XV. TTW Participation by Beneficiaries in Adequacy of Incentives Groups

In passing the Ticket Act, Congress acknowledged that providers might be unwilling to accept Tickets from some beneficiaries because the TTW performance-based payment system might not cover the cost of services. Of particular concern was the possible inadequacy of payments such that providers would be unable to serve beneficiaries who either want to work but need long-term or expensive services or who are less likely to work at a level that will result in a payment. As part of an effort to address this concern, Congress required SSA to conduct a study on the adequacy of the incentives for providers to serve the following four groups of beneficiaries:

• Group 1: Beneficiaries who require ongoing support and services to work

• Group 2: Beneficiaries who require high-cost accommodations to work

• Group 3: Beneficiaries who work but earn a subminimum wage

• Group 4: Beneficiaries who work and receive partial cash benefits

We refer to these groups as Adequacy of Incentives (AOI) groups.

In this chapter, we use data from the 2004 NBS to analyze the characteristics and TTW participation behavior of the AOI groups. The use of the NBS data, which contain survey questions designed to identify the AOI groups, represents an improvement over previous reports’ reliance on administrative data as a proxy for the groups. As in previous chapters, the analyses presented here focus on beneficiaries residing in Phase 1 states. Beneficiaries residing in Phase 2 or 3 states did not have enough time to use their Tickets, or TTW had not been rolled out at the time the NBS sample was drawn. Thus, we are unable to use the first round of NBS to examine participant characteristics for AOI groups in these states. Subsequent rounds of data collection will include Phase 2 and 3 TTW participants in the NBS participant sample.

The NBS data show that 72 percent of all beneficiaries fall into one of the four AOI groups. The high percentage of AOI members is consistent with the expectations of the Ticket to Work Adequacy of Incentives Advisory Group, is in line with the findings based on the administrative definitions used in earlier reports, and reflects the definition of disability used to administer Social Security disability programs (Adequacy of Incentives Advisory Group 2004). In fact, a large share of beneficiaries has severe impairments that include mental illness, mental retardation, or other developmental disabilities and musculoskeletal conditions, all of which have been shown to be associated with the characteristics of the AOI groups (Salkever 2003; Wehman and Revell 2003).

AOI Group 1 (needs ongoing support) and Group 2 (needs high-cost accommodations) represent the vast majority of AOI group members; more than one-third of the individuals in the groups are members of both Groups 1 and 2. The NBS data show that a smaller share of these groups report employment-related goals and expectations that would result in TTW outcome payments when compared with the non–AOI group, and the groups’ characteristics might influence a provider’s willingness to serve members of the groups. We found that individuals in Groups 1 and 2 as well as individuals in Group 1 only have statistically significant low participation rates. While the rates may suggest that TTW incentives are not adequate for providers to serve the groups, they may also reflect other factors such as demand for TTW services among the groups.

The NBS data show that small shares of beneficiaries fall into AOI Groups 3 (works at subminimum wage) and 4 (works and receives partial benefits), reflecting the low employment rate among all beneficiaries. A larger share of Group 4 members report employment goals and expectations that would result in TTW outcome payments; in fact, we find that Group 4 members are almost three times more likely to participate in TTW than non–AOI members. The higher participation rate may reflect greater demand for services among Group 4 members. By definition, Group 4 members have decided to engage in work that is not necessarily subminimum wage employment and thus may be more likely to seek services from providers to obtain employment.

Compared with other participants, individuals in both Groups 1 and 2 and those in either Group 3 or 4 are somewhat more likely to assign their Tickets to SVRAs and to be assigned to the traditional payment system. This pattern may reflect provider incentives in that ENs may be unwilling to accept Tickets from people unlikely to work at levels that terminate benefit payments. However, it could result from SVRAs’ direct connection to programs in which beneficiaries are employed at subminimum wage (such as sheltered workshops).

While participation in TTW is the first step a beneficiary may take toward leaving the SSA payment rolls, it is important to emphasize that Ticket assignment does not indicate successful use of services or certainty of entering “benefits not payable” status for AOI beneficiaries. Future analyses will focus on the outcomes of service use for these beneficiaries.

In the sections that follow, we describe the criteria used to define the AOI groups based on the survey data and provide an overview of how beneficiaries are distributed across AOI groups. Next, we compare the characteristics of beneficiaries in each AOI group and assess their levels of TTW participation relative to non–AOI beneficiaries. We then examine the provider and payment types to which TTW participants in each of the AOI groups are assigned and conclude with a summary and discussion of the findings.

