Health Affairs, 22, no. 3 (2003): 203-213 doi: 10.1377/hlthaff.22.3.203 © 2003 by Project HOPE	New Online   New Issue: China & India   Obesity In China   Pay Cuts For Medicare Docs   Access To Care Woes

This Article Abstract Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Kapur, K. Articles by Marquis, M. S. Search for Related Content PubMed PubMed Citation Articles by Kapur, K. Articles by Marquis, M. S. Related Collections Insurance Coverage Insurance - Employer-Based System

DataWatch

Health Insurance For Workers Who Lose Jobs: Implications For Various Subsidy Schemes

Abstract

 
A number of proposals have been made to help laid-off workers purchase health insurance. We use data from the 1996 Medical Expenditure Panel Survey to profile the insurance status of workers who left a job. Our descriptive analysis suggests that it might be difficult to design policies that target those who would otherwise be uninsured and that large subsidies might be needed to help laid-off workers.

Under federal law, workers who leave a job with insurance benefits have the option to continue group coverage—known as COBRA (Consolidated Omnibus Budget Reconciliation Act of 1985) coverage—for up to eighteen months by paying 102 percent of the premium. The federal law applies to workers in firms with twenty or more workers, but thirty-eight states have enacted laws that also require continuation provisions for workers in smaller firms. Only 20–25 percent of those eligible to purchase COBRA coverage do so.³ The high cost of COBRA coverage, especially for those who have just lost a job, may be a deterrent for many. Estimates for 2001 place the annual cost of COBRA family coverage at about $7,000–$8,000, or as much as two-thirds of a typical unemployment check.⁴

The economic downturn and rising unemployment have prompted proposals to help those who have lost their jobs to afford transitional insurance coverage. Such transitional subsidies can also help reduce job lock—employees’ reluctance to change jobs because of the possibility of loss of insurance—and thereby improve productivity.⁵

Some of the proposals have explicitly focused on subsidizing COBRA coverage. Others would link the subsidy with eligibility for unemployment insurance and would allow the subsidy to be applied to the purchase of individual coverage, as well as continuation coverage. The Trade Act of 2002 provides a refundable tax credit for dislocated workers eligible to receive Trade Adjustment Assistance benefits that can be applied to the purchase of continuation coverage and certain other qualified health insurance coverage. Still other proposals, including the administration’s proposal, would provide tax credits to the low-income unemployed and those whose employers do not provide coverage.⁶

To design and assess the cost of such proposals requires information about the number and characteristics of people who lose a job; COBRA eligibility and take-up rates; the number and characteristics of people who become unemployed and find other sources of coverage; and the duration of coverage under transitional policies and how it varies among subpopulations. Analysis of these issues has been limited for several reasons. First, answers require panel data to identify those who lose a job and to track their subsequent insurance coverage. Many studies have used cross-sectional data to determine how many people would be eligible for COBRA should they lose the job.⁷ However, job loss rates are likely to be higher for people in firms that do not offer insurance and for workers who are less likely to take up coverage; earlier studies have confirmed that jobs with insurance last longer than jobs without insurance do.⁸ Thus, we need to study people when they lose a job and observe their sources of insurance in the period following the job loss—that is, we need longitudinal data. A panel study of the unemployed in 1977 concluded that only about 8 percent of the unemployed lost insurance because of job loss and that job loss was the direct cause of uninsurance for only about 30 percent of the uninsured unemployed.⁹ If similar results are found today, policies that focus only on those who lose insurance, such as COBRA subsidies, would not help many of the unemployed uninsured. Second, few data sources provide direct information about COBRA coverage. We overcame these two problems by using data from a recent panel study that explicitly identified COBRA as a source.¹⁰

Data And Methods

Top Data And Methods Study Findings Discussion NOTES

 
Data. Our data come from the 1996 Medical Expenditure Panel Survey (MEPS), a stratified, nationally representative sample of about 21,500 noninstitutionalized people in 10,500 households. Each family participated in five rounds of data collection over a two-year period. During each round, information on family members’ employment and health insurance coverage was collected, along with a detailed set of variables on health, demographics, and health care use. The data from the study are released as a variety of public use files.

