The latest Administration tweak to the landmark healthcare law – an additional year’s delay to the employer mandate on businesses with between 50 and 100 full-time-equivalent employees – has predictably excited Republican opponents of the law. Coming on the heels of the CBO report that the law would reduce overall hours worked  by the equivalent of 2 million full-time positions, the charge that the law is a “job-killer” has got some wind in its sails, fairly or no. But what these events really show is how weak the ambient employment environment remains, and the cost of same for policymaking.

The purpose of the employer mandate in the first place, apart from raising revenue necessary to make the bill deficit-neutral, was to counteract the disincentive for employers to not provide health insurance to new hires, or to drop health insurance entirely where the benefit they provided previously was no longer compliant with the minimum coverage standards. The goal, in other words, was to reduce the disruption associated with the introduction of the ACA, reduce the number of employees who would be shunted from employer coverage onto individual market via the exchanges.

In a tight labor market, this regulation would be more likely to work as planned, but would also be relatively less-necessary since prospective employees would be in a good position to shop for the employer offering the best overall package. In a persistently slack labor market with weak overall demand, such as we live in now, the regulation instead creates an incentive for employers to reduce costs by reducing full-time and full-time-equivalent staff. Precisely because employers are in a stronger position to bargain with employees, they are in a stronger position to bargain with the government to reduce the regulatory burden of the law. Which is what has happened.

Suppose the employer mandate were scrapped permanently? This would shift the burden of providing health insurance for a certain class of employees off of the back of business and onto the back of individuals (subject to the individual mandate) and the taxpayers (who provide the subsidies that lower-income individuals receive). Since those subsidies phase out with income, they create a high effective marginal tax rate on income, which in turn creates a disincentive to increase income (by working more hours) within a certain income range. This is one of the effects identified by the CBO. (The other effect works in precisely the opposite direction, positing that some employees have a target income; anything that makes it easier to achieve that income with fewer hours of work serves to reduce employee hours.) In other words, there’s going to be some negative effect on employment regardless of where the burden falls (though the effect would undoubtedly be more diffuse in the absence of an employer mandate, offset by the fact that reductions in employer coverage would be larger).

What the episode illustrates is how difficult it is for the government to do what those concerned about inequality want it to do: add its thumb to the side of labor in the battle with capital for relative share of the economic pie.

This new focus is apparent on both sides of the ideological divide – albeit, one may question the degree to which reform conservatives’ interest is driven by the need to compete with a similar focus on the left. On the left (and in quirky right-wing places like this magazine), there’s increased interest in substantially increasing the minimum wage. The reform conservative counter is to advocate wage subsidies, which would cost more taxpayer money directly but would impose no regulatory burden on employers, and hence no direct disincentive to employment. Both, though, are efforts fundamentally to increase the effective return to employment – rather than to increase employment directly.

Jim Antle does a fabulous job of pointing out how the contours of the debate about work limns a cultural change since the mid-1990s. Then, the heart of the debate was about welfare reform, which is to say, how to move people to work. Now the heart of the debate is over how to make work pay adequately, both to combat inequality and, frankly, to make room for other obligations and pursuits.

Antle focuses on the cultural implications of the shift, but I would argue that a key reason for the shift is economic. Whereas in the 1990s, employment growth was robust but a segment of a population was left out of the boom, focusing on removing disincentives to employment made sense, along with finding ways to ease the transition (it’s worth recalling that when welfare reform was pursued most seriously, as for, example, in Tommy Thompson’s Wisconsin, real money was spent on helping welfare recipients transition to work). The goal was to leverage a strengthening labor market and ensure that a rising tide really did lift all boats. Now, there is no strong labor market to leverage, and policymakers are trying to adapt to and address the costs of that reality.

But slack labor markets themselves pose real risks to both approaches at raising effective wages. In tight labor markets, an increase in the minimum wage would be less necessary, but it would also be more effective in producing an incentive to invest in training and equipment to achieve higher labor productivity, which redounds to the benefit of all. In a slack labor market with weak demand, a higher minimum wage creates incentives to evade the minimum, evade an increase in costs that would make the business less-competitive, whether by reducing hours per employee, or cutting back on marginal lines of business, or offshoring, or creating categories of employee who are not subject to minimum wage rules (interns, freelancers, piece-rate workers, tip- or commission-based employees, off-the-books employees, etc.). Once average wages are up, yes, employers would respond to an expected increase in demand by increasing investment. But if they don’t forecast that increase ab initio, then the bootstrapping desired by the advocates of a higher minimum wage may never get going.

A wage subsidy approach doesn’t burden employers. Instead it burdens taxpayers – and employees. Why employees? Because the subsidies phase out, they effectively become a high marginal tax rate on wage income at the low end of the scale. Employers would be cognizant of that fact, and it would affect the wages they would be willing to offer. Again, the effect would be different in slack markets versus in tight ones. In tight labor markets, employers couldn’t afford to put a ceiling on wages because of the fear of losing out in a competition for available employees. A wage subsidy would therefore largely be captured by employees themselves. In slack labor markets with weak demand, however, the subsidy would make it possible for employers to reduce effective wages, and allow the subsidy to pick up the slack. An employee who was willing to work for $10/hour before a subsidy went into effect would, if jobs are scarce, certainly still be willing to work for $10/hour afterwards, even if now $9 came from the employer and $1 from the government. Thus there’s the possibility that wage subsidies would be largely captured by employers rather than employees. The subsidy might not increase effective wages, and might create an effective ceiling on income because of the high marginal tax rates.

None of this is intended to be an argument against the ACA, which I largely favored at the time and still do. It’s an argument that the perverse incentives created by redistribution schemes are more intractable in a weak labor market such as we have today than they are in a strong one. Those redistribution schemes may be worthwhile for other reasons nonetheless – for example, because they improve health outcomes overall, or alleviate specific social ills. All of which means we should focus more on how to strengthen that labor market (whether by improving the long-term outlook for real growth or the short-term outlook for nominal growth; the approaches are not mutually-exclusive) than on trying to perfectly optimize redistribution schemes undertaken for other purposes.