BY ALEX NICHOLLS
Some claimed benefits of Social Impact Bonds remain unproven. But they tackle long-term weaknesses in public service delivery by concentrating on outcomes, early intervention and collaboration.
Governments around the world are increasingly funding health care and other public services through Social Impact Bonds. SIBs claim to deliver better, more cost-effective public services by recruiting private investors who gain their returns by achieving pre-agreed outcomes. Earlier this month, for example, Prime Minister Theresa May promised to fund improved mental health services through this mechanism.
It’s easy to find serious flaws in Social Impact Bonds – particularly with respect to their main selling points. Like many social policy innovations, SIBs are championed for their efficiency and effectiveness, and for having superior - and better evidenced - impact on social problems. Experience so far doesn’t justify these assertions to any great extent.
Claims of greater effectiveness and rigour for SIBs remain unproven. Their supposedly enhanced efficiency – saving public money – looks largely spurious. However, we shouldn’t simply dismiss SIBs because of these doubtful claims. SIBs are a force for some considerable good, albeit in the ways that their champions tend to emphasise less. That’s because SIBs seem in practice to advance three important developments where conventional public services typically struggle: outcomes, prevention and collaboration.
SIBs and key public policy objectives
First, like Payment By Results (PBR), a central principle of SIBs is that they concentrate on the delivery of outcomes, on the consequences of public services rather than on their processes or outputs. This shift represents one of the most important – and potentially impactful – reforms in public service delivery globally. Utilitarianism may have been oversold as a moral philosophy but, as a general approach to deploying public expenditure, there’s a lot to be said for it.
Second – and this is connected to my first point – SIBs are focussed on shifting practice towards prevention and early intervention, because that’s how they can identify long-term savings.
Third, SIBs help increase collaboration in welfare delivery – they are designed to align incentives for multiple stakeholders to deliver a collaborative goal. Cooperation, with everyone pulling together, has bedevilled complex welfare systems. However, it is needed by SIBs if they are to achieve more with less and thus free up savings that can be translated into investor repayments. As a result, their approach offers a breath of fresh air to public service delivery.
So, although the headline claims for SIBs are largely unsubstantiated, some emerging practice is encouraging and much needed. Nevertheless, the verdict for now on SIBs should be “undecided”, given what’s at stake: the supposed benefits of SIBs and similar financial vehicles underpin arguments for the use of private capital in the delivery of public welfare services. I’d advise policy makers to avoid leaping too quickly to adopting SIBs on a grand scale, given what we know so far. Much of this is examined in a book that I’ve co-edited, entitled “Social Finance,” in which leading scholars in the field examine the issues, and in work done by Oxford University’s Government Outcomes Lab.
In addition to the questionable hype about SIBs, there remain further, serious questions about making profits out of vulnerable people, particularly in developing countries. There are also worries about sustainability - that SIBs will fail providers and users once contracts are completed. Finally, as with all PBR approaches, there are doubts about whether SIBs will be value for money in the longer run.
Such a shaky verdict may seem dissonant with a policy option whose popularity is sweeping the world. SIBs are everywhere. If your home country is not on the SIB map, it will be soon. Take Japan for example. It’s poised to join the scores of countries that are laying the foundations for SIBs. At the end of the summer of 2016, its government announced $200m for Development Impact Bonds (DIBs) (see below). So lots of Japanese SIBs are on their way.
Four myths about Social Impact Bonds
Returning to the “hype” I mentioned initially, let’s consider, first, the four core claims, or myths, that underpin the popularity of SIBs: that they are more effective and also more efficient in delivering superior social outcomes; they can make money for investors; and they are more rigorous in their measurement of change.
1. Superior social outcomes?
As far as superior social outcomes are concerned, the most famous UK SIB - for reducing reconvictions among ex-prisoners in Peterborough – did well. Its early performance was promising, suggesting that - once completed - the SIB would exceed its targets and deliver the superior social outcomes that were expected. It was just unfortunate that the experiment was terminated early for other reasons.