Detailed information about COBRA coverage is available only for the 1996 year of the panel members’ participation. Because this information is important in evaluating new policies related to transition coverage, we focus our analysis on 1996. However, to provide some information about the insurance status of the unemployed over a longer period, we also discuss results for the two-year period (including 1996 and 1997) of the panel. Our analysis sample from the 1996 MEPS includes people under age sixty-five who reported leaving a job during the year.¹¹ Our final analysis sample includes 1,821 job leavers. Our study year of 1996 was a time of low unemployment and economic prosperity, in contrast to the economic conditions of today. Therefore, we also used data from the 1996 and 2002 February Current Population Survey (CPS) to profile the changes in the population of job leavers since the time of our study. We also used estimates of the newly unemployed in 2002 to estimate the number of workers who might benefit from alternative policies to help them gain access to insurance.

Study methods. Voluntary and involuntary job terminations. The objective of many proposals is to help laid-off workers. Some proposals for extending continuation health insurance coverage target them explicitly—for example, by restricting eligibility to those receiving unemployment benefits. Therefore, we examined voluntary and involuntary job terminations separately. The latter are those who reported that they were laid off, the job ended, the business dissolved, or they were receiving unemployment insurance benefits during the year.¹² Involuntary unemployment may be more prevalent in today’s economic environment than it was in 1996, so separating out voluntary and involuntary terminations is important for extrapolating findings from 1996 to today’s economy. As noted above, we use data from the 2002 CPS to help in this extrapolation.

Insurance following job loss. We measure insurance following job loss by looking at reported coverage in the following month. Because workers can elect COBRA coverage at any time within sixty days after losing or leaving a job, COBRA coverage in the month following the job loss may understate participation rates. Therefore, we also look at reported coverage three and six months later. In addition, looking at multiple points in time provides some information about the duration of immediate post–job loss insurance status. The sample size is reduced for these analyses, because it is limited to those who lost a job early enough in the year so that the three- or six-month follow-up period is during 1996. We were able to observe a three-month follow-up period for 1,515 workers leaving or losing a job and a six-month follow-up for 871 workers.¹³

Insurance status in the period after a job loss is classified as COBRA coverage from lost job, group coverage from a new or another employer of the policyholder, group coverage from the old employer, group coverage through another family member, other private coverage, public coverage, or uninsured.¹⁴ These categories were defined to be mutually exclusive by imposing an insurance hierarchy in the order specified above. Some employers provide continued subsidized coverage in the group plan after a worker is laid off, and so we distinguish between coverage in a new or old group plan to measure the extent of this practice.¹⁵

We also examined the length of time that employees who elect COBRA coverage continue with it. The cumulative loss rate for each month, t, following the purchase of COBRA coverage is the proportion of enrollees who drop coverage by time t. To estimate this rate in each period t, we selected policyholders and dependents that purchased their coverage at least t + 1 months prior to the end of 1996 and measure the proportion of them who have dropped coverage by time t.

Statistical significance. Unless otherwise noted, all contrasts that we mention in the text are statistically significant. However, small sample sizes limit some of our contrasts, and so we also give the 95 percent confidence interval for selected important rates; these are estimated taking into account the complex survey design of MEPS. All analyses were weighted by the sampling weights provided in MEPS.

Study Findings

Top Data And Methods Study Findings Discussion NOTES

 
Characteristics of job leavers. Most people who left a job did not have insurance from that job (Exhibit 1). More than 70 percent of voluntary job leavers did not have coverage from the employer. The share of voluntary and involuntary job leavers in jobs without insurance benefits did not differ significantly.

Almost 40 percent of all job leavers were in low-income families (below 200 percent of the federal poverty level), even though such workers accounted for only 21 percent of employees (not shown). Low-income job leavers were less likely to be eligible for COBRA than higher-income workers were, primarily because they were less likely to have left jobs with insurance. Only 16 percent of low-income job leavers were in a job with insurance, in contrast to 35 percent of high-income job leavers. Conditional on leaving a job with insurance, low-income workers were not significantly less likely to be eligible for COBRA.

Job leavers in 1996 and 2002. Involuntary unemployment is somewhat more prevalent today than in 1996, but the characteristics of job leavers have remained fairly constant over time (Exhibit 2).¹⁶ The share of all job leavers who left a job involuntarily increased about 9 percent between 1996 and 2002. In both time periods, involuntary job leavers had lower income and less education than voluntary job leavers had and were more likely to be nonwhite and less likely to have a working spouse. Because the characteristics of job leavers in 2002 and 1996 are similar, our conclusions from the 1996 MEPS are likely to apply in today’s environment.

