John DiNardo hates Freakonomics

Freakonomics by Steven Levitt and Stephen J. Dubner is an extremely popular book that has made economics a (somewhat) sexy topic of discussion. Levitt’s research makes economics exciting and his quirky, controversial studies make interesting reading.

John DiNardo, however, thinks that even Freakonomics is “interesting” and “entertaining,” it may not be revealing truths. Dr. DiNardo has written three critical reviews of the book. DiNardo’s criticisms call into doubt the meaning of some of the conclusions derived from Levitt’s research. For instance, DiNardo discusses the logical meaning the causal effect of obesity on health.

Nonetheless, I would argue that it is unlikely that anyone will devise a severe test of the proposition that obesity causes an increase in all-cause mortality. Simply put, the effect of obesity (or of ideal weight) is inextricably implementation specific. That is, it is not helpful to think about the “effect” of obesity for the same reason it is not helpful to debate the “causal effect of race on income”(Granger 1986). Many of us suspect, for example, that encouraging obese individuals to “starve themselves” for short periods of time might help one lose weight, but wouldn’t necessarily promote longevity (although it might, who knows? ).

Similarly, we might expect weight loss that results from increased physical activity to be more protective than
weight loss that results from increased life stress. The experience in the U.S. with the drugs fenfluramine and dexfenfluramine (Redux) is a case in point. Despite good evidence that the causal effect of taking Redux was weight loss, the drugs were pulled from the market because a “side effect” of the medication was an increase in potentially serious heart problems (Food and [Drug] Administration 1997) . Indeed, it would appear that the presumption that obesity is a cause of ill health made it virtually impossible to debate whether non–obesity was the cause of the increased heart problems. Rather, the consensus seems to be that the heart problems were not caused by non–obesity, but rather by Redux’s “side effects.”

My point is simple: when each way of “assigning” obesity that we can imagine would be expected to produce a different effect on all–cause mortality or other outcome, it is not at all clear that it is helpful to debate the “effect of obesity.” It seems more intelligible (and more policy relevant) to discuss the effect of Redux or exercise than it is to talk about the “effect” of obesity.

One study that DiNardo does hold up as an example of fine research is Cullen, Jacob and Levitt (Econometrica 2006). This paper was written by Levitt as well as my dissertation advisor Julie Cullen.

• John Dinardo (2007). “Interesting Questions in Freakonomics .” Journal of Economic Literature, Volume 45, Number 4, pp. 973-1000(28)
• Julie Berry Cullen, Brian A Jacob, Steven Levitt (2006) “The Effect of School Choice on Participants: Evidence from Randomized Lotteries.” Econometrica 74 (5) , 1191–1230 doi:10.1111/j.1468-0262.2006.00702.x