The Facts about Tax Progressivity

This is a guest post from political scientist Lucy Barnes, Prize Postdoctoral Research Fellow at Nuffield College, Oxford:


The recent exchange between Jonathan Chait and Veronique de Rugy over the progressivity of the U.S. tax system in comparative context, and subsequent contributions (e.g. from Brad de Long) exemplifies the widespread confusion over what tax progressivity actually means. The distribution of the tax burden is obviously a politically loaded question, where those on the right tend to have an interest in claiming that the US tax system is already (too) progressive, while those on the left wanting to increase the share of the burden shouldered by the rich. However, the question of how progressive the system is an empirical question. In fact, the system of taxation in the United States is relatively progressive. What makes this fact surprising is that tax progressivity and fiscal redistribution (the reduction of inequality by government action) are often conflated, and it remains true that redistribution in the US is low, due mainly to the relatively small size of the US government. The American case typifies one pole of a robust negative relationship between tax progressivity and overall redistribution, the cause of which remains a contested question in the literature on comparative political economy.

So, what are the facts about tax progressivity? The most comprehensive scholarly work on this question to date comes from sociologists, Monica Prasad and YingYing Deng, who use data on individual incomes to calculate total tax burdens at different levels of the income distribution (gated, ungated). Their main takeaway finding is illustrated in the two figures below, which show the Kakwani index of taxation (a measure of the progressivity of the system that parses out the impact of income concentration on the concentration of the tax burden) for a number of advanced democracies.

These data are from the most recent year available (around 2000 in most cases). The top figure shows those taxes that are paid directly by individuals: income, wealth, property and employee social security contributions, while the bottom figure shows measures (albeit for a smaller set of countries) that incorporate the sales tax burden based on the individual household expenditure data. What is clear from both figures is that the United States’ tax structure is more progressive than that of its peer countries. The Bush tax cuts (enacted since these measurements) will have decreased progressivity somewhat, but not enough to alter these rankings.

The main reason for this is the relatively low reliance of the American fiscal system on sales taxes, which constitute a much higher share of revenues in other countries (even when including state and local taxes in the US case, as the figures above do). General sales taxes place a disproportionate share of the tax burden on the lower end of the income distribution because expenditures are higher as a share of income as we move down the distribution. In addition, specific excises tend to be levied on goods representing a larger share of poorer households’ budgets, such as taxes on alcohol, tobacco and gasoline. Again, these taxes tend to be higher in other advanced democracies than they are in the United States.

This claim, that the American tax system is progressive compared to those of its advanced economy peer countries, is hard for many (in both the US and in Europe) to accept. The conventional wisdom is that the United States intervenes comparatively little in redistributing income from rich to poor. It is not that these stylized facts are untrue: the figure below shows the reduction in inequality accomplished by government intervention (again using data from the Luxembourg income study)—both taxes and transfers.

Here the United States takes up its more accustomed place at the bottom of the pack (and it is data like these that have been used in the recent debate to refute the claim that US taxes are progressive). How is it possible that American taxes be progressive, while achieving so little redistribution? The answer is that redistribution is not driven primarily by the structure of taxation, but by its level. The next figure shows the close relationship between the level of taxation (total government revenues in GDP) and reductions in the Gini.

In fact, the association between tax progressivity and overall redistribution across countries is negative, as illustrated below .

This stylized fact is the focus of a large emerging literature on the relationship between the size of government and the structure of taxation.  Why are political demands for progressivity and redistribution, widely seen as interchangeable, channeled so differently into public policy? The existing literature on the topic focuses on the economic and political constraints that high spending imposes on politicians seeking to alter tax structures (for further reading see here, here and here, to start). But the politics of taxation, as distinct from those over redistribution writ large, remain generally understudied in political science (see also here.)

12 Responses to The Facts about Tax Progressivity

  1. ML February 16, 2012 at 1:13 pm #

    I would think the disconnect between the progressivity of the tax system and the lack of redistribution can be explained in part by the large number of transfers that go to the non-poor. I’m thinking of the big ones like Medicare and Social Security, but tax expenditures like the state/local deduction, employer health insurance exemption and the mortgage interest deduction all play a part as well.

  2. Rob February 16, 2012 at 1:53 pm #

    I agree with ML and would add that much of the redistribution that takes place in the US is upward and flows to corporate interests through government contracts, tax breaks, etc…. As a percent of GDP I think the U.S. is near or at the bottom in social welfare spending, though I haven’t checked the numbers in a few years.

  3. CF February 16, 2012 at 2:49 pm #

    Social Security and Medicare while not targeted to the poor, are redistributed to the poor through the structure of the two programs. And yes, tax expenditures draw over 1.3 trillions dollars away from the most progressive aspect of the U.S. tax system, the federal income tax, and distribute these funds to very wealthy households.

  4. Alan T February 16, 2012 at 4:56 pm #

    Thanks for an enlightening and thought-provoking post.

    Are your conclusions consistent with claims (see, for example, Krugman or figure 2 of Carasso and Steuerle that the U.S. tax system is barely progressive at all?

