| January 2004
Volume 5, Issue 1
Earlier, MRRC researchers John Bound and Julie Cullen provided evidence that the SSDI program seems to offer a fairly satisfactory level of coverage (www.mrrc.isr.umich.edu/research/publications/Brief/ib_024.pdf). Recent work by Olivia Mitchell and John Phillips featured in this newsletter, however, examines the lengthy and somewhat arduous eligibility determination process for disability claims and suggests that in some cases, people in greatest need may not be eligible for disability benefits. Other work by MRRC researchers Elizabeth Powers and David Neumark featured in this issue studies the interdependence of Social Security programs, suggesting that changes in one program may have implications for others.
MRRC will continue to support research addressing important policy issues
related to Disability Insurance, and program interactions, simulating
potential program changes, and studying possible behavioral implications
of modifications to retirement programs. Findings from this kind of research
will have important implications for solvency of the system.
Eligibility, Applications, Denials, and Appeals for Social Security Disability Insurance by Olivia S. Mitchell and John W. R. Phillips
Researchers and policymakers have been interested in the extent to which older workers who are in poor health may use Social Security Disability Insurance (DI) as a path out of the labor force. Another path such workers might take is early Social Security retirement. This has raised the question of what effect raising the age of early retirement might have on applications for DI. It is possible that older workers in poor health might apply for DI rather than waiting for delayed Social Security benefits. Identifying the characteristics of workers who might seek to use Social Security disability insurance can provide policymakers with insight regarding the well-being of the “at risk” population.
This Issue in Brief describes an empirical investigation using data from the University of Michigan Health and Retirement Study (HRS) linked with administrative earnings records for a cohort of individuals nearing retirement. With these linked data, we examined the sociodemographic and health characteristics of workers along the continuum of eligibility, application, denial, and appeal for disability benefits. Regarding eligibility, we find that, while the safety net program covers most American workers, a substantial group of older women remains without coverage. We find that older people in poor health and with low economic status are more likely to apply for DI, as compared to those reporting no health problems and with more assets. Few factors distinguish between applicants awarded versus denied benefits and between those who appeal rejected applications versus those who do not.
American workers earn “Quarters of Coverage” (QCs) based on annual Social Security payroll tax contributions. Eligibility for Social Security retirement benefits requires that a worker must have earned 40 QCs over her career, a total that can be earned in as few as ten years. Once a worker acquires 40 QCs they are “Forever Fully Insured (FFI). Eligibility requirements for disability insurance are a bit more complex. In what we call Test 1, or the recency condition, if the worker is over 30 she must have at least 20 QCs in the last 10 years; if she is younger than 30, her QCs must total at least her current age minus 22 and times 2. In Test 2, the “fully insured” test, her total QCs at any age must be at least equal to her current age minus 22. For test 3, she must have at least 6 QCs by a given age. A given worker could be both forever fully insured and disability insured, either, or neither. As a result of these program rules, workers may move into and out of DI eligibility over their lifetimes, depending on their work patterns.
Medical Eligibility:The Disability Determination Process
The Social Security definition of disability can be characterized by three factors:
Whether a worker is determined to be disabled, under the Social Security definition, is established by the Disability Determination Services (DDS). A five-step process determines benefit eligibility using the criteria listed above. Overall, an application is rejected if 1) the applicant earns $700 or more per month, 2) the condition is determined not severe enough to limit the applicant’s capacity to perform work, 3) the condition is not on the SSA list of disabling conditions or the condition is not judged to last at least 12 months, 4) the condition does not prevent the applicant from performing his old job, and 5) the applicant could participate in other occupations with this condition given his/her education, age, and work experience
If an applicant is denied benefits, he may appeal the DDS decision. This appeal process has several stages, like the disability determination process. Each stage of appeal must be performed within 60 days of judgment. The rejected applicant may request reconsideration by the DDS. The DDS reconsiders the application. If the appellant’s claim is again rejected by the DDS, he may request a hearing before an administrative law judge. Should the ALJ deny the appeal, an appellant can request review by an Appeals Council. Both the Appeals Council and the ALJ operate within the Social Security Administrations Office of Hearings and Appeals. Should both the ALJ and the Appeals Council deny the claim, the last avenue for appeal is in U.S. Circuit Court of Appeals.
