Crystal Palace striker Jean-Philippe Mateta defends Millwall goalkeeper Liam Roberts, saying "too much emotion makes you do crazy things".

Wales bring in Joe Roberts and Aaron Wainwright as two changes for the Six Nations finale against England in Cardiff on Saturday.

Yesterday’s 5–4 Supreme Court decision requiring the United States Agency for International Development to start making payments that the Trump administration had frozen was immediately hailed as a signal of the justices’ discomfort with the administration’s efforts to feed “USAID into the wood chipper,” as Elon Musk colorfully put it. It was also said to suggest skepticism of President Donald Trump’s claim that he has the constitutional authority to impound federal dollars and ignore Congress’s spending commands.

Perhaps. But the optimism may be premature. The reprieve that the order offers is brief, the basis for the decision is narrow and procedural, and the eventual outcome remains uncertain.

In a dissenting opinion, four of the conservative justices said that the federal courts ought to play a highly circumscribed role in policing Trump’s efforts to dismantle agencies by preventing them from spending money. Although Chief Justice John Roberts and Justice Amy Coney Barrett sided with the liberal justices this week, that’s no guarantee of their vote when the case comes up in a different procedural posture, which it almost certainly will. The plaintiffs’ temporary victory could still curdle into defeat.

[Stephen I. Vladeck: The Supreme Court foreign-aid ruling is a bad sign for Trump]

The key to understanding how the justices will think about the case is the Administrative Procedure Act, an 80-year-old law that allows injured parties to sue federal agencies that act in an unlawful or arbitrary manner. The APA is generally solicitous of lawsuits against the government, and is said to have created a strong presumption in favor of judicial review.

The challengers—USAID contractors who haven’t been getting paid—brought their claim under the APA. That’s natural. Although there’s some confusion about the precise source of the command to stop paying out on existing contracts—an executive order? a now-withdrawn Office of Management and Budget memo? DOGE? a directive from the secretary of state?—there’s no question that a blunderbuss spending freeze has been instituted.

If that freeze is illegal or arbitrary—and there’s a good argument that it’s both—the APA empowers the courts to set it aside and, if necessary, to enjoin the federal government from freezing the funds. Seen that way, the case is a bog-standard challenge to unlawful agency action.

But the APA is limited in some important respects. Of particular relevance here, a plaintiff can’t seek “money damages” under the APA. So if a government employee runs a person over and he wants damages for his injuries, or a government agency breaches a contract with a business owner, those parties can’t bring an APA suit. Instead, they have to take their case to the Court of Federal Claims, a special court that handles claims of money damages against the federal government.

The four dissenting justices, in an angry opinion by Justice Samuel Alito, insist that that’s what the plaintiffs should have done. Sure, the plaintiffs say they’re challenging a general spending freeze. But what they’re really challenging is the refusal to pay out on their contracts. The “relief” that they seek, Alito wrote, “more closely resembles a compensatory money judgment rather than an order for specific relief that might have been available in equity.”

That’s one way to understand what the plaintiffs want. After all, they do want money. Plus, the courts are generally reluctant to entertain broad-brush challenges to agency policy, especially when an agency is accused of not doing something that it’s supposed to do. Otherwise, as the Court explained back in 2004, there’s a risk of “injecting the judge into day-to-day agency management” of the agency’s affairs. The courts don’t want to be in the business of micromanaging all of USAID’s contracts.

[Read: Trump tests the courts]

So which is it? Is the lawsuit best seen as an APA challenge to an illegal funding freeze? Or as a demand for money damages arising from specific contractual breaches that should go to the Court of Federal Claims?

That question has no intrinsically correct answer. It’s a matter of emphasis and judgment. A person’s preferred characterization may depend on their sense of just how aberrant and troubling the Trump administration’s actions are. The closer the case seems to a conventional breach-of-contract dispute, albeit at scale, the more appropriate sending it to the Court of Federal Claims may seem.

That parsimonious approach has all the virtues of judicial modesty. It also has all the vices.

