Local Justice Reinvestment: Strategies, Outcomes, and Keys to Success


Local Justice Reinvestment: Strategies, Outcomes, and Keys to Success


Over the past six years, 17 local jurisdictions across the country have implemented policies to reduce their jail populations and costs, improve public safety, and increase the efficiency of their justice system. Through the Justice Reinvestment Initiative, sites implemented policies to address frequent front-end users, improve pretrial strategies, apply evidence-based practices in community supervision, and enhance data systems and capacity. This brief summarizes those strategies, identifies common themes across sites, documents outcomes to date, spotlights successful efforts, and highlights keys to successful reform.

Note: This brief was changed in October 2016. Figure 1, a map of local JRI sites, was updated to improve the accuracy of the locations of San Francisco City & County, Santa Cruz County, and Yolo County.

Full Publication

Over the past decade, state and local jurisdictions across the country have implemented innovative strategies to improve public safety and produce a better return on their public safety investments. The Justice Reinvestment Initiative (JRI) has played a lead role in this reform effort, facilitating a data-driven approach in which jurisdictions examine the factors—or drivers—that contribute to their corrections populations and costs and identify and implement evidence-based reforms to address these drivers, enhance public safety outcomes, and generate cost savings to be reinvested in high-performing public safety strategies. Between 2010 and 2016, the Bureau of Justice Assistance (BJA) funded JRI in 17 local jurisdictions (figure 1). In November 2014, Urban released a brief that summarized interim findings of an assessment of activities and outcomes in these local sites. This brief documents work through the completion of the initiative, identifies common themes across the local sites, summarizes policy strategies implemented, documents outcomes to date, and highlights keys to successful reform.

Figure 1. Local JRI Sites
Local JRI Collaborators


The JRI Model


The local justice reinvestment model is built on collaboration, stakeholder engagement, and strategic planning. The core of this effort is the working group, which consists of key criminal justice leaders such as law enforcement officers, public defenders, prosecutors, pretrial agency staff, judges, probation officers, county commissioners, and representatives from community organizations. This group guides the jurisdiction through its justice reinvestment activities and is led by a site coordinator who is the main liaison between the working group and the technical assistance provider and partners.


The interagency working group collaborates with a technical assistance provider, funded by BJA, to collect and analyze systemwide criminal justice data. Through data analysis, a jurisdiction identifies the drivers that are contributing to its corrections populations and costs then identifies strategies to reduce the impact of these drivers. Once the JRI working group agrees on its strategies, it can request additional seed funding from BJA for implementation. The jurisdiction implements its new policies or programs and documents the related outcomes and savings. The ultimate goal of JRI is for these reforms to reduce corrections populations and generate cost savings that can be reinvested in further effective public safety strategies.


JRI is iterative, requiring ongoing assessment of the implemented strategies to assess whether they are yielding intended results and to make appropriate corrections if not.

Technical Assistance


Local JRI Policy Strategies


Data analysis identified numerous common population and cost drivers across the 17 jurisdictions that implemented local JRI. Among these were practices that resulted in the excessive incarceration of overlapping groups of people, including those who

  • returned to jail time and again, often presenting with significant mental health, substance abuse, and/or housing needs;
  • were held before case disposition and had not yet been convicted of a crime; or
  • returned to jail for violating the conditions of supervision on probation or parole.

JRI localities implemented various strategies to address the reasons these groups of people were incarcerated (table 1). The implemented strategies are categorized below by four summarized policy goals: address frequent front-end users, improve pretrial strategies, implement evidence-based practices for community supervision, and improve data systems and capacity. Although these domains are not a comprehensive look at local justice reinvestment strategies, they provide an in-depth look at successful reforms and the challenges that accompany them.

Table 1. Local JRI Public Safety Strategies


Address Frequent Front-End Users


Four local jurisdictions identified particular people who were associated with a disproportionate number of jail bed days each year. Sometimes called “frequent front-end users,” they often struggled with homelessness, mental illness, and drug or alcohol addiction. As a result, they also disproportionately used other local resources, such as hospital emergency rooms and detox centers. Jurisdictions recognized that by targeting this relatively small group, they could reduce jail bed use and general system costs while providing services and treatment to the people who needed them most. Denver, Mecklenburg, Travis, and Santa Cruz Counties each addressed this population slightly differently, using a combination of supportive housing, case management, mental health and substance abuse services, and community outreach.




Addressing the comprehensive needs of people who frequently returned to jail required collaboration among several agencies, including the courts, jails, police, service providers, and housing authorities. In many cases, housing proved the greatest challenge. Travis County proposed to create supportive housing units with integrated services but had trouble both recruiting the targeted people and convincing housing providers to approve them as residents. Denver implemented a housing-first policy but struggled to find available housing that could be converted into appropriate supportive housing. For the first several years of its program, Denver housed many participants in hotels. The jurisdiction is working to build dedicated housing units for this population, funded independently of JRI.


Santa Cruz’s Bob Lee Community Partnership for Accountability, Connection and Treatment (PACT) program focused on connecting people to substance abuse treatment and mental health services through intensive community outreach. Providing these services was an important part of the implementation in Denver, Travis, and Mecklenburg Counties as well.




