welfare dynamics under time limits jeffrey grogger charles michalopoulos by: tien ho
TRANSCRIPT
Welfare Dynamics Under Time Limits
Jeffrey GroggerCharles Michalopoulos
By: Tien Ho
Introduction
Prior to 1996: AFDC PWRORA of 1996: TANF replaced AFDC Eligible families had child younger than 18 Time Limits imposed:
-Federal: 5 years
-State: varied
Introduction
The Controversy Over Time Limits Pro: direct route of getting people off welfare;
people who need to preserve welfare look for jobs
Cons: kids lose benefits when parents do; low-income job or short-term joblessness
Lang (2007)
Main Issue
How did time limits affect a family’s decision to stay on welfare?
Did time limits reduce welfare use?
Why is this interesting?
Are time limits an effective measure?
Why do they succeed? Why do they fail?
Time limits: cruel policy or “tough love”?
Previous Studies Council of Economic Advisers. Technical Report: explaining the
decline of welfare receipt, 1993-1996. Washington: Council Econ. Advisers, May 1997.
--mentioned time limits as a possible factor in reduction of caseloads
Swann, CA. “Welfare reform when agents are forward-looking.” Manuscript. Charlottesville: Univ. Virginia, December 1998.
--suggested that time limits motivate people to preserve their benefits
Moffitt, RA. Incentive effects of the U.S. Welfare System: a review. J Econ. Literature 30 (March 1992):1-61.
--aggregate state-level caseload analysis
Improvements
Consider the age-dependence issue-claimed this is essential to
understanding time limits
Used a random, “natural” experiment
In depth analysis of data from Florida program
Florida Family Transition Program
Location: Escambia County in May 1994 under waiver Randomized into two groups: experimental (time limit)
or AFDC (no time limit)
-new families: randomized when entering
-previous families: randomized at renewal
Followed families for 2 years after random assignment
Data collected from administrative records and survey
Key Assumptions
The effects of individual reforms are additive The effects of the financial work incentives
and enhanced services are age-invariant Time limits have no effect on parents with
children above threshold age Parents with younger children are forward-
looking, expected-utility-max consumers Prediction: parents with younger children
should reduce welfare consumption more than parents with older children
+0.25%
-8.3%
+17.5%
+27.5%
-27.75%
-35.8%
-10%
Are financial incentives and enhanced services truly age-invariant?
There are a lot more mothers with younger children on welfare versus mothers with older children which suggests there are differences between them.
Limitations
Support uses tables with different age groups Control and treatment are not the same Mothers with younger children may be affected
differently by financial incentives and enhanced services than mothers with older children
Large differences in age groups (no justification for why they chose that grouping)
Regression Estimates
Variable Definition
i=1,…,n n is number of persons in sample
t= 1,…,24 Time measured in month
Ajit =1 If youngest child in ith family at time of random assignment falls into age group j; j=0,1,2,3,4; otherwise Ajit =0
Ei =1 if in experimental; =0 if in AFDC
Xit Exogenous regressors: mother’s age @ t; # of children in family, mother’s years of schooling, # of months of welfare receipt prior to randomization; # of quarters of employment prior; vector of year dummies; dummy for black (1) or not (0); dummy for 3y time limit (1) or 2y limit (0)
Joint effect of financial work incentives and enhanced services
“Linear Interaction” Model
What is the effect of prior welfare use?
Follow-up results and additional tests
Linear Regression
Only takes into account observable characteristics
Experimenter bias; didn’t consider other factors-subsidized daycare, etc.
Other unobservable factors not accounted for-individual heterogeneity
Implications
Families with younger children appear to be affected by time limits (16% reduction overall)
Imposing time limits may encourage people to find a job => leading to reductions in welfare payments
Could have adverse results on younger children (education, parental care)
The Bad Too many assumptions (age-invariance, forward-looking
consumer, etc.) People knew they were in a study (may not have actually
believed their benefits would stop) Florida Data very unreliable and not generalizable; social
workers had smaller caseloads and could be more proactive
Subsidized daycare for children 12 years and younger may have had a greater effect on mothers with younger children
Does not tell us how time limits would effect entry into welfare
Time limit of the FTP program different from other time limits
Lang (2007); Fang and Keane (2004)
The Good Offered age-dependent analysis of time limits Significant results from data analysis Accounted for many factors (age, race, etc.) Defended assumptions with data from other
studies Data from two nationwide surveys similar to data
here (Grogger 2002, 2004)-relative to states without limits, welfare
participation rates dropped more rapidly among households with younger children than in homes with older children
Discussion
Do you think time limits actually work (in an ultimately beneficial way)?
If time limits work, should the federal government require all states to impose them?
What problems/critiques do you have with this study? Any ideas for a better one?