Sorry, Ezra Klein: The ‘one tax graph you need to know’ needs more context
By Michael Strain and Stan Veuger @ AEI (September 26, 2012)
The above graph has been floating around and receiving a lot of attention. Labeled “the one tax graph you really need to know” by Ezra Klein on his popular blog, this graph was produced by the Citizens for Tax Justice.
This graph shows the tax bill — importantly, federal and state and local taxes — as a percentage of income for several income groups. The bottom income quintile pays 17.4% of their income in taxes, while the top 1% pays 29%. The middle quintile pays just 3.8 percentage points less than the top 1%.
The point of this graph is correct: State and local taxes are much more uniform than federal taxes, so total taxes are less progressive than federal taxes. Indeed, according to Citizens for Tax Justice, in 2011 state and local taxes were actually regressive, with the bottom twenty percent of the income distribution paying a higher share of their income in state and local taxes, 12.3%, than any other income group, and the top 1% of earners paying less than all other income groups at 7.9%.
The problem with the graph that Mr. Klein has popularized is not that it includes taxes which it shouldn’t. Instead, it excludes taxes that it shouldn’t. Which taxes does it exclude? Negative taxes, which are also called transfers. In other words, the graph includes payments from households to the government, but it does not include all payments from the government to households.
As N. Gregory Mankiw — Harvard economist and current adviser to Governor Romney — noted on his blog back in July, using data from the Congressional Budget Office it is very easy to calculate income after taxes and transfers as a percentage of market income.
Note that this includes both federal and state transfers, but only federal taxes. As we pointed out above, state and local taxes are close to uniformly distributed across income groups. So if state and local taxes were included in this picture, the bars for the lowest quintiles would be a touch lower relative to the bars for the highest quintile. But when you’re talking about 10 or 15 percentage points this hardly matters — including state and local taxes doesn’t change the qualitative interpretation.
What does this chart show? For the lowest income group, income after taxes and transfers was four times that of income before taxes and transfers. Only for roughly the top half of income earners do taxes and transfers result in an after-tax-and-transfer income that is less than their market income — for which the ratio is less than 100%.
As the next chart shows, this relationship is not perfectly stable over time, but the qualitative story is. In economic downturns, more people qualify for transfers like food stamps and unemployment insurance payments. But the significant progressivity of the system is clearly not driven by the business cycle.
In popularizing “the one tax graph you really need to know,” Mr. Klein hasn’t even artificially truncated the distribution over taxes at zero. Instead, he seems to include some transfers from government to individuals — for example, the EITC — but not others, like food stamps, Social Security, TANF, veteran’s programs, worker’s compensation, Medicaid, CHIP, and many other transfers. Not including these transfer payments is a significant omission.
…But the point remains that our tax system when properly analyzed is extremely progressive.
It is important to state plainly that in broad strokes we believe the tax code should be structured in this way: It should be the case that folks facing hard times are net beneficiaries of the government, and likewise that the most fortunate among us are net contributors.
In this presidential election, when taxes and transfers are such an important issue, the public is best served when the media presents an accurate and comprehensive description of the realities of public policy. We wouldn’t want to argue that ours is “the one tax graph you really need to know.” Really, you need to know more than one.