PERSPECTIVES: Changes to the Ticket to Work Regulations in 2008 Attracted Providers and Participants, but Impacts on Work and Benefits Are Unclear

Introduction

The Social Security Disability Insurance (DI) and Supplemental Security Income (SSI) programs, administered by the Social Security Administration (SSA), provide income support to individuals who have long-lasting medical impairments and are unable to work at a substantial level. In August 2015, around 13 million working-age adults received benefits from one or both of these programs (SSA 2015). This article presents new statistics on the extent to which beneficiaries have given up their benefits to return to work since the introduction of the Ticket to Work (TTW) program in 2002, particularly in the period before and after July 2008, when SSA significantly changed the program's regulations to spur participation among both beneficiaries and employment service providers.

Many DI and SSI beneficiaries are interested in working, even if they are not able to do so at a significant or sustained level. Eligibility for federal disability benefits is partially based on the inability to engage in substantial gainful activity (SGA), which in 2015 is defined as equivalent to monthly earnings of $1,090 for nonblind beneficiaries and $1,820 for blind beneficiaries. Despite this criterion, several program provisions are designed to allow participants to test their ability to return to work. Under DI, beneficiaries are granted a 9-month trial work period (TWP) within a rolling 60-month window during which they can earn an unlimited amount and yet retain benefits. Following the completion of the TWP, benefits are suspended for work in any of the following 36 months in which beneficiaries engage in SGA (except for a 3-month grace period). After this 36-month period (and any remaining grace-period months), benefits are terminated in the first month of SGA.¹ SSI rules are quite different; after a small earnings disregard, benefits are reduced by $1 for every $2 in earnings, meaning that many beneficiaries may earn approximately twice as much as the federal benefit rate and retain some level of benefits.² Both programs include provisions that allow beneficiaries to maintain associated health insurance coverage (from Medicare in the case of DI and from Medicaid in the case of SSI) even after cash benefits have been terminated because of SGA.

Because of the strict and sometimes lengthy determination process required to prove inability to engage in SGA, beneficiaries often fear losing their disability benefits if they become employed and earn above certain thresholds. Moreover, once individuals with disabilities have left the labor force and met either program's eligibility criteria, they may suffer skills deterioration and loss of human capital that may complicate labor force reentry. A large body of literature has explored the magnitude of the labor-supply disincentive effects of the DI program; two of the most recent examples are Maestas, Mullen, and Strand (2013) and French and Song (2014).

Recognizing that many beneficiaries feared losing benefits and lacked knowledge of program rules and work supports, Congress enacted the Ticket to Work and Work Incentives Improvement Act of 1999 (Ticket Act). That legislation put into place a number of new policies and programs designed to support the return-to-work efforts of disability program beneficiaries. The Ticket Act focused on increasing the extent to which beneficiaries forgo cash disability benefits, in whole or in part, because of work. The centerpiece of the Ticket Act is the TTW program; its implementation began in February 2002. TTW expanded the ways in which SSA pays service providers for supporting beneficiaries in their employment efforts. Under TTW, providers receive compensation when beneficiaries achieve certain specified earnings levels or, in the case of “outcome payments” (described later), benefit cessation because of work.

This article presents new statistics on TTW participation and participant work activity. First, we present annual statistics from 2002 through 2010 on TTW participation to show that enrollment growth was initially slow, but accelerated after revised program regulations went into effect in July 2008. We also document patterns of TTW participant earnings during this period, focusing on the longitudinal pattern of earnings across successive annual new-participant cohorts. Next, we document TTW participants' likelihood of forgoing cash benefits because of sustained work activity and the duration of the nonpayment periods, comparing the outcomes with those for nonparticipant beneficiaries. Finally, we examine in detail the statistics for participants immediately before and after the revised regulations went into effect. We do not attempt to formally estimate the direct impacts of the 2008 regulatory changes, given the lack of a suitable counterfactual and the confounding effects of the recession during 2007–2009. Nonetheless, the statistics are informative about the effects of the regulatory changes on program participation and participant outcomes.

An important feature of our analysis is the use of a constructed monthly variable that indicates whether the beneficiary is in nonpayment status following suspension or termination of benefits because of work (NSTW). This indicator aggregates information from a variety of administrative sources to provide the first available measure of its kind. It was developed to support the evaluation of the TTW program and has been used extensively for that purpose (Schimmel and Stapleton 2011; Liu and Stapleton 2011; Ben-Shalom and others 2012).³

This article consists of seven sections, beginning with this introduction. In the second section, we describe the history and features of the TTW program. In the third section, we describe our data sources, our study population selection criteria, and the NSTW measure in detail. The fourth section presents statistics on TTW participation and the achievement of SGA-level earnings by successive cohorts of TTW participants. In the fifth section, we compare outcomes for TTW participants with those for nonparticipant beneficiaries, examining first-time attainment of NSTW status and the likelihood of sustaining that status. The sixth section compares statistics from before and after the 2008 regulatory changes. In the concluding seventh section, we synthesize our findings and discuss their policy implications.

The TTW Program

The TTW program is the result of extensive deliberations that have been taking place since before the Ticket Act was passed. This section highlights the program's underlying principles, its implementation, and subsequent revisions to its initial regulations.

Initial Program Design

The TTW program was established to offer new or broader access to employment services and supports that might enable SGA for participating beneficiaries. Under the program, SSA mails to the beneficiary a ticket that he or she may “assign” to a provider in exchange for employment services. The provider then collects payments from SSA when the client beneficiary reaches certain earnings thresholds. In essence, TTW is a performance-based voucher system; SSA agrees to pay certain amounts to the beneficiary's service provider over a long period based on the beneficiary's attainment of specific objectives, primarily the suspension or termination of benefits because of work.

Before TTW, state vocational rehabilitation agencies (SVRAs) were virtually the only providers eligible to receive payments from SSA for serving beneficiaries.⁴ TTW sought to augment that system with additional providers called employment networks (ENs), which could be public or private entities. TTW would thereby limit SSA's direct role and recognize market forces by compensating providers for successful beneficiary outcomes (Berkowitz 1996; Livermore and others 2003). According to the final rules issued December 28, 2001, TTW was meant to “enhance the range of choices available to Social Security disability and disabled and blind SSI beneficiaries when they are seeking employment services, VR [vocational rehabilitation] services and other support services to obtain, regain or maintain self-supporting employment” (SSA 2001).

In addition to expanding service options, a critical element of the original TTW legislation was that providers would be paid only if their beneficiary clients actually gave up benefits for work, in which case they would receive a share of the benefit savings for a lengthy period (Berkowitz 1996). The National Academy for Social Insurance (NASI) formed a Disability Policy Panel that summarized that feature as follows:

Under the plan, new beneficiaries would receive an RTW [return-to-work] ticket, akin to a voucher, that they could use to shop among providers of rehabilitation or RTW services in either the public or private sector. Once a beneficiary deposited the ticket with a provider, it would constitute an obligation for the Social Security Administration (SSA) to pay the provider after the beneficiary returns to work and leaves the benefit rolls. Providers whose clients successfully return to work would, each year, receive in payment a fraction of the benefits savings that accrue to the Social Security trust funds because the former beneficiary is at work and not receiving benefits…The Panel believes its incentive-based system of financing rehabilitation and RTW services could bring a doubling or tripling of the rate at which beneficiaries leave the benefit rolls because they have successfully returned to work. While those numbers are small in relation to the size of the beneficiary population, they represent a significant improvement over current experience. (Mashaw and Reno 1996, 101 and 108; emphasis added)

The NASI Panel recognized that initial payments based on earnings objectives that fall short of benefit savings might be needed to entice providers to participate, but recommended starting with payments based on benefit reductions alone, with payments equal to half of the benefit savings over 5 years (Mashaw and Reno 1996, 117).

Program Implementation and Early Results

Reflecting the negotiations that preceded passage of the Ticket Act, the TTW program that rolled out between 2002 and 2004 was substantially different from and more complex than the version recommended by the NASI Panel (Berkowitz 2003; Mashaw and Reno 1996). First, although the Panel recommended offering tickets only to new beneficiaries until the program was well established and showed signs of success, nearly all DI and adult SSI-only disability beneficiaries were eligible to use tickets from the program's inception.⁵ Second, SSA established two new payment systems. One system, outcome-only (O-O), followed the Panel's recommendation of paying providers only when the beneficiary gave up benefits for work; the other, milestone-outcome (M-O), allowed providers to receive some payments when clients achieved earnings milestones that did not result in a loss of benefits. Each EN was required to choose one of those two systems. Third, SVRAs were allowed to accept tickets as an EN under either of TTW's new payment systems, or to use the traditional SVRA cost-reimbursement system on a case-by-case basis. Details of the initial versions of the O-O and M-O systems appear in the top panel of Table 1 and are discussed in more detail later.⁶

Fourth, instead of basing provider payments on benefit savings for the individual beneficiary, SSA based them on average savings for all beneficiaries in the previous year, with values for DI and SSI-only beneficiaries tabulated separately. As a result, the payment received by a provider for a DI beneficiary was the same amount regardless of whether potential monthly savings to SSA were $500 or $1,500. Further, although the NASI Panel suggested paying providers an amount equal to 50 percent of the benefit savings for the individual over 5 years, payments under the initial O-O system were equal to 40 percent of mean DI benefit savings in the previous year (Thornton and others 2004).

In summary, the TTW implementation opened the market for beneficiary employment services to more providers and, for the first time, explicitly offered to compensate providers for reductions in disability benefit payments; but it did not go as far in the latter direction as recommended by the NASI Panel. Nonetheless, in the program regulations announced on December 28, 2001, SSA pursued the same goals promoted by the Panel. Livermore and others (2003) noted that “overall, the expectation is that TTW will increase the likelihood of self-sustaining employment for DI and SSI beneficiaries, resulting in decreased government expenditures on benefits, increased tax revenues, and a general strengthening of communities and the workforce.”

The initial response to TTW was quite limited in terms of the number of tickets assigned, the number of providers accepting tickets, changes in the services offered, and the extent to which beneficiaries earned enough to forgo benefits (Stapleton and others 2008). The vast majority of assigned tickets were accepted by SVRAs under the traditional payment system. Only a few providers elected the O-O payment system and, as a result, most assignments to ENs were under the M-O system. It also became clear that very few providers were likely to find TTW economically attractive (Thornton and others 2007; Stapleton and others 2008). ENs cited administrative burdens and low payment levels as significant barriers to their TTW participation (Prenovitz, Bardos, and O'Day 2012). Findings from a recent evaluation show that the introduction of TTW increased beneficiary enrollment for services, but had no measurable impact on the number of months in NSTW status that beneficiaries accumulated (Stapleton, Mamun, and Page 2014).

Revised Program Regulations

TTW's early years provided evidence on the extent to which beneficiaries would prefer earning enough to forgo benefits. Many beneficiaries do in fact work, but only a small minority of them earn enough to leave the program rolls (Liu and Stapleton 2011; Livermore 2011; Schimmel and Stapleton 2011; Ben-Shalom and others 2012). To encourage provider participation in TTW, SSA implemented significant regulatory revisions in July 2008 that increased the amounts that providers could be paid before beneficiaries forgo benefits for work, accelerated the payment schedule for DI-beneficiary clients, and brought payments for SSI-only clients more in line with those for DI beneficiaries. Although those changes made TTW more financially attractive to providers, they also reduced provider incentives to produce benefit savings.

From 2002 through June 2008, the O-O and M-O payment rules shown in the top panel of Table 1 were in effect (the table shows the payment amounts that SSA established for 2008, expressed in 2012 dollars). The maximum payments under the O-O system totaled $25,800 for DI beneficiaries (including those receiving concurrent DI and SSI benefits) and $14,760 for SSI recipients. The maximum total payments under the M-O system were $21,900 for a DI client (including $4,380 in milestone payments) and $12,600 for an SSI-only client (including $2,512 in milestone payments).

The 2008 revisions, shown in the lower panel of Table 1, had four primary effects. First, they shortened the payment period for DI clients so that ENs could receive full payment within as few as 36 months. Second, they created two phases for milestone payments, allowing providers to receive payments sooner and for lower levels of client earnings. Third, they increased M-O payment amounts so that the maximum would be closer to the maximum O-O payment amount. Fourth, they brought payment amounts for SSI-only recipients more closely in line with those for DI beneficiaries. Under the new regulations, SSA bases phase 1 milestone payments on the TWP income amount—in 2012, equal to gross earnings of $720 a month—whereas phase 2 milestone payments are based on the original (higher) SGA amount. Under the M-O system, the maximum value of milestone payments is $9,409 for DI beneficiaries (more than double the pre-2008 amount) and $9,148 for SSI-only recipients (3.6 times the pre-2008 amount).

