My last post consisted of the sort of Sisyphean snark about The Australian that I’d like to cut back on, but can’t resist writing. I was a little taken aback that the same paper that labelled a modest trim to family payments for high-income households as ‘class warfare’  would unashamedly lament that “the old principle that welfare should exist only for those who genuinely need it appears no longer to hold.”

Beyond the exasperation and incredulity, though, there’s a serious point to be made. Designing welfare policy involves a range of unavoidable trade-offs between worthy principles. Three central principles in the Australian welfare system are:

  • Payments should be sufficient to protect people from poverty.
  • Payments should be means tested and targeted to people on low incomes.
  • ‘Poverty traps’ should be avoided. It should be financially worthwhile for people to take a job, or to increase their hours of work.

These three principles are actually discussed in the editorial that I excerpted. The Australian implicitly acknowledges that Newstart is too low to protect people from poverty (“lifting the unemployment benefit, which is low by international standards, may be part of any comprehensive reform package”). It also calls for stricter means testing (“the old principle that welfare should exist only for those who genuinely need it appears no longer to hold”) and for policymakers to avoid creating poverty traps (“welfare reform must not reduce the incentive to work”). The problem is that the editorial writer makes the common mistake of believing that these three principles can all be rolled together in some tension-free way. They can’t.

Pretend, for a moment, that we have a system in which the full-time minimum wage is $600 per week, and the unemployment benefit is $250 per week. If you earn a dollar from working, then you lose 50c of your benefit. This means that the payment is phased out entirely by the time your earnings reach $500 a week. We’ll call that scenario A.

Suppose that policymakers then increase the unemployment benefit from $250 to $300 per week. $250, it’s believed, is not sufficient to achieve the bedrock goal of protecting people from poverty. As a result of the $50 increase, a whole new range of people will now be eligible for the payment. The benefit is still phased out at the rate of 50c in the dollar, but the higher benefit means that your earnings have to be a bit higher before you lose the payment entirely. The unemployment benefit would now phase out at $600 per week, so you’d have to be working full time before you ceased to be eligible for the payment. Call this scenario B.

Some people might not like scenario B, because it brings a whole lot more people into the welfare system. It might be seen to violate what The Australian called “the old principle that welfare should exist only for those who genuinely need it”. To avoid that eventuality, you could increase the payment by $50, but also increase the rate at which the payment is withdrawn as a person’s earnings rise. Rather than losing 50c of unemployment benefit for every dollar of earnings, you could take away the benefit at the rate of 60c in the dollar. This would mean that the payment would still cut out completely when earnings reach $500 a week, so it would make sure that eligibility for the payment doesn’t extend further up the income scale. This is scenario C. Here’s what they look like:


Do you see the problem here? Scenario A involves an inadequate payment; it violates the ‘protection from poverty’ principle. Scenario B includes an adequate payment, but the payment is less tightly targeted to those who need it most. Scenario C tries to fix this by withdrawing the payment a bit more rapidly as your earnings rise. The problem with this is that it means that people get to keep a smaller proportion of each additional dollar that they earn – their ‘effective marginal tax rate’ rises, which can discourage work and can make it harder to make ends meet.

Scenario A Scenario B Scenario C
Adequate X
Targeted X
Encourages work X

There is no choice facing policymakers that makes the payment ‘better’ as far as one of the principles is concerned, like improving its adequacy, that doesn’t also make it ‘worse’ in terms of another principle, like targeting the payment only to the lowest income earners. This doesn’t mean that no change should be made to our payments system. Personally, I’d favour something like scenario B, because I put less weight on the ‘targeted’ principle than I do on improving adequacy and having low EMTRs. The point is that that’s a judgement based on my own preferences regarding the relative weight of the three principles; I don’t pretend that there’s a proposal on the table that would magically improve policy along all three of these dimensions.

If you want to argue that our welfare system, which already has less ‘middle class welfare’ than any other in the OECD, should be even more tightly targeted to low income earners then that’s fine. Just don’t pretend that by doing so you can also call for an adequate payment and one with low barriers to work. Pick any two out of three; you can’t have it all.

NOTE: My example scenarios are obviously simplified and stylised – there’s no income tax, the benefit is withdrawn from the first dollar of labour earnings at a constant rate, etc. The point is just to illustrate the trade-offs between the three principles, not to produce numbers that reflect current or proposed policy settings. Check out figure 29 and figure 30 in this submission if you’re interested something closer to reality.