Z:\grogger3\ddrive\Duncan\Results\Work Conditioned Welfare\APSA Version\Work Conditioned Welfare 7.doc
Printed On: 08/14/03
2
Reforms to state and federal welfare programs during the last decade have substantially
altered the way in which the U.S. provides transfer payments to the poor. One of the most
important changes in recent years has been the increase in the degree to which state benefits
encourage or are explicitly conditioned on employment. The Earned Income Tax Credit (EITC),
for example, which only provides benefits to people who work, has outstripped more traditional
forms of social assistance in recent years and has become the largest transfer program for the poor.
Reforms to more traditional aid programs in the 1990s also increasingly required work as a condition
of aid and changed the structure of benefits to increase work incentives.
That said, the literature has not properly explained the causes of this trend. Empirical
studies of trends in U.S. welfare policy generally focus on measures of program generosity or
spending, which do not fully capture important aspects of recent reforms. And the comparative
political economy literature has some interesting predictions about what factors affect demand for
welfare policy, but these studies have not taken advantage of variations in state policy that could be
used to test these predictions. This leaves an important gap in our understanding of what factors
have been behind recent shifts in U.S. welfare policy.
In this paper, I pursue two different empirical approaches to determine what caused
policymakers to increase work incentives in U.S. welfare benefits. The first, more traditional
approach, is a difference model that compares changes in AFDC/TANF maximum benefits to
changes in inequality, unemployment, government ideology, and other variables between 1980 and
1999. The second, more novel, approach directly examines how these factors affected a specific
aspect of U.S. welfare policy in the late 1990s, the fraction of earnings that welfare recipients can
keep when they work. Both these measures, I argue, provide information on the degree to which
states were concerned about work incentives.
Convention