Its time for dynamic scoring

Posted: May 02, 2002 12:00 AM
"It is a standard tactic to take someone's proposal, misstate it and then disagree with it. It's very effective. I've used it myself." So said Sen. Edward Kennedy, Massachusetts Democrat, at the National Press Club recently. Nowhere has this technique been put to better use than with the issue of "dynamic scoring." This refers to the idea of taking into account the economic effects of tax changes in order to calculate their impact on revenues. The notion may seem commonsensical, but in fact it is quite controversial. For many years, it has been standard practice for the Treasury Department and Joint Committee on Taxation to calculate the revenue effects of tax changes on a static basis -- that is, without fully accounting for their impact on key economic variables. This is equivalent to a business planning an increase in the price of its product without considering the possibility that sales will fall as a consequence. Any business that did this would soon go bankrupt, because there would be no logical reason not to raise prices to infinity. Nevertheless, the JCT and Treasury historically have assumed that if tax rates go up, say, 10 percent, then revenues will automatically rise by 10 percent. Indeed, so rigid is this methodology that when the chairman of the Senate Finance Committee once asked the JCT how much additional revenue would be generated by a 100 percent tax rate, it dutifully replied that significant revenues would be obtained. Of course, anyone without a Ph.D. in economics knows that no revenue at all would be raised. Who in their right mind would work if the government took 100 percent of their earnings? If zero revenue is obtained at a 100 percent tax rate, then clearly a tax rate reduction at this point would raise revenue. To be sure, further rate reductions might lose revenue. But somewhere between a zero percent tax rate and a 100 percent tax rate there is one that maximizes revenue. Above this point, tax rate reductions will increase revenue; below, they will lose it. One goal of tax policy should be to identify particular tax rates that can be cut at no revenue cost to the government. These would include those on economically sensitive forms of income, such as capital gains, and on people whose attachment to the labor force is weak, such as spouses whose mates earn more than they do. It is also important to look at how tax rates affect the use of legal tax avoidance devices, such as tax-free municipal bonds, and illegal tax evasion opportunities. It may not be practical or politically feasible to take advantage of every situation that exists for cutting tax rates at no revenue cost. But they do exist and represent a kind of "free lunch" for the government. Having said this, it certainly does not mean that all tax cuts will have no revenue cost. Many will lose as much revenue as static estimates calculate. These are principally those that do not affect incentives -- i.e., those that do not change relative prices, such as the tradeoff between work and leisure or saving and consumption. Unfortunately, most tax cuts of recent years, such as the child credit and last year's tax rebate, fall into this category. On the other hand, there are tax increases that will cause revenue to fall. In general, it is reasonable to say that tax increases don't raise as much revenue as static estimates show, and tax cuts do not lose as much. For years, many economists, such as Harvard's Martin Feldstein, have argued that the failure of the Treasury and JCT to incorporate behavioral and macroeconomic effects in their revenue estimates creates a bias in favor of tax increases and against tax cuts. I would say that this methodology also biases the political system against tax cuts that would have a strong, positive effect on growth, in favor of those that have virtually none. It also encourages enactment of tax increases that are more harmful to growth than necessary to obtain a given amount of revenue. In particular, static scoring encourages increases in marginal tax rates and discourages marginal rate cuts. It is too easy to characterize the former as "fair" and the latter as giveaways to the rich when their true economic and revenue impacts are not presented to policymakers. The solution to this problem is to incorporate all economic effects of tax changes in revenue forecasts. Although the Treasury and JCT do try to include some behavioral effects, macroeconomic effects -- such those on employment and gross domestic product --- are still ignored. On May 7, the House Ways and Means Committee will meet to discuss the weaknesses in current revenue forecasting methods. It should press for the use of dynamic scoring wherever possible.