The Mathews Test

The Mathews Test

Goldberg v. Kelly left a practical problem behind it. Its package of hearing rights (oral presentation, confrontation, counsel, and a decision on the record) was justified by the stakes of a welfare cutoff, yet the opinion raised the concern that courts would soon impose trial-type procedures in many other programs, including programs in which courtroom methods would be ineffective or too expensive.¹ A fixed package invites that spread, because once a court has named the elements of a hearing, it has no principled way to explain why a different program should receive fewer of them. Mathews v. Eldridge answered by reversing the presumption that trial-type hearings are the due process norm.² The Court began from the premise that "due process is flexible and calls for such procedural protections as the particular situation demands."³

Flexibility needs a method, and the Court supplied one in the form of three factors. The first is the private interest affected by the official action. The second is the risk of an erroneous deprivation through the procedures already used, together with the probable value of any additional or substitute safeguards. The third is the government's interest, including the function involved and the fiscal and administrative burdens that the added procedure would impose.⁴ Taken together, the factors operate as cost-benefit balancing, so a court asks whether the value of the procedures a claimant demands beyond those the government already provides justifies the cost of adding them.⁵ Because the question concerns that increment, each requested safeguard is analyzed separately, and a court may settle on an intermediate procedure instead of choosing between a full trial-type hearing and the informal process the agency already uses.⁶ The weights shift from program to program, which is why the results span a wide range: an oral evidentiary hearing before AFDC benefits end, a written exchange before disability benefits end, and a brief informal meeting before a school suspension.⁷

The Court has continued to invoke the three factors in a wide variety of settings, including the detention of suspected enemy combatants in Hamdi v. Rumsfeld, so the same questions organize most of what remains in this chapter.⁸ The subsections below take them in order. The first covers the case itself, where the Court set a disability cutoff against Goldberg's welfare cutoff and weighed the factors for the first time. The next three teach the private interest, the risk of error, and the government's interest one at a time, and the closing section of this chapter returns to the limits of the framework. The first subsection, Mathews v. Eldridge, begins with the facts of the disability termination and the procedures the agency had already given the claimant before his benefits stopped.

Mathews v. Eldridge

Mathews v. Eldridge looks, at first glance, like a second chapter of Goldberg v. Kelly. A federal benefit program had found the claimant eligible, the agency proposed to end payments on the ground that eligibility had lapsed, and the claimant demanded an evidentiary hearing before the money stopped. The District Court and the Court of Appeals, treating Goldberg as controlling, gave him that hearing, and the Supreme Court reversed.⁹ The reversal shows that finding a protected interest does not carry Goldberg's trial-type package along with it. The Court reached that result by pointing to particular facts about disability benefits. Those facts are what this subsection draws out, and the formal weighing that organizes them is taken up in the subsections that follow.

The claimant was George Eldridge, who was first awarded disability benefits in June 1968. In March 1972 the state agency charged with monitoring his medical condition sent him a questionnaire, and he returned it indicating that his condition had not improved and naming the doctors and other sources who had recently treated him.¹⁰ The agency obtained reports from his physician and from a psychiatric consultant, and after reviewing them it wrote to tell him that it had tentatively determined that his disability had ceased in May 1972.¹¹ The letter stated the reasons for the proposed termination and advised him that he could request reasonable time to submit additional information about his condition.¹² Eldridge answered in writing, disputing one characterization of his medical condition and asserting that the agency already had enough evidence to establish his disability.¹³ The agency made its determination final, the Social Security Administration (SSA) accepted it, and the SSA notified him that his benefits would end and that he could seek reconsideration.¹⁴ Eldridge skipped reconsideration and sued, attacking the constitutionality of the procedures established by the Secretary of Health, Education, and Welfare and asking for reinstatement of his benefits pending a hearing on whether he remained disabled.¹⁵ The Court held that the District Court had jurisdiction despite the skipped step, because the procedural challenge was collateral to the claim for benefits and the claim had been presented to the Secretary.¹⁶

The procedures Eldridge attacked are best understood as a sequence with a dividing line at the cutoff, since what the Court weighed was how much the agency had done before payments stopped and what remained available afterward. The sequence ran as follows:

1. A state agency team made up of a physician and a nonmedical person trained in disability evaluation communicated with the worker, usually by mailed questionnaire, gathered information from the worker's treating sources, and arranged an examination by an independent consulting physician when the sources conflicted.¹⁷
2. When the agency's tentative assessment differed from the worker's own, it informed the worker that benefits might be terminated, provided a summary of the evidence supporting the proposal, and let the worker review the medical reports and other evidence in the file.¹⁸
3. The worker could respond in writing and submit additional evidence, after which the state agency made its final determination, which an SSA examiner reviewed.¹⁹
4. When the SSA accepted the determination, it notified the worker in writing of the reasons, and benefits ended two months after the month in which medical recovery was found to have occurred.²⁰
5. After the cutoff, the worker could seek de novo reconsideration by the state agency and then an evidentiary hearing before an SSA administrative law judge, a nonadversary proceeding in which the SSA was not represented by counsel, followed by discretionary review by the Appeals Council and judicial review.²¹
6. A worker later found to have remained disabled past the cutoff date was entitled to retroactive payments, although the statute allowed the Secretary to try to recoup payments that a recipient was later found not to have been owed.²²

The narrowness of the dispute follows from that sequence. Eldridge agreed that the review available before the initial determination became final would be adequate if benefits continued until after the evidentiary hearing stage, so the only question was what process was due before the initial termination, while the review was pending.²³ The Secretary did not contest that continued receipt of disability benefits is a statutorily created property interest, so the Court proceeded on that footing.²⁴ The claim arose under the Fifth Amendment, which forbids deprivation of property without due process of law and applies here because the procedures were prescribed by a federal official rather than a state.²⁵

The positions of the parties set up the contrast with Goldberg. Eldridge relied exclusively on Goldberg, which established a right to an evidentiary hearing before termination of welfare benefits.²⁶ The Secretary answered that Goldberg did not control for two reasons: eligibility for disability benefits, unlike eligibility for welfare, does not depend on financial need, and issues of credibility and veracity play no significant role in a disability decision that turns primarily on medical evidence.²⁷ The District Court rejected that answer. It viewed the disability recipient's interest in uninterrupted benefits as indistinguishable from the welfare recipient's, and it reasoned that disability determinations may involve subjective judgments based on conflicting medical and nonmedical evidence.²⁸ It therefore ordered a pretermination evidentiary hearing of the type required for welfare recipients, and the Fourth Circuit affirmed in reliance on the District Court's opinion.²⁹

The Supreme Court began from a premise that disposed of the automatic extension of Goldberg: the fact that an interest is protected does not determine what process is due, because the requirements of due process are flexible and depend on the circumstances of the particular case.³⁰ It then placed Goldberg on the map. In only one case had the Court held that a hearing closely approximating a judicial trial is necessary, and in the other decisions requiring some pretermination hearing the Court had spoken sparingly about the procedures, requiring only a probable-cause determination of fault before a driver's license was revoked and sustaining a federal employee's dismissal procedure that offered a written response and an oral appearance before an evidentiary hearing that followed the dismissal.³¹ Goldberg was therefore the single case at the far end of a range, and the Court's effect was to retire the presumption that trial-type hearings are the due process norm.³² Measured against that range, the SSA's procedures sufficed, because an evidentiary hearing is not required before the termination of disability benefits.³³ Written notice and an opportunity to respond in writing before termination, followed by a full evidentiary hearing with retroactive benefits if the claimant prevails, satisfy the Fifth Amendment.³⁴

The reason the Court gave is that the judicial model of an evidentiary hearing is neither required nor even the most effective method of decision in every setting, because administrative agencies differ in origin and function from courts and cannot simply borrow court procedure wholesale.³⁵ What is essential is that a person in jeopardy of serious loss receive notice of the case and an opportunity to meet it, with procedures tailored to the capacities and circumstances of those who are to be heard.³⁶ The Court also gave substantial weight to the good-faith judgment of the officials who administer such programs, particularly where, as here, the claimant has an effective way to assert his claim before any administrative action and a right to an evidentiary hearing and judicial review before a denial becomes final.³⁷

The procedures in the two cases had the same basic architecture, which is what makes the pair instructive. In Goldberg the agency stated its reasons and accepted a written statement before payments stopped, and an oral hearing with back payment for a prevailing recipient followed afterward.³⁸ The disability procedure followed the same pattern and added a summary of the evidence and access to the file before the cutoff, yet the Court held it adequate where the welfare procedure was not.³⁹ The difference in result lies in three facts about disability benefits, and each previews a question that the next subsections develop.

