22 October 2007

Another hypothesis for the crime drop in the late 90's

Idea Lab: Criminal Element
Reyes found that the rise and fall of lead-exposure rates seemed to match the arc of violent crime, but with a 20-year lag — just long enough for children exposed to the highest levels of lead in 1973 to reach their most violence-prone years in the early ’90s, when crime rates hit their peak.

Many, many reasons have been floated as to why crime dropped in the late 90's. The one I remember best was the one put forth in Freakonomics by Steven Levitt. His hypothesis had to do with thousands of unwanted children not being born to mothers who couldn't provide for them thanks to the legalization of abortion. This hypothesis is based upon another absence. Not of people this time but the absence of toxic lead.

Now this idea doesn't usurp Levitt's idea, in fact the two might fit together well. Together, the two predict the birth of children who are more likely to be wanted, more likely to be fully provided for and healthier. More than just the levels of lead in the air have improved since the mid-seventies. There is little doubt in my mind that the formation of the EPA and the implementation of the Clean Air and Clean Water acts has improved the lives of those born since then. But how much can you blame the criminality of past generations on the amount of lead they were breathing? Rather quickly we reach a broader discussion about nature, nurture and the free will of the criminal. But thats not where I want to take this. I want to talk about the science, this is Science Monday after all.

To understand why people put lead in their gasoline in the first place, you have to understand a little about how a internal combustion engine works. The best animation is this one, the 4-stroke at Wikipedia. Lead reduced "knocking" which is premature ignition, the gasoline ignites during compression in the second stroke rather than at the boundary between the second and third stroke. The form of lead used was tetra-ethyl lead. Thats lead with four two-carbon ethyl groups. In the engine, the TEL was combusted to CO2, H2O and lead oxide, which was picked up by a chlorine based scavenging compounds. So out the tailpipe came lead (II) chloride.

In the end it doesn't really matter what form the lead takes when it enters your body, lead oxide, lead acetate, or lead chloride. Your body has no use for lead and the lead will replace other metal atoms that form the active sites in many enzymes in your body. So what does that have to do with crime levels here? Well thats an interesting question.

Lead exposure early in life has been shown to have a link to learning disabilities later in life. Having a learning disability alone doesn't make one a criminal. But if you didn't learn in school, couldn't sit still, and no one was trying to help you learn, what are the chances you would be able to make it through high school untouched by violence?

In the end the really science-y part of this article isn't about the chemistry of lead, the biology of lead poisoning or the psychology of learning disabilities. The really science-y part comes near the end.

If lead poisoning is a factor in the development of criminal behavior, then countries that didn’t switch to unleaded fuel until the 1980s, like Britain and Australia, should soon see a dip in crime as the last lead-damaged children outgrow their most violent years.
......

The theory will be put to the test as children grow up in Indonesia, Venezuela and sub-Saharan Africa, where leaded gasoline has just recently been phased out. Meanwhile, the list of countries that still use lead in gas — Afghanistan, Serbia and Iraq, as well as much of North Africa and Central Asia — does not rule out a connection with violence.

That right there is the most science-y part of the whole thing. It makes a prediction. It says "hey, I have this data and this really neat conclusion, plus I have a way to test this!" In a few decades we'll be able to see if Jessica Wolpaw Reyes, the economist at Amherst College who's research is the basis for the article, found a correlation that holds up to its own prediction. That ladies and gentlemen, is science!

Reyes working paper is available as a pdf here. Page 69 has a very nice graph of violent crime and kilos of lead with a 22 year delay. Its really a pretty graph, as far as graphs go.

No comments: