### Thursday, February 05, 2009

## Cost benefit analysis of stimulus

Arnold Kling points to CBO estimates of the Senate version of the stimulus package, which is being debated today. A couple of points jump out at me. First, much of the expected improvement in GDP if one passes this bill is relying on there being a lot of excess supply right now. The low and high estimates for GDP growth this year if you pass the bill are 1.4% more growth to 4.1% more growth. Assuming a one percent increase in employment (technically that should be hours worked, but let's assume we don't use each worker more intensely here) generates a 0.7% increase in output, as most estimates would use, that says the contribution of capital and productivity should be between 1% and 2.9%. Given the estimates of labor productivity are thought to be around 1.5%-2%, that seems a little optimistice. Those ratios fall for 2010. So there must be some catch-up effect being estimated here. Maybe that's right.

Suppose we take, however, the cost estimates and the estimates of additional production to be created. The latter can be measured by taking CBO's guess of the closing of the GDP gap (in the letter to Sen. Gregg yesterday) and estimating them out through 2011. Compare it to the increase in the budget deficit, and you should have some cost and benefit. I have to deflate the deficit costs, so I take the CBO's forecast for GDP inflation and BEA's current 2008 estimate.

If CBO's high estimate was right, your trillion dollar investment would generate an increase in GDP over the next three years of about $350 billion, or 2.7% of one year's potential GDP. In per capita terms, the most optimistic estimate generates less than $1200 additional income for each American. (I could, I guess, do a more careful estimate of net present value and make this a little more elegant but I'm not trying to be that precise; I note that most of the benefits are in calendar year 2009, which is the point of the Senate version.) But if the low estimates are correct, there is a net cost of about the same magnitude. On net, we don't know from CBO's estimates if this stimulus is a net gain or loss to the economy. Given Kling's sensible comment on protecting the creditworthiness of the U.S. government, it seems like we're gambling a lot on a very dubious proposition.

This file contains the data and estimate.

Suppose we take, however, the cost estimates and the estimates of additional production to be created. The latter can be measured by taking CBO's guess of the closing of the GDP gap (in the letter to Sen. Gregg yesterday) and estimating them out through 2011. Compare it to the increase in the budget deficit, and you should have some cost and benefit. I have to deflate the deficit costs, so I take the CBO's forecast for GDP inflation and BEA's current 2008 estimate.

If CBO's high estimate was right, your trillion dollar investment would generate an increase in GDP over the next three years of about $350 billion, or 2.7% of one year's potential GDP. In per capita terms, the most optimistic estimate generates less than $1200 additional income for each American. (I could, I guess, do a more careful estimate of net present value and make this a little more elegant but I'm not trying to be that precise; I note that most of the benefits are in calendar year 2009, which is the point of the Senate version.) But if the low estimates are correct, there is a net cost of about the same magnitude. On net, we don't know from CBO's estimates if this stimulus is a net gain or loss to the economy. Given Kling's sensible comment on protecting the creditworthiness of the U.S. government, it seems like we're gambling a lot on a very dubious proposition.

This file contains the data and estimate.

Labels: economics