ANALYSIS
The Government is readying Social Investment 2.0, aiming to target society’s most vulnerable from an early age before they’re lost to the wrong side of the tracks. It should be win-win - for the person whose life trajectory is altered for the better, and for saving money otherwise spent on welfare, housing, health, prison time. But what sounds good in theory has pitfalls in practice. Senior writer Derek Cheng explains.
“I want Social Investment in the bloodstream of how we do things in Government,” Prime Minister Christopher Luxon said when he recently unveiled his public service goals.
Championed by Sir Bill English under the previous National-led Government, the model is about using data to target the right people with the right interventions as early as possible to break the cycle of poverty and deprivation.
It’s such a high priority that it has been put into the hands of Finance Minister and Social Investment Minister Nicola Willis, who detailed who it targets in a recent speech.
“The growing numbers of young people not attending school regularly, the children [who] bounce from shattered home lives to state care, and too often into gangs and eventually the criminal justice system, those who start on a youth benefit, and who are expected to spend an average of 24 years of their working life on an unemployment benefit,” she said.
It involves an actuarial assessment, a term borrowed from the insurance industry; future risk is measured in the costs to the state of a person’s life gone awry, while success is the quantum of future savings if that life stays on track. Data is continually collected so effective programmes can be supercharged, while the “dumb stuff” - in Luxon’s words - is cut.
It’s also about harnessing private sector money through social bonds, and returning a profit if the programme ticks the right boxes - a fund the Government wants all New Zealanders to be able to invest in.
But while it might look solid on paper, the results from Social Investment 1.0 were mixed - though it was still in its infancy - and the Labour-led Government moved in a different direction.
Concerns included the wrong targets (helping individuals with poor outcomes rather than the system that created those outcomes in the first place), wrong priorities (numbers over people), and the potential for perverse outcomes, like a big tick for taking someone off welfare no matter what happens to them afterwards.
Then there are the Trojan Horse theories: That it’s really about shrinking the size of government and privatising social services.
So what happened with Social Investment 1.0, and how will 2.0 be any better?
The ‘who’: Our most disadvantaged
A small slice of the population experience intergenerational trauma and violence, and are stuck in cycles of deprivation and poverty.
The size of the slice depends on the severity of hardship: some 20 per cent of kids might need help, 5 per cent are likely facing a life of dysfunction, and 1 per cent will almost certainly have severe behavioural issues and end up in prison.
As they grow up, they have constant interactions with the state through welfare or housing or healthcare, or when they become entangled in the state care or justice systems. The interactions mean the Government knows who they are, and could - in theory - step in with effective intervention from an early age.
But this is exactly the kind of support that is sorely lacking, successive governments have been told in multiple reports across a range of topics including justice, education and employment, social deprivation, and local government.
Here’s a typical example in an expert report on child offenders: “A child who has been in the Family Court for five years, maybe longer. As soon as he turned about 10, he started offending but of course didn’t get taken into Youth Court because of his age. And he’s now in Youth Court in a major way and resources are just being started to apply to him at age 14. He needed it when he was 5, not 14.”
This child and family will probably have intersecting issues - across education, employment, housing, poverty, crime - so the best way to help is with cross-agency, wrap-around support. A provider offering food coupons, for example, would be a lot more helpful if it could also provide secure housing and addiction support.
But the system rarely does that. A number of systemic barriers have been pinpointed in numerous reports:
- Government silos and fragmented ministries and agencies that focus on their own outcomes and don’t co-ordinate;
- Short-termism, meaning programmes don’t get funded for long enough to make a real difference;
- Institutional discrimination, which feeds inequities across all those intersecting issues.
Social Investment 2.0, like countless government attempts before it, will do little to shift the dial if these barriers can’t be overcome.
Big Data - Finding the ‘who’.
The launching pad is the Integrated Data Infrastructure (IDI), which links individual-level data from different government agencies, including tax, health, education, and social services. It covers all New Zealand residents, is longitudinal, and regularly updated.
The data is anonymised, and the model can predict how often you’ll miss school, or be on welfare, or in court. “We could pull out 1000 6-year-olds that we know are most likely to cause harm and end up in jail,” a source close to Social Investment 1.0 told the Herald.
An example of its sophistication is the Treasury, in 2016, using it to identify four risk indicators of poor long-term outcomes: abuse or neglect as a child, a mother with no formal qualifications, a parent with a prison or community sentence, and being mostly in a welfare household since birth.
Compared to children with zero indicators, those with at least two are 10 times more likely to be in jail before age 21, and three times more likely to have no school qualifications. This analysis helped to develop risk scores for every school student, which formed the basis of the Equity Index that determines school funding.
The issue - as with any dataset - is that any policy is only as good as the information it’s based on.
Further number-crunching from the Treasury showed that a third of those with at least two indicators (42,5000 children) avoid poor outcomes. This prompted academics to say the link was “at best tenuous”, while Sir Peter Gluckman, as chief science adviser to the Prime Minister, offered this caution: “The data-driven approach is not a ‘silver bullet’.”
