Narrow social policy targets don't work. Take a goal like "halving the number of deaths from road accidents in the city centre". One response could be for city authorities to make driving attractive in all parts of the city other than the centre, with the same number of deaths occurring outside the target area. Another response (and one which would appeal to drivers in my neighbourhood) would be to encourage driving on the sidewalk, and exclude deaths occurring there from 'road' death statistics.
Even broad goals, such as improved health or educational outcomes, could conceivably be subject to the same manipulation or gaming, under a Social Policy Bond regime. Without very careful targeting, investors in the bonds could make unforeseen, negative trade-offs between societal goals.
Why not then target a single ‘quality of life’ indicator for the whole of society, taking into account all quantifiable social and environmental objectives: quality of life, physical and mental health, education level, environmental pollution, crime, homelessness unemployment, leisure time and any others? Surely targeting one single aggregated ‘social welfare’ indicator would be the optimal approach?
The more obvious objection to doing this is the daunting practical problem of defining a meaningful and measurable indicator of social welfare. The second is even more fundamental. Aiming for an increase in a single social welfare indicator carries with it an assumption that society’s needs can be traded off against each other. But for many of the needs for which government usually assumes responsibility such trade-offs cannot be made. For the neediest beneficiaries of government’s welfare programmes, a massive increase in priority for, say, health care would be unlikely to compensate for a total withdrawal of government funds from, say, basic education. ‘Safety net’ programmes in particular are scarcely amenable to trade-offs. In the same way a lowering of the crime rate, say, however welcome it might be, could hardly compensate for the total collapse of a country’s physical environment.
So experimentation and continuous refinement of the Social Policy Bond principle are going to be necessary. Issuers, public or private sector, will have to be vigilant to ensure that any particular bond issue does not break the spirit of society's stated and unstated goals, as well as the letter of the redemption terms. Of course, this sort of monitoring is necessary in today's policy environment as well - and it's not often practised.