From defunding to refunding police: institutions and the persistence of policing budgets

Abstract

Dozens of municipalities in the United States pledged to defund the police after Minneapolis police officers murdered George Floyd, an unarmed Black man, while he was in their custody. We first consider whether the municipalities that promised to defund the police actually did so. We find that they did not: municipalities that promised to defund the police temporarily reduced police budgets, only to later increase them beyond what they were previously. We then argue that two mechanisms—the electoral incentives of city politicians to provide jobs and services (what we call allocational politics) and the strength of police unions—explain why the predominant political equilibrium is one with protected police officers as a barrier to reform. We discuss several additional reforms suggested by public choice scholars interested in the problem of predatory policing.

Defund does not mean abolish policing.

—Rashawn Ray, advocate of police reform (2022)

Yes, we mean literally abolish the police.

—Mariame Kaba, advocate of police reform (2022)

1 Introduction

The movement to defund the police started several years before Minneapolis police officers murdered George Floyd on May 25, 2020, but it picked up steam afterward. One of the most significant ideas behind the movement to defund the police is that police budgets should be reduced, if not eliminated. It is more radical than previous efforts to reform American policing: defunding the police, for many of its advocates, means ending policing as it is currently practiced. One reason for a more radical approach is that police misconduct, especially misconduct directed at racial minorities, appears to be an equilibrium in US policing (Davis 2021). Thus, a significant question is to what degree the movement has changed policing institutions and whether it, like many prior reform efforts, has confronted a lack of political will to implement the reforms cities promised.

We argue that public choice theory is well suited to analyze defunding the police. Public choice scholarship has long been concerned with budgets, including police budgets. Previous studies have found that financial incentives explain the expansion of the carceral state, especially prisons (Surprenant and Brennan 2019); that federal programs to provide inexpensive military surplus equipment to departments contributes to militarization of policing, including the spread of SWAT (special weapons and tactics) teams (Coyne and Hall 2018); and that federal subsidies to local police departments have led the police to prioritize federal objectives (Boettke et al., 2017; Skarbek 2021).¹ A conclusion of this research is that subsidies contribute to some of the most pernicious problems with American policing.

We extend these insights to the movement to defund the police. First, we ask whether municipalities that pledged to defund the police actually did so, since there is disagreement over whether police were actually defunded (Hochman, 2022). To answer the question, we start by reviewing the contrasting meanings of “defunding the police.” We find that reducing police budgets is a common denominator among advocates of defunding the police. We collect data on cities where city leadership promised to reduce or reallocate police budgets, including data on what was implemented. We find that police were only defunded temporarily at most and that budgets rebounded fairly quickly.

This leads to a second question: why does refunding the police appear to be the unique equilibrium in municipalities that pledged to defund them? We argue that two explanatory mechanisms are especially plausible. The first is what we call allocational politics. Paul Peterson (1981) argues that city politicians control the purse strings, which is what they wield to get elected. Thus, party competition—which has been emphasized in much of the public choice literature on elections, going back to Anthony Downs’s theory of electoral competition—is less relevant to the politics of budgets than is politicians’ funding of civil servants, including the police. Politicians may promise to defund the police, but what gets them reelected is allocating resources to special interests. Since local politicians control the purse strings, they have institutionalized incentives to keep spending high even when that spending may be supporting predatory and racist policing.

Police unions’ strength is a second mechanism providing pressure for spending on the police. Police departments, like other bureaucracies, seek to maximize their budgets (Niskanen 1996). But what makes police especially formidable is that their unions are highly concentrated interest groups. The unions ensure that previous amounts spent on policing will be persistent. They are not only effective special interests (Olson 1965) but particularly powerful ones (Fisk and Richardson 2017; Walker 2008). Their role as a barrier to police accountability and reform is increasingly recognized (Harmon 2012; Levin 2020; Rushin 2017).

Our paper is organized as follows. Section 2 describes the defund movement and the municipalities that promised to defund the police, and it makes general observations about the extent to which police budgets were reduced in these municipalities. Section 3 argues that allocational politics and police unions’ strength explain why refunding the police, rather than defunding them, appears to be the unique equilibrium in municipalities that promised to defund them. Section 4 concludes by discussing how defunding the police (specifically, efforts to control them by reducing their budget) contrasts with several reforms that, like defunding the police, have proven challenging to implement.