Other regulatory and administrative changes implemented by SSA in 2008 sought to reduce the administrative burden on ENs for participating in TTW, further enhancing the program's financial attractiveness to potential providers. SSA removed the requirement that SVRAs accept actively assigned tickets in order to receive payments under the traditional payment system. Instead, SVRAs would now only need to document for SSA that a ticket was “in use,” meaning that the beneficiary was receiving employment services from the SVRA. When it was introduced, this change applied retroactively, meaning that SVRAs were asked to provide information about client beneficiaries they had served as early as 2002 so that those tickets could be deemed in use. Although SSA requested data on all beneficiaries served, many SVRAs did not provide identifying information for all client beneficiaries from the early years of TTW.

The revised regulations also allow a beneficiary to use SVRA services under the traditional payment system and subsequently reassign his or her ticket to an EN. This option is called Partnership Plus. If a participating beneficiary works, both the SVRA and EN are eligible to receive payments, although the EN is not eligible for phase 1 milestone payments if the beneficiary is employed at vocational-rehabilitation closure.⁷ In all, the revised regulations were meant to encourage more providers to become ENs, more ENs to accept tickets, and more beneficiaries to assign tickets; to provide more complete and timely records on SVRA delivery of beneficiary services, even if tickets were not formally assigned; and to bring about better employment outcomes.

Data Sources, Study Population, and Measurement of Beneficiary Work Activity

The primary data source used in this study is SSA's Disability Analysis File (DAF), which contains administrative information on all adults with at least 1 month of DI or SSI benefits from 1996 onward. We conducted an initial analysis with the 2010 version of this file (then called the Ticket Research File, or TRF10). When the 2011 data (DAF11) became available, we updated the information on beneficiary work activity. We were also able to take advantage of substantial improvements to the NSTW variable.

We used the TRF10 to identify as TTW participants all beneficiaries aged 18–64 who assigned a ticket during the period 2002–2010.⁸ We categorized beneficiaries with multiple ticket assignments according to the most recent assignment. For instance, if a beneficiary assigned his or her ticket in 2002 but subsequently withdrew it, then reassigned it to another EN in 2008, we used only the information from the 2008 assignment. Practically speaking, withdrawals were rare during the study period, so categorizing by only the most recent assignment should not substantively affect our findings.⁹ To categorize TTW participants by the payment system (M-O, O-O, and traditional SVRA) under which their ticket was assigned, we used the payment system recorded in the month of assignment. When comparing TTW participants with nonparticipants, we stratified our results by benefit type (DI-only, SSI-only, and concurrent DI-SSI) to account for differences in propensity to work and prior work histories across those groups. Because nonparticipants do not have an assignment month, we determined their benefit type in a given calendar year as that which applied in their first month of current-pay or NSTW status.

The NSTW variable was developed separately for DI and SSI to account for each program's rules. For DI, we regarded benefits as suspended when the administrative data indicated that the beneficiary engaged in SGA in any of the 36 months that followed the TWP and a 3-month grace period. We considered DI benefits to be terminated for work in all months for which the administrative data indicated SGA after month 36 (and any remaining grace-period months). We considered SSI payments to be suspended for work when the recipient had some countable earnings and total monthly countable income exceeded the federal benefit rate, which in 2012 was $698 for an individual. SSI rules count most income, including DI benefits, dollar-for-dollar after a $20 monthly disregard; however, only half of earnings above an additional $65 disregard are countable. There are also disregards for various other expenses related to the SSI recipient's efforts to return to work or become more self-sufficient. Hence, the minimum earnings amount that results in SSI payment suspension for work varies among recipients, depending on other income received and earnings disregards; for many, that level exceeds the SGA amount.¹⁰

The DAF data contain a combined NSTW indicator that aggregates information across the two programs for cases in which the beneficiary has received benefits from each program, but not necessarily concurrently. We used this variable in our analysis. In a given month, for DI-only beneficiaries, this combined indicator simply takes on the value of NSTW status in the DI program; for SSI-only recipients, it reflects that status in SSI. In the case of participants who received benefits from both programs at some point, months in which either program is in current-pay status are not counted as months in NSTW status. This approach is consistent with TTW's outcome-payment rules, under which payments are not due if the beneficiary receives a benefit from either program. Further, the combined indicator is coded as “suspended for work” if benefits have been suspended in one program and terminated in the other.

In our analysis, we did not distinguish between benefit suspensions and terminations. Once benefits were terminated for work, we continued to count the beneficiary's status as NSTW until, according to the administrative record, the beneficiary returned to current-pay status, attained full retirement age, or died. Moreover, beneficiaries are not necessarily engaged in SGA in every month counted as NSTW. In many months, we know only that the beneficiary left current-pay status because of work in a previous month and has not returned to current-pay status, reached full retirement age, or died. Similarly, we cannot verify that the beneficiary would continue to meet DI or SSI eligibility criteria if he or she were not engaged in SGA.

Many retroactive adjustments to NSTW status occur because of the time it takes SSA to receive and process evidence of work. For that reason, we used DAF11 to allow for as much time as possible between the end of our observation window (December 2010) and the date when the data were pulled from administrative sources (March 2012). Nonetheless, our investigation (described in detail in Schimmel and others 2013) revealed that data-processing lags are still an important consideration for the latest years in our analysis. Specifically, from a preliminary analysis of partial data pulled in May 2013, we calculated NSTW values among participants whose tickets were assigned from July 2008 through June 2009 that were approximately 10 percent higher than our findings using DAF11 (derived from data pulled in March 2012). The difference is much smaller for earlier TTW cohorts; using the May 2013 data, we estimated NSTW values among participants who assigned tickets from July 2006 through June 2007 that were only 2 percent higher than our findings based on DAF11. We did not explicitly correct for these data-processing lags, but we discuss them later in the context of our findings.

Because the NSTW indicator reflects participant earnings only to the extent that they affect cash benefits, we also considered the prevalence of beneficiary work activity as measured by the presence of any earnings. Participants who receive DI or concurrent DI-SSI benefits are quite likely to have earnings before attaining NSTW status because they can use the TWP and grace period. For their part, SSI-only recipients may have earnings indefinitely without attaining NSTW status. Ultimately, it is total earnings—not just the impact reflected in reduced payment of DI or SSI benefits—that captures the social value of beneficiary work activity. Benefit payment reductions net of EN payments provide a narrow measure of savings to the DI Trust Fund and to general revenues, but total earnings offers a picture of the broader benefits to the program.¹¹

To provide us with beneficiary earnings data, SSA linked the records from our analysis sample to its Master Earnings File. That file contains annual wage data derived from Internal Revenue Service W-2 tax forms, quarterly earnings records, and annual income tax data.¹²

TTW Participation Rates and Participant Employment

TTW was rolled out to groups of states in three waves from 2002 through 2004. After growth during the initial implementation, the number of new ticket assignments under the EN payment systems (M-O and O-O) remained fairly stable from 2005 through 2007 (Table 2). For the rest of the study period, however, the number of new assignments under the EN payment systems sharply increased each year, rising from 4,168 in 2007 to 19,913 in 2010. Although some of this growth simply reflects increasing numbers of beneficiaries eligible to use TTW, those numbers had been climbing in the years prior to 2008 as well. From 2007 to 2008 alone, the number of new assignments under the EN payment systems increased by 144 percent, from 4,168 to 10,154, while the number of beneficiaries eligible for TTW rose by only 6 percent. The program continued to grow in subsequent years even as the economic recession took hold and finding employment for beneficiaries became more difficult. Given those circumstances, the timing and the magnitude of the growth in TTW participation clearly suggest that the 2008 changes in the program's regulations were the primary cause of that growth. As described below, the changing regulations encouraged providers to accept tickets—even from beneficiaries who were not likely to leave the disability rolls for work.