Because the people targeted by these strategies have diverse criminal justice, mental health, and substance abuse needs, the four sites faced challenges with recruiting, housing, and retaining participants. Nevertheless, Denver, Mecklenburg, Santa Cruz, and Travis Counties have all tracked performance measures that show progress toward jail bed and cost savings.

  • As of the end of March 2016, 50 percent of the 30 people housed by Travis County’s program had been rearrested, compared with 81 percent of the 80-person control group.1
  • In the first eight months of Santa Cruz County’s PACT program, participants had 70 percent fewer arrests and citations and served 50 percent less jail time than in the eight months before the program started.2 These reductions meant the county avoided paying $76,937 for program participants’ jail bed days.3 In 2015, the National Association of Counties granted PACT an Achievement Award in recognition of the county government’s work to improve services for residents.
  • Mecklenburg County had housed 60 people through its MeckFUSE4 program as of December 2015, after 2.5 years of program implementation. During that time, participants were arrested an average of 0.87 times each; they had previously been arrested at least four times over five years, and often much more frequently.5

Denver's Recovery Court


Improve Pretrial Strategies


Pretrial defendants make up 60 percent of the country’s jail population (Minton and Zeng 2015). In local JRI sites, that share reached as high as 84 percent (Cramer et al. 2014). Nine sites found that pretrial populations were a primary driver of jail growth and costs.


Three of these sites—Alachua, Santa Cruz, and Yamhill Counties—determined they had no evidence-based way to reduce their large pretrial populations because they had no pretrial risk assessments in place. San Francisco was using a validated tool that required a lengthy face-to-face interview with qualitative questions. Alachua and Yamhill Counties also found that delays and inefficiencies in case processing contributed to unnecessary pretrial detention. Other jurisdictions, including New York City and Allegheny, Eau Claire, and Milwaukee Counties, found that many pretrial defendants could be deferred from detention altogether with little risk to public safety. Reliance on monetary bail was also identified as a common problem, but it was often difficult for localities to address because of legislative and political constraints. Johnson County was one of the few jurisdictions to address monetary bail, after finding that 70 percent of people awaiting trial and 75 percent of those released pretrial were required to pay monetary bonds (Revicki, Brooks, and Bechtel 2015). The county developed release and detention guidelines that based decisions on defendants’ risk factors rather than their ability to pay.




Local sites implemented pretrial reform in several ways, including (1) implementing pretrial risk assessment tools and using them to guide decisionmaking, (2) streamlining case processing, and (3) improving pretrial diversion programs. These three main strategies included efforts to review the status of cases for people detained before trial, improve pretrial supervision, reduce reliance on monetary bonds, and implement cite-and-release strategies (i.e., issuing citations instead of arresting and transporting someone to jail).


Implementing pretrial risk assessment tools. Santa Cruz County, San Francisco, and New York City focused on implementing risk assessment tools. San Francisco had an instrument in place but had been unable to fully integrate its results into release decisions without greater stakeholder buy-in and a more efficient tool. JRI support was used to implement a new tool, the Laura and John Arnold Foundation’s Public Safety Assessment,6 and train criminal justice stakeholders—including judges, defense attorneys, and prosecutors—on how to incorporate risk assessment in their release decisions. Before implemen-tation began in May 2016, a consultant team, the Arnold Foundation and San Francisco stakeholders completed a fidelity check of the Arnold tool and confirmed its utility and implementation readiness. The consultant team is reviewing release decisions before April 30, 2016, and after tool implementation and expects to have an outcomes study drafted in fall 2016. Santa Cruz County also worked with the Arnold Foundation to pilot the Public Safety Assessment; the additional staff support and mobile computer solutions funded through JRI have helped the county increase the number of people recommended for pretrial release fivefold since the tool was implemented in August 2015.7 Finally, New York City implemented a pretrial screening tool used by supervised release program providers in all arraignment courts across its five boroughs, aiming to release defendants arrested for misdemeanors and nonviolent felonies who are not high risk for felony rearrest during the pretrial period. This supervised release strategy, launched in March 2016, uses the risk assessment to guide eligibility and the supervision level for people who may otherwise have been placed in jail.


Streamlining case processing. Several sites also focused on improving case processing. Generally, case processing strategies were designed to expedite cases, streamline processing, improve court data capacity, and identify good candidates for pretrial release. Alachua and Yamhill Counties hired a release coordinator to facilitate pretrial release, make case processing more efficient, and connect people to services. Yamhill’s pretrial release coordinator was part of a larger pretrial justice strategy in which the county implemented a risk assessment tool, created a decision matrix, and increased its supervision capacity.


Improving pretrial diversion programs. Several sites chose to improve diversion programs as alternatives to detention. Allegheny County funded a treatment coordinator to divert pretrial defendants with substance abuse needs to treatment. From January 2013 to March 2016, 702 people were referred to the diversion program; 399 completed it.8 Eau Claire County used JRI funding to implement a voluntary pretrial diversion program for methamphetamine users. Milwaukee County created the Central Liaison Unit to coordinate assessment, supervision, and case management for low- and medium-risk people diverted from detention.




Though some sites are still implementing their strategies, several sites have begun reporting outcomes.

