Ask a Trained Parrot
Betsey Stevenson and Justin Wolfers have long been among my favorite social scientists thanks to their meticulous, incisive work about economics and the family. Their paper “Bargaining in the Shadow of the Law: Divorce Laws and Family Distress” is one of the most horrifying pieces of non-fiction I’ve ever read. I can’t believe anyone who reads it will ever think about divorce the same way again. And “Marriage and Divorce: Changes and Their Driving Forces” is the sort of work family policy needs more of, rather than the instincts, gut feelings, and greeting-card sentimentality it usually runs on.
But I think we may have had quite enough family policy here lately — not just same-sex marriage, but child abuse, abstinence, population growth, and all the rest — so I’d like to talk about something different… also by Justin Wolfers.
In this NPR Marketplace segment, Wolfers talks about the abject failure of macroeconomic modeling, as studied by two of his students. Simply put, models don’t work. The old ones don’t work. The new ones don’t work. Computers don’t help. Unlike weather forecasting, macroeconomics hasn’t made any meaningful progress at modeling the future. Macroeconomists make big bucks and get a lot of respect, but their predictions aren’t statistically any better than those of a trained parrot.
Which raises some interesting questions: First, why not just find some trained parrots? What do the modelers get us that the parrots don’t?
The answer here I think is simple. Much of what government does is about signaling and allocating status, not about doing any actual good. But to get the process started, we have to signal the professionalism and the integrity of the government — its objectivity; its science. If you can’t do that, then it really is just transparently playing favorites. And if there is any profession that exists solely to achieve the task of enhancing government credibility, macroeconomics might just be it. (Social workers are a close contender, for those still hankering after family policy.)
Second, we have to signal that certain groups are valued and loved by the government — farmers, auto companies, large foreign financial institutions, Pakistan — and here we hit a problem. How do we do it? Politically, these things can be justified only in certain very limited ways. Thanks to our political culture’s lingering classical republicanism, one can’t very well say “Oh here, GM, we’re doing this because we love you and want you to feel valued,” even if that’s precisely what’s going on. No, we have to tie our favoritism, however structured, to some claim about the general welfare.
Enter the macroeconomists. Anthropologically, they act as soothsayers, interpreting the random walk of experienced history (and experienced favoritism) as part of an eternal cycle, only dimly seen, of transcendent justice — understood, à la Rawls, as fairness, or à l’Hayek, as conforming with abstract, impersonal rules. The latter being harder, but only by a little.
Much of what government does is about signaling and allocating status, not about doing any actual good. But to get the process started, we have to signal the professionalism and the integrity of the government — its objectivity; its science. If you can’t do that, then it really is just transparently playing favorites. And if there is any profession that exists solely to achieve the task of enhancing government credibility, macroeconomics might just be it. (Social workers are a close contender, for those still hankering after family policy.)
You just arrive here from no place. If the point of the post is just more about the perfidy of government and how it co-opts a (struggling) academic profession, great. But you say the post is about macroeconomics. Last time I checked, Goldman Sachs employs a number of macroeconomists, and pays them handsomely. You don’t seem to even be attempting to treat your subject in anything like completeness, even for a blog post. Your contention is really that macroeconomics exists only to provide a fig leaf for government’s acts of favortism in the market?Report
@Michael Drew, To be clear, I don’t deny this function. But it’s a fact in the world to be demonstrated, not just assumed. And the language here suggests this is the exclusive reason macro exists, and that’s extreme.Report
@Michael Drew,
I wouldn’t say I arrive at it from nowhere. I owe a debt here to Robin Hanson, and I thought it was such an obvious debt that I didn’t need to acknowledge it.
Do I think macroeconomics exists only to provide a fig leaf for favoritism? Not really. It was created to try and solve many vexing and important theoretical questions.
Do I think that providing a fig leaf for favoritism is nearly all of what it accomplishes in practice? Absolutely.Report
@Jason Kuznicki, Not from nowhere in terms of your private study. From nowhere in terms of this post. Show don’t tell!Report
@Michael Drew,
Try here and here. Also, here’s a podcast, which I recall being informative. I could be wrong about that one though, I haven’t listened to it in a while.Report
From what I read of the paper (and most of it is way over my head), the modelling fails in large part because it gives such an accurate picture of the present as to change policy.
I came up with this from the following quote taken from the Abstract: “In the 1997-2006 period, in particular, in?flation has been unforecastable by any method but this is in line with the standard New Keynesian model prediction that infl?ation will not be systematically varying when monetary policy responds to expected in?flation aggressively.”
If anticipated inflation is essentially zero, then the fluctuations around the zero point will be random (and thus impossible to predict). But if the policy goal is to keep inflation at zero, then the policy response will be predictable and non-random.
So the better question is to what purpose are these models being put? Is it to pull out the necessary data, as to drive a short-term policy response? It appears that the answer is yes, and that the models are useful for that purpose.
Given the Great Recession, the even better question is to ask where are the larger scale models that would have persuaded policy makers (and U Chicago macroeconomists) of the existence of the bubble.
Perhaps hoping for such things is naive; as we have seen in the debate over AGW, any model that results in an unpleasant policy will be attacked as unproveable, biased or “only a model”.Report