Simon Chapple has been studying the rhetoric and policy implementation of what Bill English calls ‘the social investment approach’. Expecting social spending in Budget 2017 to be framed with this concept, Julienne Molineaux interviewed Simon about the term.

A short version of this interview was published on The Spinoff website:

Julienne: ‘Social investment’ is a term Bill English uses to describe his approach to social spending. What does it mean exactly?

Simon: It is important to distinguish between the rhetoric – social investment is a lovely soft attractive phrase across the political spectrum – and what is actually being done under this banner. I believe that the political focus is reducing the long-term size of government.

To understand this we need to go back in time. Bill English entered parliament in 1990. The newly elected National government, of which he was part, cut government spending through the ‘mother of all budgets’ in 1991. Reducing the size of government is a standard centre-right prescription, but that particular slash-and-burn exercise burnt a lot of political capital. The hope was a very quick bounce back and a rapid political dividend. But the bounce back didn’t happen and the policies were unpopular with the public. In part, the cuts contributed to the introduction of MMP, a system giving less power in the hands of any government for radical policy change.

Bill English learned from that episode. In a sense the ‘social investment approach’ is pursuing the same goal – smaller government – but in a more subtle way. English’s claim is that social investment is a win-win by reducing the size of government while simultaneously enhancing people’s outcomes. However, when the nuts and bolts are examined, the principal win that government is measuring and incentivising is fewer fiscal dollars, not the secondary win which is better outcomes for people.

The other contextual fiscal issue doubtless on a centre-right mind is the issue of long term rising fiscal costs as the population is ageing. Ageing means increased superannuation payments, and also rising health spends. Health spends also have an independent upward momentum due to technological changes in health care. The challenge for a fiscal conservative is how to fit these rising costs within a stable taxation envelope – what spending can we reduce, so that we don’t have to raise taxes?

Correctly and positively, in the past Bill English and Paula Bennett have both talked about social investment as being centrally about improving well-being outcomes for disadvantaged New Zealanders. But look at the process of implementing social investment. Look at where the work has gone on. Look at all the incentives in place. They’re all about reducing fiscal spending over time and assuming – assuming, not testing – that the consequences of that reduction in spending over time is better well-being outcomes for disadvantaged New Zealanders. Now if the National coalition genuinely and centrally wanted better outcomes for the disadvantaged – which is their rhetoric – I think that they would be ruthlessly focussed on examining changes in disadvantaged people’s outcomes arising out of their social investment approach. But the only metric government show any serious interest in, and the main incentives in place in terms of agency behaviour, are solely to do with fiscal outcomes.

What are the key features of the ‘social investment approach’?

It is moving and evolving. We’ll get a better idea of this with the 2017 budget.

That said, the key unifying feature is managing and incentivising the welfare system in terms of reducing the future fiscal liability – that is, fiscal spending on people on the government books today and into the future – within tightening rules of entitlement and surveillance.

This approach began with the 2011 Welfare Working Group. They were was instructed by government to look at lessons to be learned from the private insurance system for how to operate welfare. Private insurers insure people against adverse events, and this creates a liability that extends into the future if they have to pay out on those events. But insurers also have a corresponding asset, which is income paid to them now and into the future as people’s insurance premiums. So the private insurance sector makes decisions to maximise the difference between their expected future assets and future liabilities.

The government is using this insurance analogy to make decisions on how well the public service is doing and about what to invest in. But how good is the analogy? The answer is not very good.

Obviously the government is not a private insurer. There isn’t a market where people can choose, say, their welfare cover and take their money elsewhere if they find a better deal. The government is only observing one component of how insurance works: the future fiscal liability. But if the system is solely based around liabilities, why do we even have a welfare system? What is our asset and how do we value it? How do we build up the asset as much as reduce the liability? Lastly, what effect does it have on people needing welfare and the way they are treated by officials if they are conceptualised as a liability?

What are some examples of problems arising from having such a narrow focus?

There are many – here’s three.

If they are fulfilling their duties under the Social Security Act, Work and Income should ensure that everyone entitled to a welfare benefit gets one. There is good evidence from the OECD that benefit non-take-up is a serious issue internationally, including New Zealand.