  • Alachua County hired a jail release coordinator in 2012; from the coordinator’s arrival to 2014, the county’s average end-of-month jail population declined 14 percent.9
  • Allegheny County’s Drug and Alcohol Diversion Program served 464 clients between January 2013 and March 2016; 81 percent of the 399 who left the program did so by completing inpatient drug treatment.10
  • Eau Claire County’s diversion program for methamphetamine users served 34 people between its implementation in June 2015 and the end of March 2016, none of whom were issued bench warrants for failure to appear.11
  • Johnson County assessed 1,227 defendants between December 2014 and September 2015. The courts concurred with the recommended release guidelines or were more lenient on release decisions 80 percent of the time, and 93.2 percent of defendants were released on bond, of which 93 percent appeared for their court hearings and 89 percent had no new criminal activity.1 These figures improve on the Bureau of Justice Statistics’ national estimates in 2009, in which 83 percent of people appeared in court and 84 percent had no new arrests on pretrial release (Reaves 2013).
  • During the two-borough pilot, the number of people accepted into supervised release programs in New York City increased from 32 in November 2014 to 102 in August 2015, for a total of 602 participants.13 In March 2016, New York City expanded supervised release citywide; in the first four months of citywide implementation 962 defendants were diverted to the program, saving an estimated 93 jail beds. Supervised release is on track to surpass its goal of serving 3,048 defendants a year.14
  • Santa Cruz County assessed and released 1,183 people pending trial and found that 88.6 percent appeared in court and 97.2 percent had no new criminal activity,15 well above national estimates.
  • Recent analysis of Yamhill County’s pretrial justice strategy (implemented with funds from both JRI and the National Institute of Corrections’ Evidence-Based Decision Making [EBDM] Initiative)16 found the program reduced the county’s daily average pretrial population from 45 percent to 36 percent and that pretrial program participants had a 96 percent appearance rate for court dates compared with a 77 percent appearance rate before the project.17

Milwaukee County's Pretrial Diversion and Deferred Prosecution Program


Implement Evidence-Based Practices for Community Supervision


Six localities found that people booked for technical violations of their probation accounted for a sizable portion of jail admissions. Delaware County found probation violations accounted for nearly 14 percent of jail admissions and were the primary reason people with felony convictions were in jail. People in jail for probation violations also had longer stays—more than three times the average length of stay for other people in jail.18 Charlottesville-Albemarle also found that probation violations were affecting county jail admissions, accounting for 10 percent of the jail population.19 Likewise, probation violations accounted for 8 percent of bookings in Yolo County20 and 6 percent of bookings in Grant County.21


In addition, sites noticed that community supervision resources could be more efficiently allocated based on the risk level and needs of people under supervision. This is important as research has shown that interventions are most effective when targeted to individuals at high risk of recidivating and matched to their unique criminogenic risk and need factors (Andrews, Bonta, and Wormith 2006;Lowenkamp, Latessa, and Holsinger 2006;). San Francisco found that its standard three-year term of probation supervision exceeded the average time to failure on probation by more than 18 months.22 In Lane County, the rate of felony probation violations was relatively high; 19 percent of people violated their supervision from 2010 to 2011, and 49 percent of those violations occurred during the first year of supervision.23 San Francisco and Lane County realized that probation officers’ time could be better spent focusing on people in the critical 12 to 18 months after release, when interventions can make the most difference.




After extensive data and systems analyses, local sites implemented evidence-based practices for community supervision to curb jail overcrowding, respond more effectively to violations, and use resources more efficiently. Local sites adopted several strategies, including risk assessment tools, violation response matrices, increased staff training, and caseload reallocation.


Delaware County pursued policy strategies to reduce probation violations by at least 10 percent each year over four years through several evidence-based practice innovations, including risk assessment, improved supervision, substance abuse intervention, and a violation and incentive matrix. Charlottesville-Albemarle County implemented an administrative response matrix, which listed consistent, neutral, and proportional responses to probation violations, in order to promote behavioral change and reduce jail overcrowding. Lane County developed more specialized supervision practices tailored to risk levels determined by validated assessment tools.24 People at the highest risk level received supervision that included intensive contact standards, exclusion zones, electronic monitoring, and risk-reduction programming, while medium-risk people received programming designed to reduce recidivism, and low-risk people were assigned to the new Reduced Supervision Unit. Its staff use an abbreviated tool (the Level of Service Inventory-Screening Version) to identify stable risk and need factors and inform decisions about supervision strategies with this low-risk population (Ozanne 2015a).25 Grant County similarly reallocated probation caseloads by risk level.


Yolo County and San Francisco County worked with George Mason University’s Center for Advancing Correctional Excellence to examine how well available programs matched needs. After seeing the results of this system assessment, Yolo County set out to improve supervision and increase capacity to meet substance use treatment needs. To implement this change, the probation department received training for evidence-based risk assessment and supervision practices. San Francisco expanded its early termination strategy to shape probation terms based on risk while protecting public safety.26 This program is described in the Spotlight box below.




As reforms have gotten under way, sites have seen reductions in probation violations and broader positive systems changes.

