This post has been cited in the 21 January 2010 edition of the Health Wonk Review.
I just read the best paper I am aware of on what economists believe and know about employer-based health insurance. In Modeling Employer Decisions to Offer Health Insurance, Sherry Glied and Joshua Graff Zivin review the relevant theoretical and applied literature up to 2004 (the year of publication). The paper is very readable with no equations and only a little economics jargon. Most policy wonks should be able to follow along.
Beyond review of the literature, the authors point to areas where it is or the data upon which it is based are weak or nonexistent. They also have nice sections on stylized facts from the empirical evidence, implications of theory, and what both suggest for future areas of research.
Because I do applied work I’ll just highlight the stylized facts from empirical research. Glied and Zivin write:
The decision to offer coverage varies with the price of coverage, particularly for small firms; characteristics of the labor force, including average wage levels and the distribution of wages within the firm, affect the offer decisions; and that over time, firms have not responded to rising costs by ceasing to offer coverage. The costs of coverage fall on workers in aggregate, and perhaps at the group level, but higher cost individuals do not appear to receive lower wages in exchange for health insurance (with the possible exception of workers in very small firms). Worker take-up decisions are sensitive to the employee share of premiums and employers charge employees a share of premiums, even though this sometimes exposes them to greater taxation.
In sum, the extent and nature of heterogeneity of preferences over health insurance within the workforce has strong implications for whether firms offer coverage and whether individuals take-up coverage that is offered to them. Low wage workers who work with higher wage workers appear to receive greater compensation (measured as the sum of wages and employer health insurance payments) than do those who work only with other lower wage workers. Employers do appear to manipulate the employee share of premiums in order to reduce participation rates among workers who do not value benefits (at least up to a point). To date, there is little evidence on how employers vary these worker shares with policy changes.
That may not sound terribly informative to some readers. A high-level take away is that marrying health insurance to employment introduces a vast amount of variation on who is offered and takes up what. Details of health insurance offers, plans, enrollment, and allocation of premium payment between employer and worker have more to do with firm strategy, the profile of a firm’s workers in several dimensions, and the market for labor than about health. Details and supporting evidence on all these points are found in the papers Glied and Zivin reference. I encourage those interested to give the paper a read. Let me know what you think.