It is too early to judge social bonds, writes Jenesa Jeram.

For a concept few people had heard of before, social bonds have attracted a fair bit of criticism very quickly.

Some of the criticism is steeped in ideology.

But there is also criticism - potentially warranted criticism - that highlights the risks associated with performance-based contracting.

For those unfamiliar, the Government is developing a pilot scheme for social bonds (known as Social Impact Bonds internationally).

Social bonds are a new way of funding and delivering social services.

They harness the finance and expertise of the private sector through performance-based government contracts.

The first social bond announced will focus on delivering employment services to people with mental health conditions.

So far, so uncontroversial.

The private sector already provides social services in many forms: through NGOs, social enterprises, charities, and public-private partnerships.

Social bonds are just part of a wide range of initiatives the Key Government has introduced, including Better Public Services targets, Children's Teams, Whanau Ora, Social Sector Trials, and social housing reform.

Critics have characterised social bonds as experimenting on the mentally ill, or using some of the most vulnerable people in society as guinea pigs. But all new social services begin as an ''experiment''.

That is how we get innovation, and figure out what does and doesn't work.

What seems to be getting people worried, then, is the financing model of social bonds.

Social bonds require funding by private investors.

Investors then enter into a performance-based contract with the Government that is fully or partially linked to achieving certain social outcomes.

If successful, private investors will receive their principal back, as well as a return.

Ideological opposition to the model is inevitable. Various unions have criticised the model as ''putting profit before people'', ''private profit crowding out public goods'', and that ''services for those experiencing mental illness are best provided by the state''.

There are some risks with the model that need to be addressed in the contracting stage.

One concern is that the private sector would simply cherry-pick clients most likely to trigger a payout.

There is a risk the private sector would choose to take on the easiest cases (and reject the hardest cases), in order to make a profit.

This leaves the people who most desperately need the help uncared for.

It would be undesirable for either the Government or service provider to have the upper hand in determining the participants in the programme.

It would also be undesirable to compel people to take part in ''experimental'' projects.

A random lottery of willing participants to be part of the social bonds programme could solve this problem.

This would also mitigate the problem of self-selection bias, where participants are not representative of the general population.

The contracting stage will need to ensure fair assessment.

Another plausible concern is that of expense: that this social bonds project could end up costing more than the Government would pay for providing the same service itself.

Establishing a pilot scheme for any new programme is expensive.

And much of the infrastructure that social bonds require has not been sufficiently established.

However, there are many spill-over benefits of the social bonds model.

As well as saving taxpayers' money, if successful, the programme can generate experience and evidence in the areas of performance-based contracting, different financing arrangements between the public and private sector, and the measurement of social outcomes.

Finally, there is the concern of ''gaming the system'', where service providers meet the letter of the contract but not the spirit.

For instance, mentally ill people may be forced into work before they are ready, so the investors can achieve their outcomes and make a profit.

This, too, must be dealt with at the contracting stage, which will outline exactly what outcomes will be achieved, and how they will be measured.

Participant wellbeing could very well be one performance-based outcome that will be monitored and recorded.

The New Zealand Initiative has recently released a report on how social bonds have worked overseas, and the key for their success in New Zealand.

One thing we will be looking at closely is the contractual detail of the social bonds programme.

Risks such as cherry-picking, gaming and cost blowouts must be mitigated.

We will also be looking at how political risks (such as a change in government) are dealt with in the contract, and how much input (beyond providing funds) private investors will have in social service provision.

If the risks are adequately addressed, a successful social bonds pilot scheme will not just deliver results for the programme recipients.

It could also grow the social investment sector, offer lessons for other forms of performance-based contracting, and grow the evidence base for outcomes achieved.