LOPEZ MARTINEZ v. U.S. Attorney General, Respondent. (2025)

Isaac Lopez-Martinez and Martha Martinez-Lara have lived in the United States for more than two decades but now face removal to Mexico. They each applied for cancellation of removal, arguing that deportation would cause their U.S.-citizen son an “exceptional and extremely unusual hardship.” 8 U.S.C. § 1229b(b)(1)(D). An immigration judge denied their applications, and the Board of Immigration Appeals affirmed. Isaac and Martha now ask us to vacate the Board's decisions. We hold, though, that we must review the Board's application of § 1229b(b)(1)(D)’s hardship test under the deferential substantial-evidence standard and, having done so, that there is no basis for rejecting the Board's determination. We therefore deny the petitions for review.

I

A

Under the Immigration and Nationality Act, a noncitizen who violates the immigration laws and faces removal has “several avenues for discretionary relief.” Wilkinson v. Garland, 601 U.S. 209, 212, 144 S.Ct. 780, 218 L.Ed.2d 140 (2024). One option is to convince an immigration judge to “cancel” his removal. Id. The cancellation analysis “proceeds in two steps.” Id. First, the judge “must decide whether the noncitizen is eligible for cancellation under the relevant statutory criteria.” Id. If the noncitizen has never received a green card, he “is eligible for cancellation of removal ․ if he meets four requirements.” Id. at 213, 144 S.Ct. 780. Only the fourth is relevant here: The noncitizen must “ ‘establish[ ] that removal would result in exceptional and extremely unusual hardship to [his] spouse, parent, or child,’ who is a U.S. citizen or lawful permanent resident.” Id. (second alteration in original) (quoting 8 U.S.C. § 1229b(b)(1)(D)).¹ If the noncitizen satisfies all four criteria—including the hardship requirement—the judge moves to the second step, at which he “decides whether to exercise his discretion favorably and grant the noncitizen relief.” Id. at 212–13, 144 S.Ct. 780.

The federal courts have limited jurisdiction to review cancellation-of-removal decisions. As relevant here, Congress has prescribed, and circumscribed, our jurisdiction in a three-part zigzag. First, the Immigration and Nationality Act grants us the general authority to review final removal orders. See 8 U.S.C. § 1252(a)(1). But second, the Act divests us of jurisdiction over “judgment[s] regarding” cancellation of removal. Id. § 1252(a)(2)(B)(i); see id. § 1229b. But then third, the Act partially reinstates our jurisdiction—to review “constitutional claims or questions of law.” Id. § 1252(a)(2)(D).

Our jurisdiction over “questions of law” extends to whether a removal will cause an “exceptional and extremely unusual hardship” within the meaning of § 1229b(b)(1)(D). In Guerrero-Lasprilla v. Barr, the Supreme Court held that the phrase “questions of law” in § 1252(a)(2)(D) encompasses “the application of a legal standard to undisputed or established facts”—i.e., so-called “mixed questions.” 589 U.S. 221, 227–28, 140 S.Ct. 1062, 206 L.Ed.2d 271 (2020). And more recently, in Wilkinson v. Garland, the Court further held that “the application of the ‘exceptional and extremely unusual hardship’ standard” is a “mixed question”—and “therefore a ‘question of law’ that is reviewable under § 1252(a)(2)(D).” 601 U.S. at 222, 144 S.Ct. 780 (alteration accepted). The upshot is that, on a petition for review of a final order of removal, our jurisdiction extends to immigration authorities’ conclusion that an applicant for cancellation of removal hasn't satisfied § 1229b(b)(1)(D)’s exceptional-and-extremely-unusual-hardship standard.²

B

The petitioners here are a married couple, Isaac Lopez-Martinez and Martha Martinez-Lara. Isaac and Martha are both citizens of Mexico who have long lived in the United States. But the couple's presence here is legally tenuous: Neither is a U.S. citizen or a legal permanent resident, and they have unlawfully entered the country several times, most recently in 2007.

Isaac and Martha have two U.S.-born (and thus U.S.-citizen) minor children. One of them, I.L., has a learning disability, has been diagnosed with ADHD, and has been in specialized classrooms since the second grade. I.L. regularly sees doctors for his conditions, takes medication, attends therapy sessions, and has an Individualized Education Program at his school.

