A new approach to SSDI reform: more people with disabilities would return to work if they faced better incentives.

Gokhale, Jagadeesh

The Social Security Disability Insurance (SSDI) program is rapidly approaching insolvency. According to the Social Security Trustees, the program-s trust fund will be exhausted sometime in early 2016, forcing a reduction in financial support for individuals with a disability. Although many lawmakers in Congress appreciate this problem, most of them appear unwilling to propose reforms to the program. One explanation for this inaction is the availability of an easy short-term fix: temporarily transferring funds from SSDI's larger companion trust fund, the Social Security Old Age and Survivor Insurance (OASI) program. According to the Trustees, the OASI Trust Fund will not be exhausted until 2034. Were a temporary transfer made from the OASI to the SSDI trust fund, the combined (OASDI) trust funds would not be exhausted until 2033.

Another explanation for congressional inaction is that the two major political parties are so far apart on how to reform SSDI that there is little chance of developing a workable coalition. Some Republicans and conservatives may view the program as riddled with loopholes, subjective decisionmaking, and a predilection for unwarranted allowances on the part of adjudicators using flawed evaluation procedures. (See "What Should We Do about Social Security Disability?" Fall 2011; "What We Should Do about Social Security Disability," Spring 2012.) They may also believe that many undeserving claimants exist on SSDI rolls and many work-capable individuals are being allowed onto the program. Some Democrats and liberals, however, appear to believe that the program is operating as intended, the overall allowance rate is not excessive, and claim procedures take too long. They may also believe that benefit levels are too low and individuals with disabilities are ill-served by having to wait many months before becoming eligible for monetary benefits and two years before obtaining access to health care under the program's rules.

Although SSDI keeps many people out of poverty, it continues to be beset with poor performance along many dimensions. Applications and enrollments have been growing from the after-effects of the recession, population aging, and poor (ex-post) policy compliance, including inadequate Continuing Disability Reviews (CDR) of beneficiaries' medical status to determine if they should continue to receive SSDI benefits. This does not reflect poor intentions and ability on the part of SSDI adjudicators as much as the difficulty of determining whether a person is truly disabled within a system that allows only a "yes or no" distinction and one that involves subjective evaluations of mental and musculoskeletal impairments. As a result, many commentators today believe that a significant number of SSDI beneficiaries have work capabilities, but beneficiaries are unwilling to participate in the work force because they face a steep "cash cliff": the loss of benefits and valuable health care coverage if they work and earn above the Substantial Gainful Activity (SGA) level, currently $1,040 per month.

[ILLUSTRATION OMITTED]

The unwillingness of work-capable SSDI beneficiaries to join the labor force implies losses to the economy in terms of lost output, reductions in tax revenues for the government, and most significantly the loss of gainful careers for younger work-capable beneficiaries. Those lost careers mean the loss of a number of important benefits, including higher incomes and living standards, participation in community activities through work, and the psychological benefit of enhanced self-worth that accompanies gainful employment.

Encouraging the Return to Work

The Social Security Administration (SSA) is charged with discovering beneficiaries whose medical condition has improved and exploring ways to incentivize those beneficiaries to seek gainful employment. SSA's CDR program is intended to examine and detect cases of medical improvement among SSDI beneficiaries and remove those who have recovered sufficiently so that they no longer meet SSDI's definition of disability. SSA has, however, accumulated a substantial backlog of CDR cases as other workloads increased when the recent recession began.

The SSA provides the Ticket to Work program to new beneficiaries, offering counseling and vocational rehabilitation services to those who may wish to pursue job opportunities despite their disabilities. As was expected, the take-up rate on Ticket to Work has been very low. However, a sizable number of SSDI beneficiaries appear to have work abilities and could re-engage in the work force if not for the disincentive created by the cash cliff. Most work-capable beneficiaries are unwilling to jeopardize their eligibility for SSDI's monetary and health benefits by working beyond the SGA earnings level as insurance against a relapse of their disabling medical conditions or loss of labor-market opportunities.

To examine the likelihood of re-engaging SSDI beneficiaries in the work force through a better incentive system, the SSA has been conducting demonstration projects to assess the work capabilities of SSDI beneficiaries. These demonstrations replace the cash cliff with a more gradual reduction in benefits (a "benefit offset") as SSDI beneficiaries' earnings increase. The projects assess increases in employment rates, average earnings, and the fraction of beneficiaries earning above the SGA level.

