Scaling Restorative Community Conferencing Through a Pay for Success Model

Scaling Restorative Community Conferencing Through a Pay for Success Model: A Feasibility Assessment Report www.nccdglobal.org MARCH 2015  Table of Contents Highlights . . . . . . . . . . . . . . Overview of PFS Contracting . . Overview of Restorative Justice PFS Feasibility Criteria . . . . . . Preliminary Project Blueprint . . Considerations and Next Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 .......................................................4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 Acknowledgments This report was produced through the generous support of The California Endowment. The Akonadi Foundation also provided support to NCCD for data collection and analysis. NCCD promotes just and equitable social systems for individuals, families, and communities through research, public policy, and practice. 2 Highlights In December 2013, The California Endowment funded the National Council on Crime and Delinquency (NCCD) and Third Sector Capital Partners, Inc. (Third Sector) to conduct a feasibility study on restorative community conferencing (RCC) to better understand its potential for a Pay for Success (PFS) project. RCC shows promise as an intervention that could be scaled through PFS given its preventive focus and the promising initial outcomes of youth served through RCC. NCCD and Third Sector focused their feasibility analysis on Alameda County, California, where an RCC intervention currently serves 75 youth annually. An analysis of available data gathered since 2012 has revealed that of the young people who completed Alameda County’s RCC program, 26.5% were rearrested compared with 45.0% of a matched sample of youth whose cases were processed through the juvenile justice system. Notably, only 11.8% of the RCC youth were subsequently adjudicated delinquent— that is, determined by the court to have committed another delinquent act—compared with 31.4% of the matched sample of youth whose cases were processed through the juvenile justice system. Of participating crime victims, 99% stated they would participate in another RCC. This program also carries significant cost- saving potential, as these lower rates of reoffending combine with a one-time cost of $4,500 per RCC versus $23,000 per year for a youth on probation. With such promising data, NCCD and Third Sector sought to better understand how RCC could be scaled through a PFS project and what capacity building would need to take place for such a project to be feasible. The results of this analysis are detailed in the following feasibility report. 1 Overview of PFS Contracting PFS and Stakeholders PFS and social impact financing have emerged as mechanisms to measurably improve outcomes for those in need by changing the way that government allocates and invests its resources. PFS contracting is a new form of performance-based contracting through which government reimburses effective social service providers for achieving agreed- upon outcomes. This type of performance-based contracting stands in contrast to traditional social service procurement, which generally reimburses providers for costs incurred when delivering services and mandates what services must be delivered as a precondition for reimbursement. Social impact financing (SIF) is a funding tool to bridge the payment timing delays inherent to PFS contracting, where measurement of the agreed-upon outcomes can take years to accomplish. SIF also shifts the risk of potential failure away from service providers and, through private funder participation, introduces rigorous due diligence and a longer time frame of attention to the social services arena. In a PFS project, funders provide initial working capital for services and the government agrees to make payments to the project only when outcomes are achieved. By focusing on results and outcomes and funding only what measurably works, this type of performance-based contracting reduces risk for government and institutes a rigorous feedback loop for social services. A PFS project includes five key partners. 1. Government • Determines which outcomes in a given area it wants to pay for • Agrees to make success payments upon achievement of those outcomes • Works with all project parties to develop an implementation plan for services 2. Service provider(s) • Implements effective services to achieve agreed-upon outcomes • Calculates the budget necessary for scaling and delivering these services • Builds systems and capacity to enable rigorous evaluation of outcomes 3. Independent evaluator • Designs rigorous evaluation to determine impact • Assesses providers’ intervention outcomes 4. Funders • Conduct due diligence on project • Provide the upfront working capital necessary to implement the project through traditional grants, mission-related/project-related investments, credit enhancements, loan guarantees, or commercial loans 5. Intermediaries • Provide project functions, moving a PFS project from procurement through launch, including: facilitating multi-party negotiations, data analysis, financial modeling, deal construction, and fundraising • Can also provide daily oversight and management of project implementation and serve as fiscal agent for the project 2 Benefits of PFS Contracting PFS is designed to shift government resources toward more effective and high-impact programs. As a tool, a PFS contract is meant to empower service providers to deliver results, allow government to understand what results it pays for, and have funders take on performance and implementation risks that government may not be able or willing to take with taxpayer funds. PFS also may help fund programs that avoid the need to spend taxpayer money in the first place by, in just one of many examples, reducing crime and thus the number of days that high-risk youth spend in prison. Figure 1: PFS Project Flow of Funds and Services Funders: Philanthropy, Commercial Loans 1 Services Payments Success Payments Government Payer Provide upfront working capital Repays funders if outcomes met 65 Makes success payments if outcomes met Based on evaluation results, determines payment amount Service Provider(s) Delivers services Provides payment for services 2 Intermediary Population to Be Served Designs evaluation and measures outcomes 4 Evaluator 3 Table 1: Benefits of PFS Across Stakeholders Stakeholder Benefits of PFS Government • Only pays for desired outcomes; reduced performance risk • Rigorous impact evaluation provides valuable information for future resource allocation decisions Service Providers • Multiyear, flexible funding to deliver services • Rigorous impact evaluation of services Funders • Impact investing opportunity to align financial and social returns • Recyclable grant funding • Innovative form of funding in issue areas of interest Taxpayers • Improvement in social outcomes • Potential offset of costs through government savings High-Risk Population • Measurable improvement of outcomes • More effective services 3 Overview of Restorative Justice Restorative Justice Concept Despite falling crime rates and evidence that racial minorities do not commit more crime than Whites, the United States has experienced a rising incarceration population and racial and ethnic disparities over the last several decades.1 Crime victims also express disappointment with justice system outcomes; even when convictions are secured, crime victims’ needs often remain unmet.2 As a result, restorative justice has grown in popularity as a viable alternative, capable of reducing recidivism and incarceration, decreasing spending on public safety, increasing community involvement, and improving victim satisfaction. Our current criminal justice system operates by asking three guiding questions: 1) What law was broken? 