Under and over the benefit in NZ

How should the new(ish) Government improve the New Zealand welfare system? There are some important areas where improvements should be made. Reorientating the social investment performance model is one of them.

Imagine a world where everyone entitled to a welfare benefit receives one. Then imagine that everyone receiving a benefit is also actually entitled to one. Given the rates paid to different beneficiaries, our imaginary welfare system is operating perfectly in terms of anti-poverty effectiveness and minimising wasted resources.

However, our world is not perfect. The reality is that some people eligible for a benefit do not take one up (which I describe here as “unders” below). Others are ineligible but do take up a benefit (“overs”). Both unders and overs are undesirable. They demonstrate a less than perfect Work and Income performance, as unders involve less, and overs more, welfare spending than is desirable.

Overs include, but are far from, restricted to fraudsters. Other reasons for overs include misunderstandings about benefit obligations or various administrative errors on the part of Work and Income.

In terms of the $12 billion plus we spend on working age welfare and tax credits in New Zealand, very little is known about system performance in terms of overs and unders. Surprising, really, given the vast sums involved. What little we do know, however, is consistent with international OECD country evidence suggesting the proportions of both unders and overs are substantial. Evidence shows unders range from 12 to 67 percent, and overs, generally much smaller, range from two to 20 percent. In other words, Work and Income performance errors may be large, likely in the order of billions of dollars.

The social investment approach of the previous National coalition provided a strong incentive for Work and Income to attack overs. That incentive was provided by the first of the now-defunct Better Public Services goals to “reduce long-term welfare dependence”, and in particular its two performance targets. Lowering overs reduces the number of welfare beneficiaries and also the actuarially measured forward liability – defined as the sum of the now and future benefit payments of this year’s beneficiaries. These incentives were central to the social investment performance model in the welfare system.

Unfortunately, National’s social investment approach also provided an incentive for Work and Income to increase numbers of unders. Creating more unders undermines the anti-poverty function of the welfare safety net, but reduces the forward liability. Because neither unders nor overs were separately measured by the social investment approach, the consequences of perverse incentives in the system were hidden by the old social investment performance model.

Following 2011 welfare reform, increases in the numbers of unders are a likely consequence of imposing higher surveillance and information provision costs on legitimately eligible beneficiaries. Unders are also likely to have increased because of a range of reform-related “soft-signalling” from both Work and Income and National politicians, which functioned to increase stigma associated with being on a welfare benefit.

In addition, while being notionally about finding jobs for beneficiaries, the social investment performance model did not actually directly incentivise Work and Income to find people work. Yes, benefit exit was incentivised. This incentive would be well and good if benefit exit was a good proxy for getting a job. Unfortunately it is not – leaving a benefit and getting a job are not the same thing. Fewer than one half of benefit exits are to work. This fact has been well known for decades, from Work and Income’s own data.

Additionally, even if benefit exits induced by policy choices were due to people finding work, the social investment approach treats all jobs as equally good, all other things constant. However, some jobs pay more, have acceptable commutes or have hours that suit child raising. Waiting on a benefit a bit longer until a more suitable job comes up may be a better outcome than taking up the first job that results in benefit exit. The social investment model is blind to such a rational need.

In the book Social Investment. A New Zealand Policy Experiment, published by Bridget Williams and Victoria University of Wellington’s Institute for Governance and Policy Studies and launched tomorrow, experts consider different dimensions of social investment. Based on my chapter, which looks at the welfare system, a range of suggestions can be made to improve governance.

Better arrangements for Work and Income – ones that go beyond the Better Public Services goal – require building sound empirical knowledge on the numbers of unders and overs and an associated understanding of their causes. Work and Income’s performance can then be regularly assessed in terms of its ability to minimise both unders and overs.

Furthermore, the effectiveness of Work and Income’s performance in getting people into jobs that are good for those people and their children should be directly examined and incentivised via governance structures, rather than relying on the poor proxy of adult benefit exit.

This knowledge-building can be funded by ceasing spending on the misdirected actuarial work that underpins the old social investment approach.

There is much to do. The part of government best placed to address these knowledge gaps and incorporate them into the Work and Income governance arrangements is the new Social Investment Agency. I look forward to seeing real progress in this area.

This article first appeared on Newsroom, on 14th February 2018.