Bonds won't work for every social service

Academics warn there is not enough evidence around social impact bonds to put much faith in the fashionable financial instruments yet. Even the New South Wales Government’s successful Newpin pilot raises questions about how widely they can be applied.

The trial of Australia’s first SIBs — the NSW government calls them “social benefit bonds” — has worked “exceptionally well” for the Newpin program, according to Sally Cowling, who runs research, innovation and advocacy for UnitingCare Children, Young People and Families and helped set up the special funding arrangement. It was also a complex, difficult, and time-consuming process, she adds, and going down the same road might not be worth the effort in many other cases.

“It’s like nothing we’ve experienced as an organisation,” Cowling told the recent Power to Persuade Symposium in Canberra. “This was the first social impact investment in the out-of-home care space in the world. It was a really, really hard thing to work out. Intellectually, it did my head in. The transaction costs were huge … but we hope they will go down over time.”

NSW Treasury associate secretary Caralee McLiesh expressed a similar view, speaking before Cowling at the forum on a panel that discussed the provocative question of whether SIBs are in fact “a wolf in sheep’s clothing”.

“We’ve learnt from the pilots that they are incredibly complex and time consuming, and they’re simply not [always] appropriate,” McLiesh said. “Perverse incentives and measurement challenges … are very real [problems], and it’s not the case that [every] social service is going to be appropriate for a bond.”

The NSW Government’s social impact investment policy, McLiesh explained, does not focus solely on SIBs. “But there are some underlying principles behind the bond that can be applied in different types of transactions: outcomes-focused grants, or ‘payment by results’ contracts, or layered investments, as we have seen in the successful GoodStart transaction that Social Business Australia championed.”

In Cowling’s view, it would be difficult to identify a large number of social services that would benefit from funding through the bonds, which only work for programs that create positive social outcomes and save the government more money than they cost.

“Social impact bonds can hardly work for anything, because you need a really strong correlation between the intervention and the outcomes, and when you realise the outcomes, they’ve got to start generating returns quickly,” she told the forum.

“The majority of programs that my organisation runs, we generate revenue … 15 years down the track, when children we do good preventative work with when they’re three and four-year-olds finish school, rather than ending up in [juvenile detention].”

“Those things don’t work under a social bond.”

Healthy scepticism

Throughout the day, several academics pointed out a serious lack of evidence to justify the current enthusiasm for SIBs and suggested similar results might be achieved through best practice approaches albeit with simpler funding models

“There’s no doubt there’s an infatuation with social impact bonds in particular in many parts of the policy world,” said University of Melbourne public administration professor Janine O’Flynn, who chaired the panel discussion.

“Many people are arguing they offer some sort of ‘Holy Grail’ which can bring together the public, for-profit and non-profit sectors to address some of our most intractable problems.

“Like any hot new trend, many jump on the bandwagon, regardless of whether it’s going to suit their needs or not.”

From her own reasearch, O’Flynn said all she could find out so far about SIBs was a lot of faith but little evidence: “There’s a lot of talk about how they will be collaborative, they will focus on outcomes and they will be innovative. Again, time will tell.”

The director of Swinburne University’s Centre for Social Impact, Professor Jo Barraket, said the available evidence in favour of SIBs was “mixed, verging on weak” in her assessment.

“They are not social innovation,” she said. “Right now, all they are is a financial novelty. There’s a difference between novelty and innovation. Innovation is new and improved; we can’t possibly know whether we can make a claim on improvement at this stage.”

Barraket added that when SIBs fail, even if the obvious financial cost to the taxpayer is limited, the real cost is mainly felt by those in the community whose needs are not met and could be obscured from view.

O’Flynn’s University of Melbourne colleague Helen Dickinson said she also noticed a lot of enthusiasm in the public sector for SIBs and could see why.

“I guess they do sound like the best of all worlds in a number of senses,” she said. “We’ve got private capital coming into the system, governments thinking about outcomes, community organisations involved in delivery and there’s talk about collaboration and innovation and co-production, an investment approach, and there’s claims about what this does in terms of risk as well.”

But when it comes to the evidence, Dickinson agreed with the others. She points out there are less than 50 SIBs currently operating worldwide, and significant evidence to suggest they increase transaction costs, as suggested by Cowling and McLiesh.

“Personally, I have reservations about them as a model, and I haven’t been entirely swayed by the kind of evidence that exists so far today,” Dickinson said.

“I think … they’ve got very particular applications; they’re suitable for some contexts and not more broadly. One of my concerns in the enthusiasm for it as an idea is it will be taken up and applied in areas where it doesn’t necessarily make sense and isn’t entirely appropriate.”