Unfortunately, the results from these assessments are not as conclusive as was initially expected. One reason for this may be that the benefit offset design does not provide a sufficiently robust work incentive to SSDI beneficiaries. Moreover, as will be explained below, eliminating the cash cliff makes existing SSDI benefits inframarginal to beneficiaries' decisions about whether and how much to work and earn, which may increase their reservation wages for entering the job market.

This article proposes a different approach to providing work incentives to SSDI beneficiaries, one that involves benefit offsets but provides greater flexibility to beneficiaries in selecting their level of work activity. The Generalized Benefit Offset (GBO) program that I propose would eliminate the cash cliff, but it conditions the change on observed earnings. GBO allows reversion to beneficiary status when labor force attachments cease, at the full discretion of SSDI beneficiaries.

The sections that follow describe the results from the SSA's benefit offset demonstrations, provide analytical justification for GBO's new approach, and discuss how to operationalize GBO by introducing its benefit schedule several months after a claimant has been allowed onto SSDI's rolls.

SSDI Pilots and Demonstration Projects

Two investigations conducted by the SSA explore the effects of offering a "benefit offset" to SSDI beneficiaries: the Benefit Offset Pilot Demonstration (BOPD) and its broader follow-up, the Benefit Offset National Demonstration (BOND).

BOPD studied the effect of replacing the steep loss of benefits consequent to working above SGA with a"$1-for-$2" offset. That is, benefits would be reduced by $1 for every $2 of earnings above SGA after a nine-month trial work period. The gradual reduction of benefits eliminated the cash cliff and BOPD's designers believed it would incentivize more SSDI beneficiaries to enter the work force at higher-than-SGA earnings levels.

The SSA staff certainly expected to see strong effects from providing the $1-for-$2 benefit offset rate. Unfortunately, the results from BOPD projects that were implemented in four states were not stellar and were more variable than seems reasonable.

Responses to the $1-for-$2 benefit offset from the treatment group compared to the control group were very weak in Wisconsin's BOPD: average earnings differences (treatment minus control) were mostly negative, overall employment rate differences were close to zero, and the rate of above-SGA earnings was higher for the treatment group by just 2 percentage points (17 versus 15 percent). In Vermont's BOPD, the $1,256 per-month average earnings for the treatment group were marginally higher (by $51); the rate of above-SGA earnings was also only marginally higher (18 versus 15 percent), as was the overall employment rate (49 versus 47 percent). The Utah BOPD yielded larger outcome differences between the two groups: by the end of the trial, treatment group mean earnings were higher by more than $245 per month; the rate of above-SGA earnings was higher by 5 percentage points (19.5 versus 14.5 percent); and the overall employment rate was also higher by 4 percentage points (47 versus 43 percent). The strongest results, by far, were obtained from the Connecticut BOPD, with the rate of above-SGA earnings for the treatment group being 12 percentage points above that of the control group (29 versus 17 percent), on average. At $2,115, mean earnings were $410 higher for the treatment group, but the difference in employment rates was just 3.5 percentage points (64 percent versus 61 percent).

The four-state BOPD results exhibit high variability and therefore do not seem as reliable as policymakers are likely to demand. The general conclusion from the BOPD study was that a benefit offset can have a significant, large, and enduring effect on the SGA rate of certain beneficiaries, but the effect may be limited to a subset of individuals who are more able and/or more motivated to work than the average SSDI beneficiary. The pilot also had drawbacks: Those SSDI beneficiaries who were more than 72 months beyond the end of their trial work period--a grace period during which the SSA does not eliminate benefits despite working and earning more than the SGA level--were excluded from enrollment. This may have eliminated the most persistent earners among SSDI beneficiaries. However, beneficiaries knew that BOPD was time-limited and may, therefore, have been unwilling to commit to higher-paying career paths. Finally, BOPD's implementation of the offset treatment was hampered by technical issues and beneficiaries may have avoided earning more because of fears of repeated errors that would endanger critical state and federal benefits on which they depended.

The four-state BOPD trials were marginally successful and suggested the need for a broader demonstration project. Unfortunately, the SSA's follow-up BOND investigation has been a failure. Out of more than 77,000 treatment group participants, the BOND take-up was a miserable 21 cases. This project is known to be riddled with methodological and technical defects that compromise the quality of its treatment and control beneficiary samples, ultimately generating poor results. Several groups of disability advocates, policy practitioners, and even nonpartisan groups, including the Social Security Advisory Board, are now advocating that BOND be terminated.