  • Charlottesville saw total unsuccessful terminations for local (mostly misdemeanor) probation cases drop 12 percent between 2014 and 2015. Revocations for state (mostly felony) probation cases supervised in Albemarle and Charlottesville Circuit Courts fell 22 percent.27 Further, preliminary findings following the first year of implementation of the probation violations response matrix found that jail admissions resulting from both probation violations and revocations declined from fiscal year 2014 to fiscal year 2015 following increases in the previous three years.28 Analyses also documented a precipitous drop in the amount of time individuals spent in jail for violations and revocations in that year. For example, though the total number of jail bed days used for probation violations had been declining between 2011 and 2014, it dropped 75 percent from 2014 to 2015.
  • Delaware County reduced probation violations 5.5 percent.29
  • In Grant County, the reallocation of caseloads contributed significantly to a 13 percent reduction in the active probation population over the past five years.30
  • While Yolo and Lane Counties have not reported concrete reductions in probation violations, both sites trained probation officers and supervisors on risk assessment and are moving toward a more risk-based approach to community supervision.31

San Francisco's Risk-Based Probation Terms and Uniform Early Termination Protocol


Improve Data Systems and Capacity


A critical component of justice reinvestment is data analysis and data-driven decisionmaking. Many sites found it challenging to efficiently extract, collect, and analyze data from agencies’ data systems. To address this cross-cutting challenge, five JRI sites identified strategies to improve their data systems and increase their data analysis capacity.


New York City stakeholders recognized the need to improve data and information sharing among justice agencies and service providers. Greater information sharing, especially about people detained before trial, could help inform and improve pretrial release decisions. One of New York City’s strategies, therefore, was to develop a justice provider system intended to

  • pull together information across multiple city criminal justice agencies to present a more complete and comprehensive view of defendants and automate calculation of a defendant’s pretrial risk score;
  • facilitate enrollment of defendants in supervised release, alternative to detention, and alternative to incarceration or other similar programs in New York City by suggesting appropriate program matches for defendants at every point in their criminal case; and
  • create a system that provides solutions for more standardized, detailed data collection from program providers regarding client and program outcomes, to help staff in the Mayor’s Office of Criminal Justice evaluate current programs and in conducting future planning.




To improve data capacity, local sites developed data warehouses, integrated data systems, data dashboards, and jail population and cost-benefit projection tools. Mecklenburg County developed a data warehouse that integrates arrest, jail, court system, and other information from multiple justice agencies. The data warehouse also allows users to generate daily, weekly, and monthly reports and monitor system-level trends such as arrest rates and changes in the jail population. Local stakeholders, from justice and non-justice agencies, use the data and reports to inform and guide policy decisions.


Johnson County stakeholders, who had been using the county’s Justice Information Management System to collect and track data on people involved in the criminal justice system, wanted to expand the county’s reporting capability and monitor people and system trends in near-real time. To do this, the county purchased IBM Cognos business intelligence software to generate reports and data dashboards on important justice outcomes. The Crime and Justice Institute, in partnership with an economist, developed the Jail Population Policy Impact Tool.32 Johnson County is using the tool to assess and identify population drivers and cost-effective strategies for addressing them. Allegheny County also implemented the tool, adapting it to estimate the value of county recidivism-reduction programs. Allegheny worked with a second consultant to improve the county’s data-collection capacity for tracking outcomes and using them to develop and validate a county-based risk assessment.




Though the outcomes related to improving data capacity are challenging to quantify, stakeholders have provided anecdotal feedback that speaks to the effectiveness of their strategies. Data capacity strategies have informed stakeholder conversations and decisionmaking about justice issues and policy. The ability to conduct in-depth analysis has allowed stakeholders to better understand and more meaningfully discuss justice population trends. Also, data availability empowers local stakeholders to routinely monitor and evaluate the implemented programs and policies. In addition, stakeholders have found that using performance measurement tools and data dashboards help hold the collaborative accountable to working toward shared goals and outcomes.

Grant County's Data Dashboard


Documenting Outcomes at the Local Level


The local JRI sites have identified and implemented strategies to reduce local corrections populations and costs. However, documenting outcomes—particularly savings and reinvestment outcomes—has proved challenging. Preliminary findings are generally positive, but many localities only began implementation in fall 2014. Thus, though support through JRI has ended for most sites, many sites continue to collect performance metric data and consider their current reported outcomes preliminary.


One key metric of success for JRI is the savings generated by policy reforms and the dollars reinvested in public safety strategies. Most sites have not been able to identify real savings and reinvest those in the system. Nevertheless, many documented success in other forms that will live on well beyond participation in the initiative.


Capture and Reinvest Savings at the Local Level


Local sites faced significant challenges in identifying cost savings and reinvesting those funds. Though specifics varied across jurisdictions, many sites faced similar challenges at each stage of the process. These included (1) identifying savings generated by JRI reforms, (2) documenting those savings, (3) making that information public, and (4) reinvesting any savings in strategies that could further reform efforts.




Capturing savings at the local level attributable to JRI reform is not an easy task. Accurately assessing the impact of a policy requires understanding how the relevant criminal justice outcomes were projected to change without any intervention, developing a method for measuring the fiscal impact of a change in outcomes, and attributing that change to a specific intervention or policy.