The possibility of fixing the now equally famous “wicked problems” at the same time as exercising strict fiscal constraint is hard for political representatives anywhere to ignore. But Dickinson warns that these problems have underlying causes — such as entrenched socioeconomic disadvantage in specific locations — that SIBs do nothing to address.

“Also, are there alternative models that we could use other than social impact bonds that would give us similar sorts of results? So are the results that you’re seeing … because of the social impact bonds, or is it because we’re starting to think about outcomes in a really serious way?

“When they work, is it the social bonds that are working, or is it the focus on outcomes?”

‘We did it to save Newpin’

Cowling added that UnitingCare was assisted in the complex job of getting the Newpin bond working by its significant size and financial weight, which allows it to invest in the research and innovation centre she leads. It applied for the NSW social bonds pilot because it was facing the decision to shut the program down in the absence of continuing government grant funding.

“We did it to save Newpin,” Cowling said. “We didn’t do it because we believed that social bonds were in any way ‘the answer’. There is no ‘the answer’ to gaps in social policy.”

In many ways it was a stroke of luck, she said, that the Newpin program was suited to social impact investment.

McLiesh also felt the need to clarify the government had not simply jumped on the latest policy bandwagon, either.

“Yes, there are risks, no doubt, with social impact bonds, and social impact investment more broadly, and they are not a panacea,” said the Treasury official. “We do not pretend that they are a panacea.”

McLiesh pointed out the $17 million invested in SIBs was small compared with the state’s $45 billion social services budget. “They’re pilots; they’re trials; we’re keen to learn,” she said.

In that case, O’Flynn later asked, why does it make sense to put so much more scrutiny on the pilot programs than the much larger area of directly funded social services? And if the SIB-funded programs achieve good outcomes, could they do the same with a simpler form of funding?

Adapted from a program of the same name designed in the United Kingdom, Newpin works with troubled families to break cycles of abuse and neglect. It aims to keep families together, which means less children in foster care. This makes it well suited to a social impact bond, Cowling explained, because the cost of running it is less than the support payments that would otherwise go to foster carers.

Cowling suggested that some of her colleagues in the not-for-profit sector were misguided in their concerns about the program, which has previously been characterised as “cash for kids” by critics. The first questions that UnitingCare looked at, she said, were around the consequences for ethics and social justice.

“Our biggest concern was about perverse incentive effects. Would this funding model cause us to go out  chasing ‘easy families’ because there were payments to be made?

“Would it encourage us to focus just on outcomes per se, so to get the ‘cash for kids’ as opposed to the sustainability of the outcomes? If our restorations don’t hold for a twelve month period, we pay the money back to government.

“They were incredibly important things for us to grapple with, as was protecting and preserving the fidelity of the Newpin program model. The most exciting thing has been what has happened in the first two years. The Newpin program has restored 66 children in care to their families. It’s prevented children from another 35 families entering out-of-home care.”

Cowling said she had been surprised by how many people in the charitable sector look down on the idea. She said not-for-profit service providers were “not particularly accountable for making change” with the traditional grant funding they received, while the bond had made Newpin more focused on outcomes than before.

“It’s been a big cultural shift for the organisation,” she said. “The thing that excites us about the bond is this program is delivering better results for a much higher-risk group of families than it was in the ten years where we self-funded it.”

McLiesh added: “The fact that under social impact investment we’re paying on the basis of outcomes really sharpens the focus on measurement, on what we’re trying to achieve. There’s a lot more transparency on what’s working and what isn’t, and more accountability for the funding.”

The associate secretary, who heads up NSW Treasury’s agency budget and policy group, said that focus on outcomes also led to small but important changes in the practices of the not-for-profit service provider, including a greater reliance on evidence, and adjustments to the Newpin program to make sure it targeted the highest-risk individuals.

Falling back on the same standard selling points of SIBs, which the academics say are largely without evidence, she said the arrangements recognised that the public, not-for-profit and private sectors could all achieve more than they could individually.

Organisations like UnitingCare are “closer to the client and better able to meet their needs than government” while investors could bring “more commercial discipline, risk management and governance” to the partnerships, according to McLiesh. She also expressed hope that SIBs could underpin longer-term solutions:

“Government budgeting is very much focused on the incremental, the annual. Social impact investment forces us to look at the question of if we invest now — and we are investing almost $50 billion a year in social services — what is the impact of that over the long term? How do we look at the costs and benefits over the long term, and how can we change the trajectory according to the different interventions we do now?”