Social impact bonds (SIBs) were supposed to revolutionise how governments fund social services. Private investors put up the capital. Service providers deliver programs. If predetermined outcomes are achieved, the government pays investors a return. If outcomes aren’t met, investors lose their money.

The theory is elegant. The practice has been more complicated. Australia was an early adopter of SIBs, with the NSW government launching the first Australian bond in 2013 — the Newpin Social Benefit Bond, focused on family restoration for children in out-of-home care. Since then, several more have been launched across multiple states.

So what do the results show?

The successes

Some Australian SIBs have delivered genuine results. The Newpin bond, operated by Uniting, consistently exceeded its outcome targets. The family restoration rate was significantly higher than the comparison group, meaning more children were safely returned to their families. Investors received positive returns. The government saved money on out-of-home care costs.

The NSW Benevolent Society Bond, focused on keeping children safely at home rather than entering care, also showed positive results. And the On TRACC bond in NSW, working with people on community corrections orders, reported improved outcomes on recidivism.

These are real achievements. The programs helped real people, investors got returns, and governments avoided costs they would otherwise have incurred. The SIB model worked as intended.

The challenges

But the story isn’t uniformly positive. Several issues have emerged.

Transaction costs are enormous. Structuring a social impact bond is expensive and time-consuming. Legal fees, financial structuring, outcome measurement design, and stakeholder negotiations can take years and cost hundreds of thousands of dollars before a single service is delivered. For smaller programs, the transaction costs can be disproportionate to the investment.

The pipeline has been thin. Despite early enthusiasm, the number of SIBs launched in Australia has been relatively small. The complexity of the model, combined with the time required to negotiate each bond, has limited scalability. Several planned bonds never made it to market.

Outcome measurement is contentious. Agreeing on what constitutes a successful outcome, how to measure it, and what the counterfactual looks like is deeply complex. In some cases, the outcome metrics have been criticised as too easy to achieve, meaning investors get returns for outcomes that might have happened anyway.

Government capacity varies. SIBs require government agencies that are willing and able to engage with complex financial instruments, negotiate outcome metrics, and manage ongoing relationships with investors and service providers. Not all government agencies have this capacity, and building it takes time.

Why they haven’t scaled

The original promise of SIBs was that they would attract private capital to social services at scale, reducing government risk and improving outcomes. That hasn’t happened, in Australia or globally.

The fundamental challenge is that the characteristics that make SIBs work — clearly defined outcomes, measurable counterfactuals, willing investors, capable service providers, and engaged government agencies — are relatively rare. Most social services operate in contexts that are too complex, too long-term, or too hard to measure for the SIB model to apply neatly.

There’s also a philosophical question: if a program works, why does the government need to pay investors a premium to fund it? Several commentators have argued that the transaction costs and investor returns built into SIBs mean that effective programs cost more to deliver through the SIB model than they would through direct government funding.

The counter-argument is that SIBs shift risk to investors and create accountability for outcomes. But in practice, the risk has often been mitigated by government guarantees or philanthropic backstops, reducing the genuine risk transfer.

What comes next

The social impact bond model isn’t dead in Australia, but it’s clearly not the silver bullet that early advocates hoped for. The more likely evolution is toward simpler outcomes-based contracting, where governments pay service providers based on results without the full SIB structure of private investment and financial instruments.

Some Australian governments are already moving in this direction. Outcomes-based commissioning models that share some features of SIBs — payment for results, independent outcome measurement — but without the complex financial engineering are gaining traction.

There’s also growing interest in outcomes funds, where government sets aside a pool of money for outcomes in a particular area and multiple service providers can access funding based on their results. This is simpler than a SIB and potentially more scalable.

The honest assessment

Social impact bonds have been a useful experiment. They’ve demonstrated that outcomes-based funding can work, that private capital can be mobilised for social purposes, and that rigorous measurement can improve program performance.

But the model is too complex and too expensive to be the primary vehicle for social service funding. It works in specific contexts with the right conditions, and Australia’s experience has helped identify those conditions more clearly.

The real legacy of SIBs may not be the bonds themselves but the outcomes-based thinking they’ve promoted. If Australian governments continue to shift toward paying for results rather than activities, that’s a positive development — regardless of whether the funding comes from private investors or public coffers.