What Can We Learn About Health Insurance From Benefit Plans Filed By Employers?

Summary

Estimates of the uninsured range between 20 and 40 million, depending on the definition. Over 60 percent of the adult nonelderly population cite the lack of health insurance through work as a key reason for being uninsured. Existing data provide important insights on the percent of workers who have health benefits available through work and who participate. However, we know much less about the firms themselves, particularly how their benefit offerings evolve over time, how costs of health benefits and cost-sharing with workers are changing within firms, and whether workers are aware that their employers offer benefits.

This study examines the potential of a new data set to answer questions like these, as well as to triangulate evidence on employer-provided benefits with other data sources. The data are based on benefit plans filed by employers within the Department of Labor’s 5500 data files, linked with the Census Bureau’s Longitudinal Employer Household Dynamics (LEHD) data.⁽¹⁾  This linkage is expected to improve the match between benefit plans and firms, add significantly more information on firm characteristics, and track firms longitudinally over time. The linked data also are matched to SIPP and CPS survey responses for employees within these firms.

The stand-alone 5500 data have many known limitations, but it is hoped that these linkages will overcome some of the issues, while also providing a richness of detail not available before on employers offering benefits. Existing ASPE work with LEHD provided a ready opportunity to explore these enhanced data. Assessing their quality and representativeness, as a first step in determining whether they merit further exploration, is the primary purpose of this report.

Research Questions

To better assess these data’s potential value, the project explores:

• What proportion of the health benefit plans on file in the 5500 data can be accurately linked to specific employers?
• Does the percent of firms offering health plans as observed in these data coincide with findings from other data sources?
• Are individual worker reports of access to employer-provided health benefits consistent with their employers’ filing of health plans?
• What types of research and policy questions are these data best suited to inform?

Results

As hoped, the Business Register increased our ability to match health plans in the 5500 data back to specific employers, with a match rate of roughly 90 percent. Most of the remaining ten percent were sponsored by firms that were defunct or nonprivate, or that already had been identified as sponsors elsewhere in the data. Thus, for the purpose of identifying offering firms within the private sector, these nonmatches appear of limited consequence.

However, in spite of the high match rate, the data reflect lower rates of employer-offered health benefits compared to what we observe in the Medical Expenditure Survey (MEPS), even for large firms. MEPS indicates that, for workers in firms with over 1,000 employees, 97 percent have employers who offer health benefits, compared to 86 percent in the 5500 data. Thus, the data do not appear well suited to measuring the prevalence of employer offered health plans.

Nonetheless, patterns of health plan offerings appear similar to those seen in other data. Offer rates are higher among older and larger firms and within particular industries, consistent with other data. Also, there is encouraging concordance between employer health plans in the 5500 data and their workers’ reports of access to employer-provided health benefits in survey data. The graph to the right shows that, among large private firms, roughly 84 percent of SIPP workers who reported employer-provided health benefits worked for an employer with a health plan on file in the 5500. This varied by industry, with a low of around 70 percent for services and a high of 94 percent for manufacturing.

These findings provide preliminary evidence that the 5500 data are representative of firms that provide health plans, at least for large private firms. Additional analyses, proposed in the full report, could give a more definitive assessment, and seem merited based on these results. Given the data’s detail on firms, longitudinal tracking and employee links, they have the potential to answer questions about employer health benefits not before possible, including:

• Are changes in benefit offers by industry or market sector due primarily to new firms entering the market, existing firms dying, or continuing firms changing their offers?
• Do changes in worker reports of employer benefits seem to primarily reflect changes in access, eligibility, or take-up?
• When an employer’s cost of health plans rises, how much of the increase is passed along to workers, and how does this relate to worker participation?
• How much of a worker’s reported lack of coverage is due to their lack of awareness of employers’ benefit offers?

Endnote

1.  Firms are required to file Form 5500 for all benefit plans with 100 or more participants, and for a few smaller plans. The LEHD data contain a number of survey and administrative data files on workers and employers, including the Census Business Register. The Business Register pertains to all private sector firms. Thus, the scope of these linked data are limited to larger, private sector firms.