Modeling Employer Decisions to Offer Health Insurance: Results using Synthetic Firms based on Newly Available Tax Data

Federal policymakers consider a variety of proposals to encourage Americans to have health coverage and are interested in how those proposals would affect insurance coverage and its cost to the federal government. One of the main tools used by the Congressional Budget Office (CBO) and the Joint Committee on Taxation (JCT) to estimate the cost and coverage effects of such proposals is a newly revised health insurance simulation model (HISIM) that models individual decisions to enroll in coverage based on their expected utility from available options using data from the Current Population Survey (CPS). Since more than half of nonelderly people in the U.S. obtain health insurance coverage through their employers, simulating employers’ decisions to offer coverage and their responses to various policy proposals is an important part of HISIM.

Employer responses to new policies are assumed to be influenced by the characteristics of their workforce. However, data that fully describe firm-level employee characteristics are not readily available in the CPS. This paper describes how CBO has improved its method and data source for building synthetic firms for each worker in the CPS data. Synthetic firms consist of a set of coworkers drawn randomly to match certain traits. The realism of the synthetic firms is enhanced by using an analysis of recently available tax data on how workforce characteristics vary by firm. In 2015, we observe over 14 million workers with their W2s matched to Form 941. We drew coworkers based on wage, age, offer status, and firm size.

The preferences and characteristics of an individual worker’s coworkers can influence whether her employer will offer insurance, at what price, and how an employer will respond to changes in policy. The new method of building synthetic firms allows employers with high concentrations of young or low income workers to respond differently than other employers to certain policies. This advance is used to illustrate how employers of different sizes and workforce characteristics are projected to change their decisions to offer coverage in future years.