There’s something deeply artificial about treating the case like an everyday spat over the terms of a contract for, say, military equipment. The funding freeze reflects a comprehensive, deliberate effort to destroy an agency that Congress established and President Trump dislikes. That freeze can be appropriately viewed as a discrete agency action that’s properly subject to APA review.

Contra Alito, just because the case is about money does not make it a case about money damages. The distinction may seem fine, but it’s got a long pedigree. “The fact that a judicial remedy may require one party to pay money to another,” the Supreme Court reasoned in 1988, “is not a sufficient reason to characterize the relief as ‘money damages.’”

Money damages, the Court explained, aim to redress an injury that’s already happened. They are meant to soothe past harms, not prevent them. Most APA suits, in contrast, are anticipatory. They allow courts to prevent agencies from harming plaintiffs in the first place. That’s what the plaintiffs are seeking here—not a financial remedy for a breach of contract, but an end to a funding freeze that causes them ongoing injury.

Moreover, conceiving of the case as a routine breach-of-contract dispute would have troubling consequences. If the case is forced into the Court of Federal Claims, the plaintiffs might eventually get a money judgment against the government, perhaps a hefty one, especially if they bring a class action. But the Court of Federal Claims likely won’t enter an injunction that ends the spending freeze. That’s not what it does.

And Trump won’t care that Congress will have to shell out cash down the line. His goal is more immediate and more destructive.

The outcome of this arcane jurisdictional dispute may thus effectively determine whether Trump has the power to impound federal funds and dismantle federal agencies. If he does, expect him to exercise that power again. And again. And again.

[Adam Serwer: Why Trump thanked John Roberts]

Right now, all we know for sure is that four conservative justices are okay with that outcome, whatever the damage to Congress’s power to control federal spending. The three liberal justices probably aren’t, whatever the risks of excessive judicial interference in government administration.

That leaves Roberts and Barrett.

We don’t know what they think. The Supreme Court’s very short opinion turned on the case’s very hurried procedural posture. Once the lower court enters a more durable order, the case will likely wing its way back to the justices, probably within weeks.

At that point, we’ll find out whether the Supreme Court intends to serve as a bulwark against a president who is hell-bent on asserting the unilateral power to control federal spending. If not, yesterday’s order may come to look like a momentary, ephemeral reprieve in Trump’s ongoing assault on Congress’s power of the purse.

The key to understanding this morning’s Supreme Court ruling unfreezing American foreign aid is that two different rulings are at issue here, and teasing apart those technicalities reveals a loss that is perhaps more significant for the Trump administration than is first apparent.

The two orders both come from U.S. District Court Judge Amir Ali. There’s his underlying temporary restraining order (TRO), which remains in effect (and which the government has neither tried to appeal nor sought emergency relief from), and then there’s his more specific order, which purported to enforce the TRO by obliging the government to pay somewhere from $1.5 billion to $2 billion of committed foreign-aid funds by February 26. It was that order that the government tried to appeal, and from which it sought emergency relief first in the D.C. Circuit Court and then in the Supreme Court. By issuing an “administrative stay” last Wednesday night, Chief Justice John Roberts temporarily absolved the government of its obligation to comply with that order—but not with the underlying TRO, which generally requires the government to spend money that Congress has appropriated for foreign-aid funding.

Against that backdrop, the Court’s ruling today is more than a little confusing. Let’s start with what’s clear: A 5–4 majority (with Chief Justice Roberts and Justice Amy Coney Barrett joining the three Democratic appointees) denied the government’s application to vacate Judge Ali’s enforcement order. The Court’s ruling contains only one meaningful sentence, and it is maddeningly opaque:

Given that the deadline in the challenged order has now passed, and in light of the ongoing preliminary injunction proceedings, the District Court should clarify what obligations the Government must fulfill to ensure compliance with the temporary restraining order, with due regard for the feasibility of any compliance timelines.

This sentence (or, perhaps, an earlier draft of it) provoked a fiery and more than a little hypocritical eight-page dissent from Justice Samuel Alito, joined in full by Justices Clarence Thomas, Neil Gorsuch, and Brett Kavanaugh. But before getting to the dissent, let me try to read a couple of tea leaves out of this cryptic but important passage.

