Effect of Non-Prosecution on Racial Disparities

The Issue 

Approximately 45 million Americans, roughly 14 percent of the United States population, have misdemeanor convictions. Eighty percent of state criminal case dockets are misdemeanor offenses, with over 13 million cases filed each year. Individuals in poor communities, especially communities of color, are disproportionately affected by misdemeanor convictions. And while racial disparities are prevalent throughout the criminal legal system, they are even starker among misdemeanants. Because mass misdemeanor criminalization is one major driver of racial disparities in arrests, prosecutions, and convictions, declining to prosecute nonviolent misdemeanor offenses should result in reductions in racial disparities in the prosecution of such cases as well as in new criminal complaints. It is unclear, however, whether this is the case.

The Research Question 

What is the differential impact by race/ethnicity of the nonprosecution of nonviolent misdemeanor offenses on current and future penal system involvement? Will nonprosecution of nonviolent misdemeanor reduce racial gaps in rates of nonprosecution and convictions between Black and Latinx people, on the one hand, and White people, on the other? Will the nonprosecution of nonviolent misdemeanor reduce racial gaps in new misdemeanor and/or felony criminal complaints and convictions?

The Study 

Our analysis will focus on two jurisdictions in which District Attorneys have recently implemented presumptive nonprosecution policies: Suffolk County, Massachusetts and Durham County, North Carolina. We will analyze administrative data from each jurisdiction’s court system to explore the extent to which presumptive nonprosecution policies impacted prosecution and conviction outcomes among impacted defendants in their current and subsequent cases.

Analysis will primarily leverage statewide court databases for MA and NC, each covering the impacted county as well as others that were never impacted by similar policies. Each database includes records for relevant cases beginning several years before the policies in question were implemented (starting in 2017 for MA and 2014 for NC), continuing through the end of 2021, the most recent calendar year with complete data. Analyzing statewide administrative data that extends to before the policies in question were implemented will allow us to examine the policy’s implementation in the impacted counties compared to earlier time periods and compared with other counties in the same states that did not face similar policies. These comparisons will be central to our analytical strategy for isolating the impact of presumptive nonprosecution policies from other factors that may have impacted outcomes over time. 

Identifying the causal effect of presumptive nonprosecution policies on prosecution, conviction, and future criminal legal system contact presents several challenges. The policies were not implemented at random, but rather resulted from concerted political campaigns in specific places, targeting specific types of conduct. As a result, we might worry that other factors relevant to trends in these outcomes and racial disparities within them are at play and that prosecution, conviction, and future criminal legal system contact might have changed in the absence of the presumptive nonprosecution policy. We plan to use a variety of empirical strategies to address these concerns including a differences-in-differences (DD) framework that leverages multiple possible control groups.

The research team will also collaborate with community partner organizations in both Suffolk County and Durham County to host a series of focus groups comprised of people disproportionality impacted by policing, prosecution, and incarceration, to finalize the research plan, and to draft the policy implications for the research briefs. In Suffolk County our community partner is the Transformational Prison Project (TPP), a Massachusetts nonprofit focused on restorative justice and community engagement. The Durham County community partner is All of Us or None NC. The research team will attend the focus groups to document the feedback from participants. 

This work is supported by Grant #21-06300 from Arnold Ventures.  Any opinions expressed are those of the principal investigator(s) alone and should not be construed as representing the opinions of the Foundation.