Secondly, there are likely to be many working people who are entitled to but don’t get the Accommodation Supplement. There are also an unknown number of people who have been made redundant who appear eligible for a benefit but are not receiving one. Yet there is a strong disincentive under the investment approach to ensure a good take-up performance, because Work and Income is tasked only with reducing welfare payments.

A third example is that there is evidence that beneficiaries are increasingly leaving benefits for tertiary study (see Statistics New Zealand, Table 2.1 here ). Now, if a beneficiary goes off welfare and onto student assistance, they build up a student debt. Effectively income support is privatised, as some student assistance (in the form of a student loan) has to be paid back at the end of study. Even if the education is a complete failure, so no private asset is created, the person is off welfare and out of the liability calculations for three years. For Work and Income, this is an unambiguous win, and they have an incentive to push people into tertiary education. But for the student/beneficiary, who may now have a large debt that they can struggle to repay, they may be worse off. Work and Income bear no consequences of this investment error.

Are government agencies being given a reward or incentive to reduce spending or the number of people on their books?

Yes, absolutely. If you look at the government’s relevant Better Public Services goal, the performance target is about reducing the fiscal liability of the welfare system and the number of beneficiaries. The target is not about getting people into jobs and ensuring that their social outcomes are better.

Two points. Firstly, leaving a benefit is not the same as getting a job. Secondly, not all jobs are created equal; some jobs are better for people’s lives than others: think a 9am to 5pm job compared to a 9pm to 5am job. Yet the social investment approach assumes that benefit exit means that people move into jobs and it presumes that all jobs are quality ones. I think both of these are highly challengeable assumptions and should be empirically tested. Already, we know overall that many benefit exits are not because people get work.

Well-designed social spending always had an investment element – for example, investing in maternity care so children get a good start in life. How does this approach differ from what has been done in the past?

In the past, rhetoric about investment has mostly lacked data to examine its truth-content. The maternity care example is intuitively plausible, but it may or may not be true in empirical practice. We haven’t been very good at quantifying longer term outcomes in a sufficient number of our decision-making processes regarding spending. To that extent that we can now quantify long-term outcomes, we should be able to make higher quality investment decisions, better informed by evidence.

In terms of commenting on the current approach, we know that most of the gains from social investments – if we do them well – are to disadvantaged people themselves. So while there may be both future fiscal gains and gains to people from better maternity care today, the government’s social investment model as it currently exists, if it were applied to maternity care, would only count and reward gains to taxpayers, not mothers and babies.

Social investment includes a highly-targeted approach to social spending, relying on data to pinpoint both recipients of spending and the interventions or programmes used, against a goal or outcome measure. This can be contrasted with a more universal approach to social services, which was the more traditional approach to social spending. What are the pros and cons of this shift to targeted interventions?

The classic problem with targeting is that you have false positives and false negatives. The more you target, the more likely you are to miss out on people who need the service – and you won’t completely solve the problem of providing the service to people who don’t need it.

If we’re looking at vulnerable children, whatever means we have currently for identifying vulnerable children, we’re going to miss a lot of them in the targeting process. In addition there will be some children we identify as vulnerable who won’t be and there may well be adverse consequences of this – for them and for their families. There are ethical issues about access to information needed to target well. Lastly there may be perverse effects on potential clients if they are required to divulge information. Trust between the client and the provider could be undermined, to the detriment of the client’s outcomes.

The other negative aspect of targeting, compared to a universal model, is that middle-class’s voice and commitment to the system can be lost - the sense of all of us as citizens being in this waka together. The lack of an articulate middle-class voice keeps service quality low. Given higher degrees of middle-class political engagement, tight targeting can also make it politically easier to gradually drain needed resources out of the system. Absent the middle-classes, the system is also more likely to become a source of stigma for the targeted. Stigma can reduce non-compulsory but beneficial service take-up.

But aren’t there false positives and false negatives now. For example, leaving judgements about vulnerable children up to social workers, they don’t always get it right as it is. This way might be more systematic and learning takes place from the evaluations?