[Adam Serwer: Why Trump thanked John Roberts]

First, I think it’s meaningful that the majority denied the government’s application rather than dismissing it as moot. In English, that is the majority signaling that the government likely still must comply with the “pay now” order—the second of the two—albeit not on the original timeline. If the majority thought that the “pay now” order was no longer live because the deadline had come and gone, then the proper disposition would have been to dismiss the application as moot, not to deny it. (Indeed, although there are good reasons to not rely upon dissents to figure out what the majority held, Alito’s dissent seems to reinforce this reading.) This may seem like a very thin reed, but it’s a distinction I can’t imagine was lost upon the justices. The majority (and, apparently, the dissent) seems to agree that the government remains under not just the general obligation of the original TRO but the specific obligation of the “pay now” order.

Second, the clause about the district court clarifying the obligations that the government must fulfill to comply with the TRO strikes me as an invitation to Judge Ali to do exactly that—to issue a more specific order that (1) identifies the particular spending commitments that he believes the government must honor to comply with the TRO and (2) gives the government at least a little more than 48 hours to do so. The upshot is that, even if the Trump administration doesn’t have to pay the money immediately, it will have to do so very soon. That’s small solace to the organizations and people who have already had their lives upended by the spending freeze, but it’s a bigger loss for the Trump administration than the text may suggest.

Third, the timing of the ruling is striking. The Court handed down the order right at 9 a.m. this morning—less than 12 hours after the end of President Donald Trump’s address to Congress last night. It is just about impossible to imagine that the ruling was still being finalized overnight (or that the chief justice was somehow influenced by his awkward moment with Trump). If not, then there appears to have been at least some choice on the Court’s part to hand down the ruling after the president’s speech and not before it at the close of business yesterday—perhaps to avoid the possibility of Trump attacking the justices while several of them were in the audience. I’ve written before about the problem of the Court timing its rulings—and how it underscores the extent to which the justices are, and ought to admit that they are, playing at least some politics even with what should be a straightforward procedure for releasing rulings when they’re ready. This at least seems like it might be another example.

And fourth, here’s that 5–4 lineup again. Back in January, I wrote about how this particular 5–4 alignment (the chief justice, Justice Barrett, and the three Democratic appointees) is starting to show up in cases “in which the Chief Justice’s elusive but not illusory institutional commitments, and Justice Barrett’s emerging independence, are separating them from the other Republican appointees. For a host of reasons that I suspect are obvious, we may see more such cases sooner rather than later.”

On one hand, it’s a bit alarming that Kavanaugh joined the dissent. On the other hand, for those hoping that the Court is going to be a bulwark against the (mounting) abuses of the Trump administration, it’s a cautiously optimistic sign that there may well be at least five votes to support lower-court rulings attempting to rein in those abuses.

In many ways, the dissent is far more illuminating than the majority’s order. As is unfortunately often the case with respect to Alito’s dissents from emergency applications, this one combines a remarkable amount of hypocrisy with statements that are either materially incorrect or, at the very least, misleading.

[Read: ‘Constitutional crisis’ is an understatement]

On page three of the ruling (page two of the dissent), for example, Alito writes that “the Government must apparently pay the $2 billion posthaste—not because the law requires it, but simply because a District Judge so ordered.” Of course, this completely misstates both the theory of the plaintiffs’ lawsuits and the gravamen of Judge Ali’s order. The whole point is that the law does require it—that Congress has mandated the spending and that the contractual obligations have been fulfilled. Indeed, Judge Ali’s “pay now” order is about work already completed for which the money was already due. If there is authority for the proposition that the government is not legally obliged to pay its bills, Alito doesn’t cite it. Yes, there may be separate questions about the courts’ power to compel the government, but that’s not the same thing as whether the “law requires” the government to pay its bills. Do the dissenters genuinely believe that the answer is no?

