Lead and Crime

Kevin Drum over at Mother Jones has stirred up a bit of a storm with his most recent article on how exposure to lead was the leading cause of the high crime rates in the late 70's, 80's and early 90's. As he writes,

Needless to say, not every child exposed to lead is destined for a life of crime. Everyone over the age of 40 was probably exposed to too much lead during childhood, and most of us suffered nothing more than a few points of IQ loss. But there were plenty of kids already on the margin, and millions of those kids were pushed over the edge from being merely slow or disruptive to becoming part of a nationwide epidemic of violent crime. Once you understand that, it all becomes blindingly obvious. Of course massive lead exposure among children of the postwar era led to larger numbers of violent criminals in the '60s and beyond. And of course when that lead was removed in the '70s and '80s, the children of that generation lost those artificially heightened violent tendencies.

This argument is supported Karl Smith, and I am a particularly big fan of his work over at Modeled Behavior. Drum elaborates on his simple rule of thumb for understanding epidemic-like phenomena in economics. "If it spreads along lines of communication, [...] the cause is information," Drum writes. "Think Bieber Fever. If it travels along major transportation routes, the cause is microbial. Think influenza. If it spreads out like a fan, the cause is an insect. Think malaria. But if it's everywhere, all at once—as both the rise of crime in the '60s and '70s and the fall of crime in the '90s seemed to be—the cause is a molecule."

Despite my high opinion of Smith in particular, I remain skeptical about Occam's Razor-style arguments applied to social phenomena. People and societies are just too damn complex to have one factor explain something like crime rates. When looking at crime, it's clear that also sorts of factors, including policing, drug laws, the state of the economy and the availability of family planning, contribute. Some obviously play a stronger role than others, but all should be included in an explanation of an effect. Economic models are supposed to parsimonious, but under-fitting (including too few variables) is just as big as a problem as over-fitting (including too many). 

This is a criticism that Drum himself is willing to acknowledge, in part. While some of the pull-out text in the article states that as much as 90 percent of the rise and fall of violent crime over the last half century can be explained by lead, this can easily be misinterpreted. As Drum writes, "Even if the 90 percent number is correct, it doesn't imply that lead is responsible for 90 percent of all crime. It only implies that it's responsible for 90 percent of the postwar rise of crime above its natural level, which is determined by a variety of other factors. Later, the drop in lead emissions was responsible for 90 percent of the decline of crime back to its natural level."

In other words, there is a least a substantial portion of the crime rate that could not be explained by lead alone, and Drum is only trying to explain the "crime bubble" that occurred in the US at the end of the 20th century. While this is still a more reasonable suggestion, it turns out that there's a lot of questions concerning the statistical treatments of lead and crime that remain unsolved. Take crime among 14-17 year-olds, for example.

As Steve Sailer explains the preceding chart,*

When looking at the crime fall in the 1990s, this study appears to have the same flaw that dragged down Levitt's abortion-cut-crime theory -- a failure to look carefully at crime rates by age cohort, combined with an intoxication with analyzing complex state-level data that leads to a failure to do simple national-level reality checks. (She was clearly influenced by Levitt, so this shortcoming is not surprising.) Wolpaw Reyes simply assumes a 22 year lag between lead poisoning around the time of birth and the violent crime rate. But, we can easily look at more detailed data for different age cohorts, which shows that the crime decline of the 1990s began among older individuals, not among the younger people supposedly benefiting from lower lead or higher abortion.

If you assume a 22 year lag to violent crime, then this graph looks great because the murder rate started to fall after about 1991.

But, that's the same mistake Levitt made way back in 1999: he forgot to look at the crime rates among narrower age cohorts. For the 17-and-under crowd, the two worst years were 1993-94. In other words, they were born when lead pollution had already fallen by almost half (just as they were born when legal abortion was close to its peak, which is a problem for Levitt's theory).

That said, I could find very little stuff out there refuting this theory, other than Sailer's criticism of the statistics. Instead, I mostly found articles supporting it. Discovery wrote about it back in 2008, and The Washington Post covered all of this back in 2007. While not quite as bullish on the crime-lead link as Drum, they nonetheless reach the same conclusion: although we do not fully understand the mechanism for this, there is obviously a strong correlation between lead exposure and crime rates.

Even if Drum is overreaching here, I'm willing to forgive him since the general message is right: lead is really bad for you. And it might take a bit of an overreaction before anything is done to clear up the many lead-contaminated sites left all over the US. As Deborah Blum points out in the Knight Journalism Blog at MIT, that's the most important point that needs to be made:

In other words, it's not that tetraethyl lead is some special form of the poison - the point is that lead exposure in any form is dangerous. And that's the real message here. The connection with crime remains somewhat complicated but there's nothing complicated about the fact that lead, in all its forms, remains one of the most troubling of all industrial exposures. And whether Drum gets it perfect here or not, he does get the main point right. We need to keep reminding ourselves - and our government - that we all benefit by, as they say, getting the lead out.

Call it a useful misconception, even if you remain a skeptic.

* Those familiar with Sailer's work would point out that his disagreement is at least partially motivated by a very specific ideological disposition. Like Charles Murray, Sailer consistently argues that different races experience different levels well-being and achievement for primarily genetic reasons. This blog does not agree with that position, and the reference to Steve Sailer's article is to illustrate the statistical issues at stake in this debate, nothing more.