The first fact is need. Welfare goes to people on the very margin of subsistence, so a wrongful cutoff may take away the very means to live while the recipient waits. Disability benefits are not means-tested, so a terminated worker may have other resources and may even qualify for welfare.⁴⁰ The second fact is the kind of question the agency must answer. A disability decision asks whether a physical or mental condition prevents work, an inquiry that usually rests on routine and unbiased medical reports and is sharply focused and easily documented, so a decisionmaker can assess it on paper.⁴¹ A welfare decision, by contrast, can depend on patterns of behavior, such as whether a mother is living with a man able to support her children, so witness credibility is often critical and the chance to watch a witness and cross-examine matters more.⁴² The third fact is who does the communicating. Welfare recipients usually have limited educations that make written advocacy difficult, while the main participants in disability proceedings are doctors, who typically write well.⁴³ Because the worker also received the agency's tentative assessment, its reasons, and a summary of the evidence, the written channel let the worker answer the precise issues the decisionmaker regarded as crucial, which is the function Goldberg assigned to oral presentation.⁴⁴ On the other side of the ledger, the Court observed that the cost of protecting those whom the preliminary process has identified as likely undeserving may come out of the pockets of the deserving, since program resources are limited.⁴⁵

Justice Brennan, who would have affirmed, rejected those premises. He called the idea that a disability cutoff causes only a limited deprivation speculative, and he reasoned that Congress's decision to pay disability benefits without any prior determination of need itself presumes a need that the Court should not denigrate.⁴⁶ The oral argument indicated that when Eldridge's benefits ended his home was foreclosed upon and the family's furniture repossessed, leaving his wife and children sleeping in one bed, and Brennan added that the possibility of seeking other public assistance is no answer.⁴⁷ The dissent points to the soft spot in the majority's reasoning: the distinctions are generalizations about how disability programs work, and critics of the decision have questioned whether they hold.⁴⁸ Those generalizations carry the result only so far as they are true of the program before the court, so a claimant who can show that a program's eligibility decisions in fact turn on credibility, or that its beneficiaries are in fact desperate, can use the same reasoning to argue for more process.

The usual exam error is to read Mathews as overruling Goldberg or as holding that disability recipients are owed no hearing at all. The claimant in Mathews received notice, reasons, access to the evidence, and a written chance to respond before the cutoff, and a full evidentiary hearing with retroactive pay afterward, and the Court held that sequence sufficient. A second error is to stop after finding that the benefit is property, which answers only whether due process applies and leaves open how much process is due. On facts with a need-based benefit, an eligibility dispute that turns on credibility, and a recipient who cannot easily write, Goldberg is the closer analogy, and where the benefit is not need-based and the decision rests on documents, Mathews is.

The Private Interest

The first Mathews factor looks at the private interest that will be affected by the official action, and it supplies one side of the balance that the other two factors complete.⁵⁰ Earlier sections asked whether an interest counts as property or liberty at all, and for that question the size of the loss was irrelevant once Roth replaced the older grievous-loss approach with an inquiry into the nature of the interest.⁵¹ Here the interest is already protected, and its weight becomes the question, because the amount of process due depends on how serious the loss would be. Goldberg put the point in these terms: the extent of the process owed is influenced by the extent to which a recipient may be "condemned to suffer grievous loss."⁵² The reason lies in the cost-benefit structure of the test, which asks whether the value of the added procedures a claimant demands justifies the cost of providing them.⁵³ An added safeguard is worth more the larger the harm it guards against, so a weighty private interest justifies more process and a slight one justifies less.

Mathews begins by defining the interest narrowly. Because a recipient whose benefits are terminated receives full retroactive relief if he ultimately prevails, the Court said, "his sole interest is in the uninterrupted receipt of this source of income pending final administrative decision on his claim."⁵⁴ Retroactive payment turns the question from whether the worker permanently loses the benefit into what the worker endures while waiting, so the first factor measures the cost of the gap in income. Measured that way, Eldridge's potential injury resembled the injury of the welfare recipient in Goldberg, the nonprobationary federal employee in Arnett v. Kennedy, and the wage earner in Sniadach v. Family Finance Corp.⁵⁵ The kind of harm was therefore shared, and the comparison with Goldberg turned on degree, with the Court finding the degree smaller for several reasons that the rest of this section takes in turn.