2) Who broke it? and 3) What punishment is warranted? Restorative justice invites a fundamental shift in the way we think about and address crime. This alternative model asks: 1) Who was harmed? 2) What are the needs of those affected? and 3) Whose obligation is it to meet those needs? Thus, restorative justice differs from the adversarial legal process because the latter focuses on the actions of the person who caused harm, while the former prioritizes the people and relationships harmed. 1 US Department of Justice. (2013). Smart on crime: Reforming the criminal justice system for the 21st century. Attorney General Eric Holder’s remarks to American Bar Association’s Annual Convention in San Francisco, CA. Retrieved from http://www.justice.gov/sites/default/files/ag/legacy/2013/08/12/ smart-on-crime.pdf 2 See, generally, Herman, S. (2010). Parallel justice for victims of crime. Washington, DC: National Center for Victims of Crime. 3 Bolitho, J., Bruce, J., & Mason, G. (2012). Restorative justice: Adults and emerging practice. Annadale, New South Wales: Federation Press. 4 Ibidem. 5 Ibid. From New Zealand to Oakland, CA In 1988, New Zealand’s government commissioned a report identifying government practices that resulted in institutionalized racism.3 This was evidenced by the “higher risks Māori children and young offenders faced compared to non-Māori. The report confirmed that Māori [were] over-represented in negative statistics relating to health, education, housing and unemployment, and concluded that this [was] the result of a ‘monocultural bias’ which favor[ed] non-Māori culture.”4 In response, New Zealand passed the Children, Young Persons, and Their Families Act of 1989, transforming the nation’s juvenile justice system, which now relies entirely on Family Group Conferencing (FGC) to address youthful offending. FGC is a form of restorative justice whereby a young person who has offended, their family, victims, and others (e.g., the police, a social worker, youth advocate, etc.) talk about how to help the young person own up to what he did wrong...learn from his mistakes.5 During the FGC, participants agree on a plan through which the youth can make up for her offense...turn her life around. The plan becomes legally binding, and the Department of Child, Youth and Family Services monitors the young person to ensure it is completed. New Zealand has found that FGCs reduce recidivism, increase victim satisfaction, and promote a sense of responsibility in offenders. Following in New Zealand’s footsteps, Oakland, California, now has a restorative justice program modeled after the FGC approach, called restorative community conferencing. When an offense occurs, legal proceedings can often be intensive and time-consuming for the responsible party, the person harmed, and interested family and community members. By contrast, restorative practices encourage constructive responses to wrongdoing by bringing those who have harmed, their victims, and affected communities into processes that repair the harm and rebuild relationships. At its best, through face-to-face dialogue, this approach results in consensus-based plans that meet victim- identified needs in the wake of a crime. In applications with young people, restorative justice can prevent both contact with the juvenile justice system, and school expulsions and suspensions. Several restorative justice models have been shown to reduce recidivism and, when embraced as a larger- scale solution to wrongdoing, can minimize the social and fiscal costs of crime. Introduced in US cities such as Louisville, Kentucky, and Baltimore, Maryland, and in larger international contexts, restorative programs have proven immensely effective. For instance, this approach has even rendered youth incarceration nearly obsolete in New Zealand, as detailed above. Restorative Community Conferencing The principles of restorative justice have led to the creation of a number of programs designed to address and resolve conflicts in different contexts. Examples of such programs include victim-offender dialogues, circles of support and accountability, and peacemaking circles. With respect to crime, an assortment of restorative models have been introduced at every stage of the legal process, from pre-arrest to post-adjudication. While restorative justice takes a number of forms, perhaps the most Figure 2: Juvenile Justice Process Crime Apprehension Police dismiss Arrest Probation RCC* District Attorney (DA) DA dismisses DA charges Court Court dismisses Youth adjudicated delinquent and sentenced *If the youth successfully completes the RCC process, charges are never filed. If, however, the youth fails to complete, his/her case is returned to the referring agency. 6 Sherman, L., & Strang, H. (2007). Restorative justice: The evidence. Retrieved from http://www.iirp.edu/pdf/RJ_full_report.pdf prominent is the RCC approach, which, according to a 2007 meta-analysis of data from several countries, is effective at reducing recidivism, among other significant benefits.6 Modeled after the New Zealand FGC model, RCCs involve an organized, facilitated dialogue in which young people, with the support of family, community, and law enforcement, meet with their crime victims to create a plan to repair the harm done. It is most effective with serious crimes in which there is an identifiable victim, such as in the case of robbery, burglary, car theft, assault/battery, arson, and teen relationship violence. 5 There is power in the simplicity of the RCC process. When police or school authorities apprehend a young person for committing a crime, rather than sending the case through traditional juvenile justice processes, the referring agency (school, police, probation, or district attorney) contacts a nonprofit organization trained in the RCC approach. The organization reviews the file and, if they accept the case, the referring agency places the case in a holding pattern, neither dropping nor charging it. Next, the organization’s RCC coordinator sends out letters and program brochures to the accused youth and his/her parents. The letters are followed by a phone call and a home visit to answer questions and encourage participation in the program. If the young person accepts responsibility and agrees to participate, letters and brochures are sent to the victim, again followed by phone calls and visits. No fewer than two meetings are held with both parties to determine amenability and safety and to allow youth and their victims to independently assess the harms, needs, and obligations resulting from the crime. By agreement with the district attorney, all communications in RCC, in preparation for the RCC, and in the completion stage are confidential and cannot be used against the youth. This encourages complete honesty about the crime and its causes and effects. It also encourages the participation of some victims who would like to hold youth accountable but are unwilling to engage directly with legal systems. Figure 3: Breakdown of a Successful RCC Within a few weeks following preliminary meetings, the RCC takes place at a neutral location, such as the nonprofit organization’s office. Through the conference, the young person, his/her victim, supporters of both, and community members, as well as school personnel, law enforcement, or the district attorney, come together to discuss the crime and its causes and effects. Occasionally, if the victim