2 Defunding the police: an empirical investigation

2.1 “Defunding the police” is a menu of policies

The defund movement mostly aims to influence local governments since, in the United States, the federal government does not directly regulate local police (the police power is held by state governments).² Federal influence over policing typically comes through offering of grants to encourage certain practices. For example, as a presidential candidate, Joe Biden promised an additional $300 million to the Department of Justice’s Community Oriented Policing Services program.³ Since federal politicians can exercise little power by defunding the police, few of them support it.⁴

Activists called for abolishing the police after Michael Brown was killed in Ferguson, Missouri, in 2014, but “Defund the police” would not become a rallying cry until George Floyd and Breonna Taylor were killed by police in 2020 (Movement for Black Lives 2020). The movement’s focus has been on reducing funding for traditional policing, reallocating those funds to social services, and delegating certain services traditionally handled by police, such as mental health and homelessness responses, to different service providers (Thompson 2020).

One of the most prominent advocates of defunding the police is sociologist Alex Vitale, author of The End of Policing. Vitale (2017) argues that the problems currently affecting American policing cannot be successfully addressed through reforms that do not alter the foundations of the problems. He asserts that measures such as implicit bias training, increasing the racial and gender diversity of police forces, and attempting to improve relations between communities and the police have all failed. Thus, starting from scratch is necessary.

But while some defund advocates, such as Vitale, favor abolishing the police, others favor partially reducing its funding. A shared goal appears to be reallocating at least some of the funding currently spent on policing to other ends, including mental health, housing, medical care, and job training. The justification is twofold. First, it is primarily police departments, the only public agencies that take service requests 24/7, that are called upon to pursue these ends. By dedicating alternative organizations to these problems, the need for police would be reduced. The second part of the justification is that advocates hope that the reallocated funds will address the causes of crime (such as educational and income inequality), thereby further reducing the need for police.

In addition to reallocating police budgets, defund proponents advocate reducing the scope of police activities. For the abolitionists, the scope would obviously be reduced to nothing, while for non-abolitionists, it would shrink less drastically. Thus, according to sociologist Patrick Sharkey, “Two questions that could guide the reform movement are ‘what is it that police actually do?’ and ‘why do we need armed police to do it?’” (Thompson 2020). Given the frequency of calls for service in response to mental health-related issues and the number of times police responses have ended in tragedy, defund advocates frequently cite mental health as an area in which nonpolice organizations should play the primary role (Movement for Black Lives 2020).

Proposals for altering the role of police include removing the police from schools (La ColectiVA 2020) and having alternative agencies respond to certain kinds of problems such as substance abuse (Vitale 2017), domestic violence (McHarris and McHarris 2020), and homelessness (Thompson 2020). While the police’s lack of specific training in addressing these issues and their tendency to rely on violence are reasons given by defund advocates for why police should not be involved, the primary reason is that, by their nature, they do not have the tools to solve these problems. Giving lawful commands, issuing citations, and making arrests are inadequate tools for addressing the causes of those problems, so it is too much to ask that the police solve these problems. At best, they can respond to the manifestations of these problems when the manifestations (behavior that may be considered criminal by police) result in violence or public nuisances. At worst, they can exacerbate the associated harms.

While advocates of defunding the police seem to agree about what police should not do, there is some disagreement on what police should do. For some, police should focus on tasks that are traditionally thought of as their domain, particularly preventing and solving violent crimes. Ray (2020) notes that police officers “respond to everything from potholes in the street to cats stuck up a tree” and suggests that “other government actors should be responsible for these and receive adequate funding for doing them.” According to this line of reasoning, it makes sense for police to focus on violent crimes because these are a serious problem; after all, according to the FBI, 37.7% of homicides, 66.6% of reported rapes, 69.6% of reported robberies, and 47.5% of reported aggravated assaults were uncleared by police in 2018 (FBI 2018).

By contrast, those who see little role for the police appear to envision a world in which criminal law enforcement would become unnecessary with the right mix of social spending. For example, when Vitale was asked in an interview what would happen if a car were stolen in a world in which policing is abolished, he replied, “If we want to reduce vehicle thefts, the first time that we come in contact with [the thief], we’ve got to start trying to address what’s driving their problematic behavior,” such as drug abuse (Pauly 2020).

