More security may actually make us feel less secure

Even as the federal government increasingly deposits its surplus military supplies with local police departments and as Americans are more exposed to militarized police forces, study of its historical political development (1) or examination of its effects for American communities has mostly escaped scholarly attention. No national data source across the nation’s 18,000 police agencies tracks the incorporation of tactics, personnel, or gear traditionally used in military operations abroad. Without such a repository, scholars can neither measure the prevalence of police militarization and how it varies across American communities, nor develop empirical insights about how these developments affect the security of Americans.

In his PNAS article, Jonathan Mummolo expands scholarly understanding of the consequences of increasingly militarized police forces in the United States, its politics, and its racial geography (2). Compiling detailed administrative data in one state of one kind of militarization, “special weapons and tactics” (SWAT) deployments, a national panel of which states acquired a special tactics team, and an original survey experiment of different levels of militarization, Mummolo examines whether militarization contributes to the safety of police officers, local crime reduction, and confidence in police among Americans. Militarized policing does not result in the anticipated public safety gains nor does it abet officer safety; it does impair confidence in police and elevate perceived crime.

The Consequences of Policing

The logics and capacity of American police institutions have undergone a dramatic shift over the past half century (3, 4) spurring debate about the consequences of policing for public safety, trust, and other aspects of well-being. A prominent through-line of this work is that police interactions cause several negative outcomes: involuntary encounters with police have been linked to an increase in posttraumatic stress disorder (5), declining grades and test scores among American youth (6), legal estrangement (7), strategic avoidance of people, places, and institutions (8, 9), isolation from peer networks (4), and aggravated perceptions of racial discrimination (10). In the wake of high-profile acts of police violence and uprisings to contest accountability deficits, studies have demonstrated that police-inflicted fatalities can lead to residents of a community lowering their calls to police (11). Sustained focus has been on police stops of pedestrians, given the rise of proactive policing models, such as “broken windows.” However, one of the key shifts in our era is the rise of police tactics that borrow from the techniques and equipment of the military. Mummolo’s (2) study advances our understanding of this important development while building on the well-worn approaches in studies of traditional policing.

One debate is that policing itself, at high levels, can be criminogenic (12). Mummolo (2) finds not only that militarized policing strategies fail to build public and officer safety (having no significant effect on crime reduction and in one model actually increasing violent crime incidence), but through survey experiments finds that it may work to its own detriment, making the public perceive they are less safe and undermining their confidence in local police. With the implications of a recent study of calls to police in view (11), some Americans may become wary of enlisting police for help if they anticipate that an armored car may show up to their request, which may work at cross-purposes with public safety goals.

The Politics of Policing

Mummolo’s (2) study provides us a solid foundation for additional examinations of the politics of police militarization. He uses data from the solitary state to legally require the tracking of SWAT deployments (Maryland) to examine the rationales for and effects of the over 8,000 deployments. Descriptive patterns indicate that SWAT teams were overwhelmingly used to serve search warrants and in other “nonemergency scenarios”; multivariate patterns indicate no significant positive effect for officer or public safety. Work on police stop-question-frisk finds that the majority of such encounters produce no contraband and no evidence of wrongdoing (13); if the “hit rate” in these deployments is similarly low [Mummolo (2) doesn’t reveal] and the officer safety benefit is nonexistent, which both his Maryland and national findings indicate, policymakers and police executives might question their utility.

SWAT activities also vary significantly by place. This heterogeneity could be leveraged to examine the political explanations for why some local governments opt for specialized units, tactics, and equipment used in military operations and others forgo a militarized approach. Since 1990 the federal government’s 1208 (and later renamed 1033) program has encouraged local governments to acquire surplus military equipment from the Department of Defense. Political explanations, like partisanship and electoral pressures, help explain local governments’ receipt of COPS grants under Clinton’s program in the 1990s (14). Whether similar political forces are at work in explaining (i) why some local governments seek out and (ii) which localities the federal government ultimately transfers military technology and equipment to is an important direction for future studies.

Political scientists have long recognized that “policy creates politics.” Public policy, beyond responding to public problems and public desires, shapes them as well. This idea has traditionally been applied to the domain of social policy; policies that distribute material benefits create or embolden constituencies and institutions and give rise to citizen ideas (15). Mummolo’s (2) experimental study is a keen elaboration, demonstrating that activities of government bureaucracy can also highlight a social problem that is divorced from reality among mass publics; seeing officers that are equipped with heavy artillery and dressed for war amplifies perceptions of crime severity (on the order of 8–15 percentage points). And given that perceptions of disorder—not actual disorder—are powerful determinants of behavior (16), such escalation may ironically undermine support for further expansions of police.

