Social bonds involve private and non-profit organisations partnering with the Government to fund and deliver services to improve social outcomes.
If the contracts achieve agreed results, investors get paid back their investment plus a return.
The return depends on the level of results, up to an agreed amount.
So, the government is going to take a vital social service - e.g. caring for the mentally ill - financialise it, then bet on failure to provide an "incentive" for companies to "innovate" and "find solutions" to serious mental illness. Quite apart from market magical thinking, it sets some very perverse incentives and its obvious to see how it can go wrong. The government, while betting on failure, is accountable to Parliament and voters, so incentivised to see the bond-issue and the bond itself succeed, so it is incentivised against tough performance measures (just as it was for private prisons). And the contractor could be like Serco, and just juke the stats to get its performance bonuses. Oh yes, the Health Minister says it won't happen, but that's not what the overseas evidence shows. Stat juking and cherry picking are pervasive in UK contracted services, because its easier than actually performing the contract.
But the fundamental problem is that while this is pitched as somehow saving us money, what it really is is taking a social service performed by the government (or NGO charity) at essentially cost, and whacking a profit margin on top of it. And because its financialised, that will be a finance-sector profit margin, which means enormous. But the "innovation" the financial sector will deliver is almost certain to be nothing more than taking the bond, then hiring the same people who used to do the job to do it for less money, while taking a fat cut as middle-man. So we end up paying more for, or getting less of, the same service, by the same people, just so we can pay the banker's vig.