declines to participate directly, he or she may choose a surrogate victim. In each RCC, all parties engage in self-reflection, firm yet supportive accountability, and apologies, all culminating in a commitment to help a young person overcome obstacles and mend social ties. During the RCC, participants produce a consensus-based plan for the young person to repair the harm done. Such a plan typically encompasses four objectives: to “do right” by one’s victim, family, community, and self. If the RCC participants are unable to come to agreement on the plan or the youth fails to complete the plan, the case is sent back to the referring agency (e.g., police, probation, or the court). The RCC coordinator monitors the plan during the completion stage. An agreements/case manager may also verify and assist in completion of plans and coordinate services needed beyond the scope of the RCC. The program director receives the cases from the district attorney or other referring agency and provides monthly status reports to them. The plan is generally completed within three to six months, at which point the case is closed without charges ever being filed. A facilitator holds separate preliminary meetings with all participants to prep for the RCC The DA or probation officer refers the youth’s case to the restorative justice nonprofit The young person commits a crime The youth completes the plan and no charges are filed The youth completes the plan to repair the harm over the course of six months The RCC takes place and the participants agree on a plan for the youth to complete 6 RCC in Alameda County RCC has been operating in Alameda County for more than six years through positive relationships with community and criminal justice stakeholders. RCC’s application in Alameda County is the first of its kind and scope to address youth crimes in a major US urban area with an explicit goal of reducing racial disparities in diversion and incarceration while producing concrete data. These efforts began when sujatha baliga,7 a nationally recognized restorative justice expert and RCC legal consultant, secured a 2008 Soros Fellowship to garner institutional support from all necessary partners (the Alameda County Juvenile Court; Oakland’s chief of police and several other police departments; the county’s public defender, district attorney, and probation departments; victim- and youth-serving organizations; and other community-based organizations). The initial operationalization of the program proved promising: harms were repaired, youth made amends, and persons harmed felt heard and vindicated. The Oakland-based organization Community Works West (Community Works) ultimately took on the task of running the RCC program, providing conference facilitators and handling case referrals. Community Works received funding as of January 1, 2012, to start a small pilot program and replace the prosecution of Alameda County youth with RCC. Community Works has demonstrated its ability to successfully and reliably implement RCCs as a pre-adjudication diversion program for youth. Through years of trust building, Community Works now receives RCC cases from across Alameda County, including from schools, the probation department, multiple police departments, and the managing district attorney who heads Alameda County’s Juvenile Division. A growing number of stakeholders within Alameda County’s criminal justice system have had an 7 Lowercase intentional. opportunity to learn about and gain skills in restorative practices. In 2014, NCCD and Community Works, in conjunction with other community-based organizations, collaborated to facilitate a two-day restorative justice training for the Oakland Police Department (OPD). As a result, OPD and other police departments are prepared to send more cases to the RCC program once Community Works has the capacity to expand its caseload. While RCC is poised to grow in Alameda County, it currently operates through individual relationships and bilateral agreements with referring agencies. PFS would involve government at a high level in RCC’s success, convene all stakeholders to coordinate efforts to scale services, rigorously evaluate RCC’s impact in the community, and raise financing to serve more at- risk youth. Additionally, RCC is particularly promising for PFS from the goal of scaling effective social interventions that produce significant, quantifiable results. In addition to measurable benefits, including reducing truancy and rates of reoffending, RCC proponents believe this model carries a number of social advantages such as increasing victim satisfaction and offender accountability, strengthening community relationships, and disrupting the cyclical nature of overincarceration and related racial disparities. Figure 4: Youth Served Through RCC, 2012–2014 300 250 200 150 100 50 0 2012 2013 Referred 2014 Eligible for RCC Enrolled 7 PFS Feasibility Criteria PFS Feasibility Criteria High-Risk Target Population Proven Intervention Benefit to Government Ability to Scale Services Available Data and Evaluation Methods Safeguards for the Target Population During PFS feasibility, parties research the considerations for undertaking a PFS project according to the following criteria. After undertaking a feasibility analysis, all parties then collectively decide whether or not to move forward with developing and launching a PFS project. Below, we evaluate Alameda County’s current RCC process to determine the feasibility of scaling through a PFS project, whether in Alameda County or in other jurisdictions. We use the key criteria and considerations for a PFS project and detail our assessment of RCC on each criterion and also detail the capacity-building and information gaps that need to be remediated in order to move forward with PFS contracting. Key Criterion: Target Population RCC Eligibility Criteria Community Works operates RCC as a post-arrest, pre- charge model. This approach allows for the individual accused of a crime and the respective victims and community to reap the benefits of the restorative process without having to suffer from the debilitating and direct collateral consequences associated with judicial system involvement. Moreover, a pre-charge restorative program allows the county to keep costs as low as possible by avoiding the use of court time and resources. Eligibility criteria for enrollment in a restorative justice process vary among programs. Some counties target crimes that young people of color are most often incarcerated for, such as robbery, larceny, and assault, whereas others make eligible any crime punishable by a period of confinement. Ultimately, the eligibility criteria for a particular program will depend on the agreement between the applicable jurisdiction’s district attorney and probation offices and the organization carrying out the restorative process. In all cases, programs should minimize discretion in the decision to refer a case to a RCC in order to reduce the risk of racial and ethnic disparities and to streamline the referral process for a PFS project. One way to minimize or eliminate discretion is for the referring agency to determine which offenses are eligible for the restorative diversion program and to refer all individuals who commit those offenses to the restorative process. This would need to be balanced with a rigorous evaluation design in a PFS project, which may require only partial referrals of the full eligible population in order to measure impact. In Alameda County, the district attorney has complete discretion to determine which cases to refer to Community Works’ RCC process. Community Works then focuses on accepting cases involving serious crimes in which there is an identifiable victim (e.g., robbery, burglary, car theft, assault/battery, arson, and teen dating violence) and the responsible youth would otherwise be exposed to significant contact with the juvenile justice system. PFS projects best serve a clearly identified, referrable, and high-risk target population. 