Summary of Major Findings
Who is Eligible for DI and Who is Not?
Patterns for Men
Patterns for Women
Reasons for Ineligibility
Which Factors are Most Important for Coverage?
Who Applies for DI? Over an 8-year period of observation 8 percent of respondents applied for DI benefits. Of these, about half were initially awarded benefits. Of those who applied and were rejected, about half appealed. Of those who appealed, 44 percent were subsequently awarded benefits. Thus 8 percent of respondents applied, and ultimately 60% of these applicants were awarded benefits.
Accounting for relevant baseline characteristics such as age and occupation, we find that older workers who are poorer, in worse health, and have less education are the most likely to apply for DI benefits.
Who is Awarded or Denied Benefits? Although we expected that the same workers who were more likely to apply (poorer and in worse health) would also be more likely to be awarded, we did not find this to be the case. In fact, there were few statistical differences between those awarded and denied benefits. Workers with a middle level of lifetime Social Security earnings were more likely to be awarded than those with high lifetime earnings.
Who Appeals and Who is Awarded on Appeal? We expected that those with more to gain by appealing (with lower income and wealth) would be more likely to appeal. However, we found little to distinguish those who appealed from those who did not. Higher non-Social Security wealth was associated with award upon appeal and more black applicants appear to be awarded through appeal than through initial award.
A significant number of Americans-- about 20 percent of women in later life--are uninsured for disability under Social Security. Family structure seems to play an important role in the insured status of women but has little effect for men. It appears that marriage and childbearing, in this cohort of workers at least, make women more susceptible to losing DI insurance coverage than men. Wealth, income, and good health improve the likelihood of being DI insured for both men and women.
Of particular concern is our finding that those who are in poor health and are not DI insured have the lowest levels of financial well-being. In general we find strong evidence that workers who were initially poor, in poor health, and had fewer years of education were more likely to apply for DI benefits, but few factors differentiate individuals beyond the application phase. Even collapsing initial and secondary DI awards into a single award category yielded similar results. One caution should be noted. Although the initial sample size was quite large, recall that only 8 percent applied for DI benefits. This reduced the sample to such a degree that our power to explain differences between those who were awarded and denied and then between those who did and did not appeal may be limited.
Some Social Security reform proposals posit that raising the early retirement
age under Social Security could improve the system’s solvency. However
this would also alter the opportunity set available to older workers.
If some early retirees respond to the policy by applying for DI benefits,
Trust Fund savings are mitigated. Our research identifies the characteristics
of older workers who apply for DI under current rules, those who are rejected
after application, and those who then go on to appeal, which provides
policymakers with insight regarding the potential well-being of the "at
risk" population if the early retirement age were to be raised. Our future
work will incorporate results from this study into more elaborate models
of Social Security benefit take-up patterns.
The Supplemental Security Income Program and Incentives to Claim Social Security Retirement Early by Elizabeth T. Powers and David Neumark
The Supplemental Security Income (SSI) provides a uniform federal safety net for the elderly and disabled. SSI benefits combined with supplemental state benefits potentially comprise a substantial source of income for the elderly poor. Because many elderly SSI recipients also receive Social Security, characteristics of the SSI program may influence use of social security, and vice versa. One interesting question is whether the design of SSI encourages "early retirement" in the Social Security system. In this Issue in Brief, we provide results of an investigation of this potential SSI-Old Age Insurance (OAI) program interaction. Because SSI is a welfare program, this topic is of general interest for understanding the retirement process of very low-income people and the potential supports that enable their retirement. It may also be helpful in understanding how ongoing and planned changes in the social security system affect the lowest-income beneficiaries.
