I think it would be incredibly difficult just for a simple social impact bond by the Peterborough project, which involved trying to reduce recidivism among people who had been released from short sentences. You have so many different factors that can impact it. If you happen to have a social impact bond project to reduce recidivism that starts up just after there's been a major downturn in the economy, then obviously it's going to be harder for people coming out of prison to find work. That's going to have an impact on the number of people who are reconvicted, and that's for a very simple project.
If you look at most social problems, there are so many different factors that come into play, you end up with an incredibly difficult model if you actually want to measure whether that specific project had an impact. For most social scientists, to try to measure the influence of particular factors, they use control groups and it can be a massive undertaking. You would essentially be having to do a major social science research project to measure the effectiveness of every single social finance project.