C

Several years ago, the Department of Homeland Security charged Isaac and Martha with inadmissibility and commenced removal proceedings. The couple conceded inadmissibility and filed applications for cancellation of removal under § 1229b(b)(1)(D)’s discretionary-relief provision. They both argued that removal would cause I.L. exceptional and extremely unusual hardship; in particular, they said, I.L. would have to accompany the couple to Mexico, where he wouldn't be able to obtain the proper health care or education.

An immigration judge held a hearing and then, in two separate (but materially identical) decisions, rejected Isaac's and Martha's applications. The judge found both Isaac and Martha to be credible and agreed that they met the first three step-one criteria for cancellation of removal. See 8 U.S.C. § 1229b(b)(1)(A)–(C). She concluded, though, that they didn't satisfy the fourth criterion, the hardship standard. See id. § 1229b(b)(1)(D). The judge acknowledged that removal would be difficult for I.L., but she found that the difficulty didn't meet the “very high” exceptional-and-extremely-unusual-hardship bar. Because Isaac and Martha weren't eligible for cancellation at step one, the judge didn't reach the step-two discretionary determination.

The Board of Immigration Appeals affirmed. In two separate single-judge decisions, the Board agreed with the immigration judge's conclusion that removal wouldn't cause I.L. an “exceptional and extremely unusual hardship” within the meaning of § 1229b(b)(1)(D).

Isaac and Martha petitioned us for review. Their arguments focus on the Board's application of the “exceptional and extremely unusual hardship” standard. Following the Supreme Court's decision in Wilkinson, which clarified our jurisdiction, we ordered supplemental briefing regarding the standard of review that applies to the Board's application of § 1229b(b)(1)(D)’s hardship standard. We now consider the couple's consolidated petitions for review.

II

There is a surfeit of standards by which appellate courts review other tribunals’ decisions. The most familiar are the standards that apply when appellate courts review lower-court decisions. In this “court/court” context, one of three standards typically applies, depending on the issue. An appellate court reviews a lower court's resolution of legal questions “de novo,” its determination of factual questions for “clear error,” and its handling of assorted discretionary tasks for “abuse of discretion.” Pierce v. Underwood, 487 U.S. 552, 558, 108 S.Ct. 2541, 101 L.Ed.2d 490 (1988). A murky fourth category of questions are “mixed”—they include both factual and legal aspects. See Bufkin v. Collins, ––– U.S. ––––, 145 S. Ct. 728, 739, 221 L.Ed.2d 192 (2025). The standard of review applicable to a mixed question depends on whether the question is primarily legal, or primarily factual, id.—more on that to come.

Perhaps less familiar—but no less important—are the standards applicable to “court/agency” review. These standards govern courts’ review of administrative agency actions, like the Board of Immigration Appeals’ decision here. There are a host of complexities in the court/agency context generally—some of which we'll need to explain briefly—but for the most part, standard-of-review law applicable in immigration cases, in particular, is well settled.

Here, for instance, are a few things we know for sure: Our review extends both to the Board of Immigration Appeals’ decision and, to the extent that the Board “expressly adopts or agrees with” it, to the immigration judge's decision. A.P.A. v. U.S. Att'y Gen., 104 F.4th 230, 236 (11th Cir. 2024). We review legal conclusions de novo and (when we have jurisdiction) findings of fact for “substantial evidence.” Kazemzadeh v. U.S. Att'y Gen., 577 F.3d 1341, 1350 (11th Cir. 2009). Nonetheless, as already explained, we lack jurisdiction to review factual determinations underlying an immigration judge's or the Board's decision to deny an application for cancellation of removal. Wilkinson, 601 U.S. at 225, 144 S.Ct. 780; 8 U.S.C. § 1252(a)(2)(B)(i), (a)(2)(D).

This case presents one unknown: Neither we nor the Supreme Court have ever settled on the standard of review that applies to mixed questions raised in petitions for review of the Board's decisions. In particular—and as relevant here—we've never settled on the standard that applies to a § 1229b(b)(1)(D) exceptional-and-extremely-unusual-hardship determination. We do so today. We proceed in three steps: First, we identify the available options; second, we explain our methodology for choosing among them; and third, we apply that methodology to select the proper standard. Spoiler alert: We hold that a determination regarding § 1229b(b)(1)(D)’s hardship standard is reviewable for substantial evidence.

