Taxing the poor to help the rich?

Rob Salmond has written a post claiming that New Zealand’s tax system is unfair on poor people and generally inefficient. His evidence boils down to this chart of tax rates across incomes:

Rob’s an expert on tax systems so I trust that the figure is accurate, but there is so much it doesn’t say that bears on his conclusion. There are a few points that immediately spring in to my mind, although I’m sure you can think of plenty more.

  1. Most importantly, a tax system’s incidence should be judged by net taxes, rather than gross revenues. Taxes don’t disappear into a bottomless pit; they accrue to someone as a benefit. Looking at the net tax people pay, once government services are taken in to account, shows a different picture. As you can see, lower deciles receive more services and transfers from the government than they pay for in taxes, and the reverse is true for wealthier deciles. So, even if there is a flat effective total tax rate, that is not the same as a flat tax incidence. I have no idea how this compares to tax incidence across similar nations, so maybe we still have a high relative incidence on poorer people.
  2. We might also ask why it is that Rob believes it so intrinsically unfair that tax rates are flat. From the same publication by the Institute of Policy Studies, here is the average income tax paid by each decile: Now we can have different views about what fair is, but it isn’t obvious to me that that distribution is unfair without a lot of normative judgments being mixed in.
  3. Rob also claims that the high GST in New Zealand is unfairly regressive, which has been discussed by Matt numerous times previously. To summarise, GST is not regressive over a person’s lifetime but it may affect the welfare of low income people more than the welfare of high income people.

Rob finally concludes that the tax system is bad for efficiency and the economy. He doesn’t draw any causal links between his discussion and conclusions, and it’s not immediately clear to me why a fairly constant average tax rate across income groups generates any of the outcomes he describes. I haven’t read Rob’s book, so I probably don’t see the connection because it’s complicated enough that you need a whole book to explain it. At least, I hope so because no effort is made to draw the connections in his blog post. This is really the nub of what bothered me about Rob’s post: it suggests a lot more than it shows and the content doesn’t appear to support the conclusion.

Maybe I’m being unfair because he’s trying to summarise a lot of material in a very short post. But, when you’re a really smart political scientist, you don’t need to provide charts without context and conclusions without justification in order to convince people of something. Particularly if you’re so familiar with the arguments that you wrote a whole book about it! I really hope that this post is just a teaser and we’ll see more in this series to back up the hefty conclusions that have already been drawn. Or, perhaps, this is just a ruse to get us out to buy the book 😛

  • Richard 29

    Re: 1.
    To my knowledge New Zealand is not known for being massively generous in welfare support payments. The one exception to this is our generous and non means tested universal pension from 65. I understand we have one of the highest rates of child poverty in the OECD and pretty much the lowest rate of pensioner poverty in the OECD. It’s not that redistribution is not occuring – it’s just that it seems to be more intergenerational than needs based. Housing is a big factor here as well – a lot of pensioners live in a rent free home. Capital is untaxed in NZ and the redistribution does not account for the existence of capital so it won’t show in these stats but has a massive impact on quality of life. Most families with young children do not live in a rent free home.

    Re: 2.
    Of course the top deciles will pay more tax even in a flat tax system. You would need a massively regressive tax system for that graph to be flat. I fail to see on what basis you are supposed to know how much to redistribute without making normative judgements. The fundamental purpose of redistribution is to ensure an acceptable standard of living for all participants in society, otherwise why do it? Defining an ‘acceptable’ standard of living is entirely normative.

    • Of course you need normative judgments to determine a fair distribution. It feels to me like Rob is saying “people of all incomes pay the same tax rate, which is obviously unfair.” What I’m saying is that if you frame the facts differently then it becomes far from obvious what is fair. Indeed, there may well be disagreements among equally knowledgeable people on the matter. More than an appeal to common sense is required to establish the unfairness of NZ’s tax system.

