Tax changes have very large effects: an exogenous tax increase of 1 percent of GDP lowers real GDP by roughly 2 to 3 percent.
How do changes in the level of taxation affect the level of economic activity? The simple correlation between taxation and economic activity shows that, on average, when economic activity rises more rapidly, tax revenues also are rising more rapidly. But this correlation almost surely does not reflect a positive effect of tax increases on output. Rather, under our tax system, any positive shock to output raises tax revenues by increasing income.
In The Macroeconomic Effects of Tax Changes: Estimates Based on a New Measure of Fiscal Shocks (NBER Working Paper No. 13264), authors Christina Romer and David Romer observe that this difficulty is just one manifestation of a more general problem. Changes in taxes occur for many reasons. And, because the factors that give rise to tax changes often are correlated with other developments in the economy, disentangling the effects of the tax changes from the effects of these underlying factors is inherently difficult.
To address this problem, Romer and Romer use the narrative record -- Presidential speeches, executive-branch documents, Congressional reports, and so on -- to identify the size, timing, and principal motivation for all major tax policy actions in the post-World War II United States. This narrative analysis allows them to separate revenue changes resulting from legislation from changes occurring for other reasons. It also allows them to classify legislated changes according to their primary motivation.
Romer and Romer find that despite the complexity of the legislative process, most significant tax changes have been motivated by one of four factors: counteracting other influences on the economy; paying for increases in government spending (or lowering taxes in conjunction with reductions in spending); addressing an inherited budget deficit; and promoting long-run growth. They observe that legislated tax changes taken to counteract other influences on the economy, or to pay for increases in government spending, are very likely to be correlated with other factors affecting the economy. As a result, these observations are likely to lead to biased estimates of the effect of tax changes.
Tax changes that are made to promote long-run growth, or to reduce an inherited budget deficit, in contrast, are undertaken for reasons essentially unrelated to other factors influencing output. Thus, examining the behavior of output following these relatively exogenous tax changes is likely to provide more reliable estimates of the output effects of tax changes. The results of this more reliable test indicate that tax changes have very large effects: an exogenous tax increase of 1 percent of GDP lowers real GDP by roughly 2 to 3 percent.
These output effects are highly persistent. The behavior of inflation and unemployment suggests that this persistence reflects long-lasting departures of output from its flexible-price level, not large effects of tax changes on the flexible-price level of output. Romer and Romer also find that the output effects of tax changes are much more closely tied to the actual changes in taxes than to news about future changes, and that investment falls sharply in response to exogenous tax increases. Indeed, the strong response of investment helps to explain why the output consequences of tax changes are so large.
Romer and Romer also examine the behavior of output following changes in other measures of taxes. Using broader measures of tax changes, such as the change in cyclically adjusted revenues or all legislated tax changes, the estimated output effects are substantially smaller than those obtained using the new measure of exogenous tax changes. This leads the researchers to conclude that failing to account for the reasons for tax changes can lead to substantially biased estimates of the macroeconomic effects of fiscal actions.
When they consider the two types of exogenous tax changes separately, Romer and Romer find suggestive evidence that tax increases to reduce an inherited budget deficit have much smaller output costs than other tax increases. This is consistent with the idea that deficit-driven tax increases may have important expansionary effects through expectations and long-term interest rates, or through confidence.
Romer and Romer find interesting changes in the motivations for tax changes over time. Countercyclical changes were frequent from the mid-1960s to the mid-1970s, but were unheard of before that time and from the mid-1970s until 2001. Tax changes motivated by spending changes were commonplace in the 1950s, 1960s, and 1970s, but have virtually disappeared since then. Tax increases to address inherited deficits were common from the late 1970s to the early 1990s, but rare before and after this period. Only tax changes motivated by long-run considerations have been a constant feature of the fiscal landscape since World War II.
This analysis might be extended to investigate the importance of the characteristics of tax changes for their macroeconomic effects. There are strong reasons to expect the effects of a tax change on output to depend on such features as its perceived permanence, its impact on marginal tax rates, and how it affects the tax treatment of investment. Romer and Romer plan to extend their analysis to see if the output consequences of tax changes depend not only on their size, but also on these other characteristics.
-- Les Picker