Search Type Title Author/Investigator Search Advanced Search

Table of Contents | Previous | Next

Conclusions

The following conclusions can be drawn from all components of the ACT evaluation:

1. Impacts on public assistance receipt from the ACT demonstration were small but significant.

In the two pilots with state time limits, use of TANF by any family member declined while enrollments in transitional Medicaid increased. Average monthly TANF benefits in these sites also decreased slightly. In the two pilots that measured the effects of the PRA against earlier AFDC rules, adult TANF use and TANF Medicaid increased because the PRA financial penalties (unlike the earlier AFDC sanction policy) allowed the caretaker to stay on TANF while penalized. In these sites, use of food stamps declined while penalties increased. The pilot that combined state time limits and the PRA produced the strongest impacts on public assistance measures.

2. The ACT demonstration increased employment rates for some groups but had no impact on caretaker earnings.

In the two experiments with state time limits, persons subject to time limits had slightly higher rates of caretaker employment than did persons not subject to time limits. These impacts were concentrated among persons who had received TANF for less than 30 months. Neither time limits nor the PRA provisions increased the overall earnings of ACT participants compared to persons who did not have to meet these requirements.

3. Weak overall impacts often masked subgroup differences, especially by tier group.

As shown by the subgroup analyses in the impact evaluation, state time limits and the personal responsibility agreement produced different impact for members of each tier group for public assistance and child support measures but not for employment and earnings. All three pilots showed great variation in impacts by tier, suggesting that both the state time limits and PRA policies affected persons with varying levels of education and work experience in different — and sometimes opposite — ways.

Short-term TANF recipients responded to ACT’s provisions more strongly than persons with longer periods of prior TANF use. This was particularly true in the sites that combined time limits and the PRA, which resulted in lower monthly TANF benefits, fewer penalties, and higher rates of employment and earnings for this group. Persons with less than 30 months of prior TANF receipt are the only ones for whom overall caretaker earnings increased as a result of this demonstration. Policy changes enacted in 1999-2000 did not substantially affect one-year outcomes from this demonstration, as documented by the similar impact results for families assigned to the ACT demonstration prior to and following the implementation of these new policies. Except for increasing Choices participation in the RER Choices pilot and use of child care in the pilots with time limits, the more recent TANF provisions — including federal time limits, the expanded earned income disregard, and the tightening of the ‘age of child’ exemption — did not change families’ responses to ACT policies. This suggests that the impact findings from this demonstration are still relevant in the current Texas policy environment.

4. Staff and clients support the concept of the time limits but the rules are too complex to understand.

As documented both in the process evaluation and the detailed interviews, the complexity in the structure of the Texas time limits and the differences in rules between the state and federal time limits caused quite a bit of confusion among both staff and clients. While staff improved in their understanding and explanation of these provisions over time, clients interviewed in 2001 still did not comprehend the implications of these provisions. This complexity caused time limits to lose power as a tool to influence behavior and may have reduced the size of the impacts in pilots with time limit provisions.

5. The interaction of state and federal time limits will result in the most disadvantaged families being the first to reach lifetime limits on Texas TANF receipt.

As documented in the process evaluation, the differing structures of the periodic state time limits enacted by HB 1863 and the lifetime federal time limits adopted in December 1999 mean that families will reach the more stringent federal limits at different rates. Under the current design, families of caretakers who are exempt from state time limits would be among the first to reach federal time limits because the federal clock ticks for these families even when they are not advancing toward a state time limit.

Also, persons who actually reach a state time have their TANF grants reduced but stop both state and federal clocks for a five-year period in which the caretaker is removed from TANF. Because of this, adults with the shortest state time limits (i.e., Tier 1) typically would be among the last to have their entire families removed from TANF by the 60-month federal limits, while those with the largest state time limits (i.e., Tier 3) would advance toward the federal limit more quickly.

6. The increased use of financial penalties did not change most of the behaviors governed by the personal responsibility agreement.

With the exception of some child support measures, the impact analysis found that the increase in penalties for failure to comply with Choices participation, school attendance or immunization requirements of the PRA did not improve performance in these areas for the groups subject to the penalties.

The process evaluation found that clients understood that they had to comply with PRA but did not know the implications of failure to comply. This occurred even in the non-Choices sites where the rules were simpler (due to the absence of time limits), and workers had more time to explain things and knew the families better. Interviews with low-income families in the sites with both the PRA and time limits found that clients often did not know why the amounts of their TANF grants were changing. Because of due process provisions that needed to be followed before a penalty could be imposed, penalties typically did not reduce TANF grants until 2-3 months after an infraction occurred. This, combined with the numerous reasons that TANF recipients could be penalized, often made it difficult for clients to know why their TANF grant was changing.

7. Most of the TANF caseload decline from 1995-1999 was not due to ACT waiver provisions.

The rapid caseload declines occurring in the early years of the ACT demonstration affected both experimental and control groups in all sites, and the very small impacts in this demonstration suggest that most of these declines were not caused by ACT policies. Caseload declines may have been partially due to other welfare reform provisions not measured by this demonstration, particularly TANF diversion policies. Texas families have a long history of cycling between welfare and work, and diversion policies may have kept some cyclers from returning to TANF. General publicity about welfare reform, changes in other related policies (such as the expansion of the Earned Income Tax Credit) or the strong economy that was in place throughout the operation of this demonstration could also have accounted for the steep caseload decline in the early years of ACT.

8. Impacts from this demonstration differ from those in other states but the mix of services is not comparable.

As mentioned earlier, the other state waiver evaluations measured a package of welfare reform provisions. Unlike the Texas evaluation, they were not designed to isolate individual policy provisions. None of the other states isolated either state time limits or PRA-like requirements from the entire package of reforms measured by those evaluations. Unlike Texas, all but one of the other states included financial work incentives or earned income disregards as part of the package of benefits included in their waiver evaluations.

9. A number of factors may have contributed to the small size of the impacts from this demonstration.

In demonstrations such as this one, small impacts typically occur because the experimental and control group policies do not differ sufficiently from each other, the experimental policies are not applied intensively enough, or the new policies are not any more effective at changing the outcomes than the old ones were. The most likely reasons that the ACT demonstration produced such small impacts include:

• the structure of Texas time limit policies with its many exemptions from state time limits;
• client confusion about the details and intent of both the time limit and PRA policies;
• clients paying more attention to their immediate needs rather than future consequences imbedded in these policies; and
• the exclusion of policies that supported employment from the set of policies included in the ACT demonstration.

The analysis of time limit clocks showed that Texas time limits did not apply to over half of TANF recipients and stopped often even when they did apply. The next two of these factors were well documented in the process evaluation. Local eligibility workers reported and detailed interviews with clients confirmed that clients were confused about ACT policies and were more focused on their immediate financial needs than any future consequences imbedded in TANF penalty or time limits policies. Finally, unlike most other states, the Texas waiver did not include clear positive incentives for increasing employment. In particular, the change in Texas’ earned income disregard policy, which the Texas Families in Transition study documented as increasing the rates of employment among Texas TANF leavers, was not a policy measured by the ACT demonstration.

 

 

Table of Contents | Previous | Next