    I would like to point out an additional reason why progressive taxation of income and redistribution of wealth are not at all the same thing. Much wealth comes from unrealized capital gains, which are not considered income. Warren Buffett’s wealth increased by $3 billion in one year, but his taxable income was “only” $40 million.

  5. Guest February 17, 2012 at 4:35 am #

    Thank you for this highly interesting post. It is certainly true that the politics of taxation are under-researched. However, I’d like to hint to a project of my university that is doing research on “The Tax State and International Tax Policies”. A small contribution to fill the gap..

    But since there is a collaboration with Oxford on welfare state research, you might already know this program.

  6. Lucy Barnes February 17, 2012 at 6:27 am #

    Thanks all for the comments.

    Quickly re: Alan T: the question as to whether the US tax system is progressive overall is somewhat different from the question as to whether it is progressive in comparison to other advanced industrial countries. If you look at the second figure where sales taxes are included in the calculation of the tax burden it’s true that the progressivity of the US system is low in absolute terms. I did a quick calculation using the data in the CTJ paper Krugman links to to calculate the Kakwani index for 2010, giving a figure of 0.056, higher than the 0.02 in my figure, so I’d say that there’s not a whole lot of inconsistency there.

    Whether a Kakwani index of 0.05 is `barely progressive at all’ is one that I am less confident has a good empirical answer, as it depends on what you think the relevant comparison should be. The more limited claim that I wanted to make was that whatever you think about the overall level of US tax progressivity, the claim that it is low relative to other countries is not one that is supported in the data once we understand progressivity properly, as distinct from redistribution.

    Re: Guest– I think that some of the best exceptions to my claim of `political science hasn’t studied taxation enough’ have come in this arena- the implications of international constraints on the politics of taxation- including a lot of the work that has come out of this project.

  7. Eric February 17, 2012 at 6:11 pm #

    This is a nice breakdown of a mistaken conflation of ideas. What I would love to see is the results of taking the existing tax structures and equalizing them for per capita tax paid. E.g. take the total tax paid by a person in Belgium and then adjust the US taxes across the board so a comparable person in the US was paying the same amount of tax, to show how much redistribution would occur if US citizens paid as much tax as other countries.
    I have heard the argument that the reason Europe can charge more taxes per capita specifically is because each taxpayer in Europe feels like their personal taxes are not being redistributed as much, even though overall government spending, per capita, on social safety nets, education, infrastructure and foreign aid is much higher.

  8. Peter A February 18, 2012 at 8:45 am #

    The real question is Where does a country’s wealth end up? How unevenly is it distributed? And what is the trend? Isn’t the proportion of wealth held by the top 1% in the United States much greater than in the other countries mentioned here? Surely tax progressivity and redistribution policies should be measured against a starting point, since the success or failure of these programs depends on the change they bring to their own economies, rather than on relative changes across diverse economies with vastly differing starting points. Those economies which at the outset have the least unevenly distributed wealth among their citizens naturally have less space to show striking results from their tax and redistribution policies. And hardly need to.

  9. Simon February 20, 2012 at 11:49 am #

    The claim that U.S. taxes are more progressive than those of Sweden and France is not at all surprising if you have lived in these three places. The thing with big-government systems is that they require the poor to pay high taxes. There are only so many rich people, and high-tax economies tend to further reduce their numbers.

    In Sweden, by the way, the governing center-right coalition lowered the income taxes in recent years. This actually reduced by 12% the number of Swedish children growing up in poverty.

    • Peter A February 25, 2012 at 8:42 am #

      Your dual claims that taxes both reduce the number of rich people and increase the number of poor people suggest net redistribution is to the middle classes, paid for by both the very rich and the very poor. In fact the so-called poor in Sweden are the benefactors of government taxing and redistribution, not the victims of it. And what taxes they pay will hardly define to which class in Swedish society they belong. But the level of redistribution will. Your claim that tax cuts have directly resulted in an end to child poverty for so many defies belief. A booming economy is more likely the cause of rising living standards, and Sweden’s recent boom was well under way before the marginal tax reductions you cite were proposed.

  10. John Thacker March 12, 2012 at 1:44 am #

    “The Bush tax cuts (enacted since these measurements) will have decreased progressivity somewhat, but not enough to alter these rankings.”

    Are you certain? The Bush tax cuts, on the whole, were progressive. ( among other sources.) The lower quartiles had their taxes cut by a greater percentage than the upper quartile. It’s possible that they were not quite progressive enough, but I think that further measurements would be needed. (The 2002 and 2003 tax cuts were regressive, but the 2001 tax cut was definitely progressive, with converting the lower range of the 15% bracket to 10% among other things, and the overall effect of all three laws was still progressive.)

    State taxes are often, however, less progressive than US federal taxes, and US state taxes are quite large compared to many other countries.

    But the biggest difference with redistribution with the USA remains that so much of the spending is aimed at the middle or upper middle class.

  11. Kevin November 14, 2012 at 2:02 am #

    So because we’re not (yet) as screwed as Europe, we should screw ourselves further. Got it.