The local justice systems of many jurisdictions were projected to keep growing without JRI, so some sites were able to avoid costs they would have incurred if they had not implemented JRI strategies. Eau Claire County, which faced an increasing jail population at the outset of its JRI work, avoided the need to open a new housing unit in its jail.33




Some sites aimed to identify savings from jail population reductions but struggled to achieve stakeholder consensus on marginal jail bed costs (the actual cost of housing an additional person in a facility) and actual dollars saved. Marginal cost estimates are not routinely available and are critical because average costs typically overstate savings as they include fixed costs (e.g., building operations, maintenance, etc.) that remain stable regardless of the average daily population. Stakeholders often estimate the costs of specific components (e.g., food and medical treatment) differently, further complicating calculation of the actual cost of a jail bed day. Further, marginal costs are typically modest. Significant savings occur only when a locality reduces staff and facility costs by closing a jail unit.




Jurisdictions that captured real savings felt political pressure to not publish those savings for fear of future budget cuts or reallocation to other agencies. Many JRI sites planned to reinvest their savings in additional reforms, and they feared losing that opportunity if dollars were reallocated to other county priorities. Despite these political sensitivities, most local stakeholders continued to work toward capturing cost savings.




Finally, sites that did achieve savings were not always able to reinvest those funds as planned to further JRI priorities. Though Mecklenburg County was able to reduce the jail allocation in the county budget by $4 million in fiscal year 2016 owing in part to JRI reforms, those savings were absorbed into the budget and were not reinvested in reform strategies (Mecklenburg County 2016b).




Despite these challenges, some sites identified cost savings related to the implementation of their JRI strategies. Sites that focused their efforts on a specific subset of their jail population were able to better estimate the savings of their policies. Denver’s Recovery Court reduced costs for some of the jail’s most frequent users by 67 percent during the first 11 months of the program, leading to over $1.6 million in avoided costs.34 Santa Cruz’s PACT program targeted people who frequently violated local ordinances, resulting in $76,937 of avoided jail bed costs for program participants.35 Preliminary outcomes from the Albemarle-Charlottesville Regional Jail indicate that incarcerations for probation revocations decreased 63 percent, leading to a cost avoidance of nearly $2.2 million.36


Find Alternative Measures of Success


Though many local JRI sites faced challenges in identifying savings as a result of their reforms, some localities have documented success in other ways. In addition to reducing justice system spending and encouraging reinvestment, JRI has encouraged systems change and the creation of new, collaborative roles within agencies, as well as ongoing data analysis, increased training and capacity, and implementation of evidence-based practices.


For example, sites created staff positions with JRI funds that are now absorbed within agency budgets: Alachua County hired a jail release coordinator, Allegheny County embedded caseworkers, and Yamhill County hired a pretrial officer. Sites also created new programs and programming space: Denver established its Recovery Court, Travis County created a housing program, Alachua County implemented an inmate transition program, and Milwaukee created its CLU as a central location for community services and treatment for program participants and for managing diversion and deferred prosecution cases.


A number of local agencies have also fundamentally changed the way they do business. Community supervision agencies have incorporated evidence-based practices, such as supervision response grids in Charlottesville-Albemarle, Grant, and Delaware Counties and risk assessment in Lane and Yolo Counties. Jurisdictions have also built capacity within agencies by hiring data analysts and focusing on collecting and analyzing data routinely and systematically, as with Grant County’s data dashboard, Johnson County’s use of Cognos business intelligence software, and Mecklenburg’s data warehouse. Further, JRI has facilitated local systems change and helped jurisdictions develop new collaborations between agencies.

Meck;enburg County's Driver's License Restoration Clinic


Keys to Successfully Implementing Reform


Regardless of the measure of success, application of the JRI process required adaptation at the local level, as site experiences did not always reflect the structure of the JRI model. Each local site formalized the JRI process by convening a working group and comprehensively analyzing the drivers of its local jail population and costs. However, many local sites struggled to use those findings to identify and implement targeted policy solutions in a timely manner. For states, that process is dictated by the legislative cycle; they are forced to identify viable policy strategies and move forward quickly to develop and introduce legislation during the current session. Local jurisdictions did not have comparable external time pressure, and many remained in the policy development phase for a year or more. In fact, most jurisdictions did not finalize their implementation strategies until BJA imposed a deadline for submitting funding requests for support. By that time, the data analysis in many sites was considered outdated, and it was unclear whether selected strategies aligned with the most pressing current drivers.


Ultimately, sites varied significantly throughout the JRI process, and some localities identified and implemented reforms more successfully than others. The mechanism for codifying policy changes inherent in legislation at the state level was also missing at the local level. As a result, some sites were not able to formalize their programs and policy changes sustainably. Despite these challenges, particular elements emerged as keys to successful implementation, including a dedicated site coordinator, a strong champion for reform, ongoing collaboration, internal data capacity, working within the state context, and patience and persistence.


Use a Cross-Agency Site Coordinator


One key element of success in local JRI was a dedicated site coordinator, ideally someone tasked with overseeing implementation and maintaining accountability while working across agencies and building coalitions. In sites with a strong coordinator, the strategies were generally closely managed and monitored, stakeholders demonstrated strong consensus for strategies, and the site coordinator could focus on JRI without neglecting additional job duties. A number of sites used site coordinators well, including Mecklenburg, Johnson, and Milwaukee Counties.


