Take it From the Cops*

(Written with Melissa Schober)

This post has been updated. For details, go to the end.

Let’s talk about cutting police budgets.

The Baltimore City Council is preparing to hold hearings on the next fiscal year’s budget in the midst of massive social resistance to police brutality. Now is the time for us to fundamentally re-think the fact that we devote nearly half of the city’s financial resources to the police department.

There is an exceptionally strong moral case to be made for dramatically reducing, if not entirely eliminating, social investment in police forces as we have them. Policing, at least as it is done in the United States, tends to disrupt communities, terrorize marginalized people, and reinforce white supremacy.

Even leaving these critical concerns aside, the financial impacts of policing do not make a strong case for continued high levels of social investment. The most obvious concern that someone might have about cutting police spending would be that doing so would ultimately hurt the city by increasing crime rates.

The evidence for police cost-effectiveness

One of the best studies of the impact of police spending on the social cost of crime was published by the RAND corporation in 2010. The study is quite sympathetic to increased police spending, concluding that “investments in police personnel generate net social benefits” (18).

Its conclusions, however, need to be examined in light of alternative options for spending scarce resources. Let us grant that the conclusions of the RAND study are true, at least for the sake of argument, and that increased police spending tends to reduce crime. Even if this is the case, it is not the best use of city funds.

The case that the study considers in the most detail is a 2005 proposal for a 10% increase in the size of the Los Angeles police force. The study estimates that this would cost $125-150 million per year, and reduce the costs of crime by $475 million per year (see table 5, p. 12). Taking the midpoint of the cost estimate, this would mean that for every dollar Los Angeles spent increasing its police force, it could expect to save about $3.50 in reduced costs of crime. This calculation assumes that police are as effective at reducing crime as the average effectiveness found in studies of their impact, and only calculates the reduction in costs associated with reported crimes. The RAND study examines a variety of different scenarios and assumptions (including in cities other than Los Angeles), which result in different cost-benefit estimates. The highest ratio in the study is based on the assumption that increasing police presence would have the same effect on unreported crimes that it does on reported crimes, and that police are as effective at reducing crime as the highest available estimates. Under these very favorable assumptions, the yearly savings for Los Angeles are estimated to be about $1 billion, meaning that the city could expect to save about $7.40 in costs associated with crime for each dollar it spends increasing police. The study does also look at the reverse, expected effects of reducing police. For instance, it estimates that laying off 12% of the police force in Toledo, Ohio in 2009 saved the city $6 million in policing costs but inflicted $32 million in increased costs of crime, losing the city $5.30 overall for every dollar it saved in policing costs.

The RAND study does an admirable job in attempting to capture the full costs of crime. These costs are social costs – they are costs to the people of the city as a whole, not just to the government. These costs also include both “tangible” costs (lost or stolen property, cost of incarceration, cost of court procedures, etc.) and “intangible” costs (such as fear and trauma that result from crime victimization). The RAND study draws on cost estimates of these intangibles from other research that used approaches like asking people how much they would pay in increased taxes to reduce their risk of being victimized by certain crimes, or the effects of crime rates on what people are willing to pay for a home in certain neighborhoods rather than others. We are not wholeheartedly endorsing an approach that attempts to reduce something like trauma to a dollar value, but the point is to take the purported benefits of policing in defenders’ own terms.

It should also be noted that the numbers that the RAND study uses and which we discuss here are all not only best-guess estimates but also represent an estimate of the marginal impact of increasing or decreasing police presence, from our current situation. If Los Angeles had had twice as many (or half as many) police, estimates of the impact of additional/fewer police would be different (the author notes this – see p. 10 and note 14 on p. 11). That said, we do not have reason to believe that Baltimore is exceptional among US cities for the ratio of police to total population. As of 2016, Baltimore had 40.6 officers (47 total personnel) per 10,000 residents. This was the 8th-highest ratio in the country, and higher than the average for cities with 500,000 or more people (24.3 officers/10,000). If anything, we should expect marginal returns on additional police to decline, and so with Baltimore’s already-high ratio of police to population, the effectiveness numbers we use may well overstate the likely impact of increasing or decreasing the police budget.