Although the experimental findings are largely convincing, I wonder whether what appears to be an unintended manipulation may have contributed to them. Mummolo (2) assumes that the experimental conditions are similar except for the manipulation, the level of militarization. In the control, subjects are exposed to rank-and-file police wearing the traditional uniform; in the two low militarization conditions, subjects see officers bedecked in extensive face masks and padded uniforms or with military-grade assault rifles; in the final and most extreme condition, subjects see police officers on and around an armored tank. Mummolo (2) concludes from his effect sizes and significance tests that the highest militarization treatment “caused support for police funding in the United States to fall” by between two and four points.

The problem is that Mummolo’s (2) experiment contains an additional source of subject manipulation that could have influenced resulting police perceptions. Specifically, subjects are exposed to the race of the police officers (a clearly visible black officer among three to four white officers) but only in two of the four conditions: the control and one of the low militarization conditions. In the other low militarization and the high militarization conditions, subjects are not exposed to a racial cue or information about the diversity of the police force, either because officer faces are obscured by riot face masks or too distant to clearly detect.

That the racial cue is present in some but not all treatment groups introduces a potentially serious source of bias: whether the racial cue interacted with the level of police militarization (or independently worked) to shape subject evaluations. In the worst case, the diversity of the police force in question (and not or not only militarization level) could have led subjects to be less supportive of the police and triggered respondents to think about crime in racially laden ways. This is no idle speculation; a voluminous literature in the social sciences has found substantial effects of visual racial cues on the evaluation of political actors through the activation of respondent stereotypes, especially strong in the domain of crime and violence (17). Given that this additional manipulation occurred in the control and one treatment group and given that it may have divergent effects based on subject characteristics and racial predispositions, the potential for

In his PNAS article, Jonathan Mummolo expands scholarly understanding of the consequences of increasingly militarized police forces in the United States, its politics, and its racial geography.

it to change the magnitude or direction of findings is concerning and this factor should be addressed in future studies; whether it led Mummolo (2) to over- or underestimate the effects of militarization is unclear, however.

State Violence Against Black Americans

Mummolo (2) is also contributing to expansive work in the social sciences and history that theorizes policing as part of state projects of social control and race–class subjugation (18–20). Indeed, the nation’s inaugural SWAT team first deployed in Los Angeles in 1969 to surveil and undermine black political resistance, sending 300 officers (and tear gas, battering rams, and army tanks) into Black Panther Party headquarters (20).

A central finding is that SWAT interventions are concentrated in black neighborhoods, and consistent with most existing work, the spatial pattern is not explained by crime rates. Police interventions are significantly higher (and more severe) even after taking into account differential crime involvement and criminal history at both the individual (21–23) and neighborhood levels (24, 25).

This important finding raises two questions: What is the mechanism for why SWAT deployments are patterned by racial geography? And what are the implications of this finding for black Americans, specifically? Mummolo’s (2) survey experiment focuses on the effect of police militarization for support for police patrols, spending, and confidence in the job police are doing as a general matter. But given that a central component of this militarization is more pervasive in black neighborhoods and in view of the negative effects of involuntary police encounters for well-being, engagement, and mental health briefly canvassed above, scholars should probe the consequences for black Americans and neighborhoods specifically.

Given SWAT’s historical origins and its contemporary use in black homes beyond what crime levels would dictate, it stands to reason that black Americans perceive it not only as an objective crime-fighting tool, but as a tool of black oppression. Militarized police may not just inform the public about the extent of crime and the job police are doing, but may also convey racialized messages and be perceived as a part of a larger way blacks are governed (3). As one West Baltimorian describes the “blue zones” in her neighborhood and the conclusions she draws about who militarization is targeted at: “…it’s constantly under police surveillance. There’s huge police lights. I’m talking about, like, military grade equipment in people’s neighborhoods. And not just anyone’s neighborhoods. Black people’s neighborhoods, Latino people’s neighborhoods, low-income people’s neighborhoods…People can’t live like this.”* Another describes how the political system responds to black grievances by positioning them as an enemy combatant: “How this government, our government, responds to riots and when black people are hurt and we feel like we have to uplift our voice. Instead of like being empathetic and compassionate, they like sent the military or their SWAT teams at us and like, you know, it’s almost like we’re, we’re invaders, at that point… in your own city... any time you can be killed by your own soldier. That’s how it feels like in America.”*

Mummolo (2) has given strong empirical support to the idea that exposure to militarized police has unforeseen consequences beyond public safety. Future research should seek greater understanding of the lived experience of militarized police and whether it exacerbates perceptions of anti-black racism, distrust of state authorities, and fear of calling police.