What Explains the Gap in Welfare Use Among Immigrants and Natives?

Background: The 1996 welfare reform and subsequent state and federal changes in immigrant eligibility for welfare programs spawned a large body of research on the role of welfare policies but only a few examined the trends in welfare program participation among native and immigrant households. We contribute to this literature by documenting long-term trends in safety net and public health care programs (TANF, SNAP, SSI, Medicaid or SCHIP) participation from 1994 to 2017, covering various stages of the business cycles to study the differences in welfare program participation among immigrant and native families. We also use an Oaxaca-Blinder decomposition analysis to explain the gap in the usage of these programs.

Objectives: This paper aims to investigates the factors that explain the variation in long-term trends in receipt of safety net and public health programs between foreign-born and US-born households: Specifically we study to what extent the gap in receipt of safety net and public health care programs is due to differences in demographic characteristics of the two groups; to what extent due to the business cycle and its diverse impact on the two groups; to what extent on account of changes in immigration enforcement, to what extent is it due to welfare policy; and to what extent due to differences in migration patterns captured by the geographic locations of the two groups. 

Methods: Our main data source for individual level variables are from the Current Population Survey from 1995-2018 that provide welfare use data for 1994-2017. First, we use logistic regression to study the gap in welfare use among immigrants and native families. We sequentially control for a rich set of variables capturing state welfare policy variations, household characteristics, state-level economy, local immigration enforcement policies, geographic and business cycle factors. Next, we apply Oaxaca-Blinder decomposition method to study to what extent the gap in welfare use between the two groups over years is on account of differences in demographic characteristics, macro-economic trends and welfare policy variations, local enforcement policies.

Results: Prior to reform, foreign-born families were more likely than similar US-born families to receive assistance. Both groups experienced sizable reductions in receiving assistance under these programs in the following years, with the decline being greater among the foreign-born, such that in the early years of the 2000s the gap between the two groups almost disappeared. Use of these programs inched upward for both groups, followed by a sharp rise during the Great Recession, and a modest fall in the recent recovery. After adjusting for household and state characteristics, the swings remain large, but the gap turns negative across all years. The usages for all programs among immigrants are relatively higher in more generous states for all programs except for TANF. The decomposition results suggest that immigrant households would use less safety net and public health care programs if they had the same endowments in household characteristics as native households. And state policy variations significantly attribute to differences in program participation between immigrant and native households.