Understanding Persistent Differences in Insurance Coverage between Poor Adolescents and Their Younger Peers in the United States

Presenter: Lindsey Leininger, University of Wisconsin-Madison

Abstract

Rationale: Uninsurance rates among poor teenagers have historically been higher than the analogous rates for poor younger children. The reform efforts enacted with the creation of the State Children’s Health Insurance Program (SCHIP) in 1997 held great promise of reducing the disparity in coverage between teenagers and younger children; SCHIP in concert with the concurrent expansions to Medicaid vastly expanded the number of low-income teenagers eligible for public health insurance coverage. And while these efforts did indeed increase the proportion of publicly insured low-income teenagers while simultaneously decreasing the proportion of uninsured poor teenagers, it is not the case that the disparity in coverage rates between poor teens and their younger counterparts was eliminated. Illustratively, data from the 2005 Medical Expenditure Panel Survey (MEPS), demonstrate that teenagers ages 13-17 with family incomes below the federal poverty level (FPL) spent an average of 2.
0 months without insurance coverage over the course of the year while poor younger children spent 1.2 months uninsured on average, reflecting a difference of 40%. The magnitude of the gap is striking given that public insurance eligibility has been harmonized across poor children of all ages since the beginning of the decade.

Objective: This work seeks to provide new evidence regarding the disparity in health insurance coverage between poor children of different age groups using a regression decomposition approach.

Data and Methods: The dependent variable of interest is insurance status, which is measured as follows: the number of months spent with public coverage, the number of months with private coverage, and the number of months spent uninsured. The primary data source is the 2005 round of the MEPS; the 2006 round will also be used pending data release. The MEPS data contain information on health insurance coverage and a wide array of socioeconomic indicators. Linked citizenship status data from the National Health Interview Survey will also be included in the analyses. The sample will include children ages 1-17 living in families with incomes less than 100% FPL (n=2,806 in the 2005 MEPS), qualifying all of them for eligibility under Medicaid per federal mandate. The Oaxaca-Blinder approach will be employed to decompose the gaps in coverage into the portion explained by the differences in observable characteristics and the portion driven by unobservable differences between poor tee
ns and their younger peers.

Conclusions: Wide-ranging efforts at the state and federal levels to increase insurance coverage among poor teenagers have enjoyed considerable success. Nonetheless, worrying disparities in uninsurance between poor teenagers and younger children remain. The aim of this study is to contribute to the literature that seeks to understand the nature and determinants of these persistent disparities.