Social impact bonds: a tale of three Newpins

The Newpin family reunification program has been deployed under the Social Impact Bond (SIB) mechanism in three states: one successful landing; one termination; one taking flight. SVA has been a key part of the crew on all three. So, what have we learned?

• SVA Quarterly
• SIBs , Impact investing

The Newpin program has been operated by Uniting in Australia since 1998 and was based on a successful model that originated in the United Kingdom. The program was developed in response to the needs of new mothers experiencing issues such as mental illness, family violence, drug and alcohol misuse, and for those who were at risk of physically or emotionally harming their children.

In the Australian context, Newpin now primarily works to restore children in out-of-home care to the care of their parents. It works with parents – mothers and fathers – through a trauma-informed lens to enable them to create a supportive and safe family environment. Participating parents attend a Newpin centre two days a week for up to 18 months, where they receive therapeutic, non-judgemental support to address emotional issues, improve bonding with their children and develop positive parenting skills.

A new era

The commencement of the Newpin Social Benefit Bond (SBB) in NSW in 2013 marked a new era for the program. The SIB arrangement brought a much sharper focus on measuring impact, and a more deliberate approach to supporting families who could most benefit from the program. This is the core of how social impact bonds work – a clearly defined target population and robust evaluation of outcomes, because payments are explicitly linked to measured performance. This brings accountability, but also risk for service providers, which is in turn shared with investors.

The SIB arrangement in NSW also shifted the emphasis to restoration, or returning children home from state care, rather than preservation, or keeping children out of state care. This was partly because of the greater level of savings that could be generated for government by working with those already in the care system.

The Newpin SBB was a success, achieving significant positive impact for children and families, and delivering for investors through an innovative social finance model. The Newpin SBB matured in September 2020, having restored 391 children to the care of their families at an overall restoration rate of 61%, and delivered a financial return of 10% per annum to investors. (Read more in the final Newpin SBB Investor Report.)

The Newpin NSW SBB was an important proof of concept. It not only demonstrated the efficacy of the Newpin program itself, but illustrated how private capital can be used to support and share risk with service providers entering outcomes-based contracts with government. The results have also given the NSW Government the confidence to continue funding the program. The follow-on contract will retain an outcomes-based component, but investor capital is no longer required to co-fund and share the performance risk of the program. Robust measurement under the SIB contract has enabled Uniting and the Government to evaluate and appropriately share performance risk. Newpin continues its work to restore children home and keep families together where it is safe to do so.

Extension to Queensland

The Newpin SIB model was replicated in Queensland in 2017, with the first centre established in Cairns and a second in Logan. Run by UnitingCare Queensland (UCQ), the program prioritised working with First Nations families and tested the model’s efficacy in working in a remote location. Unfortunately, the bond terminated in June 2020, only three years into the planned seven-year term. The key issues which led to its termination are explored in the learnings below. Early termination of the Newpin Qld SBB was clearly a disappointing outcome, however it did demonstrate that the SIB mechanism worked, by bringing transparency and accountability to decision making. In short, despite positive outcomes for some families, the project was not producing results at a level required to deliver value for all parties, so it was wound up.

Next stop, South Australia

Next up was South Australia, with the Newpin SA SIB launching in early 2021. Uniting Communities, SA Government and SVA had the benefit of observing the Newpin program in NSW and Queensland and reflecting on the learnings from those states, and so could integrate those learnings into the South Australian context.

So, what have we learned?

Our experiences across three states have led to a deeper understanding of the Newpin model, where it works and where it does not, as well as the nuances of outcomes-based contracting. In this section we share some of our insights for organisations and governments developing outcome-based funding models.

This article assumes the reader has some familiarity with how SIBs work and are developed. For background information, you can read more here.

1.  If you build it, will they come?

One of the virtues of outcomes-based contracts is that there is a high level of focus on enrolling the ‘right’ (targeted) participants in the planned numbers. Outcomes-based contracts work best when there is a sufficiently large population of eligible people located within a defined catchment area. A large participant group reduces statistical volatility in outcomes measurement and has operational benefits in terms of critical mass.

There can be a gap between the theoretically eligible population and those that can be reached and are willing to engage. For centre-based programs like Newpin this is particularly relevant as parents are expected to physically attend two days per week.

In NSW, the original plan was to grow the program to 10 centres, but it proved difficult to identify sufficient numbers to support a centre in regional areas. In the end, only seven centres were operational. Referral pathways also had to be constantly managed, and turnover within the department’s caseworker team created the need to regularly provide information about how the Newpin program operates and which families are eligible.

A critical challenge in Queensland was that an insufficient number of families participated for the program to be able to meet its targets. This made it difficult for the Cairns centre in particular to build positive momentum in its work. Referrals were made solely by the Queensland Government through existing child protection system pathways. One of the contributing factors to the low number of enrolments was a lack of clarity about Newpin’s role in the Queensland child protection system, partly due to a shift in the service delivery and policy landscape, with a greater focus on family support services being delivered by First Nations community-controlled organisations.