Comparing policing to other ways to help our communities

When we decide how to spend money, however, we need to look not just at the cost-benefit ratio of a particular program, but at its benefit relative to other things on which we could spend our money. One limitation of the RAND study is that it does not compare the cost-benefit analysis of police spending to other possible spending. There are a number of options that seem to be better than police spending, and address critical needs in Baltimore.

For instance, Governor Hogan’s recent unconscionable veto of legislation that would have funded the Kirwan Commission’s recommendations has again brought to the fore inequities in funding for children and education. Investing in youth, perhaps especially early childhood interventions, promises greater benefits than investing in police. A recent study published in The Journal of Political Economy used a randomized control trial, which followed participants from infancy through their mid-30s, to evaluate early childhood programs for disadvantaged children in North Carolina. The programs included services such as free childcare for parents, parental education, health and wellness screenings and referrals for children in care, social-emotional learning programs for children, and nutritious meals and snacks. The study concludes that the state saw $7.30 in benefits for each dollar invested in the program. This is more than twice the expected benefit of policing on the RAND study’s core assumptions, and approaches even the absolute most optimistic estimates of the benefits of police spending.

Baltimore also faces well-known challenges with homelessness, housing instability, and the quality of people’s homes. “Housing first” approaches, which provide stable housing to people as well as supports for those who have mental health challenges and/or substance use disorders – but which do not require that individuals stop using substances as a condition of receiving housing – have shown great promise. A 2016 assessment of a “housing first” pilot program in Ireland that provided people who were homeless and also struggled with mental health issues or addiction with stable housing as well as support for other challenges found that for every pound invested in the housing first service in 2014, there was a social benefit of 15.06 pounds (about $23.60; p. 110).

A 2019 review article in The Journal of Housing and the Built Environment summarized a number of recent studies of how housing programs could affect health outcomes. Baltimore City has a disproportionately high rate of asthma. An asthma program in Massachusetts was estimated to save $4.00 in social costs for every dollar spent on the program (sec. 3.2.1.). Maryland’s HIV diagnosis rates are the 6th highest in the nation, with a large number of those cases in Baltimore. An assessment of a program that supported housing for people living with HIV found a savings of $4.80 for every dollar spent (sec. 3.2.2.).

Some people in Baltimore are especially concerned with crime committed by youth. The perception that youth crime is rampant may well be driven more by fears, perennial discussions of non-violent activities that some residents find a nuisance (like dirt bikes and “squeegee kids”), or racist narratives about young Black people than reality.** Even for young people who do commit crimes, the evidence is that we have better ways to respond to them than by using the police, courts, and detention centers of the juvenile “justice” system. The Washington State Institute for Public Policy (WSIPP) has published cost-benefit analyses of a large number of juvenile “justice” interventions. Like the RAND report, these attempt to consider the total social costs and benefits of the programs, including the impact of reduced crime. WSIPP found that for every dollar spent on functional family therapy for youth released from confinement, society reaped $18.75 in benefits and for every dollar spent on problem-solving drug courts, there were $53.66 in benefits.

Early childhood education/intervention, housing programs, and therapy are only three possible areas where money could be spent otherwise than on policing in Baltimore, but all of them promise returns to our city that are comparable to or higher than the 3.5:1 ratio that is the core estimate in the RAND study, and comparable or even higher than the much more optimistic 5.3:1 or 7.4:1 estimates.

