If it’s really true that lead reduction was responsible for most of the post-1990 decrease in crime, and if changing demographics played a role as well, doesn’t that mean that everything else probably had almost no effect at all? Broken windows, open-air drug markets, three-strikes laws, CompStat, bulging prison populations, etc. etc. — all of them together couldn’t have had more than a minuscule impact if lead and demographics explain almost everything.
1. That’s not a necessary conclusion from the data. Some things push crime up, others push crime down. Effect sizes are mostly unknown. So it could be the case both that lead reduction explains most of the crime decline and that crime was lower than it otherwise could have been as a result of improved crime-control strategies. That would imply that the other forces pushing on crime would have otherwise generated a crime increase. Given our level of ignorance about what drives crime levels, that could easily be true.
2. Even if it were true that improved crime control strategies had little or no effect overall, that near-zero impact might have been the algebraic sum of better strategies leading to less crime in some places and worse strategies leading to more crime in other places.
There have been some important breakthroughs in crime control over the past fifteen years or so; that’s why I’m convinced that we could have half as much crime and half as many people in prison five years from now as we have today, while spending less money. But none of those breakthroughs has been widely imitated. One advantage of competitive markets is that they drive companies that want to survive to be “fast followers” of others’ innovation. Since police departments don’t compete with other police departments or probation agencies with other probation agencies, there is less natural pressure for the rapid diffusion of innovation. For that we need to rely on journalists, academics, and politicians.