Need carries the most weight in the Court's comparison. Goldberg called it the crucial factor: ending welfare while an eligibility dispute is pending may deprive an eligible recipient of the very means by which to live while waiting, and because that recipient lacks independent resources the situation becomes immediately desperate, which in turn weakens the ability to seek redress.⁵⁶ Mathews found that disability benefits differ, because eligibility is not based on financial need and is wholly unrelated to the worker's income or to support from sources such as a family member's earnings, savings, private insurance, or pensions.⁵⁷ The structural reason is that a means-tested program selects its recipients by their lack of other resources, so every wrongful cutoff lands on a household with nothing to fall back on, while a program that pays regardless of income includes households able to absorb a gap. Welfare is a program of last resort, and a disability recipient may have other income or may even qualify for welfare while the dispute continues.⁵⁸ The Court added that other forms of government assistance become available if a termination pushes a worker's family below the subsistence level, which left less reason than in Goldberg to depart from the ordinary principle that something less than an evidentiary hearing suffices before adverse administrative action.⁵⁹

On the remaining considerations the Court conceded ground to the claimant. The potential deprivation in a disability case is generally likely to be less than in Goldberg, but the Court cautioned that the degree of difference "can be overstated."⁶⁰ A recipient must be unable to engage in substantial gainful activity to remain eligible, so unlike the discharged federal employee in Arnett, a terminated disability recipient has little prospect of finding even temporary work to soften the interim loss.⁶¹ That fact removes one of the usual cushions against a wrongful cutoff, so the Court's lower weighting rests on the absence of a need test and the availability of other aid.

Duration is the second consideration that moved toward the claimant. The possible length of a wrongful deprivation is an important factor in measuring the effect of official action on the private interest, the Court had said in Fusari v. Steinberg.⁶² The Secretary conceded that the delay between a request for a hearing before an administrative law judge and a decision was then between ten and eleven months, and because a terminated recipient must first obtain a reconsideration decision, the delay between the cutoff and a final decision after a hearing exceeds one year.⁶³ A deprivation that the law calls temporary lasts as long as the review process takes, and every month of a slow process adds costs that a later lump sum does not undo. Given that slow review and the typically modest resources of a disabled worker's family, the Court acknowledged that the hardship on an erroneously terminated recipient "may be significant."⁶⁴ It nonetheless concluded that the worker's need is likely to be less than a welfare recipient's, and the sum of the considerations left less reason to require an evidentiary hearing before the cutoff.

Retroactive relief deserves separate treatment because it shrinks the interest without erasing it. Full back payment is why the Court treated Eldridge's stake as the interim receipt of income, and a household that can absorb the wait loses little when the money later arrives.⁶⁵ Where the cutoff itself does the damage, though, repayment at the end cannot supply what the recipient needed during the wait. Goldberg located the harm in the period while the recipient waits, when the lack of income undermines even the effort to contest the decision, and a payment made after the dispute ends reaches the recipient too late to prevent that damage.⁶⁶ Retroactive payment therefore lowers the weight of a private interest to the extent that the loss is purely financial and the household can bridge the gap, and it lowers the weight very little when the gap produces losses that money cannot reverse.

The factor also reaches interests that are not measured in income. In Stanley v. Illinois the state declared an unmarried father's children wards of the state without any hearing on his fitness as a parent, and the Court recognized his substantial interest in the companionship, care, custody, and management of his children.⁶⁷ Stanley predates the Mathews factors, but it shows the heavy end of the range: the stake is the dismemberment of a family, and no later payment can restore the time a parent and child lost together. An interest of that kind weighs heavily under the first factor because the harm of an erroneous decision cannot be compensated afterward, and that is the consideration the factor measures when it asks how serious a wrongful deprivation would be.