A Better Approach

The relatively weak boost to employment and earnings generated by the $1-for-S2 offset policy is not surprising. That benefit offset design actually provides a relatively strong work disincentive at the margin. The attempt is to trade-off elimination of a huge marginal disincentive to work--the cash cliff prospect of losing all SSDI benefits for exceeding SGA earnings by even $1--against the prospect of losing only $1 for every $2 earned above SGA. The problem is that one marginal disincentive to work, the cash cliff, is replaced by another, the $1-for-S2 benefit offset that represents a high (S0 percent) marginal tax rate on beneficiaries' earnings.

The BOPD's weak results should have indicated that, at the margin, a 50 percent marginal tax rate on earnings also generates a sizable disincentive to work, especially for those with a disability. This switch from the cash cliff to the $1-for-$2 deal makes existing SSDI benefits (including health care coverage) inframarginal, thereby worsening the rate at which beneficiaries would be willing to exchange more work for more dollars. Even at low resource levels, those with disabilities are likely to require substantial compensation at the margin to induce labor force attachments.

[FIGURE 1 OMITTED]

Applying standard consumer-preference analysis, Figure 1 explains this reasoning in greater detail: The x-axis in Figure 1 represents the daily time endowment, T, of 24 hours. Points further to the left on the x-axis represent greater time devoted to non-work activities, called "leisure" for short (though it includes "personal time" that is relatively inflexibly devoted to sleep and personal care). The vertical line marked A0 reflects the assumption that even those without disabilities require leisure time of at least 10 hours per day, cutting the daily time endowment that is potentially available for market work to 14 hours.

In Figure 1, the disability of an SSDI beneficiary with residual work capacity is characterized as an expansion of the inflexible "personal time" portion of the day and a corresponding shrinkage of the hours potentially available for market work. This reflects the additional, inflexible "personal time" required to take care of disabling conditions. Two alternative cases of this type are depicted as the vertical lines marked A1 and A2. Compared to A1, a person at A2 has a disability that leaves less time for market work.

Under current SSDI rules, a beneficiary who is receiving benefits of $B per day can work and earn up to the SGA level of earnings and continue to receive $B from the program. However, if earnings in excess of the SGA level are sustained beyond the trial work period (which lasts for nine months after the first excess-earnings occurrence), SSDI benefits are suspended and eventually terminated. This policy implies that SSDI beneficiaries eventually face a cash cliff at the SGA level of earnings. Figure 1 shows this at point P with daily income at $B + $SGA: At the market wage rate available (shown by the slope of the budget line WW'), working for S hours yields the SGA level of earnings. Increasing earning beyond SGA by even $1 more by working for more than S hours (moving leftward on the x-axis from point T - S) pushes the beneficiary to point CC (the cash cliff point on the budget line TT', which is parallel to WW' but starts at $0 instead of $B) as the daily SSDI benefit of $B is eliminated. The cash cliff, therefore, induces many work-capable SSDI beneficiaries to "park" at point P--that is, they work and earn no more than the SGA level. Were the cash cliff absent--that is, were they unconditionally eligible to work as much as they could with no loss of eligibility to SSDI benefits, those who could (those able to work more, shown by the vertical line A1) would choose work at point [E.sub.0], with a much higher welfare level shown by the welfare curve [IC.sup.3] . Note that the welfare level [IC.sup.3] is higher than that passing though point P ([IC.sup.1]).

Providing a benefit offset for earnings beyond the SGA level implies offering the tradeoff represented by the budget line OO' (to the left of point P). Additional work and (gross) earnings along the budget line WW' are offset by reducing benefits by $1 for each $2 of additional earnings. Because OO' incorporates a marginal tax rate of 50 percent, its smaller gradient implies tangency with the beneficiary's indifference curve IC2 at a point ([E.sub.1]) involving less work effort than at [E.sub.0]. Thus, a benefit offset of the type incorporated in the BOPD and BOND demonstrations would not be expected to produce strong labor market responses by beneficiaries. First, to the extent that beneficiaries generally require larger compensation increases to elicit a given amount of work effort (their indifference curves are likely to be steeper), the offer of a benefit offset of OO' would cause their optimal choice point ([E.sub.1]) to be closer to P. Second, the fact that the existing SSDI benefit ($B) is rendered inframarginal under BOPD and BOND also reduces their most preferred work choices in response to the benefit offset (points such as [E.sub.1]) closer to P.