A. National Beneficiaries Survey AOI Definitions and Overview of AOI Groups

This report is the first to use the NBS data to identify beneficiaries who are in AOI groups. In previous reports, only administrative data were available to identify beneficiaries who might belong to one of the groups. We identified members of AOI Groups 1 and 2 by using the functional and health status measures within NBS instead of the medical condition(s) documented as the reason for the disability benefit award, as was necessary when using the administrative data in previous reports. In addition, AOI Groups 1 and 2 are no longer mutually exclusive (as they were with the previous methodology’s use of administrative data). Indeed, many beneficiaries need both ongoing support and high-cost accommodations. The NBS data also allow us to define Group 3 (works at subminimum wage) based on actual monthly wages reported rather than on an estimate calculated on annual earnings from SSA administrative records. AOI Group 4 (works and receives partial cash benefits) is defined by using primarily administrative data that directly identify benefit amounts.

Beneficiaries in the four AOI groups are identified by using the NBS data as follows:¹

Group 1: Beneficiaries Who Require Ongoing Support and Services to Work. We define members of AOI Group 1 as beneficiaries with service use or a level of functioning that suggests frequent need for personal assistance or job coaching and/or a tendency to be able to work only episodically. The group includes beneficiaries who satisfy at least one of the following criteria:

1. The need for assistance from another person, such as an interpreter or attendant

2. At least three ADL or IADL limitations where the beneficiary requires the assistance of another person and/or the presence of at least three severe physical limitations

3. The need for assistance from someone at work or the need to discuss employment goals with a job coach

4. The need for the assistance of a proxy respondent to complete the survey because of poor memory, confusion, not knowing how to answer, or another mental condition

5. Poor mental health defined as a mental health summary score (based on the SF-8 TM)² in the bottom decile for the U.S. population

6. Alcohol use, drug use, or treatment that points to a substance abuse or dependence problem

The first four criteria are directly related to the need for personal assistance to perform daily tasks or activities. The inclusion of the fifth and sixth criteria is based on a large body of research showing that persons with mental health or substance abuse problems represent “hard-to-employ” populations.³ Job retention is a major issue for members of these groups, and successful employment programs emphasize the need for ongoing support and services to maintain employment.

Group 2: Beneficiaries Who Require High-Cost Accommodations to Work. We define members of AOI Group 2 as beneficiaries who indicate the need for potentially using high-cost accommodations. The group includes beneficiaries who report that they:

1. Currently use or formerly used an accommodation at work

2. Need to use assistive technology

3. Have a severe sensory limitation and/or require assistance or a proxy to complete the survey because of a hearing or speech problem

4. Use mobility aids

5. Have mobility limitations that make it difficult to get around both inside and outside the home

The definition of what does and does not constitute the need for a high-cost accommodation is somewhat controversial. Some studies simply have used the price of a specific accommodation or assistive technology without consideration of the potential high costs that may be associated with integrating the accommodation into the workplace.⁴ For TTW, consideration of the broader costs of integrating accommodations into the workplace would appear to be critical to supporting the program’s employment goals.

Pinpointing the exact costs of purchasing and integrating accommodations into the workplace is difficult. The criteria used in this report capture the potential need for high-cost accommodations. Appendix E shows that, for most of the criteria, fewer than 5 percent of all TTW-eligible persons report the need for an accommodation. The exceptions are those who require mobility aids, which include 13 percent of all TTW-eligible persons, and those with severe sensory limitations, which include 16 percent of all TTW-eligible persons.

Group 3: Beneficiaries Who Work but Earn a Subminimum Wage. Survey responses to questions about wage, salary, and hours worked at a beneficiary’s primary and other jobs provided the basis for calculating an hourly wage rate for each job. If the wage rate was less than the federal minimum wage of $5.15 per hour at all of a respondent’s reported jobs, the respondent was classified as a member of AOI Group 3.⁵

Group 4: Beneficiaries Who Work and Receive Partial Cash Benefits. Group 4 comprises beneficiaries who received SSI benefits in the month before their interview (based on both administrative data and self-reports) and had self-reported earnings in that month. DI-only beneficiaries are not included in Group 4 because they are not eligible for partial cash benefits.