From the beginning, most ENs chose the M-O payment system, and that system dominated the number of assignments; but after the 2008 regulatory changes, provider preference of the M-O system over the O-O system for their new ticket assignments only widened. From 2007 to 2010, tickets assigned to ENs using the M-O system increased by 473 percent (from 3,417 to 19,564), while the number of new assignments for which ENs chose the O-O system—far lower to begin with—fell by 54 percent, from 751 to 349. That sharp divergence most likely reflects the increased number of potential milestone payments and the increase in the maximum value of M-O payments relative to the maximum O-O payments, as shown in Table 1.

There was a small decline in traditional SVRA assignments from 2004 (the last TTW rollout year) through 2007, followed by modest annual growth through 2010 (Table 2). Thus, the increase in M-O assignments does not simply reflect a shift from SVRA acceptance of assignments under the traditional payment system toward more acceptances under the M-O system. Even though SVRA assignments rose much less rapidly than M-O assignments after 2007, they remained the dominant type of assignment in 2010. One likely factor contributing to the growth in assignments to SVRAs in the later years is a substantial increase in federal grant funding to SVRAs authorized by the 2009 American Recovery and Reinvestment Act (Stapleton and Martin 2012). That temporary expansion in funding allowed SVRAs to respond to an influx of applications from unemployed workers with disabilities and thereby shorten their waiting lists. The administrative change noted earlier, in which SVRAs needed only to document that a ticket was “in use”—in many cases, retroactively—also likely led to the counting of a greater number of tickets assigned to traditional SVRAs than were actually newly assigned.

Charts 1 and 2 track the annual earnings experiences of TTW participants in each of the 2002–2010 ticket-assignment cohorts, followed through 2010 (regardless of whether the ticket remained assigned during the entire period).¹³ Chart 1 shows the percentages of cohort members with any positive earnings; Chart 2 shows the percentages of cohort members with adjusted annual earnings of at least $12,000, a threshold that approximates the annualized SGA level (note that the scales of the two charts' vertical axes differ). One distinct pattern that emerges from these charts is that the labor market outcomes of participants appear dampened following 2008: In Chart 1, existing downward trends accelerated; in Chart 2, downward trends replaced the preceding years' generally upward or level tracks. That shift has two possible causes. The first is the effect of the recession on labor market outcomes of participants; workers with disabilities fared particularly badly during the economic downturn (Kaye 2010; Livermore and Bardos 2015). The second is that the revised TTW regulations, which increased provider's incentives to participate but reduced their incentives to help clients to earn enough to forgo benefits, encouraged providers to serve beneficiaries who were less likely to give up their benefits for work. How this pattern has changed since 2010, as the economy started to recover, is not yet known.

Chart 1.
Percentage of TTW participants with any positive earnings, by ticket-assignment cohort: 2002–2010

Show as table

Table equivalent for Chart 1. Percentage of TTW participants with any positive earnings, by ticket-assignment cohort: 2002–2010

Year	Ticket-assignment cohort
	2002	2003	2004	2005	2006	2007	2008	2009	2010
2002	49.1	. . .	. . .	. . .	. . .	. . .	. . .	. . .	. . .
2003	47.1	53.2	. . .	. . .	. . .	. . .	. . .	. . .	. . .
2004	43.7	54.6	52.2	. . .	. . .	. . .	. . .	. . .	. . .
2005	41.7	51.4	53.1	54.6	. . .	. . .	. . .	. . .	. . .
2006	40.5	49.6	50.7	56.0	55.7	. . .	. . .	. . .	. . .
2007	38.9	47.1	48.5	52.4	55.8	54.9	. . .	. . .	. . .
2008	35.9	43.9	45.3	48.0	49.7	52.2	51.7	. . .	. . .
2009	31.4	38.7	39.3	41.2	42.2	42.9	45.0	45.1	. . .
2010	28.6	36.2	36.6	38.3	38.8	38.8	40.1	43.5	45.3
NOTE: . . . = not applicable.

SOURCE: Authors' calculations based on TRF10 and DAF11.

NOTE: Cohort is determined by the most recent ticket assignment for participants aged 18–64 during the month of assignment.

Chart 2.
Percentage of TTW participants with earnings of $12,000 or more, by ticket-assignment cohort: 2002–2010

Show as table

Table equivalent for Chart 2. Percentage of TTW participants with earnings of $12,000 or more, by ticket-assignment cohort: 2002–2010

Year	Ticket-assignment cohort
	2002	2003	2004	2005	2006	2007	2008	2009	2010
2002	3.6	. . .	. . .	. . .	. . .	. . .	. . .	. . .	. . .
2003	6.7	4.3	. . .	. . .	. . .	. . .	. . .	. . .	. . .
2004	7.8	8.2	4.3	. . .	. . .	. . .	. . .	. . .	. . .
2005	8.8	10.2	8.3	4.6	. . .	. . .	. . .	. . .	. . .
2006	9.3	10.9	9.9	8.6	4.4	. . .	. . .	. . .	. . .
2007	9.4	11.5	10.7	10.0	7.8	3.8	. . .	. . .	. . .
2008	9.3	11.6	11.0	10.4	8.8	6.5	3.7	. . .	. . .
2009	8.7	10.9	10.4	10.0	8.6	6.8	5.5	2.9	. . .
2010	8.1	10.2	9.8	9.4	8.1	6.5	6.0	4.6	3.0
NOTE: . . . = not applicable.

SOURCE: Authors' calculations based on TRF10 and DAF11.

NOTES: $12,000 approximates an annualized equivalent of SGA.

Cohort is determined by the most recent ticket assignment for participants aged 18–64 during the month of assignment.

By showing the percentage of TTW participants with positive earnings, Chart 1 in essence shows the percentage employed in each year, for each annual ticket-assignment cohort. In every cohort before 2010, the two highest rates of work occurred in the year of assignment and the following year; the rates all declined thereafter. Cross-cohort comparisons indicate that the recession had a negative impact on beneficiary employment. The cohorts can be grouped visually into three categories. The first “group” consists of the 2002 cohort alone, which stands apart from the others with its relatively low percentage of members with earnings by the end of its assignment year and far lower percentages with earnings in each subsequent year.¹⁴ The second group consists of the 2003–2007 cohorts (the solid lines in Chart 1), each of which had assignment-year work rates in the 52–56 percent range. After peaking 1 year after assignment, the 2003–2005 cohorts experienced modest work-rate declines in the subsequent years. For all five cohorts, the percentages employed dropped sharply in 2008 and beyond. The third group comprises the 2008–2010 cohorts, which all exhibit worsening employment outcomes almost immediately after assignment. The assignment-year employment rate for the 2008 cohort was 3 percentage points lower than that for the 2007 cohort, and the corresponding value for the 2009 cohort was approximately 5 percentage points lower still. As noted earlier, this decline might reflect the economic downturn apart from the regulatory changes, but it also might highlight an increased propensity for providers to accept assignments from beneficiaries who were less likely to achieve earnings levels that would lead to benefit suspension or termination.

Although it is encouraging that a majority of participants had some earnings in their first year after ticket assignment, it turns out that earnings levels for many were very low, either because they earned very little each month or because they had earnings in only a few months of the year. Participants who earn above the SGA amount are of special interest, both because that amount is the conceptual standard for DI eligibility and because of the implications for forgoing benefits. Because the SGA amount is a monthly value ($1,000 for nonblind beneficiaries in 2010) but Master Earnings File data are only reported annually, we compare earnings to the annualized equivalent of the SGA amount for a nonblind beneficiary in 2010 ($12,000).

Compared with the share of TTW participants who had any earnings, relatively few had earnings above the annualized SGA level (Chart 2). However, the increasing share of participants with earnings above the annualized SGA level suggests that those who did achieve that threshold tended to sustain it. In each cohort, 3–5 percent of participants had earnings of $12,000 or more by the end of the assignment year. Although the percentage with earnings of at least the SGA level remained relatively low, it increased in the first year after ticket assignment for every cohort followed at least 1 year. Participants who assigned their tickets after 2006 did not achieve the same measure of success reached by those who assigned their tickets earlier. The 2009 and 2010 cohorts especially had lower percentages of beneficiaries with earnings of at least the annualized SGA amount in their assignment year, and the trajectories for the 2007 and 2008 cohorts are lower than those for earlier cohorts. It is difficult to know the extent to which this outcome reflects weakness in the labor market. Once again, the 2008 regulatory changes may have brought about an increase in ticket assignments from beneficiaries who were less likely to attain SGA-level earnings.¹⁵

The contrast in patterns between any earnings and SGA-level earnings is interesting. The downward trajectory of any earnings in Chart 1 suggests that a considerable share of participants initially works at very low levels but is unable to sustain work. The general upward trajectory in Chart 2 shows that participants who are initially able to engage in SGA tend to continue to do so. For each cohort, the set of beneficiaries earning at the SGA level might have changed from year to year, but it seems likely that most who attained that threshold in 1 year also did so in others. Thus, it appears that a small minority of participants achieves and sustains SGA-level earnings.

NSTW Status Among TTW Participants and Nonparticipants

The main goal of TTW is to help beneficiaries to return to work at a level that results in benefit suspension or termination. To ascertain participants' success in achieving that goal, for each calendar year we identify beneficiaries who experienced NSTW status for the first time. We define that as occurring in the month in which a beneficiary first attains NSTW status following a full calendar year in current-pay status. In other words, a beneficiary with an initial month in NSTW status in 2006 is counted as a “first NSTW” beneficiary only if he or she was in current-pay status in every month of 2005.¹⁶ This definition allows us to consider participants and nonparticipants analogously, and it excludes participants who were already in an extended period of NSTW status when they assigned their tickets.

TTW participants achieved their first NSTW month in higher percentages than did nonparticipants in each calendar year (Table 3). For example, in 2007, 3.4 percent of TTW participants experienced their first NSTW status, compared with 0.7 percent of nonparticipants. As a result, a disproportionate share of first NSTW months accrued to participants—10.6 percent of first NSTW months in 2010, although participants constituted only 2.6 percent of all those eligible for first NSTW status in that year (not shown).¹⁷ That difference reflects demographic factors and the work orientation of TTW participants relative to other beneficiaries in addition to any impact of services delivered. The frequency of first NSTW months increased steadily among participants and edged slightly upward among nonparticipants from 2002 through 2007, before declining for both groups thereafter. The decline in the later years could be a product of the recession, or could in part reflect the data-processing lags mentioned earlier.

TTW participants are more likely than nonparticipants to attain first NSTW status, but those who do so are not necessarily more likely than nonparticipants to retain NSTW status in each subsequent month (Chart 3). To make this observation, we followed all beneficiaries with a first NSTW month in 2002 and determined their NSTW status in each subsequent month through 2010. For example, Chart 3 shows that 52 percent of DI-only nonparticipants who achieved a first NSTW month in 2002 were in NSTW status 36 months later. In fact, for those with a first NSTW month in 2002, DI-only nonparticipants were in NSTW status in higher proportions than were DI-only participants in each subsequent month for 8 years. We find a similar pattern among concurrent DI-SSI participants and nonparticipants through almost all of the 60 months (5 years) after the first NSTW month. The opposite is true for SSI-only recipients, however; participants were in NSTW status in greater proportions than were nonparticipants in each of the 96 months (8 years) after the first NSTW month. Interestingly, for DI-only participants (and, to a lesser extent, for SSI-only and concurrent DI-SSI nonparticipants), the percentage in NSTW status remained approximately constant from about month 36 to month 72, suggesting sustained employment. The decline in the percentages resumed after the recession—from months 72 to 96, which correspond to calendar years 2008 through 2010.

Chart 3.
Beneficiaries who first attained NSTW status in 2002: Percentage retaining NSTW status in each of the 96 months following attainment, by benefit type ^a and TTW participation status

Show as table

Table equivalent for Chart 3. Beneficiaries who first attained NSTW status in 2002: Percentage retaining NSTW status in each of the 96 months following attainment, by benefit type ^a and TTW participation status

Months after first NSTW month	TTW participants			Nonparticipants
	DI-only	Concurrent DI-SSI	SSI-only	DI-only	Concurrent DI-SSI	SSI-only
0	100.0	100.0	100.0	100.0	100.0	100.0
1	90.6	86.4	48.7	89.5	70.9	50.0
2	84.6	72.9	46.0	85.8	65.5	43.9
3	77.8	59.3	43.9	82.5	60.9	38.5
4	73.5	64.4	39.6	79.5	56.8	34.4
5	71.8	52.5	41.2	77.3	54.8	35.1
6	68.4	50.8	35.3	75.2	52.4	33.1
7	67.5	49.2	33.7	73.3	50.3	30.4
8	65.0	44.1	34.8	72.0	48.1	28.2
9	65.8	45.8	33.7	70.3	47.6	27.7
10	61.5	42.4	31.6	69.1	46.9	26.6
11	61.5	40.7	32.6	68.2	46.6	26.7
12	59.0	44.1	35.3	67.2	45.9	27.6
13	57.3	40.7	31.0	65.9	43.7	24.2
14	56.4	40.7	31.0	65.1	42.9	23.8
15	56.4	42.4	31.0	64.1	42.0	22.6
16	54.7	40.7	35.3	62.9	41.7	22.0
17	55.6	39.0	32.1	62.6	41.1	22.8
18	55.6	32.2	28.9	61.7	39.9	21.0
19	49.6	30.5	29.4	61.1	39.2	20.6
20	48.7	25.4	31.0	60.4	39.0	20.0
21	47.9	28.8	27.3	59.8	38.7	19.5
22	47.0	23.7	29.4	59.0	38.2	19.4
23	47.9	25.4	27.3	58.6	38.2	20.2
24	47.0	27.1	25.1	58.1	38.2	19.6
25	46.2	27.1	26.2	57.3	37.3	19.0
26	45.3	28.8	26.7	56.7	36.8	18.7
27	42.7	25.4	23.0	55.9	36.0	18.2
28	42.7	27.1	25.1	55.5	36.2	18.4
29	43.6	25.4	23.0	54.9	35.8	18.2
30	41.9	27.1	25.1	54.6	35.4	17.8
31	41.0	27.1	25.1	54.0	35.0	17.8
32	41.9	27.1	26.7	53.5	34.6	17.4
33	42.7	30.5	24.1	52.4	34.2	17.3
34	41.9	28.8	25.1	52.1	33.6	17.2
35	43.6	28.8	25.7	51.8	33.7	17.5
36	44.4	28.8	21.9	51.5	33.5	17.1
37	43.6	27.1	21.9	51.1	33.1	17.0
38	44.4	25.4	20.3	50.8	33.0	16.6
39	44.4	22.0	17.6	50.5	32.7	16.5
40	42.7	23.7	21.4	50.1	32.4	16.7
41	42.7	28.8	18.7	49.8	32.1	16.4
42	42.7	25.4	18.7	49.4	31.8	16.2
43	42.7	23.7	21.9	49.1	31.6	16.2
44	42.7	25.4	20.3	48.8	31.3	15.9
45	42.7	23.7	20.9	48.5	31.1	15.9
46	43.6	22.0	21.4	48.2	31.1	16.2
47	41.9	25.4	20.9	47.9	30.9	16.0
48	41.0	27.1	19.3	47.6	31.1	15.9
49	41.0	25.4	19.8	47.3	30.7	15.7
50	40.2	25.4	21.4	47.0	30.3	15.4
51	40.2	22.0	21.9	46.8	30.0	15.7
52	40.2	22.0	20.9	46.5	29.9	15.7
53	40.2	25.4	20.9	46.2	29.6	15.4
54	41.0	23.7	20.3	46.0	29.4	15.3
55	41.9	23.7	22.5	45.6	29.6	15.3
56	39.3	27.1	21.4	45.3	29.3	15.0
57	39.3	25.4	21.4	45.0	29.1	15.2
58	40.2	25.4	21.9	44.8	29.1	15.4
59	40.2	25.4	20.9	44.5	28.5	15.1
60	40.2	25.4	21.4	44.3	28.6	15.2
61	40.2	27.1	20.9	44.1	28.4	14.7
62	39.3	27.1	19.8	43.8	28.3	14.8
63	39.3	27.1	20.9	43.6	28.3	15.1
64	39.3	28.8	21.4	43.3	28.1	14.6
65	38.5	25.4	20.3	43.1	27.8	14.6
66	37.6	22.0	18.7	42.8	27.6	14.4
67	37.6	18.6	17.6	42.6	27.4	14.3
68	37.6	23.7	18.7	42.3	27.1	14.5
69	36.8	25.4	18.7	42.0	27.1	14.4
70	36.8	27.1	18.2	41.8	26.7	14.2
71	37.6	27.1	17.6	41.5	26.5	14.1
72	37.6	27.1	19.3	41.2	26.4	14.1
73	37.6	27.1	18.2	41.0	25.9	14.0
74	37.6	25.4	19.3	40.7	26.0	13.7
75	36.8	25.4	18.7	40.4	25.7	13.6
76	35.0	23.7	17.6	40.1	25.6	13.3
77	35.0	22.0	17.6	39.9	25.3	13.2
78	35.0	22.0	17.6	39.5	25.0	12.9
79	35.0	22.0	16.0	39.2	24.7	12.8
80	35.9	20.3	16.0	38.9	24.7	12.6
81	35.9	22.0	15.5	38.6	24.5	12.6
82	35.9	23.7	15.5	38.3	24.1	12.3
83	36.8	20.3	15.5	38.0	24.1	12.3
84	36.8	22.0	15.0	37.7	23.9	12.3
85	36.8	22.0	17.1	37.4	23.6	12.2
86	36.8	22.0	18.2	37.1	23.4	12.0
87	35.9	20.3	16.6	36.8	23.1	11.9
88	35.9	23.7	16.6	36.5	22.9	11.7
89	35.9	22.0	15.5	36.3	22.8	11.6
90	35.0	20.3	13.4	36.0	22.5	11.5
91	35.0	20.3	13.4	35.8	22.4	11.4
92	35.0	18.6	13.9	35.5	22.3	11.2
93	35.0	22.0	13.4	35.2	22.1	11.0
94	35.0	22.0	11.8	35.1	21.8	11.0
95	34.2	20.3	10.7	34.8	21.7	10.8
96	34.2	20.3	11.2	34.7	21.6	10.8

SOURCE: Authors' calculations based on TRF10 and DAF11.

NOTES: Data are for beneficiaries aged 18–64 who are eligible for DI or SSI benefits. TTW participants are those whose most recently assigned tickets remain in use, regardless of the year of assignment. Nonparticipants are all other beneficiaries, including former or future TTW participants with tickets not currently assigned.

