Association for Public Policy Analysis & Management

Three states have obtained federal waivers that permit them to have work requirements for some Medicaid beneficiaries. Seven more states have applied for waivers to do so. While other studies have used cross-sectional data to examine the health and employment status of Medicaid beneficiaries, this paper provides policymakers with information about the complex dynamics of the relationship between employment, Medicaid enrollment, and health status. We use the nationally representative Medical Expenditure Panel Survey (MEPS)-Household Component two-year longitudinal data for 2014-2015 to describe these relationships for adults ages 18 through 64 who were not receiving disability-related income or covered by Medicare.

First, we describe the association between health and work for beneficiaries that were continuously enrolled in Medicaid throughout the two-year period. In preliminary results, we find that approximately forty percent worked for pay during each of the 5 rounds of MEPS data collection in the 2-year period, 32% worked some of the rounds, and 28% did not work at all. We also find that those who did not work at all reported worse health along a number of dimensions than those who worked in all of the rounds. For example, compared with Medicaid beneficiaries who worked during all 5 rounds, those who did not work were more likely to have major activity limitations (an additional 27 percentage points) and reports of fair or poor physical or mental health status (an additional 28 percentage points).

Second, for adults who gained Medicaid during their two years in the MEPS-HC, we measure the extent to which negative changes in employment and health status preceded their enrollment in Medicaid. In order to provide a sufficiently long look-back period to study the relationship between (1) changes in employment and health status and (2) subsequent Medicaid enrollment, we focus on those who gained Medicaid in the second round of MEPS or later. In preliminary analyses, we find that prior to Medicaid enrollment, 8% of enrollees had lost a job, 8% had a decline in their perceived health status and reported being in fair or poor health, and 9% had a decline in their perceived mental health status and reported being in fair or poor mental health.

Third, we use monthly insurance data to classify the health insurance status of enrollees prior to their Medicaid enrollment and examine the sensitivity of our results to different look-back periods. In preliminary analyses conducted for adults enrolling in Medicaid in June of their first year in the MEPS or after, we find that 60% were continuously uninsured in the six months prior to their enrollment, 28% were covered by employer-sponsored insurance at some point prior to their enrollment, and 12% had some other type of coverage. The estimates in this paper will help inform policymakers about the health and work characteristics of the population likely to be subject to Medicaid work requirements where they are incorporated into the program.