Cover: Effects of Access to Post-Retirement Health Insurance on Retirement Behavior and Insurance Coverage

Effects of Access to Post-Retirement Health Insurance on Retirement Behavior and Insurance Coverage

Published 1998

by Lynn A. Karoly, Jeannette Rogowski

Research Brief

Insured workers might hesitate before deciding to retire early (before age 65), because of concerns about post-retirement health insurance. For many in their age bracket who might face high medical costs, the need to continue health insurance coverage can be especially important. For the near elderly who are considering early retirement, what are the choices in health insurance coverage, and how are those choices likely to influence their decision about early retirement and their health insurance coverage if they retire?

RAND researchers Lynn Karoly and Jeannette Rogowski find that policy initiatives that aim to increase access to affordable post-retirement health insurance are likely to motivate some of the near elderly to retire early. In addition, with more options for obtaining post-retirement insurance, insured workers who do retire early are more likely to be insured after retirement.

What Are Health Insurance Choices Now if Insured Workers Retire Early?

For insured workers contemplating early retirement, the choices in health insurance coverage after retirement may be limited. Before workers become eligible for Medicare at age 65, they could purchase an individual health insurance policy when they retire early. But if they do so, they might encounter some problems. First, the costs of individual health insurance can be prohibitively expensive, especially if premiums are based on the retiree's age and health status. Second, in some cases, insurers may consider such individuals "bad risks" and refuse to extend coverage to them or agree to cover them but only by excluding preexisting conditions.

Other options may be available for early retirees, however. For example, instead of purchasing individual insurance, workers contemplating early retirement might opt for health coverage under their spouse's employer-provided health plans. Or, their own employers might extend health insurance coverage as a fringe benefit for early retirees. However, the value of this benefit can vary widely. Employers differ in the extent to which they impose eligibility requirements (e.g., age or tenure rules), provide the same level of health benefits (e.g., deductibles, copayments, or services covered) for retirees as for active workers, and share in the cost of the health insurance premium. For some retirees who qualify for the benefit, their coverage is less generous than it was when they were working, and they pay the full cost of premium payments. For others, the employer pays the full premium cost of coverage, which is equivalent to the benefit provided to active workers.

Another option for insured workers is COBRA (Consolidated Omnibus Budget Reconciliation Act, 1985) continuation coverage, at least for the first 18 months after retirement. Under COBRA, workers in firms with 20 or more employees who have health insurance while working are eligible to continue health benefits at 102 percent of the group rate for 18 months after retirement.

Recent legislation further expands continuation coverage options for insured workers. The 1996 Health Insurance Portability and Accountability Act (HIPAA) allows workers with COBRA benefits to purchase private individual policies with no preexisting-condition exclusions after the 18 months of COBRA coverage. Workers in small firms (fewer than 20 workers) with health insurance while working also have guaranteed purchase privileges into the individual market immediately upon retirement.[1] However, premiums in these individual plans may be subject to state regulation. In the absence of community rating, premiums are likely to be higher than those paid while the worker was employed or covered by COBRA.

Today's retirees rely on all of these options (and are expected to use the portability provisions of HIPAA). Coverage through a former employer is most prevalent. Data from the 1992 Survey of Income and Program Participation (SIPP) reveal that nearly half of all nonworking retirees age 55 or over have employment-based coverage, most in their own name through a former employer. Among early retirees, the vast majority—approximately 70 percent—have such coverage either as an employer-provided retiree health benefit, as a continuation coverage benefit, or as a dependent on their spouse's employer plan. Despite the various options for post-retirement health insurance coverage, one in fourteen early retirees is uninsured after retirement.

Policies Can Affect Insurance Coverage After Retirement

Health insurance options after retirement have implications for the health insurance status of retirees who leave their jobs before becoming eligible for Medicare. In a recent study, Karoly and Rogowski find that policies mandating access to continuation benefits such as COBRA and HIPAA can increase the likelihood that workers who retire before Medicare eligibility will be covered by health insurance upon their retirement. More specifically, because continuation and portability mandates allow more workers to buy coverage without preexisting-condition exclusions, and, in the case of COBRA, at group rates, these policies can increase the number of workers who will be insured after retirement.

