A Note On The Return Of Depression Economics

by: Brad DeLong

Monster Island News Godzilla

Over the 1948-1990 period, the U.S. unemployment rate converged 1/3 of the way back to its sample average level each year. Thus the good that expansionary monetary and fiscal policy can do is limited: reduce unemployment by 0.1 percentage point this year through additional expansionary policies, and you reduce the integral of expected excess unemployment relative to trend by a total of 0.3 percentage point-years. And stimulative macroeconomic policies have costs: the policies implemented will not be the policies initially proposed, there will be reduced confidence in long-run price stability, there will be the tax wedge and tax uncertainty burdens of financing a higher national debt, and there will be the opportunity cost of not using the limited attention span and bandwidth of the political system for other policy goals.

FRED Graph  St Louis Fed

Over the 1990-2007 period, the point estimate is that the U.S. unemployment rate converges only 1/15 of the way back to its sample average level each year. Deviations of unemployment from its presumed natural rate are thus to a first approximation permanent. And taking action now to reduce the unemployment rate by 0.1 percentage points reduces the cumulative expected integral of future unemployment by 1.5 percentage points. Because there is no good reason to think that the market system will fix the problem in any reasonable span of time, the benefits to acting now are much greater than they were back before 1990: five times as great.

Thus whatever you thought the benefit-cost calculation for expansionary fiscal and monetary policies to fight high unemployment was over 1948-1990, you have to think that the benefits are much greater today.

And, analogously, you have to think that the costs are much lower today. The risk--very real in previous decades--that expansionary monetary policy would destabilize expectations and erode confidence in the Federal Reserve's commitment to effective price stability is simply not an issue today. And in previous decades the U.S. government had to pay to borrow, and so the cost of financing additional government debt was a real issue. Now investors pay the U.S. to take their money: the excess burden of debt finance is certainly no greater than zero.

You could argue in previous decades that the benefits and costs were such that a policy of macroeconomic neglect was benign, at least at the margin. I genuinely do not see how you can argue the same today.