Editor’s Note: This is the first of three essays in a symposium on the executive removal cases, Trump v. Slaughter and Trump v. Cook. The following essays will be published on Tuesday and Wednesday.
Chief Justice John Roberts’s twentieth year on the Supreme Court demonstrates the contradictions of his career. There is Chief Justice Roberts, the institutionalist, who strategically guides the Supreme Court by not deciding more than is necessary to decide. Then there is Chief Justice Roberts, the institutionalist, who strategically guides the Supreme Court by deciding more than is necessary to decide. Which Roberts do we get in each case? This tale of two chiefs depends entirely on whether Roberts determines the decision will lead to the best of times or the worst of times. If the Chief fears rough roads ahead, less will be decided. If the Chief sees smooth sailing, more will be decided. No more, no less.
This tale can be told in two cases decided simultaneously on the final day of the Supreme Court’s term, each involving President Trump and Chief Justice Roberts. First, in Trump v. Slaughter, Roberts overturned a nine-decade-old precedent that protected “independent” agencies, allowing Trump to fire a commissioner. Second, in Trump v. Cook, Roberts reaffirmed the independence of the Federal Reserve, and blocked the president’s removal of the board governor. Each case represented what could only be called judicial maximalism. The Court reached out to issue sweeping rulings on issues that did not need to be decided. This duo of decisions from the chief justice is confounding, and conflicts with every fiber of Roberts’s purported institutionalism and minimalism over the course of two decades.
This ninety-year-long saga over presidential removal powers began during the New Deal with Humphrey’s Executor v. United States (1935). President Franklin Delano Roosevelt fired William Humphrey, a commissioner of the Federal Trade Commission. After Humphrey died, the executor of his estate continued to seek backpay. The Supreme Court, in a huge blow to Roosevelt, and presidential power more broadly, ruled that Congress had the power to make commissioners independent from the president, such that they could only be fired for certain causes of misconduct.
Humphrey’s Executor was the law of the land for generations. Congress routinely invoked that precedent to protect agencies from presidential control. And for nearly a century, the Supreme Court never even questioned Humphrey’s Executor—that was, until Chief Justice Roberts wrote two majority opinions a decade apart. First, in Free Enterprise Fund v. PCAOB (2010), Chief Justice Roberts ruled that the president must be allowed to fire the head of an obscure accounting board. At the time, it was unclear how broad this decision would sweep. The next shoe would drop ten years later. Seila Law LLC v. Consumer Financial Protection Bureau (2020) held that the president must be allowed to fire the single director of the important financial agency. Neither case purported to disturb Humphrey’s Executor, which concerned commissioners on multi-member boards, even though Roberts had great expectations for that precedent.
With the benefit of hindsight, it is easy enough to retell the Chief Justice’s two-decade “long game” as a coherent narrative, but the actual chronology was never a guarantee. Had Hillary Clinton won the 2016 election and filled one or more Supreme Court vacancies, the 5-4 vote in Seila Law almost certainly would have gone the other way. Indeed, even if Trump made his three appointments to the Court, and Kamala Harris prevailed in 2020, an FTC Commissioner would not have been fired, and there likely would not have been a square opportunity to overrule Humphrey’s Executor. Yet, all the pieces providentially fell into place. Trump prevailed in 2016, made three appointments to the Supreme Court, and fired Commissioner Louis Slaughter.
Friends don’t let friends be the chief justice’s fifth vote.
Thus, the stage was set to overrule Humphrey’s Executor. Still, there were many offramps that a restraint-minded chief justice could have taken. Slaughter demanded an unprecedented remedy: Unlike William Humphrey, who only sought backpay, he sued Trump seeking reinstatement. There were unresolved questions about whether the president could be sued for firing a commissioner, whether a court could reinstate a member once she is removed, and whether a court could order the president to recognize an officer he fired. The Court could have simply held that the courts could not afford Slaughter any remedy against the president, other than the award of backpay. Such a ruling would have been effectively the same as a ruling that the president had an absolute removal power. Where there is no remedy, the existence of the right becomes a theoretical question. Indeed, Chief Justice John Marshall chose exactly this path in the landmark case of Marbury v. Madison (1803). But Roberts saw no problem reaching the merits. Even then, there was another offramp.