2.2 Municipalities that promised to defund the police refunded them

Having acknowledged the diversity of perspectives within the movement to defund the police, in what follows we focus on two aspects of defunding the police: reducing funding and reallocating resources to other services. Both aspects admit a range of potential choices. Reducing funding ranges from full abolition of traditional policing to modest reductions in police budgets. It could include reducing expenditures on military-grade police equipment and SWAT teams, both of which are associated with declining liberties (Coyne and Hall 2018). Reallocating funding can mean redirecting it to social services (such as public housing, mental health programs, drug addiction and counseling programs, and education programs) or alternative public safety initiatives and organizations (including a dedicated mental health response) (Ray 2020; Sharkey 2020).

We collected data on municipalities that pledged to defund the police. We note where the commitments were uncertain. For example, on June 7, 2020, members of the Minneapolis City Council pledged to disband the Minneapolis Police Department, only to later change their minds. By June 2021, 16 cities pledged to reduce their police budgets (Holder 2020a). This includes cities with some of the largest police departments in the country (listed in Table 1). The mayor of Los Angeles pledged to cut $150 million from the police department,⁵ the mayor of New York City $1 billion,⁶ and the mayor of Austin, Texas, $21 million.⁷ Much of the budget reduction came in the form of postponing future hires and cutting overtime pay. Cutting school resource officers (SROs) was another popular form. The superintendent of Portland public schools discontinued the use of SROs, and Rochester, New York, cut five of its twelve SROs (Holder 2020a). Washington, DC, public schools now consult with a third-party security provider rather than the police.⁸ Our focus is on the how spending changed for the year following the promises that were made, so the analysis is focused on the relatively immediate impact of George Floyd’s murder on spending on policing services.

Table 1 Implementing the defund agenda

Reallocation occurred in several cities. Austin, Los Angeles, and New York took the most substantial steps to reallocate funding. Austin cut its police budget by 35% from $434.5 million to $290 million (more than what was promised) and reallocated those funds to social services, with the largest amount going to improving housing, followed by COVID-19 response and health care, general social services and community programs, family services, and mental health services (Holder et al. 2020). Money was taken from cadet classes ($13 million), overtime payments ($2.8 million), records management ($1 million), and other police services (approximately $4.7 million).

Since these cities made these promises, however, most have changed course and increased police budgets. Instead of reallocating resources, they have simply provided additional government spending for nonpolice ends. In FY 2021, Minneapolis passed a budget that maintained police staffing but allocated about $8 million to mental health interventions and alternatives to police responses (Holder 2020b); in San Francisco, the Dream Keeper Initiative, intended to reduce disparities in income and involvement in the criminal justice system between Black and White residents, was financed through additional spending.⁹ While most of the cities who pledged to cut police budgets reduced those budgets initially, this likely had more to do with falling revenues in 2020 than a commitment to defunding the police. Several cities who received federal COVID-19 relief funds spent a portion of them on policing, such as Los Angeles, which increased the police budget by 3% after receiving federal aid (Rao 2021), and New York, which used $200 million of its relief funds for additional police spending (Mays 2021).

An additional issue for defund activists is how to ensure that funds are reallocated to their own ends. The $150 million reallocated in Los Angeles, for example, has been spent primarily on back pay for police, and some proposals would spend it on sidewalk repairs, tree trimming, and other maintenance projects (Akinnibi et al., 2021).

In several cities, the primary budget items to be cut were future cadet classes and hiring and training of new officers. For others, the cuts were combined with reallocations of resources and in some instances, cuts to social programs. Consider the touted $1 billion in cuts to the New York Police Department (NYPD). According to Goldenburg and Anuta (2020), about $300 million was cut by not replacing 2300 expected departures from the police force; $500 million in spending on SROs was transferred to the Department of Education; and $20 million was cut from homelessness response and mental health response. Moving crossing guards out of the NYPD and canceling upcoming cadet classes are examples of the large portion of supposed cuts that were really transfers to other parts of the city budget; the same personnel were paid to do the same tasks.

2.3 Several of the cities implementing defund experienced large increases in crime

Critics of defunding argued that crime would increase if budgets were slashed and there was no clear alternative to standard policing. Further, critics suggested that even the mere proposal might lead to a “Ferguson effect” (Mac Donald 2016), in which officers disengage from proactive policing, leading to higher crime. While we take no position on any causal connection between reducing police budgets and subsequent increases in crime, nor do we claim that the increases were limited to these cities, it is undeniable that crime increased starkly. Table 2 displays the number of homicides in defund cities from 2019 to 2021.

