Brad DeLong, Tyler Cowen, and Justin Fox all make much of Christina Romer's expertise on the causes of the Great Depression. And, indeed, Romer does seem to be an expert on the Depression. She even wrote the Encyclopedia Brittanica entry on the subject. This is a point you frequently hear in Ben Bernanke's defense, too: He's also an expert on the Great Depression. But can this really be so useful? It's hard to believe that a complex financial crisis in 2008 is so similar to a complex economic crisis in 1929 that you need an incredibly subtle understanding of the latter to effectively apprehend the former. Presumably most economists know enough not to repeat the basic mistakes of the 1930s, and the question is which economists know enough to avoid the possible mistakes of the 2000s. Is there any real reason to assume otherwise?

Eric Rauchway at blog edgeoftheamericanwest:

Granting that I have a vested interest in the position that knowledge of history is relevant, here’s the specific case I’d make on behalf of Romer (and Bernanke) and then the general case.

(1) The Specific Case. The Great Depression remains unique in degree and perhaps in kind. We want to keep it that way. Inasmuch as the present crisis bears some resemblance to the early phases of the Great Depression we want experts of that unique catastrophe to be on the lookout for ways to halt the slide into disaster.

Specifically, Bernanke and Romer have studied how the financial crisis of the late 1920s turned into the broader economic crisis of the early 1930s. It looks like this is where we are now, with the financial crisis beginning to affect consumer spending.

People generally know that this happened, but Bernanke and Romer have specific ideas about why, and are familiar with the historical data supporting those theses. Assuming they have pretty sharp minds they should be able to spot related trends in the current data and formulate or recommend policy accordingly. Which leads to …

(2) The General Case. Ezra’s concluding question seems to boil down to, now that we have the social science, why do we need the social scientists?

Well, to some degree you don’t, of course. The whole point of social science is to draw general conclusions from specific data, and those general conclusions should be able to predict further findings. So Bernanke and Romer sift through the data, and produce a model that explains what happens, and the model should have predictive value no matter who applies it.

But that “incredibly subtle understanding”, which comes from poring over the data, is potentially quite valuable, especially in a complex crisis that may not unfold exactly like those of the past. Someone who’s familiar not only with the model—the “basic mistakes”—but also with the data that went into the model will know all of the caveats and qualifications that don’t make it into the general statement.

And such a person would be the best qualified person to catch exceptions and nuances in current flows of information, so as to say, “wait. We don’t want slavishly to follow the model, here, because….”

For the same reason, I hope the people crafting the president-elect’s infrastructure-investment program have a good empirical understanding of how the New Deal infrastructure-investment programs did and didn’t work. If you want to do more than derive the basic benefit of just paying people to dig holes and fill them again—that is to say, if you want lasting infrastructure as well as short-term stimulus—you need to take into account the weirdnesses of American federalism, the business of contracting and hiring, and the existing state of plans for infrastructure development. It would be good to know the details of how that panned out last time.