Steven Levitt, famous for his Freakanomics, shows that being an economist is not just mouthing supply and demand.

Anyone can call themselves an ‘economist’. Many do, despite having no qualifications in economics and hardly any formal training; they often make elementary errors. That is the result of a conscious decision of the economics profession which resists barriers to exit and entry. In contrast other professions have restrictions, often for good reasons; I am glad my medical advisers are not only qualified but also registered. However, claiming to be expert on economics to contribute to the public commentary without any expertise, is confusing to those with more humble understandings.

On the other hand there are those who have a high reputation in the economics profession but who don’t seem to be really economists. Consider Steven Levitt, who has written of himself, ‘I am having trouble mastering the tools of my own profession. If you ask my students whether I know calculus, they will say “not very well”. I’m not proud of the fact, but I’m a realist. If you ask the really great economic thinkers like Gary Becker and Kevin Murphy how often I’m right when I try to apply Chicago price theory, they will simply tell you that I am showing a lot of improvement, because they are kind.’

And yet in 2003 Levitt won the John Bates Clark Medal, awarded every two years by the American Economic Association to the most promising US economist under the age of 40. Today he is a Distinguished Service Professor of Economics at the University of Chicago.

He says, ‘the only thing I am good at, really and honestly, is asking questions which people seem to find interesting, and figuring out how to trick the data into answering those questions.’ (I explain what he means by ‘trick’ below.)

His most famous and controversial finding, with John Donohue, is that the Roe-Wade Supreme Court decision which legalised abortions resulted in lower crime rates a couple of decades later. One issue was whether the statistical analysis was correct. A lively debate led to some further analysis which strengthened the conclusion.

Levitt’s claim that he ‘tricks the data’ is misleading. He does not torture it until it confesses, nor does he confuse correlation with causation. Rather, he uses advanced statistical tests. A key one in this study was that some US states increased access to abortion some years before the Supreme Court decision legalising it and their decline in the crime rate was earlier. States with higher abortion rates had greater falls in their crime rate.

Of course there are those who claim the research finding supports increased accessibility to abortion. Actually it does not: research rarely leads to a policy conclusion on its own. For instance, someone opposed to abortion might conclude that the research suggests that those who have abortions which lead to criminal behaviour are living in adverse circumstances (it seems likely that the women involved take these circumstances into consideration when making their decision). The policy conclusion may well be that the state should provide more support to those who do not have abortions. (One of the research findings is that abortions do not reduce the birth rate in the long run; rather, they lead to mothers delaying the birth to a better stage of their life cycle.) Such cerebration has not been prominent in the anti-abortion lobby, which is sad.

Levitt makes a point about race and crime:

It is true that, on average, crime involvement in the US is higher among blacks than whites. Importantly, however, once you control for income, the likelihood of growing up in a female-headed household, having a teenage mother, and how urban the environment is, the importance of race disappears for all crimes except homicide. (...the homicide gap is partly explained by crack markets.) In other words, for most crimes, a white person and a black person who grow up next door to each other with similar incomes and the same family structure would be predicted to have the same crime involvement.

One wonders whether that is true for Māori and Pakeha? I don’t think anybody has done the statistical comparison. It seems likely that if the comparison of crime rates were adjusted for socioeconomic differences the gap between Māori and Pakeha would be markedly reduced, but would it be eliminated? If it were, it would make various (woolly) explanations such as ‘racism’ and ‘colonialism’ redundant.

Instructively, almost every study Levitt has done would not be classified as ‘economic’ and would not be referred to in a standard economics text. The list includes:

Abortion;

Automobile anti-theft devices;

Cheating by teachers in schools;

Cheating in sumo wrestling;

Criminal age;

Discrimination in game shows;

Drunk driving and accident rates;

Effects of electoral campaign spending;

Federal spending benefits congressional incumbents;

Finances of a drug gang;

Having a distinctively black name is primarily a consequence rather than a cause of poverty and segregation;

Penalty kicks in soccer;

Police hiring;

Politics – the median voter theorem;

Prison populations;

Ridesharing.

So is Steven Levitt an economist? Clearly, high-powered economists think he is. What makes him one, I think, is the quality and thoroughness of his empirical investigations, including a willingness to review a study when someone comes with a criticism of it. What makes him an attractive one is his humility. I bet you he is much better at calculus and price theory than he claims.

A properly trained economist will have done some ‘econometrics’, that is, the training in empirical investigation. Many who claim to be economists would not meet that hurdle, which is why their opinions are frequently detached from reality.

Freakonomics: A Rogue Economist Explores the Hidden Side of Everything by economist Steven Levitt and journalist Stephen J. Dubner.