Fiscal Arithmetic and the Little Depression
Matthew Yglesias is puzzled by John Cochrane:
Paul Ryan Dinner Buddy John Cochrane Offers A Sum Of All Right-Wing Fears | ThinkProgress: Apparently when Paul Ryan went out for dinner and a $350 bottle of wine, his dining companions were some hedge fund jerk and University of Chicago economist John Cochrane. Not coincidentally, Cochrane published a paper relatively recently (PDF) that offers a novel model of fiscal and monetary policy in a recession that has the convenient property of affirming all of Rep Ryan’s political views. In particular, Cochrane argues that contra everyone on the Krugman-Bernanke axis, it’s simply not possible for either fiscal or monetary authorities to halt a deflationary trend. And he also argues that it’s not necessary to attempt to do so. And he warns that not only is it plausible to think that runaway inflation is right around the corner even though there’s no evidence of elevated inflation expectations, he argues that runaway inflation is likely to be sparked by tax increases and can best be combatted by gutting Social Security and Medicare.
I heard John Cochrane give this paper when he came through Berkeley. And it seemed to me to be simply... wrong.
In Cochrane's framework, he holds the long-run future paths of government purchases and of taxes constant and looks at the effects of shocks to the interest rate on Treasury securities on total spending. The Little Depression, he claims, started with a sharp fall in the long-run real interest rate on Treasury securities. This magnified the real value today of the future government primary surpluses that provide the ultimate backing for government debt. People were thus willing to pay much higher prices for government debt relative to currently-produced goods and services than they had before. The switch of demand from currently-produced goods and services to government debt put downward pressure on the price level, production, and employment.
This seemed to me when Cochrane presented it, and seems to me now, to get the causal chain wrong.
The expected future government primary surpluses are not constant: when interest rates on Treasury bonds fall, people expect that taxes in the future will fall as well because the size of the future debt will be lower and the taxes needed to amortize it will be lower as well. Indeed, this has to be the case arithmetically: the real value of the government's outstanding debt in terms of currently-produced goods and services did rise during our current Little Depression but by only about 1/10 as much as the shift in discount rates and the duration of net debt amortization would imply. 90% of the adjustment in Cochrane's debt-amortization equation came on the side of changing expected future primary surpluses; only 10% on the side of the changing real value of the government debt today. Overwhelmingly the story is not that changing discount rates drive changes in the real value of the debt today but rather that changing discount rates drive changes in expected future primary surpluses.
Cochrane indeed notes that his model does not work:
Fluctuations in ‘‘aggregate demand’’ are somewhat mysterious, and do not easily line up with other ways we might measure expectations of future surpluses...
But his answer? This:
But accounting for the history of U.S. stock prices by news about expected dividends has been an even more catastrophic failure...
I do not feel that the argument "my model does not work, but it is not as bad as other finance models" is terribly persuasive.
I would argue that the Little Depression did not start with a fall in the long-run real interest rate on Treasury securities that greatly raised the present value of expected future government primary surpluses and caused a stampede as everybody tried to switch from spending on currently-produced goods and services to buying up today's government debt which carry ownership rights over those expected future government primary surpluses.
I would argue that the Little Depression started with a sharp rise in expectations of the degree of risk carried by private, risky assets; that this rise produced a huge and immediate increase in spreads vis-a-vis safe Treasury securities; and that the fall in the real interest rate on Treasury securities was a result of the Federal Reserve's attempts to neutralize the effects of the confidence shock on the private economy.
The Federal Reserve could have, I maintain, pegged the Federal Funds rate at 5% and kept it pegged there throughout 2007-2009 had it wanted to--and such a peg would have created expectations of continued high Federal Funds rate policies and kept long-term Treasury rates higher than they have been. The Federal Reserve could have simply expanded the supply of bank reserves at a slower rate by buying fewer Treasury bonds for its own account, and that would have by standard money market mechanisms have kept the Federal Funds rate at its 2006 level. Those Federal Fund rates and the expectation of their continuance would have induced Treasury bond traders to sell Treasuries as their prices rose, and kept long-term Treasury interest rates at significantly higher values than we see today.
According to Cochrane's framework, such a high interest rate Federal Reserve policy would have avoided the Little Depression completely: no fall in long-tern Treasury rates, no rise in the present value of expected future government primary surpluses, no switch out of spending on currently-produced goods and services.
Everybody else--or nearly everybody else--however, thinks that had the Federal Reserve taken action to peg the Federal Funds rate at 5% from 2007 on that the Little Depression would have been another Great Depression indeed. Everybody else--or nearly everybody else--sees low long-term Treasury interest rates as the result of and not as the principal cause of the downturn.
But here I should turn the mike over to people who make these points better than I do--like Andrew Harless and Mark Thoma and Nick Rowe. In order to do economics right, you need not just to have hold of an equilibrium condition, you need to know in which way you must turn the crank in order to get sense rather than nonsense.
