"Golden Oldies (Wonkish)"

Paul Krugman:

Golden Oldies (Wonkish), by Paul Krugman: Mark Thoma jokes that “I’ve learned that new economic thinking means reading old books.” And Brad DeLong says that it’s no joke: the macroeconomics of 1936 — or even of 1873 — is a much better guide to current events and their policy implications than all the macro theory of the past few decades.

All this is very much in line with what I’ve been saying about a Dark Age of macroeconomics. But what happened, really? A few more thoughts.

Economics is basically about incentives and interaction — or, as Schelling put it, micromotives and macrobehavior. You try to think about what people will do in certain circumstances, and you try to understand how individual behavior adds up to an overall result.

What economists have known since Bagehot (with regard to financial markets) and since Keynes (with regard to goods and labor markets) is that under some circumstances seemingly reasonable individual behavior adds up to very unreasonable macro outcomes. ...

But notice that I’ve framed this in terms of “reasonable” behavior; it’s a lot harder to tell these stories in terms of perfectly rational, maximizing behavior.

One response — a pretty good response — is, “So?” After all, maximization isn’t a fact about human behavior, it’s a gadget — an assumption we use to cut through the complexities of psychology and all that, one that can be very useful if it clarifies your thought, but by no means an axiom or a law of nature.

But maximizing models have a special appeal for modern academic economists: they require solving equations! They’re rigorous! They make it easy to show that you’re doing “real research”. And so maximization tends to acquire a bigger importance in economic thought than it deserves.

To be fair, applying maximizing thinking has achieved some major successes even in macroeconomics. ...

But from the 1970s onwards, what happened was that the drive to base everything on maximizing behavior narrowed the profession’s thinking — and, crucially, led first to a de-emphasis, then to a total forgetting, of the great insights about interaction. We created an economics profession which believed that Keynesian economics, and for that matter Bagehotian finance, had been “proved wrong”; whereas all that had really happened was that those things proved hard to model in terms of perfectly rational maximizing agents. Again, so?

And there’s a sense in which even New Keynesian economics was wasted effort, at least from a social point of view, because it was mainly a way of showing New Classical types that we can too ground the concepts we already knew in maximizing models. Actually, I don’t think that’s entirely fair: I find that New Keynesian models, especially on the liquidity trap issue, do deepen my understanding. Still, you can understand why Larry Summers says that none of that stuff proved useful in actual policymaking.

The point, though, is that something went terribly wrong. Put it this way: if all we had known when this crisis struck was 1950-vintage macroeconomics, we would probably have done a better job of responding.

Someone asked me how I came to embrace the need for new economic thinking. The crisis had a loot to do with it. Before that, I was pretty defensive about economics and what it had to offer.

When the crisis hit, I felt I had to say something about what happened, and what we needed to do. But when I reached into the bag of models that had been popular over the last few decades, very few had much to offer, and the standard models were all but useless.

I can remember feeling guilty that I had been teaching my graduate students the wrong things all these years. They needed to know the latest models and latest technical tricks, so I don't mean it was a wasted effort to teach those things, but it wasn't augmented with the things from the past that would help us to understand what happened. Names like Minsky, Kindleberger, and Bagehot to name just a few were nowhere to be found, and because of that the ability to place what was happening in the correct historical context, to understand and evaluate what had been done in the past -- what could we learn and what mistakes should we avoid? But I hadn't introduced my students to any of this, and to be honest in the ten weeks I have with graduate students, every moment is crammed with the mathematics they'll need to do the dynamic optimization discussed above, and there is very little room to include it. So I think separate courses are needed, and as I said yesterday, they should include both history of economic thought and the economic history of the US and other countries (fitting these courses into the curriculum and finding a way to pay for the requisite faculty are big hurdles to doing this).

But here's the thing. The students have to understand that these courses are a crucial part of their education, not something only of interest to historians, and that will require a change in both the sociology and culture of the profession.