Panel Documents

Legislation

Reports

Official Correspondence

Useful Links

Briefing Papers

Long Range Plan and Yearly Priorities

Advice Report to the Commissioner of the Social Security:

Statutory Requirements and Design Issues Related to SSDI $1 for $2 Benefit Offset Research

August 2002

The Ticket to Work and Work Incentives Advisory Panel

Table of Contents

Executive Summary and Recommendations	1
Background	4
Summary of Social Security Administration Draft Implementation Plan for the $1 for $2 Benefit Offset Demonstration Projects	7
Summary of Experts Roundtable, November 16, 2001	9
Summary of Public Comments	11
Discussion of Issues and Recommendations	12
Appendices Ticket To Work And Work Incentives Advisory Panel Design And Evaluation Committee List Of Experts, Experts Roundtable, November 16, 2001 Selected Scholarly Studies	19

EXECUTIVE SUMMARY AND RECOMMENDATIONS

In this report, the Ticket to Work and Work Incentives Advisory Panel (the Panel) provides advice to the Commissioner of Social Security on the most effective designs for demonstrations and other research associated with a proposal for a $1 for $2 benefit offset under the Social Security Disability Insurance (SSDI) program. The $1 for $2 benefit offset is intended to eliminate the "cash cliff" associated with the SSDI program, thereby encouraging work and providing savings to the Trust Fund. A demonstration is authorized in Section 302 of the Ticket to Work and Work Incentives Improvement Act of 1999 (the Ticket Act). The Ticket Act requires the Commissioner to conduct demonstration projects and compile other data to evaluate the $1 for $2 proposal. It also requires the Commissioner to take into account the advice of the Panel (Title III, Section 302 (b) (1)).

This advice report is based on relevant documents and testimony, including several Social Security Administration (SSA) reports considering SSA's draft evaluation plan for the $1 for $2 program. In addition, the report is based on an Experts Roundtable held in Washington, D.C., on November 16, 2001, and on public comments made before and after the roundtable. The Panel gratefully acknowledges the expert advice of Robert Moffitt, Ph.D., who drafted earlier versions of this report.

This report discusses both demonstration methodologies and nondemonstration methodologies. By demonstration methodology, we mean a methodology that uses random assignment to place current beneficiaries into treatment and control groups and measures the effects of the program by comparing outcomes for individuals in the groups. By nondemonstration methodology, we mean a methodology that involves a particular data set that is available or can be gathered, together with a statistical methodology that uses the data set to generate estimates.

PANEL RECOMMENDATIONS

Panel Recommendation 1: $1 for $2 Demonstration for Current Beneficiaries
The Panel endorses the intent of the SSA, consistent with the Ticket Act, to use a demonstration methodology to evaluate the effect of a $1 for $2 benefit offset in encouraging current beneficiaries to earn more and exit the SSDI program, as well as to evaluate associated outcomes and budgetary costs.

Panel Recommendation 2: Employment Supports
The Panel recommends that the following employment supports be in effect for both the treatment and control groups throughout the duration of the demonstration:

• Access to local community-based benefits planning services (or their reasonable equivalent).
• Access to local community-based protection and advocacy services (or their reasonable equivalent).

Top of Page
1

• Access to responsive local work incentive specialists (or their reasonable equivalent) within SSA.
• Access to ongoing, understandable information on the treatment and its interaction with other programs and services administered by SSA.

Panel Recommendation 3: Disregard Level for the $1 for $2 Current Beneficiaries Demonstration
The Panel recommends that the indexed Substantial Gainful Activity (SGA) amount be used as the disregard, consistent with the principle that beneficiaries should be held harmless relative to the current SSDI program. The Panel encourages SSA to pursue options to encourage work and reduce costs consistent with this principle.

Panel Recommendation 4: Sample Size for the Exit and Work Incentives Demonstration
The Panel endorses the proposal by SSA to use 5,000 individuals in each treatment group and in each control group for the demonstration.

Panel Recommendation 5: Accounting Period
The Panel recommends a monthly accounting period for the demonstration projects. To simplify accounting for both the beneficiaries and SSA, the accounting system should be consistent with that used for the Supplementary Security Income (SSI) program.

Panel Recommendation 6: Deferring the Induced Entry Evaluation
The Panel recommends that SSA undertake the current beneficiary demonstration on the $1 for $2 benefit offset, deferring the induced entry evaluation. This alternative was raised in the Tuma report dated November 5, 2001. After lengthy deliberation with experts and beneficiaries, the Panel thinks this approach is cautious and cost effective. The Panel further recommends that SSA request a technical amendment from Congress to permit a delay in the evaluation of induced entry.

Panel Recommendation 7: Induced Entry Evaluation Proposal
The Panel recommends that, when SSA decides to undertake the evaluation of induced entry, it seek at least five independent designs from outside experts and organizations, each proposing the most cost-effective methodology (either demonstration or nondemonstration) to measure induced entry.

SSA should report the proposed designs to the Panel so that the Panel can compare them with SSA's draft recommendation involving the National Survey of Health and Activity (NSHA) as the data set and a structured statistical modeling approach to translate the data into induced entry estimates. This comparison will permit the Panel to make a more informed judgment on nondemonstration and demonstration evaluation alternatives in providing its advice to SSA. Further, the Panel recommends that SSA seek a technical amendment from Congress to permit SSA to implement either a demonstration or a nondemonstration methodology in studying induced entry associated with the $1 for $2 benefit offset.

Top of Page
2

Panel Recommendation 8: Other Studies, Including Studies of the Determinants of Return to Work
The Panel recommends that SSA's Office of Policy (OP) follow and report on parallel research projects across the country that inform on determinants of return to work and related issues, and incorporate relevant findings into SSA research on the $1 for $2 benefit offset or on topics related to Section 302 in general.