Inflicted upon persons who had broken no law, disobeyed no order, made no threat***

In addition, it is also important to note the costs that the RAND study does not include. The study notes that “some police activities, such as actions involving excessive use of force or racial profiling, may generate social costs” (15). Given recent events, this reads like a horrific understatement. Unfortunately, the RAND study makes no effort to estimate these potential costs, and we are not aware of any existing studies that do. However, the intense reaction to recent incidents of police brutality speak to the pervasive terror that our current police institutions inflict on many people, especially members of marginalized groups, and of course especially Black residents of Baltimore. For too many people, police are a source of threat rather than a source of safety. Given that the study estimates that the vast majority of costs of violent crime are represented by the intangible burdens of fear and trauma – 66% of the cost of homicide, 80% of the cost of serious assault, and 94% of the cost of rape (table 3) – it is plausible that if we could place a dollar value on the social costs of police terror, it would be frighteningly high.

Nor does the RAND study include all the costs of law enforcement even when it is working “well.” The tangible costs it uses include the costs of courts and incarceration, but do not take into account the ways in which the carceral system disrupts families and lives, and it is not clear whether they include the costs incurred when formerly incarcerated individuals face difficulty re-entering the workforce. These costs are, again, difficult to calculate and we are not aware of anyone who has made an attempt to assign dollar values to them. But they are quite plausibly large. The Fragile Families and Child Wellbeing Study at Princeton University found in 2008 that children with an incarcerated father were 25% more likely to experience material hardship, 19% more likely to receive public assistance, and 44% more likely to display aggressive behavior, among other impacts.

Furthermore, we have followed the RAND study in focusing on crimes that are uncontroversially bad, and that we assume anyone would want to reduce or eliminate – surely no one wants to see people hurt or killed by violence, and even if our current property relations are deeply inequitable, this is probably not a reason to desire more robberies. But police are used to prevent things besides assault, robbery, and murder. Large amounts of police resources are devoted to preventing things which are crimes, but do not seem to carry serious social costs (and might even have social benefits, especially if they were not forced into illegal, clandestine, or precarious markets), such as marijuana use or illegal forms of sex work. In addition to preventing things we might have no real need to prevent, police are often used to prevent things that we, as a society, should be actively promoting. Police around the country have been using violence these past two weeks to repress legitimate protests against serious injustices – how much more effectively might we be able to redress injustice if people were not afraid to make their voices heard? What is the social cost of this?

The RAND study does suggest that there may be benefits to police not captured in cost-benefit analysis of crime reduction as well. For instance, police help provide traffic control and assistance during natural disasters. This may well be true, but we have two concerns about basing an argument for police spending on them. First, we find it difficult to believe that these benefits would be as high as the unmeasured costs of policing discussed above. Second, both are functions that could just as well be provided by personnel other than police. If Baltimore City needs to reduce traffic accidents, non-sworn individuals could be hired for traffic control. If Baltimore City needs natural disaster first responders, let us invest in more firefighters, EMTs, and similar personnel. Similar things could be said for other non-police benefits that some police may provide. We have no doubt that some police officers provide benefits to the community through things like mentorship of youth, or connecting individuals in need to social services. But these benefits could be provided as well if not better by people who are not armed and do not bring with them the threat of the state’s entire apparatus of violence. Let us thank police who try to mentor youth for their good intentions, care for their community, and attempt to make the best of a problematic position; but, then spend the money on social workers instead.

Their existence is a crime****

Perhaps some people would be inclined to say, “regardless of this cost-benefit analysis, society should act to stop crime.” Crimes violate people’s rights, and so their badness maybe cannot be fully captured in their costs, even including intangible costs. We are sympathetic to this concern – we believe, for instance, that it would be worth reducing police brutality even if to do so were costly – but we do not think it provides a strong argument for increased police spending, either. Policing is not unique in being an intervention that can reduce crime. Baltimore City, for instance, has a long-standing Safe Streets program, based on the Cure Violence model for reducing violent crime. We recently invested in the Roca program, which reaches out to youths hardest to reach with other social services, and who are most at risk of involvement in violent crime.

Nor does policing and the carceral system seem to be uniquely good at reducing crime. This is implied by the cost-benefit ratios of some of the non-police interventions that we noted above. The bulk of the benefits realized from the juvenile programs analyzed in Washington state came from crime reductions. WSIPP estimated that the savings per participant in drug courts was $2,775 – $2,490 (90%) of which was from the effect on crime; $142,814 (98%) of the total $146,222 savings from the Family Therapy program were because of reductions in crime.