A substantial proportion of those who lost a job were uninsured even when they had access to subsidized group coverage from another family member. Although half of involuntary job leavers who had another family member with group coverage were enrolled in that plan one month after the job loss, almost 30 percent of them were uninsured in that time frame. Most of the remaining 20 percent had coverage through their old or new job (data not shown).

Eighteen percent (95 percent CI ± 5) of voluntary job leavers eligible for COBRA coverage enrolled within the first month following a job loss (Exhibit 3). But an almost equal number continued to have subsidized coverage from their former employer. Among involuntary job leavers eligible for COBRA coverage, COBRA take-up was similar to take-up among voluntary leavers eligible for COBRA. However, slightly more of these former workers continued to have subsidized coverage from their prior employer. This coverage is a short-term benefit, and many fewer workers report this source of insurance six months later.

Involuntary job leavers were somewhat less likely than voluntary job leavers were to move directly to a new job with insurance benefits (9 percent versus 12 percent).¹⁸ COBRA-eligible job leavers, who all left a job with insurance, were more likely than other job leavers were to move to a new job with insurance within a month. These differences persisted over time, and within eighteen months of a job loss, 69 percent of workers who were eligible for COBRA benefits held employer-sponsored health insurance (data not shown).

During the six months following a job loss, more workers arranged for coverage, and the uninsurance rate fell. Nonetheless, more than a third of voluntary job leavers and 40 percent of involuntary job leavers remained uninsured six months following a job loss. Eighteen months later, the uninsurance rate was 31 percent for both voluntary and involuntary job leavers (data not shown). For voluntary job leavers this was a small decline in the uninsurance rate.¹⁹ However, for involuntary workers this constituted a nine-percentage-point drop in the uninsurance rate.

Income differences in post-job insurance status. Low-income workers who left a job (those in families with an income of 200 percent of poverty or below) were more likely to be uninsured than higher-income workers were (Exhibit 4).²⁰ More than half of low-income workers who voluntarily left a job were uninsured (95 percent CI 48–58 percent), and almost two-thirds of low-income involuntary job losers were uninsured (95 percent CI 54–70 percent). However, most of these workers were uninsured prior to the job loss; those who lost insurance accounted for 3 percent and 9 percent of the population, respectively. The uninsurance rate for higher-income voluntary job leavers was 28 percent (95 percent CI ± 4); for higher-income involuntary job leavers, 35 percent (95 percent CI ± 7).

Low-income workers who left a job were less likely than higher-income workers were to have insurance from another job and were less likely to get coverage through a family member in the month following a job loss. However, they were more likely to have public coverage. These differences persist over time.

Estimated beneficiaries under alternative subsidy schemes. We used our results and the February 2002 CPS data to estimate how many involuntary job leavers would benefit from various subsidy schemes. Based on the number of new lost jobs in February 2002, we estimated that about 17.9 million workers would be involuntarily terminated from a job during the year.²¹ COBRA coverage subsidies would be available to 3.9 million of them (or 22 percent, Exhibit 1). Only about 0.9 million of these would be workers with family income below 200 percent of poverty (5 percent). In contrast, tax credits targeted to all low-income involuntary job leavers would be available to about 7.2 million workers (40 percent). If the credit were restricted to uninsured low-income workers (62 percent of involuntary low-income leavers), the credit would be available to 4.4 million workers.

Duration of COBRA coverage. As intended, COBRA coverage provides short-term coverage for most who purchase it. About 17 percent (95 percent CI ± 8) of participants held the policy for a month or less, about 41 percent for three months or less, and about 67 percent (95 percent CI ± 14) for fewer than six months.²² About 46 percent of those who dropped COBRA coverage became uninsured.

Discussion

Top Data And Methods Study Findings Discussion NOTES

 
Most lost jobs do not carry insurance benefits. Our analysis does not identify the factors that account for this, but the result is consistent with job lock—that is, a reluctance to leave a job when it would also result in loss of insurance. Alternatively, jobs without insurance may be more likely to be "bad" jobs with higher turnover.

