The Answer Depends on What You Call the Question: Review of Questions of Law, Fact, and Mixed Questions

A buyer closes on a house Friday and walks into the dining room Saturday to find the antique chandelier gone. The seller unscrewed it from the ceiling Thursday night. During a showing, she’d mentioned it had been in her family for generations and she planned to take it with her. The buyer said nothing. Neither mentioned it in the contract. The trial court applies a fixture test — manner of attachment, adaptation to the property, objective evidence of intent — and rules for the seller. The buyer appeals, arguing the court got the answer wrong. Everyone agrees on the legal test. Everyone agrees on what happened. The buyer just argues those facts mean the chandelier is a fixture. What does the appellate court do? That answer depends entirely on what kind of question the appellate court decides it’s answering — whether the dispute is about what the law means, what the facts are, or how to apply one to the other. And that classification, not the answer, often decides the case.

This is the first merits-review episode after the administrative record page established what the court gets to look at. The next question is what the court is allowed to do with it. APA § 706 tells courts to “decide all relevant questions of law,” and separately to review agency factual findings under substantial evidence (in certain proceedings) or arbitrary-and-capricious review. Particular statutes supply their own instructions — Social Security § 405(g) says the Commissioner’s factual findings “if supported by substantial evidence shall be conclusive.” That word conclusive is strong. Congress didn’t tell the district judge to consider every fact and decide which account she prefers. Universal Camera (1951) made clear substantial evidence review isn’t a rubber stamp — courts must consider the whole record, including evidence that detracts from the agency’s conclusion — but a judge still can’t replace a supported factual determination with the conclusion she’d have reached herself. Congress gave the agency the initial responsibility to find facts. Federal courts review those findings within the limits Congress set.

The distinction between law and fact matters because the court’s authority differs on each side of it. A legal question gets the court’s independent judgment. A factual question gets review for support in the record. What makes administrative law interesting — and what makes this episode necessary — is that enormous swaths of agency decisions sit on neither side. They’re mixed questions: is someone an “employee” under a labor statute? Does an immigrant’s family face “exceptional and extremely unusual hardship”? Do particular events amount to “persecution”? In each case, the decisionmaker applies a general legal standard to specific facts. The classification of such questions as primarily legal or primarily factual decides how much authority the reviewing court has — and sometimes decides whether the court can review the question at all.

The Supreme Court addressed mixed questions directly in U.S. Bank v. Village at Lakeridge (2018), a bankruptcy case about whether a creditor who’d bought a claim from someone romantically involved with a corporate insider should be treated as a non-statutory insider. The Ninth Circuit used a two-part test: the relationship had to be comparable to the statute’s enumerated insider exceptions, and the transaction had to be negotiated at less than arm’s length. The bankruptcy court found the purchase had been conducted at arm’s length. Should the appellate court independently reconsider that application, or review it only for clear error? Justice Kagan, writing for a unanimous Court, framed the answer this way: “the standard of review for a mixed question all depends on whether answering it entails primarily legal or factual work.” That’s not much of a test — it describes the problem rather than solving it. But the institutional point underneath it is real. If resolving the dispute requires clarifying what a legal standard means for future cases, the appellate court is well-positioned. If it requires examining one complicated transaction and the people involved, the court that heard the evidence has the advantage. In Lakeridge, the legal test was taken as a given; the remaining work was evaluating whether these particular parties had actually dealt with each other at arm’s length. Clear error review applied. But Lakeridge was appellate review of a bankruptcy court, not an administrative agency, so its holding doesn’t automatically control standards for reviewing agencies.

The outer limit on characterization comes from Pullman-Standard v. Swint (1982). A Title VII case where Black employees challenged a seniority system; the district court found no discriminatory intent behind its adoption. The Court of Appeals treated discriminatory intent as an “ultimate fact” it could reconsider independently. The Supreme Court rejected that move. Federal Rule of Civil Procedure 52, the Court said, “does not divide facts into categories. In particular, it does not divide findings of fact into those that deal with ultimate and those that deal with subsidiary facts.” The fact that a finding determines the outcome doesn’t transform it into a question of law. Intent was factual, and the appellate court had to respect the trial court’s fact-finding role. Even when legal error distorts the fact-finding, the ordinary solution is to send the case back for proper findings — not to have the appellate court make them itself. That’s a meaningful limit. The importance of an answer doesn’t change what sort of question it is.

Wilkinson v. Garland (2024) applied these principles in a setting where the classification decided whether the court could review the question at all. Situ Kamu Wilkinson had come from Trinidad and Tobago on a tourist visa and stayed after it expired. Years later he faced removal, and sought cancellation based partly on hardship to his U.S. citizen son, who was seven and suffered from severe asthma requiring hospitalization several times a year. The statutory standard required showing “exceptional and extremely unusual hardship” to a qualifying family member. The immigration judge credited Wilkinson’s testimony but concluded the combined hardship didn’t meet the threshold. The Third Circuit held it lacked jurisdiction — treating the hardship determination as an unreviewable discretionary judgment under the immigration statute’s jurisdiction-stripping provisions. The Supreme Court drew a distinction. Applying the statutory hardship standard to established facts is a mixed question of law and fact, and under this particular immigration provision, mixed questions fell within the exception preserving review of “questions of law.” The federal court can’t reconsider findings about the child’s medical condition or Wilkinson’s support. It reviews whether those established facts satisfy the statutory standard. Wilkinson won the right to judicial review, but because the hardship inquiry requires primarily factual work, the Court said that review would be “highly deferential.” The question is legal enough to get into court, but factual enough to receive deference. That separates two things that had been treated as one: whether a court has authority to review an issue, and how independently it can decide the issue once it gets there.

