By Simon Johnson
Can tax cuts “pay for themselves” – inducing so much additional economic growth that government revenue actually increases, rather than decreases? The evidence clearly says no.
Nevertheless, a version of this idea, under the guise of “dynamic scoring,” has apparently surfaced in the supercommittee charged with deficit reduction – the joint congressional committee with 12 members. Dynamic scoring sounds technical or perhaps even scientific, but here the argument means simply that any pro-growth effect of tax cuts should be stressed when assessing potential policy changes (e.g., reforming the tax code). For anyone seriously concerned with fiscal responsibility, this is a dangerous notion.
Economists disagree about almost everything, of course, and the effect of tax cuts is no exception. One reasonable way to assess the evidence is to begin with the highest plausible effects, then see what happens if some of the more extreme assumptions are relaxed (this is a nice way of saying that we don’t believe everything the authors are trying to tell us).
I would start with a study by Gregory Mankiw, former chair of George W. Bush’s Council of Economic Advisers – and therefore presumably on the tax-cutting side of American politics – and Matthew Weinzierl that shows the economic growth caused by a tax cut can at best offset a portion of the revenues lost by that tax cut. (Unfortunately, you need a subscription to the journal to read the study.)
Specifically, Profs. Mankiw and Weinzierl calculated that 32.4 percent of the “static” or direct revenue loss of a capital gains tax cut and 14.7 percent of the static revenue loss of a labor tax cut could be offset in present-value terms by additional growth, ignoring short-term Keynesian effects (i.e., any immediate stimulus provided to the economy).
Now 32.4 percent is a lot, but it is far less than 100 percent. And a critical assumption for Profs. Mankiw and Weinzierl is that government spending falls to keep the budget in balance. In their framework that’s a good thing – as they are effectively assuming away any productive effects of government spending (e.g., what if less spending on schools means less education and this hurts “human capital” and therefore productivity down the road?)
Sticking for a moment with just with their view of the world, if instead the tax cuts are financed by additional debt – as was our collective experience during the 2000s – the ultimate effect of those cuts can be to lower economic growth in the long term, depending on whether the larger debt eventually leads to lower government transfers, lower government consumption, higher taxes on capital or higher taxes on labor. (See Eric M. Leeper and Shu-Chun Susan Yang, “Dynamic Scoring: Alternative Financing Schemes,” Journal of Public Economics 92 (2008): 159-82, pp. 166-69. Again, a subscription is needed to read the article.)
More broadly, in 2005 the Congressional Budget Office – then headed by a Republican appointee, Douglas Holtz-Eakin – estimated that the economic effects of a 10 percent cut in income taxes would offset between 1 and 22 percent of the revenue loss in the first five years; in the following five years, the economic effects might offset up to 32 percent of the revenue loss — but might also add 5 percent to the revenue loss.
This is an entirely reasonable assessment – the C.B.O. exists in order to provide balanced analysis for the budget process. The bottom line is that betting that tax cuts will pay for themselves is a high-risk strategy – and not a good idea at our current levels of government debt relative to gross domestic product. We do not have a large margin for error. (Disclosure: I’m on the Panel of Economic Advisers for the C.B.O, but I didn’t have anything to do with that study).
Of course, economic studies do not necessarily have a direct effect on political discourse. For example, President George W. Bush asserted in 2007, “It is also a fact that our tax cuts have fueled robust economic growth and record revenues.” But this is nothing more than an assertion. Growth during the 2001-7 expansion was only 2.7 percent compared, for example, with 3.7 percent during the 1990s expansion (when tax rates were higher).
And much of the growth during the Bush period turned out to be illusory; it was based on our corporate and national accounting system, which measures profits (an important part of G.D.P.) but not on a risk-adjusted basis. When the risks materialized in the financial crisis of 2008-9, we lost so much output that G.D.P. per capita in real terms today is only at about the level of 2005.
To assess growth properly, you should look “over the cycle,” meaning roughly 10 years for the modern American economy. It is hard to argue that the last decade was any kind of growth success. Of course, other things happened during the 2000s – including further financial sector deregulation not directly related to the tax cuts.
That’s why we have the economic analysis, particularly by the C.B.O. – to disentangle what tax cuts can really do. If the supercommittee buys into dynamic scoring for tax cuts, at best this would be wishful thinking. At worst, it would represent yet another round of fiscal irresponsibility at the top of American politics.
And if anyone is seriously considering altering the rules under which the CBO operates, they should stop and think again. Changing the scorekeeping guidelines at this stage would amount to undermining the credibility of the Congressional Budget Office – one of the few remaining impartial and well-informed observers. Perhaps this strategy might yield some short-term political gains but the damage to our creditworthiness would be immense, and the consequences would be felt sooner rather than later.
The nightmare downward spiral and fiscal implosion in the eurozone began with a few countries cheating on their numbers – first to get into the currency union and then to avoid various forms of official criticism. Do not start down the same path.
An edited version of this post appeared on the NYT.com’s Economix blog; it is used here with permission. If you would like to reproduce the entire post, please contact the New York Times.