COBRA subsidies. Proposals to help laid-off workers with subsidies for the purchase of COBRA coverage or other continuation of benefits would benefit only a small fraction of involuntary job leavers and would not be target-efficient. About 59 percent of COBRA-eligible involuntary job leavers purchase private insurance (Exhibit 3). Some of these people would take advantage of the subsidy. Those who now purchase COBRA coverage (about 15 percent of eligible involuntary job leavers) would be expected to take advantage of the subsidy. Those who obtain new group insurance on their own or through a family member might also take advantage of the public subsidy if it exceeded the group subsidy. The substitution of public money for private money among involuntary job leavers might be desirable on equity grounds, but it would increase the cost of the program.

Subsidies to low-income involuntary job leavers. Tax credits or other subsidies restricted to low-income involuntary job leavers would be available to 40 percent of this group (Exhibit 1) and would be more target-efficient. About 55 percent of uninsured involuntary job leavers would qualify for such subsidies. Moreover, there would be little substitution of subsidies for currently purchased private insurance, even if available to all low-income involuntary job leavers, because only 20 percent of these workers purchase private insurance in the month after losing a job. Almost two-thirds of them are uninsured (Exhibit 4). However, subsidies may have to be quite substantial to make private coverage affordable for these workers; expansion of public programs might reach more of them. This possibility is illustrated by the high rate of uninsurance among involuntary job leavers who have access to subsidized coverage through a family member.

Duration of subsidy. The period of participation in transitional coverage policies may suggest that temporary subsidies need not be of long duration. On the other hand, the short duration of COBRA coverage may be in part a result of its high cost. Many who drop COBRA become uninsured. Moreover, the majority of COBRA-eligible beneficiaries obtained new jobs with insurance within a relatively short period, but this was not true of others who left or lost a job.

Behavioral effects. Subsidies might alter workers’ behavior, leading to increased periods of unemployment.²³ On the other hand, they might also increase mobility and improve productivity.²⁴ These effects need to be factored into any estimates of the costs and benefits of alternative policies.

Subsidies also might alter employers’ behavior. Almost 30 percent of involuntary job leavers in COBRA-eligible jobs have subsidized employer benefits that continue for several months after the job loss. But this could change with government subsidies. Earlier simulation analyses of subsidies to the unemployed have indicated that up to a third of subsidy benefits would be paid to those who otherwise would have had continued employer-subsidized benefits.²⁵

Limitations. Our analysis presents new data about insurance coverage for laid-off workers, COBRA take-up, and the implications for current policy proposals. Our data, however, are limited by small sample sizes and limited follow-up duration. They also rely on self-reports of type of coverage that may be subject to misreporting—especially COBRA coverage. Moreover, the data pertain to a period of economic good times, and the characteristics of workers losing a job and their patterns of insurance may differ during an economic downturn. Further exploration of these issues with larger samples and longer periods of time would help in formulating effective policy to help laid-off workers obtain affordable coverage.

Editor's Notes

 
Kanika Kapur is an associate economist at RAND in Santa Monica, California. Susan Marquis is a senior economist at RAND in Arlington, Virginia.

This study was supported by Contract no. J-9-P-7-0045 from the Pension and Welfare Benefits Administration, U.S. Department of Labor (DoL). Any views expressed herein are solely those of the authors, and no endorsement by the DoL or RAND is intended or should be inferred.

NOTES

Top Data And Methods Study Findings Discussion NOTES

 

