and Social Research. Paul Johnson is director of the Institute for Fiscal Studies. John Van Reenen is director of the Centre for Economic Performance, London School of Economics.

The economic consequences of leaving the EU have naturally been a central focus of the referendum campaign. As June 23 draws near we bring together the conclusions from our research on the likely consequences, and reflect on some of the claims made.

First and foremost a vote to leave the EU would almost certainly make us financially worse off compared with staying in the EU, quite possibly by a substantial amount. Analysis by the CEP and NIESR suggests that were we to leave the EU the economy would be between 1% and 3% smaller by 2020 and between 2% and 8% smaller by 2030 than if we stay in. A 1% drop in GDP is a fall of £19 billion, equivalent to £720 for each household currently in the UK.

The reasons to expect lower national income if the UK leaves the EU are well-established – prolonged uncertainty, reduced access to the single market, and reduced investment from overseas. Each of these would be highly likely and the overwhelming weight of evidence is that each would be damaging for the living standards of UK households. Consequently, leaving the EU would, relative to staying in, be likely to result in:

  • Lower real wages;
  • A lower value of the pound – and hence higher prices for goods and services;
  • Higher borrowing, lower public spending or higher taxes;
  • In the short run, higher unemployment.

It is not just the research undertaken by our institutes that predicts these effects, but the work of almost all those who have looked seriously at this issue. In our lifetimes we have never seen such a degree of unanimity among economists on a major policy issue. The precise effect, in terms of a numerical percentage, is of course uncertain. But that we would be financially worse off outside the EU than in it is almost certainly true.

Considering some of the biggest economic claims made by each side in this campaign about what would happen if we were to vote to leave the EU:

"There would be £10 billion more to spend on public services and tax cuts."

Almost certainly untrue. While we would get to keep our current net £8 billion contribution to the EU budget, overall the public finances would almost certainly be weakened by leaving the EU as a result of a negative impact on the economy. Hence in the long run taxes would have to rise, spending fall and/or public borrowing would have to rise;

"Households would be £4,300 a year worse off by 2030."

Uncertain – the effect might be bigger or smaller and will fall unevenly across households. And they would not be worse off than they are now. But households would in all likelihood be, on average, significantly worse off if we left the EU than if we stayed;

"The UK would be able to trade with other EU nations on equally good terms to those we currently have."

Almost certainly untrue. Membership of the single market on something like current terms might be available if we were to continue to make budget contributions and accept free movement of labour. It would not be available otherwise. Trade with EU countries would continue, but it would become more difficult and costly, there would be less of it, and we would be worse off as a result;

"There would be immediate big tax increases."

Unlikely. The public finances, even accounting for the return of the net contribution to the EU, would be badly affected as the economy would likely be smaller. That would require tax rises or spending cuts at some point. But our judgement is that the government would allow borrowing to rise in the short term rather than implement further tax raises or spending cuts immediately.

Clearly non-economic arguments matter as well. Leaving the EU would allow us more freedoms over some aspects of sovereignty such as law-making and possibly greater control over immigration. There is a trade-off to be made. Voters need to make a judgment. That judgment should be informed by the fact that our research, and that of every reputable economic research organisation, suggests that leaving the EU would almost certainly harm our economic prospects.

Given the importance of this referendum we feel we need to reiterate that the CEP, the IFS and NIESR are all independent, impartial organisations which exist to conduct and promote high quality economic research. What we publish and say is not, and never has been, influenced by our funding. None of our funding is dependent on  being either pro-EU or pro-government – as our frequent vocal criticisms of each in the past should make clear. None of the three current directors of the organisations ever supported joining the Euro, nor did the organisations themselves.

Jagjit Chadha, director of NIESR said: “Our research shows that in the short run a vote to leave the EU would increase uncertainty and, at best, reduce growth. In the long run more costly trade and less foreign investment would hold growth back.”

John van Reenen, director of the CEP said: “Research at both CEP and NIESR shows that the UK economy would do worse outside the EU than in it. Trade as an engine of growth would stall. Indeed virtually all the economic work of which we are aware tells the same story – there really is no serious doubt that leaving would be taking a big risk with the economy.”

Paul Johnson, director of IFS said: “Given that leaving the EU is likely to reduce growth the public finances would suffer. That would mean higher borrowing in the short term and higher taxes or lower spending in the long-term.”


Baker, J., O. Carreras, M. Ebell, I. Hurst, S. Kirby, J. Meaning, R. Piggott and J. Warren. (2016). The short-term economic impact of leaving the EU, National Institute Economic Review, No. 236, pp108-120.

Dhingra, S., G. Ottaviano, T. Sampson, J. Van Reenen. (2016). The consequences of Brexit for UK trade and living standards, Centre for Economic Performance, Brexit Analysis no. 2.

Ebell, M. and J. Warren. (2016). The long-term economic impact of leaving the EU, National Institute Economic Review, No. 236, pp. 121--38.

Emmerson, C., P. Johnson, I. Mitchell and D. Phillips (2016). Brexit and the UK's Public Finances, IFS Report 116.

]]> Mon, 20 Jun 2016 00:00:00 +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[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[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[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[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[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[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[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[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[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 th