Saturday, January 06, 2007

Mankiw vs Bush

An editorial in today's Washington Post uses my research with Matthew Weinzierl to challenge my former boss:

PRESIDENT BUSH wrote in a Wall Street Journal op-ed Wednesday that "it is also a fact that our tax cuts have fueled robust economic growth and record revenues." The claim about fueling record revenue is flat wrong, and it is shocking that the president should persist in making such errors. After all, tax cuts are the central plank of his domestic policy. How can he fail to understand the basic facts about them?

This is not just our opinion. Harvard's N. Gregory Mankiw, an economic conservative who served as chairman of Mr. Bush's Council of Economic Advisers, has tested the hypothesis on which Mr. Bush's claim is based: He looked at the extent to which tax cuts stimulate extra growth and the extent to which that growth generates extra tax revenue that offsets the initial loss of revenue from the tax cut. Mr. Mankiw's conclusion: Even over the long term, once you've allowed all of the extra growth to feed through into extra revenue, cuts in capital taxes juice the economy enough to recoup half of the lost revenue, and cuts in income taxes deliver a boost that recoups 17 percent of the lost revenue. So a $100 billion cut in taxes on capital widens the budget deficit by $50 billion, and a $100 billion cut in income taxes widens the budget deficit by $83 billion.

A few comments:

1. The editorial is wrong when it says we "tested the hypothesis." In fact, my paper with Weinzierl was data-free, and any scientific test requires data. What we did was address the policy question by calibrating a standard intertemporal general equilibrium model using parameter estimates drawn from the literature. In other words, our paper presented a model simulation, not an empirical test.

2. The 17 percent figure cited in the editorial is not for income taxes, but for labor income taxes. Because income taxes are a combination of labor income taxes and capital income taxes, the right figure for income taxes would be between 17 percent and 50 percent--I would guess about 25 percent.

3. The results cited are for our baseline model. The paper also presents a variety of alternative simulations that lead to different conclusions. For example, if one believes in large social multipliers, then the elasticity of labor supply could be larger than we assumed, and the revenue feedback effects would be greater. Similarly, if one believes that capital investment generates significant positive externalities (as suggested, for example, by the research of DeLong and Summers), then the revenue feedback effects would be much larger. Tax cuts could even be self-financing. I don't view that outcome as the most likely, however, mainly because I am not convinced of the evidence for such positive externalities. But I have enough respect for DeLong and Summers as economists that I would not completely rule out the possibility.