The weighing in Mathews was contested then and has been contested since. Justice Brennan, who would have affirmed the lower courts, called the Court's view that a disability cutoff may cause only a limited deprivation speculative.⁶⁸ He reasoned that Congress's decision to pay disability benefits without any prior determination of need presumes a need on the recipient's part that the Court should not denigrate, and he noted that the oral argument indicated Eldridge's home had been foreclosed upon and the family's furniture repossessed after the cutoff, leaving the family sleeping in one bed.⁶⁹ The availability of other public assistance, he added, is no answer to a worker who has been wrongly denied disability benefits.⁷⁰ Critics of the decision make a similar point: although a disability beneficiary is somewhat better off financially than the typical welfare recipient, most disability beneficiaries would still face an extremely difficult financial situation after an erroneous cutoff.⁷¹ The majority's own wording shows what is being contested, since it said only that the disabled worker's need is "likely" to be less than a welfare recipient's.⁷² The conclusion therefore rests on a judgment about how typical recipients fare, and that judgment is the point on which the dissent and the critics disagree.

The second step separates welfare from disability in the decided cases, and the fifth step explains why a promise of back payment does not close the inquiry.

Two mistakes recur when this factor is applied. The first is to treat the finding of a property interest as fixing its weight, when that finding only establishes that due process applies, and Mathews itself held a protected interest to require less process than Goldberg did.⁷⁴ The second is to treat a promise of full back payment as the end of the inquiry, when Goldberg shows that the harm of a wrongful cutoff falls during the wait. A need-based benefit, no independent resources, and a long wait for review are the facts the first factor weights most heavily, while a benefit paid regardless of income, available fallback resources, and full retroactive relief point toward Mathews. Neither result decides the case alone, because the weight of the private interest is carried into the second factor, which asks how likely the existing procedures are to produce error and what an added safeguard would be worth.

Risk of Erroneous Deprivation

The second Mathews factor shifts from the size of the claimant's loss to the reliability of the decision that causes it. It directs a court to weigh the risk of an erroneous deprivation through the procedures the government already uses, together with the probable value, if any, of additional or substitute safeguards.⁷⁶ The Court described the inquiry as one into "the fairness and reliability of the existing pretermination procedures, and the probable value, if any, of additional procedural safeguards."⁷⁷ Two questions are folded into that phrase. The first asks how often the current procedures are likely to produce a wrong answer, and the second asks how much that rate would fall if the claimant's requested procedure were added. The second question carries the weight because procedures exist to prevent mistaken deprivations, so a safeguard earns its place only to the extent it would change the accuracy of decisions. An added procedure that would rarely alter a result protects almost no one while costing the government money and time, and the test asks whether the value of the procedures a claimant demands beyond those already provided justifies the cost of adding them.⁷⁸

The value of any safeguard depends on what kind of question the decisionmaker must answer, which is why the Court called the nature of the relevant inquiry central to the evaluation of any administrative process.⁷⁹ Continued eligibility for disability benefits required a medical assessment of the worker's physical or mental condition, which the Court found a more sharply focused and easily documented decision than the typical determination of welfare entitlement.⁸⁰ A welfare decision draws on a wide variety of information, and issues of witness credibility and veracity often are critical. That is why Goldberg treated written submissions as a wholly unsatisfactory basis for decision in that setting.⁸¹ Goldberg itself shows the pattern: whether a mother lives with a man able to support her children depends on a caseworker's account that she disputes, and a decisionmaker reading a file cannot test that account, while oral presentation and questioning can.⁸² The same procedure, an oral hearing, therefore has high value in one program and little in the other, and the second factor exists to measure that difference.

The disability side of the pair works for a reason that can be traced through its steps. A report from a physician who personally examined the claimant is itself the evidence, so a decisionmaker can read it, compare it with other reports, and weigh it without watching its author. The Court expected the decision to turn, in most cases, on "routine, standard, and unbiased medical reports by physician specialists" about a person the doctors had personally examined, and it concluded that the potential value of an evidentiary hearing, or even of oral presentation to the decisionmaker, was substantially less than in Goldberg.⁸³ A disagreement between doctors can be aired by exchanging documents, because each side's reasoning is on the page. A dispute over whether a witness is telling the truth cannot be aired that way, because the page shows neither demeanor nor the answer to a follow-up question. Credibility is therefore the feature that moves an inquiry from the paper side of the line to the oral side.⁸⁴