The analysis, based on data from the 1984 to 1987 SIPP, shows that since the enactment of COBRA, which became effective in 1986, early retirees are less likely to become uninsured than are their counterparts who retired before COBRA was in effect. In the post-COBRA period, the probability of retaining one's own employer coverage after retirement increased compared to the pre-COBRA period, from 72.0 percent to 78.5 percent. At the same time, COBRA appears to have reduced the probability of moving from own-employer coverage to being uninsured. Before COBRA was enacted, this probability was 7.5 percent; in the post-COBRA period, it is 2.9 percent.

Figure 1. Uninsured Retirees Previously Covered by Own Employers

SOURCE: Jeannette A. Rogowski and Lynn A. Karoly, "COBRA Continuation Coverage: Effect on the Health Insurance Status of Early Retirees," Health Benefits and the Workforce, Volume 2, Department of Labor, 1998.

Among early retirees who were uninsured after retirement, the proportion previously covered by their own employers declined significantly in the post-COBRA period, from 52.1 percent to 25.6 percent (see Figure 1). Nonetheless, more than one-fourth of uninsured retirees in the post-COBRA period had previously been covered through their own employers. Because these retirees, with the exception of those in small firms, should have been eligible through COBRA to continue their employment-based health insurance for 18 months after retirement, it is likely that they are uninsured because they could not afford the health insurance premiums.

A multivariate analysis of the loss of health insurance upon retirement confirms that COBRA does, in fact, reduce the likelihood of becoming uninsured upon early retirement. Nonetheless, a primary reason for not taking advantage of the continuation mandate is low income.

Retirement Health Benefits Affect Early Retirement Behavior

The timing of retirement is known to be a function of both the health and the financial status of older workers. Given the importance of continued health insurance coverage in retirement for ensuring greater financial security, one would expect access to post-retirement health insurance to influence retirement behavior.

Figure 2. Probability of Retiring over a Two-Year Period for Men Age 60

SOURCE: Lynn A. Karoly and Jeannette A. Rogowski, "The Effect of Access to Post-Retirement Health Insurance on the Decision to Retire Early," Industrial and Labor Relations Review, Vol. 48, No. 1, October 1994, pp. 103—123.

By observing the retirement behavior for a sample of men in the SIPP from 1984 to 1988, Karoly and Rogowski document that health insurance is indeed an important determinant of early retirement among male workers. A multivariate analysis shows that for men ages 55 to 62, the offer of continued health insurance coverage through an employer had a positive effect on the likelihood of early retirement. For example, the mean predicted probability of retiring in the course of roughly a two-year period was 12 percent for 60-year-old men whose employers did not offer retiree health benefits, compared to 24 percent on average for men of the same age with employer-provided retiree health insurance (see Figure 2). According to this study, access to retiree health benefits increases the likelihood of early retirement for men, raising the baseline probability of their early retirement by 50 percent on average. Access to health insurance coverage through a spouse's employer also increases the likelihood of early retirement.

Another factor in retirement decisions is the generosity of the retiree health benefit plan. From preliminary work that extends their study to a cohort of men surveyed in the 1992 to 1994 Health and Retirement Study, Karoly and Rogowski find that both access to post-retirement health insurance and the extent of premium cost sharing affect labor force transitions more generally, with an increase in premium cost sharing dampening the incentive to leave full-time employment. The researchers also replicate their finding that coverage through a spouse can affect labor force transitions away from full-time employment.

Thus, it appears that retiree health benefits, like pension benefits, are a financial incentive that can influence the retirement decision. Moreover, this research suggests that access to continued health insurance coverage after retirement can serve as an inducement to retire early.

Policies Can Affect Access to Affordable Retiree Health Insurance and Early Retirement Behavior

According to this research, public policy initiatives have implications for retirement behavior if they alter the relative costs of health care pre- and post-retirement. Five recent or pending policy changes are especially likely to affect some employees' access to retiree health benefits and, therefore, to modify retirement incentives by either encouraging or discouraging early retirement. These policy initiatives have implications for the behavior of potential retirees because they change the financial incentives associated with early retirement. At the same time, the policy environment can affect employers' behavior and their willingness to offer retiree health benefits.

