On August 9, 2014, Michael Brown, a Black teenager, was stopped by police officer Darren Wilson in Ferguson, Missouri. Within minutes, Brown had been shot and killed. In the wake of the killing—and as more deaths of Black men and women at the hands of police officers made headlines—the Black Lives Matter movement, which had begun as a social media campaign the previous year, grew into a nationwide protest movement.
The federal government responded in 2015 with the Police Data Initiative (PDI). Among other initiatives, the PDI required participating police departments to release some internal data on traffic stops.
“In the eyes of Black Lives Matter activists, this transparency initiative may have been a way to bring department practices in line with their demands through external accountability,” says Minjae Kim, an assistant professor of organizational behavior at Yale SOM. Such external accountability measures are also common in pursuit of other goals, such as closing gender pay gap or decreasing carbon emissions. However, Kim says, “we haven’t really figured out if changes that take place in response to these accountability measures last over time.”
In a new study, Kim, with Tae-Ung Choi of South Korea’s Yonsei University, tracks the PDI over 30 weeks and find that transparency initially did what activists may have hoped, shifting police departments toward fewer stops in neighborhoods with more Black residents. But this change didn’t last. These “treated” departments that opted into the PDI reversed their gains, eventually making more traffic stops in those neighborhoods than departments that never joined the PDI.
The researchers came to these conclusions by merging two sources. The Stanford Open Policing Project has used Freedom of Information Act requests to assemble internal records on nearly 100 million traffic stops, releasing them publicly beginning in 2017. Matching those records against a list of PDI participants yielded a before-and-after picture for 36 local departments. Six of these opted into the initiative in May 2015—a decision made by mayors and police chiefs, not rank-and-file officers; 30 departments did not opt in and served as a control group.
Rather than deciding in advance what counted as “short-term” or “long-term” effects, Kim and Choi let the data set the terms. They fixed a 6-week window before the PDI’s launch and compared it against a series of expanding windows afterward: first the 6 weeks immediately following implementation, then 10 weeks, 14 weeks, and so on, out to 30 weeks post-implementation.
The effect, at first, was clear. “In police departments that adopted the PDI there was initially a large decrease in the number of stops,” Kim says. “This took place in the first six weeks and was especially true in neighborhoods with more Black residents.” Relative to the control group, departments that opted into PDI made about 2.81 fewer stops per 10,000 residents each week, or roughly an 11.7% reduction.
But then the effect eroded. As the researchers looked further out, the decline shrank, lost statistical significance, and finally flipped. By 30 weeks out, the departments in PDI were making about 1.82 more stops per 10,000 residents than the control group, or a 7.8% increase.
This trajectory was concentrated in precisely the neighborhoods targeted by the PDI. The initial drop was steepest in census tracts with the largest share of Black residents, but by week 30, tracts with more Black residents saw the steepest increases in traffic stops.
It’s hard to make change. But this work shows that the boundaries of organizations might be more porous than we think.
Kim and Choi also sorted departments by their pre-PDI conduct and found that this backsliding appeared only among the departments that had started out furthest from the PDI’s goals, meaning those making the most stops in neighborhoods with more Black residents. Departments with lower baseline stop rates in such neighborhoods, by contrast, did not see this reversal during the study period. In other words, the departments that may have been most subject to the reform pressure were the ones that snapped back most dramatically.
Though the authors can’t be sure what drove the reversal, they suspect the initial pressure from activists gave way to a resurgence of pressure from other and perhaps more entrenched interest groups that pushed departments backward. Fading public attention and simple organizational inertia could explain why compliance decayed, Kim notes, but neither of these mechanisms explains why stops rose above where they had been. That points to a different constituency—“possibly those with more political capital, with resources to hold police departments accountable”—getting its way.
The implications are sobering. Because departments volunteered for the PDI, these may have been the organizations most inclined to comply. The findings suggest that transparency may be necessary but not sufficient. Disclosure can nudge an institution in the right direction initially, when pressure for reform is greatest. Mandatory measures may require durable oversight and real enforcement to avoid generating resistance. Sustaining reform also appears to require continued attention from the people who demanded the change.
That said, Kim doesn’t read the findings as reason for activists, or anyone interested in organizational change, to despair.
“Yes, it’s hard to make change, but to me this work shows that the boundaries of organizations might be more porous than we think,” he says. “This is certainly not reason to be less hopeful if your goal is to elicit organizational change using some accountability initiative. Despite the later backsliding, we also saw that enough pressure and attention on an issue translated, at least for a time, into a real change.”
“The Yale School of Management is the graduate business school of Yale University, a private research university in New Haven, Connecticut.”
Please visit the firm link to site