impact bond seems flavour of the month. The US budget on 14 February was
sprinkled with references to the concept, there known as a “pay for success”
bond, and as much as $100m could be spent on them in the next few years. Other
countries are also interested.
to like the idea, because economically everyone’s incentivised in the right way
– the government only pays out once you’ve proved you’ve saved them cash,
charities get freedom to work the way they want to, and investors get what they
always want – a chance to turn cash into more cash.
Two key issues
need to be addressed, though.
One is whether
the bond can be priced to attract commercial investors. After all, this will
need a lot of up-front capital. Foundations and philanthropists can’t provide
all of it. And commercial investors will go where they can get the best reward
for the lowest risk, so the bond has to have an attractive profile for them.
original bond, working with offenders in Peterborough, there’s a lot of risk.
If the project goes very well, investors get a return of around 10 per cent a year.
back-of-the-envelope reckoning, that seems a slightly better return than you’d
get on the stock market over a similar period, and a significantly better one
than the bond market. But you’re very unlikely to lose all your cash investing
in those markets, and you can sell up if you start to get nervous.
The key question
is: how often do social impact projects go wrong? If the risk is very low, 10
per cent a year is good compensation. If high, you won’t find many investors
taking a punt for that return.
Until there’s more
data, it’s not surprising foundations are the ones carrying all the risk,
because they also care about social outcomes, and are used to funding projects
like this with no prospect of a return at all.
We won’t know
the risk associated with these projects until we’ve seen a lot of them, and
once we do, we’ll find out what sort of return commercial investors will
government can probably afford to pay whatever the market wants, because early
interventions save it huge amounts of cash, and it only needs to pay for those
potentially thornier issue which could derail the social impact bond is that of
metrics. How do you measure success?
This has proved
surprisingly tricky to nail down on the pilot project, even though it’s based
on an apparently simple metric: have released offenders committed more crimes.
This seems easy
to check, but there are a lot of potential issues. What if reoffending has gone
down because of something unconnected to your intervention, like a change in
the job market or an increase in benefits? How do you stop the people you’ve
contracted cherry-picking easy wins, and leaving difficult cases worse off than
And are you
reducing reoffending in a way that helps society? After all, if you offered
courses in how to rob houses, you might have less clients back before the beak,
and you’d get paid. But it wouldn’t cut crime.
The pilot seems
to have found a way around these issues. But a reduction in crime is easier to
measure than an improvement in other fields like children’s services or
economists are spending a lot of time thinking about this problem, so it seems
likely, in time, that answers can be found. But it seems likely that devising
good mechanisms for social impact bonds will keep a lot of bright guys in work
for a long time to come.