Table 2 Homicides in defund cities, 2019–2021.

2.4 Defunding as a political move

One politician whose career was not hurt and perhaps was even helped by supporting defunding the police was Austin City Council member Greg Casar. He prominently supported the defund movement, was endorsed by Congresswoman Alexandria Ocasio-Cortez, and won the primary on March 1, 2022, giving him the Democratic nomination to represent the 35th district of Texas in the US House of Representatives. While most city politicians who supported defund have not yet had to face reelection, some have, and the evidence is mixed: support for defund neither always tanked electoral chances nor always ensured reelection. In Minneapolis, for example, of the three city council members who famously spent thousands of taxpayer dollars per day on personal private security while supporting defunding the police (Lyden 2020), one was reelected, one lost reelection, and one did not run for reelection but was replaced by a candidate who favored a ballot initiative (later rejected by voters) that would have replaced the police department with a department of public safety (Panetta 2021). Three other council members who favored the ballot initiative were defeated, but two others who favored it were reelected and one who opposed it was replaced by a candidate who favored it. Jacob Frey, the mayor when George Floyd was murdered, also won reelection. Clearly, where one stood on questions of police reforms was not a decisive citywide factor in Minneapolis.

The same is true for other cities. Members of the Los Angeles City Council who supported budget cuts and who faced primary elections on June 7, 2022, were all reelected (LA Times 2022). Several of them faced challengers who were strongly in favor of defund (and featured in an article in the Los Angeles Times) (Zahniser 2022). The winner in the primary for mayor was Rick Caruso, who served on the police commission and promised to add 1500 officers to the Los Angeles Police Department (Oreskes 2022). In New York City, former NYPD officer Eric Adams replaced Bill de Blasio as mayor. The 2022 recall of San Francisco district attorney Chesa Boudin may indicate voters’ lack of enthusiasm for certain progressive reforms. Generally speaking, while voters have rejected defund and supported anti-defund candidates, having supported defund has not been a political death sentence and may have been helpful to some politicians seeking higher office in particularly progressive regions.

3 Explaining the refunding equilibrium

All economic explanations assume that people maximize: they do what is in their interest (Leeson 2020). This goes for everyone, including politicians. As Buchanan (1984) argues, politicians are like anyone else: they do what is in their interest, not necessarily what is tin the public interest.

The reason to assume maximization is to focus on the institutions that structure individual incentives. An economic analysis of police reform thus starts by recognizing that the socially costly equilibrium (namely, racist, predatory policing) reflects maximization. Saying that does not mean that it is good for society but invites consideration of the ways in which institutions give rise to it. Here, we offer a few mechanisms that contributed to the refunding equilibrium, including police unions’ strength and political incentives to provide policing as an infrastructural good.

3.1 Political incentives: allocational politics and spending on police

Even though defund was not especially popular with voters, many politicians at least outwardly supported it; the empirical evidence above makes that clear. Most likely, they honestly believed that they needed to do something about the use of excessive force by police. But why did they generally refund the police?

One answer has to do with the incentives for local politicians to spend on what they see as the public good. In City Limits, Peterson (1981) develops a public choice perspective on city politics. Since cities do not control borders, people and capital move freely between them. As a consequence, Peterson contends, city politicians and public administrators have incentives to choose policies that benefit all city residents, such as infrastructure improvement, and all benefit from policies that increase the tax base. Peterson argues that the important policies in cities are “allocational” policies, such as those determining which groups get public sector jobs and contracts and which benefit from public infrastructure investments.

If Peterson’s theory is correct, then we should expect that city politicians choose to support policies that increase spending on police. Voters ultimately want certain outcomes (Ansolabehere and Snyder 2000). Those outcomes include reduced crime, even if it comes with police misconduct.

The problems arise when we consider that more spending on crime does not necessarily bode well for those most likely to interact with the criminal justice system. The desire to protect those most likely to experience predatory policing, including Black people, comes into conflict with broader citizen support for policies to increase funding to police: the typical voter wants higher levels of spending on police because reducing crime is a public good, and if they do not experience the bias in policing themselves, they have few self-interested reasons to support reductions in spending on policing. They may even think that defunding the police undermines provision of a public good, even though the provision involves clear harms for many groups who experience bias in policing.

