Yesterday I commented on the 2008 OECD study that shows the richest 10% of Americans pay more in tax, relative to their share of income, than anywhere else in the developed world. That story has been working its way through the econ blogosphere, with Scott Sumner arguing yesterday that the study's results show that countries with higher income inequality institute more progressive tax systems.
Sumner's point draws on the positive correlation visible between share of income earned by the rich and share of taxes paid by them. As the rich earn a larger share of the economic pie (Sumner uses this a proxy for income inequality), their share of the tax burden also grows. I've reproduced my version of the graph showing these two variables (Sumner and others have been looking at a slightly different one that has the axes switched).
The connection seems pretty clear. But is share of income captured by the top 10% necessarily the best indicator here to stand in for income inequality? Intuitively it makes sense -- we'd generally expect countries where the rich earn more of the wealth to be less "equal." Still, let's take a look at what the graph looks like if we replace share of income earned by the top 10% with the Gini coefficient, a well-known measure economists use to compare income inequality across societies.
Now the picture is a little less clear. Though some countries (U.S., Italy, Poland) are in roughly the same places as they were in the graph using the top 10%, the neat correlation we see in the first graph has given way to an amorphous, globular mass. It seems that when a more robust measure of income inequality is used, the correlation with progressivity tends to disappear.
I highlighted some particular cases that buck the "more inequality -> more progressive tax system" idea -- the Netherlands and Australia, both of which are low-to-average in terms of income inequality but have highly progressive tax systems.
It seems there might be more to the relationship of tax system structure to income inequality than the original graph (income of richest 10% vs. taxes paid by richest 10%) might let on. Countries that have a more even income distribution can still have highly progressive tax systems or, as Poland demonstrates, vice versa.