First NSTW status occurs in the first month in which a beneficiary's administrative record is coded for suspended or terminated benefits following the full calendar year 2001 in current-pay status. (For recipients of concurrent DI-SSI benefits, this definition requires current-pay status in both programs for the entire calendar year 2001.) Eligibility for first NSTW status is restricted to beneficiaries aged 18–64 who had at least 1 month in 2002 in current-pay or NSTW status.

a. As of the first month that benefits are in current-pay, suspended-for-work, or terminated-for-work status.

Among TTW participants, throughout the study period, SSI-only recipients were the most likely to enter NSTW status for the first time, with entry rates varying from 2.3 percent to 4.1 percent per year (Table 3). Despite higher entry rates for SSI-only recipients, the likelihood of retaining NSTW status after initial entry is highest among DI-only beneficiaries, followed by concurrent DI-SSI beneficiaries and then SSI-only recipients (Chart 3; note the isolated brief reversals between concurrent DI-SSI beneficiaries and SSI-only recipients). As shown by the sharp initial drop in the chart's blue lines, many SSI-only recipients in first NSTW status returned to current-pay status almost immediately; regardless of TTW participation, more than half of them left NSTW status in the month after they first attained it. The stronger retention of NSTW status among DI-only and concurrent DI-SSI beneficiaries likely reflects differences in SSI and DI rules. Although crossing the threshold from current-pay to suspended status in SSI might result in a very small change in payment amounts, crossing that threshold in DI produces a large reduction in benefits. Hence, an informed DI beneficiary might avoid crossing that threshold before securing sustainable earnings above the SGA level, whereas an SSI-only recipient might give an earnings increase that crosses the threshold no more consideration than any other small change in earnings.

Although TTW nonparticipants are somewhat more likely to retain NSTW status once they first achieve it, they are much less likely than participants to attain a first NSTW month. In fact, the share of all eligible participants who entered NSTW status in 2002 and were in NSTW status 48 months later more than doubles the share for nonparticipants (not shown).

TTW Participation and NSTW Outcomes Before and After the 2008 Regulatory Changes

The revised TTW regulations were intended to increase overall program participation, use of the EN payment systems, and the number of participants who work and earn enough to forgo their reliance on benefits. To assess whether the latter occurred, we compare the experiences of two TTW cohorts, respectively comprising members who assigned their tickets before (July 2006–June 2007) and after (July 2008–June 2009) the regulatory change. These “before” and “after” cohorts are relatively similar in size and composition to the calendar year 2007 and 2009 cohorts described in Table 2, but we have redefined the assignment periods so that neither cohort substantially straddles the regulatory changes, which took effect in July 2008.¹⁸ For each cohort, we identify the number and proportion of participants with a first NSTW month in the 18 months following assignment.¹⁹ We also report the number of NSTW months observed during that period per 1,000 participants and per participant with at least 1 NSTW month.

Despite high growth in both TTW participation overall and EN payment-system use after June 2008, there was no comparable growth in the number of participants achieving NSTW status, which translates to a cross-cohort decline in the proportion of participants experiencing NSTW status (Table 4). Specifically, members of the after cohort who assigned tickets to an EN were 43 percent less likely to experience at least 1 NSTW month in the 18 months following ticket assignment than were those in the before cohort. If we apply the data-processing factors described earlier to account for lags in NSTW determination, we calculate a 37 percent decline—slightly lower, but still substantial.

The recession almost certainly explains some and possibly most of the adjusted decline in the percentage of participants with at least 1 NSTW month, but other explanations may also apply. In particular, the regulatory changes increased incentives to serve beneficiaries with a lower likelihood of sustaining high levels of earnings—a likely cause of the large increase in M-O assignments relative to O-O assignments. Note, however, that the percentage of participants attaining NSTW status dropped at similar rates for both traditional and EN payment-system participants. Presumably, the regulatory changes produced no direct effect on outcomes for those served under the traditional payment system, although they might have caused a change in the composition of participants within that group.²⁰

After the regulatory changes, the proportion of participants in EN payment systems that attained NSTW status declined; but because of the large increase in the number of participants in EN systems overall, the number that attained NSTW status actually increased. Specifically, the proportion of participants that achieved at least 1 NSTW month declined 43.0 percent (from 14.2 percent to 8.1 percent) but the overall number of EN-system participants increased by 234.4 percent, which translates to a 90.8 percent increase in the number of participants with at least 1 NSTW month.

At the same time, the number of traditional (SVRA) payment system users who achieved NSTW status decreased by 32.4 percent, reflecting the relatively low growth in the number of such participants and the large unadjusted decline in the proportion that achieved NSTW status. In combination, these changes significantly shifted the influence of the new payment systems on participant achievement of NSTW status. SVRA users represented 84 percent of participants with NSTW status in the before cohort, but only 65 percent of them in the after cohort.

If the regulatory changes improved employment outcomes over time, the likelihood of remaining in NSTW status after initial entry should have increased. Once again, the recession's effects would of course confound any apparent effect of the regulatory changes. For this analysis, we counted all the NSTW months accrued in the 18 months following assignment and generated two separate statistics: the number of NSTW months per 1,000 participants and the number of NSTW months per participant with at least 1 such month during the period.

As measured per participant with at least 1 NSTW month, the regulatory changes do not appear to have substantially changed how quickly participants accumulated NSTW months. That statistic fell by about 9 percent for participants in both traditional and EN payment systems. The apparent decline may be due to data-processing lags; as described earlier, we found NSTW status to be underestimated by 2 percent in the before cohort and 10 percent in the after cohort. The much more substantial decline in NSTW months per 1,000 participants seems almost entirely due to the declines in the proportions of participants with at least 1 NSTW month.

Changes in NSTW months per 1,000 O-O participants were qualitatively different from those for users of the other payment systems. We observe a 9.7 percent increase in the percentage of O-O participants with at least 1 NSTW month, as well as a 7.2 percent increase in NSTW months per participant with any such month. Those increases might be as much as twice as large if not for the difference in data-processing lags. The increases may reflect the shift of assignments from the O-O system to other payment systems. That is, because total provider payments under the M-O system became almost as high as those available under the O-O system after July 2008, the cases that providers now accept in the O-O system are more likely to be for clients whom the providers judge to have a very high probability of achieving NSTW status over a sustained period. As shown elsewhere, almost all providers that continued to accept tickets under this system after the regulatory change used a consumer-directed service model that is attractive by design to participants who are ready to give up their benefits quickly (Schimmel and others 2013).