Alito also makes much out of the argument that sovereign immunity bars the claims against the government. But the Supreme Court has already held that relief under the Administrative Procedure Act can run to whether the government is obliged to pay expenditures to which the recipients are legally entitled. Alito asserts that actually ordering the government to pay those expenditures is something else entirely; suffice to say, I think that’s slicing the bologna pretty thin. His argument would have more force if Judge Ali’s “pay now” order was about funds for which the administrative processes haven’t fully run. But here, they have. And so it’s just a question of whether federal courts have the power to force the government to … enforce the law.

In that respect, contrast Alito’s analysis here with his dissenting 2023 opinion in United States v. Texas—in which he would have upheld an injunction by a single (judge-shopped) district judge that effectively dictated to the executive branch what its immigration-enforcement priorities must be. In explaining why the Biden administration should lose, he wrote:

Nothing in our precedents even remotely supports this grossly inflated conception of “executive Power,” which seriously infringes the “legislative Powers” that the Constitution grants to Congress. At issue here is Congress’s authority to control immigration, and “[t]his Court has repeatedly emphasized that ‘over no conceivable subject is the legislative power of Congress more complete than it is over’ the admission of aliens.” In the exercise of that power, Congress passed and President Clinton signed a law that commands the detention and removal of aliens who have been convicted of certain particularly dangerous crimes. The Secretary of Homeland Security, however, has instructed his agents to disobey this legislative command and instead follow a different policy that is more to his liking.

In 2023, Alito dismissed the view that courts could not push back against the president in such cases as a “radical theory.” In 2025, apparently, it’s correct. I wonder what’s changed?

Finally, Alito offers what I would euphemistically call a remarkable discussion of why the harm that the plaintiffs are suffering is insufficient to overcome the government’s case for a stay:

Any harm resulting from the failure to pay amounts that the law requires would have been diminished, if not eliminated, if the Court of Appeals had promptly decided the merits of the Government’s appeal, which it should not have dismissed. If we sent this case back to the Court of Appeals, it could still render a prompt decision.

In other words, the plaintiffs are being harmed not by the government’s refusal to pay them but by the D.C. Circuit’s refusal to exercise appellate jurisdiction over Judge Ali’s “pay now” order. I don’t even know what to say about this argument other than that, if that’s how irreparable harm worked, well, emergency relief (and the role of intermediate appellate courts) would look a heck of a lot different.

Alito closes by accusing the majority of imposing “a $2 billion penalty on American taxpayers.” This comes back to the central analytical flaw in the dissent: The “penalty” to which Alito is referring is the government’s underlying legal obligation to pay its debts. Debts aren’t a penalty; they are the literal cost of doing business. And if this is the approach that these four justices are going to take in all of the spending cases to come, that’s more than a little disheartening.

[Read: Trump tests the courts]

As for what comes next, well, I’m not entirely sure. We know that Judge Ali is scheduled to hold a preliminary injunction hearing tomorrow. It is very possible that before then (or shortly thereafter) he will reimpose some kind of “pay now” mandate that, with the hints from the Supreme Court majority, is a bit more specific and has a slightly longer timeline. Of course, the government could seek emergency relief from that order, too, but I take today’s ruling as a sign that, so long as Judge Ali follows the Court’s clues, at least five justices will be inclined to deny such relief. That doesn’t do anything immediately for the plaintiffs and other foreign-aid recipients who are continuing to suffer debilitating consequences. But it does suggest that, sometime soon, the government really is going to have to pay out at least some of the money at issue in these cases (and, as important, perhaps other funding cases too).

The broader takeaway, though, is that this is now the second ruling (the first was Dellinger) in which the Court has, in the same ruling, moved gingerly but at the same time denied the relief that the Trump administration was seeking. Two cases are, obviously, a small data set. But for those hoping that even this Supreme Court will stand up, at least in some respects, to the Trump administration, I think there’s a reason to see today’s ruling as a modestly positive sign in that direction.

Yes, the Court could do even more to push back in these cases. But the fact that Trump is already 0–2 on emergency applications is, I think, not an accident, and a result that may send a message to lower courts, whether deliberately or not, to keep doing what they’re doing.

This article was adapted from a post on Steve Vladeck's Substack, One First.