8 Current Outcomes for Target Population Using data from the Alameda County Probation Department’s July 2013 report and the results of a recidivism analysis by NCCD on the youth population on probation in 2010, we can provide some background on juvenile crime and outcomes in Alameda County. In July 2013, 2,147 youth were on probation in Alameda County,8 348 females (16.2%) and 1,799 males (83.8%). The racial composition of the probationers was African American, 1,238 (57.7%); Latinos, 579 (27.0%); White, 182 (8.5%); Asian, 97 (4.5%); and and youth of other ethnicities, 51 (2.4%). The average time on probation was 12 months for females and 18 months for males. The average time African-American youth spend on probation was 20 months, with Latino youth spending 14 months; White youth, 10 months; Asian youth, 12 months; and youth of other ethnicities, 6 months.9 The majority of youth were placed on probation for property (28.0%) or person offenses (26.0%) or failing to obey a court order (26.0%). NCCD conducted a recidivism analysis of youth in Alameda County with a cohort of 758 youth who had received formal or informal probation as a disposition in 2010. The youth were tracked for two years for new offenses, petitions filed by the prosecutor, and petitions sustained by the court. The cohort had a rearrest rate of 46%, with 72% of petitions filed and 73.6% of petitions sustained by the court.10 Key Criterion: Proven Intervention In order to merit consideration for social impact financing, an intervention must demonstrate the ability to achieve positive outcomes. 8 Alameda County Probation Department. (2013). A look into probation: Monthly report—July 2013. Retrieved from http://www.acgov.org/probation/ documents/July2013Report.pdf 9 Because average length of stay is calculated by the probation department by those cases that closed during the month, it is expected that these numbers are actually higher. 10 The method used to determine rearrests only counts individuals and not number of times arrested. The total number of rearrests for the 2010 cohort may be greater than this due to multiple arrests for one individual during each year. “Petitions filed” is analogous to charges being filed in adult court; “petitions sustained” is analogous to being found guilty in adult court, since judges in the juvenile justice system determine guilt rather than a jury. 11 Strang, H., Sherman, L., W., Mayo-Wilson, E., Woods, D., & Ariel, B. (2013). Restorative justice conferencing (RJC) using face-to-face meetings of offenders and victims: Effects on offender recidivism and victim satisfaction. A systematic review. Campbell Systematic Review, 9(12), 1–59. See also Sherman, L., & Strang, H. (2007). Restorative justice: The evidence. Retrieved from http://www.iirp.edu/pdf/RJ_full_report.pdf 12 Ibid. Evidence Base for Restorative Justice The available research on restorative justice both in the United States and internationally has found that this approach is effective at reducing recidivism rates while improving victim satisfaction compared with traditional, adversarial court processes. For instance, a systematic review of programs in the United States, Australia, and the United Kingdom found restorative models decrease reoffending, especially for violent crimes.11 The researchers found that restorative processes also benefit victims in a number of ways, including reducing post-traumatic stress symptoms, increasing satisfaction with the resolution of their case, and lessening the desire for violent revenge. Finally, the review determined that restorative justice was more economical than conventional justice systems because it not only prevents crime but also costs less to administer.12 9 Evaluation Methodology Timeline San Francisco RCC Random Assignment of Youth December 2015 Long Beach Matched Pairs September 2016 San Diego Matched Pairs December 2016 Evaluations in Progress In addition to the program based in Oakland, three cities in California— Long Beach, San Diego, and San Francisco—have begun RCC programs with the support of NCCD and The California Endowment. San Francisco began its random assignment pilot program in October 2013. Long Beach also began its program in October 2013, and San Diego’s start date was May 2014. San Francisco will limit its pilot to 25 treatment youth receiving RCC each year, whereas Long Beach and San Diego anticipate 40 treatment cases per year for three years. The latter two cities will not be using random assignment and will implement a matched comparison group design. San Francisco’s block randomization design will result in one treatment group (for those who were assigned to a treatment group and agreed to participate), a comparison group made up of youth randomly assigned to the group, and a second comparison group made up of youth assigned to the treatment group but who rejected the offer of RCC and were processed through the court system. In Long Beach, cases will be assigned to RCC by the police department and the courts. San Diego will be assigned cases by the police department, probation department, and the district attorney’s office. Long Beach and San Diego will use a matched pairs design in which treatment youth will be compared to youth who were processed through the court process. They will be matched by age, race, gender, offense, zip code, and number of prior offenses. Preliminary Results From Impact Measurement In order to more accurately measure RCC’s impact, NCCD is currently matching each of the RCC youth with Alameda County youth who were adjudicated through the traditional court process. Each pair is matched by gender, race, birth date, offense, zip code, and number of prior offenses. Data on new offenses, probation violations, petitions filed and sustained, and dispositions for all youth in the two cohorts will be compared as part of the analyses. This enables NCCD to compare youth who enrolled in RCC with very similar youth who were processed through the regular court process to understand their different trajectories. The primary outcome in the comparison of the RCC youth with those who were processed through the court system is whether the RCC youth committed fewer, more, or the same number of new offenses as those who were court adjudicated. If a difference is generated, a second consideration is whether the difference between the RCC youth and the court- adjudicated youth is statistically significant.13 The Community Works RCC program has made improvements in Alameda County similar to other restorative programs demonstrating evidence-based success. Between January 2012 and December 2014, 172 youth completed or were in the Community Works RCC program. Of those youth, 102 have completed the program, and 44 of those were enrolled as of December 31, 2014.14 Of youth who 13 A result or finding is considered “statistically significant” if it is likely attributable to a specific cause as opposed to a random occurrence. Statistical significance, evidenced by the p-value, can be strong or weak. The p-value represents the probability of error involved in accepting the observed result as valid, that is, as “representative of the population.” As p-values typically range from .01 to .05, the higher the p-value, the greater the likelihood of error and, thus, the less reliable the observed relationship is between the selected variables. 