As shown in the first section of Exhibit XV.1, 72 percent of all Phase 1 beneficiaries fall into at least one of the four AOI groups. The second section shows the percentage of beneficiaries who meet the criteria for membership in each of the four AOI groups. Unlike the case of our earlier analysis, which was based on administrative data, beneficiaries may be members of more than one group; for example, they may require both ongoing support (Group 1) and high-cost accommodations (Group 2). Most beneficiaries fall into Group 1 (63 percent) and Group 2 (35 percent). Only very small shares of beneficiaries fall into Group 3 (3.3 percent) and Group 4 (2.5 percent). The final section of Exhibit XV.1 shows the distribution of those in one group only, those in at least Groups 1 and 2, and those in some other combination of AOI groups. Many beneficiaries fall into at least Groups 1 and 2 (27 percent), accounting for more than one-third of the 72 percent of beneficiaries in the AOI groups.⁶

The distribution of Phase 1 TTW participants across AOI groups differs from the distribution across all Phase 1 beneficiaries. Relative to all Phase 1 beneficiaries, a smaller share of TTW participants falls into one of the AOI groups (67 versus 72 percent). In addition, a much greater share of TTW participants is in Group 4 (10 versus 2.4 percent), and a somewhat smaller share falls into Group 1 (54 versus 63 percent).

1. Characteristics of AOI Group Members

Exhibit XV.2 presents selected beneficiary characteristics by AOI group status for those in Group 1 and Group 2. The two groups are then disaggregated into members of both Groups 1 and 2, members of Group 1 only, and members of Group 2 only. Each AOI group shares some characteristics that differ in many respects from those of beneficiaries not in any of the four AOI groups. Chapter III showed that some of the characteristics distinguishing the two groups (AOI versus non–AOI) are associated with TTW participation. Section 2 describes the distinguishing characteristics of each group. ⁷

2. Characteristics of AOI Group Members and TTW Participation Rates

Groups 1 and 2. Due to the large overlap between Groups 1 and 2, members of AOI Group 1 (needs ongoing support) and Group 2 (needs high-cost accommodations) share many characteristics. Relative to beneficiaries not in any of the AOI groups, members of Groups 1 and 2 have been on the rolls somewhat longer and are more likely to be white, Hispanic, not to have earned a high school diploma, to be in poverty, and to report poor or very poor health. The more frequently reported limitations are mental health and musculoskeletal conditions.

Another important difference between AOI Groups 1 and 2 and the non–AOI groups is that a smaller percentage of Group 1 and 2 members (1) have set forth goals that include work/career advancement, (2) see themselves working for pay in the next five years, and (3) see themselves working enough to stop disability benefits in the next five years.⁸ Groups with these lower employment goals and expectations tend to be less likely to work at levels that result in TTW performance-based payments (Mashaw and Reno 1996). Thus, the groups may pose a challenge to providers because, by definition, they tend to require expensive services and are less likely to have employment goals and expectations that may result in performance-based payments.

Given the large degree of overlap between Groups 1 and 2, we disaggregated the two groups into three alternative subgroups in order to compare their characteristics. The three subgroups include those in both Groups 1 and 2; those in Group 1 only; and those in Group 2 only. Comparisons across these alternative subgroups indicate that those in both Groups 1 and 2 differ in some important respects from those in only one of the two groups. For example, relative to beneficiaries in only one of the two groups, those in both groups have been on the rolls longer, are male, are more likely to live alone or with unrelated others, and are considerably less likely to see themselves working in the future. Only 22 percent report that they see themselves working in the next five years, and 9 percent report that they see themselves working enough to stop disability benefits as compared with the corresponding estimates of more than 30 and 16 percent in the two other groups.

The comparisons also show that Group 1 and Group 2 members differ from each other more dramatically on some characteristics when those in both groups are excluded. For example, relative to Group 1-only members, Group 2-only members are more likely to be DI-only, older, white, to have education beyond high school, and to report sensory disorders. They are less likely to be SSI-only, African American, Hispanic or Latino, without a high school diploma, living alone or with unrelated others, poor, with a disability since childhood, or diagnosed with mental illness or mental retardation.

Group 3. Exhibit XV.3 presents descriptive characteristics for AOI Groups 3 and 4. Relative to non–AOI beneficiaries, members of Group 3 (work at subminimum wage) are somewhat younger (45 years, on average, compared with 49 years) and have been on the rolls considerably longer (200 versus 136 months), perhaps partly due to the fact that a very large share of Group 3 members experienced disability onset during childhood (73 versus 16 percent). Group 3 members are much more likely to report excellent or very good health (40 percent compared with 8 percent), much more likely to report mental retardation and sensory limitations (24 and 19 percent, respectively, compared with 2 and 5 percent, respectively), and less likely to report a musculoskeletal condition as a reason for limitation. Relative to non–AOI beneficiaries, Group 3 members are also less likely to be SSI-only, female, and in poverty and considerably less likely to be members of a racial or ethnic minority.

A very large share of AOI Group 3 members see themselves working in the future, but relatively few see themselves earning enough to leave the disability rolls. The latter finding may affect a provider’s willingness to accept Tickets from Group 3 members. Under both new payment systems, payments are either totally or partly contingent on working enough to terminate disability payments. If Group 3 is less likely to work at such a level, providers may be less likely to accept Tickets from group members because payments may not be adequate to cover the cost of services.

