Actuarial modelling is not evidence of effective social welfare: NZ academic


Using actuarial modelling of future costs as the key measure of the welfare safety net’s effectiveness is a “deeply flawed” idea imported from New Zealand, according to a former public servant turned academic.

The Australian government is barking up the wrong tree in copying New Zealand’s reliance on actuarial modelling to measure the future cost of welfare and judge the success of its reforms, according to an Auckland academic who is in Canberra for a conference.

Lowering the government’s future fiscal liability for welfare does not necessarily indicate a better functioning social safety net, argues Michael Fletcher, a former public servant who spent over 20 years working on welfare and employment policy before swapping Wellington for the Auckland University of Technology, where he is now a senior researcher hoping to shortly be awarded his PhD.

“It’s simply a cost-saving approach,” he says of the NZ government’s attempts to reduce long-term welfare dependency, guided by actuarial reports, rather than more detailed evaluation that considers other indicators around employment rates or general wellbeing.

Michael Fletcher

He argues the NZ approach is “deeply flawed” because it focuses “purely” on welfare costs, which is not an adequate proxy for the kind of outcomes social assistance programs are designed to achieve.

The lifetime fiscal liability does not even include all the wider savings to other areas of the budget that would be expected from more effective social assistance, he points out.

“I just think that it is the wrong measure to use,” he told The Mandarin. “A welfare system needs to be focused on people’s welfare, people’s wellbeing, and promoting employment, not on measuring fiscal savings.”

Fletcher will sit on a panel at tomorrow’s Power to Persuade symposium discussing “changing concepts of evidence” alongside Tara Oliver, the managing director of the Behavioural Economics Team of the Australian Government, and Australian National University academic Daniel Reeders.

The discussion will follow the keynote presentation on “evidence-based policy in a post-truth environment” by former Australian Human Rights Commission president Gillian Triggs. This should be a good introduction, as most critics of the NZ-style investment approach to welfare see it as a form of post-truth policy.

Conservative governments on both sides of the Tasman have pitched the idea as earlier, evidence-based interventions help disadvantaged people achieve better outcomes over the long term, meaning they require less government support.

But at the same time, both have been cracking down on recipients, enforcing stricter rules to make it harder and slower to apply, and easier to be cut off, and generally trying reduce the number of payments as much as possible.

“And I think that those two things are interacting with each other in a way that has a negative effect,” said Fletcher.

In NZ, he says, there’s been a “dramatic increase” in sanctions like temporary suspensions and cancellations of benefits, “mostly for things like failing to turn up to an interview on time – things that are relatively minor”.

The numbers of people on welfare are clearly going down in NZ but it is not clear why, or what else is happening to those people after they have stopped receiving benefits. The actuarial modelling of the lifetime liability of the welfare safety net cannot provide the answer.

“The numbers on the rolls are going down but there’s no evidence that these people are leaving to go into jobs or that their lives are better off, no evidence relating to that at all,” the researcher said. Some critics have suggested they are simply giving up and going it alone.

“We’re still in a situation — which is quite shocking in its own way really — that we have no evaluation.” Fletcher added. “So this has been running in New Zealand for five years or so … and we still have no evaluation of the welfare reforms.”

Fletcher, who urged the Australian government not to adopt the NZ approach in 2015, said there had been no discernible improvement in the government’s ability to help unemployed people find jobs. The unemployment rate hasn’t moved and child poverty is either slightly up or slightly down, depending on how you measure it.

“So the types of things that you would think would be showing up as a result of improved employment placements for people on benefit, we’re not seeing. But we don’t have a formal evaluation.”

The government of course, presents its actuarial reports as all the evidence anyone needs to show its welfare reforms are having a positive effect. “And they, by themselves, are being treated as sufficient, but they are not the same thing at all as a proper evaluation,” added Fletcher.

Actuarial modelling underpins the insurance industry, and is probably more appropriately applied to something like the National Disability Insurance Scheme than social welfare in either Australia or New Zealand.

“The New Zealand style of welfare is not insurance; we don’t have a European-style social insurance with a flat-rate social assistance underneath,” Fletcher said. “We’re like [Australia], just the flat-rate social assistance. It’s a safety net, it’s not an insurance scheme.”

It is hard to disagree with the idea of early interventions and social assistance that aims to improve people’s lives over the long term, getting them out of welfare dependency to the benefit of everyone. But managing such systems requires more than just the latest fad.

“I think it’s fair enough to use actuarial fiscal-liability modelling as one tool amongst many – it’s a very expensive tool, so I wouldn’t be at all sure that it’s value for money — but used as one tool among many, and using your standard, orthodox cost-benefit analysis that is the standard way of deciding whether a government intervention is warranted, and is the best intervention to choose, is the way they should be going,” Fletcher says.

“And using this an alternative to that is just flawed; it’s just fundamentally wrong. But using it as one measure amongst many, you could argue it has some value.”


Tomorrow’s Power to Persuade symposium is still open for registration — bringing government, academics and the community sector together to share knowledge across sectors and explore more effective ways of working with each other.

  • RWJ

    Our welfare expenditure is exploding, and will worsen with an ageing population; the reforms may well be geared primarily (or even only) to fiscal restraint but that doesn’t make that an invalid approach to a solve a real problem.

  • ATLANTIS

    “The Australian government is barking up the wrong tree in copying New Zealand’s reliance on actuarial modelling to measure the future cost of welfare and judge the success of its reforms”

    This is outrageous. The Australian Government uses modelling from the UK and US as proxies and applies the data to Australian society, not NZ. Taking our cues from non US/UK sources can only lead to disaster