The Department of Justice maintains a website of interventions demonstrated to have an impact on crime and related issues. They found that a youth diversion program in Michigan significantly reduced recidivism for youth, compared to young people subjected to the traditional juvenile “justice” system. This not only reduced crime relative to the traditional police-and-courts approach, but saved $1.8 million per year across the 144 youths served by the program – just in direct court-related costs, not in total social costs.

Similarly, the RAND report estimates that increasing the number of police by 1% is expected to lead to an 0.9% decrease in homicides (table 4). By comparison, a 2004 study published in the American Economic Review estimated that an increase of 1% in the high school graduation rate would decrease homicides by about 2% (table 11).

On the grassroots level, the Ceasefire program has been shown to reduce gun violence. The study estimated that Ceasefire reduced all shootings by 52% on Ceasefire weekends. Baltimore had 309 homicides in 2018 and 348 in 2019. If we assume that Ceasefire is not disproportionately good at stopping non-fatal shootings (there is no reason to believe that it is), then it appears likely to be much more effective, dollar-for-dollar, than policing. Assume that homicides are roughly equally distributed throughout the year – this is a gross oversimplification, but the study found that the main trends in homicides were seasonal, and Ceasefire operates quarterly, so seasonal effects are likely to cancel out). Ceasefire weekends are 3 days, quarterly, for a total of 12 days per year. Taking the average of 2018 and 2019 homicide numbers, gives us an average of 0.9 homicides per day in Baltimore. That means that a very rough estimate of the number of homicides prevented by Ceasefire would be 5.6 per year. The study found no statistical evidence that these homicides were being “displaced” to other days (i.e., people do not seem to have been saying to themselves, “it’s Ceasefire weekend, let me make sure to murder this person next Tuesday rather than today). This is a 1.7% decrease in the number of homicides, compared to the 0.9% change expected from a 1% change in police. In Baltimore, the police budget has recently hovered in the neighborhood of $500 million (it is proposed at $550 million in the FY21 budget). Not all of this is personnel costs, though that is the largest component, and of course many other costs scale with staffing levels. We do not know Ceasefire’s operating budget, but even including the value of many hours of labor from dedicated volunteers, it is almost certainly at least an order of magnitude smaller than the $3-5 million or so that it would cost to increase police staffing by 1% (or that could be saved by decreasing the police force). Massively increasing Ceasefire’s resources would likely be both cheaper and more effective than marginally increasing police numbers.

All evidence points to the greater efficiency even at directly reducing violent crime of spending on supportive, preventative, and community-based programs like family therapy, education, and grassroots violence prevention, than on policing.

Implications for Baltimore’s Budget

All of these considerations apply to the likely general effects of redirecting funds from policing to other services that are more cost-effective, carry fewer social costs, and move away from a militarized and white supremacist approach to addressing social challenges. We believe the evidence shows that even redirecting funds away from “core” police services to non-police solutions is likely to lead to better outcomes.

But, as the RAND study right notes, the devil is in the details – “not only the absolute number but also the composition of personnel changes may affect crime” (16). We would not be hopeful, for instance, about cuts to the police department that eliminated all detectives but left SWAT teams untouched.

Fortunately, a review of the proposed FY2021 budget reveals a number of especially promising places for specific cuts to the Baltimore Police Department. We hope that some of these proposed cuts will be palatable even to policy-makers who do not yet see the need to dramatically reduce policing overall.

In Service 621, the Administrative Bureau, the FY21 budget removes nine sworn positions (Operations Officer or Operations Manager) for a total cost savings of $1.3 million. However, the same budget nearly doubles the number of police officers under this Bureau (from 104 to 200), an increase of  $7.3 million (nearly double the FY20 amount of $8.3 million). We question why it is necessary to move 100 officers into the Administrative Bureau? What function(s) are these officers performing that could not be done by civilians? The movement of officers into administrative roles is especially concerning given that the FY21 budget reveals that the BPD has failed to meet several effectiveness targets: The percent of homicide cases cleared remains well below the target of 60%. Similarly, the percentage of violent cases cleared is also low — less than one in five in FY19.   

