Tax Rules Raise Firm Health Benefits

"For firms with fewer than 100 employees, each10 percent rise in the subsidy raises health insurance spending levels by almost7 percent."

When a business provides its employees with health insurance, that benefit is not subject to either individual income tax, or to the payroll tax that finances the Social Security and Medicare programs. This "tax subsidy" benefits both the company and its employees. So, what would firms do if it were to disappear or diminish in size? Would there be a reduction in employee-provided health insurance coverage and spending? In How Elastic is the Firm's Demand for Health Insurance? (NBER Working Paper No. 8021), NBER Research Associate Jonathan Gruber and his coauthor Michael Lettau conclude that firms indeed are fairly responsive to changes in the tax subsidy given to health insurance.

The subsidy is large, costing the federal, state, and local governments more than $100 billion in lost revenues. Individuals who buy their health insurance directly do not benefit from the subsidy and thus, in effect, pay more. The tax subsidy is also regressive, providing the largest tax break for the most well off employees. And, the subsidy may be leading to "over-insurance" and perhaps to cost inflation in the health sector. At the same time, removing all or part of this subsidy could lead to large reductions in insurance coverage.

Gruber and Lettau assess the impact of the tax subsidy on firms' insurance decisions by comparing firm insurance-offering rates and spending levels to the tax subsidies to insurance for that firm's workers. They use a unique dataset, the Employee Compensation Index (ECI) data collected by the Bureau of Labor Statistics (BLS). These data present information on both firms and a representative sample of workers within those firms. Using these ECI data from 1983 to 1995, the authors compute the tax subsidy to insurance for the median worker in each firm in each year. This tax subsidy varies significantly across workers of different income levels, over time, and across states, allowing the authors to assess how the subsidy affects firm behavior.

Gruber and Lettau find that firms are fairly sensitive to tax subsidies in their decisions to offer insurance. In particular, for each10 percent rise in the tax subsidy to insurance for the firm's median worker, the rate of insurance offering rises by about 3 percent. The responsiveness of small firms in their offering decisions is particularly large; for firms with fewer than 100 employees, each 10 percent rise in the subsidy leads to a rise in offering rates of 6 percent. And, such firms are much more responsive in their spending decisions to tax subsidies: each10 percent rise in the subsidy raises health insurance spending levels by almost7 percent.

The authors also use their data on the distribution of workers within the firm to ask: which workers' preferences appear to be relevant to firm decisionmaking? Do firms pay attention just to their most highly compensated workers, who have the largest tax subsidies? Or does the "typical" worker have the largest vote? The authors conclude that the right answer is some mix of these two views: firms' insurance decisions are influenced jointly by the tax price of the worker with the median wages in the firm, and by the tax price of the highest paid workers.

The implications of these findings are that tax reforms can have major impacts on employer-provided insurance. For example, a radical reform that included health insurance spending by firms in both the state and federal income tax base (but continued to exclude that spending from the payroll tax base) would eliminate 60 percent of the existing subsidy to employer-provided insurance. Such a change would lead to 10-19 percent fewer firms offering insurance, and a reduction in insurance spending among those offering insurance of 20-28 percent. Assuming that those firms that stopped offering insurance were spending the average amount, the total spending per individual on insurance would decline by 30-42 percent, or $760-$1080.

Such a decline would mean that 16- 29 million people would lose their employer-provided coverage. That amounts to 36-66 percent of the size of the current pool of uninsured Americans. Of course, not all of these individuals would become uninsured; some would seek coverage from other sources. But the net result could be a large rise in the number of persons without access to health insurance.

Should income tax rates be cut 10 percent, which is comparable to the proposal of President Bush, the insurance tax price would rise 3.4 percent. Firms would reduce their offerings of insurance by 1.1-1.9 percent, and spend 3.3-4.7 percent less. "Thus, even this very modest tax reform could have a non-trivial impact on the spending of employers on health insurance in the U.S.," Gruber and Lettau note.

-- David R. Francis

The Digest is not copyrighted and may be reproduced freely with appropriate attribution of source.

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us