The exchange was so awkward, it should have been followed by the Curb Your Enthusiasm theme song. While President Donald Trump was shaking hands down the aisle, exiting the House chamber after his address last night, network cameras caught him as he turned to Chief Justice John Roberts, patted him on the back, and said, “Thank you again. Thank you again. I won’t forget.” Roberts, whose back was to the camera, then headed for the exit.

We can’t know precisely what the president meant, but Trump does have a lot to thank Roberts for. After all, the chief justice and the other conservatives on the Supreme Court helped rewrite the Fourteenth Amendment, completely gutting the ban on insurrectionists holding office in order to allow Trump to run for president again following his attempt to seize power by force after the 2020 election. Then Roberts and the other conservative justices manifested an absurd, imperial grant of presidential immunity, with no textual basis in the Constitution, to shield Trump from criminal prosecution, and in so doing set the stage for a despotic second term during which Trump will try to ignore court efforts to impose limits on his power.

In fairness, Roberts has not been as supplicant as some of his colleagues. He has been willing to occasionally refuse Trump demands; this morning, Roberts and Justice Amy Coney Barrett sided with the three Democratic appointees in declining to overturn an order from a lower court to unfreeze $2 billion in USAID funding. The underlying dispute here is more high-stakes than it might sound; the Trump administration is publicly, though not yet in court, claiming the right to usurp Congress’s constitutional authority over spending, which, if sustained, would bring the country closer to dictatorship. The dissent was so unhinged that one might conclude that there are only five votes on the Supreme Court to uphold the basic constitutional structure. But even though Roberts went against the president on this occasion, he is unlikely to be a reliable check on Trump’s lawlessness. Trump may well have more to thank Roberts for in the future.

[Read: Trump tests the courts]

Any casual observer of the Supreme Court can see what many prestigious constitutional lawyers can’t, which is that the conservative justices are frequently accomplices to Trump’s assault on democracy—a flag signaling support of the January 6 insurrection flew outside Justice Samuel Alito’s house. (Alito, vital specimen of right-wing masculine energy that he is, blamed his wife.) That sort of open partisanship is a bit inconvenient for Roberts, however, who during his confirmation hearing famously compared justices to umpires calling balls and strikes in a baseball game. A more appropriate sports analogy for how Roberts and his right-wing comrades approach cases appeared a few months later, when several referees in the Italian soccer league were implicated in fixing matches for top teams during the 2006 Calciopoli scandal.

Trump has threatened to criminally prosecute those who criticize the Court, declaring that they should be “put in jail,” consistent with the right-wing belief that the right to free speech allows people to say only what conservatives want them to say. But as is often the case, no critic of the Court could implicate the conservative majority’s partisanship as effectively as Trump’s own behavior.

In his own way, the president agrees with the liberal critique that the Roberts Court is a partisan institution, with a majority that will generally do what he wants. He just believes that this is both good and exactly how it should be. Perhaps the only person who is still in the dark about what the Supreme Court has become is Roberts himself.

The Football Association says Millwall keeper Liam Roberts' three-match ban for a challenge on Crystal Palace striker Jean-Philippe Mateta is "clearly insufficient".

Watch as Millwall goalkeeper Liam Roberts is sent off after a 'dangerous' challenge on Crystal Palace's Jean-Phillipe Mateta which lead to the Eagles' top scorer being stretchered off early in the first half of their FA Cup fifth round tie.

This article was featured in the One Story to Read Today newsletter. Sign up for it here.

When colleges began announcing the makeup of their incoming freshman classes last year—the first admissions cycle since the Supreme Court outlawed affirmative action—there seemed to have been some kind of mistake. The Court’s ruling in Students for Fair Admissions v. Harvard had been almost universally expected to produce big changes. Elite universities warned of a return to diversity levels not seen since the early 1960s, when their college classes had only a handful of Black students.

And yet, when the numbers came in, several of the most selective colleges in the country reported the opposite results. Yale, Dartmouth, Northwestern, the University of Virginia, Wesleyan, Williams, and Bowdoin all ended up enrolling more Black or Latino students, or both. Princeton and Duke appear to have kept their demographics basically stable.

These surprising results raise two competing possibilities. One is that top universities can preserve racial diversity without taking race directly into account in admissions. The other, favored by the coalition that successfully challenged affirmative action in court, is that at least some of the schools are simply ignoring the Supreme Court’s ruling—that they are, in other words, cheating. Finding out the truth will likely require litigation that could drag on for years. Although affirmative action was outlawed in 2023, the war over the use of race in college admissions is far from over.

History strongly suggested that the end of affirmative action would be disastrous for diversity in elite higher education. (Most American colleges accept most applicants and therefore didn’t use affirmative action in the first place.) In the states that had already banned the practice for public universities, the share of Black and Latino students enrolled at the most selective flagship campuses immediately plummeted. At UC Berkeley, for example, underrepresented minorities made up 22 percent of the freshman class in 1997. In 1998, after California passed its affirmative-action ban, that number fell to 11 percent. Many of these schools eventually saw a partial rebound, but not enough to restore their previous demographic balance.

Something similar happened at many selective schools in the aftermath of the Supreme Court’s 2023 ruling. At Harvard and MIT, for example, Black enrollment fell by more than 28 and 60 percent, respectively, compared with the average of the two years prior to the Court’s decision. But quite a few institutions defied expectations. At Yale, Black and Latino enrollment increased, while Asian American enrollment fell by 16 percent compared with recent years. Northwestern similarly saw its Black and Latino populations increase by more than 10 percent, while Asian and white enrollment declined. (In Students for Fair Admissions, the Court had found that Harvard’s race-conscious admissions policies discriminated against Asian applicants.)

[Rose Horowitch: The perverse consequences of tuition-free medical school]

Figuring out how this happened is not easy. Universities have always been cagey about how they choose to admit students; the secrecy ostensibly prevents students from trying to game the process. (It also prevents embarrassment: When details have come out, usually through litigation, they have typically not been flattering.) Now, with elite-college admissions under more scrutiny than usual, they’re even more wary of saying too much. When I asked universities for further details about their response to the ruling, Dartmouth, Bowdoin, and Williams declined to comment, Yale and Northwestern pointed me toward their vague public statements, and a Princeton spokesperson said that “now race plays no role in admissions decisions.” Duke did not reply to requests for comment.

The information gap has led outside observers to piece together theories with the data they do have. One possibility is that universities such as Yale and Princeton are taking advantage of some wiggle room in the Supreme Court’s ruling. “Nothing in this opinion should be construed as prohibiting universities from considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise,” Chief Justice John Roberts wrote in his majority opinion. This seemed to provide an indirect way to preserve race-consciousness in admissions. “It’s still legal to pursue diversity,” Sonja Starr, a law professor at the University of Chicago, told me. Her research shows that 43 of the 65 top-ranked universities have essay prompts that ask applicants about their identity or adversity; eight made the addition after the Court’s decision.

Another theory is that universities have figured out how to indirectly preserve racial diversity by focusing on socioeconomic status rather than race itself. In 2024, Yale’s admissions department began factoring in data from the Opportunity Atlas, a project run by researchers at Harvard and the U.S. Census Bureau that measures the upward mobility of children who grew up in a given neighborhood. It also increased recruitment and outreach in low-income areas. Similarly, Princeton announced that it would aim to increase its share of students who are eligible for financial aid. “In the changed legal environment, the University’s greatest opportunity to attract diverse talent pertains to socioeconomic diversity,” a committee designed to review race-neutral admissions policies at the college wrote.

Some evidence supports the “socioeconomics, not race” theory. Dartmouth announced that it had increased its share of low-income students eligible for federal Pell grants by five percentage points. Yale has said that last year’s incoming freshman class would have the greatest share of first-generation and low-income students in the university’s history. Richard Kahlenberg, a longtime proponent of class-based affirmative action who testified on behalf of the plaintiffs challenging Harvard’s admissions policies, told me that, by increasing economic diversity as a proxy for race, elite colleges have brought in the low-income students of color whom purely race-based affirmative action had long allowed them to overlook. (In recent years, almost three-quarters of the Black and Hispanic students at Harvard came from the wealthiest 20 percent of those populations nationally.) “While universities had been claiming that racial preferences were the only way they could create racial diversity, in fact, if we assume good faith on the part of the universities, they have found ways to achieve racial diversity without racial preferences,” Kahlenberg said.

[Richard Kahlenberg: The affirmative action that colleges really need]

If we assume good faith—that’s a big caveat. Not everyone is prepared to give universities the benefit of the doubt. Edward Blum, the president of Students for Fair Admissions, the plaintiff in the case that ended affirmative action, has already accused Yale, Princeton, and Duke of cheating. And Richard Sander, a law professor at UCLA and a critic of affirmative action, said that if a university’s Black enrollment numbers are still above 10 percent, “then I don’t think there’s any question that they’re engaged in illegal use of preferences.”

The skeptics’ best evidence is the fact that the universities accused of breaking the rules haven’t fully explained how they got their results. Yale, for example, has touted its use of the Opportunity Atlas, but hasn’t shared how it factors information from the tool into admissions decisions. Before the Court’s ruling, a Black student was four times more likely to get into Harvard than a white student with comparable scores, and a Latino applicant about twice as likely.

To keep numbers stable, race-neutral alternatives would have to provide a comparable boost. According to simulations presented to the Supreme Court, universities would have to eliminate legacy and donor preferences and slightly lower their average SAT scores to keep demographics constant without considering race. (In oral arguments, one lawyer compared the change in test scores to moving “from Harvard to Dartmouth.”) With minor exceptions, selective universities have given no indication that they’ve made either of those changes.

Even the data that exist are not totally straightforward to interpret. Some universities have reported an uptick in the percentage of students who chose not to report their race in their application. If that group skews white and Asian, as research suggests it does, then the reported share of Black and Latino students could be artificially inflated. And then there’s the question of how many students choose to accept a university’s offer of admission, which schools have little control over. Wesleyan, for example, accepted fewer Black applicants than it had in prior years, Michael Roth, the university’s president, told me. But a larger share chose to matriculate—possibly, Roth said, because even-more-selective schools had rejected them. The University of Virginia similarly had an unusually high yield among Black students, according to Greg Roberts, its dean of admissions. He couldn’t tell whether this was thanks to the school’s outreach efforts or just a coincidence. “I think what we’re doing is important, but to the extent it will consistently impact what the class looks like, I have no idea,” he told me. (Both Roth and Roberts, the only university administrators who agreed to be interviewed for this article, assured me that their institutions had obeyed the Court’s ruling.)

None of those alternative explanations is likely to sway the people who are convinced the schools cheated. With Donald Trump back in office, colleges that don’t see a meaningful uptick in Asian enrollees will likely face civil-rights investigations, says Josh Dunn, a law professor at the University of Tennessee at Knoxville. “If everything ends up looking exactly like it did prior to SFFA,” he told me, then the courts will “probably think that the schools were not trying to comply in good faith.”

Blum, the head of Students for Fair Admissions, has already threatened to sue Yale, Princeton, and Duke if they don’t release numbers proving to his satisfaction that they have complied with the law. (Blum declined to be interviewed for this article.) A new lawsuit could force universities to turn over their admissions data, which should reveal what’s really going on. It could also invite the Court to weigh in on new questions, including the legality of race-neutral alternatives to affirmative action that are adopted with racial diversity in mind. A resolution to any of these issues would take years to arrive.

In many ways, the endless fight over affirmative action is a proxy for the battle over what uber-selective universities are for. Institutions such as Harvard and Yale have long been torn between conflicting aims: on the one hand, creating the next generation of leaders out of the most accomplished applicants; on the other, serving as engines of social mobility for promising students with few opportunities. It will take much more than the legal demise of affirmative action to put that debate to rest.

Anna Roberts, 11, is one of just 39 people in the UK with the CLN3 strain of Batten Disease.

David Roberts spent six years protecting the late superstar as she turned the world.