Top of Page
3

BACKGROUND

Currently, an SSDI beneficiary may have 9 months of "trial work" within a 60-month rolling window-this is known as the Trial Work Period (TWP). Any month during which a beneficiary earns more than $560¹ counts as one of the 9 months of the TWP. A 36-month Extended Period of Eligibility (EPE) follows the TWP. During the TWP, the first month of the EPE in which a beneficiary earns more than the Substantial Gainful Activity (SGA), and next 2 months (known as the grace period), the beneficiary receives full cash benefits regardless of earnings level. For the remainder of the EPE following the 3-month grace period, no cash benefits are paid in a month when the beneficiary earns more than the SGA amount (currently $780 per month),² but eligibility for cash benefits and health insurance continues. Following the EPE, benefit eligibility ends after the first month a person earns more than the SGA, and the person's case with SSA is closed. Any subsequent disability benefit payments require a new application.

Thus, after the TWP and the grace period, a beneficiary who earns even $1 more than the SGA receives no cash benefits. Many consider this so-called "cash cliff" to be a major deterrent to work. Under the $1 for $2 benefit offset proposal, as earnings exceed a specified disregard amount, benefits would be reduced by $1 for each $2 of earnings. Thus, a beneficiary whose earnings exceed the SGA would no longer be subject to a sudden loss of all cash benefits.

The Ticket to Work and Work Incentives Improvement Act (TWIIA) of 1999 includes a requirement that SSA conduct a demonstration project to evaluate the cost and other effects of an SSDI $1 for $2 benefit offset designed to eliminate the cash cliff. This mandate is found in Section 302 of the Act:

Sec. 302. DEMONSTRATION PROJECTS PROVIDING FOR REDUCTIONS IN DISABILITY INSURANCE BENEFITS BASED ON EARNINGS-

(a) Authority.-The Commissioner of Social Security shall conduct demonstration projects for the purpose of evaluating, through the collection of data, a program for Title II disability beneficiaries (as defined in section 1148(k)(3) of the Social Security Act) under which benefits payable under section 223 of such Act, or under section 202 of such Act based on the beneficiary's disability, are reduced by $1 for each $2 of the beneficiary's earnings that is above a level to be determined by the Commissioner. Such projects shall be conducted at a number of localities which the Commissioner shall determine is sufficient to adequately evaluate the appropriateness of national implementation of such a program. Such projects shall identify reductions in Federal expenditures that may result from the permanent implementation of such a program.

¹The threshold used in the Trial Work Period is indexed annually.
²The Substantial Gainful Activity amount is the level of earnings beyond which the recipient is regarded as having the capability for work. The SGA is indexed annually.

Top of Page
4

(b) Scope and Scale and Matters to Be Determined.-

      (1) In general-The demonstration projects developed under subsection (a) shall be of sufficient duration, shall be of sufficient scope, and shall be carried out on a wide enough scale to permit a thorough evaluation of the project to determine-

1. the effects, if any, of induced entry into the project and reduced exit from the project; the extent, if any, to which the project being tested is affected by whether it is in operation in a locality within an area under the administration of the Ticket to Work and Self-Sufficiency Program established under section 1148 of the Social Security Act; and
2. the savings that accrue to the Federal Old-Age and Survivors Insurance Trust Fund, the Federal Disability Insurance Trust Fund, and other Federal programs under the project being tested.
   
   The Commissioner shall take into account advice provided by the Ticket to Work and Work Incentives Advisory Panel pursuant to section 101(f )(2)(B)(ii) of this Act.

This Advice Report represents the Panel's effort to bring expertise to bear in advising the SSA on the complex issues underlying the design of research related to the $1 for $2 benefit offset. Over recent months, the Panel has carefully reviewed these issues. This review has involved consideration of the following SSA reports:

• $1 for $2 Benefit Offset: Exploring the Feasibility of Measuring Induced Entry and Evaluating Potential State-Level Project Designs (Weathers, Muller, and Hennessey, Office of Policy, September 14, 2000);
• $1 for $2 Benefit Offset Demonstration Projects: Draft Implementation Plan (Office of Policy, April 25, 2001);
• $1 for $2 Benefit Offset: Exploring the Feasibility of Measuring Induced Entry and Evaluating Potential County-Level Project Designs (Muller, Weathers, Hennessey, and Bellemore, Office of Policy, June 12, 2001); and
• Approaches to Evaluating Induced Entry Into a New SSDI Program with a $1 Reduction in Benefits for Each $2 in Earnings (Tuma, Office of Policy, November 5, 2001).

The Panel's review has also included consultation with SSA, other Federal agencies, and various experts; opportunities for public comment; and Panel deliberations. The Panel's Design and Evaluation Committee convened an Experts Roundtable of nationally recognized experts, SSA experts, and Panel members on November 16, 2001. The Roundtable addressed issues related to the specifications for the $1 for $2 demonstration and, by implication, the features of a possible future benefit offset provision that would

Top of Page
5

be implemented nationwide. Subsequently, a well-known expert, Robert Moffitt, Ph.D., assisted the committee and the Panel in drafting this report. The Panel looks forward to continuing its collaboration with SSA as the $1 for $2 research goes forward, including regular progress reports.

Top of Page
6

This section is a brief summary of the SSA report $1 for $2 Benefit Offset Demonstration Projects: Draft Implementation Plan (Office of Policy, 2001).

The central element of SSA design is a random assignment trial, or demonstration, intended to estimate the effects of the $1 for $2 program on current beneficiaries in a variety of dimensions: Perhaps the most important aspect is estimating the effects of the program on employment, wages, occupations, benefits, income, and program entry/exit for beneficiaries who are offered the program. In addition, the design is intended to estimate costs, Trust Fund savings, the determinants of return to work, and the extent to which the $1 for $2 program will interact with the Ticket Program. A second component of the design is a plan to estimate induced entry, for which nondemonstration methods are proposed.

The basic components of the SSA design are as follows. Approximately 5,000 current beneficiaries will be selected for each of two treatment groups, and 5,000 will be selected for a control group. One treatment group will have a limited duration of eligibility for the $1 for $2 benefit offset, possibly the length of the current EPE, and the other treatment group will have an unlimited duration of eligibility. The participants will be randomly selected from several areas around the country to make the sample representative of the nation as a whole. The sample will be stratified into groups-recent SSDI beneficiaries and longer-term beneficiaries. The long-term effects of the program are best estimated using the former group, and the short-term effects are best estimated using the latter group. Outcomes will be measured by periodic surveys of the treatment and control groups to gather information on health, economic situation, household composition, work expectations, and use of employment services, one-stop centers, vocational service providers, and Employment Networks. A second, silent control group of 5,000 individuals will be included to compare with the main control group to determine if there are any effects of the monthly accounting and surveys per se. The surveys and other data-gathering activities that measure the outcomes for the treatment and control groups will continue for at least 5 years, perhaps longer. Participants who have received lifetime guarantees of the program will continue to receive the $1 for $2 treatment even after the evaluation has ended. Hence, the total number of participants will be 20,000, but 5,000 of these will constitute a silent control group that will neither receive a treatment nor be surveyed.

SSA has taken the position that it will not specify many of the key details of the design at this time but rather will hire a contractor to work on those details with SSA, in light of whatever cost constraints are imposed on the project design.³ Among the many key

³After the April 25, 2001, report was written, SSA informed the Design and Evaluation Committee that it would select the contractor by issuing a Request for Task Order Proposals.

Top of Page
7

details not yet specified are whether a second, lower level of earnings disregard will be tested in addition to the earnings disregard set at the SGA; the exact time limit for eligibility; the accounting period for adjusting benefits to earnings; and the statistical power of a demonstration with 5,000 experimentals and 5,000 controls.

The study of induced entry will be conducted separately using a nondemonstration methodology. Estimates of the number of eligible nonparticipants will be obtained from the planned 5,000-sample National Survey of Health and Activity (NSHA), which will be taken as an upper bound for the magnitude of induced entry. Estimates of the fraction of eligible nonparticipants in the NSHA who would apply for SSDI and be accepted into the program will be based on modeling techniques, including structural modeling. SSA has not worked out the details of the modeling method. Some direct questioning of individuals and focus group analysis will be used to supplement the modeling estimates.

Top of Page
8

An Experts Roundtable was convened by the Design and Evaluation Committee of the Work Incentives Advisory Panel on November 16, 2001, in Washington, D.C. The Panel assembled experts in several fields, with the main focus being a discussion of the SSA plan for evaluating the effects of $1 for $2 benefit offset on current beneficiaries and on induced entry. (For a list of participants, see Appendix C.) Experts from SSA's Office of Policy (OP) were also in attendance and described the evaluation plan.

EXPERT COMMENTS ON EXIT DEMONSTRATION

• There was a concern that increases in income among $1 for $2 participants could jeopardize their eligibility for Medicaid benefits.
• It was noted that SSDI recipients fear changes in the program that may cause them to lose eligibility down the line when the program is changed again. This fear needs to be addressed.
• There was some concern that the sample sizes of experimentals and controls planned by the OP are not large enough for subgroup analysis.
• Some experts were concerned that the disregard choice option⁴ may not be what Congress wanted to have studied.
• One expert opposed testing the disregard choice option and recommended only disregards at or above the current SGA.
• The experts generally believed that testing disregards at or above the SGA would not allow reliable extrapolation to disregards significantly below the SGA.
• It was suggested that disregards significantly below the SGA would not work as long as recipients had the freedom to resign from the experiment and return to the existing program.
• Although the program uses a single definition of disability, it was noted that the definition might be operationalized somewhat differently in different States, a situation that would complicate the choice of demonstrations if only a few areas were used.
• An expert recommended that SSA include in its design more consideration of what it takes to get disability insurance (DI) recipients to go to work and whether a more comprehensive set of benefits and services is necessary.

EXPERT COMMENTS ON INDUCED ENTRY

• One expert believed that a 200-county demonstration should be pursued and proposed to Congress even though its estimated cost is $500 million, and that Congress should be asked to appropriate such funds.

⁴The disregard choice option, explained in more detail below, is the option of offering beneficiaries in the demonstration the choice of receiving either benefits under the existing formula with current SGA disregards or benefits under a $1 for $2 sliding scale with a lower disregard.

Top of Page
9

• It was argued that the entry demonstration would have to test disregard levels at or above the current SGA to avoid harming any current or potential beneficiary.
• Some experts expressed concern with the response rates expected for the NSHA, arguing that estimates based on a survey with such response rates would lack prima facie credibility and have potential for bias.
• It was noted that procedures exist in surveying a population to test for the presence of bias due to nonresponse.
• There was a concern that the 5,000-observation sample size would be too small to allow estimates of induced entry, because the latter would occur only in a narrow subset of eligible nonparticipants.
• It was suggested that the questions on disability planned for the NSHA could have a deleterious effect on the measurement of disability and that the medical exam planned as part of the NSHA would not sufficiently replicate the SSA test for some impairments, such as those related to mental function.
• One expert suggested that the NSHA should oversample young people, because that is where the costs of induced entry are likely to be highest.
• One expert asserted that hypothetical questions about whether people would apply for DI if a $1 for $2 program were in place would be unreliable.
• One expert stated that neither demonstration nor nondemonstration methods for estimating induced entry have acceptable margins of error.
• More than one expert recommended that OP consider data sets other than NSHA, such as the Medical Expenditure Surveys, Health and Retirement Survey, National Health Interview Survey, census reports, and others.
• The experts suggested that data sets that are longitudinal and follow individuals over time have advantages over the NSHA.
• One expert recommended having focus groups of benefits counselors to estimate induced entry.
• One expert argued that induced entry itself is not a bad thing; that cost is the major issue, and that costs of other programs might be reduced if those entering DI would have been on other programs instead.

Top of Page
10

There was no time for public comment at the November 16 meeting, but there was public comment at a Panel meeting on November 15, 2001. The public made the following points at that meeting:

• The induced entry issue deserves primacy because the SSA Office of the Chief Actuary and the Congressional Budget Office (CBO) estimated that it constitutes 80 percent of the cost of a $1 for $2 program.
• Evidence exists from a switch of plans at the UNUM Company that there would be no induced entry from the $1 for $2 benefit offset.
• The demonstration should establish how many current beneficiaries would earn more than the SGA; many more do so in the SSI program than it is estimated would do so if $1 for $2 were implemented in DI.
• The offset tested in the demonstration should begin at the current SGA, not below it, although SSA could consider holding the nominal dollar SGA fixed over the lifetime of the demonstration to allow some SGA variation.
• Allowing beneficiaries to choose different SGA options would be too confusing to the beneficiaries; not all would be fully informed, and the choice would really be no choice at all.
• The SSA plan does not consider the blind SGA.
• One commenter said that a demonstration must be conducted to assess induced entry.
• Econometric models do nothing to advance the cause of enacting $1 for $2.

The public was also invited to submit comments on the Ticket Panel listserv. Most of the comments were not on the evaluation design per se and thus were outside the scope of this report. Most of those comments were instead speaking in favor of or against $1 for $2, with the majority-at least of those who properly understood how $1 for $2 would work-in support of it, often enthusiastically so. A few of the public comments did deal with evaluation issues, including the following:

• A suggestion that the exit experiment not be randomly assigned across the country but rather be focused on areas where strong work support services are already in place.
• A recommendation that OP study the SSI program more carefully for ideas about problems with implementation (e.g., the accounting period) that might hinder successful $1 for $2 effects.
• A concern that SSI work incentives do not work very well and that this issue needs study along with the direct DI evaluation project.

Top of Page
11

Panel Recommendation 1: The Panel endorses the intent of the Social Security Administration, consistent with the Ticket Act, to use a demonstration methodology to evaluate the effect of a $1 for $2 benefit offset in encouraging current beneficiaries to earn more and exit the SSDI program, as well as to evaluate associated outcomes and budgetary costs.

ISSUE 2: EMPLOYMENT SUPPORTS
Panel members have struggled with the parameters of treatments offered in a random assignment demonstration. The treatment used in the demonstration will be of less scientific value if it does not "imitate" the treatment SSA plans to implement in a prospective national program. Because the Ticket Act establishes specific supports, including benefits planning, protection and advocacy services, and work incentive specialists (Subtitle C, Section 121), the $1 for $2 demonstration should include such supports.

Panel Recommendation 2: The Panel recommends that the following employment supports be in effect for both the treatment and control groups throughout the duration of the demonstration:

• Access to local community-based benefits planning services (or their reasonable equivalent).
• Access to local community-based protection and advocacy services (or their reasonable equivalent).
• Access to responsive local work incentive specialists (or their reasonable equivalent) within SSA.
• Access to ongoing, understandable information on the treatment and its interaction with other programs and services administered by SSA.

ISSUE 3: DISREGARD LEVEL FOR THE $1 FOR $2 CURRENT BENEFICIARIES DEMONSTRATION In this report, "disregard" is defined as the maximum monthly earnings that a beneficiary can have with no reduction in benefits. For example, the disregard amount implicit in the

Top of Page
12

current program is the SGA amount. The amount of the disregard has a major impact on the effects of a benefit offset program in terms of the work incentive effects of the program, the costs to the Trust Fund, and-for those who do not increase earnings for whatever reason-poverty outcomes. In coming to terms with its disregard recommendation, the Panel has struggled to balance these outcomes.

One option for the demonstration would be to disregard earnings up to the SGA and apply a $1 for $2 benefit offset for earnings above the SGA.⁵ However, some preliminary estimates suggest that adopting the SGA as the disregard would limit the number of individuals ever leaving the DI program. For example, the average cash benefit paid to SSDI beneficiaries in 2000 was almost $800 per month, implying that (assuming the SGA was adopted as the disregard) the beneficiary could earn up to roughly $2,300 (approximately $28,000 annually) before losing all cash benefits.⁶ In general, because many disability beneficiaries work at low wages during their working careers, a substantial number would probably not be able to return to work at earnings higher than their break-even amount.

Thus, using the SGA as the disregard has two implications. First, the cost of such a program may be high compared with a program that uses a lower disregard, because relatively few beneficiaries would exit the SSDI program and because there would be no savings in benefit costs until earnings increased above the SGA level. Second, many beneficiaries-especially those with low predisability earnings-may work while receiving partial cash benefits.⁷ In considering this outcome, the Panel believes that policymakers should consider partial benefit reduction not only as a successful employment outcome but also as a reasonable way of accruing additional savings to the Trust Fund.

Given the importance of the cost issue, the Panel considered disregards below the SGA for both the demonstration and a possible permanent program. However, if lower-than-SGA disregards are used, beneficiaries at some earning levels who-for whatever reason-would not increase their earnings could be worse off financially than under the existing program. The Panel is especially concerned about beneficiaries, including many with significant disabilities, who work part-time or at low wages. Thus, the Panel tried to define specifications that would hold beneficiaries harmless under a new benefit offset program to ensure that they would be no worse off financially than under the current program.

⁵Consistent with current program procedures, the $1 for $2 benefit offset would be applied after the expiration of the TWP and the 3-month grace period.
⁶The break-even point-the point at which benefits reach zero as the beneficiary increases earnings-will vary from one beneficiary to another depending on the size of his or her benefit. The annual break-even figure is calculated as 12 x (disregard + (2 x monthly benefit)). A fraction of 1 percent of all SSDI beneficiaries could have earnings over $60,000 before losing all benefits: Specifically, this could occur for high earners with one or more dependents.
⁷The Panel points out that Ticket Program policies would have to be modified in order for Employment Networks to provide employment supports for those with partial benefits.

Top of Page
13

There are additional concerns about the use of lower-than-SGA disregards for a demonstration. One is a human subjects problem, which would be formally addressed by an Internal Review Board. While the individuals could give informed consent, some beneficiaries-because of the nature of their disabilities-may not be fully able to assess their work capabilities and may make poor choices when deciding whether to participate in a demonstration involving a disregard below the SGA. A second problem is that, because participation in a demonstration would necessarily be voluntary, any beneficiary who found himself or herself in a worse position under the demonstration than under the existing SSDI program could quit the demonstration and return to the SSDI program. This would result in biases and incorrect estimates from the demonstration, because the treatment group would no longer represent the full set of beneficiaries who would be subject to a $1 for $2 program if one were implemented nationally at that disregard level.

Nevertheless, from a scientific point of view, most of the experts from the Roundtable agree that it would be preferable to test disregards below the SGA. Estimates of the effects of $1 for $2 at disregard levels below the SGA could be obtained by extrapolating downward from tests of various disregard levels at and above the SGA. However, the experts believe that such extrapolation would be statistically unreliable.

An alternative that would achieve some cost savings but still be feasible would be to offer beneficiaries the choice of two programs: the existing program or a $1 for $2 program at a disregard level below the current SGA. The Panel has explored offering this choice to a treatment group under a demonstration or to all beneficiaries as part of a permanent program. The advantage of the choice option is that it would cost less than a benefit offset program that uses the current SGA as the disregard. Implemented for a demonstration, this option would largely allay human subject and ethical concerns, because a beneficiary who found the $1 for $2 program disadvantageous could switch out of it. However, a limit would have to be established on how often beneficiaries could switch programs.

A possible disadvantage of using the choice option in a demonstration if the option is not proposed for the permanent program is that it would not yield estimates of the effect of a $1 for $2 program at the current SGA disregard level or a lower level; that is, it would not allow estimates of either of the two elements of the choice option alone. However, the seriousness of this disadvantage depends partly on whether Congress is interested in this type of program. Using the choice option would also require another treatment group in the demonstration. If cost is a consideration, OP could drop one of its existing treatment groups and test only one duration.

The choice option poses another potential problem, whether it is implemented for a demonstration or as a permanent program. If a person chooses the $1 for $2 alternative and his or her earnings do not increase as expected, the issue of unequal treatment may arise. For example, one person, having chosen the $1 for $2 program with a disregard lower than the SGA, will have his or her benefits reduced when earning at the SGA level. A second person with the same earnings, having chosen the current program, will have no benefit offset. Because both are disabled under program criteria and they have equal

Top of Page
14

earnings, this can be considered unequal treatment. Some Panel members saw this as problematic, while others concluded that such outcomes are inevitable results of public programs that offer choice.

The Panel has tried to find ways to offset the work incentive and cost effects of a $1 for $2 program that uses the SGA as the disregard, consistent with the constraint that beneficiaries be held harmless relative to the current program. One idea is to consider a higher benefit offset rate, such as $3 for $5 or $2 for $3. Another possibility considered by OP is to impose an earnings cap for eligibility. A two-tiered structure could also be considered, with high earners subject to a higher benefit offset rate for earnings above a specified threshold.8 A two-tiered structure might be preferable to an earnings cap because it avoids the cash cliff.⁸

Panel Recommendation 3: The Panel recommends that the indexed SGA amount be used as the disregard, consistent with the principle that beneficiaries should be held harmless relative to the current SSDI program. The Panel encourages SSA to pursue options to encourage work and reduce costs consistent with this principle.

ISSUE 4: SAMPLE SIZE FOR THE EXIT AND WORK INCENTIVES DEMONSTRATION

A sample of 5,000 each for the two treatment groups is likely to be sufficient if no subgroup analysis is conducted. However, subgroup analysis would be of some interest, because those with different kinds of employment barriers and difficulties may respond differently to the $1 for $2 program. However, a sample of 5,000 is enough to obtain an average estimate for the beneficiary population, and expanding the sample size to obtain subgroup impacts would be costly.

Panel Recommendation 4: The Panel endorses the proposal by SSA to use 5,000 individuals in each treatment group and in each control group for the demonstration.

ISSUE 5: ACCOUNTING PERIOD IN THE DEMONSTRATION

A monthly accounting period has several advantages for the $1 for $2 program. The most important is that it would allow beneficiaries to immediately see the gains from working, as their benefits would be adjusted soon after their work increased. This reinforcement is likely to increase the work incentives estimated in the demonstration. Another advantage is that a month is often used as the unit of time in the DI program-the SGA is calculated by the month, as are the Trial Work Period and the Extended Period of Eligibility. Also, a monthly accounting period is easier to coordinate with the Ticket Program, under which Employment Networks (ENs) are paid on a monthly basis.

A disadvantage of a monthly accounting period would be the greater administrative burden on the SSA system. In addition, without a carryover or retrospective accounting

⁸The Panel is grateful to Scott Muller and John Hennessey of the Office of Policy for sharing their thoughts on these issues.

Top of Page
15

period, benefits would fluctuate considerably if earnings do. A carryover or retrospective accounting period would smooth out those fluctuations. Quarterly or annual accounting periods are administratively easier to implement, but they make it difficult to recover overpayments, and beneficiaries do not immediately see the fruits of their increased earnings.

An accounting period must take into account concurrent beneficiaries, who may be paid under particular formulas. For example, SSI recipients are paid with a retrospective rather than prospective accounting for income. It would not be appropriate for the $1 for $2 and SSI accounting systems to be different.

Panel Recommendation 5: The panel recommends a monthly accounting period for the demonstration projects. To simplify accounting for both the beneficiaries and SSA, the accounting system should be consistent with that used for the Supplementary Security Income (SSI) program.

ISSUE 6: DEFERRING THE INDUCED ENTRY EVALUATION

Most of the experts at the Roundtable concurred with SSA experts and SSA-commissioned experts on several concerns with respect to an induced entry demonstration. While demonstration methodologies ordinarily are preferred to nondemonstration methodologies, demonstrations face unique and severe difficulties for the analysis of induced entry effects of social programs in general and the $1 for $2 program in particular. Demonstrations would have to be conducted on the city or county level, offering the $1 for $2 program to the entire population of the area. To reduce the influence of random events in individual areas to acceptable margins, a demonstration would have to include at least 200 counties or other areas. The cost of such a demonstration, according to the Tuma report, would be approximately $500 million. In addition, even with this large, costly demonstration, risks associated with the methodology could invalidate the results and produce estimates with unacceptably wide margins of error. The Panel recognizes that in some instances the gains in knowledge from an experiment may not be worth the cost.

It is possible that none of the alternatives would achieve acceptable margins of error. It may be best in that case to ask Congress to postpone the study of induced entry. The disadvantage of a postponement is that it would leave the CBO and the Office of the Chief Actuary estimates of induced entry as the only existing estimates, although the experts agree that they can be improved upon using the NSHA or other new data sets. The advantage of a postponement is that it would reflect the judgment of the Panel, based on the scientific evidence, that new estimates made from current draft plans, while possibly superior to existing estimates, would still be below the scientifically acceptable margin of error.

Panel Recommendation 6: The Panel recommends that SSA undertake the current beneficiary demonstration on the $1 for $2 benefit offset, deferring the induced entry evaluation. This alternative was raised in the Tuma report (dated November 5, 2001).

Top of Page
16

After lengthy deliberation with experts and beneficiaries, the Panel thinks this approach is cautious and cost effective. The Panel further recommends that SSA request a technical amendment from Congress to permit a delay in the evaluation of induced entry.

ISSUE 7: INDUCED ENTRY EVALUATION PROPOSALS

Producing better estimates than what we have now on induced entry and its impact on costs is of critical importance. Past estimates have suggested that the costs arising from induced entry would constitute 80 percent of the total costs of a $1 for $2 program.

Most of the experts at the Roundtable concurred with SSA experts and SSA-commissioned experts on several concerns with respect to an induced entry demonstration. While demonstration methodologies ordinarily are preferred to nondemonstration methodologies, demonstrations face unique and severe difficulties for the analysis of induced entry effects of social programs in general and the $1 for $2 program in particular. Demonstrations would have to be conducted on the city or county level, offering the $1 for $2 program to the entire population of an area. To reduce the influence of random events in individual areas to acceptable margins, a demonstration would have to include at least 200 counties or other areas. The cost of such a demonstration would be approximately $500 million. In addition, even with this large a demonstration, risks associated with the methodology could invalidate the results and produce estimates with margins of error too wide to be acceptable. Another issue for SSA's consideration is whether a state-centric design for a $1 for $2 benefit offset treatment should be considered, if such a design could be conducted without a serious threat to its validity.

Given the difficulties of feasible demonstration designs for induced entry, together with the serious threats to the validity of those designs even at their estimated budget levels, nondemonstration alternatives must be considered.

SSA has proposed to evaluate induced entry using a nondemonstration methodology. However, the SSA plan to rely heavily on the NSHA carries some risk. The NSHA will not have high response rates, which creates the risk of nonresponse bias. The medical exam to be administered to respondents may not accurately mimic the DI program's medical exam and eligibility criteria. The sample size of the NSHA is not large, and it may be difficult to obtain accurate estimates of induced entry if it occurs only in a narrow segment of the eligible nonparticipant population. Questions exist regarding the statistical methodology to be used by SSA for estimating the size of induced entry from the sample of eligible nonparticipants in the NSHA data. Because OP has not spelled out the methodology, its reliability cannot be judged. There is a risk associated with using statistical methods that rely on too many untestable assumptions.

For these reasons, it would be preferable for OP to consider additional data sets for its analysis and consider a wide range of statistical methodologies. Additional data sets could include the Medical Expenditure Surveys, the National Health Interview Survey, the Health and Retirement Survey, the Current Population Survey, the Survey of Income

Top of Page
17

and Program Participation, the National Longitudinal Surveys of Youth, and others. Statistical methodologies could include structural modeling, descriptive modeling, and variations on each.

Because of the questions about both demonstration and nondemonstration alternatives, as well as the limitations on OP staff, SSA should solicit proposals to evaluate induced entry that involve either methodology. These proposals can be compared with OP's proposal for using the NSHA and its preferred statistical modeling method, to determine the best alternative and whether multiple alternatives should be pursued.

Panel Recommendation 7: The Panel recommends that, when SSA decides to undertake the evaluation of induced entry, it seek at least five independent designs from outside experts and organizations, each proposing the most cost-effective methodology (either demonstration or nondemonstration) to measure induced entry.

SSA should report the proposed designs to the Panel so that the Panel can compare them with SSA's draft recommendation involving the National Survey of Health and Activity as the data set and a structured statistical modeling approach to translate the data into induced entry estimates. This comparison will permit the Panel to make a more informed judgment on nondemonstration and demonstration evaluation alternatives in providing its advice to SSA. Further, the Panel recommends that SSA seek a technical amendment from Congress to permit SSA to implement either a demonstration or a nondemonstration methodology in studying induced entry associated with the $1 for $2 benefit offset.

ISSUE 8: OTHER STUDIES, INCLUDING STUDIES OF THE DETERMINANTS OF RETURN TO WORK

Other private, State, or Federal research projects may involve the study of SSDI beneficiaries, the use of comprehensive employment supports, or the determinants of return to work for SSDI beneficiaries. Examples of such research include efforts by the Social Security and Rehabilitation Services Administration State Partnership Initiatives, the Robert Wood Johnson Foundation related to the Medicaid buy-in, the Medi-Cal Policy Institute, the Lewin Group, and researchers working under Medicaid infrastructure grants. Findings from such research should be incorporated into SSA research on the SSDI $1 for $2 benefit offset that will eventually be reported to the President, Congress, and the Advisory Panel.

Panel Recommendation 8: The Panel recommends that SSA's Office of Policy follow and report on parallel research projects across the country that inform on determinants of return to work and related issues, and incorporate relevant findings into SSA research on the $1 for $2 benefit offset or on topics related to Section 302 in general.

Top of Page
18

APPENDIX 1:
TICKET TO WORK AND WORK INCENTIVES ADVISORY PANEL

Sarah Wiggins Mitchell, R.N., M.S.W., J.D., Chair
President and Executive Director
New Jersey Protection and Advocacy, Inc.

Richard V. Burkhauser, Ph.D.
Professor of Policy Analysis
Chair, Department of Policy Analysis and Management
Cornell University
Ithaca, NY

Kristen E. Flaten, M. Div.
Mental Health and Benefits Advocacy Specialist
Lifetrack Resources
St. Paul, MN

Thomas P. Golden, M.S., C.R.C.
Faculty
Program on Employment and Disability
School of Industrial and Labor Relations
Cornell University
Ithaca, NY

Frances Gracechild
Executive Director
Resources for Independent Living, Inc.
Sacramento, CA

Christine M. Griffin, J.D.
Executive Director
Disability Law Center
Boston, MA

Larry D. Henderson
Executive Director
Independent Resources, Inc.
Wilmington, DE

Jerome Kleckley, M.S.W., C.S.W.
Director
Hospital Services
Eastern Paralyzed Veterans Association
Jackson Heights, NY

Top of Page
19

Stephanie Smith Lee
Governmental Affairs
National Down Syndrome Society
Oakton, VA
(On February 20, 2002, Ms. Lee resigned from the Panel to accept the position of Director, Office of Special Education Programs, U. S. Department of Education.)

Bryon R. MacDonald
Public Policy Advocate
World Institute on Disability
Oakland, CA

Stephen L. Start
President and CEO
S.L. Start Associates
Spokane, WA

Susan Webb, M.B.A., S.P.H.R.,
Director, ABIL Employment Services
Phoenix, AZ

Top of Page
20

APPENDIX 2:
DESIGN AND EVALUATION COMMITTEE

Richard V. Burkhauser, Ph.D.
Professor of Policy Analysis
Chair, Department of Policy Analysis and Management
Cornell University
Ithaca, NY

Kris E. Flaten, M. Div.
Mental Health and Benefits Advocacy Specialist
Lifetrack Resources
St. Paul, MN

Thomas P. Golden, M.S., C.R.C., Chair
Faculty
Program on Employment and Disability
School of Industrial and Labor Relations
Cornell University
Ithaca, NY

Stephanie Smith Lee
Governmental Affairs
National Down Syndrome Society
Oakton, VA
(On February 20, 2002, Ms. Lee resigned from the Panel to accept the position of Director, Office of Special Education Programs, U. S. Department of Education.)

Bryon R. MacDonald
Public Policy Advocate
California Work Incentives Initiative
World Institute on Disability
Oakland, CA

Susan Webb, M.B.A., S.P.H.R.,
Director, ABIL Employment Services
Phoenix, AZ

Top of Page
21

APPENDIX 3:
LIST OF PARTICIPANTS, EXPERTS ROUNDTABLE, NOVEMBER 16, 2001

Experts

Barbara M. Altman
Special Assistant for Disability Statistics
National Center for Health Statistics
6525 Belcrest Road
Hyattsville, MD 20782

Judith Cook, Professor
Department of Psychiatry
University of Illinois at Chicago
104 South Michigan Avenue, Suite 900
Chicago, IL 60603

Mary C. Daly
Federal Reserve Bank of San Francisco
101 Market Street-Mail Stop 1130
San Francisco, CA 94105

Martin Gerry, Research Professor
University of Kansas
1315 Wakarusa Drive
Lawrence, KS 66049

Kim Hildred
Subcommittee on Social Security
U.S. House of Representatives
B316 Rayburn House Office Building
Washington, DC 20518

Robert Moffitt
Professor of Economics
Johns Hopkins University
Department of Economics
Baltimore, MD 21218

Robert Silverstein, Director
Center for the Study and Advancement of Disability Policy
1730 K Street, N.W., Suite 1212
Washington, DC 20006

Top of Page
22

Ticket to Work and Work Incentives Advisory Panel Staff

Marie Strahan, Executive Director
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Tamara Allen
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Kristen Breland
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Gordon Richmond
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Bernard Wixon
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Theda Zawaiza, Ph.D., Consultant
Ticket to Work and Work Incentives Advisory Panel
Social Security Administration
400 Virginia Avenue, S.W., Suite 700
Washington, DC 20024

Top of Page
23

Presenters

Scott Muller
Social Security Administration
Office of Research, Evaluation and Statistics
Office of Policy
Room 4-C-15 Operations
6401 Security Boulevard
Baltimore, MD 21235

Nancy Tuma
Professor of Sociology
Stanford University

Paul N. Van de Water
Acting Deputy Commissioner
Social Security Administration
Office of Policy
Room 828 ITC Building
Washington, DC 20254

Top of Page
24

APPENDIX 4:
SELECTED SCHOLARLY STUDIES

Benitez-Silva, H., M. Buchinsky, H-M. Chan, J. Rust, and S. Sheidvasser. 1999. An empirical analysis of the Social Security disability application, appeal, and award process. Labour Economics 6:147-178.

Bound, J. 1989. The health and earnings of rejected disability insurance applicants. American Economic Review 79:482-503.

Bound, J., and R. Burkhauser. 1999. Economic analysis of transfer programs targeted on people with disabilities. In O. Ashenfelter and D. Card (Eds.), Handbook of Labor Economics. New York: Elsevier.

