War on welfare: The victors, the victims and a new welfare regime. Economic consequences of welfare reform

This dissertation consists of three papers. In the first paper, I investigate the effect of welfare reform on the employment, marriage rate and fertility of low-educated unmarried women. I find that welfare reform has induced low-educated unmarried women to move from welfare to work in significant numbers. I estimate that the imposition of time limits and other administrative reforms correlated with it increased the employment of unmarried women with 12 or fewer years of education by 484,239. This represents approximately 32% of the decline in welfare caseloads for this group since 1994. I also find that women who left welfare for employment worked approximately 30 hours per week, which even at low wages may significantly improve their financial status relative to public assistance. I find little evidence that welfare reform affected the fertility of low-educated unmarried women.;The second paper investigates the effect of the 1996 welfare reform on immigrants. Federal law banned legal non-citizens from receiving any federally financed means-tested benefits in their first five years in the US. However, a number of states restored some of these benefits. I study the effect of the state level variation in policies on the employment, hours worked and marriage rate of low educated women. The results suggest that welfare reform induced native-born citizens and foreign-born non-citizens to increase their employment. It appears to have had a larger effect on the least educated native-born women and among foreign-born, a larger effect on more recent immigrants.;In the final paper, I study the effect of safety-net programs on the locational choices of new immigrants. I use a conditional logit model to obtain estimates of the effect of the policies of interest. I control for unmeasured, time-invariant characteristics of states that may be correlated with locational choice and benefit availability. To control for time-varying, unmeasured state characteristics, I use a difference-in-differences procedure. I find that safety-net programs do not affect the locational choices of new low-skilled women immigrants. This result is robust to a variety of specifications including those limited to the target group; difference-indifferences models with two different control groups; and models that take account of the number of programs available to immigrants and generosity of benefits.