To study this question, we use multiple panels of the Survey of Income and Program Participation (SIPP) that are linked to confidential Social Security Administration (SSA) data. We develop three types of evidence on the issue of whether SSI encourages earlier OAI claiming. First, we examine how the structure of SSI affects the financial return to delaying retirement from age 62 to the full retirement age (65 in our sample era). Second, we look for patterns in the timing of SSI and OAI first claims that suggest complementary program use. Finally, we estimate the probability of an early OAI claim as a function of SSI status and other household characteristics. We find that SSI recipients face quantitatively important financial incentives to claim OAI early; that SSI applications appear to vary with eligibility for OAI at age 62; and that SSI-aged program applicants are substantially more likely to make an early OAI claim, even holding family structure, health status, education, and expected Social Security replacement rates constant. While we stop short of claiming to have demonstrated a "causal" effect, due to data limitations, a credible case is made that program interactions plausibly affect retirement in unintended ways.
Supplemental Security Income
Sufficiently low income and asset units (individuals or couples) may receive SSI beginning at age 65. Benefits are equal to a maximum benefit amount minus one-half of monthly earned income over $65, all of unearned income over $20, and all of means-tested transfer income. For example, for a unit with monthly income of $165 from earnings, $50 from social security, and $100 from Veterans' Benefits, a total of $85 per month will be disregarded, and the unit will keep an additional $50 of earned income. Therefore, the benefit is determined by subtracting $135 from the maximum monthly SSI benefit.
Although people become eligible for SSI at age 65, the program can affect work effort and OAI claims earlier in life. In general, SSI likely decreases labor force participation--as we have shown in past work--because of very high implicit taxes on pension income (i.e., 100% above the $20 disregard). The asset test in SSI also discourages the additional work one would need to generate retirement savings. A separate issue is that the standard disincentive for individuals to make early OAI claims --the actuarial reduction in OAI benefits--is nullified at age 65 by the SSI program. In effect, SSI rules turn a permanent "penalty" for early retirement into a temporary state of affairs, lasting only until age 65. All of these features tilt the balance towards earlier OAI claiming. The empirical portion of our paper attempts to characterize just how much.
To test our theoretical prediction, we used interview data from the Survey of Income and Program Participation (SIPP) for 1984, 1990, 1991, 1992, 1993, and 1996. SIPP data come from large, nationally representative groups of households. We restrict attention to male-headed units (see our paper for a discussion of the special issues with women). We also use SSA records matched to SIPP data on SSI applications and payments, OAI eligibility status, OAI benefits, and social security earnings histories. Given this information, we can determine household characteristics, the incidence and timing of application to SSI, whether these applications were successful, and the timing of OAI receipt.
Descriptive Information on SSI and OAI Claimants
We first establish that OAI is a relevant concern for many SSI units. In fact, nearly 90% of our sample men who have an SSI application on record have some Social Security covered earnings since 1951. Of the group with any covered earnings, 61% are computed to be eligible to claim OAI at age 62 (i.e. have amassed 40 or more covered quarters by age 62). Therefore, it is plausible that the use of OAI and SSI may be interrelated for a significant share of aged men in SSI.
Another plausibility issue is whether our sample men take up SSI near to age 65. Our theory suggests that they should; once early OAI is taken, to delay SSI receipt is simply to stretch out the "punishment phase" (i.e., suffering the actuarial reduction) for early OAI. In general, we find that SSI payments are strongly age-dependent, with a spike at age 65. The number of people receiving their first payment at age 65 is five times larger than the number receiving a first payment at age 62, and three times larger than the number at age 66. Looking more specifically at the relationship between early OAI and SSI claiming, we show that the age-pattern of SSI receipt for men who are eligible for OAI at age 62 is more compressed around age 65 than for men who are not eligible for early OAI.