Conclusion

As documented in this study and elsewhere, the TTW program was in gradual decline from 2004 through 2007, both in terms of provider interest and the number of tickets assigned under the new payment systems. Early statistics on NSTW status showed that some participants were in fact giving up their benefits for work, but the numbers were small, and it was unclear how many of those participants would have done so even in the absence of TTW. Analyzing providers' revenues and costs also suggested that few providers were likely to find TTW economically attractive (Thornton and others 2007; Stapleton and others 2008). TTW was found to have a positive impact on employment service use, but no evidence emerged of an impact on months in NSTW status (Stapleton, Mamun, and Page 2014).

Based on the statistics presented in this article, we conclude that the 2008 regulatory changes had the intended effect of making TTW more attractive to providers and contributed to rapid growth in ticket assignments to providers using the EN payment systems. The number of new assignments increased from just over 66,000 in 2007 to nearly 94,000 in 2010 (Table 2). Likewise, Schimmel and others (2013) found substantial increases in the number of providers entering the TTW market following the 2008 regulatory changes, in contrast with stagnation in the number of new providers in the preceding years. Taken together, these outcomes support the conclusion that the regulatory changes succeeded in increasing TTW participation.

The overall growth occurred during a severe recession, making it all the more impressive and reinforcing the view that the regulatory changes made the program much more attractive to providers. The recession itself may have contributed to higher numbers of assignments, especially assignments to SVRAs. Beneficiaries who lost jobs during the downturn may have sought help to reenter the workforce. In addition, the 2009 infusion of federal funds under the American Recovery and Reinvestment Act increased the capacity of SVRAs to serve client beneficiaries. Retroactive changes in the criteria for reporting ticket assignments to SSA might also have raised the official count of participants under the traditional SVRA system without reflecting a change in the actual number served. However, that factor alone cannot account for growth in the number of ENs and the number of assignments that they accepted.

Given both the effects of the 2007–2009 recession on the labor market and the reduction in incentives for providers to help beneficiaries give up their benefits for work, it is not surprising that the percentage of beneficiaries achieving NSTW status declined following the July 2008 introduction of the revised regulations. Comparing the July 2008–June 2009 ticket-assignment cohort with the July 2006–June 2007 cohort reveals a decline in both the proportion of TTW participants experiencing NSTW status and the number of NSTW months per 1,000 participants, although the total number of NSTW months increased because of the growth in participation.

It is not feasible to separate the extent to which the recession and the revised regulations reduced NSTW-status outcomes for participants, but it seems likely that both played a significant role. The importance of the recession is reflected in the lower employment and earnings of all beneficiaries. The role of the provider incentives is reflected in the rapid growth in M-O system use for new ticket assignments and an accompanying decline in O-O system use.

Although the statistics show that the number of beneficiaries exiting the rolls after assigning a ticket is substantial and growing, the growth to date is probably not sufficient for TTW to pay for itself via reduced benefit costs. Thornton (2012) concluded that TTW might be self-financing even if the number of beneficiaries induced to exit the rolls, at least temporarily, were just a small fraction of the 100 percent increase targeted in the Ticket Act; his analysis suggests that the breakeven value for induced exits might be as low as 2,000 per year. Beneficiaries who would have exited without services, or with SVRA services under the traditional payment system, are not counted in that figure. Our data indicate that 1,120 of the beneficiaries who assigned their ticket under a new payment system between July 2008 and June 2009 (our after cohort) had at least 1 month in NSTW status within 18 months of ticket assignment (Table 4). Even allowing for some increase in that figure after 18 months, it is well below the lower end of Thornton's breakeven range. Further, it includes an unknown number of participants who, in the absence of TTW, would have exited the program rolls either without services or with services under the traditional SVRA payment system—a number that is likely to be substantial, based on the findings of Stapleton, Mamun, and Page (2014) for new young DI-only beneficiaries.²¹ On an annual basis, the number of new participants using ENs who eventually exit the program rolls is likely to exceed substantially the number we observed over 18 months for our after cohort. Yet there is no way to know whether that growth would reflect increased use of the EN payment systems by beneficiaries who would have attained NSTW status under the pre-TTW system. It also seems likely that the breakeven value for induced exits has increased substantially since 2009 because of higher TTW administrative costs and the shift toward more milestone payments. Thus, we are unable to determine the extent to which SSA's expanded investment in TTW is paying off in terms of increased beneficiary earnings and reduced government expenditures.

The lack of evidence that TTW is producing disability program savings could mean that the whole concept is flawed as an approach to managing program costs. There are two important caveats, however. The first is that TTW may be producing savings that cannot be detected with the available statistics, although we are not optimistic that this is true. The increase in NSTW months required to produce program savings is not very large (Thornton 2012)—small enough that it is hard to detect without a pilot test of some sort, because the effect is easily obscured by confounding influences such as the recession.

The second caveat is that TTW as it currently operates is far removed from the initial concept. As noted in this article's program-history section, TTW—even as initially implemented—departed from the original concept in several very significant ways, most notably by not basing provider payments exclusively on benefit savings. The 2008 regulations were another step away from the original concept of only paying a share of disability program savings, as a way to make TTW more attractive to providers. Two other important features of the NASI Disability Policy Panel's proposal were also dropped prior to implementation: pilot testing and an initial focus on new beneficiaries.

If policymakers continue to look to TTW for disability program savings, we would encourage them to consider changes that bring the program more in line with the original design. For example, pilot programs could test payment changes that make TTW more attractive to providers by increasing the share of program savings that they receive, rather than increasing payments based on other factors. Because new and relatively young beneficiaries are substantially more likely than other beneficiaries to use their tickets and return to work, any impacts of such a pilot would be easier to detect if the test followed the NASI Panel's recommendation of including only new beneficiaries.

Interestingly, the Panel's proposal did not include an idea put forth by Berkowitz (1996) in his initial description of the TTW concept that takes the proposal to test innovations on new beneficiaries one step further. That idea is to offer a version of the ticket to DI applicants, possibly accompanied by temporary benefits and counseling. Substantial evidence, mostly emerging since 1996 and in contexts other than DI, implies that providing employment support prior to DI entry would more efficiently increase employment and reduce reliance on DI than does providing supports after program entry.²² In seeking practical means of expanding such supports to address the expected depletion of the DI Trust Fund in 2016, Congress allocated funds to SSA's 2015 budget for testing. Other approaches to providing employment support for prospective beneficiaries before they enter DI have strengths and limitations relative to an expansion of TTW for selected applicants. A distinct advantage of a TTW expansion, however, is that it would build on an existing SSA infrastructure, complete with qualified providers, and could be tested on a small scale within the context of that infrastructure.