So why defund the police temporarily, only to refund them later, if indeed voters typically prefer higher levels of funding on policing services? For some politicians, leading the charge to defund the police solidified their progressive credentials and likely benefited them politically since, among Democratic voters, hardline progressives tend to vote more and contribute more money to campaigns (Pew Research Center 2021). Things seemed to have worked out well for some politicians, such as Austin’s Greg Casar. Thus, the choice facing politicians in progressive cities was either to appear to do nothing about police violence after mass protests or to do something that sounded radical and had a chance of reducing police violence (and a chance of having no effect, which would mean a fiscal gain); at worst, crime would follow the reform, in which case their support for defund would be forgiven (or forgotten) by voters. Indeed, recent evidence suggests that the Black Lives Matter protests did not necessarily have much of an impact on voters. Engist and Schafmeister (2022) find that Black Lives Matter protests did not have a large impact on registration. This evidence would be consistent with progressive politicians pledging to defund the police but not doing much to actually defund police.

These same progressive politicians had a relatively free hand and electoral incentive to turn around and increase spending. Budget decisions can be changed annually. Reducing police budgets did not involve a constitutional change or otherwise tie the hands of policymakers in the future and therefore could be easily reversed. In the face of increasing crime in 2020–2021, especially violent crime, remaining committed to a course of reduced police budgets (particularly when the major source of budget reduction came from reducing the size of the police force by replacing officers at lower rates than attrition) became increasingly politically costly.¹⁰ The public, which never showed majority support for defund, became increasingly hostile to the idea (Parker and Hurst 2021; The Economist/YouGov 2021). Politicians and strategists still in support of defund now tend to avoid the defund terminology, recognizing that it is a political loser (Chaggaris 2021). Politicians promised to defund the police, but soon their electoral incentives to spend to reduce crime came to dominate, and that also happened to be what they realized was a winning electoral strategy.

3.2 Pressure groups: police unions as barriers to reform

Police unions provide a complementary reason why it is challenging to defund police. One question to ask first is why police unions might have a disproportionate influence in city politics. Edward C. Banfield’s (1970) Unheavenly City contends that most groups in cities have some representation in city politics. But public choice theories suggest that not all groups will have equal advantages. According to the public choice theory of group competition, groups with more resources and a more unified interest will likely have advantages in politics, provided they are not too large in terms of the number of members (Olson 1965). Police unions check these boxes: there are many police but not too many of them, and they have very clear interests and substantial resources. They also, despite criticism, have an important role in providing law and order. Thus, police unions are likely to have disproportionate influence in city politics.

One way to think of the social costs of police unions is that they undermine competition in labor markets and hence increase costs. In a competitive market, wages reflect demand and supply.

Previous reform efforts provide insight into these barriers to reform. Large-scale reforms to policing are very costly for city governments because their collective bargaining agreements require them to negotiate new agreements with unions if they make changes that affect working conditions. Such agreements do not restrict state legislatures, which is why it has overwhelmingly been state governments that have implemented the 8CantWait policies. City politicians are quite limited in what they can do without having to negotiate with the police union. One thing they can do is reduce the police budget, at least on certain margins, making defunding the path of least resistance. (Other options that would be still less likely to trigger negotiation, such as additional training, would require greater spending, which was difficult to muster during the fiscal austerity resulting from COVID-19 lockdowns.) In contrast, for example, while private security agencies have long used video-recording devices to protect themselves from liability issues (McLeod 2002), many police unions have successfully impeded their adoption because changes in working conditions must be negotiated (Harmon 2012).

That police unions are public further contributes to institutional persistence. Private sector labor unions may be a source of inefficiency, but the bargaining parties include business and labor. With public sector unions, the bargaining party is part of the government. Bureaucrats do not necessarily have the same incentives to keep costs down, including when bargaining over wages for public sector employees (Niskanen 1996). But governments also face fewer constraints in what they can bargain over, and the unique nature of police leads them to bargain over protections that undermine prospects for accountability-enhancing reforms.

While both private sector and public sector unions are interested in obtaining higher wages and better working conditions for their members, governments are less constrained in the concessions they may make to unions. For police unions, this has resulted in the negotiation of job protections in lieu of higher wages, even though experts have found that such protections undermine police accountability (Police Accountability Task Force 2016). Without market competition, employers (in this case, government) cannot easily calculate the optimal mix of wages and job protection. City managers who negotiate with police unions are biased toward job protection as a form of compensation because it is budget-neutral, at least in the short run (Fegley 2020). Compared to poorer individuals, who are more likely to have adverse interactions with the police and thus favor constraints on police services (given they experience over-policing and racist policing), richer individuals are more likely to prefer lower taxes and more protections for police. Thus, the average taxpayer may be more willing to go along with higher levels of spending on police.