We provide two pieces of specific evidence on the quantitative importance of the incentives to take early OAI under SSI. First, we estimate the financial loss from delaying the OAI claim from age 62 to age 65, given use of the SSI program at 65. We find that SSI recipients on average face a very large loss of 7% per annum in old age transfer wealth because of the design of the SSI system.
The second thing we do is to estimate the probability of an early OAI claim as a function of SSI-aged applicant status, controlling for other important factors (age-in-sample, birth cohort, marital status, education, race, the state unemployment rate, spouse age and spouse covered quarters, and the age-62 Social Security benefit replacement rate). The major finding is that among the group of men that is eligible for OAI at age 62, the individuals in aged units applying to SSI are 24%-points more likely to have entered OAI early. While we find this figure plausible given the very large financial penalties to delaying retirement under SSI, we cannot rule out a competing hypothesis that early contact with SSA through OAI is an important means for learning about the SSI program with the available data.
This paper builds on our previous work, which found that SSI program rules significantly influence household behavior prior to the SSI eligibility age of 65. Quite aside from behavioral issues, the evidence presented here suggests lessons for program design. Welfare systems can unintentionally build in perverse financial penalties when they interact with other programs used by the same population. Changes in program design that would better integrate SSI and social security, such as giving SSI recipients greater credit for their social security contributions (Burkhauser and Smeeding, 1981), would reduce the SSI-induced differential in the financial terms of delaying retirement. (Alternatively, if it is thought that the impoverished elderly gain too many resources by exploiting both OAI and SSI, penalties for early retirement similar to those under OAI could be extended to SSI). Aside from changing the programs, if policymakers want impoverished groups to rely on government transfers in their early 60s, and if–as seems likely–not all potential beneficiaries take account of the SSI-OAI interaction–then SSA could advise all households with low social security earnings histories that under SSI it may be advantageous for them to make an early OAI claim.
The MRRC provided funding for the conference, Improving Social Insurance Programs, that was held at the University of Maryland College Park on September 13th and 14th. The program was organized by John Rust of the University of Maryland, John Laitner, MRRC Director, and Robert Willis of the University of Michigan. Multiple sessions met concurrently over the two days of the conference.
In September 2003, Commissioner of Social Security, Jo Anne Barnhart, presented an approach to improving the disability determination process that would shorten decision times, pay benefits to people who are obviously disabled much earlier in the process and test new incentives for those with disabilities who wish to remain in, or return to, the workforce.
Commissioner Barnhart’s approach is the result of more than a year and a half of study and discussions with groups involved in the disability process – including Social Security employees from across the country, state Disability Determination Services (DDS), Administrative Law Judges (ALJs), health care providers, the federal courts, claimant attorneys and representatives, claimant advocacy groups, Members of Congress and concerned members of the public.
Under the current process, it takes an average of 628 days for a Social Security disability application that is denied and appealed at each step to reach final agency action.
None of the suggested changes would require legislative action; none would adversely affect the employment status of current Social Security or DDS employees.
Commissioner Barnhart’s approach is predicated on successful rollout of the Accelerated electronic Disability process (AeDib). AeDib creates an electronic folder for the claimant’s application, medical information and other data. AeDib will eliminate numerous time delays and financial costs related to locating the paper file, maintaining its contents and mailing them from office to office. The Social Security Administration is currently piloting AeDib in three states, and will phase in its use nationwide during an 18-month roll-out.
At a hearing before the House Ways and Means Subcommittee on Social Security, Commissioner Barnhart described highlights of her approach. They include:
Social Security has developed a Service Delivery Plan to eliminate backlogs within a 5-year period. The President’s FY 2004 Budget Request provides significant additional funding to begin to eliminate these backlogs. Backlogs at the DDS have already been reduced. Progress is being made at hearing and Appeals Council levels.
