In a recent post I claimed that monetary policy failure associated with the zero bound, not financial turmoil was the cause of the Great Recession. Brad DeLong had correctly noted that the S&L fiasco of the late 1980s involved losses similar to the subprime crisis as a share of GDP. Where I disagreed with DeLong is that he focused on differences between the resolution of the 1980s and 2008 crises, whereas I focused on differences in monetary policy. In particular, I argued that rapid NGDP growth in the late 1980s put us so far above the zero bound that it was never an issue in the 1990 recession, and hence there was no failure of monetary policy. I argued that falling NGDP caused by overly tight monetary policy explains the severity of the 2008-09 recession.
I also argued that we could have had a severe recession in late 2008 even if the financial turmoil had been handled optimally. For instance, suppose Lehman had been successfully resolved and there was no post-Lehman banking crisis. Even in that case the tightening of lending standards would have depressed the Wicksellian equilibrium rate, perhaps below zero, and we could very well have had the same monetary policy failure. Here’s how DeLong responds:
As I have said repeatedly: I simply do not buy this. From the fourth quarter of 2005 through the fourth quarter of 2007 lending standards tightened enormously, and investment is residential construction collapsed. By the end of 2007 residential investment had fallen 2.5%-points from its peak-of-the-boom level. But the heightening of lending standards and the unwillingness of people to make further NINJA (no income, no job or assets) loans had not sent the economy into any sort of a recession. It was the spiking of risk premia in 2008 that sent us to Wicksellian natural-rate-of-interest-below-the-ZLB territory. And there is no reason to think that we would have been in such a situation but for the financial crisis–and every reason to think that the whole mishegas would have been avoided had congress simply put the too-big-to-fail banks into conservatorship in January 2008…
Let me first concede that DeLong might be correct, my claim is certainly not a necessary part of the general market monetarist worldview. But I don’t think he is correct. Consider the following:
1. I agree that the long downslide in residential real estate between late 2005 and the end of 2007 was not associated with a recession. Indeed I’ve made that argument myself on a number of occasions, as a way of pointing to the importance of monetary policy (aka NGDP growth.)
2. Here’s how I read the events of 2008. During calendar 2008 the Wicksellian equilibrium rate fell gradually, under the impact of both the housing slump, and feedback from sharply slowing NGDP growth, which was itself an indication of tighter monetary policy.
3. Let me point to two important facts for people who don’t like the Bernanke-Sumner “NGDP” indicator of monetary policy, and who insist on “concrete steppes.” The monetary base, which had been trending upward for many years at more than 5% per year, stopped growing between August 2007 and May 2008. If you prefer interest rates as your concrete step, note that stock market responses to (timid) Fed rate cuts indicated that in late 2007 that the Fed was “behind the curve,” as even Bernanke later admitted.
4. This accidental tightening of monetary policy tipped us into a very mild recession in early 2008. But still nothing severe, so at this point I still agree with DeLong.
5. The inflation surge in the first half of 2008 scared the Fed, and prevented them from doing what they otherwise would have done, as it became increasing clear we were in a recession. By May 2008 the rise in unemployment was already more than you ever observe in non-recessionary periods. And as (I seem to recall) Bernanke once observed, it’s almost impossible to be too expansionary when the economy is tipping into recession.
6. But the Fed was not expansionary, even using the conventional interest rate indicator. Fear of inflation led to them to keep their interest rate target on hold at 2% after April 2008; indeed they did not even cut interest rates at the first meeting after Lehman failed in September. Unemployment had reached an expansion low of 4.4% in 2006. Here are the unemployment rates from April to August 2008, all fully known to the Fed at its September meeting: 5.0%, 5.4%, 5.6%, 5.8%, 6.1%. And there were no rates cuts at all! Look at that data and ask yourself if a steady fed funds target represents normal Fed behavior on the edge of a recession. Then think about the fear of inflation that is so evident in the minutes of the 2008 meetings.
7. Because the Wicksellian equilibrium rate was falling during this period, the stable fed funds target tipped the economy from a mild recession into a severe recession.
8. Ignore quarterly GDP numbers, which can be very misleading at turning points. The monthly GDP estimates from Macroeconomics Advisors show GDP (both real and nominal) suddenly plunging sharply between June and December 2008. By December the plunge was almost over, even though quarterly data for 2009:1 were much lower, as the level had been falling sharply throughout the 4th quarter of 2008.
9. My claim is that accidentally tight monetary policy caused the NGDP plunge in the second half of 2008. Because of data lags (and later revisions in data) we didn’t know about that NGDP plunge until after the post-Lehman financial crisis broke. So at the time it seemed like the financial crisis caused the severe recession. And almost all economists still believe that to be the case. But markets were already sensing problems before Lehman, and falling asset prices caused by falling NGDP almost certainly helped trigger the September to December financial crisis, which occurred in the last half of the tight money-induced June to December fall in NGDP.
Let me finish by admitting that financial crises can have real effects, and thus it’s possible that DeLong is correct. It’s possible that I’ve underestimated the independent impact of financial distress on RGDP. Maybe RGDP would have fallen sharply even with a monetary policy aggressive enough to boost NGDP in late 2008. But I’d point to one observation in my favor. FDR enacted a highly expansionary monetary policy during the spring of 1933, when much of the US banking system was closed down. Both NGDP and RGDP rose very strongly. The technique FDR used (dollar depreciation) has some similarities to NGDPLT, as it raises expectations of future NGDP.
PS. I have one quibble with DeLong’s post. His use of ellipses in a quotation of material from my earlier post somewhat mischaracterizes what I wrote:
One area where I slightly disagree with Krugman is his focus on inflation. A 5% NGDPLT target path would have been enough, we didn’t need 4% trend inflation. Nor do we need fiscal stimulus…. All stabilization policies eventually fail…. The trick is to have a modest failure like Clinton or Obama, not a serious failure like FDR or Nixon. NGDPLT would have given us just that in 2008-09.
The ellipsis right before “The trick is” makes it seem like the final sentence refers to failed stabilization policies of each President. Not so, I was discussing their overall performance. Here is the complete passage:
All stabilization policies eventually fail, just as all presidents are judged failures in their 6th year in office. The trick is to have a modest failure like Clinton or Obama, not a serious failure like FDR or Nixon. NGDPLT would have given us just that in 2008-09.
It’s no big deal, and my comment was probably ambiguous. But Nixon’s big failure was Watergate, and Clinton didn’t even have a modest failure of stabilization policy, just a silly scandal.