It's difficult to specify in quantifiable terms exactly what we want, when 'we' are society acting through a government. It's even quite difficult to do so as individuals. For most of us, happiness cannot be readily expressed as a list of numerical indicators. We should probably all feel happier with incremental increases in bank balances, salaries, or years of healthy life, but for most of us our overall level of happiness is more a state of mind than a list of numerical circumstances.
This makes specifying targets for a Social Policy Bond regime difficult. Take something that seems readily quantifiable, such as climate change. Critical questions immiediately arise: do we want to mitigate or prevent climate change? climate change is likely to increase flooding, drought, storms and (in some countries) food shortages - would we be better off targeting these detrimental human outcomes, rather than climatic variables? But then what about the entire global ecology - is it to be valued solely in terms of the services it provides to humans?
Important and difficult questions to be sure, but exactly the same questions arise however we attempt to address climate change. When decision-making on behalf of our large and complex economies, let alone at global level, we do unfortunately have to look at quantifiable measures of success, otherwise we cannot reliably monitor how well we are doing. Gone are the days when a government could, for instance, recognise that Ms A receiving unemployment benefit (say) does in fact benefit from such a payment, while Mr B's long-term interests would be best served by putting some pressure on him to find a job. Perhaps the necessarily out-of-touch nature of a big, remote, government, and its expanding role in our lives, has something to do with our endemic anxiety and depression.
There are no simple answers, except perhaps to note that it's mostly at lower levels of income, nutrition, wealth, or environmental status, that well-chosen numerical variables correlate most strongly with what most of us would consider improvements. Social Policy Bonds are not perfect: they don't avoid this difficulty. In fact, they entirely subordinate policy to specified, targeted, quantifiable outcomes. While this can be irksome, especially to those who benefit from unspecified, obscure or mutually conflicting goals under the current regime, it's actually an advantage from society's point of view. Under a Social Policy Bond regime, hard decisions about what we want to achieve could not be ducked, as at present: they would, in fact, have to be made explicit to all of us, right at the beginning of the policymaking process.