The Court added a limit on how the risk is measured. It conceded that credibility and veracity may be a factor in the ultimate disability assessment in some cases, and it still held that procedural due process rules are shaped by "the risk of error inherent in the truthfinding process as applied to the generality of cases, not the rare exceptions."⁸⁵ The reason is practical. Procedures are prescribed for a whole program in advance, so a rule that gave a hearing to any claimant who could point to something unusual would turn every program into a case-by-case negotiation and defeat the purpose of setting procedures ahead of time. A claimant who wants more process must therefore show something about the program's typical decisions, and a feature of the claimant's own file does not carry the argument alone.

The first half of the factor counts everything the agency already does, and Mathews is a case about a program with many safeguards in place. The agency gave the recipient's representative full access to all the information it relied on, and before the cutoff it informed the recipient of its tentative assessment and its reasons and supplied a summary of the evidence it considered most relevant. The recipient was then free to submit additional evidence or arguments, which let the recipient challenge the accuracy of the file and the correctness of the tentative conclusions directly.⁸⁶ These procedures let the recipient "mold" the argument to the precise issues the decisionmaker regards as crucial, and that is the function Goldberg assigned to oral presentation.⁸⁷ Each such safeguard lowers the baseline risk of error, so it also lowers what an additional procedure can add. The written channel was realistic for these claimants, who could get help completing forms from the local Social Security office, and the main evidence came from treating physicians presumably able to write clearly.⁸⁸

Reversal-rate statistics raise a different question, which is what counts as evidence that the existing procedures are failing. The amici in Mathews pointed to the significant reversal rate for appealed cases as clear proof of inadequacy, and the figures varied with the base chosen, from 58.6% for appealed reconsideration decisions to an overall rate of only 3.3%.⁸⁹ The Court answered that "[b]are statistics rarely provide a satisfactory measure of the fairness of a decisionmaking process."⁹⁰ Its reason was specific to the program: the review system ran on an open-file basis, so a recipient could always submit new evidence, and fresh medical examinations were held in roughly 30% to 40% of appealed cases.⁹¹ A reversal after new evidence arrives shows that the system responds to later information, and it does not show that the first decision was wrong on the record then before the agency. The Court therefore treated reversal rates as relevant but certainly not controlling.⁹²

The factor reaches beyond hearing formats, because a standard of proof is itself a procedural safeguard. It fixes how much evidence the government needs to prevail, and so it decides how the risk of a mistaken result is divided between the government and the person affected. In Santosky v. Kramer a state could terminate parental rights on a mere preponderance of the evidence, and the Court weighed the Mathews factors against that standard. The parents' interest is large, the risk of error under a preponderance standard is substantial, and the state's interest in the lower standard is slight, so due process requires at least clear and convincing evidence. A preponderance standard tolerates error against either side about equally, which is acceptable in a money dispute where a mistaken loss can be repaid. It is a poor fit where the loss is the permanent severing of a family, the kind of harm that the first factor treats as the heaviest and that no later payment can correct.

The medical-report premise is contested, and the contest shows what the factor can and cannot do. The District Court had reasoned that disability determinations may involve subjective judgments based on conflicting medical and nonmedical evidence, and critics have called the Court's assumption that such decisions routinely rest on unbiased medical reports dubious.⁹³ The Court's answer was that professional disagreement among doctors differs from a contest over credibility, since the former can be laid out in writing and the latter needs a decisionmaker who can watch and question the witness. The factor also measures only accuracy, so the participatory value of being heard has no direct place in it, a limit that critics say makes the test too narrow.⁹⁴ The burden on a claimant follows from the factor's focus on marginal gains: the claimant must identify the specific additional procedure sought and show that its absence is a real defect in the agency's current process, because each procedural right is analyzed separately and a court may choose an intermediate procedure over the all-or-nothing choice between a trial-type hearing and the agency's own process.⁹⁵

The fourth step carries the doctrine's limit and the second step carries its main distinction, so those are where a fact pattern usually turns. The usual exam error is to answer the second factor by saying that a hearing would always be more accurate, when the factor asks how much accuracy this specific procedure adds beyond what the agency already does. A second error is to rely on a single unusual feature of the claimant's file, or on a raw reversal rate, as though either settled the risk of error for the program. Facts showing a decision that rests on routine written reports, advance disclosure of the evidence, and a usable chance to respond in writing point toward Mathews. Facts showing a decision that turns on a disputed witness account, and a claimant who cannot effectively use paper, point toward Goldberg.