Two policies—COBRA and HIPAA—support a trend toward early retirement. Both COBRA and HIPAA make insurance coverage after retirement more accessible and affordable. Therefore, both policies could increase the propensity to retire early, because both allow workers to leave their employer before Medicare eligibility and still retain their health insurance coverage, for 18 months at the employer's group rate, and after that through the individual insurance market. HIPAA should have an effect similar to COBRA, but it is likely to be smaller in magnitude because premiums are likely to be higher than the employer's group rate once conversion into an individual policy occurs.

Another recent policy change may push retirement to older ages. Under new rules issued by the Financial Accounting Standards Board in 1993, employers are required to report their retiree health liability as a cost against current earnings. Before this change in accounting rules, these costs were not treated as a liability. Partly as a result, many employers have decreased the generosity of their offers of post-retirement health insurance, a reversal of a prior trend toward increased generosity of benefits. Some firms have eliminated retiree health benefits altogether. Consequently, fewer workers will be affected by a financial incentive to retire early.

Other proposed policy changes also have implications for future retirement behavior. To reduce the number of near elderly who are uninsured, the Clinton administration has proposed to extend Medicare coverage to individuals ages 62 to 64. Individuals who choose this option would pay annual premiums estimated at $3,600, a rate that far exceeds the highly subsidized Medicare premiums for retirees age 65 and over. (Individuals age 62 to 64 who begin Medicare coverage before age 65 would also pay a higher Medicare premium after age 65.) The Medicare buy-in option increases access to health insurance after retirement and is likely to induce early retirement. However, the size of the premium paid will also affect observed labor force transitions. The premium that is currently proposed will be actuarially fair, but it exceeds the typical cost of health insurance for an employer-provided policy as either an active-worker, retiree health benefit or a COBRA benefit or through a spouse (and the employer-provided plan may be more generous). Workers with these options, therefore, are not likely to face any new incentive to change their labor force behavior.

Depending upon the state regulatory environment, the Medicare buy-in premium may or may not exceed the cost of an individual plan under HIPAA provisions or in the individual market for workers without employment-based coverage. The Medicare buy-in option should be most attractive to older workers in poorer health who are insured by their own employer and who would exhaust their COBRA coverage prior to age 65 (or would not be eligible for COBRA because they work in a small firm). There may also be changes in labor force behavior for workers without employment-based coverage, because they might face lower insurance premiums under Medicare than on the individual market, thereby reducing their out-of-pocket health care expenses both as workers and retirees.

Concern over the financial solvency of the Medicare system has led others to propose extending the age of Medicare eligibility from 65 to 67. The longer it takes to reach Medicare eligibility, the smaller is the incentive to retire at an earlier age. In addition, employers offering retiree health benefits to early retirees will now cover their former employees for a longer period until Medicare coverage takes effect. This may further weaken employers' offers of generous retiree health benefits for early retirees. Thus, if this policy is enacted, older workers will face a longer wait for Medicare eligibility and the willingness of employers to provide continuation benefits will become even more important in their retirement planning.

Policymakers Need to Monitor Policy Effects on Retirement Behavior and Insurance Coverage

In sum, the policy environment will influence both what employers will do regarding retiree health benefits and what individuals will do regarding the retirement decision. Given a decision to retire early, policies that affect access to continued health benefits after retirement also have implications for whether an early retiree has health insurance coverage.

In the future, policymakers need to continue monitoring the effects of policies both on retiree behavior and on employer decisionmaking. If reforms make insurance more accessible and less costly after retirement and before Medicare eligibility, an increased incentive to retire early will exist. Also, policies that make affordable insurance available upon retirement should reduce the prevalence of the near elderly who are uninsured. At the same time, some policy initiatives may have the opposite effect if they reduce the health insurance coverage options of early retirees, either through changes in employer offers of retiree health benefits or through changes that affect insurance access by other mechanisms.


  • [1] The HIPAA guarantees are subject to certain requirements, such as having held the group coverage for a minimum amount of time.

This report is part of the RAND research brief series. RAND research briefs present policy-oriented summaries of individual published, peer-reviewed documents or of a body of published work.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.