Marijuana legalization won’t stop racially skewed arrests. But it limits them.

One of the most common arguments for marijuana legalization is getting a little more complicated.

A new preliminary report by Mike Males at YouthFacts found that the marijuana arrest rate dramatically dropped — by 76 percent between 2008 and 2014 — in five states after they legalized or decriminalized the drug, as one would expect. In comparison, marijuana arrest rates in all other states fell by just 15 percent in the same time period.

But it also found that the racial disparities within those arrests didn’t improve, even though black people aren’t more likely to use or sell drugs than their white peers.

So in Colorado, which legalized pot in 2012, the total number of marijuana arrests fell by 60 percent from 2008 to 2014. But racial disparities remained: Black Coloradans were still more than twice as likely to be arrested for pot as people of other races. (Even though Colorado legalized marijuana, people can still be arrested for possessing an amount of pot above a certain threshold and public use.)

But this doesn’t necessarily mean marijuana legalization was an ineffective way to reduce racial disparities in the criminal justice system. After all, the report found that black people still benefited the most from the change: The marijuana arrest rate per 100,000 black people in Colorado dropped by nearly 300 (to 242.2), while the rate for people of all other races dropped by nearly 160 (to 103.8).

Why did black Coloradans disproportionately benefit? Consider implicit bias, subconscious biases against people of different races and ethnic backgrounds. These biases may drive a cop to, for example, see a black man smoking pot as a criminal who needs to be arrested and a white man smoking pot as a mere troublemaker who can be let off with a warning.

This kind of discrimination is limited under legalization: With one less crime to charge people with, cops are less able to use racially biased discretion in deciding whether to arrest someone.

The argument for marijuana legalization’s effects on racial disparities, then, isn’t really that it will undo all bias in arrests. But it will help — by limiting the discretion cops and other actors in law enforcement can use to disproportionately criminalize black people. Until systemic prejudices like implicit bias are fixed, that may be the best option.