A good economy doesn't crash. It goes bad first.

Via Mark Thoma, Claire Jones's How and why economics forgot Keynes’ warnings on panics at FT. Jones links to George Akerlof's 16-page PDF What They Were Thinking Then: The Consequences for Macroeconomics during the Past 60 Years. Akerlof's paper is a good read.

Akerlof, on financial crashes:

In the [early 1960s] push to assure acceptance of the dominant-paradigm Keynesian-neoclassical synthesis, a major macroeconomic question was left unresolved: what is the basic reason for very hard times that especially require fiscal or monetary stimulus? ... But in chapter 12 of The General Theory, on “the state of long-term expectation,” Keynes (1936) had suggested ... those bad times result largely from financial fragility. Famously, in that chapter, Keynes (1936, 140) analogized stock prices to the outcomes of a newspaper beauty competition.

Akerlof rejects the thought that the beauty contest analogy was just satire, and continues:

In concert with modern game theory, finance theorists have evolved a sophisticated understanding of why Keynes’s beauty contest is central to the theory of financial crash. A very simple game (adapted from Atkeson 2001) describes the skeleton of such models. In this game players have two choices: to continue to hold an asset or to sell it. With economic conditions sufficiently strong, the incentives to hold may be so strong that it pays to do so, even if everyone else sells. In this case, there is only one equilibrium: everyone holds, and the price of the asset remains high. With economic conditions at the opposite extreme, the incentives to hold may be so weak that it pays to sell, even if everyone else keeps it. Again, there is only one equilibrium: but this time everyone sells. But between these two extremes, in such a game, there is the possibility of an intermediate range with a threshold level of holders. If the number of holders exceeds the threshold, it pays to hold; if that number is below the threshold, it pays to sell. If such a model describes asset markets, financial equilibria are likely to be fragile...

Next, he offers three examples, among them the bank run:

In a bank run model (like that of Diamond and Dybvig 1983), if only the usual transactors are making withdrawals, there is not much reason to line up at the bank. But if others are lining up out of fears of its insolvency, there is reason to rush to be among those first in line to retrieve one’s deposit.

Like Keynes's beauty contest, a bank run depends less on what we think about conditions than on what we think other people think about conditions, and what we think they're thinking about what we're thinking. Tying this back to his topic, Akerlof points out that

such representation of financial markets had no substantial place in mainstream macroeconomics in the period leading up to the Great Recession.

Akerlof's paper is interesting, and quite convincing. Recommended reading.


The focus of economists on the major but unresolved macroeconomic question, "what is the basic reason for very hard times", is the wrong focus. A better question, and the answer too, arise from Akerlof's description of the game adapted from Atkeson:

• "With economic conditions sufficiently strong ..."
• "With economic conditions at the opposite extreme ..."
• And "between these two extremes".

What people do depends on economic conditions. When the economy is good, no one is thinking about bank runs. When the economy is bad, everyone is. If you keep the economy good, you don't have to worry about the basic reason for very hard times, because the very hard times never occur.

But here anyway is your "basic reason": Very hard times only occur as an outgrowth of plain old hard times. If you keep the economy good, you don't have hard times, so you don't get the very hard times. It's like having your cake, and then having some more.

You gonna tell me we don't know how to make the good times permanent? Well then, that's the question you need to work on, isn't it, instead of asking "what is the basic reason for very hard times?" This should be obvious, but apparently it isn't.


I wrote the above, then went looking for “Comment on Morris and Shin’s ‘Rethinking Multiple Equilibria in Macroeconomic Modeling.’”, the Atkeson PDF that Akerlof mentions. The Atkeson PDF led me to the Morris and Shin PDF.

I said:

What people do depends on economic conditions.

Morris and Shin open with this thought:

It is a commonplace that actions are motivated by beliefs, and so economic outcomes are influenced by the beliefs of individuals in the economy.

I go directly from economic conditions to economic actions. Morris and Shin go from beliefs about the economy to economic actions. Is there a difference between economic conditions and what people believe those conditions to be? Yes: It is possible that our beliefs are wrong. So the first task is to understand the economy, to get our beliefs in tune with actual conditions. Our second task, then, is to get back to the good times and stay there.

Of course, if you've been down so long it looks like up, you probably don't believe we can get back to good times. Okay. But maybe -- just maybe --that's an example of holding the wrong beliefs. Or, maybe I'm wrong. But the only way to tell is to understand the economy.


Akerlof contrasts "modern macroeconomic models and the models of financial crash", finding modern macro models inadequate. Good. But I think models of financial crash are also inadequate. You can't wait until there's a crash before you solve the problem. You have to catch the financial crash in its formative stages, years (or even decades) before the warning signs are obvious.