Let us explain.

A

Our first step is to identify the available standards, and to do so we evaluate both “traditional administrative law principles” and the Immigration and Nationality Act. Garland v. Ming Dai, 593 U.S. 357, 369, 141 S.Ct. 1669, 210 L.Ed.2d 11 (2021).

The Administrative Procedure Act supplies three possible standards of review. First, a court reviews legal issues de novo, including by applying its “independent judgment” to statutory-interpretation questions. Loper Bright Enters. v. Raimondo, 603 U.S. 369, 391–92, 412, 144 S.Ct. 2244, 219 L.Ed.2d 832 (2024); see DeKalb Cnty. v. U.S. Dep't of Lab., 812 F.3d 1015, 1020 (11th Cir. 2016); 5 U.S.C. § 706. Second, the APA prescribes a deferential “arbitrary and capricious” standard to govern a court's review of an agency's decisionmaking process. See 5 U.S.C. § 706(2)(A); cf. Dep't of Homeland Sec. v. Regents of the Univ. of Cal., 591 U.S. 1, 16, 140 S.Ct. 1891, 207 L.Ed.2d 353 (2020) (explaining that the APA “requires agencies to engage in ‘reasoned decisionmaking’ ” (citation omitted)). Arbitrary-and-capricious review extends both to agency decisions exercising discretion and to those finding facts. See 2 Kristin E. Hickman & Richard J. Pierce, Jr., Administrative Law Treatise §§ 10.4, 11.1 (7th ed. 2018 & Supp. 2025); Harry T. Edwards & Anne Deng, Federal Standards of Review 113–14 (2024).³ The de novo and arbitrary-and-capricious standards are generally applicable defaults—the APA doesn't limit them to any particular subset of agency decisions. Cf. 5 U.S.C. § 706.

The third APA-based standard of review is “substantial evidence,” which applies to findings of fact. Id. § 706(2)(E). Under the substantial-evidence standard, a reviewing court asks “[w]hether on the record as a whole there is substantial evidence to support [the] agency[’s] findings.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 491, 71 S.Ct. 456, 95 L.Ed. 456 (1951); see also Allentown Mack Sales & Serv., Inc. v. NLRB, 522 U.S. 359, 380, 118 S.Ct. 818, 139 L.Ed.2d 797 (1998). Unlike the de novo and arbitrary-and-capricious standards, the APA itself makes the substantial-evidence standard applicable in only one type of proceeding—namely, a challenge to a so-called “formal” agency action. See 5 U.S.C. § 706(2)(E). As just noted, the substantial-evidence standard (somewhat confusingly) overlaps with the arbitrary-and-capricious standard as a means of reviewing agency factfindings. Indeed, most of the time, “[w]hen the arbitrary or capricious standard is performing th[e] function of assuring factual support, there is no substantive difference between what it requires and what would be required by the substantial evidence test.” Ass'n of Data Processing Serv. Orgs., Inc. v. Bd. of Governors of the Fed. Rsrv. Sys. (ADAPSO), 745 F.2d 677, 683–84 (D.C. Cir. 1984) (Scalia, J.). We'll revisit this overlap—and try to disentangle the two standards as they apply in immigration cases—in due course.

The Immigration and Nationality Act supplements the APA's default standards of review. As particularly relevant here, it states that an administrative agency's “findings of fact are conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary,” 8 U.S.C. § 1252(b)(4)(B)—which the Supreme Court has interpreted to mean, in effect, so long as they are supported by “substantial evidence.” See Nasrallah v. Barr, 590 U.S. 573, 584, 140 S.Ct. 1683, 207 L.Ed.2d 111 (2020) (stating, in an immigration case, that “[t]he standard of review is the substantial-evidence standard” and citing 8 U.S.C. § 1252(b)(4)(B)).

Under the APA's default rules, the substantial-evidence standard wouldn't apply in a case, like this one, challenging a Board of Immigration Appeals decision. That's because, despite their court-like trappings, immigration proceedings don't qualify as “formal” adjudications.⁴ But § 1252(b)(4)(B) fills a gap left by the APA and, as interpreted by the Supreme Court in Nasrallah, makes the substantial-evidence standard applicable in immigration cases.⁵

So, taking stock: Considered in tandem, the APA and the Immigration and Nationality Act provide three standard-of-review contestants: de novo (for legal issues); arbitrary and capricious (for exercises of discretion and findings of fact); and substantial evidence (for findings of fact). We turn next to the methodology for choosing among these three standards.

B

As already noted, the Supreme Court recently clarified that § 1229b(b)(1)(D)’s hardship standard presents a “mixed question” of law and fact. Wilkinson, 601 U.S. at 222, 144 S.Ct. 780. And helpfully, in a pair of recent decisions, the Court has sketched out a method for identifying the proper standard of review for mixed questions, at least in the court/court context. First, “a reviewing court should try to break [a mixed] question into its separate factual and legal parts, reviewing each according to the appropriate legal standard.” Google LLC v. Oracle Am., Inc., 593 U.S. 1, 24, 141 S.Ct. 1183, 209 L.Ed.2d 311 (2021). Second, for an irreducibly mixed question—i.e., one whose legal and factual elements can't be disaggregated—the standard of review “all depends[ ] on whether answering it entails primarily legal or factual work.” Id. (quoting U.S. Bank Nat'l Ass'n ex rel. CWCapital Asset Mgmt. LLC v. Vill. at Lakeridge, LLC, 583 U.S. 387, 396, 138 S.Ct. 960, 200 L.Ed.2d 218 (2018)). Some mixed questions “require courts to expound on the law, particularly by amplifying or elaborating on a broad legal standard”—they should be reviewed de novo. U.S. Bank, 583 U.S. at 396, 138 S.Ct. 960. Others “immerse courts in case-specific factual issues”—they should be reviewed “with deference,” which in the court/court context means for clear error. Id. at 396, 399, 138 S.Ct. 960.

Though not squarely applicable, the Google-U.S. Bank framework is instructive for court/agency review. In fact, the Supreme Court has already suggested—albeit only in dicta, and cursorily—that Google-U.S. Bank supplies the appropriate method for choosing the standard applicable to a court's review of an agency's resolution of a mixed question. In particular, in Guerrero-Lasprilla, the Court quoted U.S. Bank to explain how it has “determin[ed] the proper standard for appellate review of a district, bankruptcy, or agency decision that applies a legal standard to underlying facts.” 589 U.S. at 228, 140 S.Ct. 1062 (emphasis added). And several of our sister circuits have gone one step further, applying the Google-U.S. Bank framework in immigration cases. See Alzaben v. Garland, 66 F.4th 1, 7 (1st Cir. 2023); Wilkinson v. Att'y Gen. U.S. (Wilkinson II), 131 F.4th 134, 138–40 (3d Cir. 2025); Williams v. Garland, 59 F.4th 620, 633–34 (4th Cir. 2023), as amended (Feb. 10, 2023); Singh v. Rosen, 984 F.3d 1142, 1154 (6th Cir. 2021); Gonzalez-Juarez v. Bondi, 137 F.4th 996, 1000–03 (9th Cir. 2025). We're convinced and follow suit.

Extending Google-U.S. Bank jot for jot to the court/agency context does pose one wrinkle that needs to be ironed out: What to do with arbitrary-and-capricious review? As already explained, in the court/court context, the mixed-question options are de novo—either for any discrete legal components or for an irreducibly law-heavy issue—and clear error—either for discrete factual components or for an irreducibly fact-heavy issue. As we've seen, things on the court/agency side are a bit more complicated—principally because there are two standards, not one, that apply to factual determinations: substantial-evidence and arbitrary-and-capricious. Which of those two should sub in for clear error in the Google-U.S. Bank analysis?

For several reasons, we hold that, at least in the immigration context, the substantial-evidence standard—rather than the arbitrary-and-capricious standard—is the best fit. First, and most obviously, the substantial-evidence and clear-error standards are quite similar in theory and essentially identical in practice. In Dickinson v. Zurko, the Supreme Court called “the difference” between the standards “a subtle one” and said that (with the exception of the case before it) it had “failed to uncover a single instance in which a reviewing court conceded that use of one standard rather than the other would in fact have produced a different outcome.” 527 U.S. 150, 162–63, 119 S.Ct. 1816, 144 L.Ed.2d 143 (1999). Second, and relatedly, unlike the substantial-evidence and clear-error standards, the arbitrary-and-capricious standard isn't really designed for review of factual findings. Rather, as then-Judge Scalia once explained, “the ‘arbitrary or capricious’ provision[ ] is a catchall, picking up administrative misconduct not covered by the other more specific” standards. ADAPSO, 745 F.2d at 683. Here, we have a more specific standard for review of factual determinations—namely, substantial evidence, because (as already noted) Congress has extended that standard to immigration proceedings. See 8 U.S.C. § 1252(b)(4)(B); Nasrallah, 590 U.S. at 584, 140 S.Ct. 1683. And third, the arbitrary-and-capricious standard is “cumulative” rather than exclusive of substantial-evidence review, in any event. ADAPSO, 745 F.2d at 683. So, adopting substantial-evidence review for a mixed question doesn't necessarily exclude arbitrary-and-capricious review. An agency action might be “supported by the required substantial evidence,” and yet “in another regard” be arbitrary and capricious—“for example, because it is an abrupt and unexplained departure from agency precedent.” Id.⁶

Accordingly, we adapt the Google-U.S. Bank framework for immigration-proceeding purposes by substituting substantial evidence in the place of clear error.

C

We turn, then, to the application of the (modified) Google-U.S. Bank methodology here. Under that framework, what standard of review governs a court's review of a hardship determination under § 1229b(b)(1)(D)? The Supreme Court's decision in Wilkinson doesn't squarely decide the issue, but it does provide three helpful hints. First, the Board's (or an immigration judge's) application of the standard presents a “mixed question” of law and fact. 601 U.S. at 222, 144 S.Ct. 780. Second, this mixed question is “primarily factual.” Id. at 225, 144 S.Ct. 780. And third, because the hardship question is primarily factual, a court's review is “deferential.” Id.⁷ Pairing Wilkinson’s hints with Google-U.S. Bank pretty clearly indicates that the applicable standard of review is substantial evidence. That conclusion follows from two premises.

First, per Google-U.S. Bank, we must, to the extent possible, separate what might appear to be a mixed question into its legal and factual components. In Wilkinson, while clearly holding that a court has jurisdiction to review the Board's § 1229b(b)(1)(D) hardship determination, the Supreme Court reiterated that “a court is still without jurisdiction to review a factual question raised in an application for” cancellation of removal. Id. at 222, 144 S.Ct. 780. So, for example, a court lacks jurisdiction to second-guess an immigration judge's “underlying factual determination” that an applicant for cancellation of removal “was credible” or that an applicant's U.S.-citizen relative “had a serious medical condition.” Id. A witness's credibility and the existence of a medical condition are related to the hardship issue, but they remain at their core factual determinations rather than irreducibly mixed questions. Accordingly, we must always take care to winnow factual predicates underpinning the hardship determination—with respect to which we lack jurisdiction—from the hardship determination itself—over which we have jurisdiction.

Second, Google-U.S. Bank instructs us to select a standard of review for an irreducibly mixed question according to whether it is primarily legal or factual. Wilkinson expressly characterized the application of the hardship inquiry as “primarily factual.” Id. at 225, 144 S.Ct. 780. And for reasons already explained, the substantial-evidence standard applies to factual determinations made in immigration proceedings. Accordingly, we review for substantial evidence the question whether the found facts in the administrative record (which, again, we lack jurisdiction to second-guess) add up to an “exceptional and extremely unusual hardship” within the meaning of § 1229b(b)(1)(D).⁸

* * *

So, to recap: When considering decisions of the Board of Immigration Appeals and (where appropriate) immigration judges, we review legal conclusions de novo and findings of fact for substantial evidence—keeping in mind that in some contexts, like this one, we lack jurisdiction to review pure findings of fact. We review irreducibly mixed questions de novo or for substantial-evidence, depending on whether legal or factual issues predominate.⁹ When the Board (or an immigration judge) “weighs ․ found facts and applies the ‘exceptional and extremely unusual hardship’ standard, ․ the result is a mixed question of law and fact,” Wilkinson, 601 U.S. at 222, 144 S.Ct. 780, which, because it is “primarily factual,” id. at 225, 144 S.Ct. 780, we review for substantial evidence.

On, then—at last—to the merits.

III

The Board has settled on a two-part test for hardship-based cancellation-of-removal applications like Isaac's and Martha's.¹⁰ “[T]o the extent that a claim is based on the health of a qualifying relative, an applicant needs to establish [1] that the relative has a serious medical condition and, [2] if he or she is accompanying the applicant to the country of removal, that adequate medical care for the claimed condition is not reasonably available in that country.” Matter of J-J-G-, 27 I. & N. Dec. 808, 811 (B.I.A. 2020) (footnotes omitted); see also Matter of Monreal-Aguinaga, 23 I. & N. Dec. 56, 63 (B.I.A. 2001) (en banc) (suggesting that a “strong” cancellation of removal applicant “might have a qualifying child with very serious health issues, or compelling special needs in school”); Matter of Andazola-Rivas, 23 I. & N. Dec. 319, 323 (B.I.A. 2002) (en banc) (considering, as part of an education-based hardship application, whether the qualifying relative “would be deprived of all schooling or of an opportunity to obtain any education”). We understand Isaac and Martha to make two different arguments under this test, and we address each in turn.

A

First, Isaac and Martha argue that the level of medical care (and related educational support) in Mexico is so deficient that if they are removed, I.L. will suffer an exceptional and extremely unusual hardship. In their briefs to us, Isaac and Martha emphasize the limited availability of healthcare and special education in Mexico. Whether those services are (or aren't) available in Mexico is a purely factual issue over which we have no jurisdiction. See Wilkinson, 601 U.S. at 222, 144 S.Ct. 780. But the couple concedes that “[t]he facts of th[is] case are not in dispute,” Br. of Pet'r in No. 23-10105 at 17; Br. of Pet'r in No. 23-12058 at 17, and so we take them to be arguing, more broadly, that the undisputed facts add up to an “exceptional and extremely unusual hardship” within the meaning of § 1229b(b)(1)(D). This is an argument about the hardship standard's application, over which we have jurisdiction and which, for reasons explained, we review for substantial evidence.¹¹

The Board's hardship determinations were supported by substantial evidence. The hardship bar is “high”: Conditions satisfying it are “substantially beyond that which ordinarily would be expected to result from the alien's deportation.” Flores-Alonso v. U.S. Att'y Gen., 36 F.4th 1095, 1098 (11th Cir. 2022) (quoting Monreal-Aguinaga, 23 I. & N. Dec. at 59), abrogated on other grounds by Wilkinson, 601 U.S. at 217 n.2, 144 S.Ct. 780. Here, the Board observed that I.L. could likely access substitute ADHD medication, and it reasoned that while I.L. might have “fewer educational opportunities in Mexico,” he wouldn't be entirely “deprived of all schooling.” These conclusions are supported by the record. The immigration judge heard testimony and reviewed documentation indicating that an education and ADHD medicine are available in Mexico. And the judge appropriately acknowledged contrary evidence, including testimony that—whatever the general availability of medication and special education in Mexico—accessing those services for I.L. in particular would be difficult and expensive. While an immigration judge “must consider all the evidence,” and while the judge here could have devoted even more attention to considerations that favored Isaac and Martha, she did “not need to discuss each piece of evidence in [her] order.” Lingeswaran v. U.S. Att'y Gen., 969 F.3d 1278 (11th Cir. 2020) (citation modified).

Isaac and Martha haven't explained why the Board's conclusion wasn't “supported by reasonable, substantial, and probative evidence on the record considered as a whole.” A.P.A., 104 F.4th at 236 (citation modified). They have shown no obvious disconnect between the Board's conclusions and the contents of the administrative record. It is of course possible that alternative inferences might have been drawn from the record that would have been more favorable to Isaac and Martha. And we recognize that removing the couple may very well cause I.L. some—even much—hardship. But § 1229b(b)(1)(D)’s hardship standard is “high,” Monreal-Aguinaga, 23 I. & N. Dec. at 60, and substantial-evidence review is deferential. We can't “reweigh the evidence from scratch.” Murugan v. U.S. Att'y Gen., 10 F.4th 1185, 1194 (11th Cir. 2021) (citation modified). In light of these considerations, we conclude that the Board didn't err.

B

Second and separately, Isaac and Martha contend that the Board misapplied its own decision in J-G-G- by requiring them to show that the services I.L. needs aren't available at all in Mexico, rather than that they aren't “reasonably available.” Their argument that the Board failed to “appl[y] the correct legal standard” is a question of law that we review de novo. Farah v. U.S. Att'y Gen., 12 F.4th 1312, 1325 (11th Cir. 2021). It's true that the Board occasionally used the term “unavailable.” But it also accurately quoted J-J-G-’s “reasonably available” standard, and the substance of the Board's analysis is consistent with that measure. As just noted, the Board explained that an alternative medication was available in Mexico and that I.L. could still obtain an education there. In other words, the Board's analysis wasn't just that some kind of care was available—instead, it concluded that I.L.’s particular needs could likely be satisfied. We are therefore persuaded that the Board applied the right standard, even if its word choices were at times imprecise.

* * *

Accordingly, we find no fault in the Board's decisions—either in its application of the hardship standard or its use of the J-J-G- test.

IV

In sum, we hold that a court should review the Board's § 1229b(b)(1)(D) hardship determination for substantial evidence. And, applying that standard to the Board's decisions affirming the immigration judge's denials of Isaac and Martha's applications for cancellation of removal, we hold that the Board's decisions were supported by substantial evidence. We accordingly DENY the petitions for review.

FOOTNOTES

1.   To satisfy the first three requirements, the noncitizen must show that he (1) “has been physically present in the United States for a continuous period of not less than 10 years immediately preceding the date of such application,” (2) “has been a person of good moral character” during that period, and (3) hasn't been convicted of certain criminal offenses. 8 U.S.C. § 1229b(b)(1)(A)–(C).

2.   Before Wilkinson, we had held that we lacked jurisdiction over the “determination of whether an applicant's relatives will experience exceptional and extremely unusual hardship.” Flores-Alonso v. U.S. Att'y Gen., 36 F.4th 1095, 1100 (11th Cir. 2022) (per curiam). That portion of Flores-Alonso was overruled by Wilkinson and is therefore no longer good law. See Wilkinson, 601 U.S. at 217 n.2, 144 S.Ct. 780 (specifically identifying Flores-Alonso as falling on the wrong side of the preWilkinson circuit split over a hardship determination's reviewability).

3.   Our cases are rarely as clear as the treatises about the domain of arbitrary-and-capricious review. But the treatises’ gloss—that arbitrary-and-capricious review may encompass both exercises of discretion and findings of fact—is borne out in the real world. See, e.g., Ferreira v. U.S. Att'y Gen., 714 F.3d 1240, 1243 (11th Cir. 2013) (applying arbitrary-and-capricious review to an “exercise of discretion”); Atlanta Gas Light Co. v. FERC, 140 F.3d 1392, 1397 (11th Cir. 1998) (same, for “findings of fact[ ]”).

4.   Under the APA, an agency adjudication is “formal” only if it is “determined on the record after opportunity for an agency hearing.” 5 U.S.C. § 554(a). Courts have interpreted this provision “narrowly,” 1 Hickman & Pierce, supra, § 6.2, and the Supreme Court has held that immigration-court proceedings aren't “formal” in the technical, APA sense, see Ardestani v. INS, 502 U.S. 129, 139, 112 S.Ct. 515, 116 L.Ed.2d 496 (1991).

5.   Two other familiar standards of review—applicable in what we've called the court/court context—are off the table. First, the “clear error” standard doesn't fit, even though the agency decisions at issue here do involve findings of fact. In Dickinson v. Zurko, the Supreme Court emphasized “the importance of maintaining a uniform approach to judicial review of administrative action” and held that APA § 706 supplies “court/agency review standards in the absence of an exception.” 527 U.S. 150, 154, 119 S.Ct. 1816, 144 L.Ed.2d 143 (1999). Here, no exception applies to immigration proceedings.Second, the “abuse of discretion” standard—which the government raised and then rejected in its supplemental brief, isn't really a distinct standard for our purposes. Although familiar in the court/court context, see Edwards & Deng, supra, at 15, 68–70, in the court/agency context abuse-of-discretion review is more or less equivalent to arbitrary-and-capricious review, cf., e.g., Alkotof v. U.S. Att'y Gen., 106 F.4th 1289, 1297 (11th Cir. 2024) (stating that immigration officials abuse their discretion when their actions are “arbitrary or capricious”); Edwards & Deng, supra, at 204 (observing that courts “rarely draw any meaningful distinctions” between acts that are “arbitrary” or “capricious” and those that constitute “abuse[s] of discretion” (citation modified)). As we will explain, the arbitrary-and-capricious standard is “cumulative” of other APA standards. ADAPSO, 745 F.2d at 683. So at least in principle, an agency action might—for example—both be unsupported by substantial evidence and be an arbitrary and capricious abuse of discretion. See id. at 683–84.

6.   One more thing: Using the arbitrary-and-capricious standard for mixed questions would create an anomaly in the deference spectrum. Substantial-evidence and arbitrary-and-capricious review overlap a great deal, see ADAPSO, 745 F.2d at 684, but to the extent that they differ, the Supreme Court has repeatedly characterized arbitrary-and-capricious review as being even more deferential than substantial-evidence review. See 2 Hickman & Pierce, supra, § 10.4; e.g., Abbott Lab'ys v. Gardner, 387 U.S. 136, 143, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967). So if we used arbitrary-and-capricious review for mixed questions, the most deferential standard would (strangely) occupy the law-fact spectrum's middle.

7.   At oral argument, the petitioners briefly suggested that our review of the hardship standard should be de novo. But because the Supreme Court said that our review must be “deferential,” Wilkinson, 601 U.S. at 225, 144 S.Ct. 780, one thing we can be certain of is that de novo review isn't an option.

8.   We're not alone. On remand from the Supreme Court's decision in Wilkinson, the Third Circuit likewise held, for similar reasons, “that the substantial-evidence standard governs review of a hardship determination in a cancellation-of-removal proceeding.” Wilkinson II, 131 F.4th at 142. And the Ninth Circuit recently followed suit. Gonzalez-Juarez, 137 F.4th at 1000–03.

9.   As always, there remains the background default that immigration authorities’ decisions are subject to arbitrary-and-capricious review. See ADAPSO, 745 F.2d at 683.

10.   No party argued that we should apply a different standard in light of Loper Bright, 603 U.S. at 412, 144 S.Ct. 2244, and so we don't reach that issue here.

11.   At oral argument, and then again in a supplemental-authority letter, the government argued that the reasonable availability of medical care in a foreign country is a purely factual question with respect to which we lack jurisdiction. But we don't take Isaac and Martha—understood most charitably—to be disputing the precise level of medical or educational resources in Mexico. After all, they expressly agree that the facts in this case are undisputed. Rather, their position is that, on these facts, the difficulties I.L. would face satisfy the hardship standard.Moreover, “reasonableness” isn't inevitably a purely factual determination. Compare, e.g., United States v. E. Air Lines, Inc., 792 F.2d 1560, 1564 (11th Cir. 1986) (“The question of reasonableness is a mixed question of law and fact.” (footnote omitted)), with, e.g., Martin v. Occupational Safety & Health Rev. Comm'n, 947 F.2d 1483, 1484 (11th Cir. 1991) (“[R]easonable diligence is a question of fact, subject to substantial evidence review, and not a mixed question of law and fact.”). At least in this case, the question whether a particular level of medical care is “reasonable” is just a different way of phrasing the underlying hardship standard, whose application we unquestionably have jurisdiction to consider. See Wilkinson, 601 U.S. at 222, 144 S.Ct. 780. Accordingly, insofar as Isaac and Martha might be understood as disputing reasonableness, we take them to be disputing the question whether the hardship standard is met.

Newsom, Circuit Judge: