50p tax – strolling across the summit of the Laffer curve?

Published on 27 January 2014

Ed Balls and Ed Milliband have cited recent HMRC statistics which show those paying the 50% income tax rate are estimated to have paid some £10 billion more in tax over the three years 2010-11 to 2012-13 than was projected to be the case back in 2012 when HMRC analysed how much the tax was raising. Is that an indication that the 50p rate was more successful in raising revenue than HMRC concluded in their analysis?

On Satuday 25th January, Ed Balls, the shadow chancellor announced that if elected, a Labour government would return the rate of income tax payable on incomes above £150,000 to 50%. What would the effect of this be? The contention of Ed Balls and Ed Miliband is that it would raise a meaningful sum of money to help reduce the budget deficit and make for a fairer tax system. The contention of their opponents is that it could result in an exodus of talent from the UK, a reduction in entrepreneurial drive, and an increase in tax avoidance and evasion, and may actually reduce the amount of tax paid.

Who is right is important not just for the lucky few who have incomes high enough that they would be directly affected. It matters for everyone because the Exchequer is, perhaps worryingly, reliant on this very small group of individuals for a very large fraction of revenue: the 1% of income tax payers with incomes in excess of £150,000 pay somewhere between 25 and 30% of all income tax. How these people would respond to a change in tax rates can therefore have big implications for overall tax revenues.

Perhaps the best evidence we have at present is that produced by HMRC, and signed off by the Office for Budget Responsibility, in 2012. This suggested that cutting the 50p rate to 45p could reduce revenues by about £3.5 billion in 2015–16 if there was no change in behaviour by affected individuals. However, once one allows for behavioural response, their central estimate was a cost of just £100 million – a very small amount of money. The best available estimate of what reversing the cut would raise is therefore about £100 million too.

However, it is important to bear in mind that there is substantial uncertainty around this central estimate. Calculating the revenue effects allowing for behavioural response requires one to estimate the “taxable income elasticity” – the extent to which taxable income changes when the tax rate changes. HMRC’s central estimate is that this elasticity was 0.45, which is broadly in line with estimates by IFS researchers based on the last time the top rate of income tax changed – in the 1980s – and with estimates from a number of other countries. If instead the true elasticity was 0.35 (which is well within the range of uncertainty), reducing the top rate of tax from 50p to 45p will have cost the exchequer about £700 million, whilst if the true elasticity was 0.55 (again, within the range of uncertainty), it will have actually raised about £600 million. In other words cutting the top rate of tax may have cost the government a bit more than it thought or actually raised a bit. This evidence led the OBR chairman Robert Chote to conclude that, whatever the precise answer, we were “strolling across the summit of the Laffer curve”.

Has anything changed since then? Ed Balls and Ed Miliband have suggested that the most recent HMRC statistics show those paying 50% income tax have paid some £10 billion more over the three years 2010–11 to 2012–13 than was thought back in 2012 when HMRC did their analysis. The statistics in question are estimates and projections of tax liabilities based on the Survey of Personal Income, which is itself based on a sample of tax records. The versions of these tables from 2012 and 2013 do show a difference of around £10 billion in the total amount of tax paid by those paying the 50p rate in the years between 2010¬–11 and 2012–13. Is that an indication that the 50p rate was more successful in raising revenue than HMRC assumed in their analysis?

Looking carefully at the notes that accompany the 2012 tables cited by the Labour Party shows that the figures for 2010–11 to 2012–13 were projected tax payments based on how much tax was paid in 2009–10, before the 50p tax rate was introduced. Had the HMRC’s analysis of taxable income elasticity been based on these original, lower, projections for tax revenues then we might have good reason for questioning their analysis. In fact it was not. In their analysis, HMRC made use of the actual 2010–11 tax returns of people paying by self-assessment (who make up the vast majority of people who were paying the 50p rate of tax). This is pretty much the same data (bar a few late filers and those few 50p rate tax payers not required to fill in a self-assessment form) that then goes into producing the Survey of Personal Income data used in the most recent 2013 tables. These tables show 50p tax payers paid £34.5 billion in income tax in 2010–11 on an income of £86.5 billion (and HMRC’s analysis assumed their income was £87 billion). So in fact there is little additional evidence to suggest that a 50p rate would raise more than was estimated by HMRC back in 2012.

Of course, even if increasing the top rate of income tax raised little or nothing one might still consider it worthwhile if one had a very strong preference for reducing inequality. Some effect has of course already been felt since the recession hit through the combination of the 45p rate itself, a major reduction in pension tax reliefs, and the withdrawal of the personal income tax allowance for those on high incomes. Our analysis of the effect of policy changes since the recession suggests people with incomes over £100,000 have on average seen a bigger percentage hit to their incomes from tax and benefit policies than people in any other part of the income distribution. But the group with incomes over £150,000 a year remains extremely well off relative to the majority.

The uncertainty around HMRC’s estimates mean it is also possible that the 50p rate would be somewhat more effective at raising revenue than their initial analysis suggests. HMRC made their calculations at great speed on the basis of one year’s data that had only just become available. Indeed only around 95% of the data was available at the time they made the calculation. By now they have data for 2011–12 too, and soon they will have data for 2012–13 as well. Given this there is certainly a case for HMRC looking again. IFS researchers also now have access to much of the relevant data and we will also be looking at this issue over the course of the year.

But at the moment, the best evidence we have still suggests that raising the top rate of tax would raise little revenue and make, at best, a marginal contribution to reducing the budget deficit an incoming government would face after the next election.