The American Spectator

Claiming that Glaeser and Sacerdote find that taller buildings cause increased crime is a step further than they themselves were willing to take in their paper “The Social Consequences of Housing.” The study is a cross-section controlling for observed characteristics of neighborhoods, and so although the authors can isolate the correlation between the heights of buildings and the level of crime in neighborhoods, it seems like a stretch to confirm the enticing but unlikely claim that taller buildings cause an increase in crime.

Harford not only makes that leap of logic but he also uses it in a passage incorporating anecdotes about two women attacked in different neighborhoods as well as race statistics about the UK — a synthesis which seems to run roughshod over the assumptions underlying Glaeser and Sacerdote’s paper.

IN ANOTHER EXAMPLE of his rush to shoehorn empirical findings into fun facts, Harford makes a mistake the borders on the costly blunder Steven Levitt and Stephen Dubner made in Freakonomics’ most famous passage.

Soon after publishing his now-famous study arguing that the legalization of abortion led to lowered crime in the 1990s, Levitt recommended that abortions were beneficial. A wave of counter-studies ensued, along with outrage from many pro-life advocates. In the final printing, Levitt omitted many of those normative claims, and further studies have proved that Levitt’s data was fatally flawed.

And now, in the chapter entitled “Is Divorce Underrated,” Harford argues that, from an economic perspective, there is a socially optimal level of divorce, and that it would signal some women’s disutility if the divorce rate fell.

One study he uses to support this claim is based on a speed-dating experiment run by Michele Belot and Marco Francesconi. The two economists find that the quantity of numbers exchanged during a speed-dating event doesn’t depend on the quality of the participants.

Thus people aren’t in fact holding out for their “one true love” — despite what they might say. Harford uses this finding as part of his argument in favor of divorce: if people don’t have one and only one significant other, they shouldn’t be bound to their spouses permanently.

Belot and Francesconi’s study, however, is limited in its conclusions because of the weakness of using data from a dating service for which the participants to register and pay a fee. Aren’t the people most likely to go speed-dating also the most likely to have lower ideals of love?

No Romeos or Juliets are likely to register for something as shallow as a speed date.

THESE ARE MERELY two examples of slight misinterpretations in a book crammed full of intricate studies, yet it is not nitpicking to mention them. They encapsulate the weaknesses that characterize pop-economic books in general and The Logic of Life in particular.

Ignoring details undermines econometrics’ predictive power and reduces its conclusions to mere opinions. Harford wants to portray economics as a brutally rational and value-free way of thought, but he thwarts himself by including examples that incorporate faulty assumptions.

Taking a page out of his book, though, I could explain Harford’s oversight using rational choice theory: It makes sense for Harford to sacrifice some precision to maintain the breezy attitude with which he sheds light on racism, the allure of cities, etc. His aloofness while giving immediately obvious explanations for phenomena once thought of as life’s mysteries leaves the reader feeling smart enough to solve all the world’s problems using the economist’s tools.

Harford states his aim in the introduction: “…it is a world in which people can generally be expected to make rational decisions, and where those rational decisions suggest some astonishing explanations for many of life’s mysteries. It is this world that I would like to show to you.” He accomplishes this goal handily. The reader comes away impressed with the brute power of choice.

But in economics, for every good the costs must be considered along with the benefits. I would suggest that Harford consider the costs of drawing too many questionable conclusions from tentative research — for how it will affect those who can’t tell the difference.