Working for Families as employer subsidy?

Matthew Hooton critiques WFF in the Herald:

In fact in 2004, the left-wing critique of Working for Families was stronger than Key's, that it would operate as a subsidy of low-paying employers.

That is, using Key's original numbers, if there was a job to do worth $60,000 a year, an employer could hire someone with two kids, pay them just $38,000 a year, and they'd end up with almost the same pay in the hand.

Union bosses rightly feared it would be difficult to get workers with children to sign up for a pay campaign if it made little difference whether they earned $38,000 or $60,0000 a year.

Worse, if Government subsidises something, there will be more of it, in this case low-paid jobs. To an employer, Working for Families screams out: "Don't buy more plant and machinery or invest in on-job training, just hire a few more low-skilled labour units and get the government to pick up a big hunk of the tab."

There is very little doubt Working for Families has led to lower productivity and wages across the economy than had Clark not launched it as her big 2004 Budget bribe to fend off Don Brash's Orewa-speech challenge.

That is bad for everyone but the most pernicious effects of Clark's bribe are on those without children trying to save for a first home, such as young nurses, teachers, doctors, and police officers.

They suffer from the economy-wide lower wages caused by Working for Families but without the top ups.

Let's think this through. Labour markets are generally competitive - employers have to compete for workers, and workers have to compete for jobs.

The incidence of the wage subsidy through Working for Families, like the incidence of any other tax or subsidy, will wind up depending on relative elasticities. The relatively inelastic side of the market draws the larger portion of the burden of a tax, or the benefit of a subsidy.

If the demand for labour were perfectly inelastic (and labour supply were other than perfectly inelastic), a government wage subsidy would only benefit employers. They would employ the same amount of labour at a lower cost (to them), with workers taking home the exact same amount as before. The supply of labour depends on the after-tax-and-transfer salary on offer, and take-home pay would not change.

But labour demand is hardly perfectly inelastic. And labour supply can be pretty inelastic. I haven't the labour demand estimates to hand [anyone who knows them off the top, point me at them!], but here's Creedy and Mok's estimates on labour supply elasticity:

Elasticities on labour supply range from very low for married men, to less inelastic for sole parents. Any number below 1 is inelastic, so all of this shows ranges of inelasticity. Is it then particularly likely that the main thing going on in WFF is a transfer to employers?

A few bottom lines:

• Unless one side of the market is perfectly inelastic, both sides of a market wind up sharing the burden of a tax or the benefit of a subsidy. The more inelastic side gets the larger share of either. Labour supply is reasonably inelastic. I haven't labour demand elasticities to hand, but it would be unlikely for labour not to be benefiting substantially from the transfer. 
• At least some of the point of Working for Families was to increase the benefit of being in work relative to being on benefit for people like single parents. The simulation models have it having increased sole parent labour supply, but reduced labour supply among married women with children - the latter due to the combination of income effects and high EMTRs.
• There are big problems yet with the Effective Marginal Tax Rates in some of the abatement ranges [which I had wished the Tax Working group had had a chance to look at]. But how else you run a programme to provide an in-work cash transfer to those with kids to build a gap between earnings in work and earnings on benefit? There are variants of Milligan's trilemma that are going to apply here. Like, pick two of {not incredibly costly; reasonable wedge; low abatement rates}. If we want lower abatement rates, either benefits have to go down or the overall cost of the programme has to go up. And if the overall cost of the programme goes up, then we're trading off a reduction in a very high EMTR affecting a small group (which can often be hurdled by working more hours) with an increase in EMTRs for a very broad group.
• Wage subsidies place the burden of supporting the employment of those with lower productivity on the tax base in general. Minimum wages place that burden on those who would employ those workers, their customers, and on workers disemployed in the process. The former seems to me both more efficient and more equitable.