24/01/2008

Unintended Consequences

I think it is great that Stephen Dubner and Steven Levitt use their latest column to bring to the attention of the public one of the most important concepts of the social sciences: unintended consequences. It is less great that they make a mess of it. Two problems:

One, they confound - as is common - three aspects: unintended, unwanted and unanticipated. All of the examples in their article combine these three features. This is understandable, as these examples are the most interesting. Yet these features do not always go together.

A consequence may be unintended yet wanted. My favourite example is legislation on helmet use for motorcyclists. In the 1970s many countries legislated that it is illegal to ride a motorcycle without a helmet. This was meant to reduce serious injuries but was also followed by a stark decline in motorcycle thefts (presumably because many motorcycle thefts used to be spontaneous, so that potential thieves did not carry helmets). It is probably safe to say that most politicians who voted for this regulation did not do so to reduce theft, yet welcomed the consequence. Other consequences an action has may be seen as neutral - I don't care about the newsagent, yet my buying at his place makes him more wealthy.

A consequence may also be unintended yet anticipated. For example, I might not like traffic congestion in the central city, and I might fully anticipate that my taking the car to go there contributes to congestion, yet I may chose to do so.

Two, they talk of a law of unintended consequences, but there is no law that says that any or most actions have unintended consequences. When I scratch my head, that does the job, and that's it.

I'm a big fan of the Freakonomics blog and the NYT columns, but in this case, it's back to the drawing board.

(Note: This post is an extended version of the comment I left at this MR post, which in turn was inspired by this post from Andrew Gelman, who sort of made the same point as point two above.)

Addendum: Bryan Caplan notes another positive unintended consequence.

No comments: