While riding with officers, conducting interviews and coding policies for my forthcoming book, Camera Power: Proof, Policing, Privacy and Audiovisual Big Data, I was struck by the influence of police unions—or lack of a strong union—in shaping body camera recording policies and limits on using the video to evaluate and discipline officers. Delving into the literature on police unions, I was impressed to read the work of prolific professors using innovative methods to systematically collect and analyze data on the influence of police unions. I would like to spotlight two recent important empirical studies on police unions.
Analyzing a large dataset of police union contracts, Stephen Rushin’s latest article illuminates how collectively bargained protections in the police disciplinary appeals process can impede efforts to address potentially problematic officers. The findings are particularly disturbing and compelling when read in conjunction with an important new study by Dhammika Dharmapala>, Richard H. McAdams and John Rappaport. This dream team of interdisciplinary scholars offers the first quasi-experimental evidence that conferring collective bargaining rights on sheriffs’ deputies is associated with about a 45% increase in violent incidents.
The son of a police leader, Rushin is often at the forefront of cutting-edge empirical questions in policing. For example, do heightened scrutiny and criticism of cops and their departments lead to de-policing—a cutback in the vigor of officers in combatting crime? Rushin has examined the question using U.S. Department of Justice (DOJ) investigations as proxies for heightened public scrutiny and Justice Department settlements as proxies for heightened external regulation of departments. He applied difference-in-difference regression analyses comparing the impact of DOJ investigation and regulation on crime rates compared to control jurisdictions that did not fall under DOJ scrutiny. Check out the article for his findings. Rushin also has analyzed a dataset of 178 police union contracts and illuminated common provisions arising from collective bargaining that can frustrate police accountability efforts. This empirical strategy complements other important work on police unions applying legal analysis and case studies, such as this excellent 2017 article by Catherine L. Fisk and L. Song Richardson.
Rushin’s latest project, Police Disciplinary Appeals, draws on an even larger dataset of 656 police union contracts to reveal how the often-extensive process police officers have to appeal disciplinary sanctions can stymie police accountability. Through compelling narratives, the article illustrates how attempts to dismiss problem officers are blocked and sanctions reduced in lengthy appeals—even when smoking-gun evidence of egregious misconduct is caught on video. Rushin is able to shed light in an area often overlooked because of his impressive data collection strategy. He amassed his large dataset of police union contracts between 2014 and 2017 through open record requests as well as searches of publicly available documents. I have been impressed by the ingenuity and persistence of JD-PhD-trained law professors like Rushin and Jordan Blair Woods, author of the excellent new article Policing, Danger Narratives, and Routine Traffic Stops, in accumulating large datasets using public records requests to offer fresh insights on policing.
Like another prolific scholar, Kate Levine, whose superb new article Discipline and Policing, was featured last month here, Rushin’s work also is valuable in illuminating the internal procedures and regulations of police departments. The article begins with an informative overview of labor and employment protections for police officers coming from (1) police union contracts, (2) law enforcement officer bills of rights, and (3) civil service laws. Rushin then explains how he generated and coded his large original dataset of police union contracts covering municipal police departments in 42 states that authorize police unionization.
Rushin found that the strong majority of these 656 contracts have a similar disciplinary appeals process. Around 73% provide for appeal to an arbitrator or comparable procedure and nearly 70% provide that an arbitrator or comparable third party makes a final binding decision. About 54% of the contracts give officers or unions the power to select that arbitrator. About 70% of the jurisdictions give these arbitrators extensive review power, including the ability to revisit disciplinary matters with little or no deference to the decisions made by supervisors, civilian review boards or politically accountable officials. Together, Rushin’s findings reveal that “arbitrators are the true adjudicators of internal discipline in the majority of police departments” in the study’s extensive dataset. Rushin observes that the police union contracts setting forth these protections are typically negotiated outside of public view or participation. He proposes reforms that would give the public more oversight power and democratic actors more of a role in the disciplinary appeals process.
Rushin’s wide-angle lens view of how police union contract provisions can pose potential hurdles to police accountability is illuminating. For a scientist, the arguments also raise hypotheses that call out for testing via an innovative research design. Dharmapala, McAdams and Rappaport’s new quasi-experimental study rises to the challenge, offering important and disturbing evidence of an association between conferral of collective bargaining rights and an increase in violent incidents in Florida sheriff’s offices.
An innovation of the trio’s study is exploitation of a change in law—a 2003 Florida Supreme Court case Coastal Florida Police Benevolent Ass’n v. Williams that gave Florida sheriffs’ deputies the right to engage in collective bargaining. Collective bargaining proceeded in such earnest that 76% of sheriffs’ deputies in Florida spread in 28 offices were under a collective bargaining agreement by 2008. In contrast, Florida police officers had collective bargaining rights much earlier, beginning in 1968. Because of the contrast, the researchers could examine violent-incident complaints before and after the conferral of collective bargaining rights on sheriffs in 2003 using a difference-in-difference approach. Changes in trends in violent incidents in the treatment group of Florida sheriff’s offices receiving collective bargaining rights could be compared to trends in violent incidents in the control group of Florida police departments unaffected by this legal change (the treatment). This approach allows us to focus on changes attributable to the intervention of conferring collective bargaining rights and control for other factors occurring in the state that might have affected rates of violent incidents.
The investigators found that before Williams the treatment (sheriffs) and control (police) groups had similar parallel trends, an important assumption that must be met for the difference-in-differences approach. After Williams, the number of incidents of violent misconduct rose in the treatment group but not the control group. Moreover, the increase in violent incidents appeared to be sustained more than a decade after Williams (2011-2016 compared to 2003-2010). Interestingly, violent incident also rose in sheriff’s offices that did not enter into collective bargaining agreements (to a smaller magnitude and with borderline statistical significance).
The findings are particularly intriguing because Florida has a statutory Law Enforcement Officers Bill of Rights containing numerous protections for officers, and also is a right-to-work state that allows employees to decline to join unions. Both factors would be expected to dilute the impact of conferring collective bargaining rights because officers already have a baseline of protections and because it is harder to fully unionize a workforce in right-to-work states. Yet the investigators’ findings regarding an increase in violent incidents were of noteworthy magnitude as well as statistical significance. The investigators are collecting Florida collective bargaining agreements to analyze how provisions go above the protections in the statutory Law Enforcement Officers Bill of Rights. Here their envisioned analytical strategies intersect with those pursued by Rushin, showing how both empirical approaches are complementary and enrich our knowledge of these important issues.