Urias-Orellana v. Bondi (March 2026) made the distinction even sharper. Douglas Urias-Orellana and his family had come from El Salvador in 2021 seeking asylum. He testified that a hitman had targeted him and his family had repeatedly moved to escape threats. The immigration judge explicitly found his testimony credible but concluded his experiences didn’t establish past persecution or a well-founded fear of future persecution. The BIA affirmed; the First Circuit upheld under substantial evidence review. He argued the federal court should independently decide whether the established facts amounted to persecution. A unanimous Court, in an opinion by Justice Jackson, held substantial evidence review applied to the agency’s persecution determination — including its application of the statutory standard to established facts. The applicants weren’t challenging the legal definition of persecution; they accepted it. They were disputing its application. And on that question, the governing immigration statute required substantial evidence review. “Whether a given issue is treated as a question of law exempt from the INA’s jurisdiction stripping provisions,” Jackson wrote, “tells us nothing about the type of review the court must afford to that issue under other provisions of the statute.” Substantial evidence review doesn’t allow reversal unless the record compels a conclusion contrary to the agency’s — a demanding standard. It isn’t enough for the federal court to think persecution is the better conclusion. The record has to compel it. Even a person whose testimony was fully credited can fail to obtain reversal because the court can’t say the statute’s demanding standard has been met.

This is where the Loper Bright question lands. Loper Bright says courts must independently interpret statutes. But if courts characterize a dispute as applying a legal standard to facts rather than interpreting the statute, they can still defer to the agency. The Alito/Roberts/Thomas dissent in Wilkinson pressed this point: the majority had extended the “questions of law” exception too far into what was essentially a factual determination. Formally, a court might say it independently determines what “persecution” means. But if it defers to the agency on whether particular conduct qualifies, the agency retains considerable practical authority over how that legal standard operates. Courts can continue deferring to agency judgments with substantial legal content, even though Chevron is gone. Urias-Orellana itself rested on a specific statutory provision governing substantial evidence review — the Court wasn’t free to simply choose deference. But in cases where the boundary between interpretation and application is less clearly determined, how a court characterizes the issue can make the entire difference. Loper Bright may have shifted the battle from whether courts defer to agencies on statutory interpretation to what courts are willing to characterize as statutory interpretation in the first place.

That struggle over the interpretation/application boundary isn’t new. NLRB v. Hearst Publications (1944) involved what the Court insisted on calling “newsboys” — a label the Court itself called improper, since they were generally grown men supporting families, working established street corners sometimes for years, buying papers at a price the publisher set and reselling them at a price the publisher also set. The publishers returned unsold copies for credit but had to pay for papers that were lost. The publisher controlled distribution areas, influenced how many papers each worker received, and had district managers who supervised the operation and assigned selling spots. The NLRB found the regular full-time workers were “employees” under the NLRA. The publishers argued the term should be confined to common-law categories. The Supreme Court rejected that interpretation — it interpreted the statute in light of its context and purposes — and then considered the board’s application of that understanding to these particular relationships. The board’s determination that specified persons are employees, the Court said, “is to be accepted if it has warrant in the record and a reasonable basis in law.” Judicial responsibility for statutory interpretation was separate from the board’s role in applying a broad term to particular facts. That’s how the Loper Bright majority later characterized Hearst. The dissent questioned how cleanly those steps could be separated in cases like this.

Congress didn’t leave Hearst alone. In 1947 the Taft-Hartley amendments expressly excluded independent contractors from the definition of “employee,” shifting the governing framework to common law agency principles. In NLRB v. United Insurance Company of America (1968), the Court considered insurance “debit agents” who collected premiums, tried to prevent policies from lapsing, and sold new insurance. The board found they were trained by company supervisors, ordinarily sold only the company’s policies, were subject to company procedures, and received benefits under company vacation, insurance, and pension arrangements. The board concluded the agents were employees. The Seventh Circuit disagreed and refused to enforce the board’s order. The Supreme Court reversed. Hearst’s broad statutory approach no longer governed, but applying the common law test still involved evaluating the actual relationship, and the board had to choose between fairly conflicting views supported by the record. The Supreme Court acknowledged applying common law agency principles didn’t require specialized administrative expertise judges lacked. Deference wasn’t a reward for technical expertise — it reflected the board’s assigned responsibilities and the governing limits on judicial review. Congress could change the legal test. That didn’t mean a court could disregard the board’s supported application of that test.

