Much is being said these days about Social Security’s expanding disability rolls. Many have sensationalized the increase in the number of recipients of disability benefits, portraying a system populated with malingerers and “welfare queens” seeking to game the system, the latest in what some have described as the “fictional genre known as the ‘undeserving poor.’…The goal is to rationalize a cut in benefits by portraying their beneficiaries as morally depraved.”
For advocates for the poor and disabled, these representations are far from reality. Award rates areamong the lowest in history and have fallen during the recession. The standard for disability is stringent, one of the strictest in the developed world. In fact, benefits are often wrongfully denied even when the standard is met and even to the most severely disabled.
For decades, thousands of claimants for disability benefits in Queens, New York, were denied by adjudicators who were improperly denying benefits in a pattern of repeated, glaring, and often intentional legal errors. In response, the Urban Justice Center, with pro bono co-counsel Gibson Dunn & Crutcher, filed the class action Padro. v. Colvin in the Eastern District of New York on April 12, 2011, for the systematic failure to provide full and fair hearings. The lawsuit charged that five of the eight administrative law judges in the Social Security Administration’s Queens hearing office—David Z. Nisnewitz (then Chief Administrative Law Judge of that office), Michael “Manuel” Cofresi, Seymour Fier, Marilyn P. Hoppenfeld, and Hazel C. Strauss—were generally biased against claimants for disability benefits.
The complaint alleged that in case after case, these administrative law judges showed a shocking hostility to applicants and invented reasons to deny benefits. It claimed that they routinely engaged in bullying and outwardly hostile conduct, systematically cherry-picked facts and manipulated evidence to support preordained conclusions, and willfully subverted and ignored established law, even in the face of explicit instructions to the contrary from federal courts and the Social Security Administration’s Appeals Council.
The Social Security Administration’s Queens hearing office is one of four in New York City. Local advocates already knew that claimants appearing in Queens—the fourth-most densely populated county in the United States and the most ethnically diverse—faced brutal treatment and markedly poor odds compared to their neighbors in the rest of New York City. Their suspicions were confirmed in 2008 when the claim-approval rates of administrative law judges were made public for the first time outside of requests under the Freedom of Information Act. (The Social Security Administration has since begun publishing administrative law judge disposition data on its own website.) Those statistics revealed that claim-approval rates in Queens were not only the worst in the city but also the third worst in the entire nation. Individually, these administrative law judges ranked very highly among the nation’s top claims deniers. Collectively, they formed what was described by the Padro complaint as a “brick wall of bias.”
Although the severity of the problem was obvious to many advocates, the problem of administrative law judge bias is a daunting one, and litigation was by no means an obvious solution. Only after encouragement from pro bono partner Gibson Dunn & Crutcher did the Urban Justice Center begin to consider a lawsuit. The Urban Justice Center had already employed various other strategies and forms of advocacy, but they had all proven unavailing or insufficient.
Employment and Independence of Administrative Law Judges
Understanding the nature of the administrative law judge position is necessary to appreciate some of the complications involved in cases charging administrative law judge bias. The Social Security Administration employs administrative law judges; however, the Office of Personnel Management is exclusively charged with their hiring. To remove, suspend, reduce in level, reduce in pay, or furlough for 30 days or less an administrative law judge, the Social Security Administration must bring a proceeding with the Merit Systems Protection Board and show good cause. The public is unable to bring a proceeding for removal; only the Social Security Administration may do so.
Thus, while administrative law judges work for and are paid by the Social Security Administration, they are subject to limited agency management. Because of the lack of a mandatory retirement age and the difficulty of removing an administrative law judge, they have, in effect, lifetime tenure. As a result, the independent nature of the administrative law judge position creates a degree of unaccountability, even when an administrative law judge’s adjudication is tainted by bias. Nonetheless, an administrative law judge’s independence from agency influence or coercion is essential to ensuring due process and the right to a full and fair hearing. The Supreme Court has recognized the importance of an administrative law judge’s ability to “exercise[ ] his independent judgment on the evidence before him, free from pressures by the parties or other officials within the agency.”