Public sector unions also get two bites at the apple: they collectively bargain with city management, and they are part of the body that votes to choose that management (DiSalvo 2015). City managers do not negotiate with service providers solely as the representatives of citizen-consumers but also as representatives of the citizens who provide the services. As the recipients of local government spending, the latter citizens form a concentrated interest in support of more spending on their service, whereas the cost of this spending is dispersed among taxpayers as a whole. This has a number of implications for police accountability. For example, compared to a residual claimant over a firm’s earnings, local governments do not seek to minimize liability costs, as evidenced by the fact that municipalities generally indemnify officers who face civil judgments even when indemnification is contrary to stated policy. The nationwide sum of these judgments can be substantial ($730 million between 2006 and 2011), but they are considered a cost of doing business for police departments (Schwartz 2014). Thus, the presence of police unions offers a plausible reason why the defund movement has not resulted in substantial reductions in funding for police departments.

This institutional environment provides insight into what happened with police spending after George Floyd’s murder in cities that promised to defund police. The cuts that occurred were minor and followed paths of less resistance. Cities made cuts to cadet classes and froze new hiring because these are margins on which cuts can be made relatively easily. More capital-intensive police activities, such as mounted patrols or marine patrols, were also relatively easy to cut. Much more difficult to cut were the wages and benefits of current employees, which were left mostly untouched with the exception of funds budgeted for overtime pay. Since the vast majority of police spending is on officer pay, the degree to which city politicians are able to make cuts is limited. For example, Philadelphia City Council member Isaiah Thomas, who supported reducing the police budget, said the budget could not easily be reduced because 96.8% of the proposed budget allocation in 2021 went to salaries (Orso and McCrystal 2021). Thus, without changing the compensation schemes that are typical in collective bargaining agreements, major cuts to police budgets are impossible in the short term. The only feasible way to achieve the amount of cutting that some activists advocated was to slow hiring to a rate lower than the rate of retirement.

3.3 Additional mechanisms that may explain police reform

Several additional mechanisms might explain the persistence of police budgets. One is revenue-maximizing behavior. Policing can be a profitable business for government. Emily Skarbek’s (2021) analysis of policing suggests that budget reductions are not likely to be as effective as proponents of defunding the police think. Skarbek contends that police budgets can never fully align with citizens’ preferences. And Skarbek shows that police in general have incentives to engage in predatory revenue-generating behavior no matter what their budgets are.

Complementary to revenue maximization is the ratchet effect in spending on police. The federal government response to “crises” is an explanation for growth of government, which Higgs (1987) called the ratchet effect. In the context of policing, the ratchet effect has taken the form of federal subsidies for militarization of municipal police departments. Once police are militarized, and police are trained to use that equipment, that training and investment can lead to persistence in budgets independent of city politics of police unions.

Another plausible reason that defunding the police might not have resulted in major changes is that it is not especially popular, even among Black Americans (Saad 2020). Thus, local politicians who promised to defund the police may have realized that it was not electorally popular, which they might not have fully appreciated in the months after George Floyd’s murder.

It is also likely that variation in the structure of local institutions explains differences in reforms. Size, institutionalization, and professionalization of city councils vary extensively (L. MacDonald 2008). Since defunding the police might not be the best institutional policing reform, larger, more institutionalized, and more professional city councils might choose more nuanced policies governing police. Another possibility is that rules that provide greater minority representation influence the scope, pace, and content of police reform.

A mechanism related to local government institutions is partisanship or ideology of local politicians. The reason we have not focused much on political parties is because most local elections are nonpartisan. Still, partisan or ideological control of local government is often significant in explaining policies chosen at the local level (Trounstine, 2010). Variation in the partisan or ideological composition of city governments may explain variation in the level of funding of police as well as willingness to explore alternatives to defunding the police.

4 Discussion and conclusion

We argued that public choice, with its emphasis on budgets, offers insight into defunding the police. Defunding the police is one of several reforms, many of which have also been analyzed from a public choice perspective. We consider two policy reforms that we see as complementary to defunding: neighborhood policing and greater reliance on quasi-markets to provide policing services. We discuss how these complementary reforms may lead the defund agenda to include decentralizing policing and contracting out policing services.

