A Republic, If You Can Keep It

by Neil Gorsuch

  But here, too, the separation of powers has come under some strain, with power gravitating away from the more deliberative and neutral processes of the judiciary toward the more energetic and political ones of the executive. As Oil States illustrates, executive employees today often work under the nom de plume of “administrative judges” and claim the power to decide cases and controversies in proceedings that bear at least a passing resemblance to judicial trials and appeals. Nor do these administrative proceedings resolve only disputes over public rights. Often enough they claim the power to strip individuals even of their vested private rights. And often enough, as in Oil States, the administrative judges charged with deciding these cases are accountable to political bosses with agendas of their own; gone in these cases is the right to a neutral judge and a jury of your peers.

  Even when it comes to the disputes that do remain within the judicial branch, executive agencies today enjoy considerable power. In relatively recent years, courts have decided to defer to agencies when it comes to the job of interpreting the law, effectively ceding to the executive a task previously thought a core judicial responsibility and competence. Perhaps the most well-known of these deference doctrines is the Chevron doctrine. Chevron says that if a statutory term is ambiguous, a judge must accept any reasonable gloss on the law the agency can supply—even if the judge is convinced it’s not the best reading of the statute’s terms. The related Auer doctrine requires judges to do much the same for an agency’s interpretation of its own regulations. So now, not only do executive agencies get to write the regulations, not only may they enforce them too, but they are even allowed to resolve any ambiguities that later emerge in favor of their preferred policy outcomes.

  In a short and recent span, as Justice Clarence Thomas has noted, courts have taken these doctrines and run with them. They’ve deferred to an agency’s interpretation of a different agency’s regulations. They’ve deferred to agency interpretations that were inconsistent with the agency’s previous interpretations of the same statute or regulation. They’ve deferred to agency interpretations advanced for the first time in litigation. They’ve deferred to agency interpretations when it comes to criminal sentencing. Under the Brand X doctrine, the Supreme Court has even said that courts must sometimes overrule their preexisting judicial interpretations of the law when an executive agency wants a different result.

  What’s the upshot of affording the executive so great a role in judicial functions? In the first place, it can make it exceedingly difficult for ordinary people to know what the law is (and even where to look for it). Under our deference doctrines, it’s not enough anymore to look to the statute books and the decisions of courts interpreting them. You also have to worry that a completely contrary and binding rule lies buried in the appendix to an agency’s guidance manual—maybe the third manual issued in as many years. Even then, you still have to predict accurately every reasonable possibility of what the law could be in the future. For now an agency can disagree with a court’s precedent as a matter of policy and demand its overruling. In a world like that, should we be concerned (as Hamilton was) about “subjecting…men to punishment for things which, when they were done, were breaches of no law,” something that “ha[s] been, in all ages, [a] favorite and most formidable instrument[] of tyranny”?

  Nor are these the only potential costs to consider. The founders insulated the judiciary from political pressures to ensure that judges would apply existing laws equally to everyone regardless of their present popularity. As Hamilton put it, this kind of independence was needed “to guard the Constitution and the rights of individuals from the effects of those ill humors, which the arts of designing men, or the influence of particular conjunctures, sometimes disseminate among the people themselves, and which, though they speedily give place to better information, and more deliberate reflection, have a tendency, in the meantime, to occasion dangerous innovations in the government, and serious oppressions of the minor party in the community.” Remember, too, that when it comes to adjudicating cases and controversies, we’re talking about assigning liability based on past actions that cannot be undone. It’s not like the job of passing new legislation that usually applies only prospectively, with fair notice given to everyone about the need to conform their conduct to the new rule. In adjudicating cases and controversies, the decisionmaker knows exactly whose ox will be gored by his decisions and the victim has no chance to alter his conduct to conform to the judgment. So when independent judges and juries give way to political decisionmakers in deciding cases and controversies over past facts, should we worry about the consequences for unpopular persons and causes? And what happens when the best reading of the law a neutral judge can muster gives way under our deference doctrines to an inferior gloss produced by administrative judges who are handpicked for each case by a political boss?

  Consider the case of Yellowbear v. Lampert. Andrew Yellowbear, in prison for murdering his daughter, is the kind of person who needs to be kept in special housing because his crime is so heinous other prisoners are likely to harm him. Given his unpopularity, it’s perhaps unsurprising that, when he sought to access the prison’s existing sweat lodge so he could exercise his Native American faith, executive officials refused. In response, my old court (the Tenth Circuit) had little difficulty applying a federal statute prohibiting religious discrimination to hold that Mr. Yellowbear was entitled to some access to the sweat lodge like other prisoners. Likewise, my court didn’t hesitate to apply the same set of statutes to protect a Muslim prisoner who required halal food, or to shield a family-run business from federal mandates that would have forced family members to violate the dictates of their faith. None of these outcomes would have been possible if the original executive decision hadn’t been reviewable by an independent judge. Can it be any surprise a Wall Street Journal study has found that, over a five-year period, the agency it studied won more than 90 percent of its cases before its own administrative judges but fewer than 70 percent of the cases it brought before federal judges?