Limited Administrative Options to Handle Generalized Bias
The procedures that the Social Security Administration has in place to address bias—the recusal, complaints, and appeals processes—have been woefully inadequate to handle the problem of generalized bias.
Recusal. Under Social Security Administration regulations, a claimant may request “at [the] earliest opportunity” that an administration law judge withdraw. In addition to requiring advance knowledge of an administrative law judge’s bias, the procedure poses the inherent difficulty of presenting a claim of bias to the very judge who is alleged to be biased.
More significantly, the recusal provision is directed at specific, not general, bias. Courts have agreed that the regulation appears to fail as “an adequate mechanism when the bias alleged does not relate to a particular plaintiff, but is based on a claim that the [administrative law judge] is generally biased against all Social Security claimants.”
Complaints. The complaint process has proven similarly ineffective. At the time Padro was filed, the complaint processes available included a set of interim procedures issued in 1992 (Social Security Administration Procedures Concerning Allegations of Bias or Misconduct by Administrative Law Judges, 57 Fed. Reg. 49,186 (Oct. 30, 1992) (interim procedures)). They noted that additional procedures were under development and were expected to issue “in approximately six months” (id.). The procedures required prompt notification to the claimant that the claimant’s complaint had been received. Complaints were to be investigated promptly and claimants notified of the result. These procedures added to those set forth in the Social Security Administration’s policy manuals, which describe a byzantine, inconsistent, and convoluted set of protocols that refer to numerous entities within the Social Security Administration.
In practice, the complaint process offers little to no information to the claimant and exists in a vacuum. If a complaint results in any corrective action toward the administrative law judge, the outcome has no bearing on any appeal that the claimant may have filed.
In Queens, possibly the most futile aspect of the complaint procedures concerned the role of the chief administrative law judge of the hearing office. The chief judge participates in investigations into allegations of misconduct or bias, responds to complaints, and conducts training. However, the chief judge of the Queens hearing office was in fact one of the five administrative law judges named in Padro.
The agency lacked an adequate method to track complaints. They were maintained largely by the hearing office with no method of centralization. The Social Security Administration created a new repository in 2010, acknowledging that it did “not have a good mechanism to track complaints about ALJs from initiation to resolution.” However, even the new system fails to capture key information—such as the multiple bases for a complaint—necessary to recognize patterns. Processing time can be unacceptably long: up to 894 days at the hearing level. Without the ability to track patterns, the system fails to prevent or resolve general bias or bias directed at a particular group.
Appeals. The appeals process is similarly insufficient. A claimant may raise allegations of bias on appeal when requesting review of the merits of an administrative law judge’s decision from the Social Security Administration’s Appeals Council. The Appeals Council may refer an allegation of misconduct to a separate entity at the Social Security Administration, but the Appeals Council itself has no authority to take action toward the offending administrative law judge. The evidence that would be available is constricted; on appeal, the evidence is limited to the record. However, general bias depends, by definition, on evidence outside the record. Without authority to make findings beyond an individual case under review or the capacity to identify and address systemic patterns, the Appeals Council can make only scattered individual findings. The fragmented design of the Appeals Council exacerbates this internal lack of coordination. Denials issued in the Second Circuit in 2011 went to one of at least 10 scattered branches—a problem that posed inherent internal coordination problems.
The actions that the Appeals Council may take are limited. When it finds errors, Social Security Administration policy is to remand the claim to the same judge to correct mistakes. If bias or misconduct is found, the case may be reassigned to a different administrative law judge. However, in Queens, a claimant’s case was likely to be reassigned to another of the five complained-about administrative law judges. And this slim relief came only after an extensive wait; at the time Padro was filed, the average processing time at the Appeals Council was 14 months.
Other Options. In Queens, advocates found some redress at the Appeals Council and in individual federal court actions, but the efforts were insufficient to reach the systemic nature of the problem. The Urban Justice Center considered a media campaign and appealing to members of Congress for help, but similar efforts in other jurisdictions had garnered little success.