The takeaway: An in-depth analysis of how many eligible people are in the catchment area is crucial. So too is effective ‘triaging’ so the right people are introduced to the right program at the right time.

2.  Accountability and control have to be in sync

Core to every outcomes contract is the metric(s) through which performance is measured and payments are determined. ‘What gets measured gets managed’, so selection of the right metrics is critical.  A challenge can arise if service providers have limited ability to manage a key driver of success.

The primary objective of the Newpin program is the safe restoration of a child to their family. In both the NSW and Queensland SIBs, payments from the government were based upon the total number of reunifications relative to the counterfactual or baseline. Breaking this metric down, there are two key drivers of the outcome:

1. the level of success in reunifying participating children with their families (efficacy); and
2. the number of families enrolled in the program (volume).

In Queensland, volume risk was generally managed by government but was to a large extent borne by UCQ and investors, who had minimal ability to control referrals, as described above. There was a minimum referral guarantee included in the contract, under which government was penalised for not meeting referral targets. However, a small participant group size impacted government’s value for money considerations, heightened statistical volatility in outcomes measurement, and created operational challenges.

In South Australia, a rate-based metric has been adopted, with payments linked to the proportion of children reunified to their families relative to the counterfactual. Under this metric, volume risk sits with SA Government, the party who controls and is best able to manage that risk. Payments to Uniting Communities and investors are cleanly tied to program performance risk, which Uniting Communities is best placed to manage.

The takeaway: The trick is to make sure parties are held accountable for what they have the power to deliver – but no more.

3.  Compared to what?

Trying to measure incremental change is what sets outcomes-based contracts apart. ‘Compared to what?’ is always the most difficult question to answer when developing a SIB. What is the relevant benchmark that program performance is measured against and how is that benchmark, or ‘counterfactual’, determined?

In the NSW SIB, program performance was initially measured against a fixed restoration rate of 25% while an administratively constructed control group was slowly established through identifying ‘Newpin families’ outside the centres’ catchment areas. In year four, performance was measured against the outcomes achieved by that control group, which had a restoration rate of 19%. Creating a control group proved to be burdensome and there were challenges finding sufficient suitable families that were exacerbated by reforms in the sector. For the last three years, a fixed counterfactual rate of 20% was adopted. A key challenge was that the population restoration data available during the development phase was not as sophisticated or readily accessible as desired, hence the need for a control group to validate the baseline assumption.

In Queensland, a fixed rate of 16.5% was adopted, based on an analysis of historical reunification outcomes for children likely to be eligible for the Newpin program there. At the first centre established in Cairns, an issue emerged that the enrolled families were experiencing more complex challenges than the general population analysis.  Parents presented with a wide range of significant social and personal issues including homelessness, pending criminal convictions and entrenched drug and alcohol misuse. The counterfactual remained at 16.5% regardless of the level of complexity of the enrolled cohort’s issues.

South Australia had access to an abundance of reunification data and a rich analysis of outcomes over several years. This meant that a more sophisticated counterfactual approach could be adopted. The analysis, commissioned by the SA Government, examined the impact of a range of factors on likely reunification rates, and it was determined that two key factors had a discernible predictive impact on reunifications for enrolled children:

• The amount of time a child had already spent in out-of-home care (children in care longer are much less likely to be reunified); and
• Whether the child identified as Aboriginal or Torres Strait Islander (if a child has Aboriginal or Torres Strait Islander heritage they are less likely to be reunified).

In SA we learnt from the challenges in Queensland with enrolling a cohort with more complex issues than anticipated. A segment-specific counterfactual approach was adopted that ‘flexes’ with the characteristics of the children who are enrolled without having to use a cumbersome and costly control group.

The takeaway: The counterfactual against which performance is measured should be as simple as possible – provided it is still fair.

4. The criticality of relationships and ‘buy-in’

The issues described above illustrate a key risk for any project that is complex, requires a new way of doing things, and extends over a long period of time – all of which describe a SIB. If personnel change, or priorities shift, or the grand purpose (and contract detail) is not clearly understood by everyone involved, then things can break down when challenges emerge. ‘Mission drift’ can also occur at an operational level if people are not clear about how an outcomes contract differs from more traditional commissioning arrangements.

In NSW, the partnership between Uniting, the Government and SVA brought diversity of thinking to the Newpin program and contributed to a stronger and shared understanding of the needs of families and solutions that work. Over the seven years of the contract, most of the partnering organisations experienced staff changes at both a management and operational level. New relationships had to be established, and ‘transaction memory’ was important in navigating issues that arose. Sometimes, new connections took time to form, but the overall structure of governance processes and operating rhythms was strong.

The takeaway: Ongoing nurturing of relationships and recapping of project objectives and mechanics should be deliberately undertaken at both governance and operational level.

5. Implementation risk

The contrasting fate of the NSW and Queensland SIBs illustrates that even though a program can be well structured and impeccably credentialled, implementing it in a new context is never straightforward.

Setting up any new program has challenges, regardless of the funding mechanism. For Newpin, finding and equipping a suitable property is a key part of the process, along with building and skilling a team, developing a centre culture, and – crucially – grounding the program in the community and the existing local network of services.

