ACF OPRE: Achieving Change for Texans

Background

Achieving Change for Texans Demonstration

The ACT demonstration included the following components:

State time limits. Adult TANF recipients were assigned to one of three time limit tiers (12, 24, or 36 months) based on their educational attainment and work history. State time limit policies took effect when adult TANF caretakers were offered a slot in the Choices program and did not apply to families who were exempt from Choices participation requirements. After reaching state time limits, caretakers could receive twelve months of transitional Medicaid and child care.

Expanded TANF eligibility rules. ACT expanded TANF eligibility for certain low-income families. When determining eligibility for cash assistance, the old AFDC work history and 100-hour work rules for two-parent families were eliminated, resource limits were increased and children’s income and resources were disregarded.

Personal responsibility agreement. Caretakers were required to sign a personal responsibility agreement (PRA) as a condition of TANF eligibility. In counties with a Choices program, state time limits also applied. The PRA states that the client will:

cooperate with child support requirements,
provide immunizations and regular health checkups for children,
refrain from voluntarily quitting a paying job,
participate in an employment services program,
refrain from drug use and alcohol abuse,
provide proof of school attendance for each dependent child, and
participate in parenting skills classes if referred.

Persons not complying with the PRA received financial penalties that resulted in smaller monthly TANF checks but were not removed from TANF, as would occur under some sanction policies.

The demonstration was designed so that the impacts of these policies could be measured separately and in combination with each other. The TL pilot tested the impacts of state time limits while holding all other policies constant. The RER Choices pilot measured the combination of state time limits, expanded TANF eligibility rules and the PRA against old AFDC program rules. The RER Non-Choices pilot operated in counties without a Choices program, meaning that no TANF recipients were subject to state time limits. Thus, this pilot only measured the effects of expanded TANF eligibility and the PRA against old AFDC rules. Table 1 summarizes the specific policies applicable to experimentals and controls in each pilot.

One-time cash payments in lieu of TANF. The One-Time program allowed TANF-eligible families who also met certain crisis criteria (typically, recent loss of employment) to obtain a one-time payment of $1,000 instead of TANF. Applicants accepting this payment were not eligible to receive TANF in Texas for twelve months. This program was intended to address short-term crises within households and to prevent these families from becoming dependent on welfare. Unlike the other ACT components, this benefit was available to all eligible families and was not designed as a randomized experiment.

Table 1:
ACT Evaluation Research Design
Waiver Component	Provisions and Start Dates	Time Limits Pilot		RER Choices Pilot		RER non-Choices Pilot
Waiver Component	Provisions and Start Dates	Exp	Control	Exp	Control	Exp	Control
State Time Limits and Transitional Benefits	Institute state time limits (6/96 in Bexar County; 1/97 in other counties)	X		X
	Provide transitional Medicaid and child care after state time limits are reached (6/96 in Bexar County; 1/97 in other counties)	X		X
	Sign and comply with the Personal Responsibility Agreement (6/96)	X	X	X		X
	Disregard child's income and resources (11/96)	X	X	X		X
	Raise resource limits (11/96)	X	X	X		X
Responsibilities, Employment, and Resources (RER)	Delete work history, 100-hour rule for two-parent families (10/96)	X	X	X		X
	Change age of youngest child exemption for participating in Choices (10/96)	X	X	X
	Increase transitional Medicaid and child care from 12 months to 18 months for caretakers and second parents who were exempt from Choices but volunteered (10/96)	X	X	X

Source: ACT Process Evaluation

The Evaluation

As a condition of the federal waiver approval, Texas was required to design and implement a comprehensive evaluation of ACT. As described above, TDHS implemented the demonstration with experimental and control groups in a number of offices across the state so that a formal evaluation could be conducted. Texas received federal funding to continue studying waiver activities after the passage of PRWORA.

The major objectives of the ACT evaluation were: to document how well this demonstration was implemented; to determine the impacts of the ACT waiver policies on a number of public assistance, economic and child outcomes; and to report participants’ views of welfare reform. To accomplish these objectives, the evaluation included three major components:

A process evaluation that described the demonstration and its implementation.
An impact analysis to compare differences in outcomes for cases assigned to either experimental or control groups in each pilot. Ray Marshall Center researchers analyzed administrative data from TDHS and six other state agencies to assess differences in employment and earnings, receipt of welfare and other types of public assistance, participation in employment training programs, use of child care, child support and several other family and child outcomes. Outcomes were tracked for five years from the beginning of the demonstration. Results were computed for all cases in the experiment, as well as selected subgroups.
Intensive interviews with participants to provide in-depth understanding of clients’ experiences with state time limits and the TANF One-Time program. The interviews focused on the clients’ economic and family situations.

Separate reports are available from each of these analyses. This summary report synthesizes results from all three analyses, places their findings in context with those from other states and develops policy implications that are relevant for the post-ACT waiver environment.