In Service 622, Police Patrol, the FY21 budget includes $354,000 for research analysts to staff four Baltimore Community Intelligence Centers (BCIC) to, among other things, analyze the “strategic deployment of police officers [and] case management services designed to prevent violence.” No new funds for patrol analysis should be made until and unless the BPD responds to the FY2016 and FY2017 performance audit which noted that the “Audit Department requested documentation to support the actual amounts reported for percentage of time patrol officers spent on proactive policing during fiscal years 2017 and 2016 of 29% and 14%, respectively. BPD was unable to provide the records to support the actual amounts reported for fiscal years 2017 and 2016” (9). As the actual amount of proactive policing declined even further – to a mere 7% in FY19, it is folly to spend a third of a million dollars analyzing data that may not be accurate, timely, or verified.

In Service 626, Data-driven Strategies, the BPD proposes to move officers from other lines to this new line, and significantly reducing the number of sworn personnel overall, from 77 to 37. However they retain the same number of Police Sergeant EID (5) and increase the number of Police Lieutenant EID (1 to 2) even as the number of officers decreases by 4.

In Service 628, the Public Integrity Bureau, we note the softening of the target of number of discourteous complaints (per 100 officers). The previous target was 3 per 100 and has been raised to 3.1 per hundred, even as the relative size of the force shrinks. Where the BPD is weaking its internal performance metrics, a budget cut or significant explanation should follow. This Bureau is staffed primarily by sworn personnel (51 sworn v. 3 civilian proposed for FY21), including 21 leadership positions. The number of non-leadership officers is reduced from FY20 (42 to 30) but the number of lieutenants remains at five without explanation. We also think it would be preferable to invest more heavily in external, civilian oversight of police integrity.

In Service 635, Recruitment Section, we again note a softening of targets. The BPD reduced the targeted number of recruitments from 3000 to 2750 and the number of recruits hired from 250 to 240 without a consummate reduction in budget. Salaries and personnel costs are increased over FY20; the reductions in this function are related to materials and supplies, equipment, and contractual services. Why is the BPD spending more to accomplish less?

FY21 creates a new service 807, the Compliance Bureau. It appears as though there are significant avenues for cost savings in this function as the BPD anticipates spending almost as much on contractual services ($16.9 million) as on salary and personnel costs ($11.8 million and $6.4 million, respectively). There are only four program compliance positions listed to oversee nearly half of this Bureau’s total funding. As this Bureau is new, over-reliance on external, contractual positions or services should be avoided. This is especially so given the BPD’s poor history of contracts management and oversight.  In addition, this new Bureau has nearly twice as many sworn personnel as civilians. If officer shortages are acute, as the BPD consistently says they are, significant cost savings could be realized by further civilianizing this new Bureau, especially in light of its accountability mission. 

Service 816, Special Operations, combines several previous functions. Nonetheless, significant cuts are warranted. In FY20, personnel costs for SWAT were recommended at $9.7 million, nearly $2 million above the previous fiscal year. SWAT failed to meet its target of assisting in 150 felony arrests per calendar year; indeed, it had not met the target since FY15. Nearly $10 million to serve a target of 20 deployments for barricade and hostage situations, 175 high risk warrant services, and 150 felony arrests is a terrible return on investment. Particularly so since it is unclear how the BPD defines “high risk” warrant. In addition, the history of SWAT and other violent crimes units in Baltimore is poor, with lax oversight and accountability, particularly with regard to overtime, asset forfeiture, evidentiary standards, supervision, hiring, and training. 