There is unlikely to be political support to provide unconditional benefits to those on SSDI's rolls (to allow access to [E.sub.0]). After all, SSDI is an insurance program to protect people with a sufficient work history against the loss of earnings (and health coverage) from developing a work-disabling condition. Providing unconditional benefits to those who were found eligible for SSDI would be in direct conflict with SSDI's eligibility requirements. And it would be a policy that could cause a surge of SSDI applications even by work-capable individuals with physical or mental impairments that may not meet SSDI's definition of disability. At best, that policy would cause SSA's processing costs to spiral from such induced entry into the SSDI program. At worst (if many of those induced to apply are also allowed onto SSDI's rolls), it would increase taxpayer burdens from providing unconditional benefits to a growing number of beneficiaries who are work-capable or who eventually recover close to full work capacity.

However, the cash cliff that the current policy incorporates discourages partially work-capable beneficiaries from participating in the work force, a situation that Congress has sought to ameliorate. The Ticket to Work program and benefit offset demonstration projects were designed to explore ways of incentivizing work-capable SSDI beneficiaries to return to work. As already noted, those efforts have fallen short of expectations, producing only marginal increases in the labor force attachments of SSDI beneficiaries beyond the SGA level.

Proposal: The Generalized Benefit Offset

The way to improve upon the benefit offsets offered to SSDI beneficiaries becomes clear once it is recognized that the $1-for-$2 benefit offset, albeit a better option compared to the current cash cliff, is itself a relatively poor work incentive because it involves a marginal tax on earnings of 50 percent. The resulting increase in work effort beyond SGA for someone with a minor disability (shown by vertical line A1 in Figure 1) is much smaller (point [E.sub.1]) than the potential increase in work effort (point [E.sub.0]). This result arises because under a simple benefit offset, current SSDI benefits become inframarginal, which increases the marginal compensation (at points northward along any vertical line from the x-axis) that the beneficiary requires to provide additional work effort. The key is to trade off inframarginal SSDI benefits for a more generous benefit offset. This is shown in Figure 2 by the budget line DD'.

Under the GBO offer shown by budget line DD' in Figure 2, SSDI benefits are reduced (point D is below point W = $B) but the benefit offset is a more generous $1-for-$4 instead of $1-for-$2. By reducing SSDI benefits by a small amount (W- D), the work-capable beneficiary would work as much as he wants above SGA along DD', the most preferred choice being the tangency point of the indifference curve [IC.sub.2] with the budget line DD' at [E.sub.2]. Figure 2 shows that this choice involves more work than under the simple $1-for-$2 benefit offset at SGA (point [E.sub.1]). This is the result of removing the "wealth effect" (by reducing inframarginal benefits) from the total "price effect" of an improved benefit offset rate. The remaining "substitution effect" (movement from [E.sub.1] to [E.sub.2]) would make the beneficiary willing to be more active on the labor market.

Under GBO, the beneficiary is offered a schedule of choices: either remain disengaged from the work force with a benefit of $B or work and earn along the budget line DD' involving a small reduction in initial SSDI benefits and a $1-for-$4 benefit offset. This more marginally generous benefit offset will induce those with the A1 work time availability to work at point [E.sub.2]. Note that at this point, SSDI beneficiaries would receive the same welfare level, [IC.sup.2], as those disabled who were offered the $1-for-$2 benefit offset beginning at the SGA benefit level (at point [E.sub.1] along budget line PO'). Those with work time availability at, A2 would choose a feasible point that is lower along the DD' budget line (at the intersection of A2 and the budget line DD' in Figure 2). The GBO schedule of initial benefits and benefit offsets offered to SSDI beneficiaries would incentivize beneficiaries to sort along the DD' budget line, providing the amount of market labor supply consistent with maximizing their welfare level.

Indeed, the GBO schedule presented thus far could be improved by adding more elements as shown in Figure 3. For example, adding the offer shown by the budget line FF' would involve a steeper reduction in initial benefits but an even more generous benefit offset. Indeed, the marginal benefit offset may even be negative in that working and earning fetches a subsidy much like the federal Earned Income Credit (EIC) program subsidizes earnings at very low levels. Along FF', SSDI beneficiaries' initial benefit would be reduced by about 50 percent, but they would receive a subsidy of $1 for each $6 earned. On this schedule, beneficiaries for whom it is feasible would choose to work and earn at point [E.sub.3].

