Fond of the social impact bond?

In an age of government austerity, there is particular focus on spending public money wisely to achieve desired social outcomes. New schemes will involve variants of payment by results and outcome based commissioning. Another similar tool is the Social Impact Bond (SIB). It involves targeted interventions to prevent social problems occurring and the consequent public spending, for example truancy resulting in NEETs and the cycle of re-offending resulting in greater criminal justice costs. Investors fund interventions up front and are paid by the national government in proportion to the spending cuts achieved. The model is likely to work best with the following conditions:

• Timing: reasonable gap between interventions and results, say seven years or less.
• Causation: there is a demonstrable strong causal link between interventions and outcomes.
• Savings: there is a potentially large cost saving for identifiable government departments.
  

An example is an SIB vehicle put in place for St Giles Trust, along with other organisations, to provide support to 3,000 short-term prisoners over a six year period. If successful, investors will receive a return from 7.5% up to a maximum of 13% in proportion to the government savings.

With great expectations of such projects come great risks that must be allayed. The Young Foundation's paper identifies 4 risks:

• Execution risk: there is the difficulty of the outstanding results of a local project run by passionate people that can not replicated on a larger scale. Selection of providers with the ability to scale is imperative.

• Measurement risk: even if a provider can show that the results are statistically significant because the sample size is large enough, and that there is no systematic bias judged by comparison to a control group, there is still the underlying danger of gaming the measure. For example, the apparently robust measure of reoffending is vulnerable because it is measurement of the criminal justice process and not fundamental change of an ex-offender. A pressurised manager could hit targets by persuading the police not to prosecute their users of relatively petty offences for exceptional reasons or by reducing the detection rate of criminal activity by their clients, for example, by moving to a different part of the country where they are less well known to police. Fundamentally, the contractor must be trusted or measurements must be changed to capture the change in an ex-offenders' maturity, social ties and personal identity.

• Basis risk: the reality may be that there are no great savings. For example, the local authority does not save money if a provider prevents young people from going to prison because young offender institutions are funded by the Ministry of Justice. Likewise, to save on prison costs a whole wing of a prison may have to close before government achieves any actual savings. Carefully agreed outcomes and sharing of rewards may make the difference here.

• Unintended consequences: where a current provider of services affecting the target group is not included in the SIB and withdraws support for what could be a host of reasons. For example, in the Peterborough prison example, prison staff not included in the SIB may obstruct its progress so that their lack-lustre performance is not highlighted by the new initiative's success.

I have high expectations that the Peterborough project will be a success and pave the way for greater uses of social impact bonds in the criminal justice system and beyond. The next challenge will be to introduce this at a micro level by overcoming high transaction fees through a simplified model.