From the TUC

Did the 50p tax rate really raise less than £1 billion in 2010/11?

22 Mar 2012, by Guest in Economics

The main justification used for the reduction in the top rate of income tax from 50% to 45% (effective from April 2013) in George Osborne’s Budget yesterday was an analysis by HMRC, published on the same day, which claimed that the top rate was raising very little money – less than £1 billion on an annual basis, rather than the £2.7 billion projected in the March 2010 Budget. In this blogpost I assess the evidence presented by HMRC in support of this claim. Based on the HMRC research as well as other recent evidence on the revenue impact of increases in top tax rates, I reach two conclusions:

(1)    the methods used by HMRC to assess the yield from the top rate are likely to produce an underestimate of the revenue-raising potential of the 50p rate in the tax system as it currently stands;

(2)     even if the HMRC estimate is correct, for the most part it reflects easily correctible anomalies in the tax system which allow a large proportion of high income individuals to avoid paying 50% tax relatively easily. Reforming the system to close these loopholes would ensure that the 50p rate raises substantially more revenue than it does now.

Why the HMRC research most likely underestimates the yield from the 50p tax rate

Based on the number of people paying tax with gross incomes above £150,000 in 2009/10 (the last tax year before the 50p rate was introduced), HMRC have estimated that the 50p rate would have raised about £7 billion if the behaviour of taxpayers was unchanged after the tax was introduced. However, it is implausible that there would be no behavioural effect whatsoever of the tax increase. While most research shows that the effect of the tax system on work effort (e.g. hours of work) is very small for the types of workers who are likely to be affected by the 50% rate, the rate does increase the incentive to undertake tax avoidance activity – that is, finding ways to avoid paying income tax at 50p by exploiting loopholes in the tax system. Some examples of possible avoidance activity are:

  • Converting income into capital gains (taxed at a maximum rate of 28% in the UK) ;
  • Allocating income to a spouse or partner;
  • Bringing forward a stream of income before the 50p rate is introduced, thus managing to pay tax at the old rate;
  • Entering into artificial avoidance schemes.

There may also be some impact in terms of high earners migrating out of the UK, although as Richard Murphy discusses in a recent report for the TUC, this impact is likely to be small.

In the event, HMRC data for 2009/10 (the year before the 50p rate was introduced) and 2010/11 (the first year of its existence) appear to show a very large increase in income tax revenue in 2009/10 and a very large drop in 2010/11. This suggests that tax avoidance through ‘forestalling’ (moving taxable income forward from 2010/11 to 2009/10 to pay tax at 40% rather than 50%) inflated tax receipts in 2009/10 and depressed them in 2010/11 (and possibly subsequent years as well). The HMRC report on the impact of the 50p rate shows that total income for individuals with incomes above £150,000 increased from £101.3bn to £115.7bn from 2008/09 to 2009/10 before falling to £87bn in 2010/11 (HMRC, Table 5.1).

Forestalling on this scale makes it pretty much impossible to estimate the ‘long-run’ yield from the 50% tax rate on the basis of the 2009/10 and 2010/11 data alone, because there is no real way of knowing how much of the drop in income in 2010/11 is due to forestalling and how much is due to other, longer run, avoidance factors depressing taxable income, or other behavioural impacts on work by top earners. HMRC’s approach uses a time series regression of data  from the last two decades to estimate the relationship between total net incomes for individuals with net incomes of above £150,000 per year, total net incomes for individuals with net incomes between £115,000 and £150,000 (used as a control group unaffected by the 50% rate) and average UK equity prices. However, the extent of the financial crisis and ensuing recession which unfolded from 2008 onwards in the UK – as well as the equity price bubble of the late 1990s – mean that the coefficients from this kind of regression are most unlikely to be stable over time. Hence it is not clear that HMRC can reliably forecast what the longer-run impact of the 50% tax rate on revenues would be using this approach.

