Working and Poor: How Economic and Policy Changes Are Affecting Low-Wage Workers

14. How Is Health Insurance Affected by the Economy? Public and Private Coverage Among Less-Skilled Adults in the 1990s

Chapter 14 How Is Health Insurance Affected by the Economy? Public and Private Coverage Among Low-Skilled Adults in the 1990s Helen Levy I n 1992, 38.6 million Americans were uninsured (DeNavas-Walt, Proctor, and Mills 2004). Eight years later, at the end of the longest economic expansion in the United States since World War II, the number of uninsured had increased to nearly 40 million (DeNavas-Walt et al. 2004). Why did the booming economy not translate into gains in insurance coverage? This puzzle is even more pronounced when we look at trends by education level, since low-skilled individuals experienced both the largest gains in employment and the largest declines in insurance coverage. Among high school dropouts between the ages of twenty-five and fifty-four, the fraction employed increased from 67 percent to 71 percent, and mean real family income increased by more than 15 percent. But the fraction of dropouts who were uninsured increased from 36 percent to 40 percent. What happened? This chapter analyzes the puzzle of declining coverage in the booming 1990s. Using data from the Current Population Survey (CPS) describing the years 1988 through 2003, I look at trends in income, employment, and insurance coverage by education and sex. I analyze the relationship at a point in time between health insurance coverage and a variety of economic indicators that reflect the health of the economy and the economic well-being of an individual’s own family: the employment of an individual and his or her spouse, family income, and the unemployment rate in the individual’s state of residence. Once these static relationships have been established, I analyze the trends in coverage during the boom and the downturn that followed it. Was the decline driven by public or private 396 / How Is Health Insurance Affected by the Economy? coverage? How much of the decline in coverage can be explained by observable factors such as income and employment? The story that emerges suggests that, during the boom, gains in employment and income increased private coverage for men and women at all levels of education , including the least-skilled. For low-skilled men and high-skilled women, who do not rely heavily on public coverage, these employment and income gains triggered small declines in public coverage that only partially offset the gains in private coverage. But for low-skilled women, who have much higher baseline rates of public coverage, these employment and income gains triggered losses of public coverage that offset the gains in private coverage. The loss of public coverage is even larger than can be explained by the employment and income gains, so that the net effect is an increase in the fraction of low-skilled women who are uninsured. It is unclear why the declines in public coverage are so much larger than what the employment and income gains would have predicted. Welfare reform may have played a role in reducing take-up of public insurance, although the evidence on this point is not conclusive. During the downturn that followed the boom, private coverage declined for men and women at all education levels; public coverage increased, but by much less than private coverage declined, so that uninsurance increased for all groups. Between one-quarter and one-half of the decline in private coverage and the increase in public coverage is explained by changes in the joint distribution of income and employment that occurred during the downturn. THE DATA The data for the analysis come from the Current Population Survey Annual Social and Economic Supplement (the “March supplement”) for the years 1989 through 2004. I restrict the sample to adults age twenty-five to fifty-four, yielding a sample of between 70,000 and 112,000 observations in each year, for a total of 1.3 million observations. Based on responses to questions about health insurance coverage in the calendar year prior to the survey date, I code each individual as having private insurance, public insurance, or no insurance. Each individual is assigned to one of three education categories: less than a high school education (“dropouts”), education exactly equal to high school graduation, and education greater than high school. I also use demographic information (age, race, ethnicity ), labor force participation variables, and information on family income. All income data presented in this chapter have been adjusted to 2000 dollars using the consumer price index (CPI) for all urban consumers. These data are supplemented with information on aggregate unemployment rates by state...