Social Security plans to conduct several demonstration projects aimed at helping people with disabilities return to work. These projects would support the President’s New Freedom Initiative and provide work incentives and opportunities earlier in the process. They would include: early intervention demonstration projects that provide medical and cash benefits and employment supports to applicants who elect to pursue work rather than disability benefits; temporary allowance demonstration projects that provide immediate cash and medical benefits for a specified period to applicants who are highly likely to benefit from aggressive medical care; interim medical benefits demonstration projects that provide health insurance coverage to certain uninsured applicants whose medical condition is likely to improve with medical treatment; and ongoing employment supports to assist beneficiaries who wish to obtain and sustain work will also be tested.
Source: Social Security On-line (www.socialsecurity.org).
In 2001, the Social Security Administration processed over three million disability claims. This chart provides a gross summary of the outcome of these initial claims Approximately forty percent of disability applications are approved at the first level of the process. Most claims are allowed (54 of 100), and most allowances are made at the initial level.
The Inter-university Consortium for Political and Social Research (ICPSR) now offers instant online data analysis for several studies (www.icpsr.umich.edu/access). Of interest is the National Health Interview Survey: Longitudinal Study of Aging. The data analysis system (DAS) enables users to perform online analysis directly over the Web, without downloading files. This is especially useful for users who do not have statistical software installed on their current workstation. In addition, DAS enables users to create and download subsets of the data files. ICPSR also offers a wealth of data on aging through their National Archive of Computerized Data on Aging (NACDA) (www.icpsr.umich.edu/NACDA).
Below is a selection of abstracts from recently released MRRC working papers. Visit our website for full papers and other current papers. www.mrrc.isr.umich.edu.
Abstract: More than 40 percent of Social Security beneficiaries continue to work after age 65. This research investigates the extent to which these individuals substitute labor across periods in response to anticipated wage changes induced by the Social Security earnings test. While we find that a disproportionate number of individuals choose earnings within a few percentage points of the earnings limit, we find no evidence that these individuals substitute labor supply between ages 69 and 70 when, in our sample, the tax on earnings falls from 50 percent to zero.
Abstract: This paper uses panel data from the Health and Retirement Study to estimate the relationship between measures of labor supply flexibility and portfolio-choice decisions by utility-maximizing individuals. Seminal research on portfolio decisions over the lifecycle, and recent research on stochastic dynamic programming models with endogenous labor supply and savings decisions suggest that, other things equal, individuals with more labor supply flexibility are likely to invest more in risky assets, regardless of their age, because of the insurance component that flexible labor supply provides. After controlling for panel sample selection and unobserved heterogeneity I find that labor supply flexibility leads to holding between 12% and 14% more wealth in stocks.
Abstract: The effects of poor health habits on mortality have been studied extensively. However, few studies have examined the impact of these health behaviors on workforce disability. In the Health and Retirement Study, a nationally representative cohort of 6044 Americans who were between the ages of 51 and 61 and who were working in 1992, we found that both baseline smoking status and a sedentary lifestyle predict workforce disability six years later. If this relationship is causal, cost-benefit analyses of health behavior intervention that neglect workforce disability may substantially underestimate the benefits of such interventions.
The Workshop on the Economics of Social Insurance Policymaking: Theories, Models, and Methods, co-sponsored by the University of Maryland, MiCDA and MRRC, was held at the Maryland Population Research Center on the campus of the University of Maryland, College Park in September. The workshop was designed for government agency and graduate student policymakers and analysts, and others interested in surveying the current trends in economic analysis of social insurance, with specific focus on social insurance policy making. The lectures surveyed the latest theories and models and provided "hands-on" demonstrations of how to solve them numerically, how to estimate and test them statistically, and most importantly, how they can be used in practice. Simulations using these models yield quantitative forecasts, important tools that assist policymakers in designing effective policies and assessing the political consequences of alternative policies (including whom policy interventions hurt and help).
University of Michigan
Institute for Social Research
426 Thompson Street
Ann Arbor, MI 48106-1248