Tax, Saving Welfare and Retirement: Have We Lost our Way? Symposium, Retirement Policy and Research Centre, University of Auckland, 16 July, 2009
<> <>Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist. Maynard Keynes <> <>Keywords: Distributional Economics; History of Ideas, Methodology & Philosophy; Regulation & Taxation; Social Policy; <> <>I was invited to give this paper because a few months ago I wrote a Listener column (4 April 2009) in which I pointed that a recent Wellington tax symposium had hardly discussed equity. Contributors advocated tax reform – as they have done for some 25 years – and their apparently sensible proposals have been ignored. When one has been advocating change for a quarter of a century and made little progress, it is time to step back and wonder why. I suggested it was because the reformers were not paying attention to equity, but their political masters had to. <> <>The Wellington Seminar is not alone. In May the government announced a tax working group to advise them. It would be fair to say that not a single one of the group (perhaps aside from its officials) has any expertise in equity issues. The IRD recently put a list of six factors contributing to a good tax system. The last was ‘A good tax system is characterised by high levels of voluntary compliance and is thus perceived as broadly fair’, a grudging acknowledgement that equity (albeit an odd notion of and justification for equity) might be taken into consideration but it was not a high priority. <> <>Yet, equity – or at least self interest – is never far from any discussion on taxation. Politicians do not over-rule reformers out of cussedness, but because equity is an important component of the tax system. So how did economics get into a muddle of ignoring equity considerations? (I leave how politics did, to another venue.) <> <>It is useful to go back to Arthur Pigou’s magisterial Economics of Welfare published in 1920, which derived many welfare rules, from the assumption that an additional dollar was less valuable to a rich man than a poor man. But, said those who came after him, how do we know that to be true; what is the scientific basis for comparing the standard of living of different people? And if we have no scientific basis to make comparisons, surely economists are no better placed than others, to make comparisons of well being. <> <>Instead economists tried to build welfare economics around another objective which seemed to be much less contentious. Pareto Efficiency, named after the Italian economist Vilfredo Pareto, says that economic situation One is superior to economic situation Two if somebody is better off, and nobody is worse off in situation One compared to situation Two. <> <>It is a strong requirement, because it is hard to think of practical policies that do not make someone worse of. There is even an economist’s principle which states this. Rabin’s Law is that all existing policies are Pareto Efficient, that is if you change the policy it will make someone worse off. Perhaps there are a few exceptions, following the introduction of a new technology or a new situation, but they are very rare. If we taught equity properly in economics we would not only teach Rabin’s Law, but as a matter of course we would assess who were better off and who were worse off when policy changed. <> <> <>An extension of the Pareto criteria is that the winners in a policy change from situation Two to situation One could compensate the losers so they were no worse off. When they applied the compensated criteria economists found that many – but not all – of the Pigou welfare rules gave Pareto optimal with compensation. <> <>However there were many other areas of policy where this new approach did not work. Yet economists wanted to give, or were asked for policy advice. So they sought to relax the rigorous requirement of Pareto Efficiency with compensation. One option was that after the policy shift from situation Two to situation One there was still a sort of gain, even though those worse off were not compensated. I think what the proponents had in mind was that it was someone else’s job to make the interpersonal comparisons and decide whether the improvements in the better-offs more than compensated for the losses of the worse-offs <> <>Where compensation need not happen, the notion is sometimes called ‘efficiency’, although I should warn that there are so many meanings of ‘efficiency’ floating around in economics that it is a very treacherous term. The basic idea of the various notions is getting more output for a given set of inputs (or resources). Is so increasing efficiency seems to be a reasonable objective. Few would advocate waste as a social objective. <> <>This narrow definition of efficiency is the underlying notion of those who advocate tax changes to increase GDP. The evidence is equivocal as to whether, within practical ranges, the average level of a tax system affects the level of output. That is a (very heated) debate for another day. To do it today would be to avoid facing up to equity, which might be an explanation as to why the debate is so heated. (It is evident from the studies that the winners will gain much more than the GDP increase, so others will be worse off. Even if there was a gain it would not be a Pareto efficient one.) <> <>(I also need to warn you that while we treat GDP as a simple measure of output, its level depends upon the relative prices in the system, and they can alter the level of GDP. This is a bit technical – those of you who puzzled over the Scitovsky criteria, in which welfare loci crossed, were struggling with this problem. I mention it as a warning; often underpinning simple economics are some very complex and subtle goings-on.) <> <>I have seen the GDP efficiency criteria – summarised as ‘a dollar is a dollar’ – called ‘Hume’s Law’. This is not a general term. David Hume is more famous for the insight that ‘is’ and ‘ought’ are conceptually different and should not be interchanged; a lesson lost on many economists, especially those who are advocating policies which they pretend to be pure value-free economics. <> <>The ‘dollar is a dollar’ formulation illustrates the problem of this GDP efficiency notion. Suppose a dollar was taken from each of the hundred poorest people in the land as was given to the richest man in with a gain of one dollar (say from administrative savings), so he was $101 better off. GDP would be a dollar higher