Did Lehman cause the Great Recession?
By Scott Sumner
Obviously not. It occurred 9 months into the recession and several months into the most severe phase of the recession. Even by August 2008, unemployment had nearly reached the peak level of the previous recession:
As we come up on the 10-year anniversary of Lehman, we can expect a flood of essays that point to this bank failure as a pivotal event in the Great Recession. Don’t believe them.
[Update: Stephen Kirchner also pushes back against the standard narrative.]
I’m very pleased to announce that Mercatus is publishing a new paper by David Glasner, which contains some quite interesting empirical work on the Great Recession. Here is the abstract:
This paper uses the Fisher equation relating the nominal interest rate to the real interest rate and expected inflation to provide a deeper explanation of the financial crisis of 2008 and the subsequent recovery than attributing it to the bursting of the housing-price bubble. The paper interprets the Fisher equation as an equilibrium condition in which expected returns from holding real assets and cash are equalized. When inflation expectations decline, the return to holding cash rises relative to holding real assets. If nominal interest rates are above the zero lower bound, equilibrium is easily restored by adjustments in nominal interest rates and asset prices. But at the zero lower bound, nominal interest rates cannot fall, forcing the entire adjustment onto falling asset prices, thereby raising the expected real return from holding assets. Such an adjustment seems to have triggered the financial crisis of 2008, when the Federal Reserve delayed reducing nominal interest rates out of a misplaced fear of inflation in the summer of 2008 when the economy was already contracting rapidly. Using stock market price data and inflation-adjusted US Treasury securities data, the paper finds that, unlike the 2003–2007 period, when stock prices were uncorrelated with expected inflation, from 2008 through at least 2016, stock prices have been consistently and positively correlated with expected inflation.
I’ve always been a fan of studies using time-varying correlations. Recall that in macro it’s very hard to establish causality, because so many things are going on at the same time. Thus if investment is correlated with GDP, it might be because investment drives the business cycle, or it might be because businesses react to changes in sales by adjusting investment.
In some cases, a model has causality claims that apply in one scenario, but not another. Thus in David’s model, higher inflation expectations don’t always bode well for the economy, but are likely to be seen as a positive indicator when in a deep recession where monetary policy is currently too contractionary (perhaps due to the zero bound problem.) David’s work focused on how the zero bound problem could prevent macroeconomic equilibrium from occurring, which is one interesting perspective. Here I’d like to emphasize a very important implication of his empirical findings, that monetary stimulus would have had a positive impact on stock prices during 2008-16, but not during 2003-07.
When David first mentioned these empirical result in a blog post, Paul Krugman was very impressed:
Scott Sumner sends us to David Glasner, who shows that stock prices have had a positive correlation with expected inflation since 2008, strongly suggesting that aggregate demand is at the heart of our problems.
I really like this kind of argument. For me, the case for a sticky-price, Keynes-Friedman view of the world rests crucially on some of the correlations one sees in the real world — between real and nominal exchange rates, for example. And this is a good modern example.
Krugman’s right that a big drop in aggregate demand caused the Great Recession. Of course that doesn’t entirely rule out Lehman playing a big role, as it might have reduced AD. But on close inspection, it was tight money that was the real problem. The early (and milder) phase of the Great Recession occurred between December
2017 2007 and June 2008. Real GDP leveled off during this 6-month period. The proximate cause was a sharp slowdown in NGDP growth, from 6.5% during the housing boom, to just below 3%:
The proximate cause of the slowing NGDP growth was a nearly complete stop in growth in the monetary base from July 2007 to May 2008 (this graph shows 12-month growth rates):
Base velocity actually increased during late 2007 and early 2008. The slowdown in NGDP was entirely caused by the Fed’s decision to stop printing additional currency during late 2007 and early 2008. (In those days the base was roughly 98% currency.)
After mid-2008, two other problems developed:
1. The Fed did not print enough base money to meet the rising demand for cash (falling velocity), which was caused by a fall in the opportunity cost of holding cash.
2. In early October 2008, the Fed began paying interest on bank reserves, further boosting the demand for base money.
If you are one of those people still unenlightened by market monetarism, someone who insists on thinking in terms of interest rates, note that during April-October 2008 the Fed held its target fed funds rate at 2%, even as the natural rate of interest fell sharply into negative territory. (See the graph below from a paper by Vasco Curdia, which shows a negative natural rate during mid-2008.) Thus in New Keynesian language, the severe phase of the Great Recession was caused by the Fed’s decision to hold their target rate at 2%, while the natural rate had plunged into negative territory.
Notice that this problem occurred well before Lehman. This doesn’t mean the failure of Lehman had no effect; it probably slightly reduced the global Wicksellian natural interest rate. But the Fed could have and should have offset that effect. Lehman was much more a symptom than cause of the Great Recession.
The cause of the recession was tight money leading to falling NGDP growth, which (because of sticky wages) led to falling RGDP growth.
I’ve included “concrete steppes” in this post because commenters always ask for them. But in truth, the concrete steps are beside the point. As Ben Bernanke pointed out in 2003, NGDP growth and inflation (not the money supply or interest rates) are the best indicators of the stance of monetary policy:
As I discussed earlier, for reasons of financial innovation and institutional change, the rate of money growth does not seem to be an adequate measure of the stance of monetary policy, and hence a stable monetary background for the economy cannot necessarily be identified with stable money growth. Nor are there other instruments of monetary policy whose behavior can be used unambiguously to judge this issue, as I have already noted. In particular, the fact that the Federal Reserve and other central banks actively manipulate their instrument interest rates is not necessarily inconsistent with their providing a stable monetary background, as that manipulation might be necessary to offset shocks that would otherwise endanger nominal stability.
Ultimately, it appears, one can check to see if an economy has a stable monetary background only by looking at macroeconomic indicators such as nominal GDP growth and inflation.
Note that if the Fed had engaged in 5% NGDPLT in 2008, then the natural rate of interest never would have fallen into negative territory.
BTW, I got my new California license plates yesterday. In an amazing coincidence, California chose a set of letters that exactly correspond to my policy preferences. I’ll take that as a sign that I’m on the right track.