I'm in two minds about Frances' critique of macroeconomics. A bit of me disagrees, but a bit thinks she might be understating the problem.
First, a disagreement. In one respect, the 2008 crisis actually strengthened economics. The investor who followed the sell in May rule, or who used foreign buying of US equities as a sell signal, would have been out of equities before the crash of 2008. In this sense, the financial crisis actually vindicated the claim that economists can fulfill the useful function of warning of trouble ahead.
Of course, these two guides told us nothing about the causes of the crisis. But we must remember Jon Elster's words:
Sometimes we can explain without being able to predict, and sometimes predict without being able to explain. (Nuts and Bolts for the Social Sciences, p8)
I also disagree with this:
The failure of most macroeconomists to foresee the financial crisis grew out of their incorrect understanding of how money is created, and perhaps more importantly, how leverage builds up.
However, most banks didn't fail merely because their debts turned bad. They did so because of bad takeovers (eg RBS-Amro or Lloyds-HBOS) and/or a reliance upon wholesale finance. These were not so much macroeconomic failures as micro ones - bad decisions by individual businesses*. As the TSC said (pdf):
The origins of the banking crisis were many and varied, including low real interest rates, a search for yield, apparent excess liquidity and a misplaced faith in financial innovation. These ingredients combined to create an environment rich in over-confidence, over-optimism and the stifling of contrary opinions.
In this sense, the bank crisis was an example of what Xavier Gabaix called the "granular" origin of aggregate fluctuations. Recessions can result from the failure of individual firms.
What matters here is the structure of economic networks: see for example this paper by Daron Acemoglu and these papers in the JEP. Bank failures matter because banks are key hubs, in three different senses:
- Banks emulate each others' strategies. As Frances says, they herd. The failure of one is therefore likely to be correlated with the failure of others.
- One bank's failure will have a chilling effect on others; they'll try to hoard cash by withdrawing credit.
- Non-financial firms can't easily switch away from bank finance. As an empirical matter, private equity, P2P lending, crowdfunding and suchlike were not sufficient to replace lost bank credit. For this reason, the collapse of RBS was not like the demise of Woolworths. People could easily get their pick n mix elsewhere; they couldn't so easily get finance elsewhere. For this reason the collapse of banks had horrible macroeconomic effects.
It's for this reason that I say Frances understates the problem. The flaw with representative agent models isn't that they ignored the financial sector or that they forgot that "expectations are driven as much by emotion as logic." These problems are fixable. The problem is that recessions sometimes arise from interactions between firms and from firms position within networks; the failure of a hub matters whereas the failure of a spoke doesn't. Representative agent models, by definition, omit this. To this extent, we need a different paradigm. And perhaps this paradigm will only ever allow us to understand events after the fact rather than predict them.
We must, however, remember something else. Even if we had a fantastic macroeconomic theory, there is no guarantee whatsoever that governments will use it for policy purposes. As Simon says, we have a perfectly good theory of what fiscal policy should be at the ZLB - and it is being ignored.
Explanation and prediction are two different things. And good policy is a third different thing.
* Caveat: a current account deficit means that domestic investment exceeds domestic saving. This implies that domestic banks' lending might well be rising faster than deposits, implying a reliance upon other sources of funding.