too, and there would be an ‘efficiency’ gain – that is an increase in GDP. The explicit notion is a dollar is just as valuable to the rich as to the poor. Is it not extraordinary how step by step one set of principles which favoured the poor were reversed into a set which favoured the rich; how in the name of being ‘ value free’ the opposite set of values were introduced? <> <>So would the society be no worse off if we took a dollar from one hundred poor people, and gave it to a rich man? Pigou would not think so. Is the marginal value of a dollar to the rich exactly the same as the value to the poor? That seems unlikely given the universally accepted principle in economics that marginal utility diminishes for each individual as they consume more. Pigou has been upside-downed, and interpersonal comparisons have been sneaked back in. <> <>Earlier I warned that efficiency is a treacherous term. For instance, is there an efficiency-equity trade-off, the notion that if the economy pursues equity there will be a reduction in efficiency? But surely the advocates of equity want to pursue their ends in a way which is efficient. So what could efficiency-equity trade-off mean? <> <>The paradox may be resolved by observing that the term ‘efficiency’ is used here in two ways. <> <>The claim of an efficiency-equity trade-off is actually a belief that an economy which pursues equity will have a lower GDP than if it ignored equity – GDP is reduced if one pursues fairness. As I have said, a comprehensive review of the research literature does not support this conclusion. But suppose the volume of output would fall is the income distribution was made a mite less unequal. Then there would be a efficiency-equity trade-off with ‘efficiency’ meaning GDP. <> <>The second use of efficiency in the expression that the advocates want to pursue equity as efficiently as possible. That simply means they want to do it in an effective way as possible. Of course. But the meaning of efficiency here is not the same as GDP efficiency. (A diagrammatic representation off the argument. <> <>There is an even greater paradox. Consider a change which is Pareto efficient when there is compensation. For instance, suppose a tariff reduction gives an increase in output, but workers in the previously protected industries have to shift to lower wage ones. A way to compensate them would be to tax the winners on part of their increment, and recycle the takings to those who are worse off. But so we are told, taxation reduces efficiency. Which means that even though the proposed policy change is Pareto efficient, it cannot be implemented in a Pareto effective way because there could be no compensation. <> <>The fact is that policy proponents have economic values which underpin their recommendations. They need admit to the values they hold. (Hume would remind them that such values do not come out of any empirical knowledge but out of an ethical belief.) That would involve a radical change to the existing policy debates. Admitting one has values which underpinned the policy recommendations means that someone with different values might recommend different policies. Better to pretend to be value free while pursuing values which are in one’s self interest or the interest of one’s client or patron. <> <>This conclusion should be no surprise. In 1971, economist James Mirrlees, a Nobel laureate, wrote ‘An Exploration in the Theory of Optimum Income Taxation’ which shows the optimal tax structure depends on the nature of the social welfare function – there are very few general results. <> <>Where there are general results, they are often uncomfortable for those who want to reform the tax system, and tend to get ignored. Here is a simple principle I learned about thirty years ago, but which is still ignored. I present it at the simplest level. <> <>Suppose one wants to give everyone a guaranteed income which is a certain proportion of the average after-tax income of the population; suppose it is funded by a proportional income tax. I shall assume that there are no government services, only transfers. (Other government spending can be added with little extra cost with no increase in insight.) <> <>Suppose we identify everyone in the population with a number i, there are n people and their market income is Yi. The total income to be allocated will be ΣYi, summed across all n observations. <> <>The average income is therefore ΣYi,/n Let’s call it Y*. <> <>We set the minimum income as mY*, where m is the target proportion of the minimum income as a proportion of the average income. <> <>Then each person’s after tax income is the sum of their guaranteed income polus their market income after tax, or <> mY* + (1-t)Yi,. <> <>Summed across all people we get the total income is Σ(mY* + (1-t)Yi,) which can be arranged to <> nmY* + (1-t) ΣYi, <>or <> (1 + m – t) ΣYi, (since nY* = ΣYi,). <> <>Since we have no government spending, the total after tax income equals the total before tax income so <> <> (1 + m – t) ΣYi,= ΣYi,, <> <> or <> m = t. <> <>That is the average tax rate equals the ratio of the minimum to average income. That is an important enough conclusion to put in capitals. <> <>THE AVERAGE TAX RATE <>EQUALS <>THE RATIO OF THE MINIMUM TO AVERAGE INCOME. <> <>When I have asked people, they have typically said they think a minimum to average income ratio of about 60 percent seems reasonable, so their logic is that the flat tax rate has to be 60 percent. <> <>This is a very high rate. across the board, but it is an underestimate. <> – it ignores that government services also have to be funded from taxation, <> – it ignores the cost of administrating the tax system <>AND (a very big ‘and’ you notice) <> – it assumes there will be no behavioural response so that individuals faced by high marginal tax rates will not reduce their effort to make market incomes (however some may have an income target and may increase their effort). <>Collectively, then, we have to conclude that a flat tax has to be at least the ratio between the minimum and average incomes. <> <>Is the problem the flat tax? Could a cunningly designed variable tax avoid the high rates? It turns out that for any income distribution with some people with low market incomes, somewhere in the tax system there has to be a marginal tax rate which is at least that of the ratio of minimum to average incomes. <> <>I have tried various ways of lowering marginal taxes for a minimum guaranteed