Demand for Health Insurance Among the Uninsured

A central goal of the Affordable Care Act of 2010 (ACA) is to cut the number of uninsured Americans from the current level of roughly 50 million. To achieve this goal, the ACA includes an expansion of Medicaid, substantial premium subsidies for low- and middle-income households, health insurance exchanges, and an individual mandate.

Projections of the effect of the ACA, much of which is not implemented until 2014 or later, rely on existing estimates of the price elasticity of demand for health insurance. Yet existing estimates are generally based on workers' decisions regarding enrollment in employer-sponsored health insurance and may not be well-suited for these projections. The uninsured are substantially poorer than the average worker who is offered employer-provided insurance and are rarely offered the opportunity to purchase insurance through their employer. Many of the uninsured have been denied coverage in the past. The decision to purchase subsidized insurance from a state-run exchange may also differ from that to enroll in employer-sponsored coverage.

In The Demand for Health Insurance Among Uninsured Americans: Results of a Survey Experiment and Implications for Policy (NBER Working Paper 16978), researchers Alan Krueger and Ilyana Kuziemko conduct a survey experiment to assess the willingness to pay for a health plan among a large sample of uninsured Americans. This study represents the first attempt to elicit such information from this population.

The data used in the study were collected as part of the Gallup-Healthways Daily Poll, a daily survey of about 1,000 individuals. During a two-week period, the poll asked all individuals who reported being uninsured-about 1,300 individuals in all-whether they would be willing to pay some specified amount (for example, $3,000) in order to obtain a health insurance policy as good as the one members of Congress have. If individuals said no, they were asked the question again with a lower amount. To avoid the "anchoring bias" that can result from starting at a particular value, the starting dollar amount was varied randomly.

The authors begin with a simple comparison of the characteristics of the insured and uninsured in their data. The uninsured are younger, more likely to be male, and less likely to be married. They also have lower income than the insured, are less likely to have a job, and, interestingly, are twice as likely to have been denied coverage when trying to purchase health insurance in the past.

Turing to the results, the authors find that the uninsured are quite sensitive to price in their health insurance purchase decisions. If offered the opportunity to purchase insurance for a $2,000 annual premium, more than 60 percent of those who are currently uninsured say they would voluntarily buy insurance.

Applying these estimates to the provisions of the ACA suggests that the law will substantially reduce the number of uninsured. Under the current configuration of subsidies, over 75 percent of the uninsured are projected to enroll, implying that 39 million individuals would gain coverage as a result of the law. Removing the tax penalty imposed by the individual mandate, a provision whose constitutionality is being challenged in federal court, would result in 7 to 12 million fewer individuals gaining coverage.

The authors note that this projected decrease in the uninsured population is larger than the estimate generated by the Congressional Budget Office, likely resulting from the fact that the price sensitivity they estimate and use is greater than that found in previous studies. The authors offer several possible reasons for this difference. First, the uninsured population is poorer than the population of workers offered employer-sponsored insurance and may be more price sensitive as a result; indeed, the authors find that within their sample, relatively richer people are less price sensitive. Second, this study is designed to evaluate sensitivity to prices in a range that is generally lower than that explored in previous work.

Turning to the issue of adverse selection (the question of whether those who are less healthy will be more likely to buy insurance), the authors find that less healthy individuals are less price sensitive, but no more likely to enroll under the ACA's subsidy schedule. However, the authors note that the results might differ under other subsidy schedules.

The authors conclude "our results suggest that extrapolating the effects of premium subsidies for the uninsured from the elasticities generated in past papers could seriously under-estimate the coverage rates these policies could achieve."

The authors acknowledge financial support from the Industrial Relations Section at Princeton University.