In addition, it is unclear how this service differs from two line items in Service 623 (Criminal Investigation Division) which includes $6.6 million for a warrant apprehension task force and $7.3 million for targeting violent criminals. Again, we note the softening of targets without a consummate reduction in budget or explanation. The FY20 target for percent of violent crime cases cleared was 40%; it has been halved to a mere 20% in FY21 even as the overall funding for this Division increased by $7 million.  Similarly the target with regard to homicide clearance was weakened, dropping from 60% to 59%.

Service 853, Patrol Support Services, includes $3.7 million for community and youth outreach but no output, efficiency, or outcome measures related to this expenditure. We note that there is only one position listed which looks related to community and youth outreach (“Social Prog Administrator II” at a salary of $66,752). Rather than expend nearly $4 million on ill-defined community and youth outreach under the BPD’s budget, this money could be cut and redirected to the Health Department to increase school health services (some of which may be Medicaid-reimbursable, which stretches scarce dollars), to the Mayor’s Office of Employment Development to improve or expand YouthWorks, to Recreation and Parks for programming, or to the Enoch Pratt Library System to assist students this Fall with materials as distance learning remains likely. Further, we note that this service includes 53 sworn positions; civilianization of outreach positions with individuals actually training to provide youth services (e.g. LCSW-Cs or LCPCs) is a better use of funds likely to have a higher return on investment. 

Let’s not just talk about it.

We know that Baltimore City is unlikely to dramatically cut its police budget. There may be room to quibble with some of the elements we’ve pointed to as likely candidates for cuts. However, we should stop pretending that resistance to cutting the Baltimore Police Department is clearly a matter of sober, realistic political and economic calculation, rather than a matter of giving in to entrenched interests, fear, and the grinding inertia of the way that things have been done for so long.

* With apologies to our friend, who tried to get this chant going at a school funding event in Annapolis, but who we’re not naming so she won’t get @-ed without warning.

** This sentence originally referenced a 2018 report by the Abell Foundation that claimed that while crimes committed by juveniles had been declining overall, violent crimes by juveniles were increasing. Jenny Egan, Chief Attorney for the Juvenile Division at the MD Office of the Public Defender, brought to our attention several problems with this report. In particular, Egan pointed out to us that the report’s claim that violent crimes by juveniles had increased was misleading. The increase was not statistically significant; and, the increase only appears if you compare cases in juvenile court without considering cases in “adult” court that originated in the juvenile “justice” system (i.e., children being tried as adults). As Egan wrote in notes she shared with us, “[w]hen you compare apples to apples and look at children whose cases originate in the juvenile there [has] been a 38% decrease crimes of violence between 2011-2018.” As the updated paragraph notes, we believe our main point stands despite our citation of inaccurate statistics in this report – if you are concerned about youth who commit crimes, there are better ways to address the issue than through police. We regret inadvertently giving power to an inaccurate and arguably racist framing of the issue, and we appreciate the analysis Egan shared with us.

*** With apologies to the 1968 Walker Report.

**** With apologies to Rage against the Machine.

Image of the a 2015 protest in Baltimore by “Veggies” via Wikimedia Commons CC-BY-SA 3.0

Oh, by the way…

If you’ve read this and liked it, and you live in Baltimore City, you should tell your City Council member and the mayor what you think about the police budget.

We’re just some well-meaning white people, but other folks have been working to reduce violence (police or otherwise) and create racial justice in Baltimore for a minute; if you buy what we’re saying here, you should give them your support (financial or otherwise):

Baltimore Ceasefire

Organizing Black

Citizens Policing Project

Restorative Response Baltimore

Baltimore Community Mediation Center

This list is, of course, not complete, and we don’t want to imply that other groups aren’t also doing good work! Just, if you’ve gotten this far, you’ve read enough.

This post has been updated

8 June 2020:

  • Revised paragraph about juvenile justice with additional information from Jenny Egan
  • Added paragraph clarifying that impact estimates are at current margins
  • Broke discussion of Ceasefire out into own paragraph, added detail
  • Added paragraph about social costs of police repression.
  • Fixed some typos, probably introduced more