14 These youth enrolled in 2012, 2013, or 2014. 10 completed the program, 26.5% were rearrested. The period in which new offenses occurred ranged from a few weeks to more than a year after completing the program. Such low rearrest rates stand in stark contrast with the county’s youth rearrest rate of 45.0% for a matched comparison group of youth cycling through the juvenile justice system (p = 0.05). Even more impressive is that petitions for subsequent crimes were filed for RCC youth in only 12.7% of cases and sustained for 11.8%. For the matched sample of youth whose cases were processed through the juvenile justice system, 33.5% of cases resulted in petitions filed and 31.4% of petitions were sustained. The difference is significant (p = 0.05). Key Criterion: Benefit to Government Government, in order to commit to success payments, must be able to associate positive outcomes in a project with an internal priority, whether prioritizing a project’s social outcomes or identifying budgetary savings through diversion or reduction of future expenditures. FIgure 5 Youth Outcomes: RCC Versus Juvenile Justice System 45.0% 33.5% 31.4% 26.5% 12.7% 11.8% Rearrest Rate Petitions Filed for Subsequent Petitions Sustained for Crimes Subsequent Crimes RCC Juvenile Justice System As of December 2014, 50% of Alameda County’s RCC participants have been African American and 30% have been Latino. This is consistent with the program’s explicit goal of reducing racial and ethnic disparities in the county. Moreover, with Community Works prioritizing serious offenses, 61% of the conferences have involved felony charges. With its unique focus on addressing victims’ needs, all RCC participants have reported in post-conference satisfaction surveys that they would recommend the process to a friend and 99% said they would participate in another conference, while 84% of families surveyed indicated the highest level of satisfaction with the process. PFS projects are driven forward on the government side by two different primary motivations. The first, budgetary savings, motivates governments that want a largely budget-neutral initiative and are focused on paying for prevention out of the reduced utilization of downstream services. This motivation drives a PFS project’s focus on a high-risk, high-cost target population and bases outcome metrics and payment on the resulting cost impact to the government payer. The second, a focus on maximizing service efficacy, seeks to make existing or new budget expenditures more performance- and outcomes-based. This motivation to find more innovative and high-impact services may result in less explicit cost savings. It will instead focus on addressing a government priority and achieving a positive social outcome regardless of contingent cost savings. 11 Government Benefits and Approach to PFS In reality, a “pure” motivation for a PFS project rarely exists, and both motivations often exist during the development of a PFS project. The important objective during feasibility work with governments is to identify which motivation is the primary driver for constructing and launching a PFS project and to understand how that motivation develops the selection of outcome metrics and target populations. For RCC, a savings-motivated project might focus on the most high-risk juvenile populations that enter the judicial and probation systems through the use of a valid and equitable risk assessment. Outcomes would also be focused on budgetary cost avoidance by reducing the heavy utilization of government services such as probation or detention. The payments related to those outcomes would then be calculated on a government’s current expenditures for a youth using the relevant service. Savings-motivated PFS projects also estimate the level of potential savings through a project; depending on a government’s funding processes, it is likely that some of these savings may not be realized even if a PFS project achieves positive outcomes that could reduce cost expenditures. If a PFS project drastically reduced recidivism, for example, it would be possible to close a jail. In reality, closing a jail may be infeasible. With juvenile justice RCC projects, while reducing probation costs could be possible due to the theoretical reduction in caseload, current caseloads may be already over capacity. A PFS project in that context would simply relieve some of the pressure on the system by diverting eligible youth to RCC. Alternatively, an efficacy-motivated project could serve a wider population as net savings would no longer be the highest priority in the project. By serving all youth entering the judicial and probation systems, it could also consider outcome metrics that have weaker direct ties to government budgets, such as rearrests (as opposed to tracking lengths of stays in judicial systems) or incidences of truancy. While payments would be entirely based on the government’s willingness to pay for improvement of those outcomes, they could also be informed by the government’s current expenditures on programming to impact those outcomes. A government would maintain levels of funding for programs and, through the PFS project evaluation, would better understand what that funding achieved. Current Cost and Savings Estimates In 2012, Alameda County had 2,874 youth arrests, of which approximately 854 could have been eligible for diversion to RCC. The eligibility criteria for RCC in Alameda County are detailed below. Table 2: Alameda County 2012 Juvenile Arrest Data Motivation Target Population High-risk populations that heavily utilize government services High-risk populations that are not well-served and are a policy priority Outcome Metrics Budgetary Savings Metrics tired to reduced use of government services Service Efficacy Metrics tied to positive societal outcomes PFS Payments Calculated on net savings Calculated on current service costs Felony Misdemeanor Total Arrests 1,308 1,566 Appropriate Category of Crime for RCC 913 497 Percentage That Could Be Diverted to RCC (Assumption) 50.0% 80.0% Available RCC Cases 456 398 Total Eligible for Referral to RCC 854 12 As shown in the diagram below, rearrested youth incur two types of avoidable costs: the costs they incur through their first offense and the costs they incur through later recidivism. The average young person arrested costs the county $23,000 in related probation costs per year.15 This estimate does not include other costs incurred through a juvenile’s involvement in the justice system, including costs incurred through the public defender’s office, district attorney’s office, court costs, and police costs post-arrest. In contrast, Alameda County’s restorative justice program carries a lower marginal cost of approximately $4,500 per case. From data collected by NCCD on youth who have participated in and completed RCC, we estimate that 26.5% of Community Work’s RCC participants will be rearrested within 18 months. In Table 3, we calculate the net savings per juvenile served through RCC rather than the current juvenile justice system. On average, a juvenile case addressed through the existing system costs the county $51,830, while serving a youth accused of crime through RCC costs $13,908, a savings of $37,922 per youth. Table 3: Potential Savings From RCC per Youth Served16 Current Costs RCC Costs Annual Cost to Probation $23,000 $4,500 Assumed Time in Probation (Years)17 1.5 0 Rearrest Rate 46.0% 26.5% Total Cost: First Offense $35,500 $4,500 Total Cost: Recidivism $16,330 $9,408 Total Cost $51,830 $13,908 Average Restorative Justice Net Savings – $37,922 Youth Commits Crime Police Department Probation District Attorney’s Office Costs Avoided Through Reduction in Recidivism Estimated Probation Costs Avoided Through Diversion Into Restorative Community Conferencing Probation Home or Camp Placement Out-of- Detention California Youth Authority 15 Estimate based on 2010 probation costs, including salaries, equipment, contract costs, and detention costs. 