UCQ faced a number of issues in establishing Newpin in Cairns, including neighbours who weren’t thrilled by the location of the centre, and transport and access barriers.  They worked very deliberately on building connections to the local Aboriginal and Torres Strait Islander communities, including establishing a reference group to help guide the cultural integrity of the program and to support collaboration with key partners. They formed a partnership with a local community-controlled Aboriginal health service provider, with staff from that organisation delivering a key component of the Newpin program. They also employed First Nations staff (including the program lead) and paid attention to building a culturally welcoming centre. However, trust-based relationships take time to form, and that factor, coupled with a growing (and appropriate) policy focus on providing support through First Nations organisations, meant that the proportion of Newpin families identifying as Aboriginal and/or Torres Strait Islander was lower than initially planned. The complexity of issues facing enrolled families also meant that Newpin was reliant on participants receiving support from multiple aligned services, and those supports were sometimes not in place.

On the ground in South Australia, Uniting Communities is ensuring that its Newpin centres are either co-located with, or located near, other services, including alcohol and other drug services and Aboriginal support services. This aims to ensure that families with other needs and layers of complexity are adequately supported during their time in the Newpin program. The development of referral pathways is well underway, and planning for the commencement of services in July started in the middle of last year.

The takeaway: From an outcomes contracting perspective, allowing time for a program to get into a rhythm is important so that early results don’t disappoint. Implementation risk needs to be specifically considered.

6.  The best laid plans oft go astray

Although all parties use the best available evidence and their deep experience to predict what will occur during the term of the contract, the assumptions agreed during the contract development phase are inevitably not quite right. This is because it is difficult to plan for the unexpected and to forecast how new services in new areas with new providers will be received; it is the nature of forecasting. This challenge highlights the need for outcomes contracts to have flexibility and review points to allow the parties to come together and adapt the contract and/or services when things don’t go to plan.

In NSW, the evolving child protection policy landscape has, happily, resulted in a decline in the number of children entering state care. This reduced the number of children eligible for the program and in particular diminished the viability of establishing a fair and reliable control group. Contractual reviews allowed the parties to come back together to navigate these challenges, based on a shared understanding that Newpin was having a positive impact. The transition from the SIB contract to the follow-on outcomes contract also involved some complexity, as the documentation of how families enrolled at the end of the SIB contract period would be treated for payment purposes was somewhat unclear. Constructive negotiation was underpinned by a strong partnership between all parties – NSW Government, Uniting and SVA – and a shared commitment to the interests of the children and families.

In Queensland, the expected enrolment pattern of families did not go to plan in Cairns and the complexity of the enrolled families’ challenges was greater than expected, as described above. In addition, the service delivery and policy landscape shifted. There were a series of reviews which brought the parties together to identify the key issues and agree on both operational and contractual changes to counter their effect. Some changes were agreed at the end of the second year but challenges persisted and, unfortunately, the parties could not find common ground at the end of the third year, and the contract was terminated. UCQ and investor interests were protected as the breakdown in negotiations occurred early in the contract’s life, before formal reunification outcomes were available.

In SA, contractual flexibility and review points have again been built into the contract, with triggers, terms and consequences honed by the experiences in NSW and Queensland. Definitions and processes are also carefully documented to ensure there is little room for ambiguity, as we are aware that shades of grey can lead to confusion at best, and dispute at worst. The partnership between Uniting Communities and the SA Government demonstrated during the contract development phase bodes well for any future negotiations; both parties have a clear commitment to the families and children involved, and recognise that in order to achieve the best outcomes for them, they need to work closely together.

The takeaway: When things don’t go to plan, clear documentation and a shared objective can go a long way to finding a positive way forward. But if common ground cannot be found, all parties need an exit path.

In summary

The three different Newpin SIBs have highlighted some of the key challenges of outcomes-based contracting: measuring impact, sharing risk, having sufficient scale, determining the counterfactual and dealing with the uncertainty and unknowns of long-term projects.

We’ve learned more about how to manage each of these challenges through the creation and operation of each Newpin SIB, as well as the other SIBs we have helped to develop – as have our partners. It is becoming faster and easier to work through these design issues during the joint development phase of outcomes contracts. The growing body of expertise is also leading to more sophisticated analysis of risks and benefits, and the exploration of how concepts and capability can be applied more broadly to government commissioning.

At a program level, we have also seen the subtle evolution of the Newpin model, as the scrutiny that comes with outcomes-based payments has tested and strengthened practice and operational management. This strengthening should lead to many more families being successfully supported to provide a safe and stable home for their children.

Newpin in South Australia will commence services in July this year, and while we think it has been set up for success, there will no doubt be new wisdom to be gained in the years to come. We look forward to continuing the journey.

Author: Elyse Sainty

Contributor: Patrick Bollen

For over nine years, Elyse Sainty led SVA’s SIB and outcomes-based contracting work. She joined SVA in 2012 to help develop Australia’s first SIB, the Newpin Social Benefit Bond, and has helped service providers develop almost 20 proposals, been at the table during a dozen contract negotiations and completed the capital raise for eight SIBs.