Services for Young People

Written Evidence Submitted by Social Finance

Could Social Impact Bonds be applied to youth services?

· Social investment in any form is not a replacement for the revenue that the youth sector is currently losing.

· Social Impact Bonds represent just one of a range of social investment instruments that will be needed to capitalise the sector – and will include debt, equity and equity-like finance. Earned income and contract income will continue to be important.

· Having said that, by raising investment from non-government investors, Social Impact Bonds do potentially offer a mechanism to pay for improved youth outcomes.

· Social Impact Bonds are potentially attractive because Government only pays for what works while service providers get paid for up front. As the financial return is tied to the social impact, the interests of Government, service providers, service users and investors are aligned.

· In many ways youth services offer good potential for Social Impact Bond financing – a range of experienced, high quality service providers exist with a good understanding of their client group and proven intervention models.

· Equally, we know that when young people aren’t supported the social consequences (and the public cost) can be great in terms of youth offending, teen pregnancy, poor educational outcomes, unemployment and mental health.

· The wide range of positive social outcomes constitutes part of the challenge of applying a Social Impact Bond to youth services.

· Often a single intervention can drive multiple outcomes that will benefit a number of government departments – for instance Teens and Toddlers run a teen pregnancy prevention programme that has also been demonstrated to have a positive impact on school attendance and achievement.

· Social Finance has been undertaking detailed feasibility studies in Essex, Liverpool and Manchester to assess the opportunities for establishing Social Impact Bonds to address the problems of some of the most vulnerable young people and their families. The failure to remedy severe behavioural problems among adolescents and the family circumstances that often lead to such problems creates enormous costs for the young people themselves, their wider communities and the public services that deal with the consequences.

· We consider that if a Social Impact Bond were to provide upfront investment and ensure the rigorous implementation of intensive therapeutic and family support programmes such as Multi-Systemic Therapy it could improve the lives of young people and their families and generate savings sufficient to more than repay investors.

· Our analysis suggests that the majority of outcome payments could be covered by reductions in the costs of foster and residential care alone. However, the programmes should also reduce demand on services such as health and the criminal justice system, and on welfare expenditure. To make the model work in a number of areas these benefits need to be reflected in the contract.

· No single government department is the obvious candidate to pay success payments on youth-focused outcome-based contracts like Social Impact Bonds.

· Without a number of departments (including Justice, Education, Health and DWP) contributing to outcome payments it is doubtful that the payout from Local Authorities would be large enough to enable investment in services.

· Developing Social Impact Bond contracts focused on youth services is therefore likely to require an expert intermediary and, in all likelihood, would be greatly helped by an interdepartmental agreement to jointly fund positive outcomes within both local and central government agencies.

Would Social Impact Bonds need to be targeted towards specific groups (e.g. young offenders) or could they be used to fund broader services (e.g. open access youth centres)?

· Part of the problem with existing funding for services in the youth sector and elsewhere is that it tends to be targeted towards preventing or encouraging specific behaviours (anti-teen pregnancy, pro-school attendance, etc.) rather than meeting the needs of individuals as a whole.

· While data in the sector is often limited, there does seem to be evidence that the same young people are often at risk of multiple negative outcomes – teen pregnancy, poor school performance, anti-social behaviour, substance abuse, poor mental health, etc.

· This implies that in fact there could be significant value to focusing outcomes-based contracts, like Social Impact Bonds, around multiple outcomes and hence broader services for those most at risk, although it may be possible to structure the contract itself around a single outcome that is tightly correlated to other positive outcomes and would reduce the complexity of contracting.

· One example could be a stated outcome of preventing entry into care for adolescents. Research highlights the poor social outcomes experienced by young people who have been looked after. A Social Impact Bond with the stated outcome objective of reducing care entry could be linked to a variety of positive social outcomes such as improved school attendance and reduced youth offending.

· To generate a social and financial return on investment, however, it seems likely that any open access services would need to be within targeted geographic areas – e.g. a defined geographic area known for poorer than average youth outcomes.

How closely defined would outcome metrics need to be for Social Impact Bonds to work? Could SIBs be based on counting heads rather than actual outcomes?

· Getting the outcome metric right is central to getting a Social Impact Bond to work.

· Changes to the outcome determine how much the public sector pays to investors and the financial return that investors make.

· The definition of outcome metrics is consequently determined by two factors:

o The confidence of investors that the metric will capture the value created by their investment; and

o The confidence of the commissioner(s) that the outcome metric is a reliable indicator of value to the public sector (be that savings, revenue or broader public goods).

· A Social Impact Bond focused on youth outcomes could potentially be based around metrics that the public sector already collects – teen pregnancies, school truancy rates, entry into care, grade averages, youth crime, etc. These could be benchmarked and independently audited.

· The bigger issue is therefore whether sufficient data around the impact of specific interventions exists (i.e. around what works to deliver outcomes) to give investors confidence that they will see a return on their investment when funding a particular set of interventions.

· In our work to develop a SIB with the objective of reducing entry into care by adolescents, we have studied a range of relevant interventions. The evaluation data of such services is varied in existence and quality. It is often difficult to determine which services are consistently effective in delivering improved outcomes and which are able to be replicated. The way in which services are implemented is key to their success. There needs to be a positive interface with Local Authority services.

· Data around outputs – numbers of young people worked with / literacy courses delivered, etc. – that are not linked to the outcomes that SIB contracts are based on are unlikely to give investors sufficient confidence to invest although it may be possible to persuade a small number of youth focused trusts and foundations to support a pilot on this basis.

