Did Banning Lead Lower Crime?

The major news media (see for instance here and here) have been reporting recently on the hypothesis that banning lead from gasoline caused a reduction in crime. This follows a similar article in the Washington Post a few months ago, which I blogged about at the time.

Since a lot of people have written to us lately asking for comment about the lead/crime connection, I’m going to go ahead and reprint here what I wrote back in July:

————————————————————————————

Over the weekend, the Washington Post published an article suggesting that much of the decline in crime in the 1990s may have been due to the reduction of childhood lead exposure after the removal of lead from gasoline and house paint.

This is an intriguing hypothesis. There is evidence on an individual level that high exposure to lead is harmful to both IQ and the ability to delay gratification, two traits that could enhance the attractiveness of crime. There is also some suggestive time-series evidence of a relationship: the rise and fall in lead exposure at the national level match the rise and fall in crime. Still, although both Post reporter Shankar Vendantam and the cited economist, Rick Nevin (whom I’d never heard of), appear quite convinced by the time-series data, I am not. When you have a variable like crime that goes up for a long time then goes down for a long time, it is easy to find other variables that share that pattern and appear to have a causal impact, even though the relationship is completely spurious.

About seven years ago, Michael Greenstone and I tried to look into this same issue using airborne lead measures at the local level, as well as other approaches. We ultimately gave up without finding anything. That largely soured me on the lead/crime link.

Recently, however, Jessica Wolpaw Reyes at Amherst has put together what appears to me to be the most persuasive evidence to date in favor of a relationship between lead and crime. Rather than looking at a national time-series, she tries to exploit differences in the rates at which lead was removed from gasoline across states. I haven’t read her paper with the care that a referee would at an academic journal; but, at least superficially, what she is doing looks pretty sensible. She finds that lead has big effects (and, for what it’s worth, she also confirms that, when controlling for lead, the link between abortion and crime is as strong or stronger as in our initial study, which did not control for lead.)

Roger Masters, a professor at Dartmouth, has also been doing interesting research on this subject, although I am also not very familiar with his work.

It will be very interesting to see how this research agenda plays out. If it can be shown here and in other areas that environmental factors have powerful and long-lasting impacts on human behavior, it may dramatically change the way we think about public policy.