That doesn’t mean data shouldn’t be used, according to ImpactLab, a company set up by Sir Bill English that’s been applying Social Investment principles in the NGO space for the last four years.
“Often the data is really poor and really messy, but once you start using it, you identify where the gaps are, and the best places to invest for better visibility,” says chief executive Maria English, Sir Bill’s daughter.
Just one piece of data can be profound, she adds, pointing to ImpactLab’s work with a community housing provider.
“The local former DHB shared data with the provider about the hospitalisations of the tenants - people with severe mental health conditions. Just that little bit of data shows that, prior to getting the housing, those tenants were spending 100 days a year in hospital. After they got that appropriate housing, it dropped by 90 per cent - and it stayed low.
“That’s only one metric - avoidable hospitalisations - but it’s the tip of a pretty important iceberg about what’s going on for those people, and how we can effectively support them.”
The ‘what and the how’: The ‘Social Investment Experiment’
Social Investment 1.0 sprouted from the 2011 Welfare Working Group but, by early 2017, officials said it was still “in its infancy”.
It was first embedded at the Ministry of Social Development to get people off long-term welfare from as young an age as possible. Beneficiaries faced stringent sanctions and work-related conditions, with a particular focus on 16 and 17-year-olds, and 18-year-old parents. Some of their welfare payment was set aside for essentials such as food, and they had a case manager to help with budgeting and meeting benefit conditions.
Financial incentives were introduced for parenting classes and education/training because, in the words of a source close to the reforms, “if you’re uneducated and young, you’re going to find it much harder to get off welfare”. It’s even harder for teen mothers with a baby to care for, so women on benefits and their daughters were offered free long-term contraception in Budget 2012.
From 2011 to 2016, the number of those on a main benefit fell by almost 45,000, and as a proportion of the working-age population, from 12.1 to 9.8 per cent. A 2015 Ministry of Social Development report said that Social Investment-inspired reforms resulted in $12 billion in savings over four years - and much more over a lifetime, should they stay off welfare.
But academics raised questions about the dangers of too narrow a focus at the expense of the bigger picture, as did the Productivity Commission.
“Slavish application of an investment approach based purely on costs and benefits to government might lead to perverse outcomes,” the commission said in a 2015 report, noting studies about saving money due to obese people dying more quickly. “A health system that sought only a reduction in future health costs might therefore do little, if anything, to discourage obesity.”
There’s no suggestion this might happen, but actual examples exist where the financial focus was too narrow. Take the disability employment charity that partnered with ImpactLab to find work for long-term beneficiaries with disabilities.
The contract incentivised placing ‘X’ number of people in jobs, but the challenge was keeping a job because most of them returned to welfare “quite quickly”, Maria English says.
“So this charity was doing a lot of things - mentoring people on the job - to help them sustain employment, but there was no reward for that in the government contract. In fact, the incentive of the contract was: ‘Don’t focus on that, just focus on putting bums in seats.’
“The main value they [providers] care about is the families themselves: improvements in people’s subjective wellbeing, other benefits that aren’t so clearly defined and that can get overlooked.”
This is exactly the scenario that sceptics of Social Investment feared, but it isn’t necessarily a characteristic of the model; the government contract in the above scenario was under Labour’s watch.
The danger of easy ticks
Willis is aware of the tricky balance in getting the outcomes right.
“The danger, if you’re too ‘big picture’ about it all, is that everything might tick some boxes,” she says. “‘Did I enjoy the programme, and I felt good about it?’ Tick, successful. Secondary impacts are good and they’re nice to have and there’s some value in evaluating them but, ultimately, we have to be accountable for the result we are investing to achieve.”
The overarching test should be whether someone’s life is being improved, she says. That means it’s “not a win” to have lower welfare numbers because a beneficiary or someone leaving emergency housing ends up in prison, or by preventing those eligible for a benefit from getting one.
How can this be avoided? “Clarity about the what and the why” in the contract, she says.
How that clarity is decided will be crucial. It would ideally minimise the chances of gaming the system by collecting some self-evaluated data to show some self-selected “return on investment”.
Sometimes the outcomes you want are straightforward, such as for the Healthy Homes Initiative, where the metrics were demonstrable in hospitalisations per person (down 19.8 per cent), school absenteeism for children (down 3 per cent), and welfare savings (down 9 per cent for those aged 24 to 64).
At other times, it’s not. Should “subjective wellbeing” be part of evaluating the impact of recreational physical activity? How about “feeling worthy and confident” following a parenting programme? These evaluations (which included several other measures) were part of Labour’s direction to measure “social return”.
Social Investment 1.0 might have ended up in a similar place anyway if it hadn’t been redirected. A stakeholder told a 2021 academic paper that “social and economic improvements for vulnerable and complex families” were being measured for 1.0 work in the justice and child protection areas.
Other challenges highlighted in the paper, which drew on interviews with Sir Bill English and several anonymous “key informants”, were with the “who” and the “how”.