  Finally, consider the loss of civic involvement. The judiciary isn’t just made up of life-appointed judges who have considerably more freedom to depart from majoritarian norms and political pressures than their counterparts in the other branches. It’s also composed of juries. In this way, the judiciary might be said to be simultaneously the least and the most democratic of the branches. Yet like judges, juries are notoriously independent. They do not stand for election and they answer to no one. So while they are of the community, they do not have to answer to the community. It’s the very independence and common sense of juries, their ability and willingness to cut through dogma and doctrine and technical legalese, that the founders and so many generations before and since have found so valuable. Of course, litigants in court may opt out of a jury trial, but they may also demand it, something they cannot do in administrative tribunals.

  Nor is the loss only the litigants’. Participation in juries is the closest many people ever get to participating in their government. Studies suggest that serving on a jury may lead to increased electoral engagement, especially among those who have not previously participated in elections. Studies suggest, too, that participation in the jury system increases significantly participants’ evaluation of our legal system. All of these civic goods disappear when cases are moved from judicial courtrooms to administrative conference rooms.

  * * *

  —

  THE CONCERNS I’VE DISCUSSED here are nothing new. As long ago as 1933, the American Bar Association created a Special Committee on Administrative Law. Soon the committee voiced concerns that the cascade of new agencies was commingling executive, legislative, and judicial functions in ways that could prove problematic. Eventually, Congress adopted the Administrative Procedure Act as a step toward addressing these concerns. The APA created two ways for agencies to act—rulemaking and adjudication—and created two modes of proceeding for each—form
al and informal. Unsurprisingly, formal proceedings offer considerably more procedural protections.

  Generally speaking, in informal rulemakings an agency suggests a new rule, the public comments on that proposal, and then the agency makes adjustments accordingly before issuing a final rule—or deciding not to issue a final rule at all. When the agency does issue a final rule, that rule can and does often look different from the original proposal; to be permissible, a final rule only needs to be a “logical outgrowth” of the proposed one. Formal rulemaking, by contrast, requires an agency official to preside over a verbal or written proceeding featuring cross-examination and forbids that official from engaging in ex parte communications. The agency also bears the burden of proving that the rule is supported “on consideration of the whole record…and supported by…substantial evidence”—something the agency may not need to prove in an informal rulemaking.

  The story is much the same when it comes to adjudication. In informal adjudication, the requirements are relatively few. Proceedings are presided over by officials who are often subject to oversight by others in the agency (just as in Oil States). These officials may even perform other duties within the agency in addition to their “judicial” responsibilities—including prosecuting and investigating other cases. By contrast, formal adjudications afford more protection for those involved. Parties may present their case by verbal or written evidence, conduct cross-examination, and submit rebuttal evidence. Parties are also entitled to a hearing where an administrative law judge (or the agency head) will preside. Administrative law judges do not enjoy the same independence as federal judges, but unlike some decisionmakers in more informal adjudications, these administrative law judges generally cannot have their pay cut without approval from a separate entity called the Merit Systems Protection Board.

  When Congress adopted the APA, most everyone expected that formal proceedings would be quite common. That understanding held true for a while, as formal devices initially “dominated the administrative law landscape,” in the words of the Administrative Conference of the United States. But today formal proceedings are a vanishing breed; formal rulemakings in particular, as Professor Aaron Nielson has explained, “ha[ve] been effectively exiled from administrative law.” Nor can it come as a surprise either that agencies have pushed back against formal processes given the many advantages informal processes hold for them—or that courts have acquiesced in these agency efforts given our deference doctrines.

  Agencies not only enjoy a great deal of leeway today when choosing whether to proceed informally or formally, they also have considerable discretion when choosing whether to proceed by rulemaking or adjudication. In a case called Chenery II, the Supreme Court held that an agency can sometimes even create a brand-new rule in an adjudication and apply that new rule retroactively to the parties before it (who, of course, lack any notice that might permit them to adjust their conduct to conform to the new regulatory demand). To be sure, this retroactive-rulemaking-by-adjudication has rested uneasily in our due process jurisprudence and has proven a frequent target of criticism. Justice Robert Jackson—President Franklin Roosevelt’s attorney general and himself an architect of the administrative state—penned a forceful dissent in Chenery II. He argued that the Court was upholding an “assertion of power to govern the matter without law,” something he thought would (in the words of Justice Benjamin Cardozo) reduce “[l]aw as a guide to conduct…to the level of mere futility.”

  In fairness to the Chenery II majority, it didn’t think the retroactivity problem would prove significant in many cases because it didn’t expect agencies to make a lot of new law through adjudication. Indeed, the majority wrote that agencies should use rulemaking procedures “as much as possible.” But if that was the Court’s expectation, reality has fallen far short. Agencies today routinely announce new rules while adjudicating individual cases. In fact, some agencies almost never engage in rulemaking and create virtually all of their new rules through adjudication. It’s a result that may be at least a little ironic considering that Congress created some of these agencies precisely so that they could formulate clear ex ante rules for people to follow. But it’s a result that cannot come as much of a surprise. For, again, what rational agency would choose to shoulder the burden of providing notice and allowing public comment from the whole of the people on a proposed rule when it could avoid that burden by simply announcing the same nationwide rule in the course of an informal adjudication of a discrete case against a single private party?

