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

It is difficult to estimate the extent to which these types of barriers influence our findings. It is interesting to note that low-income individuals do not always take advantage of other public benefits for which they are eligible. For example, the participation rate of eligibles for food stamps is 74 percent, below the participation rates observed here. Whatever these barriers to participation in benefit programs, they seem to be significant for low-income individuals.

What does seem clear is that while modest premium reductions may be more palatable than mandates, they will not achieve universal coverage for certain population groups. Moreover, much of the benefit will accrue to workers who already purchase coverage (or their employers). Given the relatively small response to premium reductions observed in this study and by other investigators, premium subsidies are likely to be very expensive relative to the modest reduction in the number of uninsured that will be achieved.

ACKNOWLEDGMENTS

The research support of The Robert Wood Johnson Foundation is gratefully acknowledged. Two anonymous referees provided useful comments.

NOTES

1. A demonstration program in Washington State that provided large subsidies directly to low-income individuals had a greater impact than the subsidy programs aimed at employers (see Helms, Gauthier, and Campion 1992, or Hoare, Mayers, and Madden 1992).

2. Our measure of "another source of coverage" is based on the response to the question, Does the worker have another source of coverage? Although some workers could purchase coverage in the individual market, this option generally will be less desirable because employer-based coverage is purchased with pre-tax dollars and priced below individual-based coverage. Some individuals may also have access to government-provided coverage and would hence be omitted from the analysis.

3. Although 2,239 firms responded, 297 firms that participated in a health insurance demonstration project were not randomly selected and were dropped from the sample.

4. For salaried workers, the hourly wage was computed based on their salary and hours worked.

5. We assume that the ratio of the premium for a family plan to the premium for an individual plan is a constant 2.5 for all employer groups. This is the ratio observed from data gathered directly from a sample of insurers in four of our sites. Using this ratio (r) and the knowledge of the type of plan chosen by each enrolled employee, we calculate an individual plan premium for each firm as: [P.sub.i] = [P.sub.t]/[F(r-1) 1], where Pi denotes the individual plan premium, Pt denotes the total firm premium per enrollee, and F denotes the percentage of enrollees with a family plan.

6. Data were available for each individual indicating whether the employee paid "all," "some," or "none" of the premium, but not percentages for each employee. However, we did ask at the firm level what portion of the individual plan premium was paid by the firm. When collapsed into the "all," "some," or "none" categories, the responses corresponded closely with the employee-specific data.