  • The single most regressive tax in NZ is not the GST (Matt’s right on that one). Look instead to tobacco excise. When O’Dea did his cost report for ASH a few years back, he had some incidence-by-decile analysis of the likely effects of big tobacco tax increases, tabulating the benefits to those who then quit against the costs to those who keep smoking. Worth a read.

  • Rob Salmond

    Thanks for the link. My blog post is really making two points rather than one. They are:
    1.     Poor folk in New Zealand pay more tax than they would pay in most other rich countries, while rich folk pay less than they would pay in most other rich countries. That is the part that strikes me as normatively troubling;
    2.     Poor folk and rich folk pay pretty similar overall rates of tax in New Zealand. That is the part that strikes me as socially inefficient.
    You wondered about why flat taxes might be inefficient in welfare terms. I think the answer is pretty straightforward – if I take 30% of a poor person’s income I am lowering their utility more than if I took the same 30% from a rich person. Therefore the more tax we take from poor people to fund our services, the more we are draining the community’s stock of utility. And flat taxes are at the end of the real-world distribution on that dimension. (Now of course I realize that if we sought only to correct that particular inefficiency, we would end up in a very punitive “soak the rich” situation. My concern is that we spend a lot of time fretting about the technical inefficiencies, but not really any time considering the other side of the coin.)
    In terms of your own points, I think that when you start examining net taxes, you are really assessing the “tax and public service system,” rather than just the tax revenue system as I did. International comparisons are what interest me most here, and comparisons of only the revenue system are complex enough, and even the OECD folk can’t do it as thoroughly as we might like. OECD-wide comparisons of net-taxes are another level of complexity again.
    I was confused by your use of the raw dollar totals of tax paid by decile. Tax systems are usually assessed using rates, not raw totals, and I do not see what we can learn from the figures you presented, other than the trivially true point that people with large incomes pay more absolute tax.
    I specifically address the “lifetime GST” argument in my book. I understand the point, but I do not think lifetime GST is a good metric. In all other areas of taxation, we assess progressivity / regressivity against current period income, not lifetime income. I fail to see why we should treat GST differently. The point of progressive taxation is to levy a higher rate **at the time a person has high income.** The timing is an important consideration.

    • Thanks for coming over and clarifying your points, Rob. I have no quibble at all with the factual points you make; facts are facts, after all. However, I still think it is very difficult to make any normative judgment about the equity of a tax system without looking at the net incidence of it. I understand that data is very difficult to come by on that count. Looking only at tax revenues is far easier and there are still plenty of interesting things to say about it. Yet, it’s difficult to make the leap from that to a comment about how our tax/public service system treats the poor, since they are also the greatest beneficiaries from those revenues. I fear that looking at revenues alone is searching under the spotlight, as economists themselves are often accused of doing!

      I think that same point is why I struggle with your idea of social inefficiency. If I take 20% of a low income and almost simultaneously give 400% of that money back then it’s misleading to focus only on the tax and claim that I’m reducing their utility. Without a comparison of net tax incidence it seems hard to make any normative claims about social efficiency, although comparisons between nations may still be interesting.

      Hopefully that summarises the substance of what I wanted to say. The use of raw dollars was only to illustrate that a constant tax rate is not as obviously inequitable if you frame it differently. There are plenty of proponents of flatter tax systems and I don’t think their moral compass is so fundamentally different to the average person’s. I wasn’t suggesting that it added to the substance of your analysis, only that the framing matters for the inference that people draw from the chart.

      • Rob Salmond

        Thanks James for the response. Certainly I agree with the idea of looking at combinations of policies rather than single policies in as much as we can – it gives us a better sense of the overall policy terrain. That is what underlies my hope that we can move away from thinking about our “GST policy” or our “income tax policy” to instead our “tax revenue policy.” If it were feasible, I would like nothing more than to look at international comparisons of the entire tax and social services system considered together. Sadly, I do not have a good idea how to accomplish that task at this point. I guess I would rather look under the spotlight than shut my eyes.
        I think what I can say, though, is that **for any given social policy** there are various methods, or combinations of methods, of raising the required revenue. Some of those combinations are more efficient, in welfare terms, than others. Flatter tax regines tend to be less efficient on this dimension. I hope this is all I ever claimed in the post – that is certainly all I intended to claim.