Litigation had been used to some degree in the past, but the results were inconsistent, and almost all of those cases were long and protracted. Some had taken decades to resolve. Only after the encouragement and commitment of extensive resources from the Urban Justice Center’s pro bono partner, Gibson Dunn & Crutcher, did litigation become an option.
Federal Court Claims of Generalized Bias
The federal courts have found subject matter jurisdiction in cases charging general bias by an administrative law judge. Examples include Pronti v. Barnhart (remanding case and holding that “the Court retains jurisdiction of the action”), Small v. Sullivan (ruling that court had subject matter jurisdiction over general bias claims), Kendrick v. Sullivan (finding subject matter jurisdiction in case charging systematic bias against disability benefits claimants), and Grant v. Sullivan (finding exhaustion requirement of 405(g) jurisdiction was waived).
Although jurisdiction is available, the extent of discovery available remains in question. The Social Security Act permits a federal court to order the Social Security Administration to take additional evidence. The district court in Grant had done so, but on appeal, the Third Circuit found that the evidence was limited to the existing record.
These cases demonstrate the importance of fact development. The most damning evidence sometimes comes from depositions, such as one in Grant that featured detailed testimony that the administrative law judge invoked racial slurs, referred to claimants as “no-goodniks,” and believed that certain ethnic groups faked mental illness. In Pronti, plaintiffs submitted compelling affidavits with data and analysis of denial rates correlated with evidence of legal error, remands, and adverse credibility determinations. These cases illustrate that statistical evidence alone is not enough to prove bias. For example, in one unsuccessful challenge, a court found that the anecdotal evidence offered by the plaintiffs was vague and that the statistical evidence was insignificant and lacked context or connection to other evidence (see Doan v. Astrue, 2010 WL 1031591 (S.D. Cal. March 19, 2010)).
Allegations in the Padro Complaint
Compared to prior lawsuits, Padro was significantly broader in scope. It was the first to complain of bias by more than one administrative law judge and to include allegations of bias by the chief administrative law judge of the hearing office.
The Urban Justice Center brought the Padro lawsuit under the Social Security Act, the Mandamus Act, and the Due Process Clause. The lawsuit sought declaratory and injunctive relief, including a permanent injunction barring these administrative law judges from hearing claims for disability benefits. The Urban Justice Center initially filed the case with eight named plaintiffs, many of whom had been denied more than once and had had the administrative law judge ignore instructions from the Appeals Council on remand. Three additional plaintiffs joined the case on May 4, 2011, in an amended complaint.
The amended complaint cited the nonadversarial nature of Social Security claims adjudication, noting that it was rooted in the beneficent and remedial purpose of the Social Security Act and that the Act was to be “broadly construed and liberally applied.” The amended complaint noted that a “stricter application” of the impartiality requirement applies in the administrative context, where there is an “absence…of procedural safeguards normally available in judicial proceedings ….”
The judges in the Eastern District of New York had become increasingly admonishing of the five named administrative law judges. In framing the lawsuit, class counsel analyzed the findings of scores of district court decisions criticizing the five administrative law judges as evidence of a pattern of repeated errors indicative of bias. The complaint alleged that the decisions and the facts pertaining to the individual claims of the named plaintiffs demonstrated a pattern of the following errors: (1) the failure to apply well-settled law, including the treating-physician rule; (2) the failure to develop the record; (3) improper credibility determinations; (4) hostile and unprofessional conduct; and (5) failure to follow instructions on remand.
The complaint cited the district court’s language describing decisions issued by the five administrative law judges, much of which was quite harsh: opinions that were “deficient,” “incoherent,” “plucked from thin air,” “arbitrary,” and “illogical” and conduct that was “particularly egregious,” “exhibit[ing] bias,” showing “inappropriate hostility” toward benefit claimants, peppering claimants with “combative questioning,” showing “serious negligence,” and “trivializing” claimant impairments.
These decisions reflected a remand rate of over 80 percent. The complaint cited this and other figures, including the high denial rates. While a powerful indicia of bias, statistics were only one element of the case, as prior bias cases instructed that “statistics in and of themselves may have limited probative value.” As in Pronti, plaintiffs submitted affidavits from six attorneys and advocates from both nonprofit and private law firms describing their experiences with the five administrative law judges.
How the Litigation Unfolded
Padro initially appeared on track to last at least as long as the prior bias cases, which is to say that it was expected to last a very long time. Plaintiffs moved for expedited discovery early in the litigation to obtain evidence for a preliminary injunction motion. The Social Security Administration fought discovery aggressively. The court granted the plaintiffs’ motion, and the Social Security Administration appealed. The agency was forced to produce some documents before oral argument on its motion to dismiss on February 3, 2012, where the court stayed discovery temporarily and ordered the parties to enter settlement negotiations. No further discovery was taken, and the court never ruled on the pending motions because the parties filed a proposed settlement agreement on January 11, 2013. The parties then moved for preliminary approval, and the plaintiffs moved for class certification on February 8, 2013.
The settlement drew objections from the administrative law judges’ union and from several individual administrative law judges who complained, among other things, that the five were not afforded due process and that the allegations had not been adequately challenged. The court found that the objections lacked merit. After holding fairness hearings on June 16 and September 19, 2013, the court granted final approval of the settlement on October 18, 2013.
The Terms of Settlement
The settlement in Padro gives both retrospective and prospective relief. Thousands denied in the past will have the opportunity to have their cases reheard by a different administrative law judge. Any claimant whose claim was denied by one of the five named administrative law judges during the period from January 1, 2008, to October 18, 2013, the date of the settlement’s approval, will have the right to a new hearing by an administrative law judge other than one of the five named. The class member must affirmatively request the relief within 60 days of receiving notice of potential entitlement to relief. Some exclusions apply, such as if the person’s claim had already been reheard by an administrative law judge other than the five named. Approximately 4,000 individuals may be eligible for readjudication.
Prospectively, the settlement includes measures to protect the rights of thousands more in the future. For 30 months, future denials issued by the five administrative law judges will be monitored. The settlement requires review by a special, newly created unit charged with review of all such denials. Ordinary appeal standards apply, but in addition, the unit must review for the type of conduct complained about in the lawsuit. Unfavorable decisions will be automatically reviewed. In a change from ordinary agency policy, any subsequent hearing on appeal cannot be heard by any of the five named administrative law judges. Additionally, the Social Security Administration agreed to conduct administrative law judge training on the issues complained about in the lawsuit.
The settlement required the Social Security Administration to issue a new agency ruling on its procedures for addressing bias—something the agency had promised to do over 20 years ago. The ruling was published on January 29, 2013.
Many significant changes, not included in the terms of the settlement, have evolved since the filing of the lawsuit. The chief administrative law judge of the hearing office, accused of setting a tone and fostering a culture of anti-claimant bias, was replaced. Denial rates have changed dramatically. In the fall of 2013, the five administrative law judges were approving over 81 percent of the claims before them. Advocates noted that a broader roster of experts began to appear at hearings, and those experts were no longer permitted in the interior of the hearing office. All experts appearing in person now wait in the general waiting room for cases to be called.
The only relief not obtained in the settlement is removal of the five named administrative law judges—a result the court suggested it did not have the authority to grant. The Social Security Administration has initiated proceedings with the Merit Systems Protection Board in past bias cases while the claimant’s lawsuit was pending, but we are unaware of any proceedings being brought against these five administrative law judges.
Implementation of the settlement is now underway. Individuals have begun requesting readjudication, and the Social Security Administration is notifying class members of their eligibility for relief. Challenges for class counsel now include explaining various aspects of Padro relief and ensuring that class members are able to obtain the relief to which they are entitled.
We hope that we have achieved lasting change in the Queens hearing office, beyond the 30-month monitoring period. This remains to be seen, as does the efficacy of the new bias policy. For now, we are grateful that the class action avoided decades of prolonged litigation and that thousands of New Yorkers with disabilities obtained relief that may not have been achieved had the case proceeded to trial, including the opportunity to have their claims reheard and additional due process protections as they pursue their claims for disability benefits.