San Francisco’s Police Commission voted 4-2 this week to place limits on what circumstances in an effort to cut down on police disproportionately targeting people of color. Called the “pretext stop” policy, the new rule limits police from primarily stopping drivers over nine different low-level offenses, like failing to display registration tag or failing to use a turn signal, reports.

Low-level traffic stops are often used by police as fishing expeditions targeting a . In San Francisco, for instance, Black people make up five percent of the population, but account for 19 percent of all traffic stops made in the city, . Countless studies reinforce the trend of pretextual stops disproportionally affecting people of color while such stops do little to cut down on crime, as reported in the .

“There are a cluster of low-level traffic stops that are just not yielding any public safety benefit for the city,” Commission Vice President Max Carter-Oberstone told the . “But they do take up a lot of time and they do cost a lot of money and by curtailing those stops we can reallocate all of those law enforcement resources to other strategies that we know are effective.”

Critics say , but the Police Commission notes the new policy is merely a test balloon that will hopefully allow officers to focus on my serious crimes. The policy change now goes to the San Francisco Police Officers Association, which could slow down adoption of the limits considerably, the Standard reports. Police unions in general are not supportive of a restriction in pretexual stops, the Times reports.

The non-profit , which is devoted to ending mass incarceration, notes many high-profile deaths of Black people by police started with a low-level traffic violation:

Sandra Bland was pulled over because a police officer said she failed to signal a turn. Philando Castile was stopped for a malfunctioning brake light. Daunte Wright was driving a car with an expired registration tag.

[...]

Police have great discretion in making non-public safety stops, leaving space for implicit and explicit racial biases to impact their decisions. Numerous studies show that than white people. found that Black drivers were less likely to be stopped after sunset, when a “veil of darkness” masks race, suggesting bias in stop decisions. This same study examined the rate at which stopped drivers were searched and the likelihood that searches turned up contraband, and found evidence that the bar for searching Black and Hispanic drivers was lower than it was for searching white drivers.

When researchers examined statistics surrounding “must stop” situations that presented a clear danger and non-public safety stops that did not, they found that virtually could be attributed to non-public safety stops.

The vast majority of non-public safety stops do not result in the discovery of drugs or weapons. , for example, only 2 percent resulted in the discovery of a weapon. Meanwhile, victims of such stops can suffer . Non-public safety stops also expose police officers to unnecessary danger. The most common proactive policing activity preceding a fatality is an officer-initiated traffic stop.

San Francisco joins cities like Pittsburgh, Los Angeles, Seattle, and the entire state of Virginia in restricting traffic stops for low level offenses or what are known as “pretexual” stops. In about a dozen places in the U.S., where police have failed to change policies, prosecutors have said they won’t prosecute such tickets and offenses, according to the .