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

BACKGROUND

In 1994, approximately 40 million Americans had no health insurance coverage (U.S. Bureau of the Census 1996). A disproportionate number of uninsured individuals have low incomes and do not qualify for Medicaid. Of persons under the age of 65, 35 percent with a family income of less than $14,000 and 30 percent with a family income between $14,000 and $25,000 lack coverage. In contrast, approximately 5 percent of non-elderly individuals with a family income over $50,000 are uninsured (National Center for Health Statistics 1996). Most of the uninsured, low-income adults, are employed. Some are not offered the opportunity to participate in employer-sponsored health plans; others elect not to participate despite being eligible for coverage. Evidence suggests that the lack of coverage reduces consumption of medical care services (Manning, Newhouse, Duan, et al. 1987; Lohr, Brook, Kamberg, et al. 1986). Moreover, although identification of the health consequences associated with little or no insurance coverage is difficult, Lohr, Brook, Kamberg, et al. (1986) report some adverse health effects concentrated among low-income individuals.

Because of strong political opposition to expanding coverage through government mandates, efforts to reduce the number of uninsured Americans increasingly focus on stimulating insurance purchases by reducing premiums through subsidies and various market reforms, such as formation of employer purchasing cooperatives. Although reform efforts at the federal level have been largely unsuccessful, several states have proposed providing direct financial incentives to encourage businesses to offer insurance and individuals to purchase it. Other states are encouraging, and in some cases subsidizing, the development of voluntary purchasing cooperatives, with the hope that they will result in reduced premiums (McLaughlin and Zellers 1994). Many of the state proposals have focused on small firms or their workers, in part because these firms are much less likely than large firms to offer their workers the opportunity to purchase insurance through the firm. Proposed programs often target low-income individuals, thereby directing benefits to those thought most in need.

Much of the literature investigating the response to reductions in health insurance premiums examines the behavior of employers. Survey results reveal that over two-thirds of small businesses that do not offer insurance at present said they would be influenced to do so by a subsidy (McLaughlin and Zellers 1994). However, several demonstration programs and econometric studies of the potential impact of programs designed to reduce premiums to small businesses have concluded that premium reductions, even as substantial as 50 percent, would not greatly increase provision of insurance (Leibowitz and Chernew 1992; McLaughlin and Zellers 1992; Thorpe, Hendricks, Garnick, et al. 1992).

Firm-level analyses are most relevant when addressing the availability of insurance, which generally is determined at the firm level. However, such analyses may not capture the effectiveness of reduced premiums if participation in firm-sponsored plans is voluntary or if workers could purchase subsidized coverage outside of the firm.(1) Firm-level analysis typically cannot capture underlying demand from low-income workers because decisions of all workers are aggregated, masking any heterogeneity of preferences within the firm. In addition, the price faced by workers at the time of the participation decision is rarely observed in firm-level studies.

Evidence based on analysis of individual-level data clearly demonstrates that the probability of purchasing coverage increases as the price of coverage falls (Farley and Monheit 1985; Taylor and Wilensky 1985; Gruber and Poterba 1994; Thomas 1994; Marquis and Long 1995). Although evidence also suggests that the probability of purchasing coverage rises with income, relatively little is known about how income affects the price sensitivity of insurance purchase.

In this study we estimate the responsiveness of currently uninsured, low-income, small business employees to premium reductions, providing insight into the likely effectiveness of many proposed programs aimed at reducing the number of medically uninsured. Our estimates indicate that significant numbers of currently uninsured, low-income individuals would not purchase insurance even at substantially subsidized rates. One of the main strengths of this work is that we observe the explicit price of participation and can therefore explore alternative hypotheses regarding which price employees perceive to be the relevant price when deciding whether or not to participate in their employer's plan.

DATA AND METHODS

This study relies on a unique data set to estimate the demand for health insurance by low-income workers. We observe a wide range of prices and are therefore able to simulate the effect of even large subsidies without out-of-sample forecasts. We analyze the decision to participate in an employer-sponsored plan, focusing on low-income employees who are eligible for that plan.