Academic journal article Economic Inquiry

Cross-State Variation in Medicaid Programs and Female Labor Supply

Academic journal article Economic Inquiry

Cross-State Variation in Medicaid Programs and Female Labor Supply

Article excerpt


Using a pooled cross-section data set from the 1980 through 1993 Current Population Survey March Supplements, we test if different Medicaid benefit levels across states impact the labor supply behavior of female heads of households. The ordinary least square (OLS) results support the prediction that Medicaid expenditures reduce labor supply. Controlling for state fixed or random effects alters the effect of both AFDC and Medicaid on the decision to participate as well as the number of hours worked. We also find that while the effects of program generosity are sensitive to the inclusion of state effects those of variation in eligibility thresholds are not. (JEL J22)


The issue of the design and effects of government policies aimed at providing support for children is at the center of an ongoing national debate. Social scientists, policy makers, and the public at large have all focused their attention on the benefits and, perhaps even more so, the costs to society of this support network. These costs are perceived to take not just monetary forms but also to have adverse incentive effects in terms of work, out-of-wedlock births, divorce, etc.

Much of this debate has focused on the adverse effects of the cash components of welfare (Aid for Families with Dependent Children [AFDC] and Food Stamps), but recently other programs have become embroiled in this debate. As the costs of health care have risen, government-provided Medicaid benefits have also come under increasing scrutiny. Medicaid expenditures are one of the fastest rising components of state budgets. State Medicaid spending rose by 80% in real terms between 1980 and 1990, rising from $19.1 billion to $33.7 billion. Further, state Medicaid expenditures increased even more dramatically in the past few years, rising about 19% in 1990-91 and 22.5% in 1991-92. Up until the mid-1980s, Medicaid coverage was primarily limited to recipients of AFDC and indigent senior citizens. Over the last ten years however, Medicaid coverage has been altered to include individuals with incomes above the AFDC maximum, primarily pregnant women or children whose parents have incomes at or slightly above the poverty line. Despite the fact that Medicaid is a federally mandated program, there is a large amount of latitude as to the number and types of individuals and services that each state can choose to cover. This state freedom implies that there are basically 51 different Medicaid programs in the United States (Washington D.C. also participates in Medicaid). Because of this, there is a large variance in per-recipient Medicaid expenditures across states. In 1992, for example, annual per-recipient Medicaid expenditures for AFDC adults ranged from a high of $2,937 in Louisiana to a low of $575 in Mississippi. [1]

Recently, economists and other social scientists have used the cross-state variation in these programs as a form of "natural experiment" that allows them to access the effectiveness and adverse incentive effects of these programs. In this article we use this approach to add to the literature on the effects of these programs on labor supply. First, we construct three new measures of the value of Medicaid benefits to adults, children, and families which vary across states and time. These alternate measures will allow us to test for the sensitivity of the conclusions from previous studies to the measure of Medicaid value used. Previous work on Medicaid and female labor supply has used measures of state level Medicaid spending which included spending on the elderly and disabled. Since these groups account for a substantial share of all Medicaid expenditures, and they are not randomly distributed across states, this has the potential to bias the reported results. Second, because there is substantial variation acr oss states in their labor market characteristics, demographic mixes, and other factors which are potentially correlated with both labor supply and benefit generosity, our analysis checks for the sensitivity of the results to the inclusion of state fixed and random effects. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.