The government has committed a lot of money to social impact bonds, but there are serious issues with the model, says Michael Birtwistle of NCVO.

Last November saw the Cabinet Office dedicate the not inconsiderable sum of £80m towards the creation of more social impact bonds (SIBs). While money for good causes is always welcome, it feels strange to see the government commit such a significant part of the Office for Civil Society’s budget to a largely untested commissioning model.

At the Budget, the Chancellor dedicated a further £5m to an existing SIB as part of the government’s homelessness strategy. The government appears to be proceeding under the assumption that the value of this model has already been established, when in truth, the evidence base is really quite limited.

SIBs are inherently expensive due to the costs of setting up their complex structures and the fact that the funder pays the investors interest. There are a number of supposed advantages to SIBs that are said to justify this additional cost – primarily bringing in external funding, and transferring risk away from public sector commissioners, which is believed to enable the testing of innovative approaches to solving social problems.

But there’s not a great deal of research to confirm the extent to which these results actually arise in practice, and there’s confusion over which of them is driving the government’s enthusiasm.

The government’s strategy for the model doesn’t follow through when it comes to innovation. There doesn’t appear to be any provision for developing and scaling up new interventions that are shown to have worked under a SIB, via traditional forms of commissioning – which would save paying the premium that the SIB model entails.

And as for bringing in external funding, it’s not clear whether the money being put into SIBs is genuinely additional, or whether it would have been used for traditional philanthropic funding anyway. We can at least hope that this new tranche of SIBs will try to address these issues.

The one aspect of SIBs that we do know quite a bit about, by contrast, is their payment-by-results (PbR) element. Last year the National Audit Office published an overarching review of outcomebased payment schemes, and made recommendations for commissioners to follow when designing them. The framework it published alongside the review suggests there are limited circumstances in which PbR will be the most appropriate model for commissioners to adopt.

We also know that commissioners have been making some serious design mistakes when it comes to PbR – see for example NCVO and Bates Wells Braithwaite’s work on this or Stephen Crossley’s excellent analysis of the Troubled Families programme.

This isn’t to say that SIBs can’t produce positive results. The evaluation report for the Peterborough SIB found that the programme had successfully reduced reoffending, but that with respect to the model, there was “no compelling reason to believe that SIB funding on its own fosters innovation”.

This observation brings us to the core of the issue. The real question isn’t whether they work in and of themselves, but whether they are more efficient and effective than an alternative commissioning approach in producing the desired outcome. I wouldn’t bet £80m on it.

Michael Birtwistle is a senior policy officer at NCVO