Since 1999, the Employer Health Benefits Survey has documented trends in the employer-sponsored health insurance market. Every year, about two thousand private and non-federal public employers with three or more employees have completed the full survey. Among other topics, the survey asks firms for the premium or full per-person cost of their health coverage as well as the share that workers are responsible for. The graphing tool below allows users to look at changes in premiums and worker contributions over time for covered workers at different types of firms.

Findings from the 2018 survey and supplemental information are available here. For more information on the survey methodology, see the Survey Design and Methods section. For additional questions on the Employer Health Benefits Survey or this tool, please go to the Contact Us page and choose “TOPIC: Health Costs.”

Standard Errors (SE): Like all surveys, there is a degree of uncertainty in every estimate in the Employer Health Benefits Survey. As an analysis examines more specific groups, uncertainty tends to increase. Standard Errors (SE) are a measure of how much uncertainty there is in an estimate and should be used in statistical tests to determine whether the difference between any two numbers is significant. Often even large differences in premiums or worker contributions between two groups are not actually meaningfully different. Standard errors are available for each data point in the “Export Table Data” download link above.

Not Sufficient Data (NSD): In cases in which there are too few firms in a sub-population to provide a reasonable estimate and/or protect respondent confidentiality, the abbreviation NSD is used.

Weights: In order to ensure that estimates are nationally representative, firms are selected randomly and weights are applied. Premium and worker contribution estimates are weighted to the number of workers covered by health benefits and adjusted to the number of employees in industry and firm size categories. For more information, see the Survey Design and Methods section.

Variable Definitions: Family coverage refers to a family of four. Employers offering self-funded or partially self-funded plans bear at least some of the financial risk of covering their employees directly. Employers sponsoring self-funded plans typically contract with a third-party administrator or insurer to provide administrative services for plans. In some cases, the employer may buy stop-loss coverage from an insurer to protect the employer against very large claims. For more information on self-funding, see Section 10. Firms with multiple plan types are defined as self-funded or fully insured based on the characteristics of their largest plan type; however, premiums are calculated as a weighted average of up to four plan types. Therefore, the premiums of both self-funded and fully insured plans may be included in the average premium and worker contribution for some firms. Between 2009 and 2012, respondents were asked if their largest conventional plan was self-funded, but no premium information was collected.

Industry classifications are based on a firm’s primary Standard Industrial Classification (SIC) code as determined by Dun and Bradstreet. A firm’s region is determined by the location of its primary location, according to the U.S. Census Bureau definitions. Firm ownership classifications are reported by survey participants.

Firms with Many Lower-Wage or Higher-Wage Workers:  Since 2013, thresholds are based on the 25th and 75th percentile of workers’ earnings as reported by the Bureau of Labor Statistics (BLS) using data from the Occupational Employment Statistics (OES) (2017). The cutoffs were inflation-adjusted and rounded to the nearest thousand. From 2007 to 2012, wage cuts offs were calculated using the now-eliminated National Compensation Survey. BLS OES data are available at: http://www.bls.gov/oes/oes_emp.htm#scope (Accessed: August 2018). Higher-wage firms are those where at least 35% of workers earn more than the cutoff. Lower-wage firms are those where at least 35% of workers earn less than the cutoff. To reduce the survey burden on respondents, in some cases, the survey instrument only included questions on higher-wage workers.