John Taylor's Congressional Testimony On The Stimulus

See The 2009 Stimulus Package: Two Years Later. It has some great charts. Excerpts:

"My empirical research during the past two years shows that ARRA did not have a significant impact in stimulating the economy. I do not think this finding should come as a surprise. Earlier research on the discretionary countercyclical Economic Stimulus Act of 2008— enacted three years ago this week—indicates that it too did little to stimulate the economy. Research on the discretionary countercyclical actions in the late 1960s and 1970s—the most recent period of such large interventions prior to this past decade—also shows disappointing results, including high unemployment, high inflation, high interest rates, and frequent recessions;"

"only a tiny slice of ARRA has gone to purchases of goods and services by the federal government."

"But when you look at what state and local governments did with the funds, you find that they did not increase purchases of goods and services or increase infrastructure projects."

"But state and local government purchases have hardly increased at all and they are still below the levels of late 2008 before ARRA grants began."

"What did the states do with the ARRA funds? The data show that they mainly used the funds to reduce their borrowing"

"However, aggregate personal consumption expenditures did not increase by much at the time of these sharp increases in stimulus payments. In general, the overall pattern of personal consumption expenditures seems to move closely with disposable personal income without the addition of the stimulus funds, though the decline in consumption is greater than the decline in either measure of disposable personal income."

"the effect of the temporary stimulus payments on personal consumption expenditures is much smaller than the effect of more permanent income changes and statistically insignificant from zero."

"As in the case of the grants to the states the temporary payments were mainly added to personal saving in the form of reduced net borrowing. In effect the increased borrowing by the federal government to finance ARRA was nearly matched by a decrease in net borrowing by the state and local governments and by persons."

"the percentage contribution of government purchases (federal as well as state and local combined) to the growth of real GDP is negligible. The swing in economic growth during the recession and the recovery seems largely independent of the changes in government purchases."

"The contribution of consumption is larger, but still consistent with the finding that the change in consumption was due to changes in income without the stimulus payments. Note that the largest contribution from consumption comes in the last quarter of 2010 at the time when the agreement to extend existing tax rates likely became anticipated and eventually a reality with increased expectations that they would become permanent."

Why Do Policy Evaluations Differ?

"Why do some argue that ARRA has been more effective than the facts presented here indicate? Many evaluations of the impact of ARRA use economic models in which the answers are built-in, and were built-in before the stimulus package was enacted. The same economic models that said, two years ago, that the impact would be large now show that the impact is in fact large. This is why, for example, the Congressional Budget Office finds larger effects while other researchers using different models find smaller effects. The models disagree so the policy evaluations disagree. The data I examine here place more emphasis on where the funds from ARRA actually went. The approach makes less use of simulations of existing econometric models, although it uses general theories—such as the permanent income theory or similar theories of government behavior—to analyze the data. These data point to some inconsistencies, however, in how model simulations have been conducted. For example, many model simulations assumed that ARRA would have a much larger effect on government infrastructure and other purchases than was actually the case. I believe the simulations should take account of the very small amount of funds that went to government infrastructure and other purchases. If CBO or other groups are still assuming in theirsimulations that a large fraction of grants to the states went to government purchases, then those simulations should be adjusted."

Conclusion

"In sum, the data presented here indicate that the American Recovery and Reinvestment Act was not effective in stimulating the economy. Despite its large size, ARRA did not result in more than an immaterial increase in government infrastructure and other purchases at the federal level. The large grants to the states did not result in an increase in government infrastructure and other purchases at the state and local level. And finally an analysis of the payments that temporarily increased disposable income shows that they did not significantly affect personal consumption expenditures. In contrast changes in private investment and net exports have been much more of a factor in the recovery. Currently, the increased debt caused by ARRA both directly through its deficit financing and indirectly through its de-emphasis on controlling spending—is likely a drag on economic growth."