Alan Auerbach feeds the 2001 Bush tax cut to his and Larry Kotlikoff's (1987) behavioral model, and finds out that--once again--the drain on national savings produced by the revenue losses from the tax cut more than offset any likely savings increases from the incentives provided.

This is interesting because--in my humble opinion at least--the Auerbach-Kotlikoff model leans on the side of finding large behavioral responses where they are unlikely to exist in the real world. In the Auerbach-Kotlikoff model, the people who receive last year's tax cut by and large do not perceive it as an increase in their lifetime wealth--they expect to pay the tax cut back with interest a decade or so down the road when the government raises taxes to balance its budget. Hence people decide to work more now when taxes are low, work less a decade from now, and save their extra income.

Thus Auerbach and Kotlikoff's model has a much bigger savings response than does a--in my opinion more realistic--model that sees tax cut recipients as believing the tax cut raises their lifetime wealth, and so decide to boost their consumption now. Yet even this bigger savings response is not enough to disrupt the conclusion that the tax cut is likely to lower national saving and hence economic growth over the next decade or wso.

The Bush Tax Cut and National Saving

by Alan J. Auerbach | NBER Working Paper No.w9012 | Issued in June 2002

---- Abstract -----

Following through on pledges made during his election campaign, President Bush proposed and Congress passed a substantial tax cut in 2001, the Economic Growth and Tax Relief Reconciliation Act (EGTRRA). Much has been written about the size of the tax cut, its impact on the federal budget, its distributional consequences, and its short-run macroeconomic impact. There has been less focus on EGTRRA's incentive effects; one of the most important potential behavioral effects is on saving. To analyze the behavioral effects of the Bush tax cut on saving and other macroeconomic variables, I use the Auerbach-Kotlikoff (1987) model in conjunction with the NBER's TAXSIM model. An interesting by-product of this analysis is the 'dynamic scoring' of the tax cut - the estimated feedback effects of behavior on revenue. By comparing the revenue losses generated by the model with those that would occur without any behavioral response, one can estimate how much of the static revenue loss would be recouped by expanded economic activity. The simulations suggest that dynamic scoring has a significant impact on estimated revenue losses, but that the tax cut's impact on national saving is still negative in the long run.

Posted by DeLong at July 02, 2002 06:42 AM