“Teachers, police, and nurses all had quite different tools in their head for determining whether a child was an abuse victim, and that took 12 months to sort out,” English told the paper. “They’re all talking about child abuse and we all thought they knew and we knew what we were talking about. But it turned out, we didn’t know.”
Government agencies also sidelined the Social Investment Agency because it was “underpowered” and couldn’t impose itself, an informant said.
Others were guarded about sharing data because they thought they might lose autonomy and future funding, a nod to the aforementioned problem of being too siloed. You might identify who’s in need, English said, but then “it’s just two years of chaos while agencies fight”, leaving the Government in a bind over what to do with their funding.
“The hardest job is not actually solution-finding. Those emerge,” English said. “It’s changing existing institutions to accommodate those solutions.”
Shifting deeply embedded culture in government agencies won’t be easy, Willis concedes. She’s trying to put the right “architecture” in place including, for example, making sure the Social Investment Agency is “powered up, big time” and placing it in the middle of all central agencies.
“I want it to be a magnet so that if you’re a minister dealing with a particular issue, you want to go to the agency because it’s known as an excellent place, with excellent people who will help you look at the spending, analyse it, and come up with better options.”
She also wants agencies to devolve services to communities that know their people best. That includes a “by Māori, for Māori” approach that Luxon has repeatedly said he endorses.
“We need to be far more prepared to enable community organisations and NGOs to deliver,” Willis says.
Supercharging the good stuff
The best example of this is Whānau Ora, Maria English says, which can break down siloes and build trust with those who - due to difficult experiences with police or Oranga Tamariki - are suspicious of the state.
“It’s tailored and flexible support that’s at scale,” English says. “A family can turn up and say, ‘For us this week, the priority is paying the grocery bill’. Unless you have that kind of flexibility, services can’t tend to be useful in the context the family is actually in.”
These were exactly the essential components of Whānau Ora identified in a 2019 review: flexible and whānau-centered, enabling families to be “the“architects of their solutions”, as well as a high level of support.
“In all areas we visited and across all monitoring reports we reviewed, we have seen whānau progress towards achieving their self-identified priorities,” the review said. ”However, the approach is relatively new, and we believe it is too early to form a view as to whether or not that positive change will be enduring.”
In a nod to where Social Investment 2.0 could make immediate inroads, the review said demand for Whānau Ora - “overwhelming” in some areas - exceeds available funding and resource.
Level of support - or lack thereof - is another key part of the “how”, and Willis says it’s not just about more resources, but what happens with the existing pot.
“I’m so hugely frustrated that we already spend $70b each year on social service provision. I just do not believe that we are driving maximum value from every dollar we already invest.”
There will also be social bonds, with the first one aiming to provide secure housing for those stuck in emergency housing.
Willis wants a Social Investment Fund to be set up by Budget 2025, and for all Kiwis to be able to invest and potentially make a profit, if programmes succeed. Pre-election, National said it would start with a $50 million housing bond, but Willis won’t say if that’s still the plan, citing ongoing Budget negotiations.
She’s agnostic about whether it might lead to more privatisation of social services, or smaller government, or both. The vital metric, she says, will be whether it’s working.
There’s “little evidence” social bonds are any better than traditional programmes overseas, according to studies. Of the two New Zealand pilots under Social Investment 1.0, the first - finding jobs for those with mental health issues - “ended prematurely in October 2018 due to low rates of participant referrals”, academic research showed.
The second - where investors saw maximum returns - aimed to reduce youth reoffending in South Auckland. A review said it had a “lifetime social value of around $9″ for every $1 spent, while reoffending rates were 30 per cent lower than a control group.
Participation was stifled, however, due to the change of government in 2017 and Labour’s “ministerial suspicion of social bonds”.
Apolitical thinking
Taking a long-term view is another vital part of the “how”, even if it goes against political instincts.
Willis uses an example of refusing to invest in early psychiatric support for children. “A short-term political approach would say, ‘Oh, well, we can invest in it, but who cares? We’re not going to see a reduction in ram raids next week.’ But we have to ask, ‘What are the interventions, if we take them now, that will actually make a difference over a lifetime?’”
Short-term progress should still be measured, she says, which is crucial so changes can be made to improve support.
“You’ll probably find they’re less disruptive in the classroom, or there’s a huge amount of less stress in the family, or they’re far less likely to be involved in incidents that create harm for others. There will be ways of measuring impact immediately, even if the longer-term goal won’t be able to be measured for 10 to 15 years.”
Even if everything falls into place as Willis wants, it won’t necessarily mean Social Investment 2.0 will shift the dial where successive governments have fallen short.
But she is certain about one point: the status quo isn’t working.
“You talk to any New Zealander from the left, from the right, from an immigrant community, from the Māori community, old or young, and they’ll say to you: ‘We fail too many New Zealanders’. There are too many who, despite our good intentions, end up in terrible cycles of dependence and dysfunction.”
Time will tell if this Government will be any different.
Derek Cheng is a senior journalist who started at the Herald in 2004. He has worked several stints in the press gallery team and is a former deputy political editor.