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

Finally, those beneficiaries who have very limited work potential may benefit from an offer of GG'. This offer involves an increase in the initial benefit but a much less generous benefit offset. Under GG', for example, the initial SSDI benefit is increased by 50 percent, but earnings are subject to an 80 percent ($4-for-$5) marginal tax rate. Those with limited work potential could work and earn at point [E.sub.4], involving only a slight increase in earnings above SGA. But their living conditions would be better from the inframarginal increase in SSDI benefits.

Described thus far, GBO still has a problem of inducing a separation of beneficiaries according to their work abilities. In Figure 3, the different GBO offers are intended to induce SSDI beneficiaries to separate along different work and earnings levels shown as points [E.sub.1] through [E.sub.4]. However, in the hypothetical example of Figure 3, all of those points yield the same welfare level of [IC.sup.2]. Thus there is nothing to prevent a beneficiary with high work capability to select point [E.sub.4] (with high initial benefits but with a higher marginal tax rate) instead of point [E.sub.3].

This is the so-called information revelation problem of providing a set of incentive-compatible offers that would induce beneficiaries to sort across the available schedule of GBO offers according to their work abilities and preferences. Incorporating incentive compatibility across the offer schedule requires the payment of a premium to the most work-capable to ensure that their work choices remain effectively distinct from the choices of those who are less work-capable. This calls for a set of GBO offers shown in Figure 4.

In Figure 4, the four budget lines [S.sub.1], [S.sub.2], [S.sub.3], and [S.sub.4] indicate elements of a GBO schedule, each with a different combination of an initial benefit reductions and benefit offset (or earnings subsidy) intensity. Each successive element involves a larger reduction in initial benefits combined with a more generous benefit offset. The tradeoffs are designed to provide those with larger work-abilities or capacities to achieve a higher welfare level than those less capable to work. Thus, beneficiaries who are more work-capable and could work and earn at point [E.sub.1] would not choose to be at any of the other three points, [E.sub.2], [E.sub.3], or [E.sub.4].

[FIGURE 4 OMITTED]

Of course, one could argue that such a schedule of offers could be designed only if the underlying structure of work abilities and preferences of SSDI beneficiaries--how much more they would be able and willing to work for given additional compensation at different initial benefit levels--were known with certainty. In reality, different beneficiaries will differ in terms of their work potentials (the amount of time that they could potentially devote to labor market activity) and in their willingness to trade work time for dollars (the shape of their indifference mappings between time use and compensation).

Lacking definitive distributional information about the severity of disabilities and beneficiaries' preferences structures, how could one design a GBO-style post-entitlement SSDI reform? The answer is simply to mimic private sector providers in residential and commercial property, vehicular, health, and other insurance sectors. In these industries, insurance contracts provide property owners with a range of pre-determined combinations of deductibles and premium rates. The choices are designed so that clients would sort themselves according to their types: risky, moderately risky, or safe (alternatively, in the health insurance sector, those with chronic conditions, moderately healthy, and healthy). The latter types would select high-deductible policies with low premium rates and the former types would select low-deductible policies with high premiums, provided that the schedule of choices contains sufficient diversity to induce such separation among clients. That is, the contract schedule's high deductible policies have sufficiently lower premium rates to attract the low risk (or healthy) types but not the riskier (less healthy) clients. In the case of SSDI, designing a schedule of GBO offers should also be feasible to provide better living conditions to those with almost no work ability and stronger work incentives to those capable of devoting more time to market work. Based on these considerations, the U.S. Congress should alter SSDI's benefit schedule to incorporate such potentially incentive-compatible offers for both current beneficiaries and new entrants.

Operationalizing GBO

The GBO schedule should be designed to be sufficiently generous in the benefit offset (or wage subsidy) relative to the initial SSDI benefit (increase or reduction) offered under each GBO schedule element. Such a schedule would generate a distribution of responses from beneficiaries along a range of work/leisure and earnings choices. That distribution would automatically reveal the extent of work capabilities among SSDI beneficiaries, producing clear benefits to SSDI beneficiaries as well as to the economy.