Meanwhile, the New Horizons SIB, developed for the Department for Work and Pensions, was meant to reduce numbers of NEETs - young people “not in education, employment or training”. After a rocky start, the SIB outperformed its targets, the contract was paid and the follow-on contract was awarded, albeit at a much discounted price.
There was also qualitative evidence from service users who said things such as: “If it weren’t for that practitioner, I wouldn’t be here today and that’s actually the truth. I’d be dead. Everyone I knew thought I probably wouldn’t make it until the end of the year.” So, there was impact. It might be of a young person who had been excluded from school who said: “Somebody came to see me, worked with me, got me back into school and back into education”. So, the impact was superior at least to what had gone before. But, maybe, there was no help there before, which makes it hard to compare.
So, there is some evidence of SIB successes - exceeding pre-defined targets, around reoffending, educational attainment, the securing of stable accommodation and full time employment. But SIBs have also failed to hit targets in some key areas and there are problems where there has been no baseline.
Sometimes, the performance metrics also seem artificial. In the Peterborough SIB, it’s plain that the particular targets were created to be very achievable because people wanted the programme to succeed. In short, it is possible that improved social outcomes are being achieved. There certainly are evident social outcomes. But it’s more difficult to say that better social outcomes than those achievable by other methods have been proven.
2. Do SIBs save public money?
The claim that SIBs can save public money is important. The financial logic works like this: once a SIB is in place, the cost to government of the status quo is reduced such that when you add the cost of the new intervention - plus a nice payback to investors - there is still a saving to the public sector.
Indeed, it’s a fundamental assumption, in these times of austerity, that SIBs save money. This is the simplest explanation of why SIBs are proliferating: they appear to give you something for nothing, which is very attractive if you are an impoverished service provider or an impoverished government. SIBs seem to get somebody else to pay for something up front, so you pay nothing, or, if you do eventually pay, you pay less than you would have paid anyway. Also, with any luck, repayment will be over a five or seven year period by which time, at the very least, you’re no longer the minister. So the pay back becomes somebody else’s problem.
There is certainly some technical truth in the saving money argument: services delivered now – because someone else is paying upfront – probably have more value than those delayed five years until the Government has the cash. Yet, beyond what’s called the “present net value” argument, the claim that SIBs save public money is probably the most spurious myth of all. In truth, it’s likely that the net cost of a SIB will be higher than the alternative and, even if it is lower, realising the savings may be nearly impossible.
For a start, some SIBs actually increase costs. For example, the Greater London Authority Homelessness SIB aimed to deliver support to people who were outside the welfare system, so it inevitably increased direct costs to government.
SIBs also have very high transaction costs. We underestimate the amount of management time that’s required to run a SIB and we radically underestimate the demands for information that investors make. So there are hidden transaction costs, often borne by service providers.
Realising notional savings is also difficult. If a SIB reduces reoffending how should that be expressed? Do you weigh up the numbers of sacked policeman, closed courts, shut prisons, prison officers put out of work? Realising those savings is clearly not easy. So it’s hard to see how savings estimates ever become a reality. Yet, saving money is one of the biggest sells to government regarding SIBs.
3. Making money for investors
It’s also claimed that SIBs make money for investors. There is some evidence of real returns but a lot of this depends on how the SIB is structured. The Peterborough SIB, for example, required investors to wait a long time to get any payback. In other SIBs, investors get some payback much more quickly, once there are some evidenced results. And, of course, the investor, who might be putting in a £1m, may never actually put all the money in up front. It’s typically staggered over a period. So if an investor has to put in £250,000 every three months, but starts getting back £100,000 after four months, the risk is much reduced and the returns are higher.
Financial risks in SIBs vary greatly. With Peterborough, investors took a 100 per cent risk – they would have lost everything if they had not hit their outcome targets. But this approach is less common than originally envisaged. The SIB at Riker's Island prison in the US, which Mildred Warner discusses in her contribution to this series of blogs, involved zero risk to the investor, Goldman Sachs, whose risk was underwritten by the philanthropists.