Identify a Champion for Reform


In addition to a site coordinator tasked with monitoring and implementing JRI strategies, a consistent, strong champion for reform from within one or more participating agencies was a key element of success. Several sites, including Yamhill County and Denver, had a prominent judge working with others to implement court-specific strategies. Other sites had district attorneys, court administrators, or probation chiefs as champions. Because JRI required buy-in over a significant period, sites occasionally experienced turnover in either their site coordinator or champion. The most successful sites transitioned responsibilities and authority from one individual to the next. When Denver’s Recovery Court judge retired, Denver moved to a new champion by training the incoming judge on the protocols and allowing him to build relationships with participants and other stakeholders. Johnson County transitioned smoothly to a new site coordinator when the original site coordinator, a strong proponent of JRI, retired.


Encourage Ongoing Collaboration


As mentioned on page 2, an initial step of the JRI model is to establish a collaborative, interagency working group. A number of sites with a strong history of collaboration already had such a group in place, from either EBDM or other system reform work. The ability to build from this existing group, rather than develop it from scratch, helped sites move more quickly through data analysis, strategy selection, and implementation. Although it may have required more time, many sites without preexisting working groups were still able to develop strong collaboration to make decisions. And some sites were able to leverage their TA provider as a neutral outside voice to help facilitate collaboration and consensus on divisive strategies.


Build and Maintain Data Capacity


Internal data capacity was an additional key element of reform in local JRI sites. Sites that had an integrated data system in place, such as Johnson County, were able to efficiently collect and analyze data, identify the drivers of their jail populations, and choose and implement policy strategies. They were also well positioned to evaluate the strategies implemented because data were more readily available and accessible. Some sites worked with TA providers or outside evaluators to monitor implementation of the strategies and routinely analyze data in order to identify implementation challenges and make midcourse corrections. Other sites, such as Eau Claire County, Mecklenburg County, and New York City, hired criminal justice–specific data analysts to provide support and evaluate the JRI strategies.


Work with State-Level Reforms


Because local criminal justice reform does not occur in a vacuum, sites that were able to leverage resources and build on state reforms, rather than simply react to them, were often more successful. Yamhill County developed an exemplar pretrial program and was awarded funding from Oregon’s Criminal Justice Commission, through the state’s Justice Reinvestment Grant Program, to hire additional pretrial services officers and share its knowledge with other counties in Oregon.


Be Patient and Persistent


JRI is a time-intensive process, requiring both patience and perseverance. Many sites experienced unanticipated challenges, including staff turnover, political obstacles, administrative and funding delays, data challenges, and implementation hurdles. Sites that overcame these challenges had many of the characteristics listed above, but they also were able to adapt quickly, work together, and seek assistance when needed.

JRI in Context
San Francisco's Strategy to Reduce Racial and Ethnic Disparities


Next Steps: Maximizing Local Reforms


Justice reinvestment provided an opportunity for many local jurisdictions to implement reforms and integrate data-driven, evidence-based practices in their criminal justice systems. Four sites are able to continue expanding their justice reinvestment efforts through a BJA-funded Maximizing Local Reforms grant. This grant aims to improve upon jurisdictions’ reforms and further sites’ ability to reinvest savings in high-performing public safety strategies. Denver, Milwaukee County, New York City, and Santa Cruz County were selected to continue honing and improving their justice reinvestment strategies. The Local Maximizing Reforms projects, which will take place over three years, began in October 2015. The four sites are focused on the following strategies:


Denver: Although the Recovery Court has shown it can reduce jail bed days, detox usage, and hospital emergency visits, it has been unable to provide treatment to two types of front-end users who need a higher level of care: those found incompetent to aid in their defense, who are thus released from jail without services; and those who need more substance abuse treatment than the Recovery Court program can provide. Denver will focus on improving services for these two groups by establishing alternatives for people whose competency is likely to be questioned but who are ineligible for civil commitment, increasing civil commitments for those found incompetent as appropriate, providing health insurance enrollment for those who are not currently enrolled, hiring a behavioral health care coordinator, and securing transitional beds for program use.


Milwaukee County: Milwaukee County plans to provide enhanced services for people with mental illness. While the CLU has demonstrated early success, the needs of justice-involved people with mental health concerns were not being met. Milwaukee County will use its funding to obtain technical assistance from mental-health experts, select a validated mental health screening tool to incorporate in pretrial screening, implement stakeholder and staff training, and hire one full-time CLU case manager and one part-time certified peer support specialist.


New York City: New York City plans to address three main strategies: expand supervised release as an alternative to detention; improve alternative to detention and alternative to incarceration programs through fidelity to evidence-based practices; and develop a data system to track, monitor, and evaluate justice program capacity and performance. In March 2016 the city launched its supervised release strategy as an alternative to detention for people arrested for misdemeanors or nonviolent felonies. Supervised release providers use a risk assessment tool to determine eligibility and guide supervision levels and to provide participating defendants referrals to voluntary services. The strategy is on track to divert more than 3,000 individuals a year.37 The city has also started designing and building the justice provider system, which is projected to launch by the end of 2016.


