INTERPRETING THE CAUSES OF THE GREAT RECESSION OF 2008 By Joseph E. Stiglitz1
The Great Recession of 2008 is both complex and simple. In some ways, beneath the complexity of CDS’s, sub-prime mortgages, CDO’s, and a host of new terms that have entered the lexicon is a run-of-the-mill credit cycle. As banks lent money freely on the basis of collateral, prices increased, allowing more and more lending. Real estate bubbles are a dime a dozen. Bubbles break, and when they break, they bring havoc in their wake. Perhaps the most unusual aspect of this bubble was the conviction of key policymakers (including two Chairmen of the Federal Reserve) that there was no bubble (perhaps a little froth), and the bald assertions (a) that one could not tell a bubble until it broke; (b) that the Fed didn’t have the instruments to deflate the bubble, without doing untold damage to the economy; and (c) that it would be less expensive to clean up the mess after it broke than to take preventive action. These assertions were made presumably on the basis of the “accepted” wisdom of the economic profession. Such views were reinforced by the belief in rational expectations and the belief that with rational expectations there couldn’t be bubbles. Few would hold to these views today. But even before the crisis there was little basis for these beliefs. Brunnermeier (2001) had shown that one could have bubbles with rational expectations (so long as individuals’ have different information).2 Decades ago, economists had shown that there could be dynamics consistent with capital market equilibrium (rational expectations, with the no-arbitrage condition being satisfied across different assets) for arbitrarily far into the future, but not converging to the long run “steady state,” so long as there were not futures markets extending infinitely far into the future.3 Such paths look very much like “bubbles.” There has been, in addition, a large literature on rational herding. Standard results on the stability of market equilibrium with rational expectations employed representative agent models with infinitely lived individuals (where the transversality condition replaced the necessity of having futures markets extending infinitely far into the future). But as soon as the assumption of infinitely lived individuals was dropped, there was no assurance of convergence; the economy could oscillate infinitely, neither converging nor diverging.4 Other models in the same vein emphasized the possibility of multiple rational expectations equilibria.5 These may seem theoretical niceties, but to the extent that the belief that markets were efficient, and that efficient markets precluded the possibility of a bubble, they gave confidence to the Fed’s ignoring mounting evidence that there was a bubble and are thus much more than that. From a more practical perspective, though one might not be sure that there was a bubble, surely a policy maker should ask the question if it is possible, or even likely. All decision making is made under uncertainty. Policymakers need to balance the risks: historical experience should 1
Lecture to have been delivered at BIS Conference, Basel, June, 2009.
2
Brunnermeier (2001)
3
See Hahn (1966) and Shell‐Stiglitz (1967).
4
See, for instance, Stiglitz (1973, 2008).
5
See Cass and Shell (1983) and the large literature on “sunspot” equilibrium; see also Hoff and Stiglitz (2001).