|
Abstract:
|
From 1992 to 2017, the Current Population Survey Annual Social and Economic Supplement (CPS ASEC) included modelled estimates of employer contributions to health insurance, based on a statistical data match from the 1990s. Using different methods, recent research has shown that these estimates undervalued the amount employers contributed to their employee’s health insurance, which in turn means that total compensation estimates are understated (Janicki, O’Hara and Zawacki 2013; Berchick and O’Hara 2017). In this paper, we revisit these methods and propose a new strategy to estimate employer contributions to health insurance premiums. We replicate existing imputation methods, making use of CPS ASEC and Medical Expenditure Panel Survey data to compare estimates across more recent survey data. We then employ a Bayesian imputation procedure which uses aggregate statistics to estimate employer contributions. The paper outlines the specifics of model selection and data sources, with the goal of producing a reliable measure of employer contributions of health insurance premiums for the annual CPS ASEC release.
|