An important issue concerns how to design and operationalize a GBO-type benefit schedule for the SSDI beneficiaries and incorporate it within the current SSDI system. Under current law, an application for SSDI benefits generates a "protective filing date" for each beneficiary--the date when the claimant reports the intent to apply for benefits and reports the "alleged" date of the disability's onset. Once the intention is established, current law states that a disability award is contingent upon SSA determination that the claimant's disability is sufficiently severe that it will result in death or preclude work above SGA for a continuous period of 12 months. Obviously, the applicant must immediately cease working (above SGA level) to qualify. The definition of disability and its application in determining eligibility, therefore, precludes the adoption of GBO during the process of determining eligibility to SSDI.

After eligibility has been established, benefits are paid for a maximum of 12 months retroactively from the "established" disability onset date. Indeed, retroactive benefits are also awarded for the period between that date and the date of final allowance. Under GBO, however, current-law benefits could be continued for N months beyond the date of final allowance--as few as is administratively feasible and the fewer the better. Unlike the current SGA limit on working and earning, current-law post-entitlement benefits would be payable under the new system with no limitation on beneficiaries' gainful employment activity. However, N months after the date of SSDI allowance, the benefit would be determined by comparing earnings during the previous N months against a schedule. The schedule could specify higher-than-current-law benefits at zero earnings and a graduated reduction in those benefits with higher earnings during the past N months. Under the pre-determined GBO schedule, the change in SSDI benefits would result in higher total income (benefits plus earnings) for those with higher earnings.

One consideration in setting the GBO schedule is that beneficiary earnings statements (from state unemployment insurance agencies) would only report total earnings and not hours and wage rates. In general, however, the GBO's objective is to incentivize high total earnings regardless of whether they result from high work hours by low-wage beneficiaries or high wage rates paid to SSDI beneficiaries with high skills despite working only a few hours. Thus, a schedule such as the one shown in Table 1 (or a variation thereof based on additional prior information about the distribution of work abilities among SSDI beneficiaries) could deliver better work incentives.

One noteworthy omission from Table 1 is the consideration of inframarginal health care coverage for SSDI beneficiaries. Such coverage is mainly through Medicare, although some beneficiaries are dually eligible for Medicaid as well. Under the new Obamacare law, and consistent with implementing a GBO schedule as outlined in this article, health care subsidies would automatically be reduced as beneficiary incomes increase for those younger than age 65. Consideration of whether and how Obamacare subsidies should be modified for such SSDI beneficiaries--under a new "Medicaid buy-in" allowance--is not discussed herein.

In Table 1, the first two columns show alternative ranges of a beneficiary's reported average monthly wages over the previous N months. The third column shows the worker's base SSDI benefit ors 1,000 in 2013 dollars. The base SSDI benefit shown in Table 1 is simply an example. This benefit will be different across beneficiaries depending on the length of their work history and average annual earnings. In general the schedule described in this section could be made more (or less) progressive for those with smaller base SSDI benefits.

For the case shown in Table 1, applying the earnings-contingent rates in column 4 determines the worker's SSDI benefit during the current month. This amount would decline as the worker's average earnings during the previous N months increases. But because the initial benefit reduction rate is negative at low earning rates, low-earner beneficiaries receive more ($1,100) than their base benefit level ($1,000).

Applying the earnings-contingent benefit offset rate (column 6) yields earnings net of the benefit offset as shown in column 7 (for earnings assumed to be those in column 2). The offset amount declines and earnings net of the offset increase with higher earnings. The beneficiary's total income, which equals earnings net of benefit offset plus earnings-contingent initial SSDI benefit, also increases with earnings (column 8). Because earnings net of offset increase faster than the decline in the initial benefit, total income increases at a faster rate at higher earnings levels--up to a point. This is shown in Figure 5, where total income increases at an increasing rate with the beneficiary's market earnings. Note that this path is qualitatively similar to that traced by incentive-compatible decision points [E.sub.1] through [E.sub.4] in Figure 4.

Column 9 of Table 1 shows the marginal tax rate implied by the GBO schedule computed as the change in income per dollar of earnings. In the example of Table 1, the marginal tax rate is 100 percent at the lowest earnings level. As earnings increase, the marginal tax rate declines and eventually becomes negative at earnings between $1,000 and $2,000 per month. This provides and Earned Income Credit-type tax incentive to induce more work by work-capable beneficiaries. Of course, at yet higher earnings, this work incentive may have to be withdrawn--as shown in the last four rows of Table 1 where the marginal tax rate (column 9) is positive. Table 1 also shows that a beneficiary who earns $2,750 or more per month would be, economically speaking, "off SSDI rolls." Ira relapse of an SSDI-eligible individual's disabling condition prevents high earnings during certain periods, the person could reduce labor force activity or exit the work force. In that case, SSDI benefits would automatically resume as average monthly earnings (potentially computed over rolling periods of the previous N months) decline over time.