1. S.A. Glied, "Challenges and Options for Increasing the Number of Americans with Health Insurance," Inquiry 38, no. 2 (2001): 90–105[Medline]
2. J.M. Lambrew, How the Slowing U.S. Economy Threatens Employer-Based Health Insurance, Pub. no. 511 (New York: Commonwealth Fund, November 2001).
3. P. Flynn, "COBRA Qualifying Events and Elections, 1987–1991," Inquiry (Summer 1994): 215–220; and M.C. Berger et al., "Health Insurance Coverage of the Unemployed: COBRA and the Potential Effects of Kassebaum-Kennedy," Journal of Policy Analysis and Management 18, no. 3 (1999): 430–448.[Medline]
4. G. Scandlen, Helping Laid-Off Workers Keep Insurance, Brief Analysis no. 373 (Washington: National Center for Policy Analysis, 4 October 2001); and Lambrew, How the Slowing U.S. Economy Threatens Employer-Based Health Insurance.
5. J. Gruber, "Transitional Subsidies for Health Insurance Coverage," Inquiry (Summer 2001): 225–231.
6. See S. Dorn and J.A. Meyer, What Health Coverage Would Laid-Off Workers Obtain under Recent Tax Credit Proposals?; (Washington: Economic and Social Research Institute, March 2002).
7. S. Zuckerman, J. Haley, and M. Fragale, Could Subsidizing COBRA Health Insurance Coverage Help Most Low-Income Unemployed?; Health Policy Online, no. 2, 17 October 2001, www.urban.org/UploadedPDF/410351_HPOnline_2.pdf (11 March 2003); and L. Duchon et al., Security Matters: How Instability in Health Insurance Puts U.S. Workers at Risk (New York: Commonwealth Fund, December 2001).
8. J.A. Klerman, J.L. Buchanan, and A. Leibowitz, "Labor Turnover and Health Insurance," in Health Benefits and the Workforce (Washington: U.S. Department of Labor, Pension and Welfare Benefits Administration, 1992), 43–54.
9. A. Monheit et al., "Health Insurance for the Unemployed: Is Federal Legislation Needed?;" Health Affairs (Spring 1984): 101–111.
10. The Survey of Income and Program Participation (SIPP) is another panel data set that tracks changes in insurance. However, it does not explicitly identify COBRA coverage, so we chose to use MEPS.
11. If a person had more than one job turnover, we examined insurance status following the first job loss.
12. Other reasons given for leaving a job include retired, quit, took unpaid leave, or had illness or injury.
13. We compared the insurance status in the initial month following the job loss for the full sample and these two subsamples to ensure that the patterns were similar and that the change in status we report between the initial month and these later months was not attributable to sample differences.
14. These are based on self-reports of type of coverage. We define COBRA eligibility as holding a job with insurance from a firm with twenty or more employees. A categorical measure for firm size, which groups firms with ten to twenty-five employees in a single category, was used when the continuous measure for firm size was missing. In this case, COBRA eligibility was identified for those with jobs with insurance in firms with more than twenty-five workers. If a person has COBRA coverage and non-COBRA employer coverage from the same employer in the same round, it is not possible to distinguish these two types of coverage in the MEPS data. Therefore, we use the reported job end date to determine the start of COBRA coverage.
15. If the worker reported having coverage from an employer in the month after leaving a job but did not report COBRA coverage in the month, we assume that the employer continued to subsidize the insurance. Workers on temporary layoff are dropped from the analysis.
16. We use the CPS for both years of the contrast, rather than the 1996 MEPS data, to have consistent definitions over time. We defined job leavers in the CPS as all people who were not working but had worked in the past month, and those who had changed jobs within the past month. Involuntary job leavers are those reporting that the reason for unemployment was loss of a job or the end of a temporary job.
17. This is similar to the 8 percent figure for 1977. Monheit et al., "Health Insurance for the Unemployed."
18. In some cases, this is a new job; in others, it is a previously held second job.
19. We calculated uninsurance rates for twelve and eighteen months after job loss by linking the 1996 and 1997 MEPS data. The 1997 MEPS data do not contain information on COBRA.
20. Family income is measured as total income for 1996, so it does not necessarily reflect the family circumstances after the job loss.
21. This estimate is based on the number of people under age sixty-five who reported a lost job in the month preceding the February 2002 CPS. The estimate does not include those who left a job voluntarily. It may also overcount the number of people displaced because we counted lost jobs in a month and multiplied them by 12; that is, we did not correct for multiple job losses. If COBRA subsidies were available to all job leavers including voluntary terminations, the number eligible for the subsidy would increase to 10.6 million.
22. The samples of low-income and involuntary job loss COBRA participants were too small to allow reliable estimates of COBRA durations for these groups.
23. J.R. Baumgardner, "Providing Health Insurance to the Short-Term Unemployed," Inquiry 35, no. 3 (1998): 266–279.[Medline]
24. Gruber, "Transitional Subsidies."
25. Baumgardner, "Providing Health Insurance to the Short-Term Unemployed."

                      What's this?

This article has been cited by other articles:

				K. Steenland and L. E. Pinkerton Mortality Patterns following Downsizing at Pan American World Airways Am. J. Epidemiol., January 1, 2008; 167(1): 1 - 6. [Abstract] [Full Text] [PDF]