I think in terms of the challenges around social impact bonds or pay-for-performance, my comment earlier about disproportionate attention is that they are a sexy new tool, so they get a lot of attention. But as you heard from earlier witnesses, the areas or the specific kinds of problems where they can be used are fairly limited. Not everything—social services or impacts that community groups do—can have measurable, short-term outcomes. So when we get into the types of interventions in the call groups, and the level of evaluation, the measurement, whether it's randomized, controlled trials that you need to actually demonstrate an impact, there's a huge evaluation cost that can be built in. I think that's one of the advantages in situations where they work.
A social impact bond is just about the only way I can think of that would actually resource the level of evaluation that would actually test the difference. But then, even if it demonstrates success—from what I understand some foundations in the U.S. are questioning—it becomes a recipe for that intervention. It's not always transferable to different environments, to every reality in different jurisdictions.
The other challenge is often the cost savings accrued don't go to one department. They might go to multiple departments, multiple levels of government. So having to bring together a number of investors to create the bond is the work of an intermediary, and it's extremely complicated. From our perspective, there are a lot of smaller-scale, well-established, existing programs that could have a tremendous impact beyond the social impact bonds themselves.
Just to answer your last question quickly, as Mr. Huddart said earlier, there's actually a significant amount of philanthropic capital looking and wanting to do some good. We don't see the issue so much on the demand side. If we give them a safe placement, there's probably an interest. There's really more on the demand than finding the right investments.