One major point about drug laws is that they're very widely violated. In 2002, there were some 19.5 million drug users in the United States—about 8 percent of the population—but just 1.5 million drug arrests and 175,000 people who went to jail for drug offenses. Now, that's a ton of people in prison—and most of them should not be there, period—but it's still only a tiny fraction of all drug users. Any law that's flouted that frequently is going to be enforced in a very, very selective manner.

But even though we often hear platitudes about how that's true—about how the War on Drugs disproportionately affects black people, and so on—the details here are truly remarkable. For instance, a 2002 NIDA survey found that African American teens use drugs at a slightly lower rate than their white counterparts, and that's true for a variety of specific drugs—even crack. But, that year, black youths were brought to court on drug charges at a rate of 8.2 per 1,000, compared with only 6.0 per 1,000 for whites.

Likewise, the 2002 SAMHSA National Survey on Drug Use and Health found that African Americans of all ages use illicit drugs at only a slightly higher rate than whites (9.7 percent vs. 8.5 percent). But blacks were sent to prison for drug offenses at a far, far higher rate. Okay, maybe that's because African Americans are more likely to be involved in selling drugs? But that's also doubtful: One DoJ study found that 13 percent of black youths had sold drugs, compared with 17 percent of white youth, yet, in 2003, black youths were arrested for drug crimes at twice the rate of whites.

Now, there are a slew of possible theories for why that's so. Disparate policing practices are one factor—police often focus on low-income and minority neighborhoods, and it's easier to spot and bust street deals than suburban sales. Unequal treatment in courts is another factor—white youths are twice as likely to retain private counsel as black youths, and one study found that, for instance, probation officers are more likely to see crimes by minorities as caused by personal failings, while seeing crimes by white youth as caused by circumstance. ("He's a good boy, just fell in with a bad crowd…") It's not hard to imagine that bias applies to judges and prosecutors, too.

Anyway, along those lines, the Justice Policy Institute just put out a fantastic new study adding its two cents on the issue. Their study looked at county-by-county data, and found some striking patterns. There's barely any correlation between drug use and the rate counties put people in prison for drug offenses. Rockingham County, NH, has a higher percentage of drug users than Jefferson Parish, LA, but Jefferson has a drug admission rate 36 times greater. And so on. JPI did find, though that counties with higher poverty rates, or larger percentages of minorities have higher drug incarceration rates—even after you control for crime rates.

The other noteworthy finding here is that counties that spend more on police budgets end up imprisoning people for drug offenses at a higher rate—even after you control for crime, region, poverty, unemployment, and so on. It's like you'd expect: When a county adds more police, those police spend more time going after drug offenders. And they're more likely to use their discretion to go primarily after black drug dealers—even if African Americans aren't selling drugs at higher rates, as has been the the case in Seattle, according to this in-depth study.