Your browser is unsupported

We recommend using the latest version of IE11, Edge, Chrome, Firefox or Safari.

State and Local Governments are Likely Overstating Their Use of SLFRF Support for Community Violence Interventions

August 8, 2022

By Philip Rocco and Amanda Kass 

Join our mailing list to stay updated about our research.

For the last year, the Biden administration has encouraged state and local governments to invest dollars from the State and Local Fiscal Recovery Funds (SLFRF) in alternatives to traditional policing, ranging from community-based violence intervention programs to long-term violence prevention strategies.

State and local governments have been slow to invest SLFRF dollars in these programs. As we reported in June, as of the end of 2021, state and local governments had collectively obligated only 0.14% of the first tranche of SLFRF aid (which amounts to about $79 million) on community-based violence intervention programs.1

Yet as our new analysis reveals, state and local governments’ misclassification of SLFRF-funded projects might mean that even this relatively low number is likely inflated. In fact, one in five obligated dollars classified by state and local governments as supporting community violence interventions went to support measures connected to traditional policing, including the purchase of shot spotters and sniper rifles.2

The Fiscal Classification Problem

The Coronavirus State and Local Fiscal Recovery Funds program is designed to give state and local governments an extraordinary level of flexibility to respond to numerous health, economic, and fiscal problems. Yet this flexibility, as our prior blog posts note, entails difficulties for monitoring the use of funds.

Arguably the most important difficulty is that of classifying how dollars are spent.

Under the initial Treasury regulations, all governments have to account for how they spend SLFRF dollars using specific expenditure categories created by the Treasury Department (see our first blog post for more on this classification system). Under the interim rules, which were in place for the spending data that covers March 2021 through December 31, 2021, there were 67 specific categories (starting in April 2022, the list has been expanded to 83 categories). Every project that recipient governments report has to be assigned to one, and only one, of the expenditure categories. Hence, a project can only be assigned to one expenditure category even if it could fall into multiple, and it is up to the recipient governments to assign an expenditure category to a project.

The distinction between some types of expenditures is obvious. One could hardly confuse expenditures on storm drains for an investment in a vaccination campaign.

But not all classification decisions are so easy to make. Community violence interventions (CVIs) provide a case in point.

As defined by the Bureau of Justice Assistance, CVI “is an approach that uses evidence-informed strategies to reduce violence through tailored community-centered initiatives” which “engage individuals and groups to prevent and disrupt cycles of violence and retaliation, and establish relationships between individuals and community assets to deliver services that save lives, address trauma, provide opportunity, and improve the physical, social, and economic conditions that drive violence.”

Yet classifying projects as CVIs is difficult for at least three reasons.

First, CVIs are internally heterogeneous. While the focus of these interventions is almost exclusively on gun violence (as opposed to other forms of violence), they can range from “violence interrupters” such as the Cure Violence model to hospital-based interventions designed to identify and reduce risk factors for violence.

Second, community-based strategies for violence intervention are often weakly institutionalized at the state and local level. While the practice of community-based work to reduce violence has a long history, CVIs have—until recently––struggled to find enduring institutional support in local government.

Third, the Treasury Department’s guidance documents and videos for the SLFRF program do not clearly define what CVIs are and are not. In the absence of this guidance or a strongly institutionalized definition of CVIs, it is reasonable to suspect that recipient governments may have, possibly unintentionally, applied the CVI expenditure category too broadly, including both spending on traditional policing and community-based approaches.

How We Analyzed Spending Classifications

As we previously reported, state and local governments obligated a combined $79 million in SLFRF dollars to the CVI expenditure category as of the end of 2021.3 In the first batch of data released by the Treasury Department, 88 governments reported they had obligated SLFRF aid to at least one project they coded as CVI. In total those governments reported a collective 168 projects coded as CVI.

To see how accurate that $79 million figure really was, we performed a content analysis of each of the 168 project names and descriptions submitted by state and local governments in their first Project and Expenditure Reports. Specifically, we noted whether project descriptions met at least one of two criteria, developed through consultation with subject-matter experts.

First, we assessed whether each project classified as “Community Violence Intervention” (Expenditure Category 3.16 in the first reports) proposed investing in community-driven strategies that involve at-risk or high-risk populations to reduce the incidence of violence and in an alternative to traditional policing. These strategies could include, but were not limited to, group violence interventions, community-based violence interrupters, and hospital-based interruption programs.

Second, we determined whether projects could be classified under the broader heading of community violence prevention (CVP) rather than CVI. CVP could include changes to the physical environment (e.g. Crime Prevention Through Environmental Design (CPTED), treatments such as Trauma-Focused Cognitive Behavior Therapy ®(TF-CBT) and Multisystemic Therapy®, strengthening economic supports through job training or summer youth job programs, connecting youth to caring adults and activities such as mentoring and after-school programs, as well as general income support programs.4

For those projects we did not classify as CVI or CVP, we developed additional descriptive categories to capture projects aimed at supporting traditional policing or public safety, mental health services, support for criminal courts and victim services.

Finally, we also took note of all instances in which the project description was too vague to classify. In addition to having discretion over how to categorize projects, governments also have latitude in how much of a project description they provide. Some governments included details about the program they were using SLFRF aid for, rationale for that program, and the policy aim. Others, in contrast, gave vague descriptions and/or little information; for example, one city’s project description was simply “Crime prevention services and gun violence prevention services.”

