Social impact bonds for criminal justice: emerging lessons and key challenges in the UK and Japan

Following the launch of the world's first Social Impact Bond (SIB) in Peterborough in 2010, SIBs have been deployed in criminal justice across many countries. However, these have so far achieved mixed success in preventing reoffending. Two of the newest SIB programs in the criminal justice field – The Skill Mill in the UK and Ministry of Justice (MoJ) Pilot in Japan – build upon these precedents but diverge on some key aspects.

Similarities and differences

The Skill Mill and Japanese Ministry of Justice Pilot are similar in a number of key aspects. As has been the case with previous initiatives, the primary goal for both SIBs in the UK and Japan is to prevent reoffending. However, these SIBs target high-risk juvenile delinquents, who typically face complex and multidimensional issues including mental health and marginalisation from society. Both SIBs also deliver services in multiple locations across the country, and both providers (The Skill Mill in the UK and Kumon Ltd. In Japan) have experience in conducting outcome evaluations in collaboration with academic researchers and experts.

However, beyond these similarities, these SIBs diverge on some key aspects. The SIB in Japan is commissioned by the central government (MoJ), while the Skill Mill in the UK is jointly commissioned by local authorities with support from the Life Chances Fund. The leading provider in Japan is an established education provider (large-scale industry leader). The Skill Mill is a relatively new and locally based social enterprise, established by a former youth justice practitioner. And finally, while both programs use short-term recidivism reduction as the primary outcome metric, each set other unique measures such as intervention needs (MoJ) and employment (Skill Mill).

Emerging learnings

1. The commissioner – who pays for outcomes and how – shapes subsequent collaboration processes among stakeholders. It is crucial to recognise that each ministry and organisation has its own culture underpinned by its institutional history. In the context of Japan where new service providers and investors are risk averse and the use of SIBs is at a nascent stage, active involvement of criminal justice agencies in the program as commissioners enhances collaboration. This approach embodies the principle of Kaizen – continuous and incremental improvement – by facilitating collaborative learning and subsequent adjustments. In contrast, the co-funding model in the Skill Mill SIB enables engagement from local government and small social enterprises.
2. Relatedly, discussions about commissioning and scaling the programmes should prompt us to consider who is best suited to reduce the stigma offenders face in society. Young offenders with complex needs can often be stigmatised as ‘ex-offenders’. However, they can also be understood as economically excluded youth, deprived of access to education and employment opportunities. Depending on how young offenders as a target cohort are categorised in SIBs, who can and should be involved varies significantly. In Japan, central leadership comes from the Ministry of Justice, while in the UK the central co-commissioner (the Life Chances Fund) is the Department for Culture, Media and Sport (DCMS), which are not directly responsible for criminal justice. Importantly, in both cases local networks are built upon a vision and goals shared between local stakeholders. When delivering programmes across multiple locations, whether the service delivery mechanism is centralised or decentralised will affect the ways in which these local networks are nurtured to deliver the interventions.

Key challenges, and how they may be overcome

Ultimately, there is no one-size-fits-all solution for preventing reoffending. As criminological research suggests, the effectiveness of interventions is highly context-specific. In this regard, particularly relevant to the effectiveness of these community-based programs is the local context within which they operate. Both SIBs deliver their services along with existing statutory services and businesses, necessitating effective coordination with many stakeholders. The more the scope of their service delivery expands, the more challenging it is to navigate these coordination issues and potential imbalances in influences over decision-making and implementation.

The local ecosystem for service delivery in SIBs is built on trust with a shared vision and objectives. To build and maintain these relationships, the formal contracts of SIBs must be designed in a way that takes these complex group dynamics into account. The formal contracts must allow flexible adjustment to the services to enable prompt response to changing circumstances. Ongoing evaluation and learning can help to understand whether local partnerships are functioning, and whether their local services meet the needs of participants and their families.

Practical constraints on data sharing must be identified and addressed at an early stage. For instance, handling of sensitive and personal data on criminal records may require guidelines and protocols, especially when they are shared across sectors. Another key issue to address is how best to centralise such data when the services are delivered across multiple localities. Discrepancies in information made available to service providers and probation officers can lead to fragmented service delivery and may be confusing for users. Without consideration of these constraints, both implementation and evaluation may be severely limited by missing or inaccurate data.

Appropriate mitigation measures will depend on the extent of digitalisation, as well as the organisational structure. There may be different understanding of measurement and definitions around these data, even within public agencies. Since the taxonomies and data are not necessarily interoperable, resources should be allocated to create a point of contact who oversees data management. In the long term, this may help to develop the government's in-house digital capabilities.

Robust evaluation of the impact of the programme can be challenging. Young people’s own motivation to change their life trajectories is a prerequisite for the effectiveness of the rehabilitation programmes. Furthermore, the use of randomised control trials (RCTs) may be both impractical and unethical, as it may cause distrust among those excluded from services. Striking a balance between these ethical considerations and pursuit of scientific rigour continues to be a dilemma for justice policy.

Instead of settling with pre-post evaluations and group comparisons, quasi-experimental methods such as propensity score matching (PSM) may be a way forward for outcomes evaluation. While data interoperability across institutions may be limited, individual-level data in criminal justice tend to be rich and detailed in both the UK and Japan. Given the evaluation challenges, an approach to data management must be forward-looking, considering the broader use of the new data collected during the implementation of these SIBs to enhance our understanding of the effectiveness of their interventions.

Both centralised and decentralised approaches to administration have their own risks. Centralised approaches require large-scale intermediaries that are adequately flexible in responding to changing needs among local stakeholders, while decentralised approaches need leadership and stewardship to build trust among practitioners across jurisdictions and sectors. One of the challenges of horizontal deployment of SIBs in the field of criminal justice is the development of staff capacity and commissioning frameworks for dealing with these complex group dynamics.

SIBs can act as a catalyst for collaboration and skill development, but ground work must be laid for these programmes to achieve success. In this regard, a platform for convening exchange of knowledge among practitioners and experts from varied walks of life may contribute towards collaborative learning and capacity building. As with many of today’s pressing challenges, improved outcomes for justice policy requires a truly multidisciplinary perspective. The approach taken in the Skill Mill – under the leadership of GO Lab Fellow of Practice David Parks bringing his decades-long experience in youth justice before participating in a start-up programme to launch his own social enterprise – exemplifies how such diverse experience may enrich the programmes and contribute to innovative and flexible funding model.

Outlook for the future

As both countries look towards delivering effective interventions to prevent reoffending among high-risk young people at scale, a multidisciplinary perspective that transcends the boundaries across government departments – both at local and national level – is critical. As these two SIBs progress, further comparative analysis may shed light on how complex local networks can be nurtured, ultimately to improve outcomes for service users and wider society.

This blog was written with support from former GO Lab Policy and Research Officer Yuko Ishibashi.