Group 4. Members of Group 4 (work and receive partial benefits) share several of the same characteristics that make Group 3 members different from non–AOI beneficiaries, as shown in Exhibit XV.3. Group 4 members are considerably younger, with a mean age of 36 years compared with 49 years for non–AOI group members. They have been on the rolls longer (183 versus 136 months), and the majority experienced disability onset during childhood (63 percent compared with 16 percent). Similar to Group 3 members, Group 4 members are more likely to be in excellent or very good health and to report mental retardation as a condition causing limitation. They are less likely to be female and to report a musculoskeletal condition as the reason for a limitation. Unlike Group 3 members, they are more likely to be in poverty (54 versus 46 percent).

Compared with the non–AOI group members, Group 4 members have set forth employment goals and expectations that would appear to make them good candidates for TTW. An estimated 63 percent of Group 4 members report goals that include work/career advancement as compared with only 36 percent among non–AOI members; 84 percent report that they expect to be working in the future as compared with only 33 percent of non–AOI members; and 33 percent report that they see themselves working enough to stop disability payments within the next five years as compared with only 16 percent of non–AOI members.

3. TTW Participation among AOI Group Members

As shown in Exhibit XV.4, the TTW participation rate among all Phase 1 beneficiaries classified into any of the four AOI groups is somewhat lower than the rate for all non–AOI beneficiaries (0.7 percent compared with 1.0 percent). The overall lower rate is attributable to the lower participation rates in Group 1 (0.7 percent) and Group 2 (0.7 percent). Participation rates do not appear to change substantially when Group 1 and Group 2 members are disaggregated into those in Groups 1 and 2, those in Group 1 but not in Group 2, and those in Group 2 but not in Group 1. Participation rates among members of the very small AOI Groups 3 and 4 are higher than the rate for non–AOI beneficiaries, with Group 4 participation rates more than three times as high.

Some of the differences between the AOI and non–AOI participation rates may be explained by characteristics unrelated to AOI status per se. For example, as discussed in the previous section, relative to non–AOI beneficiaries, those in Groups 1 and 2 are more likely to be white, not to have earned a high school diploma, and to have been on the disability rolls longer. Chapter III showed that, all else constant, these characteristics are negatively associated with TTW participation. To test whether the differences in TTW participation across AOI groups are statistically significant after controlling for characteristics unrelated to AOI status but likely to affect TTW participation, we estimated a multivariate (logit) model of TTW participation that includes variables reflecting AOI group status. Given that a variety of health-related criteria were used to define AOI Groups 1 and 2, the model excludes variables that directly reflect current health and functional status (Appendix Table C.32).

The findings from the multivariate analysis indicate that, after controlling for a variety of sociodemographic and program-related characteristics, the differences in TTW participation rates between AOI members and non–-AOI members are somewhat smaller. For those in AOI Group 2 only, the difference is not statistically significant. The differences for those in AOI Group 1 only and those in both Group 1 and Group 2 remain similar in magnitude to the results shown in Exhibit XV.4 and are statistically significant at the 10 percent level. This finding is consistent with the multivariate model (logit) results in Chapter III. Almost half of Group 1 beneficiaries report that they need assistance with at least three ADLs and/or IADLs, and Chapter III shows that the need for such assistance is a strong predictor of low TTW participation rates.

The findings of the multivariate analysis also indicate that, after controlling for other characteristics, Group 3 TTW participation rates do not differ significantly from those of non–AOI beneficiaries. In contrast, the greater participation of Group 4 members is statistically significant after controlling for other characteristics, with Group 4 beneficiaries almost three times more likely to participate in TTW. The higher participation rates for Group 4 may reflect the fact that everyone in the group has made the decision to work and, as a group, may be more likely to seek the services necessary to obtain or maintain employment.

Consistent with the findings of the participation analysis presented in Chapter III, the analysis incorporating the AOI variables indicates that age, education, age at disability onset, and having children under the age of six residing in the household are among the most important determinants of TTW participation.

B. Provider and Payment Types among TTW Participants in AOI Groups

Exhibit XV.5 shows the distribution of provider and payment types associated with TTW participants in the various AOI groups. Compared with non–AOI group members, AOI group members participating in TTW were even more likely to assign their Ticket to an SVRA than to an EN. AOI group members were also more likely to be assigned to the traditional payment system, which reflects the TTW rules restricting the use of the traditional payment system to SVRAs. Through an examination of differences across AOI groups, Exhibit XV.5 clearly shows the relationship between the use of SVRAs and assignments to the traditional payment system. As the percentage of persons within each AOI group using SVRAs increases, the percentage of persons assigned to the traditional payment system also increases.