Back to Kathleen, the warehouse supervisor from Season 2 whose Social Security disability denial has reached federal court. The ALJ wrote that her statements about the intensity, persistence, and limiting effects of her symptoms were “not entirely consistent with the medical evidence and other evidence in the record” — translated: he thought her symptoms didn’t limit what she could do as much as she said. He concluded she had the residual functional capacity to perform sedentary work. She argues three things: the ALJ misunderstood what sedentary work meant (legal error); the medical evidence doesn’t support his finding about how long she can sit (factual challenge); and even accepting the ALJ’s factual findings about her limitations, those limitations don’t satisfy the legal standard for sedentary work (mixed question). A single decision can contain multiple kinds of error. An ALJ who misunderstands the legal standard might also make factual determinations affected by that mistake. The reviewing court has to identify the actual dispute and determine what the Social Security Act and governing precedent require. Sometimes the problem is what the law means. Sometimes it’s what happened. Sometimes it’s what happens when you put the two together. And after Loper Bright, courts may continue to defer to agencies by treating disputes as applications of law to facts rather than questions of legal interpretation. The decision about how to characterize the question can be as important as the answer itself. The decision is the decision before the decision.

Listen Now

Apple Podcasts | Listen on our site

What We Cover

  • The chandelier hypothetical that opens the episode and why the appellate court has to decide what kind of question it’s answering before it decides whether the trial court got it wrong
  • APA § 706’s instruction to courts to “decide all relevant questions of law” and its separate framework for reviewing agency factual findings
  • Social Security § 405(g) and the “conclusive” standard for the Commissioner’s supported factual findings
  • Universal Camera (1951) and why substantial evidence review requires looking at the whole record, including evidence that detracts from the agency’s conclusion
  • Why the court can’t replace a supported factual determination with the conclusion it would have reached independently
  • The three main categories (questions of law, questions of fact, mixed questions) and why mixed questions dominate agency decision-making
  • The examples that recur: Is someone an “employee”? Does a family face “exceptional and extremely unusual hardship”? Do events amount to “persecution”?
  • U.S. Bank v. Village at Lakeridge (2018) and Justice Kagan’s formulation that review standard “all depends on whether answering it entails primarily legal or factual work”
  • Why the Lakeridge framework describes the problem rather than solving it — and the institutional-competence logic underneath it
  • Pullman-Standard v. Swint (1982) and the outer limit on characterization: appellate courts can’t convert factual findings into “ultimate facts” subject to independent review
  • Why Rule 52 “does not divide facts into categories” and what that means for how legal error gets corrected
  • Wilkinson v. Garland (2024): why applying the “exceptional and extremely unusual hardship” standard to established facts is a mixed question that falls within the statute’s “questions of law” exception
  • The separation between whether a court has authority to review an issue and how independently it can decide the issue once it gets there
  • Urias-Orellana v. Bondi (March 2026): the unanimous ruling that substantial evidence review applies to the agency’s application of the persecution standard to credited testimony
  • The “record must compel” test and why even a fully credited applicant can lose because the court can’t say the statute’s demanding standard has been met
  • The Loper Bright workaround: how courts can continue deferring to agencies by characterizing disputes as applications rather than interpretations
  • Why the Alito/Roberts/Thomas dissent in Wilkinson pressed this exact point
  • NLRB v. Hearst Publications (1944) and the “newsboys” case: why Supreme Court statutory interpretation and NLRB application were treated as separable
  • The 1947 Taft-Hartley amendments and how Congress narrowed the legal rule after Hearst
  • NLRB v. United Insurance Company of America (1968): why deference wasn’t a reward for technical expertise but a function of assigned responsibility and governing limits on review
  • Kathleen’s case: three distinct kinds of challenge she can make to the ALJ’s denial, and why the characterization of each determines the court’s authority
  • Why “the decision is the decision before the decision”

Full Transcript

Gwen: Welcome to Administrative Remedies, because you can't fix what you don't understand. Brought to you in part by the University of Tulsa College of Law's Mabee Legal Information Center. I'm Gwendolyn Savitz, an associate professor here at TU.

Marc: And I'm Marc Roark, a professor at the College of Law. We'll be breaking down complex doctrines with real-life analogies and examples to demystify the world of administrative law for everyone trying to understand how government actually works.

Gwen: Agencies are the main way the federal government gets things done. It's not through Congress for reasons we'll be addressing over the course of the series. All right, Marc, let's say that house closes on Friday and the buyer gets the keys on Saturday, walks into the dining room and discovers that the chandelier is gone.

Marc: So the seller took it with them.

Gwen: Yes. Like Thursday night, she unscrews it from the ceiling. She patches the hole and she takes it to her new house.

Marc: Yeah, that's not going to go over well. Whenever the buyer comes into the house and finds that there is no chandelier in the dining room, they're probably going to raise a dispute: breach of contract.

Gwen: So is that because the chandelier came with the house?

Marc: It's not that simple. The question is whether it was a fixture or not. And personal property can become part of the real property and generally fixtures transfer with the land. So whether there's something that qualifies depends on the applicable law and circumstances and also the contract that was executed.

Gwen: OK, so here's a little more information. The chandelier was an antique and it belonged to the seller's grandmother. She had installed in the dining room five years earlier. It was wired into the ceiling, but she'd always intended to take it when she moved.

Marc: OK, I mean, so she had private expectations about how what the longevity of that chandelier was, but those aren't necessarily known. And so the question is, did she communicate that to the buyer when they signed the contract?

Gwen: What if during the showing, she told the buyer, well, the person showing the house, told the buyer that the chandelier had been in her family for generations and she planned to take it with her? The buyer doesn't say anything, but then when they sign the actual contract, neither side mentions the chandelier.

