The Laffer Curve Straw Man

I’ve already addressed this issue, but here we go again. Commenting on the “totemic appeal” of the Laffer Curve for conservatives, Kevin Drum of Washington Monthly asserts that the concept is no longer relevant because:

…the current top marginal rate in the United States is 35%, and no one in their right mind thinks that’s anywhere near high enough to have a serious Laffer effect. When top rates are that low, lowering them further just reduces tax revenue.

Mr. Drum may know how (fairly) to mock politicians who exaggerate, but his critique of the Laffer Curve is based on the straw-man proposition that tax cuts are self-financing. Notwithstanding some of the political rhetoric, the Laffer Curve does not imply — and never has implied — that all “tax cuts pay for themselves.” That only happens if rates become sufficiently punitive to put the taxing jurisdiction on the downward-sloping portion of the curve.

The real issue is whether certain changes in tax policy will have some impact on economic activity. If an increase (decrease) in tax rates changes behavior and causes a reduction (increase) in taxable income, then revenues will not rise (fall) as much as “static” revenue-estimating models would predict. This is hardly a radical concept, and evidence of Laffer-Curve effects is very well established in the academic literature.

The reason there is a debate is that the government’s revenue-estimating bodies (the Joint Committee on Taxation on the Hill and the Office of Tax Analysis at Treasury) assume that tax policy changes have zero impact on economic performance. That’s right, zero.

For example, if the entire tax code was scrapped and replaced by a low-rate flat tax, JCT and OTA would assume no effect on macroeconomic aggregates. If tax rates were doubled, JCT and OTA would plug new numbers into their simplistic (yet totally nontransparent) models and estimate that tax revenues would double (to be fair, the government’s revenue estimators assume some microeconomic dynamic effects, such as higher tax rates causing people to shift the timing and/or composition of income, but this is akin to measuring the tail and ignoring the dog).

So what does all this mean? Two points are worth highlighting:

First, the current system is rigged against good tax policy by over-estimating the revenues that can be obtained by raising tax rates and exaggerating the revenues foregone when tax rates are reduced. This is why some of us advocate “dynamic scoring.”

Second, not all tax cuts (or tax increases) are created equal. An increase in the personal exemption or the child credit will have very little, if any, impact on incentives to work, save, and invest. As such, the “static” estimate of revenue losses will be reasonably accurate. A reduction in the tax rate on capital income, by contrast, will substantially alter incentives for taxpayers who have considerable ability to adjust their behavior. This will result in a Laffer-Curve response, though it is an empirical question whether the revenue feedback will offset 20 percent, 50 percent, or 75 percent of the static revenue loss (timing is also important since a tax rate reduction that yields a revenue feedback of 20 percent in the first year may generate a much larger revenue feedback five years in the future).

The left is very clever. Defenders of the status quo have created a straw man, and they find quotes from politicians and others with little knowledge to create the impression that advocates of lower tax rates believe in the fiscal version of a perpetual-motion machine. This tactic is then used to prop up the existing system of revenue estimating, which is based on assumptions that would earn an F if put forth by a student in an undergraduate public finance course.