Yes, I'll provide some overall comments on social impact bonds. I think you have had some comments on them from prior speakers as well.
We believe in the concept of a social impact bond, which is pay for performance, and which introduces new players and uses private capital, and we believe in de-risking it for governments, and so forth.
The challenge we see is obviously that you need to have perfect before-and-after attributions. They have to be very long term, because you don't know if a specific intervention was actually the one that was contributing to the outcome. There are a number of challenges with social impact bonds. I think it's still a very interesting concept that we should continue to try to pilot, but the program that we have constructed tries to take some of the challenging components of social impact bonds to make them more available and to utilize existing infrastructure as opposed to creating new infrastructure for the social impact bond.
In terms of the framework, I think the components of a social impact bond are very compelling, but if we can simplify the way it's implemented, we believe there would be some opportunities with it to leverage existing infrastructure and tweak it to become the social-ish impact bond.