Reforms such as defunding the police are examples of what public choice scholars refer to as starve-the-beast policies (Bartlett 2007): by reducing revenue, there is less scope for predatory behavior. But if we assume that any reductions will be partial at best, then a natural question is how to design institutions to improve police behavior, given that police will continue to have substantial resources. Elinor Ostrom’s public choice research agenda on policing suggests some of the institutional reforms that could improve police behavior: reorienting policing jurisdictions to focus patrol officers on neighborhoods; improving public participation to increase responsiveness of police to neighborhoods; improving processes for monitoring police; and establishing more robust frameworks to address conflicts between police and citizens.

Here, we briefly discuss how Ostrom’s (likely complementary) reform proposals relate to defunding. Ostrom’s research on policing questioned consolidation, or scaling up of police services. Proponents argued that scaling up policing services of several local jurisdictions into larger municipal police forces would reduce duplicative expenses and increase the quality of services by improving recruitment and training, supervision and organization, the use of technology, and the consolidation of records, among other channels (Durgin 2014; Fruecht 2018). Ostrom’s research suggested that neighborhood policing (in which police are assigned to specific neighborhoods so they can develop better relations with community members) can improve the quality of policing service, as measured by citizens’ satisfaction with police as well as other indicators of quality of policing, such as spending on policing services and crime reduction, in comparison to policing organized at the metropolitan level (Ostrom et al., 1973; Ostrom and Whitaker 1974). In their study of policing in Black neighborhoods in Chicago, Ostrom and Whitaker (1974) found that residents in neighborhoods patrolled by independent departments rated them as more responsive than those living in comparable neighborhoods patrolled by the Chicago police, even though the former spent less than one-tenth as much per capita on policing services.

Like defunding the police, neighborhood policing has proven challenging to implement. Communities seeking to scale down policing have had a hard time acquiring meaningful control over police, including their budget decisions (Boettke et al., 2016; Fegley 2021). In addition, there are downsides to neighborhood policing; for one thing, homeowners have disproportionate political power in local elections and an incentive to block “investment in serious foot patrols or ‘community’ policing in poor areas,” thus exacerbating racist policing outcomes (Lacey and Soskice 2015: 457).

Our view is that because defunding the police is often interpreted to mean there will still be police, a fruitful area of research would be to see how partial defunding can be combined with reforms envisioned by proponents of neighborhood policing. One possibility is defunding the police, Ostrom style, which would involve modest budget reductions with a focus on scaling down policing to the neighborhood level. This may enable neighborhood policing to capitalize on support for reform. Policy-process scholars emphasize the need for policy windows to open before major reforms can occur. The defund movement has quite clearly opened a policy window, but the specific reforms it pushes were not as popular, or as feasible, as many proponents envisioned. To the extent that the window remains open, it may be possible to view defunding through a more Ostromian lens.

Another possibility is contracting out to private providers in hopes of improving accountability. While several economists in the public choice tradition have treated policing as a local public good (Ostrom 1973; Tullock 1987), policing services are routinely provided privately (Clark and Powell 2019). The large and growing market for both in-house and contracted security demonstrates the demand for private policing.¹¹ Indeed, the market for policing services suggests that property owners are able to internalize the benefits of providing a safe environment.¹² Government provision of policing, rather than any inherent aspect of policing itself, is what renders it a common pool resource (Benson 1994).¹³

Like neighborhood policing, the quasi-market approach could be considered within the broad agenda of defunding the police: it reimagines policing and is motivated to reduce overall spending on it. One such approach is contracting. Contracting policing and security services, by both government and private entities, is already commonly practiced, but barriers to entry keep it from being used more widely (Benson, 1998: 18; Meehan & Benson, 2015).

One quasi-market proposal that has yet to be tried in policing but has a track record in schooling is the use of vouchers and educational tax credits. “Police choice” mimics various school choice policies and is meant to encourage a greater connection between the recipients and providers of policing services (DeAngelis 2018). For example, individuals could be provided vouchers (or tax credits) that they spend on the policing service of their choice. This would enable them to hold service providers more accountable by financially punishing those that provide unsatisfactory service. Such a measure could be ideal for defund advocates, as it would transfer to individuals the ability to defund the police.