The differences in provider and payment types between TTW participants categorized as AOI and non–AOI are greatest for members of Group 3, members of Group 4, and members of both Group 1 and Group 2. While the differences may be related to the adequacy of TTW payment incentives, other reasons may explain the difference in the distribution of providers and payment types. One alternative explanation is that Group 3 and Group 4 consist of individuals who were working at the time of the interview and may have decided to use SVRA services before rollout of TTW in Phase 1 states. Another alternative explanation is that many Group 3 members may be participating in supported or sheltered employment programs frequently sponsored by SVRAs and thus earning subminimum wages. Finally, SVRAs may be more effective in providing services to persons with several limitations such as those in both Groups 1 and 2.

C. Conclusions

While the data presented in this chapter represent a fairly early stage of TTW implementation, we find some evidence that may support the concern that the performance-based payment system discourages providers from serving beneficiaries in Group 1 and beneficiaries in both Groups 1 and 2 who might require more intensive or long-term support to become employed. Both groups have low participation rates, and members of both Groups 1 and 2 are likely to have a Ticket assigned to an SVRA and operate under the traditional payment system. But, when compared with other factors that affect participation—such as age, education, and having children under the age of six living in the household—the influence of membership in the AOI groups on participation is weak. Research (McGrew 2005) indicates that, if properly designed, performance-based payment systems can address the needs of individuals with the most severe disabilities. The problems we observed may be an artifact of the low payment rates under the current system, which may be addressed by the proposed payment system. Further study of this issue after implementation of the proposed payment system may resolve this fundamental question. In addition, it is possible that the findings related to TTW assignments may result from the early stages of TTW implementation. Thus, we are unable to determine the degree to which the findings are attributable to the adequacy of TTW incentives.

Future reports will investigate whether the new round of NBS data may be used to assess involuntary nonparticipation among AOI groups.⁹ Involuntary nonparticipation may provide a more convincing means of assessing whether low participation rates result from the adequacy of incentives. If we find only a few involuntary nonparticipants or do not find that membership in the AOI groups has an impact on involuntary nonparticipation, then it is less likely that lower participation rates are attributable to TTW incentives. If we find such an effect, then a further examination of how incentives may be changed to improve access to services for AOI groups may be required.

The second round of data from NBS will also be used to examine further both participation and outcomes for AOI group members. The new longitudinal component from the second round of NBS will allow us to track the experiences of AOI group members over time for comparison with the experiences of non–AOI group members. The analysis will provide the more detailed information that is necessary to assess the extent to which AOI group members face different experiences with ENs and whether these experiences are related to TTW incentives.

¹ Appendix E provides a summary of the survey data classification criteria, descriptive statistics, and a comparison to the AOI group classification methods using administrative data. (back)

² SF-8 TM is a trademark of QualityMetric, Inc. (back)

³ Dion et al. (1999) contains an excellent review of the literature that shows the close link between substance abuse and difficulty finding and keeping a job as well as the link between poor mental health and keeping a job. This report also reviews successful employment programs for individuals with such problems. (back)

⁴ See Delaire (2003) for a useful summary of the issues associated with measuring costs of accommodations. (back)

⁵ Although many state minimum wage rates are higher than the federal minimum wage, we use the federal minimum wage to define those in AOI Group 3, thereby recognizing that the AOI group is defined in federal legislation pertaining to a federal program and that neither administrative nor survey data indicate the state in which wages were earned. Among the 13 Phase 1 states, where the vast majority of the Phase 1 survey respondents resided at the time of interview in 2004, 5 states ( Delaware, Illinois, Massachusetts, Oregon, and Vermont) had minimum wage rates in 2004 that exceeded the federal level, ranging from $5.50 to $7.05 per hour. (back)

⁶ Author’s calculations based on data in Exhibit XV.1. (back)

⁷ Appendix Table E.2 presents additional characteristics and beneficiary subgroups. (back)

⁸ It is also worth noting that this group is more likely to be working or to have worked in 2003. It is possible that the combination of current work and poorer health could be driving their less optimistic expectations. The next report will take advantage of the new round of data to examine this possibility empirically. (back)

⁹ Due in part to the limitation to Phase 1 states, small sample sizes of involuntary nonparticipation make it difficult to assess involuntary nonparticipation among AOI groups. The additional of Phase 2 states in round two of the NBS will likely be sufficient for an analysis of involuntary nonparticipation in the next report. (back)