When unemployment drops to low levels, two instincts seem to kick in among commentators and and politicians. One is to demand industrial relations reform, out of fear that high demand for workers might drive wages up — a normal market response, but anathema to many people who are normally big fans of free markets.

The other is welfare reform, designed to ensure no one who can work chooses not to.

It’s been a fixation of policymakers for centuries. Early modern European civic leaders railed at “sturdy beggars”. The Dutch even devised a primitive form of “work for the dole” — cells that would fill with water unless the sturdy beggar, who would otherwise drown, pumped it out, thereby inculcating the habit of physical labour. It’s said they abandoned it after one participant, perhaps unconvinced by the philosophy of reciprocal obligation, preferred to drown rather than pump.

The Disability Support Pension is a particular fixation of policymakers, partly because it has been expanding rapidly for more than two decades; but also, one feels, because, when it comes to welfare clichés, someone fraudulently claiming disability is considered even worse than the stereotypical dole bludger or single mother.

The Howard Government had a go at disability reform in 2006, with some big bang changes meant to end the expansion of the DSP. They had little long-term impact. This year the DSP will cost taxpayers over $13b. And in fact the DSP itself was the product of reforms by the Hawke Government in 1991, which significantly tightened eligibility and removed non-medical factors from assessments for claimants. They were similarly unsuccessful in curbing growth.

In between, commentators and economists have lamented its remorseless rise as evidence of a particularly Australian malaise of workshyness.

In early February, the prime minister announced further reforms aimed at increasing workforce participation, including reforms to the DSP. The coming budget is likely to be loaded with welfare reforms. Today, Tony Abbott has joined in, urging reforms to the DSP and to further “incentivise” the unemployed to join the workforce. Just over a year ago, Abbott was the victim of an internal leak about his plans for a package of get-tough welfare measures, including making one-third of disability support recipients sit an annual assessment, designed to force 24,000 DSP recipients back into work. The annual assessments would have cost nearly $700m. A number of the measures didn’t make the Coalition’s election platform.

The policy imperative is correct. Even if you don’t buy the link between self-worth and employment, we need all the workers we can and we need to overhaul the tax and transfer payments system to maximise participation.

Only problem is, as the potted history I’ve just given suggests, when it comes to the DSP, no government has yet devised a way to address the problem.

That’s mostly because, as Peter Martin pointed out in an excellent piece on Ms Gillard’s ambitions, the relentless expansion of the DSP is the perverse product of a miserly attitude toward unemployment benefits — these days known as Newstart. Newstart is now so low that there’s a strong incentive to switch to the DSP, which is indexed at a higher rate.

Ken Henry’s tax review pointed this out. “…Decisions to target payment increases to particular groups has increased incentives for some people to remain on and for others to seek to qualify for higher-rate, non-activity tested payments, such as Disability Support Pension.”

These decisions are, the report noted elsewhere, inherently political. “Categorical distinctions — such as single parenthood, disability or unemployment — assist in targeting support to those with varying need and capacity to support themselves. Such distinctions also give effect to various social judgements about who should receive assistance.”

The report went on to explore, in great detail, how such “categorical distinctions” combine to produce all sorts of skewed outcomes that defy policymakers’ intent.

As IR expert Prof Mark Wooden pointed out at the time in response to the prime minister’s statements, the only really effective means of addressing the increase in the DSP is by cutting it, so there’s less incentive to switch from Newstart, and that’s unlikely to be very popular. Tony Abbott’s idea to further tighten eligibility will be no more successful than previous “tightenings”.

There’s another context for the sudden interest of our politicians in welfare reform.

The DSP will cost over $13b this year, and in the last budget was forecast to rise to over $14b in 2013-14.

But Family Tax Benefit A will also cost over $13b this year. The government actually expanded Family Tax Benefit A as part of its election pitch, via the education rebate. Family Tax Benefit B will cost $4.5b. The baby bonus will cost just under $1b, even after eligibility for it was tightened up. And the private health insurance rebate will cost over $3.5b even after it, too, was tightened up. The savings from overhauling middle-class welfare will be considerably larger than savings from tightening the DSP yet again — even if a way could be found of actually curbing its growth.

It’s odd that “welfare reform” never seems to focus on middle-class welfare. It’s all about “incentivising” the unemployed and getting people with disabilities into the workforce, and never about ending payments going to households earning over $100,000 a year. Welfare for the unemployed and the disabled is the subject of rigorous focus on whether it is achieving desired policy outcomes.

Middle-class welfare doesn’t even have a policy outcome, beyond fostering the great Australian sense of entitlement.