Changes in the Composition of the Workforce and Their Effects on the Future of Employer-Sponsored Health Insurance

Changes in the Composition of the Workforce and Their Effects on the Future of Employer-Sponsored Health Insurance By Gregory Acs and Linda J. Blumberg In this paper, we examine how projected changes in the US economy and workforce may affect employer-sponsored health insurance (ESI) for workers. We develop a multivariate model of ESI coverage and use it to forecast coverage rates in 2008 under various assumptions about the future composition of the workforce. Our forecasts indicate that changes in the workforce are likely to lead to modest declines in ESI coverage rates between 1997 and 2008. However, the model does not take into account potential changes in the health insurance market, such as escalating premiums, which may drive down insurance coverage even more. Changes in the Composition of the Workforce and Their Effects on the Future of Employer-Sponsored Health Insurance Executive Summary Changes in the US workforce have potentially large consequences for workers’ health insurance coverage under employer-sponsored plans and health insurance coverage in general. In this paper, we examine how projected changes in the US economy and workforce may affect employer-sponsored health insurance (ESI) for workers. While research shows that changes in the workforce cannot explain all the changes in ESI (Acs 1995), we can anticipate how certain changes in the distribution of workers across jobs and in the demographic composition of the workforce may affect ESI coverage in the future. Initially, we examine recent trends in coverage rates across population sub-groups and explore how coverage rates may change (1) if sub-group specific coverage trends continue and (2) if sub-group composition changes (shift-share analysis). Then we develop a multivariate model of ESI coverage and use that model to forecast ESI coverage rates in 2008 under a variety of assumptions about the future composition of the workforce. We use two supplements to the Census Bureau's Current Population Survey (CPS) — the April 1993 Survey of Employee Benefits and the February 1997 Contingent Worker Supplement (CWS) — in addition to the March CPS files for each of those years. While the supplements are the core of our analysis, we link individual respondents with additional information about them obtained from the March survey. Our sample, therefore, consists of those workers who are included in both the supplement and the March survey for their respective years. Shift-Share Forecasts of ESI Coverage. Shift-share forecasts involve dividing the entire sample population of workers into mutually exclusive groups based on a small set of characteristics. For example, workers can be divided along gender lines, age brackets, wage rates, or job characteristics. We use eleven different sets of sub-groups which we discuss in two broad categories: demographically-based groupings and economicallybased groupings. Each of the eleven sets divides the population into two to eleven mutually exclusive and exhaustive sub-groups. Our first set of forecasts holds the composition of the population at 1997 levels and asks what would happen to ESI coverage if the trend in sub-group specific coverage probabilities between 1993 and 1997 continues to 2008. When we use the 1993 to 1997 trend in group-specific coverage probabilities to predict coverage in 2008, all eleven groupings yield forecasts of a modest decline in ESI coverage. Second, we provide an alternative set of forecasts in which we examine how ESI coverage may change if trends in the composition of the workforce between 1993 and 1997 continue to 2008 while group-specific coverage probabilities remain at their 1997