The Court could also have drawn a distinction between the FTC of 1935 and the FTC of 2025. The present-day agency has far more powers than its Progressive Era antecedent. The Court could have left the husk of Humphrey’s Executor in place, perhaps leaving open the possibility that Congress could pare back the muscular powers of today’s agencies. Yet the chief justice wiped the slate clean. He wrote, “If anything more is left of Humphrey’s, we overrule it.” While William Humphrey died of natural causes (to eliminate any doubts, I obtained a copy of his death certificate), the case that bore his name met a far more unfortunate fate. Slaughter was, well, slaughtered.
To be perfectly clear, I agree that Humphrey’s Executor was wrongly decided. Justice Scalia’s iconic dissent in Morrison v. Olson (1988) laid the foundation for overruling that errant precedent. Any jurist who subscribes to Scalia’s form of originalism should have joined the majority opinion in Slaughter without reservation. The votes from Justices Clarence Thomas, Samuel Alito, Neil Gorsuch, Brett Kavanaugh, and Amy Coney Barrett were entirely to be expected. The problem, of course, is that the chief justice has never pretended to be an originalist, and routinely grounds his decision in the political realities of the day. And there were countless non-originalist grounds to retain Humphrey’s Executor.
It is useful to contrast Slaughter with one of the few recent landmark cases that the chief justice did not join: Dobbs v. Jackson Women’s Health Org. (2022). In that case, Roberts refused to overrule Roe v. Wade (1973). Instead, he chose an off-ramp that none of the parties requested. Roberts rewrote Roe to allow states to ban abortion up to fifteen weeks. Why? The chief justice warned that overruling the abortion precedent would cause a “serious jolt to the legal system.” The political consequences, he feared, were too great, so he salvaged Roe in name only. But why then overrule Humphrey’s Executor? As Justice Sonia Sotomayor observed in dissent, “the Humphrey’s rule is long established and has been easily applied for decades.” The same could not be said for Roe. Why should Roberts decide more than was necessary to resolve Slaughter’s case? Here, the Chief Justice saw no collateral consequences for overruling Humphrey’s Executor, even as he (rightly) predicted fallout from overruling Roe. Politics, and not law, explains the conflicting votes.
Slaughter, at least, was part of the plan. Trump v. Cook, however, did not fit into the chief justice’s “long game.” Under the Federal Reserve Act, a governor could only be removed for “cause.” In 2025, President Trump fired Lisa Cook, a governor of the Federal Reserve, for cause. The president alleged that Cook engaged in misconduct by listing a vacation home as her primary residence on a mortgage application. President Trump never challenged the independence of the Federal Reserve Board. Indeed, despite numerous threats, Trump never fired Jerome Powell, the chairman of the Federal Reserve that he had appointed in his first term. Cook, like Slaughter, sued the president, seeking to block her termination. Unlike Slaughter, however, the Supreme Court ruled on an interim basis that Cook could not be removed from office. This preliminary ruling in October 2025 was the first signal that Slaughter and Cook would face a different fate at the high court.
By a 5-4 vote, the Court ruled for Cook. Once again, Chief Justice Roberts wrote the majority opinion, joined by Justice Kavanaugh and the three progressives. Roberts held that the Federal Reserve Act required the president to provide Cook with notice and an opportunity to be heard before the termination. But the chief justice did not stop there. He reached out to decide that the Federal Reserve was an independent agency, and the governors could not be fired at will. Again, none of the parties disputed the independence of the Federal Reserve, so this ruling was unnecessary.
There is a longstanding principle in Supreme Court jurisprudence known as “constitutional avoidance.” As Chief Justice Roberts observed only four years ago in Dobbs, “Our established practice is instead not to ‘formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.’” The chief justice refused to revisit Roe, and fashioned a narrow constitutional ruling that only decided the present case: whether Mississippi’s fifteen-week abortion ban was constitutional. But in Cook, the chief justice went far beyond the question presented, and settled issues that no party disputed.
