Comment on ‘What taxpayers, governments and tax economists do – and what they should do’

Many economists working on tax policy, including my colleagues at the Institute for Fiscal Studies (IFS), advocate for reforms which we believe will improve the efficiency and equity of the tax system. Equalising tax rates across different forms of income rather than treating self-employment and business income more favourably than income earned by employees, a uniform VAT rate rather than extensive exemptions and zero rating, and reducing or abolishing stamp duty land tax and increasing and reforming council tax, are just three among many such reforms. We feel we are on solid ground in arguing for change, yet change does not happen.

The justifications by policymakers and politicians for lack of reform are unequivocal. The latter will point to the political difficulties. Whatever the strength of the case for putting VAT on food or increasing council tax, the sales job, with an eye to re-election, is just too hard. However generous the compensation package, a broad extension of VAT would cause a political backlash. The failed attempt to extend it in full to domestic fuel consumption in the early 1990s has not been forgotten. Stamp duty may be a staggeringly damaging tax, but it is collected at the point of a very large transaction, is not especially unpopular, and most is paid by those buying expensive properties. Why give it up? Any reform to council tax would cause losers, and it is already an unpopular tax.

Those in the Treasury or the revenue department make a similar point differently: the single overwhelmingly most important aspect of tax design so far as they are concerned is to make sure the revenue rolls in. Whatever the theoretical case for major reforms, if the current system isn't broken, don't try to fix it. It may look broken to those of us concerned with its efficiency, equity or complexity, but not to those who know that, in the UK context, it will yield them £1 trillion a year without too much trouble. I have had that conversation too many times to recount.

It's also why one of the biggest tax cuts of recent years has come in the reverse way – fuel duty has been frozen in nominal terms for 13 years, representing a huge real-terms cut. Raising it each year in line with prices is not a real increase, but it looks like one. We have discovered that the longer it is frozen, the harder it is to unfreeze. An annual inflationary increase used to be expected and accepted. That norm has changed. Which at least is evidence that norms can be changed.

Money illusion of this kind certainly influences policy decisions. Whether the use of fiscal drag tips us into a situation where ‘intentional imperfect taxpayer education becomes deliberate deception’ is perhaps moot.

This raises questions for all concerned. Are there more effective ways of making the case for rational reform? What reforms that at least move us in the right direction might be palatable? Do we have credible ways of estimating a cost of imperfect tax design, and can we make use of that in the debate?

There is a different set of issues arising from the research carried out which is aimed at understanding the likely effects of changes to marginal rates on behaviour and on economic output. We struggle with estimates and are uncertain of the consequences of changing the structure of corporation tax or raising the top rates of income tax. Here, changes are frequent and justifications confident but the consensus among researchers often considerably less clear. Slemrod's conclusion – that we should lay out our assumptions about behavioural change and views as to the appropriate degree of redistribution when commenting on such issues – is a welcome one. Transparency may not make for simple communication, but it has to be a vital first step to a better-informed debate.

Raising revenue requires minimisation of evasion in a direct sense – more evasion equals less revenue. It is also important indirectly. The sense that some ‘get away with it’ through either evasion or avoidance damages taxpayer morale. Authorities are enormously sensitive to the idea of widespread evasion because it has the capacity to breed more evasion.

In a different context, Slemrod raises the puzzle of the public scepticism of estate taxation, inheritance tax in the UK. There may be many reasons for this, but one is surely the knowledge that it is easily avoided by the properly wealthy. If, as is true of most middle-class families, by far the biggest asset you have to pass to your children is the family home, then there is little scope for avoidance. If, however, you have substantial liquid wealth, then avoidance is absurdly easy. Which helps explain why the average tax rate paid on estates over £10 million is only half that on estates of £2 million. This is not evasion, but the point is the same – if the rich are seen to be ‘getting away with it’, taxpayer morale will suffer. HM Revenue and Customs (HMRC) will point out that most actual evasion comes from cash-in-hand small tradesmen not paying income tax or VAT. This seems to many to be fair game, though illegal, whereas clever avoidance by the very rich and big companies is not, though legal.

Again the issue of honesty raises its head. Many in the public believe that – for non-PAYE income – detection of evasion is relatively likely. It is not. Tax audits are rare and the information available to HMRC is less complete than one might expect. That false impression is no doubt good for tax collection and adherence to the law.

Close to the heart of all those of us who work on public economics is the question of what the research community should do. Slemrod asserts the importance of transparency over assumptions: ‘I support an increase in the top marginal tax rate, based on my distributional values and my estimate of the top ETI of 0.2; if you told me that the ETI was greater than 0.5, I would retract my support.’ I agree.

That is exactly the sort of response though that does not go down well with policymakers. As Slemrod puts it, if you want to be asked back to a congressional hearing you need to play the ‘confidence game’ – make strong assertions about what is true rather than discuss uncertainties and ranges. As Lyndon B. Johnson is supposed to have quipped ‘ranges are for cattle’. I have spent enough time being quizzed by politicians to know how hard it can be to resist the pressure to be definitive where certainty is not warranted.

Perhaps pushing in the other direction is ‘p-hacking’ – doing everything possible to find a statistically significant effect. Papers that find a behavioural effect from tax are more likely to be published. This is not a neutral bias. Behavioural consequences tend to imply a welfare cost from higher taxes. If there is a bias in research and publication, there is a danger that it will be a bias against more progressive taxes.

Even more important to research and its quality, relevance and impact is the availability of data. Administrative data are controlled by the tax authority. Fear that research will indicate that policy is ineffective may be one of the barriers to making data available. There is always more risk to officials in making data available than in not doing so. There is without question an under-provision of data and resources for those wanting to do empirical tax research. Data availability has certainly improved in recent years, but is much worse in the UK than it could be. One doesn't want to overstate the extent to which our knowledge of tax policy and its effects would be transformed by better data, but it would be helped a lot. That would come both directly and indirectly from the fact that its mere existence would draw more researchers into the field. The preponderance of UK academics working on policy problems in other countries is depressing, and results in part from the better availability of data in those countries.

Slemrod concludes that ‘the biggest barrier to economics being taken seriously as a science is a correlation between the researcher's political views and values on the one hand, and the results of their empirical investigations on the other’. That has all too clearly been true in some economic debates in recent years in the UK, and the public may not be in a good position to judge who is speaking objectively and who is not. It is, I would argue, why institutionally independent organisations such as the IFS are so important. That institutional independence can exert influence on individual researchers to pursue as objective an agenda as possible in a way which is not possible within a university setting.

So, what is needed? More data, more transparency, and more independent and policy-focused research.