  If you’re beginning to wonder whether the APA’s procedural protections designed to safeguard individual liberty on the front end may have fallen short of initial expectations, you might also wonder about the judicial review procedures the APA sought to provide on the back end. During the APA’s drafting process, some suggested limiting judicial review of an agency’s legal interpretations (at least sometimes) to the task of asking merely “whether the administrative construction is a permissible one.” But in the end Congress chose a more robust formulation, directing that a “reviewing court shall decide all relevant questions of law, [and] interpret constitutional and statutory provisions.” That language seems to suggest that courts must decide questions of law for themselves without deference to agency interpretations. Yet, as we have seen, modern courts may have nearly re-created for themselves this discarded standard of review through Chevron and related deference doctrines.

  Nor is the problem limited to questions of law. Generally, the APA allows courts to strike down an agency’s factual findings in adjudication only if they are unsupported by “substantial evidence.” But in my time as a judge, I’ve seen at least some cases that might make you wonder if even that deferential standard has been watered down. Consider Mathis v. Shulkin, a case that came to the Supreme Court during my first term. There, lower courts seemingly created out of whole cloth a presumption that examiners at the Department of Veterans Affairs are competent to render expert medical opinions against veterans seeking compensation for service-related disabilities. So now substantial evidence can sometimes be effectively guaranteed by mere presumption. Or consider Compass Environmental, Inc. v. Occupational Safety & Health Review Commission, where the Tenth Circuit upheld an agency’s sanction against a private company even though it presented no evidence that the company violated industry standards, as the law required; instead, the agency pointed only to alleged violations of the company’s internal policies without seeking to show that they reflected (rather than exceeded) industry standards. Here, substantial evidence came from, really, no evidence at all.

  But even putting all this aside and overlooking how the APA might have failed to realize its initial promise, you might wonder whether it was ever intended to be the final word. Those who enacted the APA in 1946 were well aware that, in the words of Justice Jackson, “[e]xperience may reveal defects.” And both our world and our administrative state have changed massively since 1946. The federal civilian workforce outside the Department of Defense has grown by more than 70 percent, and the length of the Code of Federal Regulations has increased perhaps sevenfold. Not only are there far more agency rules, but they can bear far more significant impact. Today, a single federal rule can disrupt entire segments of our national life and change radically the way hundreds of millions of people work and live. Yet we still live under pretty much the same administrative law structure as we did in 1946, at least those portions of it that haven’t been effectively brushed away in practice. If the men and women who helped draft the APA were acutely aware of the challenges agencies could pose in their time to the separation of powers and the liberties our constitutional design was meant to serve, you might wonder whether we may safely ignore these same concerns in our own time.

  * * *

  —

  TO BE SURE, the practical benefits that flow from enforcing the separation of powers in an individual case can sometimes seem hard to see—maybe especially when compared
(like apples vs. oranges) against the systemic efficiencies that fusing powers in executive agencies can provide. So what if a Caring Hearts, an Oil States, a De Niz Robles, or a Mathis suffers an injustice? That may be regrettable, but accommodating their individual concerns would be costly. What matters more, some might claim, is justice-in-gross; government must be allowed to work efficiently. But the founders saw virtue in the separation of powers and a government that is deliberately deliberate. Where others see inefficiency in the separation of powers, they saw fair notice; protection for the inherent value of every individual person, including especially dissenting voices; democratic accountability; and the rule of law as administered by independent judges and juries. By transferring more and more power from the legislature and judiciary to the executive, we alter piece by piece the framers’ work and risk the underlying values it was designed to serve. Like a tower in the game of Jenga, pull out this block or that one and the tower may seem largely unaffected, especially if you do it with a bit of finesse—and the lawyers who come up with justifications for the blending of powers have plenty of that. But keep pulling out blocks, and eventually what started out as a strong and stable tower will begin to teeter.

  Madison and the founders held the view that “[n]o political truth is” more important to “liberty” than the separation of powers because the “accumulation of” power “in the same hands” inevitably leads to “tyranny.” We might ask ourselves today: Do we still agree? And if we do, how would we grade ourselves in bearing witness to that truth?

  Alexis de Tocqueville was one of the keenest students of our new republic. He offered us this insight, which I leave with you tonight:

  After having thus successively taken each member of the community in its powerful grasp, and fashioned them at will, the supreme power then extends its arm over the whole community. It covers the surface of society with a network of small complicated rules, minute and uniform, through which the most original minds and the most energetic characters cannot penetrate, to rise above the crowd. The will of man is not shattered, but softened, bent, and guided: men are seldom forced by it to act, but they are constantly restrained from acting: such a power does not destroy, but it prevents existence; it does not tyrannize, but it compresses, enervates, extinguishes, and stupefies a people, till each nation is reduced to be nothing better than a flock of timid and industrious animals, of which the government is the shepherd.