[FIGURE 5 OMITTED]

Lawmakers could tailor the GBO schedule according to the strength of the work incentive that is deemed appropriate at the target earning range. In general, however, a steeper increase in net-of-GBO income at higher earnings levels would make the "contract curve" incentive-compatible and more likely to induce SSDI beneficiaries to sort according to their work abilities rather than park at the SGA earnings level.

Conclusion

Congress has consistently sought to incentivize SSDI beneficiaries to return to work. Most efforts to date, however, have not produced significant and reliable responses by SSDI beneficiaries. Many believe that despite significant work capabilities, SSDI beneficiaries are reluctant to join the work force for fear of losing eligibility for SSDI and health benefits.

A recent pilot project and a national demonstration project to assess the effect of eliminating the cash cliff facing beneficiaries under current SSDI rules with a $1-for-$2 benefit offset have yielded clues about the existence of work abilities among SSDI beneficiaries. Unfortunately, both projects suffered from several technical and implementation problems and the information generated to date exhibits considerable variability.

One key shortcoming of these efforts was the poor design of the benefit offset. Providing a $1-for-$2 benefit offset option with full protection of existing SSDI benefits including health care coverage eliminates the cash cliff but makes those benefits inframarginal to beneficiaries' work decisions. Moreover, the 50 percent marginal tax on benefits also provides a work disincentive at the margin. In contrast, the GBO approach as described in this article would trade off reductions in initial benefits against sufficiently generous benefit offsets including federal EIC-like wage subsidies, if necessary. Under GBO, one would expect beneficiaries to sort themselves according to their work abilities along the GBO schedule rather than park at the SGA earnings level. The distribution of resulting work and earning choices by SSDI beneficiaries would generate direct benefits to the economy and to SSDI beneficiaries themselves. GBO should receive broad support from disability advocates, policy practitioners, and (most importantly) lawmakers from both sides of the aisle because it combines key elements that their constituents are demanding: better support for individuals with disabilities but also opportunities to work whenever their health impairments permit market participation. GBO eliminates the cash cliff and would, if adopted, introduce stronger and more effective work incentives for work-capable beneficiaries while retaining SSDI insurance for individuals with disabilities.

JAGADEESH GOKHALE is a senior fellow with the Cato Institute and a member of the Social Security Advisory Board.

The author thanks, without implicating, the Social Security Advisory Board staff for their helpful comments.

TABLE 1
One Potential Formula for the GBO Schedule

Average Monthly Earnings over Base SSDI Initial Benefit
previous N=12 months is: Monthly Adjustment
 Benefit * Rate
 Greater But less than
 than or equal to

 Col. 1 Col. 2 Col. 3 Col. 4

 0 250 1000 -0.1000
 250 500 1000 0.0000
 500 750 1000 0.1000
 750 1000 1000 0.2000
 1000 1250 1000 0.3000
 1250 1500 1000 0.4000
 1500 1750 1000 0.5000
 1750 2000 1000 0.6000
 2000 2250 1000 0.7000
 2250 2500 1000 0.8000
 2500 2750 1000 0.9000
 2750 3000 1000 1.0000

Initial SSDI Benefit Earnings Net
 Benefit Offset of Benefit
 [col. 3 x Rate Offset
(1-col. 4)] [col. 2 x
 (1 col.6)] **

 Col. 5 Col. 6 Col. 7

 1100 1.000 0
 1000 0.800 100
 900 0.600 300
 800 0.400 600
 700 0.200 1000
 600 0.000 1500
 500 -0.200 2100
 400 -0.400 2800
 300 -0.200 2700
 200 0.000 2500
 100 0.000 2750
 0 0.000 3000

 Total Marginal Tax
 Income Rate on
 Earnings

 Col. 8 Col. 9

 1100 NA
 1100 100
 1200 60
 1400 20
 1700 -20
 2100 -60
 2600 -100
 3200 -140
 3000 180
 2700 220
 2850 40
 3000 40

* The Base Monthly SSDI Benefit is based on the beneficiary's earnings
history and projected earnings through retirement. ** Earnings assumed
to be those shown in column 2.

Source: Example based on the author's calculations.