Funded Research
Welfare Reform and Migration of Poor Families
Source: NSF
Active: 03/15/03 - 09/30/07
Investigator(s):
Gordon F. De Jong
Deborah Graefe
Has devolution of welfare policy based on the 1996 welfare reform act created "welfare magnet" states where state policies provide more generous benefits and lenient participation requirements? Have welfare disincentive states with more restrictive policies resulted in increased out-migration of welfare poor families?
Our study encompasses the data preparation and initial state-level analyses (Phase I) in a longer-term project to answer these questions and covers the first year's work in the overall research plan. That larger project covers a period of three to four years and uses merged data from three sources - the 1996 and 2001 panels of the Survey of Income and Program Participation (SIPP), the Urban Institute's Welfare Rules Database, and a local labor market characteristics file created from decennial census and Current Population Survey data - in a longitudinal, two-stage specification of welfare-benefit "push" and "pull" impacts on poor families' migration behavior. Based upon a state welfare policy inequality framework, we use factor analysis to develop measures from textual policy manual materials to operationalize welfare benefit and eligibility rule dimensions for the post-1996 welfare reform implementation period and use these measures to test hypothesized state program effects on migration.
In the Phase I study proposed here, we use discrete-time event history analysis to predict interstate migration events in the SIPP data. Our hierarchical modeling strategy considers an integrated, and previously untested, micro-macro analysis of two determinant-of-migration hypotheses for welfare poor families. These tests evaluate effects of 1) time-varying state welfare policy characteristics and 2) individual and family characteristics, including detailed migration, work, and welfare participation histories and network ties, from the information-rich SIPP files. Following Frey et al. (1996), we separately analyze push and pull migration effects of our hypothesized co-variates through, first, a "destination model" for identifying pull effects, and then, a "departure model" which identifies push effects for potential migrants' origin locations. The combination of destination and departure model vectors for state welfare policy and individual and household indicators will ultimately provide a strong test, giving new evidence on the "salience of benefit variation to subjects" thesis (Shram and Voss 1999) regarding the welfare policy impact on migration.
The Phase I study we are proposing involves four components: 1) conceptualization and measurement of state welfare policies using the WRD textual measures; 2) preparation of an event history file documenting annual state welfare policy indicators for the 1996 through 1999 period; 3) preparation of a person-month-based event history file using the 1996 Panel (all waves and pertinent topical modules) of the public-use SIPP files; and 4) merging of these two event history files for analyses testing our hypotheses regarding the influence of lenient versus stringent state-level benefits on interstate migration. This analysis will yield preliminary estimates of interstate migration among the welfare poor which will be presented at professional meetings in September 2002. The second, subsequent phase of this study, not included in this proposal, comprises 1) preparation of the internal SIPP files (restricted-use data for which we are currently seeking permission for access and which provides individual geographic indicators of local labor market location) to predict intra- as well as interstate migration of welfare poor families and 2) preparation of a local labor market area database for merging with the restricted-use SIPP event history file. Thus, Phase I is a first step in the overall analysis plan.
