Expert calls for an overhaul of the existing evidentiary framework for Social Security disability cases.
For six-and-a-half decades, the Social Security Administration (SSA), the courts, and Congress have engaged in a tug-of-war to find a reasonably reliable, reasonably efficient, reasonably fair source of vocational evidence to support consistent disability adjudication. The latest chapter is Biestek v. Commissioner of Social Security, which the Supreme Court decided in April.
As Richard Pierce described in a recent essay in The Regulatory Review, the facts of the Biestek case are straightforward. Biestek applied for Social Security disability benefits. Based on the testimony of an employment counselor (under contract with SSA as a vocational expert, or “VE”), the administrative law judge (ALJ) found that Biestek was not disabled because significant numbers of sorter and bench assembler positions were theoretically available to him. Individuals are disabled for Social Security purposes if they are unable to engage in any “substantial gainful work” which exists in significant numbers in the national economy. The burden rests with the government to prove the existence of substantial gainful work in significant numbers.
The problem in Biestek lay with the evidence. After the claimant’s representative asked the VE about the basis for her testimony, the VE cited private job surveys she had prepared for other clients. When the representative asked the VE for the surveys, she refused, claiming client confidentiality. The ALJ declined to require her to turn over the surveys. In finding the claimant not disabled, the ALJ relied on the VE’s testimony without making the private surveys part of the record.
As Pierce writes, we know next to nothing about the existence or reliability of the private surveys, nor whether they support the VE’s testimony. In that way, Biestek reveals the real anxieties that reliance on VE has provoked—namely, that VE testimony is inconsistent, that it has an unclear factual basis, and that it is not testable or susceptible to rebuttal.
Pierce offers one solution: adopt the rule of Daubert v. Merrell Dow Pharmaceuticals, or something like it, in the administrative context. In Daubert, the Supreme Court held that judges may only admit expert testimony that is reliable and fairly based in scientific principles. Importing the Daubert rule into administrative adjudication could increase accuracy. At least in the SSA disability context, though, the rule could also increase administrative costs (by requiring additional testimony and thereby lengthening hearings in an already burdened hearing system) without resolving the underlying questions about VE expertise.
SSA has proposed another solution: the development of a new Occupational Information System (OIS). As envisioned, the OIS will collect reliable and up-to-date data on occupational requirements and job numbers from the U.S. Department of Labor’s Bureau of Labor Statistics (BLS) and other sources and make it available through an online platform. As Social Security claimants’ representative Charles T. Hall has noted, OIS has the potential to be “the most important policy development at the agency in more than thirty years.” Following Biestek, this new tool could also go a long way to quell fears about VEs and junk science in the administrative process.
Ever since Congress amended the Social Security Act to cover disability in 1954, SSA, the courts, and Congress have struggled to find an appropriate source of occupational evidence. Following the 1955 recommendation of a group of consulting physicians, public health officials, and other experts, SSA published vocational evaluation guidelines for adjudicators that addressed the vocational impact of factors such as advanced age, limited education, and unskilled or no prior work experience.
Adjudicators were also free to consult other reference resources, such as government and industrial studies. As claimants challenged this practice, courts grew increasingly skeptical of how SSA adjudicators used occupational reference resources. In response, SSA in 1966 established a Vocational Consultant Staff comprising VEs. VEs provided evidence at hearings in individual cases about occupational requirements and job availability. VEs were not agents of SSA, and the agency expected them to “remain completely objective and impartial.”
Around the same time, Congress pushed SSA toward developing rules of general applicability. In an influential 1960 report, a new House Social Security subcommittee recognized “the difficulty of developing and enunciating specific criteria for the weight to be given nonmedical factors in the evaluation of disability and the extreme sensitivity of this area.” The subcommittee stated, however, that “the time has come, if it is not well overdue, to make a determined effort to develop and refine these criteria and make them available to the evaluators and to the public in the form of published regulations.”
SSA responded by crafting regulations that required adjudicators to consider a claimant’s age, education, and skill level of previous work experience when determining whether she could perform other work in the national economy. Congress ultimately incorporated these vocational factors into the Act itself in 1967. SSA also experimented with “medical-vocational profiles” that directed findings of disability for certain combinations of age, education, work experience, and exertional capacity.
By the mid-1970s, observers criticized the inconsistent outcomes of SSA decision-making. The New York Times wrote that the “Social Security disability program has become, in the view of its critics, the most arbitrary of the Government’s programs to help the needy, one in which poor people in similar circumstances often receive vastly different treatment.” Some observers blamed inconsistent outcomes on the Act’s definition of “disability.” Others recommended that ALJs “take official notice at the hearing of vocational facts that can be established by widely recognized documentary sources or on the basis of agency experience.”
The issue attracted significant congressional and scholarly attention, which resulted in SSA adopting by regulation the Medical-Vocational Guidelines in 1978. These guidelines consist of three matrices—each corresponding to a physical exertion level (sedentary, light, medium)—that contain rules corresponding to various combinations of age, education, and work experience. Each rule points to a decision: “disabled” or “not disabled.” SSA reached these conclusions based on information about the requirements and availability of jobs in the national economy contained in authoritative occupational publications, especially the Dictionary of Occupational Titles.
Like sentencing guidelines, SSA’s Medical-Vocational Guidelines provide a conceptually elegant solution. They certainly impose a degree of consistency on decision-making compared to what existed before. Yet they also petrify adjudicative practice and sacrifice accuracy over time by encasing dynamic labor market conditions in regulatory amber. Until the agency amends them, through a rulemaking process that can be costly and time-consuming, the guidelines will continue to reflect labor conditions as they were in 1978.
The guidelines also contain significant informational gaps. They do not reflect the occupational effects of nonexertional (e.g., mental) limitations. If a claimant has a combination of exertional and nonexertional limitations, a rule directing a finding of “not disabled” is inconclusive, and additional information is needed to resolve the claim. If all of a claimant’s limitations are nonexertional, the grid rules do not apply at all. In both cases, ALJs must seek occupational information from other sources to reach a decision.
With broader societal recognition of mental impairments, many claimants are ultimately found to have mental or cognitive restrictions. In such cases, any rule directing a finding of not disabled based on physical capabilities will be insufficient to support a decisional outcome. The result, as some feared, is that the exceptions have largely swallowed the rules.
ALJs may consult the Dictionary of Occupational Titles, but it does not address many of the mental and cognitive requirements of occupations. Last updated in 1991, the Dictionary is also woefully outdated. It still documents the job tasks of elevator operators and became obsolete even before the widespread adoption of computers. It is also an imperfect resource for job numbers.
In the 1980s and 1990s, SSA experimented with a series of policy statements to provide adjudicators additional clarity on the occupational effects of various limitations. There are only a handful of such statements, though, and they are rarely definitive and rarely address the combined effects of multiple limitations.
Today, ALJs almost universally rely on VEs to determine whether there are significant numbers of jobs a claimant can perform. Some courts have practically mandated the use of VEs in certain circumstances, especially where a mental impairment is present. Under current policy, a VE may be required to explain the basis for her testimony or explain why her testimony is inconsistent with the Dictionary of Occupational Titles. If the VE bases testimony on an assumption, the ALJ may ask her to clearly describe the assumption. Claimants and their representatives have the right to question VEs so as “to inquire fully into the matters at issue.”
At Biestek’s core is the anxiety over the basis for VEs’ testimony and the concern that such testimony may be opaque or practically unreviewable.
New technologies and better data analytics now make another option possible: reliance on a more comprehensive database of vocational information. This approach promises to promote all systemic goals—fairness, predictability, accuracy, consistency, efficiency—over the status quo. Because textual data can be made widely available, the public can critique the reliability of their methods and the accuracy of their contents. Parties have notice of the evidence that will be used in their cases. They can cite data supporting their own arguments and more effectively challenge an ALJ’s factual conclusions. Reviewing courts also have access to the resource on review. As Biestek’s attorney argued before the Supreme Court, “it is certainly a different case if an expert only relies on public data.”
Traditional documentary resources have proved insufficient; the Dictionary of Occupational Titles is outdated and incomplete, and other widely-used resources may be ill-suited for Social Security disability determinations. But if no good alternative exists, why not create a new one?
SSA is working with BLS to develop the new OIS, which should replace the Dictionary and reduce reliance on less reliable vocational resources. As planned, the OIS will survey far more of the limitations frequently encountered in disability adjudication, including the mental requirements of occupations. SSA and BLS plan to update the OIS at regular intervals (about every five years) to avoid the ossification that has plagued the Dictionary of Occupational Titles for almost 30 years. Using a web-based platform, claimants and their representatives will have easy access before their hearings to the same vocational evidence ALJs will use in their cases.
If implemented correctly, the OIS could improve the accuracy of SSA decision-making. By treating like claimants alike, it could better ensure consistency across decision-making outcomes. And, by providing claimants and their representatives with easy, web-based access to the same vocational tool used by SSA decisionmakers, it has the power to promote procedural fairness.
Effective implementation, of course, depends on attention to details. To serve as the factual foundation for millions of claims each year, it is vital that data be accurate. They must be programmatically appropriate and consistent with decades of congressional, agency, and judicial developments. Such evidence should be the primary driver of policy changes, and any final rules or policy statements should by supported by the record. So far as practicable, data should be publicly transparent during their development and easily accessible and usable in the adjudication of individual cases.
A new source of occupational evidence has the potential to revolutionize SSA decision-making. Structuring mass adjudication requires balancing accuracy, consistency, efficiency, and fairness. Providing adjudicators and parties with easy access to a reliable and more efficient source of vocational information, and reducing the use of less reliable sources of vocational information, has the potential to benefit all of these goals over the status quo. Doing so can also serve as a lesson on how the thoughtful use of data and technology can benefit public administration.
Although Biestek could eventually institutionalize the use of junk science in the administrative process, as Pierce describes, that result is not inevitable. In the end, the ruling may be just one more chapter in a 65-year search for better evidence in Social Security disability adjudication.