The Government's Interest

The third Mathews factor puts the government's side on the scale. It directs a court to consider "the Government's interest, including the function involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail."⁹⁷ Two ideas sit inside that phrase. The function involved asks what the agency is doing and why its procedures are built as they are, and the burdens ask what the specific safeguard the claimant demands would cost in money, staff, and delay. As with the second factor, the question concerns the increment: the court compares the process the agency already provides with the added process the claimant wants, and only the cost of the addition belongs in the balance.⁹⁸

Cost belongs in a constitutional balance because agency budgets are finite. Money spent on hearing officers, transcripts, and staff time is money a program cannot pay out as benefits, so every procedural requirement is financed in some form by the people the program serves. Mathews put the point in these terms: the cost of protecting those whom the preliminary process has identified as likely to be found undeserving "may in the end come out of the pockets of the deserving," since the resources available for any social welfare program are not unlimited.⁹⁹ The Court was equally careful about the limit of the argument. Financial cost alone is not a controlling weight, but the government's interest in conserving scarce fiscal and administrative resources must be weighed, and at some point the benefit of an added safeguard may be outweighed by its cost.¹⁰⁰ Cost therefore enters as one weight set against the added assurance of accuracy and fairness, and the court asks whether the value of the extra procedure justifies what it would cost.¹⁰¹

The burden the Court identified in Mathews had two parts. The most visible was the incremental cost of the larger number of hearings, together with the expense of paying benefits to ineligible recipients while the hearings were pending.¹⁰² The second part follows from how payments would work under the claimant's rule. Because full benefits would continue until the hearing, the Court expected most claimants to take up an option that kept the money flowing, and the government's theoretical right to recover undeserved benefits would not, as a practical matter, offset the added outlay.¹⁰³ The Court admitted that no one could predict the size of the increase and that the parties' estimates varied widely, so it relied on experience with the constitutionalizing of government procedures, which suggested that the added cost in money and administrative burden would not be insubstantial.¹⁰⁴

The function the agency performs is the other half of the factor, and it counsels against copying court procedure automatically. The judicial model of an evidentiary hearing is neither required nor even the most effective method of decision in every setting, the Court said, because agencies differ in origin and function from courts and cannot simply borrow their procedures wholesale.¹⁰⁵ It added that substantial weight must be given to the good-faith judgments of the officials whom Congress charged with running social welfare programs, especially where the procedures give the claimant an effective way to assert the claim before any action and assure an evidentiary hearing and judicial review before a denial becomes final.¹⁰⁶ Walters v. National Association of Radiation Survivors shows how far the function argument can reach. The Court there upheld a $10 limit on the fee veterans could pay an attorney or agent for help with a benefits claim, even though Goldberg had counted counsel among the elements of a hearing, because without the cap claimants could freely retain lawyers, and legal representation would undermine the simple, non-adversarial process Congress preferred.¹⁰⁷

The weight the Court gave this factor changed between Goldberg and Mathews, and that change helps explain why similar benefit terminations produced different outcomes. New York argued in Goldberg that summary adjudication protects the public treasury by stopping payments as soon as ineligibility seems likely, and that because most terminations go unchallenged, it also saves the cost of many hearings.¹⁰⁸ The Court answered that these interests were not overriding in the welfare context. A prior hearing would cost more, and benefits paid to ineligible recipients pending the hearing probably could not be recovered because those recipients were likely to be judgment-proof, meaning too poor to repay, but the state could contain the expense through prompt hearings and skillful use of personnel and facilities.¹⁰⁹ Mathews gave the same interest much more weight. Its analysis concentrated almost exclusively on economic efficiency, and although it repeated that cost alone is not controlling, it raised the interest in conserving scarce resources well above the level it had reached in Goldberg.¹¹⁰