Santa Cruz County: While Santa Cruz County’s PACT program and other pretrial strategies have demonstrated early success, the county’s criminal justice system has since been affected by statewide legislative changes. California State Proposition 47 (Prop 47), for example, may significantly alter pretrial reform. Under Prop 47, some low-level crimes are classified as misdemeanors, and people charged for those crimes are issued a court date, not arrested. If they fail to appear in court, they may be arrested, eliminating their eligibility for pretrial release. Santa Cruz County plans to address this issue and enhance its pretrial capacity by hiring a pretrial officer, implementing an automated notification system, and providing more extensive, community-based outreach for a targeted group of Prop 47 defendants at high risk of failing to appear for their court dates. These efforts are designed to significantly expand the number of people recommended for pretrial release.




Over the past six years, 17 local jurisdictions across the country have worked diligently to implement JRI, and it appears these efforts have generally paid off. Sites as diverse as San Francisco, California; Johnson County, Kansas; and Mecklenburg County, North Carolina, have committed to using a data-driven approach to understand how their local justice systems are functioning, identify policy strategies that could produce a better public safety return on investment, implement reforms, and track their progress. These wide-ranging reforms have encompassed everything from improving data capacity to finding better ways to address the needs of frequent front-end users, reforming pretrial processes, and implementing evidence-based supervision practices. This work was not easy, and identifying and reinvesting savings proved particularly challenging for most sites. In this sense, local justice reinvestment has not uniformly generated savings that can be easily quantified and directly reinvested into other public functions. However, findings suggest that many local sites have changed the way they do business and improved their practices in other ways that will long outlive the Justice Reinvestment Initiative.



  1. Tammy Meredith, “Justice Reinvestment at the Local Level: Two-Year Follow-Up on the Travis County Supportive Hosing Pilot Evaluation” (internal document, Atlanta, March 2016).
  2. Peter Ozanne, “Local Justice Reinvestment Initiative Close Out Memo—Santa Cruz County” (internal document, Crime and Justice Institute, Boston, December 2015).
  3. Santa Cruz County, “Justice Reinvestment Initiative: Assessment Technical Assistance and Maximizing Local Reforms” (internal document, May 2015).
  4. FUSE, short for Frequent User Systems Engagement, is a national model for addressing frequent front-end users. Developed by the Corporation for Supportive Housing, it was adapted by Mecklenburg County into the MeckFUSE program.
  5. Mecklenburg County, “Mecklenburg County Justice Reinvestment Initiative Outcomes and Updates” (internal document, May 2016).
  6. The Laura and John Arnold Foundation developed the Public Safety Assessment as a pretrial risk assessment tool to help judges make release, supervision, and detention decisions. It was created using a database of 1.5 million cases across 300 jurisdictions, and uses factors related to a person’s criminal history and current charge to guide decision making. It is currently used in 29 jurisdictions, including three states.
  7. Ozanne, “Local Justice Reinvestment Initiative Close Out Memo—Santa Cruz County.”
  8. Michael Kane, “Local Justice Reinvestment Initiative Close Out Memo—Allegheny, Co, PA” (internal document, Crime and Justice Institute, Boston, March 2016).
  9. Barbara Pierce Parker, “Alachua County Justice Reinvestment Initiative Phase II Closeout” (internal document, Crime and Justice Institute, Boston, June 2015).
  10. Kane, “Local Justice Reinvestment Initiative Close Out Memo—Allegheny, Co, PA.”
  11. Sean Callister, correspondence with Erika Parks, April 2016.
  12. Johnson County, “County Justice Reinvestment Initiative: Quarterly Performance Measures” (internal document, September 2015).
  13. Michael Kane, “Local Justice Reinvestment Initiative Close Out Memo—New York City” (internal document, Crime and Justice Institute, Boston, December 2015).
  14. Emily Turner, correspondence with Samantha Harvell, July 2016.
  15. Ozanne, “Local Justice Reinvestment Initiative Close Out Memo—Santa Cruz County.”
  16. EBDM is a method of applying empirical knowledge and research-supported principles to justice system decisions made at the case, agency, and system levels. It seeks to equip local and state criminal justice policymakers with the information, processes, and tools that will result in measurable reductions of pretrial misconduct, post-conviction reoffending, and other forms of community harm resulting from crime.
  17. Oregon Knowledge Bank, “Yamhill County Pretrial Justice Program,” accessed July 8, 2016, http://okb.oregon.gov/portfolio-item/yamhill-pretrial-justice/.
  18. Delaware County, “Phase II LOI” (internal document, June 2014).
  19. Albemarle-Charlottesville Regional Jail, “Phase II LOI” (internal document, March 2014).
  20. County of Yolo Probation Department, “Phase II LOI” (internal document, July 2014).
  21. Grant County Circuit Court, “Phase II LOI” (internal document, April 2014).
  22. Lore Joplin, “Local Justice Reinvestment Initiative Close Out Memo—San Francisco” (internal document, Crime and Justice Institute, Boston, May 2016).
  23. Peter Ozanne, “Local Justice Reinvestment Initiative Close Out Memo—Lane County, Oregon” (internal document, Crime and Justice Institute, Boston, November 2015).
  24. Ibid.
  25. Ibid.
  26. Joplin, “Local Justice Reinvestment Initiative Close Out Memo—San Francisco.”
  27. Charlottesville-Albemarle County, “Administrative Response Matrix Preliminary Outcomes” (internal document, April 2016).
  28. Leilah Gilligan, “Completion of Justice Reinvestment at the Local Level Initiative—Charlottesville” (internal document, June 2016).
  29. Diane Linville, correspondence with Erika Parks, April 2016.
  30. Richard Stroker, correspondence with Erika Parks, May 2016.
  31. Michael Kane, “Local Justive Reinvestment Initiative Close Out Memo—Yolo County, CA” (internal document, March 2016); Peter Ozanne, “Local Justice Reinvestment Initiative Close Out Memo—Lane County, Oregon” (internal document, Crime and Justice Institute, Boston, November 2015).
  32. For more information on the Jail Population Policy Impact Tool, see http://www.crj.org/cji/pages/jail-population-tool.
  33. Eau Claire County Criminal Justice Collaborating Council, “Phase II LOI” (internal document, March 2014).
  34. Denver City and County, “Justice Reinvestment Initiative: Assessment Technical Assistance and Maximizing Local Reforms” (internal document, May 2015).
  35. Santa Cruz County, “Justice Reinvestment Initiative: Assessment Technical Assistance and Maximizing Local Reforms” (internal document, May 2015).
  36. Charlottesville-Albemarle County, “Administrative Response Matrix Preliminary Outcomes” (internal document, April 2016).
  37. Emily Turner, correspondence with Samantha Harvell, July 2016.