        • I guess I would rather look under the spotlight than shut my eyes.

          I’m sure we would all agree with you there, Rob, although the strength of one’s conclusions obviously depend upon how much of the solution you think the spotlight illuminates. On that question, I think we may reasonably differ.

          Some [tax regimes] are more efficient, in welfare terms, than others. Flatter tax regimes tend to be less efficient on this dimension.

          That is a very striking conclusion you have reached, and certainly at odds with some conventional wisdom among policy-maker; not that that’s a bad thing. If we might tax your patience just a little further, would you be able to comment on how consistent your conclusion is with the optimal tax theory of economists such as Mirrlees, Diamond, and Saez? I don’t know the theory well but I’m aware that a lot of the arguments about the optimality of rising marginal tax rates revolve around the assumed distribution of ability. Does your book deal with their approach?

  • Hone

    Not sure if this comment should go to Rob’s site or this one, seeing as I am commenting on Rob’s views, but anywho…
    (1) At a minimum, tax rates should be judged net of cash transfers. A reduction of 30% of a “poor person’s” notional or gross income is not going to reduce their utility disproportionately if they (both) did not earn that income and never saw the money. 
    (2) If the chart is correct, it tells a nice story in terms of declining tax rates in the lower portion of incomes – a nicely efficient result. However…
    (3) I suspect some jiggery pokery in the chart. How can it be that the GST rate is declining in the low income part of the curve (<14k where income tax is constant and GST is a constant proportion of disposable or gross income)? Is this some kind of savings assumption? 
    (4) Judging progressivity/regressivity on current income is dumb. “We” might do it but “we” are still dumb to do it. 
    I suspect, like James, that perhaps it is worth getting the book. I will do if Rob can convince me that his chart has not been jiggered and pokered. I am sure there is a simple answer.

  • Rob Salmond

    @Hone: First, I respectfully differ from your views that the declining overall tax rates in the lower part of the income distribution in nicely efficient, and that lifetime consumption is a good metric for judging consumption tax (if that is true, why not use lifetime income to judge income tax?).
    Second, the I estimate the GST incidence using Stats New Zealand data on the proportion of incomes that ultimately attract GST in a given year. The Tax Working Group use the same data in <a href=”>>its paper on GST</a> (see Figs 6 and 7).  The reason the incidence is declining through most of the chart, but not declining at very high incomes, is that the original data are based on income deciles. At income levels above the median for the highest decile, for which there was no direct data, I was forced to assume no further decline in GST exposure. That is the flat GST section you see on the right hand side of the chart. If anything, this may result in slightly overstating the GST paid by people with very high incomes.

  • Rob Salmond

    I do not consider optimal tax theory in the book. In fact, in the book I go out of my way NOT to draw any controversial policy conclusions, as my goal was to help people understand the contours of the tax regime without colouring it will too much politics. Blog posts give me much more room for those kinds of shenanigans…
    With that disclaimer, and the further disclaimer that I am not a real expert in optimal tax policy, I think this recent JEP paper from Diamond and Saez suggests their position and mine are not really that different. Here is the abstract:

    This paper presents the case for tax progressivity based on recent results in optimal tax theory. We consider the optimal progressivity of earnings taxation and whether capital income should be taxed. We critically discuss the academic research on these topics and when and how the results can be used for policy recommendations. We argue that a result from basic research is relevant for policy only if 1) it is based on economic mechanisms that are empirically relevant and first order to the problem, 2) it is reasonably robust to changes in the modeling assumptions, and 3) the policy prescription is implementable (i.e, is socially acceptable and not too complex). We obtain three policy recommendations from basic research that satisfy these criteria reasonably well. First, very high earners should be subject to high and rising marginal tax rates on earnings. Second, low-income families should be encouraged to work with earnings subsidies, which should then be phased-out with high implicit marginal tax rates. Third, capital income should be taxed. We explain why the famous zero marginal tax rate result for the top earner in the Mirrlees model and the zero capital income tax rate results of Chamley and Judd, and Atkinson and Stiglitz are not policy relevant in our view.