16 Assuming one year in probation for all arrested juveniles except when diverted to RCC or rearrested after participating in RCC. Does not include costs of juveniles referred to RCC who are already on probation. 17 Assuming 18 months on average in probation for all arrested juveniles except when diverted to RCC (but those rearrested after RCC also incur 18 months in probation). Does not include costs of juveniles referred to RCC who are already on probation. Source: Alameda County Probation Department. (2013). A look into probation: Monthly report—July 2013. Retrieved from http://www.acgov.org/probation/documents/July2013Report.pdf 18 Zehr, H. & Macrae, A. (2004). The little book of Family Group Conferences: New Zealand style. Intercourse, PA: Good Books. 19 Umbreit, M. (2000). Family Group Conferencing: Implications for crime victims. St. Paul, MN: University of Minnesota, Center for Restorative Justice & Peacemaking. In addition to the potential cost savings through both diversion from the current juvenile justice system and reduced recidivism, RCC offers other societal benefits. For instance, this model serves to ease the pressure placed on courts, correctional facilities, and probation departments, all of which are overburdened by the number of individuals cycling through the criminal justice system each year. By reducing this drain on resources through the use of an RCC diversion program, criminal justice agencies can focus on providing services to those who need them most. Another societal advantage associated with RCCs is the fact that victims are often more satisfied when their case is resolved through this approach as opposed to the traditional legal process.18 This can largely be attributed to the fact that RCCs are victim- oriented, whereby the person harmed—as opposed to a judge—is instrumental in holding the young person accountable. Furthermore, RCC participation helps to alleviate symptoms of post-traumatic stress associated with victimization.19 RCCs also help to strengthen family and community ties, have the potential to reduce racial and ethnic disparities within the criminal justice system, and can reduce truancy and improve graduation rates. Finally, because the RCC process focuses on healing harms and repairing broken relationships as opposed to punishment, many have found this approach more morally sound than adjudication and incarceration. Key Criterion: Ability to Scale Additionally, if Community Works were to scale services for this project, the current physical infrastructure that houses the RCC program is insufficient. Currently, program staff use donated space to facilitate the conferences, often in churches, community centers, and schools. If the program were to grow, at least one new dedicated physical space would be required in Oakland. In addition, an additional smaller office would be needed in south county as the program grows. Through conversations with the managing attorney for the juvenile division of the district attorney’s office, we have learned a significant number of cases could be sent from south county, and facilitators would need a reliable space to bring the parties together there. In addition to staff and infrastructure, Community Works would need to purchase and upgrade appropriate technology for staff and for data collection and analysis. Based on their past program-scaling successes, these are tasks that Community Works appears capable of performing. For example, Community Works began the Resolve to Stop the Violence program as a weekly support group for men accused of violent crimes and grew the program into a $1.3 million project that has since served 10,000 men in the 17 years it has been running in the San Francisco County Jail. The program began through a grant from the Soros Foundation, is now 90% funded through San Francisco City and County, and has been replicated across the nation. Other Community Works programs have experienced similar successes and growth. Key Criterion: Available Data If a PFS project proposes to serve a larger number of the target population, the project must be able to scale services to meet the identified need while maintaining program fidelity and a high level of impact. Infrastructure Needed for Scaling Currently, RCC serves 75 youth annually in Alameda County. RCC staff and NCCD determined that scaling while maintaining fidelity to the RCC model could be possible up to 300 cases annually through a PFS project. For RCC to ultimately receive over 800 referrals and resolve 300 cases per year, significant infrastructural enhancements are required. First and most obvious is staff. In order to scale from currently handling 75 cases a year to 200 to 300 annually, Community Works or any other organization would need up to nine facilitators; a director; an assistant director; and appropriate support staff, including translators for youth and families who do not speak English. All staff would require a ramp-up period for hiring, training, and starting to handle cases. Leadership staff would be needed to focus on relationship building and education to encourage currently nonparticipating or underparticipating police/probation/district attorney’s offices to send more cases, as well as to maintain caseloads and meet the needs of the independent evaluator. In addition to data collected by service providers, government or other external databases that track outcomes and the target population must be dynamic and accessible to the evaluator of a PFS project. Data must be able to support the project’s needs for a rigorous, multiyear evaluation. 14 Two primary data sources are used in the analysis of RCC program outcomes. Qualitative data come from surveys of participants in the conference process, and quantitative data come from the Alameda County Probation Department’s data management system. The surveys are administered immediately after the conference and are completed by the victim, responsible youth, parents, supporters, and community members who participated in the conference. The responses are entered into an Microsoft Excel database and analyzed by NCCD. NCCD requests an extract of data from the probation department on each youth who is referred to the RCC program. In addition, an extract is requested for youth who are processed through the court system and not referred to RCC. Each youth is matched to an RCC participant based on age, gender, race, and prior history. NCCD’s request covers data at each stage of processing, including arrest, outcomes from probation review, prosecution by the district attorney’s office, court process results, disposition, prior history, needs assessment scores, and demographic characteristics. Outcomes to Be Measured As stated above, PFS contracts are centered on outcomes. These outcomes and the methods for measuring them are agreed upon by all stakeholders and should be a representation of success for the project. There are a number of ways in which success may be measured for this project, but outcomes need to be carefully considered. While final outcomes to be measured in a PFS project will be determined with all stakeholders, the following