The key parameter which is important for evaluating the effect of increases in top rate tax on the amount of revenue received by the Exchequer is the Taxable Income Elasticity (TIE). This parameter measures the extent to which the taxable income of top earners falls – due to tax avoidance, or other changes in behaviour by high income individuals – as the tax rate rises. The higher the TIE estimate, the lower the yield from an increase in the top rate of tax and the lower is the ‘revenue-maximising’ rate of top tax – the rate above which net revenue from a tax increase is actually negative.  The HMRC analysis produces an estimate for TIE of around 0.46, suggesting that the revenue-maximising top rate of income tax is 48%. This is the reason why the OBR’s costing for the reduction in the 50p top rate of tax to 45p is only £100 million – because with a TIE of 0.46, a 50% top rate is actually above the revenue-maximising rate. Reducing the rate from 50% to 48% would raise revenue if HMRC is right, while reducing the rate to 45% loses very little revenue compared with 50%.

But is the estimate of 0.46 correct? HMRC provides a review of some of the estimates of TIE – mainly based on US research, but also including a study by Brewer, Saez and Shephard (2008) for the Institute for Fiscal Studies’s Mirrlees review of the tax system based on UK data, which also produces an estimate of 0.46 based on the effects of the changes to top marginal rates in the UK between 1979 and 1988. However, as explained in the IFS’s Green Budget released earlier this year, the accuracy of the TIE estimate by Brewer et al relies crucially on the assumption that the share of income going to the richest 1% of people in the UK compared to the share of income going to the next richest 4% of people in the UK would have changed by the same amount between 1979 and 1988 (when the top tax rate was reduced from 83% to 40%) as it in fact did change over that same period. There is in fact no real reason to expect this to be the case. To the extent that the income share of the top 1% was tending to increase relative to the next richest 4% anyway, the estimate by Brewer et al overstates TIE for the UK.

Furthermore, a recent review of estimates of TIE by Saez, Slemrod and Giertz (2012) suggests that many of the earlier studies included in HMRC’s review of the literature on TIE produced estimates of TIE (for the United States) that were excessively high due to flaws in the data and the methodology used. Saez et al claim, on the basis of more recent evidence, that  “the best available estimates [of TIE] range from 0.12 to 0.40”,  although they also suggest that “the TIE is higher for high-income individuals who have more access to avoidance opportunities, especially deductible expenses”. Hence, a TIE of 0.46 for UK top earners, while not impossible, would be on the high side of recent estimates.

Furthermore, recent research by Piketty, Saez and Stantcheva (2011) suggests that a high TIE estimate  may actually be a reason to increase rather than decrease tax rates, for distributional reasons. Piketty et al suggest that there are three ways in which top earners can respond to changes in tax rates:

  1. The supply response, whereby lower tax rates stimulate economic activity among top earners (through greater work effort, for example);
  2. The tax avoidance response, as catalogued above;
  3. The bargaining response, reflecting the possibilities for top earners to increase their compensation packages where corporate governance arrangements are weak and pay does not reflect performance accurately (a situation which arguably describes the UK very well). The idea here is that the lower the top tax rate, the higher the incentive for highly paid employees to engage in bargaining behaviour to increase their net incomes.

Piketty et al use data from 18 OECD countries from the 1970s to the 2000s to estimate the relationship between top tax rates and taxable income for high earners and find an overall TIE of around 0.5 for top earners. However, they also find that the tax avoidance component of the elasticity is rather small – around 0.2. The elasticity for the bargaining response  – which shifts resources from people further down the income distribution to top earners, thus exacerbating inequality – accounts for 0.3 of the elasticity. What this means is that increasing the rate of top income tax may cut the yield, but it mainly does so because the pre-tax distribution of income is being equalised. In other words, if Piketty et al are correct then increasing the top rate would redistribute pre-tax income from the highest earners to people further down the income distribution as well as redistributing post-tax income.  In other words, reductions in the top rate of tax exacerbate inequality rather than delivering increased productivity and growth.