Yes. False negatives are important in the area of vulnerable children. In the early 2000s a guy called Mike Doolan, who had been the Chief Social Worker, did some work for the University of Canterbury on child deaths in New Zealand. One of his questions was, “How many of those kids were not known at all to the authorities?” – effectively a false negatives question. Almost certainly all those kids killed had a hard life – albeit often very short – leading up to their death – but four in five were unknown to the system. So even if we had perfect government information in terms of children in the system who were going to be killed, we’d only identify 20% of those kids who had died. One would hope that we should be able to do better, but it is – and will remain – very tough to design effective targeting systems in this area.

Thinking about existing programmes under the social investment banner, to what extent are the measures of success aligned with the outcomes being sought?

If you look at the papers that have gone before Cabinet and comments from Ministers on social investment – we’ve had them saying it’s all about better outcomes for New Zealanders and that the social and economic benefits from it are much bigger than the fiscal benefits. But if you then look, for example, at what the Cabinet papers have said about the set-up of the Social Investment Unit, the over-riding emphasis is on how do we connect the fiscal dots across different areas of justice, health, education and welfare. The Unit has been tasked with ensuring that these different agencies coordinate and measure the fiscal liability. The Unit has not been tasked with measuring what happens to disadvantaged people’s well-being when policy changes, an approach which is greatly at odds with the headline rhetoric. Either there’s some massive collective cognitive block common to politicians and the public services causing them to act illogically, or it is deliberate politics.

Think of the government system as a hospital. People come into the hospital because they’ve got a problem. The liability approach is basically about counting the number of days the current occupants are likely to spend in a hospital bed, and assessing gains in terms of managing down this magnitude. However, we want – or should want – to know that when we discourage people from coming into, and when we discharge people from, hospital that their outside-hospital outcomes are better than their-inside hospital outcomes. No doctor would discharge someone from hospital until they knew that where they are going to was a better place. We want to know when we discharge them that their outcomes are better. Yet this new approach says, “Who cares where they go when we discharge them? We just want them out of hospital.” And I consider that highly unethical. No doctor would be able to get away with it.

The view in government is that the evaluation of outcomes is all in the change in fiscal liability. If liability is moving down, then the government is onto a good thing. End of story. For me it is amazing. The post-2010 welfare reforms have been the biggest changes to the welfare system since the early 1990s. Bill English talks about evidence-based policy, which is great, but hang on a minute – his government is uninterested in creating and examining any evidence on whether recent welfare changes have led to better outcomes for the people concerned. Worrying.

So in the hands of a different government we could do lots of data-driven things, lots of targeting, but if the focus was on outcomes for the recipients, this approach could lead to more spending.

In theory, yes it might. If the quality of information was higher, one might make better quality decisions, and it may be necessary to invest more than we are doing today. Again, this question of efficient investment is an empirical matter, not one primarily of political ideology.

What are the positives?

Thinking long-term is good. But the idea that what we do today matters for tomorrow – this is something we’ve known for a long time! The idea that we should use joined up data to evaluate programmes or connect across government agencies – again this is good, but not a new idea.

Unfortunately we’ve spent an enormous amount of money – probably tens of millions of dollars – on actuarial services to measure fiscal liability. That is money better spent in measuring the things that Ministers say are truly important: the outcomes of people who have been affected by social investment-driven policy change. I find it pretty disappointing that while using a people-focussed rhetoric, we continue spending money on expensive actuaries to measure fiscal outcomes.

Is there any international precedent for this? Is New Zealand being a world leader here?

The short answer is yes, we are world leaders! We are the first nation state to create a fiscal liability measure for the welfare system. Do we want to be world leaders in this? No. I regard fiscal liability measures as a waste of money, time and resources. It tells us little that we didn’t know before and creates a bunch of perverse incentives as we try to improve people’s lives. I sincerely hope we’ve passed peak actuary!

Why should we care about things like this?

Rhetorically, we have a warm and fuzzy phrase – ‘social investment’ – which is obscuring standard centre-right political goals. When people vote they should be informed on what the approach is really about, rather than what Ministers say what it is about. If they like that political goal, people can vote for it; if they don’t, then don’t.

We should also care because social investment which only values fiscal wins can go badly wrong – for others and for ourselves. The system creates a bunch of perverse incentives which can negatively impact on people’s lives. Over our lifetime, nearly half of us will be on a working age benefit. So how effectively social investment works, even just in the welfare system, is not about a small minority of Kiwis.