outcomes could accurately demonstrate project success. • Probation time. Reduction in time spent on probation for RCC participants could demonstrate how participants avoid judicial system involvement and save Alameda County money and other resources. However, youth who enter the RCC program may already be • on probation as a result of a previous offenses and therefore would not avoid probation time. Currently, only 11% of youth who enroll in RCC are on probation for a prior offense. Rearrest and readjudication. A percentage of youth will reoffend regardless of whether they are diverted to RCC or adjudicated delinquent. Rearrest alone is not a guaranteed indicator of reoffending. Youth involved in the RCC program may be arrested for reasons such as fitting the description of a suspect, being in a “high-crime neighborhood,” or having had prior contact with police. As arrests do not indicate guilt, rearrest should not be the sole indicator of recidivism. However, because arrests even without subsequent petitions sustained have attendant social and fiscal costs, rearrest in addition to readjudication data is an effective indicator of program success. The relative rates of rearrest and readjudication of those diverted to RCC and those processed through the current juvenile justice system can be measured through a double-blind study or a matched sample. Youth who completed RCC had a 26.5% rearrest rate compared with a matched sample who did not receive RCC services, where 45.0% were rearrested. RCC youth who were rearrested subsequently were adjudicated delinquent 11.8% of the time as compared with 31.4% of matched sample youth who were adjudicated delinquent. Plan completion versus restitution/probation compliance. When youth cause harm, there are direct costs to crime victims. The juvenile justice system attempts to quantify this harm in monetary terms, and youth are frequently ordered to pay restitution to their victims. Youth adjudicated delinquent are also required to uphold the terms of their release. The closest measurable analogy for these two requirements in an RCC is plan completion. An RCC plan for youth to repair the harm has four parts: repairing • 15 the harm to the victim, to the community, to one’s parents, and to oneself. The plan is created by consensus of all involved in the RCC, including the crime victim and often law enforcement or the district attorney. We suggest comparing the rate of plan completion with county data on restitution and probation compliance. • Victim satisfaction. It is not possible to do a comparative analysis for victim satisfaction with the juvenile justice system because the juvenile justice system does not gather victim satisfaction data. Nonetheless, tracking this information is important as it can reveal citizen support for RCC, which in turn may empower policy makers to support an alternative response to offending that they may have initially avoided for fear of appearing “soft on crime.” Furthermore, high victim satisfaction rates through RCC would allow the county to realize improvements in the community’s faith and trust in the way crime in Alameda County is being addressed. Such potential improvements may also indicate a greater willingness of individuals to report future crimes and a decreased desire for retaliation as a response to victimization. Evaluation Considerations PFS projects incorporate rigorous evaluations in order to determine the level of success achieved by the service provider and the commensurate success payments to be made by end payers. Community Works, in partnership with NCCD, is undergoing an evaluation to determine success as discussed earlier in the report. These evaluations can be used to determine which outcomes are of interest to project stakeholders and what baseline impact can be expected from a project. During project construction, an independent evaluator will develop an evaluation design to measure outcomes during the course of a PFS project. An important evaluation consideration is what type of methodology should be conducted to maximize the statistical validity of the impact study. In general, randomized controlled trials (RCT) are considered the gold standard in evaluation, although there are several other evaluation designs that could be employed, including matched comparison groups, propensity score matching, regression discontinuity design, and other quasi-experimental designs. In an RCT, the evaluator would create two randomized groups within the target population, referring one group to RCC (intervention) while the other group proceeds through the judicial system (control). This approach ensures researchers are comparing “apples to apples” and isolating the treatment effect. A potential evaluation design for this type of randomized evaluation is detailed below for referrals from the district attorney’s office. Eligibility would be assessed at the district attorney level and verified by the project intermediary; all eligible youth would then be randomized to be referred to RCC or to be processed through the usual pathways in the juvenile justice system. 16 DRAFT RANDOMIZED DESIGN FOR Make It Right* 10/9/13 Youth arrested and referred to district attorney for filing determination on petition 1 TREATMENT GROUP 2 CONTROL GROUP Youth assessed for eligibility by juvenile division managing attorney # total assessed for eligibility # excluded based on program criteria (petitions filed on all excluded cases) # eligible cases referred for randomization # total assigned to treatment RANDOMIZED # total assigned to control (Same as total number of petitions filed for eligible cases) # refused to participate (victims and youth) # could not be located (victims and youth) ISSUE: Due to program design, randomization must occur prior to participant screening for willingness to participate. Therefore, the control group will not match all characteristics of the treatment group. Youth assigned to the control group will not be available for follow- up questions. True random number generator from random.org used for assignment ENROLLMENT # of youth enrolled # of victims enrolled *“Make It Right” is the name of the San Francisco RCC program. # of refusals and noncontacts returned to DA for filing Key Criterion: Safeguards for the Target Population to locate referred youth, youth refusal to participate, and youth otherwise deemed inappropriate for the program, NCCD and Third Sector project that any potential PFS project will only divert half of youth cases that would be eligible for RCC.20 Thus, the county will be able to evaluate the success of this program by comparing outcome data of RCC participants to youth processed through the traditional judicial system. The evaluation could also build in “early warning” indicators to flag whether the program is producing ineffective outcomes so that stakeholders can proactively address any issues. Projects must ensure that all ethical concerns associated with a project are addressed and that those served will be left no worse off than if not served through a PFS project. Current Safeguards There are few, if any, foreseeable negative consequences associated with expanding Alameda County’s RCC program. Due to such factors as inability 20 Because restorative justice is an accountability-based model, program enrollment requires youth to take responsibility for the crime they have been accused of. In addition, a small percentage of youth who are referred to the program and agree to participate exhibit mental health issues or attitudes towards their crime victims that make them inappropriate for a face-to-face dialogue. Those cases are also returned to the referring agency. 