• The premium-wage tradeoff in the private sector

    The following is cross-posted at the HealthCare Markets and Regulation (HMR) Lab, Department of Health Care Policy, Harvard Medical School and is by Jeannie Biniek, a PhD candidate in the economics track of the Health Policy Program at Harvard University. From 2009 to 2013 she worked as professional staff for the Senate Budget Committee where she advised Chairmen Murray (D-WA) and Conrad (D-ND) on health, tax, and economic policy issues. Prior to that she worked as a consultant to companies in the health care industry on transfer pricing matters and intellectual property litigation. She holds a BA from the University of California, Los Angeles and a MA from Johns Hopkins University.

    In two previous posts (here and here) I discussed the economic theory behind the premium-wage tradeoff, as well as the literature examining this tradeoff in the public sector. In this post I turn to studies in the private sector.

    In teasing out the relationship between premiums and wages, it is important for researchers to address selection. That is, the tendency of certain workers to systematically differ in unobservable ways that lead to both higher wages and higher health insurance costs. For example, more productive workers may both be paid a higher wage and have a preference for getting more compensation in the form of health insurance (perhaps because of the tax-preferred treatment of employer-provided coverage). One way studies have addressed selection is by isolating groups with different preferences who can be identified using the following factors: 1) the specific services covered, and 2) the availability of other options for coverage.

    First, there is evidence that certain groups who have higher expected medical costs appear to be offered and accept lower wages when an employer also provides health insurance covering medical services the group expects to use.

    For example, Gruber looks at the introduction of requirements for insurance to cover costs associated with childbirth. Women of childbearing age, and their spouses, would be expected to differentially benefit from this requirement and Gruber finds that associated costs are in fact shifted onto this group after state and federal laws mandating coverage of maternity benefits went into effect.

    Similarly, Bhattacharya and Bundorf find the annual wages for obese individuals covered by employer-sponsored plans are lower (compared to obese individuals in jobs without employer coverage) by an amount that approximately reflects the level of higher expected spending for this group.

    While the studies above look at a well-defined set of workers with specific medical needs, the results hold in some more general circumstances as well. In particular, when individuals are newly covered by an average cost health insurance plan, they accept correspondingly lower wages.

    For example, Kolstad and Kowalski find that individuals who gained health insurance at firms that introduced coverage in Massachusetts after the state implemented an employer mandate earned just over $6,000 less a year; this compares to the approximately $6,100 annually it cost employers on average to provide coverage (an almost exact 1:1 wage-premium tradeoff).

    However, in the context of rising health insurance costs (rather than new or existing health insurance) researchers have not always found a 1:1 tradeoff. Looking broadly at employers who offer health insurance, Anand finds, for example, that firms reduce total compensation (wages less the employee share of health insurance premiums) by 52 cents for each dollar increase in health insurance premiums — the incremental cost is not fully passed on to workers. One explanation for this finding is the additional costs associated with rising health care expenditures did not lead to commensurate benefits valued equally by all the workers in a firm. While this study did not attempt to identify the drivers of increasing costs, for the sake of example, if we assume it was higher provider prices, then the result suggests that workers do not value the higher provider prices at its cost, so aren’t willing to trade a dollar of it for a full dollar reduction in wages.

    A second important consideration is the relative value of employer-provided coverage compared to a worker’s outside options for health insurance.

    Specifically, individuals who do not have the option to obtain health insurance through a spouse may be more willing to tradeoff wages for coverage. Olson uses husband’s union status, firm size, and own-name employer coverage to identify a group of married women unlikely to have the option to get coverage from their spouses. He finds that married women with their own employer-provided coverage are paid wages lower by between $3,500 and $5,200 annually. This compares to the approximately $4,500 it was estimated to cost employers to provide coverage in the same year, or about a 1:1 tradeoff. That is, women without an outside option for health insurance were willing to work for lower wages in jobs with employer-sponsored coverage.

    Thus, to understand how the Cadillac tax will affect wages, we need to have a better understanding of how workers value generous plans. If workers have efficiently sorted themselves into jobs based on their health insurance preferences — meaning, for example, those who are more risk-averse and thus value higher-cost coverage more work for employers providing this type of coverage — then current compensation packages may reflect the point where workers are indifferent between wages and health insurance and we should see a tradeoff of $1 in higher wages for each $1 of health insurance coverage that is reduced to avoid triggering the Cadillac tax.

    However, if workers with high-cost health insurance plans do not fully value this coverage, then we should see something less than a 1:1 tradeoff. When faced with multiple options for health insurance from an employer, each subsidized to some extent, the structure of the employer’s contribution could induce some workers to choose more generous coverage than they would otherwise opt for if faced with the full price. In addition, the tax-preferred treatment of employer-sponsored insurance could lead to similar behavior. This may be particularly important for higher-cost plans, and the higher-income individuals in those plans, where the price distortion is greater.

    In conclusion, in this context we need to examine more closely what drives the high premiums and the value to workers. The particular benefits and outside options for coverage are important for teasing this out. At the present moment, however, it is not totally clear what we should expect.

    Comments closed