What We Found

Our analysis suggests that state and local governments are likely incorrectly classifying a substantial number of projects as CVI or CVP (see table below). Less than half of the projects could be clearly classified as CVI or CVP, and nearly every fifth dollar classified as CVI spending by the governments was spent on traditional policing.

Category Number of ProjectsTotal Obligations ($)Obligations as % Total reported under EC 3.16
Victim Services31$17,118,81122%
Police 44$14,151,78118%
Mental Health4$540,0911%
Description Unclear6$7,576,92310%
Total168$78,975,001 100%

Of the 73 projects that we determined were CVI or CVP:

  • 27 (representing nearly 23% of obligations listed under Expenditure Category E.16) met our definition of CVI (see table below).
    • This included, for example the city of Memphis, Tennessee’s $4.8 million obligation to a group violence intervention program aimed at hiring “Street Intervention Workers, Hospital Violence Interrupters, provide outreach services for youth, and other wrap around services.”
  • Another 35 projects (19% of total obligations listed under EC 3.16) could be classified more broadly as community-based violence prevention strategies.
    • This included the city of Cincinnati, Ohio’s obligation of $20,000 to its “Kings and Queens” program, which enrolls at-risk youth in a 3-month education program aimed at violence reduction.
    • Similarly, the state of Illinois obligated $60 million to a summer youth employment program aimed at reducing violence by connecting youth to employers who will provide opportunities for work-based learning, career development, and pre-apprenticeship.
  • Six projects, with obligations totaling $2.9 million, contained both CVI and CVP components.5
  • The final five projects had descriptions that generally met our definition of either CVI or CVP but lacked sufficient detail to distinguish between the two subcategories.
Category Number of ProjectsTotal ObligationsObligations as % Total reported under EC 3.16
CVI and CVP6$2,929,5974%

Among the projects that did not meet our criteria for either CVI or CVP, we classified 31 (representing 22% of obligations) under the heading of victim services. This included, for example, the state of Maryland’s provision of funding to the Victims of Crime Act (VOCA) program, which “works to improve the treatment of victims of crime by providing them with the assistance and services necessary to aid their restoration after a violent criminal act, and to support and aid them as they move through the criminal justice process.” Victim services may have some ties to, or be supportive of CVI efforts; however, for the purposes of our analysis we separated these initiatives from CVI and CVP programs.

44 projects (representing 18% of total obligations) were focused on support for traditional policing measures, as the table below suggests.

  • This included 30 projects, representing a collective $6 million in obligations, aimed at purchasing/updating police equipment, vehicles, or other technology.
    • For example, Independence, Missouri classified as “community violence intervention” an obligation of $9,739 for the purchase of sniper rifles.
    • In its first Project and Expenditure report, Macon-Bibb County, Georgia reported obligating $1.97 million for ShotSpotter technology.
  • Another 12 projects, representing $6 million in obligations, focused on police department staffing.
    • Lorain, Ohio, for example, submitted a $2.97 million project description which, while classified as CVI, included funding for “salary, fringes and equipment to hire 10 new police officers.”
Category Number of ProjectsTotal Obligations Obligations as % Total reported under EC 3.16
Police Equipment, Vehicles, or Technology30$5,974,6438%
Police Other2$1,900,0002%
Police Staffing12$6,277,1388%

Making Sense of the Numbers

Our analysis does not necessarily point to any deliberate misreporting of data by recipient governments. Rather, the data could just as easily have emerged due to genuine confusion about how to classify spending on the part of the recipient government officials who are charged with submitting Project and Expenditure Reports. For example, jurisdictions new to CVI work occasionally use the terms prevention and intervention interchangeably, despite meaningful differences between the two types of work.6

In other instances, officials may simply be classifying traditional police spending as community violence intervention because they do not understand what CVI is. As we noted above, Treasury’s compliance documents contain no formal definition of CVIs that would help recipients distinguish spending on community-based approaches from traditional public safety measures.

In the absence of clear guidance, it is reasonable to expect governments will misclassify a nontrivial level of spending in reports to the Treasury Department, especially when, as in the case of CVIs, the boundaries of the spending category are not well defined.

The upshot of this analysis is that efforts to monitor SLFRF implementation must, at a minimum, drill beneath the surface of Treasury-provided categories to analyze what recipient governments are actually proposing to do with federal funds.


1 In that analysis we counted any project that a government assigned to the CVI expenditure category as CVI. It is important to note that in taking that approach our analysis does not include CVI or violence prevention projects that governments chose to assign to different expenditure categories. As we have previously explained, governments must assign projects to one, and only one, expenditure category.

2 We thank Molly Clark and Alex Hawley for excellent research assistance on this post.

3 It is important to note that this analysis was based on spending data reported by 1,756 governments, which is a subset of all governments receiving SLFRF aid. The reporting rules vary by government type and population size.

4 To ensure that our coding scheme was reliable, we had two independent evaluators classify each project description. Based on these independent ratings, we calculated a Cronbach’s alpha was calculated for our indicator of CVI and CVP projects. The raw alpha was 0.66, which is in excess of the conventional thresholds for reliability. Following this initial coding, the two Principal Investigators reviewed and resolved all discrepancies and confirmed our final coding decisions with a subject matter expert.

5 Of these projects, two also included other elements that did not meet our definition of CVI or CVP as well.

6 It is possible that governments also used funds they classified as “revenue replacement” to support CVIs. Yet because revenue replacement dollars are not classified with the same level of specificity as other projects, it is impossible to use Treasury data to analyze how funds classified as revenue replacement are supporting CVIs.