A famous social investor, whom I won’t name, was so angry about the Riker’s Island SIB that he practically had steam coming out of his ears. “This is the worst thing that could possibly have happened," he said, “because every single US SIB after this will say: ‘Why should I take any risk because Goldman’s didn’t?’ ” Elsewhere, in Australia for example, there is a two-tier risk model, with lower risk bonds underwritten by the Government and offering a low annual rate of return.
In short, there is evidence of some profitable returns from SIBs. But there is a wide range of experience already among investors from losing money to making money, to making some money and making money under circumstances that feel uncomfortable, and so on.
It is said that SIBs provide better evidence of impact. There is certainly no single approach to measuring impact across SIBs. There is a vast range from randomised controlled trials to SIBs where there is little or no comparison with other delivery options.
Peterborough was rigorous in that it measured reconviction events among its prisoner cohort, compared with the rest on the National Offenders Register. But the vast majority of SIBs lack such rigour. Most offer “validated administrative data”: somebody says this is what the intervention achieves and the commissioner says: “OK, I’ll pay on that” or “I won’t pay on that”. The metrics used also vary from quantitative to qualitative. So this claim about superior metrics is again not really proven.
SIBs do have some genuine strengths on the ground. They are part of a wider shift in programmes towards outcomes which should, if successful, deliver more meaningful public services.
The majority of SIBs that I’ve studied focus on prevention or on an early intervention. Very often they put larger amounts of capital into programmes up front than could be found using conventional funding. This is because programmes, particularly run by the third sector, are often small and don’t have the cash to upscale quickly. So, in speeding up investment and concentrating on early intervention, welfare benefits can potentially be achieved earlier than otherwise.
SIBs also allow collaboration across sectors in a way that is almost unique. In other words, if a SIB works, it should align the interests of the third sector, the government and private investors perfectly. So they should strive to achieve a particular outcome because they would benefit in different ways by that being achieved. This is an important asset for SIBs because the fragmentation of public services is a well-rehearsed obstacle to tackling pernicious problems.
It’s also probably true that SIBs are contributing, at least, to the building of a social investment or finance market – for better or worse – although that market is probably too small at the moment to have a major impact.
There is a big moral question, which Julia Morley discusses in her contribution to this blog series, of whether it is right to privatise, financialise or marketise - whichever language you prefer - social outcomes. This dilemma is perhaps best demonstrated in the case of Development Impact Bonds (DIBs), in which overseas development projects are debt-funded, like SIBs, with payments contingent on achievement of agreed targets. Making profits in this way from tackling the problems of the most vulnerable looks even starker in poor developing countries.
DIBs could be seen as representing a whole new wave of colonialism in which rich financiers say: “Not only can we solve your social problems, but we’re so clever we’re going to make money out of it, and take it back to London and Bonn and Frankfurt and New York”.
It’s also not clear how service providers will sustain their work once contracts for SIBs – or DIBs – are complete. What happens next? This issue affects service users who might say: “I was having this wonderful service and then the SIB stopped. Nobody’s picked up the slack and now I’m left with nothing.” For the service users, as well as providers, this lack of exit thinking is a real problem.
NAO question mark on value for money
But perhaps the biggest question mark brings me back to the beginning – value for money. The National Audit Office struck a note of caution recently in its report, entitled “Outcome-based payment schemes – government’s use of payment by results”. The NAO was mainly talking about PBRs but SIBs sit within that field. It advised caution in the adoption and embracing of such financial approaches. It said the evidence of impact was, at best, mixed and warned about the risk that such approaches might not, in the long-run, offer value for money.
My own overview of the subject suggests that there is very limited evidence for some of the claims made for SIBs. But they do, in reality, offer some innovative approaches – around outcomes, prevention and collaboration – about which public services have long aspired but rarely delivered.
We shouldn’t allow these benefits to be lost. Likewise, we shouldn’t be seduced by the hype.
Dr Alex Nicholls is Professor of Social Entrepreneurship within the Skoll Centre for Social Entrepreneurship at Saïd Business School, University of Oxford and co-editor of the book, “Social Finance”.