As health care costs rise, the workforce ages, and global competition intensifies, many employers seem to be concluding that they can no longer afford to offer subsidized health insurance to retirees.
Breadth of retiree coverage. In 2002, employer-sponsored retiree health insurance covered about 5.3 million adults ages 55–64 and about 10.5 million adults age 65 and older.1 For early retirees, employer-sponsored health benefits provide primary insurance coverage. Few affordable alternatives exist.
Although federal law requires employers with twenty or more employees to offer continuation coverage to separating workers, coverage must last for only eighteen months (or twenty-nine months for those with disabilities), and former employees must pay the full premium cost.
Private nongroup coverage is expensive, especially for those with health problems, and many insurers deny coverage for even relatively minor health problems.2 Adults younger than age sixty-five can qualify for Medicaid or Medicare benefits only if they are blind or disabled or have end-stage renal disease (ESRD). Medicaid recipients must also meet strict eligibility, income, and asset tests.
After age sixty-five, retiree health insurance supplements Medicare, defraying cost-sharing requirements and typically covering services excluded from the basic Medicare benefits package. For example, most retiree health plans cover prescription drugs, which Medicare did not cover until Part D was implemented in 2006.3 Beneficiaries with employer drug coverage can now avoid paying Part D premiums and still obtain help with drug expenses.
Erosion of coverage. Retiree health insurance offers dropped sharply about fifteen years ago. The share of private firms with 200 or more employees providing retiree health insurance fell from 66 percent in 1988 to 36 percent in 1993.4 Most analysts attribute this decline to a 1993 accounting rule requiring employers to recognize expected future retiree health care costs as liabilities on their balance sheets.5
Although fragmentary, the available evidence suggests that less erosion in access to retiree health benefits has occurred since 1993. Employer surveys by first KPMG and then the Health Research and Educational Trust (HRET), in partnership with the Henry J. Kaiser Family Foundation (KFF), show that between 1993 and 2003, the share of private firms with 200 or more employees providing retiree health insurance fluctuated between 40 percent and 35 percent. From 2003 to 2005, however, the share fell from 38 percent to 33 percent.
6 Data from Mercer Human Resources Consulting indicate that the share of employers with 500 or more workers offering health benefits to early retirees declined from 46 percent in 1993 to 29 percent in 2001 but remained roughly constant through 2005.
7 Hewitt Associates found that among employers with 1,000 or more workers, the share offering health benefits to early retirees fell from 88 percent in 1991 to 73 percent in 2000.8 Recent surveys of retirees and workers nearing retirement generate mixed results, with some showing a decline in coverage and offers and others showing little change.9
Recent developments. An important recent development in retiree health benefits is employers’ effort to shift costs to retirees. Employer surveys show that between 2004 and 2005, about seven in ten private firms offering retiree health benefits and employing 1,000 or more workers increased retiree contributions to premiums, and more than one-third increased cost-sharing requirements. Nearly two-thirds had caps in place in 2005, limiting firms’ future financial obligations for retiree health coverage.
10 Household surveys also reveal rapid premium increases for older adults with retiree health benefits. For example, median monthly premium contributions to employer health plans by Medicare beneficiaries age sixty-five and older in the Health and Retirement Study (HRS) more than tripled in inflation-adjusted dollars, increasing from $13 (in constant 2002 dollars) in 1998 to $50 in 2002.11
These studies have a number of limitations. The existing employer surveys are quite small. The 2005 Kaiser/HRET survey collected data from only 1,182 firms with 200 or more employees, and the Mercer survey collected data from fewer than 3,000 firms.
12 The focus on large employers ignores the many workers employed by small firms. Further, to assess the financial security of future generations of retirees, we need information on trends in the share of workers with access to retiree health benefits, which these surveys do not report. Self-reports of future retiree health benefits in worker surveys might not be reliable, because few workers are well informed about their future retirement benefits.13 Our study was designed to fill in some of these gaps.
Data source. Our data are from the Medical Expenditure Panel Survey Insurance Component (MEPS-IC), a nationally representative, repeated cross-sectional survey of private-sector establishments. We used data from 1997 to 2003, a period of rising health insurance costs when retiree plans came under strong financial pressure.
14 The sample size (approximately 25,000 establishments per year) is roughly ten times as large as the Kaiser/HRET survey sample. MEPS-IC has an average response rate of roughly 78 percent, which also compares well to other establishment surveys.15
MEPS-IC collects information on whether the establishment offers health insurance to active workers and retirees (younger than age sixty-five and age sixty-five and older), as well as data on premium levels and cost sharing. It also collects data on many establishment and firm characteristics.
One limitation of MEPS-IC is that occasional changes in the questionnaire limit the years during which we can compare some of our measures. For example, industry information is unavailable for 2003, and retiree premium information is consistent only between 2000 and 2003.
Methods. An important methodological issue is the appropriate unit of observation. Simply estimating the share of establishments offering retiree health benefits would not shed much light on workers’ future retirement security, because these estimates give as much weight to an establishment employing a handful of workers as to one employing thousands. The variation in retiree health insurance offers by firm size makes simple establishment-level estimates especially problematic.