Andrews, Donald, James Bonta, and J. Stephen Wormith, 2006. “Recent Past and Near Future of Risk and/or Need Assessment.” Crime and Delinquency 52 (1): 7–27.

Center for Effective Public Policy. 2014. “Justice Reinvestment Initiative at the Local Level: Getting to Know Milwaukee County, Wisconsin.” Silver Spring, MD: Center for Effective Public Policy.

Cramer, Lindsey, Samantha Harvell, Dave McClure, Ariel Sankar Bergmann, and Erika Parks. 2014. “The Justice Reinvestment Initiative: Experiences from the Local Sites.” Washington, DC: Urban Institute.

Denning, Shea. 2010. “Driver’s License Revocations.” Durham: North Carolina Office of Indigent Defense Services.

Lowenkamp, Christopher T., Edward J. Latessa, and Alexander M. Holsinger. 2006. “The Risk in Action: What Have We Learned from 13,676 Offenders and 97 Correctional Programs?” Crime and Delinquency 52 (1): 77–93.

Mecklenburg County. 2016. Fiscal Year 2016 Mecklenburg County, North Carolina Adopted Budget. Charlotte: Mecklenburg County North Carolina Office of Management and Budget.

Minton, Todd, and Zhen Zeng. 2015. “Jail Inmates at Midyear 2014.” NCJ 248629. Washington, DC: US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

Reaves, Brian A. 2013. “Felony Defendants in Large Urban Counties, 2009 – Statistical Tables.” NCJ 243777. Washington, DC: US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

Revicki, Jesse, Lisa Brooks, and Kristin Bechtel. 2015. “Local Justice Reinvestment: Targeting Reforms at the Front End of the Criminal Justice System.” Boston: Crime and Justice Institute.




This brief was changed in October 2016. Figure 1, a map of local JRI sites, was updated to improve the accuracy of the locations of San Francisco City & County, Santa Cruz County, and Yolo County.


About the Authors


Erika Parks is a research associate in the Urban Institute’s Justice Policy Center, where she is the task lead for the local Justice Reinvestment Initiative work. She also manages the evaluation of the Second Chance Act juvenile reentry demonstration projects and the development of the forthcoming Risk Assessment Clearinghouse.


Samantha Harvell is a senior research associate with the Justice Policy Center. She co-directs the OJJDP-funded Bridging Research and Practice to Advance Juvenile Justice and Safety project, and she oversees assessment of local and state sites involved in the Justice Reinvestment Initiative.


Lindsey Cramer is a research associate with the Justice Policy Center, where she leads the coordination and assessment of the Justice Reinvestment Initiative, working closely with technical assistance providers, national partners, and local and state jurisdictions to reduce the costs of corrections services and reinvest the savings in initiatives to enhance public safety.


Abigail Flynn is a research assistant with the Justice Policy Center. Her areas of research include state and local justice reform efforts, wrongful convictions, and perceptions of the justice system. She also works on research related to the federal justice system and provided support to the Charles Colson Task Force on Federal Corrections.


Hanna Love is a research assistant in the Justice Policy Center. Her areas of research include state and local justice reform efforts, perceptions of the justice system, and bridging research and practice in the justice system. She also works on the What Works in Reentry Clearinghouse and a number of research projects related to human trafficking.


Caroline Ross is a research associate in the Justice Policy Center focusing on justice reinvestment, systems and culture change in public agencies, and promoting collaborative responses to public safety challenges. She produces practitioner-oriented research for safer and healthier communities.




The views expressed are those of the authors and should not be attributed to the US Department of Justice or to the Urban Institute, its trustees, or its funders. Funders do not determine research findings or the insights and recommendations of Urban experts. Further information on the Urban Institute’s funding principles is available at www.urban.org/support.

Copyright August 2016. Urban Institute.

Research Area: 


To reuse content from Urban Institute, visit copyright.com, search for the publications, choose from a list of licenses, and complete the transaction.
LATEST IN Crime and Justice
To reuse content from Urban Institute, visit copyright.com, search for the publications, choose from a list of licenses, and complete the transaction.