Defunding the Police: an Evidence-Based Approach?

In the months following the death of George Floyd, there have been an increasing number of protests aimed at taking a stand against police brutality and “defunding” the police. The slogan “defund the police” has since been adopted by various activist groups and is now being seriously debated by politicians and lawmakers across the country. Despite the fact that Americans are mixed on whether they support the idea, the slogan has nonetheless become an increasingly popular political talking point. While the argument that America’s police departments are in need of reform is not without merit, that does not mean that defunding the police is the answer.

Articles going back to the 1960s have advocated for advancing a research agenda in policing that should serve as the basis for police reform. However, as a nation we have not met this standard. Similar to the rest of our criminal justice system, policing as a profession tends to be reactive based on shifts in politics or public opinion without careful consideration of empirical data. This often leads to policy change that is too rapid and not carefully assessed or planned, ultimately inhibiting us from moving forward.

Many proponents of defunding the police claim that if funds were diverted from police budgets to other evidence-based community-based initiatives, crime would decrease. But how do we know whether these programs will actually work? If a policy is truly evidence-based, it is presumed that the policy has been evaluated using empirical data to assess its effectiveness in achieving intended goals. But even with empirical data, determining the ‘effectiveness’ of a policy is not as simple as it seems.

It is true that vast data on policing do exist, some of which is publicly available. However, the data has inherent limitations and is not as comprehensive as it may seem. For example, while many agencies have historically reported annual crime numbers to the Uniform Crime Report (UCR) and the National Incident-Based Reporting System (NIBRS), important details about incidents are sometimes unknown or lacking for several reasons. First, reporting crime to UCR/NIBRS is actually voluntary and  participation rates can vary based on the police department. Second, jurisdictions vary quite a bit in terms of their criminal statutes and subsequently, what crimes would be reported to the UCR/NIBRS.

Beyond just examining the numbers of reported crimes, it’s also important to examine patterns present within case details, such as the suspect/victim demographics. When it comes to incident-level details, the data become more spotty in terms of what details are actually reported from jurisdiction to jurisdiction. For example, a department might report the number of robberies that happened but may decline to include any information about the race, gender, or age of suspects. Thus, it can be difficult to adequately assess such patterns, particularly when attempting to compare jurisdictions to each other. While some academic research has been conducted to help fill in this gap, it has not been able to do so entirely.

It’s also important to remember that there is no national-level standard for what constitutes “effective” policing. Relatedly, there is no national-level oversight or standard for governing department policies (such as use-of-force requirements or body-worn camera policies, for example). The entire policing profession is piecemeal, with each jurisdiction following its own rules and its own policies. Thus, it makes it very difficult to compare jurisdictions and truly assess the effectiveness of various initiatives, further complicating our ability to know “what works”. Unpacking this issue is not an easy task, and is one that will require funding, persistence, and dedication. With that in consideration, the federal funds dedicated to police research sound small in comparison to the federal funds dedicated to other areas of discretionary research spending. For example, in 2020, the federal government allocated $26.2 billion for a variety of research ideas; only $79 million went to the National Institute of Justice (NIJ) and the Bureau of Justice Statistics (BJS), the primary federal agencies dedicated to criminal justice research. This $79 million was expected to cover the entire breadth of courts, corrections, and policing research.

Although there is much more work to be done to refine “what works” in policing, we do have a solid body of empirical knowledge that can serve as a foundation for the way forward. In 2015, the Department of Justice released a report on 21st century policing that provided a variety of research-based recommendations for the policing profession and highlighted where gaps remain in the research. This report has served as a guide for many police professionals and researchers regarding key issues that we face in policing today and topics where future research is needed, including topics such as building trust and legitimacy, issues with policy and oversight, improving officer training, and the importance of bolstering data collection and research efforts.

One recommendation from the 21st century policing report that rings particularly familiar is the recommendation to “support programs that take a comprehensive and inclusive look at community-based initiatives addressing core issues such as poverty, education, and health and safety.” An important distinction mentioned in this report that the “defund the police” movement fails to consider is that “community-based initiatives” should be implemented in tandem with policing and not in lieu of policing. This example highlights how the current national conversation has in many respects devolved into a false dichotomy where you have to be either pro-police or anti-police. Policing is not a zero-sum game, and this mentality is not only inaccurate, it’s also destructive and self-perpetuating. Policing cannot reform itself without proper planning, resources, and dedication, which will require us to put aside our political leanings in favor of empirical research and data.