income. One solution is that each person should be taxed with a lump sum based on their particular earning ability, a result which will be no surprise to tax afficionados. Unfortunately it is impractical since in order to be equitable we have to know what each individual’s income earning capacity. <> <>One can,. of course, reduce the ratio so the minimum income is lower thereby reducing the required flat tax rate. That may be unacceptable in social terms – although one cannot help thinking that was a major driver when the real value of social security benefits were slashed in 1991. It was thought that increasing the wage-to-benefit gap would encourage people to move off the benefit in order to increase their income. However they were not able to reduce the gap sufficiently (in any case it is impractical to expect all beneficiaries to go to work), so to the pull effect of the higher income has to be supplemented by the push factor of the system trying to get individuals off the benefit. <> <>A way to get the average tax rate down would be to divide the population into those we expect not to work those we expect to work. Only the first group get a guaranteed income, the second get a guaranteed job. That is how we ran the social welfare system when there was full employment. The unemployed who turned up to the Department of Social Security were told to get a job, and they usually did. Alas that option is not so possible in today’s economy with its more dynamic labour market, and with greater difficulty maintaining full employment. Even so, I have wondered whether, in more benign times, this distinction might be useful in a redesign of the welfare state. <> <>In the interim we are stuck with high effective marginal tax rates, somewhere in the income redistributional system. It is to cry crocodile tears to argue that we should reduce top tax rates and to worry about the high EMTRs on those on low incomes. Those high rates can only be ultimately reduced by reducing the minimum accessible income or by increasing tax rates further up the income. <> <>My intention here has not been to advocate a minimum guaranteed income but to explore its implications for the tax system. Those who do advocate a minimum income are probably ‘Rawlesians’, that is followers of John Rawles’ principle that one should organise society by paying attention to the needs of the poorest. (I happen to be a Rawlesian for distributional objectives; I only mention this because I dont want to pretend I am entirely value free, even if the above analysis is.) <> <>What distributional principle is followed by those who pursue efficiency and a dollar is a dollar is harder to discern, especially as they are not forthcoming about the ideology and values which underpin their policies. Some seem to me to be so pro-rich they might be called anti-Rawlesian; the rest might just need to be more explicit. <> <>This is not a trivial matter. The tax reforms in the late 1980s and early 1990s gave a measurable boost to the after-tax incomes of the highest decile. In fact their incomes continued to rise in real terms at about the same rate as they had previously, for the effect of those tax reforms was to transfer income from the bottom 80 percent of the distribution to those on top incomes. Since the economy as a whole was contracting – GDP per capita fell every year for six years in a row – those at the bottom experienced even greater real income falls. One consequence is those on high incomes have quite a different memory of the period to the rest of the community. They were shielded from the fact that average after-tax and benefit incomes fell. For the rich the reforms were a triumph, for the rest they were an economic disaster. <> <>Another lesson from this period is that the lowering of the tax rates on higher incomes led to no discernable improvement in the economy as a whole contradicting the claim that tax cuts improve economic performance. The memory the rich have of the period shields them form this reality too, making it easier for them to claim tax cuts improve efficiency. they did not on any measure in the late 1980s and early 1990s. <> <>The last two paragraphs are based on well established estimates of the income distribution. But we dont discuss them – even though they are revealing about the political economy of the times as well as the impact of tax changes – because we are unwilling to discuss equity. <> <>That is the theme I finish on. It is time that advocates of policies were more explicit about their values and how their value-based policy proposals impact on the welfare of the various groups in the community – on the economic distributions. Recalling Rabin’s Law they need to be more explicit on who is going to benefit from their policies and who is going to be made worse off. This outing would be a big challenge to the profession since we lack many of the tools to carry out this evaluation, and where we have tools the policy advocates usually lack the skills to apply them. But we can develop them. <> <>It is a terrible criticism of the policy community to say this, but it has been lazy, unwilling to use the tools available to better understand the policies that are being advocated let alone develop better ones, or incompetent unable to use the tools. <> <>This paper began quoting Keynes’ remark that practical men and women are but slaves to some defunct economist (or other intellectual). One of the reasons I have gone through, in more detail than usual, the development of welfare criteria is because I wanted to bring to light some of the defunct economist which today’s practical policy advocates call upon. <> <>So how are we to get equity back into the tax reform discussion? One is from these kinds of presentations – but that is water dripping on a stone. There is another way. Suppose every time a policy was advocated, someone of goodwill but without a policy agenda asked the advocate – in public – that since Rabin’s Law said someone was going to be worse off from the policy change, would they list who would be worse off from their proposals? <> <>After a stunned silence (or lots of ‘ers’ and ‘ums’) the advocate may well nominate a few – an inadequate few, but it is a beginning. Persistence will eventually draw more out, and slowly and reluctantly a higher standard of identification of the winners and losers will be attained. Equity considerations would once more be included in the tax reform debate. <> <>Go to top