Is it possible to measure the social and financial value of prevention? Are there issues of outcome attribution?

· There are two elements to answering this question.

· The first is the challenge of demonstrating the counter-factual – how to demonstrate that a negative outcome would have occurred had an intervention not been provided – this can be complex.

· In our Peterborough pilot, which aims to reduce reoffending among short-sentence prison leavers, the reoffending behaviour of each individual in the target population is tracked against 10 other individuals on the Police National Computer who are matched in terms of demographics and offending history.

· This ‘matched control’ creates a real-time baseline against which to judge the impact of our interventions. It gives us a good degree of confidence that we can measure the social value of the interventions.

· Finding such a real time control would be equally important in the youth services space where Social Impact Bond financing could be needed not only to build on positive outcomes, but also to simply maintain them in a context of falling funding.

· The challenge is finding non-binary measures that capture distance travelled.

· Social Finance does not currently have a clear sense of data quality in the youth sector. Further work would be needed to discover whether a cohort comparison approach (e.g. comparing teen pregnancies per 1,000 under 18 year olds in the intervention area to a control area) would be more effective than tracking future outcomes against historical behaviour or predicted outcomes for individuals.

· In our SIB development work around preventing care entry for adolescents, we have investigated a range of attribution methods. We are taking a pragmatic approach to this and believe it would be necessary to compare the SIB cohort to a baseline which reflects a similar cohort within the same Local Authority. This will ensure that Social Care thresholds are similar as will be the "service as usual" support.

· The second is that of attribution – that of demonstrating that it was this service, not that service that delivered the positive outcomes.

· While not perfect in this respect, one advantage of the Social Impact Bond model is that it seeks to capture the value created by a number of interventions rather than assuming that the same intervention will work for every individual in the target group.

· Intuitively, for young people facing multiple issues, multiple interventions may be needed – addressing literacy, parental mental health, housing, etc. – to deliver a positive outcome in terms of truancy, anti-social behaviour or school attainment.

· The set of interventions needed to deliver the outcome may vary from individual to individual.

Is the time to results in the youth sector a potential issue for Social Impact Bonds?

· Our first Social Impact Bond in Peterborough has a 12 month outcome measurement period for each prison leaver with a further 12 month lag on success payments to allow for court proceedings to complete and outcomes to be audited.

· It is easier to structure Social Impact Bonds around areas in which there is a fairly short period between intervention and results partially because investment is easier to find and partially because attribution is easier to demonstrate, and public sector value easier to realise.

· In our work to prevent care entry for adolescents, we can measure the days in care that have been avoided due to the SIB intervention by comparing to a cohort that has not received the intervention. It will not be feasible to measure the full extent of the benefit since for a 11 year old, that might mean measuring the days in care avoided over 7 subsequent years. Therefore there will need to be a combination of actual days saved which are measured over a reasonably short time frame post-intervention, plus a proxy to reflect the future savings to the local authority in terms of care days avoided.

· In the youth sector, the time to results issue could be resolved if the public sector were confident enough about their future savings to pay out on the basis of interim indicators – for instance if school grades aged 16 were a sufficient predictor of later employment / benefit usage to trigger a success payment from DWP for interventions provided at ages 13-14.

Where did the investment come from in the Peterborough Social Impact Bond? What are your expectations for investment sources for future SIBs?

· Investors in the first Social Impact Bond are mostly High Net Worth Individuals and charitable trusts and foundations. They include the Paul Hamlyn Foundation, the Barrow Cadbury Charitable Trust, the Esmée Fairbairn Foundation and the Rockefeller Foundation.

· Investors will receive a return if reoffending among the prison leavers falls by 7.5% or more compared to a control group of short-sentence prisoners in the UK. If the Social Impact Bond delivers a drop in reoffending beyond 7.5%, investors will receive an increasing return capped at a maximum of 13% per year over an eight year period. For example, a 10% reduction in re-offending would result in a 7.5% annualised return.

· Following the launch of the first Social Impact Bond to reduce re-offending in Peterborough, Social Finance has found enormous interest from social investors in subscribing to additional Bonds across a wide range of social issues.

· While a track record of social and financial returns develops, we would expect the majority of investors in the first Social Impact Bonds to have a social mission connection.

· Nevertheless, in time, we hope that Social Impact Bonds will develop the track record and scale to attract mainstream investment. In time this might come from pension funds or social ISAs in which a percentage of your savings go to support and improve society. We think this could be extremely attractive to people.

· In the shorter-term, we anticipate that the Big Society Bank will play a central role in catalysing greater investment into the youth space through both Social Impact Bonds and more conventional debt and equity investments to build organisations’ capacity to deliver services effectively.

Would a wellbeing index be useful within a social investment context?

· The definition of outcome metrics is determined by two factors:

o The confidence of investors that the metric will capture the value created by their investment; and

o The confidence of the commissioner(s) that the outcome metric is a reliable indicator of value to the public sector (be that savings, revenue or public good).

· Whether or not a wellbeing index would be useful as a basis for an outcomes-based contract would need to be tested with each counterparty.

· A wellbeing index could be helpful in measuring one aspect of improved social outcomes for young people at risk of entering care. Even if it is not immediately linked to cashable savings, it is important to have indicators other than the primary outcome metric to ensure that the social impact of funded interventions is positive.

05/2011