• Labor Market Effects of Health Costs

    A reader asked me to explain these sentences from a recent post by Jon Cohn:

    Health care reform actually is a jobs program. A recent paper that David Cutler and Neeraj Sood published through the Center for American Progress suggested reform would generate between 250,000 and 400,000 jobs.

    The Cutler and Sood paper draws on published literature that shows that as employers’ health costs rise, employment falls. Thus, if health costs fall, greater employment should result. That makes intuitive sense since an employer cannot afford to hire as many workers when the cost per worker goes up. But, as the reader pointed out, this doesn’t seem to make sense if the effect of higher health costs is translated into lower wages, as economists generally believe. How can a dollar increase in health costs be accounted for entirely by wages and there still be anything left to affect employment? Good question!

    As with many things, there are nuances not revealed by high-level summaries. Let’s dig deeper. With a nod to one of my favorite Krugman rants I’ll note that one need go only as far as the Cutler and Sood paper itself for an answer.

    But the wage offset is not dollar-for-dollar for all workers. Firms have little ability to reduce wages for workers at or near the minimum wage or for workers with fixed employment contracts. Rising health insurance premiums will thus lead to more job losses among these types of workers while falling premiums will increase employment. Similarly, not all workers value employer-provided health insurance at its cost—either because their overall income is low or because they have health insurance from another source (perhaps a spouse). For these workers, the lower wages that rising health insurance premiums necessitate induce them to leave the labor force or move into part-time jobs (with no health benefits).

    Unpacking that a bit, it means that higher health costs do translate into lower wages when what is provided at those higher costs is of value to workers and wages are not already at the floor (minimum wage). If employers cannot reduce the wage (due to the floor) then jobs are cut. Or when what is provided at a higher premium is not highly valued, workers do not accept the compensating lower wage and, instead, lower employment results. That is, sometimes the wage required to offset the increase in health costs is too low.

    In a 2006 article in the Journal of Labor Economics titled The Labor Market Effects of Rising Health Insurance Premiums, Katherine Baicker and Amitabh Chandra make essentially the same point.

    In contrast to Gruber’s study and to the results in Gruber and Krueger (1991), we find effects on both hours and employment. These results may appear to be contradictory, but they are not: in Gruber’s study workers receive new maternity benefits, and in Gruber and Krueger they receive more generous workers compensation; both benefits are probably valued by workers, and the empirical finding of the full shifting of increased costs to wages with no effect on overall employment is consistent with the insights of Summers (1989). In our article, however, the increase in the price of health insurance premiums is driven by the medical malpractice crisis, a change that may not enhance the value of health benefits. It is therefore unsurprising that workers do not value this increase in costs as highly and that the labor market responds with decreased wages and labor utilization.

    I suppose one can argue whether or not reduced employment is a wage effect. Economists would make a distinction between wage and employment. But unemployment is a pretty severe wage cut. So, I still think it is fair to say, in broad summary, that workers absorb increases in health care costs through wage reductions. But if one is being precise, one ought to say workers pay the price either in lower wages or in loss of work. Either way, higher health costs are shifted to workers, even when the employer ostensibly pays the premium.

    Share
    Comments closed
     
    • One of my problems with this line of analysis is the partial equilibrium nature of it. What about the fact that health expenditures themselves result in hiring home health aides, nurses, physical therapists etc., these are shifts in the labor market and cause other dislocation effects.

      Remember we are on a track toward spending a 1/3 of our economy on health care if we don’t change the track. That would mean (since health care as a service is high on labor costs) that perhaps more than a 1/3 of us would be working in the health care sector. And how that ends up for gaining/losing jobs is unclear from the Cutler/Sood analysis since it is not a general equilibrium model.

      Jim

      • @Jim Burgess – It is so interesting that you raise that point. When I first saw the headlines about health care and the labor force I thought the point was that expanding insurance would increase demand for health care labor. It wasn’t until a reader brought up the question I answered that I actually looked at the paper to see that they were making a connection to worker compensation, which, really, has got to be a secondary or tertiary effect.

        However, how many micro papers are general equilibrium models anyway? (Actually I’m not sure if a paper like this qualifies as micro or macro…hardly matters.) I wouldn’t criticize a paper too strongly for exploring a partial equilibrium because, ahem, I do it all the time. However, the authors could at least acknowledge that fact. A very minor demerit in my mind. Reasonable people can disagree.