President Obama and others are demanding that we raise taxes on the “rich,” and two recent academic papers that have gotten a lot of attention claim to show that there will be no ill effects if we do.
The first paper, by Peter Diamond of MIT and Emmanuel Saez of the University of California, Berkeley, appeared in the Journal of Economic Perspectives last August. The second, by Mr. Saez, along with Thomas Piketty of the Paris School of Economics and Stefanie Stantcheva of MIT, was published by the National Bureau of Economic Research three months later. Both suggested that federal tax revenues would not decline even if the rate on the top 1% of earners were raised to 73%-83%.
Can the apex of the Laffer Curve—which shows that the revenue-maximizing tax rate is not the highest possible tax rate—really be that high?
The authors arrive at their conclusion through an unusual calculation of the “elasticity” (responsiveness) of taxable income to changes in marginal tax rates. According to a formula devised by Mr. Saez, if the elasticity is 1.0, the revenue-maximizing top tax rate would be 40% including state and Medicare taxes. That means the elasticity of taxable income (ETI) would have to be an unbelievably low 0.2 to 0.25 if the revenue-maximizing top tax rates were 73%-83% for the top 1%. The authors of both papers reach this conclusion with creative, if wholly unpersuasive, statistical arguments.
Most of the older elasticity estimates are for all taxpayers, regardless of income. Thus a recent survey of 30 studies by the Canadian Department of Finance found that “The central ETI estimate in the international empirical literature is about 0.40.”
But the ETI for all taxpayers is going to be lower than for higher-income earners, simply because people with modest incomes and modest taxes are not willing or able to vary their income much in response to small tax changes. So the real question is the ETI of the top 1%.
Harvard’s Raj Chetty observed in 2009 that “The empirical literature on the taxable income elasticity has generally found that elasticities are large (0.5 to 1.5) for individuals in the top percentile of the income distribution.” In that same year, Treasury Department economist Bradley Heim estimated that the ETI is 1.2 for incomes above $500,000 (the top 1% today starts around $350,000).
A 2010 study by Anthony Atkinson (Oxford) and Andrew Leigh (Australian National University) about changes in tax rates on the top 1% in five Anglo-Saxon countries came up with an ETI of 1.2 to 1.6. In a 2000 book edited by University of Michigan economist Joel Slemrod (“Does Atlas Shrug?”), Robert A. Moffitt (Johns Hopkins) and Mark Wilhelm (Indiana) estimated an elasticity of 1.76 to 1.99 for gross income. And at the bottom of the range, Mr. Saez in 2004 estimated an elasticity of 0.62 for gross income for the top 1%.
A midpoint between the estimates would be an elasticity for gross income of 1.3 for the top 1%, and presumably an even higher elasticity for taxable income (since taxpayers can claim larger deductions if tax rates go up.)
But let’s stick with an ETI of 1.3 for the top 1%. This implies that the revenue-maximizing top marginal rate would be 33.9% for all taxes, and below 27% for the federal income tax.
To avoid reaching that conclusion, Messrs. Diamond and Saez’s 2011 paper ignores all studies of elasticity among the top 1%, and instead chooses a midpoint of 0.25 between one uniquely low estimate of 0.12 for gross income among all taxpayers (from a 2004 study by Mr. Saez and Jonathan Gruber of MIT) and the 0.40 ETI norm from 30 other studies.
That made-up estimate of 0.25 is the sole basis for the claim by Messrs. Diamond and Saez in their 2011 paper that tax rates could reach 73% without losing revenue.
The Saez-Piketty-Stantcheva paper does not confound a lowball estimate for all taxpayers with a midpoint estimate for the top 1%. On the contrary, the authors say that “the long-run total elasticity of top incomes with respect to the net-of-tax rate is large.”
Nevertheless, to cut this “large” elasticity down, the authors begin by combining the U.S. with 17 other affluent economies, telling us that elasticity estimates for top incomes are lower for Europe and Japan. The resulting mélange—an 18-country “overall elasticity of around 0.5”—has zero relevance to U.S. tax policy.
Still, it is twice as large as the ETI of Messrs. Diamond and Saez, so the three authors appear compelled to further pare their 0.5 estimate down to 0.2 in order to predict a “socially optimal” top tax rate of 83%. Using “admittedly only suggestive” evidence, they assert that only 0.2 of their 0.5 ETI can be attributed to real supply-side responses to changes in tax rates.
The other three-fifths of ETI can just be ignored, according to Messrs. Saez and Piketty, and Ms. Stantcheva, because it is the result of, among other factors, easily-plugged tax loopholes resulting from lower rates on corporations and capital gains.
Plugging these so-called loopholes, they say, requires “aligning the tax rates on realized capital gains with those on ordinary income” and enacting “neutrality in the effective tax rates across organizational forms.” In plain English: Tax rates on U.S. corporate profits, dividends and capital gains must also be 83%.
This raises another question: At that level, would there be any profits, capital gains or top incomes left to tax?
“The optimal top tax,” the three authors also say, “actually goes to 100% if the real supply-side elasticity is very small.” If anyone still imagines the proposed “socially optimal” tax rates of 73%-83% on the top 1% would raise revenues and have no effect on economic growth, what about that 100% rate?