    And a link to the paper:

    • A wide range of authors were very unhappy with the Diamond and Saez paper – as it used a very static version of “optimal tax” theory, rather than the more widely accepted idea of using a dynamic framework.  This is one of the primary reasons why they suggested such a significant tax on capital (both in terms of a capital income tax AND in terms of high MTR’s on high income earners).

      When I’m not as heavily pinned down by deadlines I’ll have a hunt around for the critical work I saw on these papers and pop up a link 😉

  • Rob Salmond

    Honestly, I think there are real problems on BOTH sides of this, not just the S-D side. To be sure, there are a bunch of unhappy folk about this paper, but S-D are no slouches, and they have support in their corner, too. Krugman is an example, as is DeLong. The fact that some people are unhappy about some elements of the paper does not make S-D wrong on counts.
    And some of the criticism looks, at first blush, a bit shallow to me. For example, from your link themoneyillusion says “[Higher taxes on high incomes] would not put higher taxes on Warren Buffett, as the tax won’t come out of his consumption, it will come out of the investments that he no longer makes, and the charities to which he no longer contributes.” First, the assumption of lower charitable giving seems pretty presumptuous. Second, even if Buffet dials back on his investments in response to higher taxes, that does not necessarily lead to national-level welfare loss. Remember, the government does have more money with which to make investments of its own, and there is no guarantee that Buffest’s investments, with a primary aim of increasing Buffet’s wealth, are better at growing the nation’s welfare than are the government’s investments, which have that as their more direct objective. Even if the government just split up the extra Buffett tax money between 1,000 poor people, the period 1 national welfare would go way up. Of course (anticipating your objection), there would be multiple feedback effects in periods 2, 3, etc. But it is an open question whether the welfare-draining feedback loops are more powerful than the initial welfare-enhancing effects and the welfare-enhancing feedback loops combined.

    • “To be sure, there are a bunch of unhappy folk about this paper, but S-D are no slouches, and they have support in their corner, too. Krugman is an example, as is DeLong”

      Their posts supported the idea that we should have higher MTR in the US – this is due to their subjective belief about the equity-efficiency trade-off.  This is fair enough, but I doubt they would support the inference of the S-D paper regarding what the equity-efficiency trade-off was in fact.

      Posts by both of them on these sorts of issues are with regards to politics not economics – they would not want 74% MTR’s at the top, they just want to push the idea that MTR’s should be higher.

      “And some of the criticism looks, at first blush, a bit shallow to me”

      The money illusion goes into a whole bunch of issues he believes in – which is nice, and I do not agree with many of them.

      However, the S-D assumption of little impact on human capital accumulation is indeed a poor one – in reality this is an empirical question.  This is a huge criticism of their figures.

      This discussion of S-D is all much of a muchness as you say, you are interested in discussing the optimal equity-efficiency trade-off that is accepted in New Zealand.  Comparing us to other countries appears to assume that, on average, countries meet this – which in itself isn’t a fair assumption. 

      The key way to figure out what we should do is to ask “what is the trade-off” and then “what trade-off does our optimal social contract entail” – this requires a direct tackling of optimal tax theory, and then a movement into defining a subjective “social welfare function or profile” for the country.  That is just the way we’d peak at things.

      Still, I look forward to reading your book – I haven’t read the Big Kahuna yet either, so I’ll give both a go and then pop up my thoughts.  I can imagine jamesz may do similar.

  • Pingback: On GST and regressivity » TVHE()