What intervened was experience. After Goldberg, participants in many agency programs claimed a deprivation of property and demanded pre-deprivation oral hearings, lower courts often agreed, and agencies complained that they lacked the resources to comply.¹¹¹ In the welfare program itself, demands for oral hearings rose significantly once hearings had to precede the cutoff, and the incentive is easy to trace. A claimant who is unlikely to have assets or earning power from which the government could recover overpayments loses nothing by asking, and even a claimant who loses the hearing keeps roughly a year of benefits that will have been spent by then.¹¹² The added costs fell into two groups: the expense of hiring many lawyers to serve as administrative judges, and the unrecoverable benefits paid to claimants who demanded a hearing and were then found ineligible.¹¹³ Goldberg had expected the state to contain such costs through prompt hearings and efficient staffing, and the Mathews Court gave weight to the warning of Justice Black, the lone dissenter in Goldberg, that added procedure could harm beneficiaries by diverting scarce public resources from benefits to administrative costs.¹¹⁴

The factor does not run in only one direction, because the government's interest extends beyond saving money. In Goldberg the public interest pointed two ways: summary decisions saved money, but the legislature had also decided that eligible people should receive assistance, so wrongly ending an eligible recipient's aid defeated the program's own purpose.¹¹⁵ The Court said that the same governmental interests that counsel providing welfare also counsel its uninterrupted provision to those eligible, and it treated the state's own post-termination hearing in the Home Relief program as cogent evidence that the state recognizes the primacy of the public interest in correct eligibility determinations.¹¹⁶ The mechanism is straightforward. An agency exists to pay the right people the right amounts, so an erroneous termination is a failure of the agency's own function, and a safeguard that lowers the error rate advances an interest the government itself holds. A court applying the third factor should therefore ask which governmental interests the added procedure would serve and which it would burden.

Convenience carries the least weight where the stakes are highest and a shortcut would bar the very question that decides the case. In Stanley v. Illinois the Court considered the argument that unmarried fathers are so seldom fit that the state need not undergo the administrative inconvenience of an inquiry in any case.¹¹⁷ The Court accepted that prompt and efficacious procedures serve a proper state interest, and then held that "the Constitution recognizes higher values than speed and efficiency."¹¹⁸ Procedure by presumption is always cheaper and easier than individualized determination, the Court observed, yet it forecloses the determinative issues of competence and care.¹¹⁹ The reasoning has a clear structure. The state's claim to custody of the children depends on the father's unfitness, so the presumption saves money by assuming the one fact the state must prove, and where the father turns out to be fit, the state's interest in caring for his children is de minimis.¹²⁰ Stanley was decided before the three factors, but it shows how the government's interest behaves at the heavy end of the range: the weight of an efficiency argument shrinks as the stake grows and as the shortcut's risk of error rises.

The opposite pole appears when the government acts against a whole category at once and cannot wait. In Village of Orland Park v. Pritzker a village, two of its residents, and a local business challenged the Illinois governor's pandemic executive orders, which restricted business activity and movement during the height of the pandemic, as violations of procedural due process.¹²¹ A federal district court applied Mathews and concluded that the private interests could not overcome the minimal risk of decisional error and the state's overwhelming interest in protecting public health and safety during an ongoing pandemic.¹²² The reasoning joins the third factor to the second. Because everyone in the same broad category was treated alike, an error as to one would be an error as to all, so individualized hearings would not have reduced mistaken deprivations.¹²³ On the burden side, those hearings would have overwhelmed the administrative system and hamstrung the state's ability to act quickly against a spreading disease.¹²⁴ The logic resembles the Londoner and Bi-Metallic line drawn earlier in this chapter, and the separate question of when urgency permits the government to postpone a hearing until after the deprivation is taken up in the timing section below.¹²⁵

The usual exam error is to treat cost as decisive, answering that any added hearing is too expensive, or to treat it as irrelevant to a constitutional right. A second error is to assume that the government's interest always opposes the claimant's safeguard, when the state's own interest in correct decisions often supports it. Facts that point toward the government include a large volume of claims likely to demand the procedure, benefits that continue and cannot be recovered while a hearing is pending, a program deliberately designed to be simple and non-adversarial, and class-wide emergency action. Facts that point toward the claimant include a modest and containable added burden, a stake such as subsistence or family integrity, and an efficiency argument that works by presuming the decisive fact. Like the first two factors, this one is carried into the overall balance and does not decide the case by itself.