The Context

National context. ACT was one of the last state AFDC waivers approved prior to the passage of PRWORA in August 1996. Some PRWORA provisions did not apply in Texas until the ACT waiver expired in March 2002. Approximately 20 states evaluated their waiver initiatives in some way, but only six of the other state waivers both included randomized impact evaluations and either time limit or personal responsibility provisions (Table 2.) Most evaluations of the earlier AFDC waivers have now been completed, which makes it possible to compare the results from the ACT demonstration with well-crafted research studies occurring in other states.

Table 2:
Randomized State AFDC Waiver Evaluations with Time Limit or Personal Responsibility Provisions
State	Key Experimental Group Provisions	Control Group
Arizona	Time limits, stricter JOBS sanctions for noncompliance, and elimination of the 100 hour rule family benefit cap, restricted eligibility for unwed minor parents, mandatory JOBS participation for teen parents, extended transitional benefits, and individual development accounts	Old AFDC rules
Connecticut	Time limits, earned income disregard and mandatory workforce services	Old AFDC rules
Florida	Time limits, financial work incentives, enhanced workforce services and requirements, and parental responsibility mandates	Old AFDC rules
Iowa**	Earned income disregard, Family Investment Agreement, participation in PROMISE JOBS, various provisions which allow two-parent families to qualify for assistance	Old AFDC rules
Indiana	Time limits, financial work incentives, work first approach, personal responsibility agreements, and sanctions for non-compliance	Old AFDC rules
Virginia	Time limits, personal responsibility agreements, earned income disregard, mandatory job search and full family sanctions for non-compliance	Old AFDC rules

Notes:
* Wisconsin’s evaluation also contained a randomized evaluation but its implementation was fraught with design and implementation problems so is not included in this table.
** Evaluation was discontinued after the implementation of TANF. Source: www.acf.hhs.gov/programs/opre/ssfpfp.htm.

During the operation of the ACT demonstration, time limit and personal responsibility policies in other states have continued to evolve. Currently, almost all states have some sort of time-limited TANF benefits, although the specific provisions vary widely across states. Most states also impose sanctions for failure to comply with TANF requirements. Some use partial sanctions that are similar to Texas’ financial penalties and only affect the caretakers’ portion of TANF payments. Others remove entire families from TANF (known as full-family sanctions) for failure to comply with workforce, child support or personal responsibility requirements. Almost all states now have stricter sanction policies than required by PRWORA.3

Texas context. Texas enacted additional welfare reform policies during the operation of the ACT demonstration. Because these policies apply to both experimental and control group members in ACT pilot sites, they are not included in impacts measured by the ACT evaluation. A special analysis was conducted to determine whether impacts for persons entering the ACT demonstration after these policies were enacted varied from those of persons entering TANF earlier in the demonstration (described in the section on subgroup analyses).

Texas welfare reform policies not included in the ACT demonstration include:

Texas Works, implemented in November 1997, which diverts potential TANF applicants to other activities and encourages them to choose work over welfare.4
Work First program, implemented in December 1997, which requires non-exempt TANF applicants to attend a workforce orientation session before they are certified to receive TANF.
A federal 60-month, lifetime time limit adopted in December 1999 for all persons on TANF cases headed by an adult (including all clients in ACT pilots). Any TANF benefits that a caretaker or second parent received for October 1999 or later are counted toward the federal 60-month time limit. State time limit months accrued after November 1996 count toward the federal limit.
Increase of the earned income disregard for TANF recipients, beginning in March 2000. This, along with related policy changes, allows TANF recipients to be employed for a longer period of time before becoming ineligible for TANF.
Replacement of the TANF program for two-parent families with a new state-funded program in October 2001.

In addition to the continuing policy changes throughout this demonstration, there was widespread media coverage about welfare reform, particularly during the early years of this demonstration. This publicity could have led some persons in the experiment to erroneously believe that welfare reform provisions applied to them. Such contamination by news media and word of mouth may have resulted in smaller impacts from this experiment than otherwise expected.

ACT’s Policy Significance

ACT is the only AFDC waiver evaluation designed to isolate the effects of state time limits separately from other welfare reform interventions. ACT is also the only waiver experiment that separately measures the impact of PRA penalty policy on the behaviors that those penalties are intended to modify. The earlier welfare demonstrations that included time limits, PRAs and financial sanctions only measured their impacts as part of a total package of reforms. Finally, all of these states except Texas and Arizona included earned income disregards or other financial incentives as part of their demonstration. As mentioned above, Texas’ earned income disregard was adopted outside of this evaluation. The ability of the Texas evaluation to isolate the effects of time limits and the personal responsibility agreement from an entire package of benefits is a key difference between the Texas evaluation and other AFDC waiver evaluations. Thus, the results from the ACT evaluation should help to fill a void in the existing research literature on the effects of welfare reform.

3See Bloom and Winstead (2002).(back)
4Texas TANF diversion programs were examined in Schexnayder, Lein et al. (January 2002).(back)

Table of Contents | Previous | Next