Marc: Okay, that does make the problem more interesting. That kind of creates a parol evidence issue. The chandelier was physically attached to the house, but the seller had told the buyer she intended to take it. Whether that matters and whether this buyer's silence matters depends on the governing contract and fixtures. And what I'll also say is when you look at the standard real estate contract, it makes it really easy to opt out of certain things that don't come with the house. That might actually not be a great set of facts for the seller who takes the antique chandelier with her.

Gwen: Suppose at this point that neither party is disputing any of these facts and assume the dispute is covered by a fixture test that looks at the manner of attachment and the adaptation to the property and objective evidence of intent. The trial court applies that test and decides that the seller could take the chandelier and then the buyer appeals.

Marc: OK, what is the basis for the buyer's appeal now?

Gwen: The buyer is saying that the trial court got the answer wrong. Let's say in this case that everybody agrees on the legal test. Everybody agrees what happened. And the buyer just argues that those facts mean the chandelier is a fixture.

Marc: OK, so that's an interesting problem. The appellate court might be deciding what the legal test for fixtures means, whether the test itself creates a certain meaning of those things. It might be reviewing how the trial judge evaluated the circumstance of the particular chandelier.

Gwen: If the facts are established and the law is established, why shouldn't the appellate court independently decide whether the chandelier meets the legal test?

Marc: Sometimes it can and maybe it should, but that's not automatic. You have only certain kinds of cases where the appellate court gets to weigh in and apply facts to the law. Depending on the jurisdiction and the issue, appellate courts may give different degrees of deference to a trial court's application of the law to the facts.

Gwen: Which can sound a little odd if everybody agrees on both of them.

Marc: Right, because the application can require judgment. I mean, how important is the fact that the chandelier was wired in the ceiling? What weight should we give the seller's statement about taking the chandelier? Does that statement affect the legal characterization of the property? And also, did you see the way they testified? And were you able to give weight to whether their body language gave any sort of credibility to their statements?

Gwen: So the appellate court has to decide what kind of a question it's answering before it decides whether the trial court got it wrong.

Marc: Yeah, and that characterization can affect how much freedom the appellate court has to reach a different answer. It isn't an arbitrary choice, but it does matter.

Gwen: Okay, so let's circle back to Kathleen.

Marc: Okay, our warehouse supervisor whose disability claim finally reaches federal court.

Gwen: Last week, we talked about what the federal judge gets to see, her administrative record. Now the judge has that record and is reviewing the administrative law judge's decision to deny benefits.

Marc: OK. And the question is what the judge is permitted to do with it now.

Gwen: Yes. Consider two possible findings. In one, the ALJ might have written: the claimant's statements concerning the intensity, persistence, and limiting effects of her symptoms are not entirely consistent with the medical evidence and other evidence in the record, or if we try and translate that. Looking at all the evidence, I don't think her symptoms limit what she can do as much as she says.

Marc: Okay, and the ALJ is deciding how much weight to give Kathleen's account of her symptoms compared with the other evidence that's available to it.

Gwen: Then the ALJ concludes that she has the residual functional capacity to perform sedentary work. She still has the ability to do sedentary work.

Marc: Which matters because the agency uses her remaining capacities, everything she's still capable of doing, together with her age, her education, and her work history to determine whether there are jobs that she still can do.

Gwen: So at this point, say she's making two arguments. Argument number one: the ALJ misunderstood what sedentary work meant. Argument number two: the medical evidence doesn't support his finding about how long she can sit.

Marc: Okay, these are different objections. The first concerns the legal requirement, and the second concerns the evidence supporting the factual finding.

Gwen: Let's add another variation, too. Suppose the ALJ correctly states the standard and identifies her limitations, but the dispute is whether those limitations satisfy the standard.

Marc: Okay, then it's not really a dispute about what the law says or what the facts are. It's about what happens when you put those two things together.

Gwen: Yes. These are different kinds of decisions, and Congress has given courts different instructions about questions of law versus questions of fact. Let's start with 706 of the APA.

Marc: Okay. The APA says here, to the extent necessary to decision and when presented, the reviewing court shall decide all relevant questions of law.

Gwen: So this gives courts the responsibility for determining the legal meaning. The same section separately addresses review of agency factual findings and includes substantial evidence review in certain proceedings.

Marc: Okay, so the statute distinguishes the court's responsibility to decide what the law means from its responsibility to review what the agency found.

Gwen: And particular statutes can supply their own instructions. Additionally, for Kathleen's Social Security case, Section 405(g) governs. Here's what that says.

Marc: The findings of the Commissioner of Social Security as to any fact if supported by substantial evidence shall be conclusive. Conclusive is a strong word. Congress didn't tell the district judge to consider every fact and decide which account she prefers.

Gwen: Right. Although substantial evidence review isn't a rubber stamp. In Universal Camera, which we've talked about before, the Supreme Court emphasized that reviewing courts must consider the record as a whole, including evidence that detracts from the agency's conclusion.

Marc: OK, but the judge still can't simply replace a supported factual determination with the conclusion she would have reached herself.

Gwen: Yes, and that's an important division of responsibility. Congress gave the agency the initial responsibility to find facts, while the federal courts review those findings within the limits that Congress established.

Marc: Which makes the distinction between law and fact consequential. If something is a legal question, the court may have a different responsibility than if it's reviewing an agency's factual determination.

Gwen: And we can also hit hard questions that don't neatly fall into either category. This became even more important after Loper Bright. That was the Supreme Court case that said courts must exercise independent judgment about the meaning of a statute. But that doesn't mean that courts independently reconsider every agency decision that involves a statute.

Marc: Because interpreting a legal standard and applying it to particular facts aren't necessarily the same thing.

Gwen: And deciding where that matters can be surprisingly difficult. The Loper Bright majority itself distinguished statutory interpretation from fact-bound applications of broad statutory terms. And the dissent questioned how cleanly those tasks could be separated.

Marc: So even after Chevron was overruled, we still have to decide whether a dispute is actually about legal interpretation or about applying law to facts.

Gwen: Yes. That brings us to mixed questions of law and fact. Agencies encounter them constantly. Is someone an employee under a labor statute? Does an immigrant's family face a legally sufficient level of hardship? Does particular conduct satisfy regulatory prohibition?

Marc: And in each case, the decision maker has to apply a general legal standard to very specific facts.

Gwen: Yeah. And the Supreme Court addressed this general problem in U.S. Bank v. Village at Lakeridge. This was a bankruptcy case from 2018.

Marc: Okay. Bankruptcy is kind of close to property. I'm intrigued.

Gwen: All right. Here we have a corporation. It's in bankruptcy and has a creditor whose claim was purchased by someone with a romantic relationship with one of the corporation's insiders. The question is whether that purchaser should be treated as a non-statutory insider.

Marc: Which matters because bankruptcy law treats insiders very differently from outsiders.

Gwen: Yeah. The Ninth Circuit here used a two-part test. They looked at the creditor's relationship with the debtor and said it had to be comparable to relationships covered by the bankruptcy code's enumerated insider exceptions. And they said that the transaction had to be negotiated at less than arm's length.

Marc: So the relationship mattered, but so did how the transaction actually occurred.

Gwen: Yes. And the bankruptcy court found that the purchase had been conducted at arm's length. Then the Supreme Court faced a standard of review question. Should the appellate court independently reconsider that application of the test or should it review it only for clear error?

Marc: And the Supreme Court took the Ninth Circuit's legal test as a given. It didn't decide whether that was the correct substantive definition of a non-statutory insider.

Gwen: Yes. So the court addressed the standard of review for the application, not the correctness of the underlying legal test. And because this was appellate review of a bankruptcy court, its holding doesn't automatically determine the standard for reviewing administrative agencies.

Marc: But the court's explanation of mixed questions is more broadly useful.

Gwen: Yes. Here's what Kagan wrote for a unanimous court.

Marc: In short, the standard of review for a mixed question all depends on whether answering it entails primarily legal or factual work. That's not really much of a test, if I'm being fair.

Gwen: Yeah, I know. It theoretically provides at least questions you can ask to analyze it.

Marc: Okay, and it tells the court to determine whether something is primarily legal or factual by deciding whether answering it requires primarily legal or factual work. That describes the problem rather than solving it. It's sort of like a self-fulfilling prophecy.

Gwen: Okay, but consider the institutional difference. If resolving the dispute requires clarifying what a legal standard means for future cases, an appellate court could be well-positioned to do that. If it requires examining one complicated transaction and the people involved in it, the court that heard the evidence probably has an advantage there.

Marc: In Lakeridge, the legal test was taken as a given. The remaining work was evaluating whether these particular parties had actually dealt with each other at arm's length.

Gwen: Yes. And so the Supreme Court held that clear error review was appropriate.

Marc: OK, that rationale is defensible. I mean, the difficult part is deciding whether a dispute requires elaborating a legal rule or evaluating the facts. Describe the question as, what does arm's length require? And it sounds legal. Describe it as, did these two people behave like strangers? And it sounds factual.

Gwen: Yes. Although, theoretically, parties can't change the governing standard by changing the wording.

Marc: Of course, but that question generally has both dimensions. That characterization can still affect how closely the appellate court examines the decision.

Gwen: And Lakeridge isn't an exceptionless formula. Constitutional mixed questions, for example, can still call for independent review for other reasons. Or a specific statute could also govern the standard for reviewing a particular issue.

Marc: So the framework is useful, but it doesn't give courts unlimited freedom to decide how deferential they want to be.

Gwen: All right, let's circle back to the chandelier. If the appellate court is deciding what the fixture test means, that sounds like legal work. But what if everyone agrees on the test and the dispute is about the particular circumstances?

Marc: Okay, then the court might be doing something closer to what the bankruptcy court did in Lakeridge. It's evaluating how an established test applies to a particular transaction. But I'd still want to know what the governing jurisdiction says about reviewing fixture determinations. Lakeridge doesn't decide that for us.

Gwen: So you wouldn't necessarily get the same standard of review just because the two cases involve applying law to facts?

Marc: No, but there's a more basic point about our chandelier. The seller's statement about taking it might be evidence of intent, but if the dispute were instead about whether that conversation changed the written agreement, we'd have a different legal issue and the physical object hasn't changed. The question before the court, though, has very much changed. Otherwise, if we don't have these kinds of standards, every losing party would describe the decisive factual finding as the legal conclusion and demand independent review.

Gwen: Yes, that was tried. And the Supreme Court rejected that move in 1982 in a case called Pullman-Standard v. Swint. This was a Title VII case. Black employees challenged the seniority system of the company. And the district court found no discriminatory intent behind its adoption.

Marc: And the Court of Appeals treated discriminatory intent as an ultimate fact it could reconsider independently.

Gwen: Yes, but the Supreme Court said that Rule 52 of the Federal Rules of Civil Procedure didn't permit that kind of distinction. Here's how it said it.

Marc: It does not divide facts into categories. In particular, it does not divide findings of fact into those that deal with ultimate and those that deal with subsidiary facts. So the fact that a finding determines the outcome doesn't transform it into a question of law. Intent was factual and the appellate court had to respect the trial court's fact-finding role.

Gwen: Even when the legal error distorted the fact-finding, the ordinary solution should be to send the case back for proper findings, not to have the appellate court make them itself.

Marc: Which is a meaningful limit on characterization. The importance of an answer doesn't change what sort of question it is.

Gwen: But when law and fact genuinely overlap, the reviewing court still has to work out what kind of review is appropriate. In 2024, the Supreme Court confronted such a mixed question in a setting where the classification determined whether the court could review it at all.

Marc: OK, you're talking about Wilkinson v. Garland.

Gwen: I am. Good guess. Here we had Situ Kamu Wilkinson, who'd come to the United States from Trinidad and Tobago on a tourist visa and remained after it expired. Years later, he faced removal.

Marc: And he had a U.S. citizen son.

Gwen: Yes. His son was seven when he was detained, and the child suffered from severe asthma, severe enough that it required hospitalization several times a year. He helped manage his son's medications, and he provided substantial financial and emotional support.

Marc: He sought cancellation of removal based in part on the hardship his son would suffer if he were deported.

Gwen: One of the statutory eligibility requirements was that removal would cause exceptional and extremely unusual hardship to a qualifying family member.

Marc: OK, that's a pretty demanding standard. Many children suffer when a parent is removed. But Congress required something beyond the ordinary hardship associated with that kind of separation.

Gwen: Yes. The immigration judge credited Wilkinson's testimony, but concluded that the combined hardship still didn't meet that statutory threshold. And the Board of Immigration Appeals affirmed.

Marc: And the Third Circuit said it lacked jurisdiction to review the hardship determination.

Gwen: Yes. So the immigration statute generally restricts review of certain decisions concerning cancellation of removal, even though it preserves review of constitutional claims and questions of law. The Third Circuit treated the hardship determination as an unreviewable discretionary judgment.

Marc: But there are two different decisions here. Does Wilkinson satisfy the statutory eligibility requirements? And if he does, should the government actually grant cancellation as a matter of discretion?

Gwen: That's the distinction the Supreme Court drew. Applying the statutory hardship standard to established facts, it said, is a mixed question of law and fact. And under this particular immigration provision, mixed questions fell within the exception that preserved review of questions of law.

Marc: Even though deciding whether this child faces exceptional hardship requires examining all the particular circumstances of his life.

Gwen: Yes. The Supreme Court distinguished applying the hardship standard from revisiting the factual findings themselves. Here's how.

Marc: Only the question whether those established facts satisfy the statutory eligibility standard is subject to judicial review. So the federal court can't simply reconsider the findings about the child's medical condition or Wilkinson's support. It reviews whether those established facts satisfy statutory hardship standard.

Gwen: Yes.

Marc: OK, that sounds as though Wilkinson won the right to an independent legal review, though.

Gwen: He won the right to judicial review. But the Supreme Court also said that because this hardship inquiry requires primarily factual work, that that review would be really deferential.

Marc: So the question is legal enough to get into court, but factual enough to receive deference.

Gwen: Exactly.

Marc: Okay. That separates two things that we had been treating as one, whether a court has the authority to review an issue and how independently it can decide that issue once it gets there.

Gwen: Yes. And Wilkinson didn't win his underlying immigration claim. The Supreme Court reversed the jurisdictional determination, that determination that it couldn't be reviewed, vacated the judgment and remanded. It didn't decide whether his son would experience the required hardship.

Marc: Nor would establishing hardship necessarily guarantee cancellation. The ultimate decision whether to grant it remained discretionary.

Gwen: Yes. And Justice Alito, joined by Roberts and Thomas, dissented. Because they said that the majority had extended this questions of law exception too far into what was essentially just a factual determination.

Marc: OK, what strikes me is that the immigration judge credited the evidence about Wilkinson's son, including his severe asthma and the support his father provided. The dispute was over whether all those circumstances met the statutory threshold. And the classification determines whether a federal court can even consider that argument.

Gwen: Yes. And even when review is available, that degree of deference still matters. Two years later, the Supreme Court confronted another immigration case that made this distinction even clearer. This was March of this year when the Supreme Court decided Urias-Orellana v. Bondi.

Marc: OK, what was the dispute?

Gwen: Douglas and his family had come to the United States from El Salvador in 2021 and they had applied for asylum. And he testified that a hitman had targeted him and that his family had repeatedly moved to escape threats.

Marc: So the immigration judge believed him.

Gwen: Yeah. The immigration judge explicitly found his testimony credible, but the judge concluded that his experiences didn't establish past persecution or a well-founded fear of future persecution under the immigration statute. And then the Board of Immigration Appeals affirmed and the First Circuit upheld the determination under substantial evidence review.

Marc: That strikes me as a harder case than the one where the judge simply doesn't believe the applicant. Everyone accepts his account of the threats. The disagreement is about whether the legal standard covers what happened.

Gwen: Right. Which is why he argued that the federal court should independently decide whether the established facts amounted to persecution.

Marc: I can see the argument. If nobody disputes what happened, why should the agency get the benefit of deference on whether those events satisfy the legal test?

Gwen: The answer was because the governing immigration statute required substantial evidence review of this kind of determination. The Supreme Court explicitly examined it, including its earlier decision in INS v. Elias-Zacarias.

Marc: And unanimously concluded that substantial evidence review applied to the agency's persecution determination, including its application of the statutory standard to established facts.

Gwen: Yes, this was an opinion that Jackson wrote, but there is an important limit. The applicants weren't challenging the governing legal definition of persecution. They accepted the standard. They were just disputing its application.

Marc: So the court wasn't deciding that an agency received substantial evidence deference on what the legal definition itself means.

Gwen: Right. It was deciding how closely a court reviews the agency's conclusion that particular circumstances meet or, in this case, fail to meet that standard. Justice Jackson distinguished that issue from that jurisdictional question that they'd been dealing with previously. Here's what she said.

Marc: As a result, whether a given issue is treated as a question of law exempt from the INA's jurisdiction stripping provisions tells us nothing about the type of review the court must afford to that issue under other provisions of the statute. I want to push a little bit on that because if the immigration judge believed Urias-Orellana and the court still has to defer to the conclusion that what happened wasn't persecution. How meaningful is the review? I mean, where does the court have room to disagree?

Gwen: Substantial evidence review still allows reversal, but the question is whether the record compels a conclusion contrary to the agency's. If every reasonable adjudicator would have to conclude that the established events amounted to persecution under the standard, the court would not be required to uphold the agency's contrary determination.

Marc: But it isn't enough for the federal court to think that persecution is the better conclusion. The record has to compel it.

Gwen: Right. This is a demanding standard. The court cannot replace the agency's supported determination merely because it might have reached a different conclusion independently.

Marc: And that means even a person whose account has been accepted may fail to obtain reversal because the court can't say the statute's demanding standard has been met.

Gwen: So Wilkinson held that this hardship determination could be reviewed at all and that the review had to be deferential, but it didn't specify exactly which deferential standard applied. Urias-Orellana addressed the more specific question for persecution determinations and held that substantial evidence review governs.

Marc: But doesn't this give courts a way around Loper Bright?

Gwen: In what sense?

Marc: Loper Bright says courts must independently interpret statutes. But if courts characterize a dispute as applying a legal standard to facts rather than interpreting the statute, they can still defer to the agency.

Gwen: Yes, and this means that courts can still preserve substantial deference to agencies by characterizing disputed issues as fact-intensive applications rather than questions of legal interpretation. And I think there is a real argument that after Loper Bright, this is a likely response.

Marc: This is the Alito, Roberts, and Thomas approach in Urias, right? They're saying that what the court is trying to do is dig into the factual determination. Even when the agency is effectively determining where the boundaries of a legal category fall, Whether certain threats amount to persecution, for example.

Gwen: Yes. Formally, the court might say it independently determines what persecution means. But if it defers to the agency on whether particular conduct qualifies, the agency is still retaining considerable practical authority over how that legal standard operates.

Marc: So courts can continue deferring to agency judgments with substantial legal content, even though Chevron deference is gone.

Gwen: Yes. Now, Urias-Orellana itself rested on a specific statutory provision governing substantial evidence review. So the court wasn't free to simply choose deference because it preferred the agency's judgment. But in cases where the boundary between interpretation and application is less clearly determined, how a court characterizes the issue itself can make a major difference.

Marc: Which means Loper Bright may have shifted the battle from whether courts defer to agencies on statutory interpretation to what courts are willing to characterize as statutory interpretation in the first place.

Gwen: Yes, and that struggle over the boundary between interpretation and application isn't new. It was still important long before Chevron.

Marc: Okay, so Congress can determine how courts review an agency's application of a standard, but what happens when Congress changes the standard itself? Does that change who gets to decide how it applies?

Gwen: That's what makes NLRB v. Hearst Publications so interesting. It's worth noting that this case is from 1944, which predates the APA.

Marc: Newspaper sellers? Are we talking Newsies?

Gwen: Sort of. The court said these people were called newsboys, but the court said that that was an improper term. They were misnamed boys. They were generally grown men. Many of them were supporting families.

Marc: So not quite the dancing children from Newsies.

Gwen: Yes. These were men who worked at established street corners, sometimes for years. But they weren't simply receiving a salary from a newspaper publisher. They bought papers and resold them and kept the difference.

Marc: Okay, that sounds like running a small business. You buy a product at one price, you sell it at another. Why wouldn't they just be independent contractors?

Gwen: That was the publisher's position. But look at the actual arrangement. The publisher controlled the selling prices and the price the newsboys paid for the papers as well. So the publisher largely determined how much each worker could earn on the sale.

Marc: Okay, what about the unsold papers?

Gwen: The workers could return unsold copies for credit, but they had to pay for papers that were lost or otherwise unreturnable. The publisher also controlled distribution areas and had substantial influence over how many papers each worker received.

Marc: So these men had some entrepreneurial features, but they couldn't choose their own prices, and they didn't have the kind of control over supply that an independent shopkeeper would.

Gwen: Yeah. The district manager supervised the operation and had authority over who could occupy different selling spots. And the board considered all of those circumstances when it concluded that the regular full-time newsboys, again, grown men, were employees.

Marc: Which makes it easier to see why the classification was disputed. There were real facts pointing in different directions.

Gwen: Yes. First, the Supreme Court had to decide what the statutory term employee meant. The publishers argued it should be confined to traditional common law categories.

Marc: And the court rejected that interpretation.

Gwen: It did. The court itself interpreted the statute in light of its context and purposes. Then it considered the board's application of that understanding to these particular working relationships.

Marc: And that second step required evaluating the competing characteristics of the newsboys' arrangements.

Gwen: Yes. The court distinguished judicial responsibility for statutory interpretation from the board's role in applying a broad term to particular facts. Here's what it said.

Marc: The board's determination that specified persons are employees under this act is to be accepted if it has warrant in the record and a reasonable basis in law. So the court supplied the governing understanding of the statute and the board had room to apply it to particular employment relationships.

Gwen: Yes, that's how the Loper Bright majority later characterized this. The dissent questioned how cleanly statutory interpretation and application could really be separated in cases like this.

Marc: But Congress didn't leave the broad understanding adopted in Hearst alone.

Gwen: No, it didn't. In 1947, in the Taft-Hartley amendments, it expressly excluded independent contractors from the definition of employee. And the governing legal framework shifted to common law agency principles.

Marc: So Congress narrowed the legal rule, but that still left someone to decide whether particular workers were employees or independent contractors under the new test.

Gwen: Yes. So then in 1968, the court considered that problem in NLRB v. United Insurance Company of America.

Marc: So dealing with insurance agents.

Gwen: Yes. The company used what it called debit agents. They collected premiums for policyholders and tried to prevent existing policies from lapsing and sold new insurance.

Marc: So the question was whether they were working as independent salespeople or as employees of the insurance company.

Gwen: Yes. And the board found that the agents were trained by company supervisors, that they ordinarily sold only the company's policies, they were subject to company procedures, and they also received benefits under the company's vacation insurance and pension arrangements.

Marc: That sounds considerably more like an employment relationship than the label independent contractor suggests.

Gwen: The board certainly thought so. It concluded that the agents were employees and ordered the company to bargain with their union. But the Seventh Circuit disagreed. It found that the agents were independent contractors and refused to enforce the board's order.

Marc: So the Court of Appeals substituted its conclusion about the employment relationship for the board's.

Gwen: Yes, and the Supreme Court reversed. The broader statutory approach approved in Hearst no longer governed. But applying the common law test still involved evaluating the actual relationship, and the board had to choose between fairly conflicting views supported by the record.

Marc: And the Supreme Court acknowledged that applying common law agency principles didn't require specialized administrative expertise judges lacked.

Gwen: Yeah, deference here wasn't just some sort of reward for technical expertise. It reflected the board's assigned responsibilities and the governing limits on judicial review.

Marc: So Congress could change the legal test, but that didn't mean a court could disregard the board's supported application of that test. And who counts as an employee determines who receives statutory protections. So this brings us back to Kathleen. Her ALJ concluded that she could perform sedentary work.

Gwen: Yes, and if he misunderstood the governing legal requirements, she could challenge that as legal error.

Marc: If the medical evidence, though, doesn't support this finding about Kathleen's limitations, she seems to have a factual challenge.

Gwen: And if the dispute concerns whether the established limitations satisfied the legal standard, the issue involves applying law to facts.

Marc: But those are not necessarily three separate boxes. An ALJ who misunderstands the legal standard might also make factual determinations affected by that mistake.

Gwen: Yes, good point. A single decision itself could contain multiple kinds of error. And the reviewing court has to identify the actual dispute and determine what the Social Security Act and the governing precedent would require.

Marc: Which is where we started with the chandelier. Sometimes the problem is what the law means. Sometimes it's what happened. And sometimes it's what happens when you put the two things together.

Gwen: Yes. And that distinction can determine how much authority the reviewing court has to reconsider the agency's judgment. But before a court decides whether an agency got something wrong, it has to figure out what kind of question it's reviewing.

Marc: And that classification matters. Even after Loper Bright, courts may continue to defer to agencies by treating disputes as applications of law to facts rather than questions of legal interpretation.

Gwen: So this means the decision about how to characterize the question can be essentially as important as the answer itself.

Marc: And the decision is the decision before the decision.

Gwen: Yes. Next week, we will look more closely at questions of law and what independent judicial judgment actually means after Loper Bright. So that does it for today's episode on administrative remedies. Thank you for joining us today. Please, if you enjoy this podcast and enjoy this episode, give us a like on Spotify, iTunes, or whatever platform you're listening on. And be sure to tune in next time where we'll continue to dive into the contours of administrative law. Because remember, you can't fix what you don't understand.