Health Insurance Coverage in Tax and Survey Data

Abstract

The Current Population Survey provides official estimates of the number of people covered by health insurance and the number of uninsured in the United States. This type of survey data are also used to study the effects of policy changes on health insurance coverage. However, there is evidence that individuals sometimes misreport health insurance coverage, which might bias findings that use survey data. We use new administrative health insurance information from tax data to evaluate health insurance coverage in survey data across several dimensions, including age, income, and state. Our main findings suggest that although overall coverage counts are similar between survey and administrative data across all demographic characteristics, coverage rates and uninsured counts differ because of differences in population size. These similarities mask coverage differences by insurance type. Medicaid coverage is very well reported in tax data, whereas surveys tend to underreport it, especially for low-income individuals and people under the age of 40. Employer-sponsored coverage counts are higher in survey data than in administrative data. Finally, this study provides researchers that use survey data a benchmark for how to adjust Medicaid coverage to align with administratively reported levels.