Bound, J., and T. Waidmann. 1992. Disability transfers, self-reported health, and the labor force attachment of older men: Evidence from the historical record. Quarterly Journal of Economics 107:1393-1419.

Burkhauser, R., and D. Wittenburg. 1996. How current disability transfer policies discourage work: Analysis from the 1990 SIPP. Journal of Vocational Rehabilitation 7:9-27.

Dwyer, D., J. Hu, D.R. Vaughan, and B. Wixon. 2001. Counting the disabled: Using survey self-reports to estimate medical eligibility for Social Security's disability programs. Social Security Administration, Publication number ORES WP 90. Washington, D.C.: Office of Policy, Office of Research, Evaluation, and Statistics, SSA.

Golden, T.P., S. O'Mara, C. Ferrell, and J. Sheldon. 2002. Supporting career development and employment: Benefits planning, assistance and outreach (BPA&O;) and protection and advocacy for beneficiaries of social security (PABSS). Social Security Administration, Publication Number 63-003. Washington, D.C.: Government Printing Office.

Halpern, J., and J.A. Hausman. 1986. Choice under uncertainty: A model of applications for the Social Security Disability Insurance program. Journal of Public Economics 31:131-161.

Haveman, R., and B. Wolfe.1984. The decline in male labor force participation: Comment. Journal of Political Economy 92 (June):532-541.

Haveman, R., and B. Wolfe. 1984. Disability transfers and early retirement: A causal relationship. Journal of Public Economics 24:47-66.

Haveman, R., P.R. de Jong, and B. Wolfe. 1991. Disability transfers and the work decisions of older men. Quarterly Journal of Economics 106:939-949.

Top of Page
25

Hennessey, J. 1996. Job patterns of disabled beneficiaries. Social Security Bulletin 59 (Winter):3-11.

Hennessey, J. 1997. Factors affecting the work efforts of disabled-worker beneficiaries. Social Security Bulletin 60:3-20.

Hennessey, J., and J. Dykacz. 1993. A comparison of the recovery termination rates of disabled worker beneficiaries entitled in 1972 and 1985. Social Security Bulletin 56:58-59.

Hennessey, J., and S. Muller. 1994. Work effort of disabled-worker beneficiaries: Preliminary findings from the New Beneficiary Followup Survey. Social Security Bulletin 57 (Fall):42-51.

Hennessey, J., and S. Muller. 1995. The effect of vocational rehabilitation and work incentives on helping the disabled-worker beneficiary back to work. Social Security Bulletin 58 (Spring):15-28.

Hoynes, H., and R. Moffitt. 1999. Tax rates and work incentives in the Social Security Disability Insurance program: Current law and alternative reforms. National Tax Journal 52 (December):623-654.

Hu, J., K. Lahiri, D.R. Vaughan, and B. Wixon. 2001. A structural model of Social Security's disability determination process. Review of Economics and Statistics 83 (May):348-361.

Kreider, B. 1998. Workers' applications to social insurance programs when earnings and eligibility are uncertain. Journal of Labor Economics 16:848-877.

Kreider, B. 1999. Social Security Disability Insurance: Applications, awards, and lifetime income flows. Journal of Labor Economics 17 (October):784-827.

Kreider, B., and R. Riphahn. 2000. Explaining applications to the U.S. disability system: A semiparametric approach. Journal of Human Resources 35 (Winter):82-115.

Lahiri, K., D.R. Vaughan, and B. Wixon. 1995. Modeling Social Security's sequential disability determination using matched SIPP data. Social Security Bulletin 58 (Winter):3-42.

MacDonald, B.R. Issues statement, questions and concerns on $1 for $2 demonstration projects, Section 302 in the Ticket to Work and Work Incentives Improvement Act. Unpublished draft.

Muller, S. 1992. Disability beneficiaries who work and their experience under program work incentives. Social Security Bulletin 55 (Summer):2-19.

Top of Page
26

Parsons, D. 1980. The decline in male labor force participation. Journal of Political Economy 88 (February):117-134.

Parsons, D. 1984. Disability insurance and male labor force participation: A response to Haveman and Wolfe. Journal of Political Economy 92 (June):542-549.

Rupp, K., and D. Stapleton. 1995. Determinants of the growth in the Social Security Administration's disability programs: An overview. Social Security Bulletin 58 (Winter):43-70.

Schecter, E. 1997. Work while receiving disability insurance benefits: Additional findings from the New Beneficiary Followup Survey. Social Security Bulletin 60:3-17.

Stapleton, D., B. Barnow, K. Coleman, K. Dietrich, J. Furman, and G. Lo. 1995. Labor market conditions, socioeconomic factors, and the growth of applications and awards for SSDI and SSI disability benefits: Final report. Lewin-VHI.

Ycas, M. 1996. Patterns of return to work in a cohort of disabled-worker beneficiaries. In J.L. Mashow, V. Reno, R.V. Burkhauser, and M. Berkowitz (Eds.), Disability, work and cash benefits. Kalamazoo, MI: Upjohn.

Top of Page
27

Contact Information

Ticket to Work and Work Incentives Advisory Panel

Social Security Administration
400 Virginia Avenue, SW, Suite 700
Washington, DC 20024
Phone at (202) 358-6430
Fax at (202) 358-6440
E-mail to TWWIIAPanel@ssa.gov
Website: www.socialsecurity.gov/work/panel

Advisory Panel Staff

Marie Parker Strahan, Executive Director
Tamara Bibb Allen
Shirletta Banks
Kristen M. Breland
Lisa D. Ekman
Mildred D. Owens
Gordon Richmond
Bernard Wixon

Anyone requiring materials in alternative formats, information regarding this document or the Ticket to Work and Work Incentives Advisory Panel should contact the Panel staff. Records are kept of all Panel proceedings and are available for public inspection under the Federal Advisory Committee Act by appointment at the Panel office.

Top of Page
28