The HMRC estimate reflects easy opportunities for tax avoidance

We have seen that HMRC’s estimate of the taxable income elasticity is problematic for several reasons. But even if the HMRC estimate is (a) an accurate reflection of the TIE for top incomes and (b) reflects tax avoidance rather than bargaining behaviour among the high paid (which are both questionable assumptions), it is clear that much of this avoidance activity could be significantly reduced or eliminated by changing the tax system to eliminate obvious loopholes. Some good examples would be:

  • Increasing the rates of Capital Gains Tax to align them with income tax marginal rates, as suggested in the 2010 Liberal Democrat manifesto. This would eliminate an incentive to reduce tax bill by converting income into capital gains – a measure which is available to many top rate tax payers;
  • Limiting the scope for reallocating income to partners or spouses;
  • Limiting the scope for artificial tax avoidance behaviour.

Indeed, a limited set of anti-avoidance measures was announced by the Chancellor in his budget speech and the projected success of these measures this was part of the basis for his assertion that the new measures contained in the 2012 Budget would raise 5 times as much from the very rich as the 50p rate did. This suggests that the Chancellor agrees that effective anti-avoidance measures can increase revenue. But if this is the case, then closing the opportunities for avoidance would be a good way of increasing the revenue from the 50p rate without producing a massive giveaway to the top 1% of taxpayers.

In summary, HMRC’s estimate that the 50% rate of income tax raised less than a billion pounds in the 2010/11 tax year has to be viewed as a premature and uncertain piece of analysis. The department was set a near-impossible task by the Chancellor in trying to evaluate the yield from the 50% rate based on data that have been hugely distorted by taxpayers shifting their income forward from 2010/11 to 2009/10 to avoid paying tax at the higher rate in the first year of its operation. This short-run effect swamps any longer-run trends and makes it pretty much impossible to identify how much the 50% rate would have eventually raised had it been left in place. Furthermore, the justifications from previous literature used to support HMRC’s contention that the taxable income elasticity for UK top incomes is as high as 0.46 are highly questionable. Finally, the reduction in the top rate from 50% to 45% is likely to increase the inequality of pre-tax incomes as well as post-tax incomes if bargaining effects for top earners are important – exacerbating the problems faced by the “squeezed middle”.

5 Responses to Did the 50p tax rate really raise less than £1 billion in 2010/11?

  1. Rob
    Mar 22nd 2012, 1:49 pm

    Excellent and typically insightful post Howard. Presumably the behaviour distorting impact of announcing the change to be effective as of next financial year will ensure that there is a self-fulfilling prophecy in the form of high earners putting off income from 2011/12 into 12/13 to pay tax at a lower rate – the mirror image of the bringing forward meaning that the dataset for the change will be essentially meaningless. Unfortunately, there is no doubt that this will be used in neo-classical orthodox economics textbooks for years to come as an example of why high marginal tax rates are ineffective at raising revenue. It is also a good reason (from the Tory/Laffer/TPA/orthodox school of thoughtPOV) for doing it now when there is insufficient evidence rather than later when there *might* be evidence that they don’t like…

  2. Why the 50p tax-cut is based on Osborne’s own dodgy dossier | Liberal Conspiracy
    Mar 22nd 2012, 2:40 pm

    […] Reed is among many to have cast doubt on this study, and argues that past Treasury figures have actually overestimated such avoidance (TIE) rather than […]

  3. The Sloman Economics News Site » Blog Archive » Assumptions about taxable income elasticity
    Mar 23rd 2012, 1:41 pm

    […] blog (22/3/12)Budget 2012: are we really all in this together? Guardian, Polly Curtis (21/3/12)Did the 50p tax rate really raise less than £1 billion in 2010/11? Touch Stone, Howard Reed […]

  4. Britain and France differ over Tax Justice — Social Europe Journal
    Mar 29th 2012, 9:10 am

    […] pointed out by various commentators—see for example the piece by Howard Reed —either HMRC is just plain wrong, or even if HMRC is right and the marginal rate encourages […]

  5. Britain and France differ over Tax Justice – they’ve got une révolution fiscale » Tax Research UK
    Apr 2nd 2012, 9:17 am

    […] pointed out by various commentators—see for example the piece by Howard Reed —either HMRC is just plain wrong, or even if HMRC is right and the marginal rate encourages […]