Moreover, it is difficult to reconcile the decisions rendered by the chief justice just moments apart. Justice Thomas observed that the Court “endorse[d] a contradiction: ‘the Constitution vests the whole executive power in the President alone,” in Slaughter, but in Cook the Federal Reserve “Board can exercise executive power ‘independen[t] from Presidential control.’” Justice Barrett agreed that “the most significant problem is that the Court decides this issue at all—not to mention the many others covered in its opinion.” In a shot at the chief, Barrett charged that “a modest approach would have been appropriate,” but the Court instead “chooses to go big.” Worse still, the Court “sets precedent on a series of important issues, with implications that extend well beyond this case.” A recent column in the New York Times relayed a law professor joke: “I hope the guy who wrote Slaughter doesn’t run into the guy who wrote Cook. That would be a really rough conversation.”
The contradiction is not just how the chief justice decided the cases in opposite directions. The far greater contradiction is why the chief justice decided the cases in opposite directions. Why did the Court reach out to decide the status of the Federal Reserve, on the interim docket no less? The chief justice was quite direct: “We see no reason to leave the public in limbo, or to sow doubt as to the status of one of our Nation’s (and the world’s) most important financial institutions.” Roberts said reaching the merits here was “ultimately a matter of prudence, upon which reasonable minds can (and often do) disagree.” But the most relevant mind to disagree is Roberts himself. Justice Thomas responded sharply: “If the Court prefers an independent Federal Reserve Board, then its issue is not with the President but with the Constitution.” Like with Dobbs, the chief wanted to prevent a “serious jolt” to our legal system. But to avoid that jolt, the chief had to avoid the principle of constitutional avoidance.
Like Oliver Twist, the chief justice can keep asking for one more vote—and his colleagues should just say no.
Indeed, the chief justice disregarded yet another longstanding tradition. The Court’s practice is that as soon as an opinion is announced, print copies are handed to reporters and a PDF is posted online. But on the final day of the term, both Slaughter and Cook were distributed to reporters simultaneously, bound by rubber bands, as both PDFs were posted simultaneously. Why did the chief justice deviate from practice? If there was a thirty-minute gap between Slaughter and Cook, there may have been instability in financial trading. The better course for the markets, the chief decided, was to make both decisions appear at the same time. John Roberts is not an institutionalist. He is an illusionist who sometimes conjures legal principles, misdirects from longstanding traditions, vanishes legal impediments, gets applause by avoiding controversies, all based on nothing more than his amateur political judgments.
Back in June 2022, I was the first scholar to write about the so-called 3-3-3 Court, though others have since picked that term up. The thinking went that the Court had three coalitions of three justices: the progressives (Sotomayor, Kagan, and Jackson), the conservatives (Thomas, Alito, and Gorsuch), and the center-right members (Roberts, Kavanaugh, and Barrett). That tripartite alignment largely ran its course as Justices Kavanaugh and Barrett have settled in. Now, in the cases that count, there are two predominant splits. There is the 6-3 super-conservative Court, with the chief justice squarely in control. Or, there is the Roberts +1 Court: the three progressives sign onto whatever the chief justice is serving, and they are joined by either Justice Barrett or Kavanaugh.
In Cook, Justice Kavanaugh was Roberts’s +1. Kavanaugh was also worried about the potential consequences for monetary policy from weakening the Federal Reserve’s independence. Still, Kavanaugh suggested that Trump could remove Cook if he followed certain processes. In another landmark case, Trump v. Barbara, Justice Barrett provided the +1 for birthright citizenship. Barrett did not write separately, so we do not know for sure what her thinking is. But she did not join Justice Kavanaugh’s concurrence, which found that Trump’s reading of the Citizenship Clause of the Fourteenth Amendment was correct, even though his order conflicted with statutory law. Both Cook and Barbara reached out to decide far more constitutional questions than were necessary. And those rulings only set precedents because Justices Kavanaugh and Barrett signed on in full.
My humble suggestion for the future: friends don’t let friends be the chief justice’s fifth vote. Whenever the chief justice writes a majority opinion that has only four votes, the other conservatives should concur only in judgment to deny his situational jurisprudence precedential authority. In Cook, even if Justice Kavanaugh agreed that the Federal Reserve must remain independent, he should have only joined parts of the chief’s opinion that did not reach out to settle novel questions. And in Barbara, Justice Barrett could have concluded that even if the best reading of the Fourteenth Amendment is to guarantee birthright citizenship, there is no reason to definitively set a constitutional rule, as the statute simply incorporates the constitutional standard. Both cases would have turned out exactly the same way, but no new risky precedents would have been set. Like Oliver Twist, the chief justice can keep asking for one more vote—and his colleagues should just say no.