The demand for health insurance coverage by low-income workers: can reduced premiums achieve full coverage?

Health Services Research, Oct, 1997 by Michael Chernew, Kevin Frick, Catherine G. McLaughlin

Few data sets have been available that provide information on the percentage of premium explicitly paid by the employee; therefore, studies that examine employee contribution as a determinant of insurance demand are uncommon. Two exceptions include Short and Taylor (1989) and Feldman et al. (1989). Both studies addressed which of several plans employees chose, as opposed to whether an individual wanted any coverage at all.(10) Although neither study directly investigated employee response to the full premium, both studies conclude that workers' choice of plan is influenced by the employee contribution.

A third study, by Marquis and Long (1995), focuses on workers who did not have access to employer-sponsored coverage. The authors rely on regional variation in the full premium for a standardized plan to identify the demand for insurance. The elasticity of demand measured is based on a comparison between participation rates of individuals in areas where insurance is expensive and those of individuals in areas where coverage is cheap. While this is a reasonable approach given available data, the existence of potentially unobserved regional differences suggests that further examination, using alternative measures of price variation, is worthwhile.

In this study, we test three measures of price. The first measure, which follows from the theory that employees pay for the employer share through lower wages, is the total premium (employee contribution plus employer contribution) for an individual plan at the firm at which the employee works. As noted earlier, it is possible that workers do not perceive this trade-off. Therefore, we have analyzed the influence of two alternative measures of price.(11) The second measure of price is the employee contribution, on the assumption that employees might respond only to their explicit out-of-pocket contribution. Given imperfections in the labor market, variation in the extent to which employees explicitly share the premium cost is likely to provide a useful measure of the price of insurance as perceived by the workers. The third measure of price reflects a spreading theory of the wage trade-off, where wages drop for all workers when an additional worker participates. Each worker's wage is only partially affected by his or her participation decision. This theory of a wage trade-off spread across all workers is examined by estimating a model that includes the employee's contribution (the explicit price) and the estimated employee's share of the employer contribution, computed as the employer contribution divided by firm size (the implicit price).

METHODS

Like Marquis and Long (1995), we estimated the probability of participation with a probit regression. Standard errors were adjusted to reflect correlation of the error terms among workers within the same firm. We used the estimated coefficients from the probit to simulate the percentage of individuals who would participate at post-subsidy prices. The estimated participation rate, after the subsidy, is the average of the estimated participation probabilities evaluated using each worker's characteristics and the post-subsidy price.