17 Preliminary Project Blueprint Next Steps A PFS project would require the development and execution of a contract with an end payer that ties payment to specific, measurable outcomes. Given RCC’s focus on diverting youth from the judicial system, NCCD and Third Sector believe the most feasible end payer would be a county government. Once feasibility work has concluded, project stakeholders generally progress through three phases of project development. • Procurement: An interested county will procure a service provider that can provide RCC services and deliver on related outcomes that to working with government, including an intermediary organization and an independent evaluator. • Project construction: In this phase, all partners develop, negotiate, and finalize the terms and conditions of the PFS contract and set up any necessary project infrastructure. • Project launch and implementation: Once the PFS contract is executed, project partners are responsible for implementing the project and carrying out their delineated roles of delivering services, evaluating performance, overseeing the contract, or other necessary responsibilities. Year 2 * Year 3 18 the government is interested in paying for. In this phase, other project partners also commit Year 1 Observation of Outcomes Below is an outline of how a hypothetical project may work for project partners. Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Services Delivered *Outcomes verified; success payment made Timeline for one cohort; cohorts referred and enrolled quarterly Organizational Ramp-Up to Full Capacity Services Delivered Value Proposition For purposes of understanding the economics of a PFS project, NCCD and Third Sector modeled a prospective five-year PFS project using available data from existing RCC services in Alameda County. In this scenario, Alameda County would agree to pay for successful outcomes achieved by Community Works by delivering RCC to youth referred from the district attorney’s office. Community Works currently has the capacity to serve 75 youth annually. The project model projects ramping up capacity over a three-year period to serve up to 300 youth annually. Over three years, the project would enroll 700 youth, growing enrollment from 100 in the first year of the project to 300 in the third year. We modeled a project where the government payer would pay $75 per avoided day in probation, realized both through diversion into RCC and reduced recidivism after RCC. These success payments would be capped at $4 million. We calculate that government savings, using the savings estimates discussed earlier in this report, would be approximately $6.5 million after making success payments to funders.21 Funders would provide $3 million in upfront working capital in a mix of loans and grants to finance the project. The sources and uses detail the uses of the provided capital. The service provider would require about $2.5 million in cost reimbursement, and we estimate $600,000 in PFS project costs. This is a very preliminary model, and project impact as well as the operations around project referrals and enrollments would have to be refined in the project construction phase of PFS project development. However, it does demonstrate that the economics of a RCC PFS project would work based on our current assumptions. Project Summary Total Referrals 1,400 Total Enrollments 700 Total Foregone Days in Probation 166,786 Gross Savings to Government $10,510,000 Total Government Success Payments $4,000,000 Net Government Savings $6,510,000 Sources and Uses (in Millions of Dollars) Sources Senior Loan $1.00 Junior Loan $1.00 Grant Capital $1.00 PFS Payments $4.00 Interest Payments $0.02 Total Sources $7.02 Uses Senior Loan Repayment $1.00 Junior Loan Repayment $1.00 Grant Replenishment $1.00 Service Provider Payments $2.45 Project Expenses $0.60 Interest Payments $0.26 Project Surplus $0.71 Total Uses $7.02 21 We caution that the $6.5 million figure is calculated from both the savings from diverting referred youth from probation into RCC, as well as avoided probation costs from the reduction in recidivism. This calculation assumes that those referred to RCC are not placed on probation, except for the 11% already on probation. 19 Considerations and Next Steps PFS is a relatively new innovation in the public sector; to date, seven projects have been launched across the country that have attracted over $75 million in capital from philanthropic and commercial sources, and more than 15 projects are in development. PFS is not a panacea, and not every service in the social sector lends itself to the features of PFS contracts. The criteria detailed in this report should be thoughtfully assessed and considered to evaluate specific PFS opportunities. Through our assessment of RCC in the context of PFS, we believe that Alameda County should consider PFS contracting as an innovative tool to scale RCC services in the county and to rigorously evaluate RCC’s impact. Other counties could also, depending on the RCC infrastructure in place, use PFS to scale RCC, although establishing RCC and scaling services need to be carefully considered. Additionally, we note the following high-level considerations for any interested stakeholders. • Service provider capacity building is critical for success. In order for service providers to rigorously implement RCC and scale services with fidelity in a PFS project, they need to carefully plan for expansion. This includes forecasting budgets, staffing, and process evaluation in order to ensure that they scale successfully. • Governments must engage with stakeholders to fully understand the potential savings and benefits of implementing RCC. While the costs and benefits detailed in this report provide a useful proxy for interested stakeholders, each government will have its own detailed costs and priorities in regards to juvenile justice and RCC. Government engagement is key for fully understanding how RCC can benefit a community and what savings to government could be tracked through a PFS project. • All project parties must provide commitment and resources prior to PFS project construction and launch. In order to successfully develop a project, stakeholders must dedicate resources and demonstrate their commitment to the project. This includes the following. » Resources: In addition to a senior government champion, senior staff from government and the provider must serve as point persons for the project to help coordinate resources. This requires participating in regular check-ins, expediting requests, dedicating resources for data requests and analysis, and reviewing materials. » Data: Projects, particularly project evaluators, need access to government administrative data on historical recidivism and justice system-related costs to determine baseline target population outcomes and potential avoidable costs. » Commitment: If a government moves forward with project construction, it will enter into a memorandum of understanding with all project parties to outline roles and responsibilities. Governing bodies will also need to demonstrate support and eventually commit to appropriated back-end contingent success payments upon launch of the project. In conclusion, this report has demonstrated how RCC could scale in Alameda County using PFS contracting. While significant work remains to be done to build and launch a project, we hope this report serves as a guide to beginning that work and as an example for other communities considering the use of PFS to scale restorative justice interventions such as RCC. 20