If we are serious about preventing needless deaths and routine humiliation of Black and Latino drivers at the hands of police, we need to change how we promote traffic safety in the United States.

Police make 20 million traffic stops every year. That means millions of opportunities for things to go tragically wrong, as they did not only for Daunte Wright this month in Minnesota but also for Philando Castile, Sandra Bland, Walter Scott and countless others who escaped with their lives but whose lives nonetheless were forever impacted.

Are these 20 million stops worth this cost? Absolutely not. In fact, only a fraction of them are for the purpose of traffic safety. Traffic stops often are about one of two things instead: raising revenue, or using minor traffic violations as a pretext to investigate people for something entirely unrelated to traffic safety.

There is some indication that the stop of Army 2nd Lt. Caron Nazario by Windsor, Va., police — the video of which went viral this month because of the dehumanizing way Nazario was treated — may be an example of what happens when towns use police traffic enforcement to raise revenue. Windsor has a population of 2,700. It is fair to ask why it has any police — there is already a county sheriff, after all — much less seven of them. Part of the reason may be that 6 percent of Windsor’s revenue comes from traffic fines.

And Windsor pales in comparison to many other small towns in the United States that generate more than half of their revenue from fines and fees grounded largely in traffic enforcement. Cities with more Black residents rely more on traffic tickets and fines for revenue, and these systems can push people into poverty.

What happened to Nazario is an inevitable consequence of this approach: Police are encouraged to prioritize traffic enforcement; people, especially Black and Latino people, experience these stops as somewhere between annoying and terrifying and respond accordingly; police — desensitized to the harm of the intrusion and emboldened to expect submission — overreact.

It may also be that Nazario was stopped, as Wright appears to have been, as a “pretext” to conduct an otherwise legally unjustified investigation. The Supreme Court legalized this tactic in Whren v. United States. Under Whren, as long as police can find some infraction, they can stop and question the driver, often conduct a limited search of the car, require the occupants to sit on the curb, and try to get the driver to consent to a more thorough search of their bodies or their vehicle. This might not be so bad if these searches had a good chance of preventing serious crimes. But data shows that few stops serve a significant public safety purpose — usually less than 1 percent turn up any contraband at all.

What’s more, police spend a lot of time on traffic stops that even some agencies have found itnot worth the effort. One study found that in just 11 law enforcement agencies, police spent 85,000 officer hours over 10 years just on potential non-moving violations.

This means that for little public safety payoff, we’ve created a racially biased, time-consuming and dangerous traffic enforcement system, and in the process relinquished our constitutional rights. We have made our roads a Fourth Amendment-free zone in which officers literally can stop anyone they want based on random selection or, worse, racial bias. Study after study shows that Black drivers are searched more often than White drivers during stops, but found to have guns or drugs less often.

There are at least three things we can do to reduce the harm and racial disparities of traffic enforcement without compromising public safety.

First, we can take much of traffic enforcement out of the hands of police. Some places are shifting traffic enforcement to unarmed traffic safety experts. Relatedly, we can make better use of transportation design and technology. There are legitimate concerns about where red-light and speed cameras are placed, but I’ve never seen one pepper-spray a motorist or show a strange proclivity for targeting Black drivers when it was light enough to see skin color. This shift would reserve police stops for immediate threats — such as drunken driving — that arguably require a police response.

Second, cities and states should reject pretext stops, by prohibiting their use, reducing the infractions for which police can stop people, and cleaning up often-antiquated vehicle codes to remove violations that have little to do with public safety and everything to do with allowing police to stop people at will. Pro tip: Prohibiting air fresheners hanging from rearview mirrors is not motivated by a traffic safety concern.

Third, cities should shift the resources currently used for our harm-inefficient traffic enforcement system to evidence-informed programs for preventing gun violence and reducing the harm of illicit drugs.

Given that Whren came out of Washington, the D.C. police would be a worthy candidate for leading this change. Indeed, the D.C. Police Reform Commission recently provided a road map for how to reduce the harm of traffic enforcement. Other jurisdictions are innovating as well. We should encourage and learn from those efforts to create a traffic safety system that does not kill and humiliate.

Read more: