Lopez v. Bondi

Opinion

FOR PUBLICATION

UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

CHRISTIAN LOPEZ, No. 23-870 Agency No. Petitioner, A205-882-422 v. ORDER PAMELA BONDI, Attorney General,

Respondent.

Filed August 25, 2025

Before: Sidney R. Thomas, Consuelo M. Callahan, and Gabriel P. Sanchez, Circuit Judges.

Order; Dissent by Judge Bumatay 2 LOPEZ V. BONDI

SUMMARY *

Immigration

The panel denied a petition for panel rehearing and denied a petition for rehearing en banc in a case in which the panel denied Christian Lopez’s petition for review of a decision of the Board of Immigration Appeals, concluding that Lopez’s petit larceny convictions under Reno Municipal Code § 8.10.040 are crimes involving moral turpitude that made him removable. Judge Bumatay, with whom Judge Ikuta joined and with whom Judge Collins concurred as to Part II(C), dissented from the denial of rehearing en banc. In his view, the panel violated Loper Bright Enterprises v. Raimondo, 144 S. Ct. 2244 (2024), in at least three ways—each warranting en banc review. First, the panel effectively deferred to the BIA, and in doing so, abdicated its judicial role by not independently evaluating the law. Second, the panel afforded deference to the BIA even though the statute at issue contained no ambiguity. Third, the panel misread Loper Bright to preclude three-judge panels from revisiting circuit precedent based on the now-defunct Chevron doctrine.

* This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. LOPEZ V. BONDI 3

ORDER

Judge Callahan and Judge S. R. Thomas have voted to deny the petition for panel rehearing. Judge Callahan has voted to deny the petition for rehearing en banc and Judge S.R. Thomas has so recommended. Judge Sanchez has voted to grant the petition for panel rehearing and the petition for rehearing en banc. The full court was advised of the petition for rehearing en banc. A judge requested a vote on whether to rehear the matter en banc. The matter failed to receive a majority of the votes of the nonrecused active judges in favor of en banc consideration. Fed. R. App. P. 40. The petition for panel rehearing and the petition for rehearing en banc are DENIED. All pending motions are DENIED. No further petitions for rehearing or rehearing en banc will be entertained.

BUMATAY, Circuit Judge, with whom IKUTA, Circuit Judge, joins, and with whom COLLINS, Circuit Judge, concurs as to Part II(C), dissenting from the denial of rehearing en banc:

When the Supreme Court handed down Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), it announced a sea change in how federal courts must treat the Executive Branch’s interpretation of the law. The instruction was clear: Courts must independently interpret statutes and must not defer to an executive agency’s legal interpretations. See Loper Bright, 603 U.S. at 412–13. In other words, the days of our courts’ binding deference to agency interpretations 4 LOPEZ V. BONDI

under Chevron v. Natural Resources Defense Council, 467 U.S. 837 (1984), are gone. Chevron is dead and buried and the separation of powers is restored. Now our duty as judges is to “use every tool at [our] disposal to determine the best reading of the statute.” Loper Bright, 603 U.S. at 400. So in all cases, our starting point is the text and our job is to determine the best reading of the text using the traditional tools of statutory interpretation. Sure, one of those tools may include peeking at how an executive agency resolves a statutory ambiguity—generally when the question involves a matter within the agency’s factual expertise. See id. at 385–86. But that first requires determining that an ambiguity exists and, even then, it is just one of many tools. In no way is it the predominant or primary tool. After all, unlike judges, executive agencies are self-interested litigants like any other with incentives to press a reading of the law favorable to them or advantageous to their policy preferences. Thus, “courts need not and under the [Administrative Procedure Act (“APA”)] may not defer to an agency interpretation of the law simply because a statute is ambiguous.” Id. at 413. Enter Christian Lopez. Lopez, a Mexican citizen, came to the United States as a child in 2000 and was a non- immigrant visa holder until his visa lapsed in 2017. In the spring of 2019, Lopez went on a multi-week crime spree in Nevada. He was eventually apprehended and charged with multiple offenses, including trespassing, shoplifting, and carrying a firearm. He pleaded guilty to four counts of petit larceny under municipal law and served 14 months in state prison. On his release, the government took him into immigration custody and began removal proceedings. The government sought to remove Lopez as an alien “convicted of two or more crimes involving moral turpitude, not arising LOPEZ V. BONDI 5

out of a single scheme of criminal misconduct.” See 8 U.S.C. § 1227(a)(2)(A)(ii). An immigration judge sustained his removability and the Board of Immigration Appeals (“BIA”) affirmed. On appeal to the Ninth Circuit, Lopez raised three issues. First, Lopez argued that his petit larceny convictions were too broad to categorically match a “crime involving moral turpitude.” Second, Lopez asserted that his convictions were ineligible for a pardon and so he could not be removed because of the pardon waiver provision of § 1227(a)(2)(A)(vi). Finally, Lopez claimed that his offenses were part of a “single scheme of criminal misconduct,” taking him out of the ambit of § 1227(a)(2)(A)(ii). A three-judge panel of the Ninth Circuit denied Lopez’s petition for review, with Judge Sanchez dissenting only as to whether Lopez’s convictions were categorical crimes involving moral turpitude. See Lopez v. Garland, 116 F.4th 1032 (9th Cir. 2024). Unfortunately, this was no mine-run immigration case. That’s because, in denying Lopez’s petition, the panel took the extraordinary step of resurrecting Chevron under the alias of “Skidmore deference.” Id. at 1041 (simplified). On issue after issue, the panel consistently sought to find ways to “respect” the BIA’s interpretation of the law instead of simply conducting its own independent statutory analysis. Thus, Lopez violated Loper Bright in at least three ways— each warranting en banc review. First, in ruling that Lopez’s convictions were a categorical “crime involving moral turpitude”—a quintessential legal question squarely within our judicial bailiwick—the panel effectively deferred to the BIA’s view because it was “thorough,” “well-reasoned,” and “consistent 6 LOPEZ V. BONDI

with judicial precedent.” Id. at 1040. Never mind that the Ninth Circuit had disagreed with the BIA’s interpretation, that the BIA had flipped its position too, and that the BIA has no special expertise on moral turpitude. See id. at 1041. Indeed, the panel majority offered zero statutory analysis independent of the BIA’s interpretation. Id. Instead, the panel majority was convinced that the “BIA’s thoroughness, persuasive reasoning, and consistency with . . . longstanding [legal doctrine]” was enough to follow the agency’s interpretation. Id. The panel then overruled our prior precedent, Castillo-Cruz v. Holder, 581 F.3d 1154, 1159 (9th Cir. 2009), and its progeny, that conflicted with the BIA’s new-found position. Simply, the panel “asked the wrong question” by starting with whether the BIA’s interpretation was “entitled to respect.” Br. Amici Prof. Michael Kagan and Christopher Walker 14. Rather, the right question is, and always is, “what’s the best reading of the statute?” Even if an interpretation is thorough, well-reasoned, and consistent with some authorities, that doesn’t mean it’s the best one. And “[i]n the business of statutory interpretation, if it is not the best, it is not permissible.” Loper Bright, 603 U.S. at 400. So the panel abdicated the judicial role and just applied Chevron deference by another name. Whatever “respect” we give executive agencies under Loper Bright, it can’t be a deference indistinguishable from Chevron. Second, in declining to apply the pardon waiver to Lopez’s convictions, the panel majority continued to defer to the BIA’s interpretation of the law even though the law contained no ambiguity. As the panel concluded, the pardon waiver’s “plain language” forecloses Lopez’s argument. Still, the panel took pains to “afford” the BIA’s interpretation of the law “Skidmore deference.” Lopez, 116 LOPEZ V. BONDI 7

F.4th at 1044. This makes little sense. If the statutory text resolves the matter unambiguously, then we stop there. We don’t then check whether the Executive branch agrees with the plain meaning. We don’t check because we don’t care. The law governs—not agency interpretation. So deference and respect have nothing to do with this question. And it contradicts Loper Bright to say otherwise. See Loper Bright, 603 U.S. at 379 (“If . . . congressional intent is ‘clear,’ that is the end of the inquiry.”). Third, the panel misread Loper Bright to preclude three- judge panels from revisiting circuit precedent based on the now-defunct Chevron doctrine. In denying that Lopez’s convictions were part of a “single scheme,” the panel again deferred to the BIA’s interpretation of the phrase even while acknowledging that the agency’s reading was “in tension” with prior Ninth Circuit authority. Lopez, 116 F.4th at 1045 (citing Wood v. Hoy, 266 F.2d 825 (9th Cir. 1959)). The panel found itself helpless to fix the situation because our circuit had granted the BIA’s interpretation Chevron deference in a 2019 case—Szonyi v. Whitaker, 915 F.3d 1228 (9th Cir. 2019). As Loper Bright was an “intervening higher authority” that is “clearly irreconcilable with” Szonyi, the panel was well within is rights—and, in fact, duty- bound—to revisit Szonyi. See Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003) (en banc). Simply, Szonyi directly “uph[e]ld the BIA’s interpretation under the principles of Chevron deference that apply when the BIA interprets immigration laws.” Szonyi, 915 F.3d at 1231. The Supreme Court has now said that the Chevron framework violates the APA. Loper Bright, 603 U.S. at 399 (“Chevron cannot be reconciled with the APA[.]”). Thus, this should have been a straightforward case of overruling bad circuit precedent. By 8 LOPEZ V. BONDI

failing to do so, the panel violates both Loper Bright and Miller v. Gammie. What makes the panel’s decision even odder is how haphazardly it treats circuit precedent. The panel reads Loper Bright to preclude overruling Szonyi, yet it has no problem saying goodbye to Castillo-Cruz. Compare Lopez, 116 F.4th at 1045 with id. at 1041. The difference between Szonyi and Castillo-Cruz? Apparently, to the panel, it’s that the BIA now likes Szonyi but doesn’t like Castillo-Cruz. Once again, this doesn’t faithfully apply Miller v. Gammie and it misunderstands Loper Bright. In no way does the BIA get to determine when we jettison our Chevron-based precedent—that is only for this court to decide. Agency interpretations aren’t anchors tying down federal courts— only to be cast off with the greatest effort. At all times, we captain our own ship and decide the best reading of the law. This case is of rare importance. As the first to interpret Loper Bright in the immigration context, Lopez will govern hundreds of cases on the Ninth Circuit’s docket. But even more, this case will infect other areas of law—no doubt spreading to our broader administrative-law jurisprudence. We should have taken this opportunity to stop the spread. As it stands, Ninth Circuit judges will be forced to give executive agencies’ statutory interpretations near binding deference all in the name of providing “due respect.” Lopez, 116 F.4th at 1039 (simplified). Respect doesn’t mean abdication. At all times under the APA and the Constitution, judges have the duty to independently evaluate the law. And respect for the Supreme Court means disregarding precedent that directly conflicts with its rulings. LOPEZ V. BONDI 9

Because Lopez misconstrues Loper Bright and defies the separation of powers, I respectfully dissent from the denial of rehearing en banc. I. A Brief History of Judicial Deference In the more than two centuries of Article III courts’ history, deference to executive agencies was weak, bordering on non-existent, except when the executive agency had particular factual expertise or its interpretation illuminated executive practice. But for a brief blip in our history, we never subordinated the best reading of a statute to an executive agency’s interpretation. In short, binding deference to the Executive Branch’s interpretations of statutory text was a judicial invention of the twentieth century. And now, the Supreme Court has now made clear that the blip is over. Founding History Chief Justice John Marshall famously proclaimed, “[i]t is emphatically the province and duty of the judicial department to say what the law is.” Marbury v. Madison, 1 Cranch 137, 177 (1803). This was the foundational principle from the Founding of our Nation. So, for the Founding generation, “executive interpretations of statutes were relevant to judicial determinations only insofar as they embodied understandings made roughly contemporaneously with the statute’s enactment and stably maintained and practiced since that time.” Aditya Bamzai, The Origins of Judicial Deference to Executive Interpretation, 126 Yale L.J. 908, 962 (2017). As James Madison viewed it, courts at the Founding respected agency interpretations out of the belief that “the political branches would ‘liquidate[]’ the 10 LOPEZ V. BONDI

meaning of constitutional provisions” through agency action. Id. at 964 (quoting The Federalist No. 37, at 229 (James Madison) (Clinton Rossiter ed., 1961)). But before the modern era, “there was . . . no general rule of statutory construction requiring ‘deference’ to executive interpretation qua executive interpretation.” Id. at 965. And judicial independence in statutory interpretation continued throughout the 19th century. See id. at 951–956. In Decatur v. Paulding, the Supreme Court reiterated that “[i]f a suit should come before this Court, which involved the construction of any of these laws, the Court certainly would not be bound to adopt the construction given by the head of a department.” 39 U.S. (14 Pet.) 497, 515 (1840). Though the Court could consider the views of an executive department, “if [a court] supposed his decision to be wrong, they would, of course, so pronounce their judgment.” Id. Almost a half-century later, the Supreme Court again reinforced the importance of independent statutory interpretation by holding “[w]hether, if the law were properly before us for consideration, we should be of the same opinion, or of a different opinion [as an executive officer], is of no consequence in the decision of this case.” United States ex rel. Dunlap v. Black, 128 U.S. 40, 48 (1888). Perhaps the “canonical precedent of this era” was American School of Magnetic Healing v. McAnnulty, 187 U.S. 94 (1902). Bamzai, 126 Yale L.J. at 956. There, the Postmaster General determined that an educational institution whose creed stated that “the mind of the human race is largely responsible for its ills” was “fraudulent.” McAnnulty, 187 U.S. at 103. As a result, the Postmaster General barred the school from using the mail service. Id. But since “Congress [had not] intrusted the administration of LOPEZ V. BONDI 11

these statutes wholly to the discretion of the Postmaster General,” the Court held that “the decisions of the officers of the Department upon questions of law do not conclude the courts.” Id. at 108. And assuming that “the evidence before the Postmaster General” could not show a violation of federal law, the Postmaster General’s error was “a pure mistake of law” that could “not bind the courts.” Id. at 109, 111. Thus, when Congress drafted the APA’s judicial review provision in 1946, it was legislating against the backdrop of the “traditional interpretive methodology that had prevailed from the beginning of the Republic until the 1940s” and expecting courts to employ “the customary-and- contemporary canons of construction” in reviewing agency action. Bamzai, 126 Yale L.J. at 987. Beginning with the Founding era and continuing through the passage of the APA, courts consistently and traditionally constructed statutes independently rather than delegating that responsibility to executive agencies. Skidmore And generally, federal courts’ independent role in statutory interpretation remained intact even after the Supreme Court’s Skidmore v. Swift & Co. decision in 1944. 323 U.S. 134 (1944). There, employees of a packing plant sued their employer for overtime damages. Id. at 135. The question before the Court was whether the employees were “working” under federal law when they were told to monitor the plant overnight. Id. at 137. In resolving this question, the Court considered the view of the executive agency tasked with administering the law because it “constitute[s] a body of experience and informed judgment to which courts and litigants may properly resort for guidance.” Id. at 140. To 12 LOPEZ V. BONDI

the Court, “[t]he weight of” the executive agency’s view “in a particular case will depend upon the thoroughness evident in its consideration, the validity of its reasoning, its consistency with earlier and later pronouncements, and all those factors which give it power to persuade[.]” Id. Even so, the Court made clear that, “of course,” the agency’s views were not “conclusive” nor were they “controlling upon the courts by reason of their authority.” Id. at 139, 140. Chevron But judicial independence in statutory interpretation came to an interregnum with the famous (or infamous) Chevron decision in 1984. Under the Chevron framework, courts employed a two-step process for resolving statutory ambiguities. At step one, we determined “whether Congress has directly spoken to the precise question at issue.” Chevron, 467 U.S. at 842. In other words, we would ask “is the statute ambiguous?” If it wasn’t, “that [was] the end of the inquiry.” Loper Bright, 603 U.S. at 379. If the statute was ambiguous, however, we proceeded to step two. There, if a federal agency which administered the statute offered an interpretation “based on a permissible construction of the statute,” we had to accept it as binding. Id. Under this regime of statutory interpretation, our hands were tied so long as an executive interpretation was permissible, even if it were far from the best reading of the statute. Mead By the turn of this century, the Supreme Court began cabining Chevron. In United States v. Mead Corp., the Court limited Chevron’s scope but interpreted Skidmore to reiterate that even when Chevron didn’t bind courts, agency interpretations may have the “power to persuade” if “the regulatory scheme is highly detailed.” 533 U.S. 218, 235 LOPEZ V. BONDI 13

(2001). How persuasive an agency’s interpretation is, however, may depend on its “thoroughness, logic, and expertness, its fit with prior interpretations, and any other sources of weight.” Id. Though the Court retained the Chevron framework, its renewed reliance on Skidmore began to return the judiciary to its traditional role—deciding questions of statutory interpretation independently, but giving “due respect” to persuasive arguments from the Executive. The respect discussed in Mead and Skidmore merely provides another source for courts to resolve statutory ambiguity—the same as any other tool of statutory interpretation. Loper Bright Finally, last year, the judicial experiment with binding executive-agency deference ended. In Loper Bright, the Court eliminated Chevron deference and returned courts to the business of independently interpreting statutes. Loper Bright recognized that the Chevron framework conflicted with the APA, disrespected the separation of powers, destroyed reliance interests, and was unworkable. 603 U.S. at 396, 403–04, 407, 410. So under Loper Bright our task is simple. When we interpret a statute, we start with the plain meaning of the text. See id. at 393. If its meaning is ambiguous, we use “the traditional tools of statutory construction” to resolve that ambiguity. Id. at 401. These tools include the canons of construction, examining the statutory structure, and evaluating the implementation of the statute. See generally Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts (2012). Indeed, these are “the tools courts use every day.” Loper Bright, 603 U.S. at 401. An agency’s interpretation can still influence courts— 14 LOPEZ V. BONDI

particularly when its interpretation “rests on factual premises within the agency’s expertise.” Id. at 402 (simplified). But much like an amicus brief or expert witness, agency interpretations possess only the “‘power to persuade, if lacking power to control.’” See id. (quoting Skidmore, 323 U.S. at 140). Of course, Loper Bright recognized that Congress sometimes confers discretion on executive agencies, such as when Congress “expressly delegate[s] to an agency the authority to give meaning to a particular statutory term,” to “fill up the details of a statutory scheme,” or to “regulate subject to [an imposed] limit[].” Id. at 394–95 (simplified). In such cases, “subject to constitutional limits,” courts should “independently identify and respect such delegations of authority, police the outer statutory boundaries of those delegations, and ensure that agencies exercise their discretion consistent with the APA.” Id. at 404. In the end, legal interpretation “has been, emphatically, the province and duty of the judicial department for at least 221 years.” Id. at 412 (simplified). And the Court has now returned the duty to independently interpret the law to the judiciary. II. Lopez Misreads Loper Bright Lopez conflicts with Loper Bright in at least three ways. First, Lopez favored agency deference rather than the best reading of the statute. Second, Lopez applied agency deference even without any statutory ambiguity. And third, Lopez refused to revisit Chevron-based precedent that is clearly irreconcilable with Loper Bright. Each of these reasons alone warranted reconsideration en banc. LOPEZ V. BONDI 15

A. Lopez Seeks Deference Over Best Reading Immediately, Lopez goes off the rails and errs in interpreting Loper Bright. Rather than apply its independent judgment on the meaning of a “crime involving moral turpitude,” the panel instead sought to determine whether the executive agency’s interpretation of the law “is entitled to respect under Skidmore.” Lopez, 116 F.4th at 1040, 1038 (simplified). Simply because the panel thought the agency interpretation “thorough,” “well-reasoned,” and “consistent” with other authorities, it skipped the need to determine whether the agency’s interpretation was the best reading of the statute. See id. at 1040. This was wrong. Lopez first argued that his petit larceny convictions were not categorically “crimes involving moral turpitude” under § 1227(a)(2)(A)(ii) because the ordinance underlying his conviction did not specify whether the deprivation of property needed to be permanent or temporary. Id. at 1038. Crimes involving only temporary deprivations, like joyriding, are not morally turpitudinous, he argues. The BIA disagrees. To the BIA, a theft offense shows moral turpitude if it includes an intent to deprive “either permanently or under circumstances where the owner’s property rights are substantially eroded.” Id. at 1038 (quoting Matter of Diaz- Lizarraga, 26 I. & N. Dec. 847, 854 (BIA 2016)). Under Loper Bright, our task is to determine the meaning of “crimes involving moral turpitude” and see which view is the best one. This means coming to our own independent conclusion about the best reading of a statute rather than seeing if the interpretation offered by the Executive branch is acceptable. 16 LOPEZ V. BONDI

But rather than interpret independently, the panel first looked at whether the BIA’s view was “entitled to ‘Skidmore deference.’” Lopez, 116 F.4th at 1041. This analysis involved no statutory interpretation. It only reviewed “the degree of the agency’s care, its consistency, formality, and relative expertness, and the persuasiveness of the agency’s position.” Id. at 1039 (simplified). Indeed, letting the cat out of the bag, the panel openly admitted that anytime “the BIA confronted an issue germane to the eventual resolution of the case and resolved it after reasoned consideration,” the agency’s interpretation should be upheld under Skidmore. Id. (simplified). “With that guidance in mind,” the panel deferred to the BIA’s interpretation of the law. First, the panel acknowledged that the BIA had flipped its position. The BIA previously held that “temporary takings” didn’t amount to a crime involving moral turpitude. Id. at 1039. But based on “developments in state criminal law,” the BIA switched to say that even temporary takings are turpitudinous. Id. While the panel found the BIA’s position “inconsistent,” it excused the flip-flopping because the BIA “carefully explained” its “revised interpretation.” Id. The panel then found the BIA’s interpretation “consistent” with the Model Penal Code and the Supreme Court’s definition of an “aggravated felony,” even though neither of those authorities interpreted the meaning of a “crime involving moral turpitude.” Id. Based on this, the panel considered the BIA’s interpretation “thorough,” “well-reasoned” and “consistent with judicial precedent.” Id. at 1040. It ultimately found the BIA interpretation “entitled to ‘Skidmore deference.’” Id. at 1041. Only then did the panel pay any lip service to “[e]xercis[ing] [its] independent evaluation of the statute.” LOPEZ V. BONDI 17

Id. But the panel supported this exercise with zero independent statutory analysis. Id. Instead, the panel largely mimicked the BIA’s reasoning that “the statutory phrase ‘crimes involving moral turpitude’ . . . encompasses offenses that require the government to prove the defendant acted with an intent to permanently deprive an owner’s property or substantially erode the owner's property rights.” Id. (simplified). This doesn’t grapple at all with § 1227(a)(2)(A)(ii)’s text. And deferring to whatever the BIA decides is the best reading of the law at the moment gets the order of operations backwards. Next, based on the BIA’s newest interpretation of the law, the panel overruled Ninth Circuit precedent that conflicted with the government’s favored reading. In Castillo-Cruz, this court granted “substantial deference” to the BIA’s old position that only permanently depriving an owner of property amounts to a crime of moral turpitude. 581 F.3d at 1159. That was no matter to the panel, however. Setting aside Castillo-Cruz, the panel wondered whether the BIA’s new interpretation was “the correct interpretation of ‘moral turpitude.’” Lopez, 116 F.4th at 1041. In deciding that question, the panel provided no independent analysis— rather, it again rested on the “BIA’s thoroughness, persuasive reasoning, and consistency with the longstanding” legal doctrine. Id. Based on that deference to the government’s judgment, the panel “conclude[d] that the BIA’s [new] interpretation is correct.” Id. Sound familiar? This is just Chevron step two by another name. Not only is this an abdication of our judicial role, it’s irreconcilable with Loper Bright. The fundamental problem with the panel’s analysis is its framing of the interpretative question. It led with the question: Is the BIA’s new interpretation of law “entitled to ‘Skidmore deference’”? Id. 18 LOPEZ V. BONDI

at 1041. But that’s the wrong question. Instead, the proper question under Loper Bright is: what’s the best reading of the statute? Sure, in answering that question, we may sometimes look at how the agency handled it. There’s no problem in double-checking our work. But our endeavor isn’t to figure out whether to defer or respect the agency’s reading. It is to “use every tool at [our] disposal to determine the best reading of the statute.” Loper Bright, 603 U.S. at 400; see also Brett M. Kavanaugh, Book Reviews: Fixing Statutory Interpretation, 129 Harv. L. Rev. 2118, 2121 (2016) (“[C]ourts should seek the best reading of the statute by interpreting the words of the statute, taking account of the context of the whole statute, and applying the agreed-upon semantic canons.”). The agency’s reading is useful only when it illuminates the statute’s best reading. So we must always begin with the traditional tools of statutory interpretation. In most cases, that will end the inquiry. Rather than use the BIA’s interpretation as a presumptive starting position or anchor, the panel should have interpreted “crime involving moral turpitude” afresh. 1

1 Admittedly, this is no easy task. “Crimes involving moral turpitude” may sound like a hopelessly ambiguous phrase to the modern ear. But the Supreme Court has noted the phrase is a term of art with “deep roots in the law.” Jordan v. De George, 341 U.S. 223, 227 (1951). The term was used in the Founding era “by early politicians as a catchphrase to sum up traits they deemed undesirable in the new Republic.” Julia Ann Simon-Kerr, Moral Turpitude, 2012 Utah L. Rev. 1001, 1011 (2012). Founding Fathers including Jefferson and Hamilton viewed a good moral code as paramount to the success of the nation. Id. at 1013. So “deception, disloyalty, and the failure to contribute productively to society were the primary traits condemned as moral turpitude in men.” Id. In the early 19th century, moral turpitude also referred to “treachery and fraud” or “oath-breaking,” but seemingly not to “violence in defense of a man’s honor” or “simple assault.” Id. at 1012–15. By the late 19th LOPEZ V. BONDI 19

Further, Skidmore simply cannot bear the weight Lopez puts on it. So-called “Skidmore deference” was never meant to be more than giving due respect to particularly strong arguments from executive agencies. Justice Scalia described Skidmore as standing only for “a trifling statement of the obvious: A judge should take into account the well- considered views of expert observers.” Mead, 533 U.S. at 250 (Scalia, J., dissenting). The same could be said for a well-reasoned and compelling brief from an amicus with a particular expertise in the subject matter. Or, as another judge has analogized, a convincing “decision[] of our sister circuit[]”—“[i]f we are persuaded by another court’s

century, it appears “crimes involving moral turpitude” “was meant to track the traditional honor code, identifying crimes of deception, theft, and perceived sexual perversion, while focusing less on crimes of violence or statutory violations.” Id. at 1040. Seemingly, “the goal was to filter out those immigrants who were perceived to pose the greatest threat to American values.” Id. at 46. Following the phrase’s introduction into our immigration laws, the Supreme Court has wrestled with pinning down a precise meaning for the term. At the very least, it has held that “crimes in which fraud [is] an ingredient” always constitute crimes involving moral turpitude. Jordan, 341 U.S. at 232; see also Craig S. Lerner, Crimes Involving Moral Turpitude, 44 Harv. J. L. & Pub. Pol’y 71, 80 (2021) (noting that post- Civil War political references to “moral turpitude” “usually contemplate[d] fraud, but many times gestur[ed] indistinctly toward the concept of moral impropriety”). Beyond fraud convictions, what else is clearly encompassed within the meaning of this term is a challenging question. But asking what a particular legal term of art means is the type of question the Constitution (and the APA) commends to the judiciary— not an executive agency. So we should have taken this opportunity, post- Loper Bright to evaluate whether theft convictions qualify as crimes involving moral turpitude, independent of what the BIA has required us to adopt in the past. And looking to the historical underpinnings of the early immigration laws and the original meaning of “moral turpitude” in the late-19th to early-20th century should have been the place to start. 20 LOPEZ V. BONDI

reasoning, we adopt it. If we’re not, we don’t.” Dayton Power & Light Co. v. Fed. Energy Regul. Comm’n, 126 F.4th 1107, 1137 (6th Cir. 2025) (Nalbandian, J., concurring). This interpretation of Skidmore finds much support among the scholarly community as well. See, e.g., David J. Barron & Elena Kagan, Chevron’s Nondelegation Doctrine, 2001 Sup. Ct. Rev. 201, 227 n.98 (2001) (suggesting Skidmore amounts to little more “than a court saying ‘we will defer to the agency if we believe the agency is right’”); Peter L. Strauss, “Deference” is Too Confusing— Let’s Call Them “Chevron Space” and “Skidmore Weight,” 112 Colum. L. Rev. 1143, 1145 (2012) (Skidmore stands for “the possibility that an agency’s view on a given statutory question may in itself warrant respect by judges who themselves have ultimate interpretive authority”); Adrian Vermeule, Deference and Due Process, 129 Harv. L. Rev. 1890, 1901 (2016) (“Skidmore just describes the attitude of any minimally sensible decisionmaker, who listens to any relevant arguments of well-informed parties when deciding what to do.”). So by justifying its analytical abdication solely based on a doctrine that says nothing more than “judges should independently consider especially persuasive arguments” the panel opinion collapses in on itself. And what’s particularly troublesome is that this case is not even one in which the BIA possesses specialized knowledge that would make its interpretation helpful under Skidmore. Loper Bright, 603 U.S. at 401 (“When the agency has no comparative expertise in resolving a regulatory ambiguity, Congress presumably would not grant it [interpretive] authority.”) (simplified). At most, an agency’s interpretation is “especially informative” only “to the extent it rests on factual premises within the agency’s expertise.” Id. at 402 (simplified). At bottom, this is a question of pure LOPEZ V. BONDI 21

statutory interpretation—what did Congress mean by “crimes involving moral turpitude”? This is a legal term of art. See Jordan v. De George, 341 U.S. 223, 227 (1951). We as judges are best positioned to answer this inquiry. The BIA is not an expert on history, criminal law, or morality, nor does this question implicate a “technical matter.” See Loper Bright, 603 U.S. at 402. After all, “agencies have no special competence in resolving statutory ambiguities. Courts do.” Id. at 400–01. So this case is not one in which an agency’s interpretation of the statute is useful. 2 Even more, the panel minimizes the BIA’s shifting positions. Loper Bright warned that one of the main problems with Chevron was that it gave an agency license “to change positions as much as it likes . . . even when Congress has given them no power to do so.” Id. at 411. Chevron thus “destroy[ed]” reliance interests because it didn’t “produce readily foreseeable outcomes.” Id. at 410. But we open the door to instability yet again. Congress did not authorize the BIA to flip its position in 2016—it passed no new legislation altering the definition of crimes involving moral turpitude. So we greenlight agency flip flops as long as the agency “carefully explain[s]” its reversal. Lopez, 116 F.4th at 1040. All this points to why Lopez’s framework for analyzing legal questions violates Loper Bright.

2 Nor is this a case where the agency claims an express delegation of authority from Congress, see Lopez, 116 F.4th at 1039 n.2, or “controlling” authority under 8 U.S.C. § 1103(a)(1). 22 LOPEZ V. BONDI

B. Lopez Grants Deference with No Ambiguity Illuminating how far the panel majority misread Loper Bright, it granted the government deference when there was not even a statutory ambiguity. Lopez argued that he wasn’t removable because his convictions were not pardonable offenses. Under the law, removal for committing crimes involving moral turpitude “shall not apply in the case of an alien with respect to a criminal conviction if the alien subsequent to the criminal conviction has been granted a full and unconditional pardon by the President of the United States or by the Governor of any of the several States.” 8 U.S.C. § 1227(a)(2)(A)(vi). Lopez essentially argued for the atextual, negative implication of the provision—asserting that he can’t be removed under § 1227 because he cannot obtain a pardon for his petty municipal offenses. Courts have never been permitted to look to agency interpretations when Congress has spoken directly to the issue and has provided unambiguous statutory language. As the Court made clear, “[i]f . . . congressional intent is ‘clear,’ that is the end of the inquiry.” Loper Bright, 603 U.S. at 379 (simplified). Even under Chevron, courts “must give effect to the unambiguously expressed intent of Congress.” Chevron, 467 U.S. at 843. It’s only when the meaning of a statute is unclear that an agency’s interpretation “may help inform that inquiry.” Loper Bright, 603 U.S. at 413. And § 1227(a)(2)(A)(vi) isn’t ambiguous. Nowhere in the statute does it say that aliens must get a chance to receive a pardon before being removed. As the panel noted, “[w]hen the language is plain, we have no right to insert words and phrases, so as to incorporate in the statute a new and distinct provision.” Lopez, 116 F.4th at 1043 (quoting United States LOPEZ V. BONDI 23

v. Temple, 105 U.S. 97, 99 (1881)). “If Congress had intended the pardon waiver to include crimes for which no pardon was available,” the panel observed, “it could easily have said so.” Id. So the panel correctly ruled “that the plain language of the pardon waiver precludes Lopez’s interpretation.” Id. at 1044. That should have been the end of it. That’s why it was baffling to see that the panel only conducted this analysis after first reviewing how the BIA addressed the pardon waiver issue. In the panel’s view, “pursuant to Loper Bright Enterprises,” it had to “construe the statute independently” “[a]gainst” the BIA’s interpretative “background.” Id. at 1043. Despite this approach having no basis in Supreme Court precedent, the panel spilled a lot of ink tracing the BIA’s inconsistent interpretation of the provision over time. See id. Ultimately, the panel conducted some sort of hybrid analysis— “afford[ing]” the BIA’s newest interpretation “Skidmore deference” while also “independently conclud[ing], based on [its] own statutory analysis,” that the pardon waiver theory conflicts with the law. Id. at 1044. But the panel should have never even consulted the BIA’s interpretation. It should have started with the statute’s text and recognized that its meaning was unambiguous. From there, it should have ended the analysis. There was simply no place to grant the agency any deference—and there never has been. See Loper Bright, 603 U.S. at 385–86. Relying on the BIA’s interpretation here isn’t just wrong— it establishes bad precedent in our circuit, requiring panels to analyze even unambiguous statutes through the lens of agencies in the name of Skidmore “respect.” 24 LOPEZ V. BONDI

C. Lopez Precludes Revisiting Chevron-Based Precedent The Lopez panel’s handling of circuit precedent might be the most puzzling aspect of the opinion. Lopez argued that his convictions were part of a two-week crime spree that formed a “single scheme of criminal misconduct.” See Lopez, 116 F.4th at 1037, 1045. Recall that removal under this provision of law requires “two or more” convictions “not arising out of a single scheme of criminal misconduct.” 8 U.S.C. § 1227(a)(2)(A)(ii). So what does it mean for criminal misconduct to arise out of a “single scheme”? Certainly not what the BIA claims. Nearly sixty years ago, the Ninth Circuit clarified that “single scheme of criminal misconduct” does not mean “single criminal act.” See Wood, 266 F.2d at 830 (observing that courts cannot interpret this language “as ‘crimes arising out of a single act of criminal misconduct,’” because “this is not what the statute says”). This understanding is undoubtedly correct. Consider these definitions of “scheme” from around the time of the statute’s enactment:

• A “plan or program of something to be done”; a “project,” especially a “crafty, unethical project”; or a “systematic plan” comprising a “combination of thoughts, theories, or the like, connected and adjusted by design.” Webster’s 2d New Int’l Dictionary 2234 (1934). • “A design or plan formed to accomplish some purpose—a system.” Black’s Law Dictionary, 1511 (4th ed. 1951). LOPEZ V. BONDI 25

• “[A] systematic plan for attaining some object . . . an orderly combination of things on a definite plan; system . . . a plot; underhand intrigue.” Webster’s New World Dictionary, 662–63 (1956). • A “program of action; plan . . . plan of action to attain some end, often one characterized by self-seeking or intriguing.” Thorndike-Barnhart Comprehensive Desk Dictionary, 691 (Clarence L. Barnhart, ed. 1956).

These definitions all revolve around the creation of a “plan” to accomplish some end, including one that consists of “interconnected elements”—a concept that is simply “not captured by the BIA’s construction.” Szonyi, 942 F.3d at 887 (Collins, J., dissenting from the denial of rehearing en banc). So necessarily, § 1227(a)(2)(A)(ii) specifically addresses instances where a series of crimes were linked as part of a single scheme. See id. (explaining that two separate emails sent days apart in furtherance of a single scheme to defraud would, contrary to the BIA’s interpretation, naturally arise out of a single scheme of criminal misconduct). But rather than interpret § 1227(a)(2)(A)(ii) according to its plain text, the panel instead felt bound by the BIA’s contrary interpretation, which we granted deference to in Szonyi. Szonyi applied Brand X deference to the BIA’s view that § 1227(a)(2)(A)(ii) is satisfied even if multiple convictions were “part of an overall plan of criminal misconduct” if each crime was “complete, individual, and distinct.” Id. at 891, 892 (quoting Matter of Adetiba, 20 I. & N. Dec. 506, 509 (B.I.A. 1992)). As Judge Collins 26 LOPEZ V. BONDI

explained, the BIA’s interpretation was simply unreasonable. Id. at 875–77 (Collins, J., dissenting from the denial of rehearing en banc). No matter. The panel viewed Szonyi as untouchable. The opinion simply observed that “the Supreme Court has instructed that Loper Bright Enterprises does not ‘call into question prior cases that relied on the Chevron framework.’” Lopez, 116 F.4th at 1045 (quoting Loper Bright, 603 U.S. at 412). So the panel thought itself bound by Szonyi unless the BIA “promulgate[s] a new interpretation of the statute to prompt us to reconsider our precedent.” Id. The panel then declared, with little explanation, that Loper Bright is no “intervening higher authority” that is “‘clearly irreconcilable with’ Szonyi” under Miller v. Gammie. Id. (simplified). But the panel got it wrong for at least two reasons. First, Loper Bright was a clear “intervening United States Supreme Court decision” requiring us to reevaluate our holding in Szonyi. See Miller, 335 F.3d at 892. Loper Bright’s statement about “prior cases” refers to its prior cases—not ours. Second, the panel’s declaration that only a new agency interpretation permits us to revisit Chevron-based precedent lets the Executive branch—not the courts—dictate the interpretation of the law. Thus, because the government no longer agrees with Castillo-Cruz, the panel felt compelled to overrule it. But because the government still stands behind Szonyi, it’s somehow sacrosanct. This defies sound logic— and worse still, it resurrects Chevron. I address each concern in more depth. LOPEZ V. BONDI 27

1. First, the panel impermissibly entrenched the BIA’s reading of the law by misreading Loper Bright. Before getting to that, it helps to remember what Miller v. Gammie decided. In that case, we established the standard for deciding when a three-judge panel should disregard precedent due to intervening legal authority. We agreed that, where intervening authority is “clearly irreconcilable” with a circuit precedent, “a three-judge panel of this court and district courts should consider themselves bound by the intervening higher authority and reject the prior opinion of this court as having been effectively overruled.” 335 F.3d at 900 (emphasis added). “Should” implies that rejecting irreconcilable circuit precedent is a mandatory duty—not a discretionary call. And the “issues decided by the higher court need not be identical in order to be controlling.” Id. Rather, the higher authority “must have undercut the theory or reasoning underlying the prior circuit precedent in such a way that the cases are clearly irreconcilable.” Id.; see also Tapia Coria v. Garland, 114 F.4th 994, 1008–09 (9th Cir. 2024) (holding circuit precedent no longer binding, even though intervening Supreme Court decision “did not directly address” the same legal issue, because the Supreme Court decision “direct[ed] a completely different approach” to the issue). Now, return to Loper Bright. There, the Court was crystal clear that we must no longer grant deference to an executive agency’s interpretation of the law under Chevron. So any Ninth Circuit precedent that relies on Chevron to defer to an agency’s interpretation of the law is “clearly irreconcilable” with Loper Bright. And, of course, a Supreme Court decision is “higher authority.” That alone 28 LOPEZ V. BONDI

should have settled the matter. Faithfully applying Loper Bright and Miller v. Gammie dictates overruling Szonyi. But Lopez relied on one sentence in Loper Bright to gum up our precedent. In Loper Bright, the Court said, “we do not call into question prior cases that relied on the Chevron framework.” Loper Bright, 603 U.S. at 412. But the panel overread that sentence to preclude lower courts from revisiting lower-court precedents that relied on Chevron. That’s just wrong. To begin, Loper Bright doesn’t direct lower courts to preserve lower-court precedent relying on Chevron. More importantly, it misunderstands the context of the Court’s statement. In context, it’s clear that the Court was expressly discussing its precedent. After making that statement, the Court continued,

[t]he holdings of those cases that specific agency actions are lawful—including the Clean Air Act holding of Chevron itself—are still subject to statutory stare decisis despite our change in interpretive methodology. Mere reliance on Chevron cannot constitute a “‘special justification’” for overruling such a holding, because to say a precedent relied on Chevron is, at best, “just an argument that the precedent was wrongly decided.” That is not enough to justify overruling a statutory precedent.

Id. (simplified). Thus, the Court discusses its own precedent and clarifies that it was not overruling its own prior agency deference precedents in one fell swoop. See id. LOPEZ V. BONDI 29

That statement in Loper Bright makes sense given the distinction between vertical and horizontal stare decisis. Overturning Supreme Court precedent requires different considerations than overturning lower-court precedent. As judges on this court have previously recognized, “the Supreme Court’s obligation to follow stare decisis” differs from that of circuit courts. See United States v. Aguon, 851 F.2d 1158, 1173 (9th Cir. 1988) (en banc) (Reinhardt, J., concurring); see also Critical Mass Energy Project v. Nuclear Regul. Comm’n, 975 F.2d 871, 876 (D.C. Cir. 1992) (en banc) (“Circuit courts of appeal, of course, play a different role in the federal system than the Supreme Court, and this is reflected in certain differences in the manner in which the principle of stare decisis is applied to circuit precedent.”). When it comes to the high court, overturning precedent “require[s] ‘special justification,’ not just an argument that the precedent was wrongly decided.” Halliburton Co. v. Erica P. John Fund, Inc., 573 U.S. 258, 266 (2014). Not so at the lower-court level. Here, whenever the Supreme Court “undercut[s] the theory or reasoning underlying . . . prior circuit precedent in such a way that the cases are clearly irreconcilable[,]” our precedent is no longer binding. Miller, 335 F.3d at 900. That’s because our goal of “preserv[ing] the consistency of circuit law . . . must not be pursued at the expense of creating an inconsistency between our circuit decisions and the reasoning of . . . a decision of a court of last resort.” Id. So the Court’s invocation of its horizontal stare decisis doctrine in Loper Bright in no way precludes us from fixing our precedent premised on faulty Chevron grounds. Indeed, that we treat our circuit precedent differently than the Supreme Court treats its precedent makes intuitive 30 LOPEZ V. BONDI

sense. First, the Ninth Circuit only covers nine States within the country. While our jurisdiction represents an outsized portion of the country, we don’t preside over the entire Nation like the Supreme Court. Shifting the law nationwide has far greater potential to disturb reliance interests than shifting the law just within our circuit. Second, as few as two judges can set the law for the entire circuit (as was the case for parts of Lopez). And given our enormous caseload, these cases can establish circuit law with little attention or fanfare. On the other hand, the Supreme Court hears fewer cases but with the highest salience—and so the Supreme Court can be more deliberate in announcing its precedent. Further, that this is a question of statutory stare decisis underscores the need to revisit our erroneous precedent. Two primary justifications exist for adherence to statutory stare decisis—neither of which applies strongly to lower court decisions. One is congressional acquiescence. If Congress disagreed with the Court’s interpretation of a statute, it could amend the law, and congressional silence on the matter could be read “as approval of that interpretation.” Amy Coney Barrett, Statutory Stare Decisis in the Courts of Appeals, 73 Geo. Wash. L. Rev., 317, 317 (2005). The second reason is that if the Court were to change its “statutory interpretations from case to case, Congress would have less reason to exercise its responsibility to correct statutes that are thought to be unwise or unfair.” Neal v. United States, 516 U.S. 284, 296 (1996). But these justifications hold less weight when it comes to lower courts. That’s because it’s unlikely that Congress would respond to lower court orders. An empirical analysis suggests “that Congress is generally unaware of circuit-level statutory interpretations,” with one study finding that over eight years Congress only responded to .002% of all courts LOPEZ V. BONDI 31

of appeals decisions. Barrett, 73 Geo. Wash. L. Rev. at 331 (citing Stefanie A. Lindquist & David A. Yalof, Congressional Responses to Federal Circuit Court Decisions, 85 Judicature 61 (2001)). Another study found “impressive congressional activity in connection with Supreme Court decisions, . . . [but] an unimpressive knowledge of and response to the far more numerous lower federal court statutory interpretation decisions.” William N. Eskridge, Jr., Overriding Supreme Court Statutory Interpretation Decisions, 101 Yale L.J. 331, 416 (1991). These studies make sense; while the Supreme Court’s interpretation of a law “binds the whole nation,” circuit courts only have jurisdiction over a limited geographic region. Barrett, 73 Geo. Wash. L. Rev. at 344. And “[w]hen the First Circuit interprets a statute, . . . what incentive does a senator from California have to introduce or support legislation to override a judicial opinion that affects a small portion of the East Coast and Puerto Rico?” Id. So, that this is a question of statutory stare decisis doesn’t counsel against revisiting our precedent. All this is to say that neither Loper Bright nor our stare decisis factors precluded us from revisiting and overruling Chevron-based precedent. The panel erred by failing to do so. 2. My second concern is the novel proposition that executive agencies dictate when we may revisit our prior holdings. Nothing in Loper Bright nor Miller v. Gammie looks to agency preferences before deciding if we may jettison circuit precedent irreconcilable with the Supreme Court’s own. But that’s what Lopez said and did—it took the lead from the BIA to decide the fate of our precedent. 32 LOPEZ V. BONDI

Lopez expressly stated, “the BIA has not promulgated a new interpretation of the statute to prompt us to reconsider our precedent.” Lopez, 116 F.4th at 1045. Presumably, then, the BIA’s new interpretation of “crimes involving moral turpitude” allowed the panel to revisit Castillo-Cruz. That’s why Szonyi remains but Castillo-Cruz gets thrown out. The result of Lopez is that whether the Ninth Circuit is bound by Chevron-based precedent entirely depends on an executive agency’s whims. If the agency maintains the same position as our prior precedent, then a three-judge panel must follow the precedent even though it’s decided on deference grounds overruled by Loper Bright. By contrast, if the agency doesn’t like the position to which this court deferred, the agency can simply change positions, which frees us from its prior decision and allows us to adopt our own, independent interpretation of the statute. All this is just a return to deference to the Executive branch on the interpretation of law. If judicial independence means anything, it must mean that we can decide when the law is irreconcilable with Supreme Court precedent. None of this adheres to Loper Bright. III. Loper Bright represented one of the most dramatic changes in how courts should do statutory interpretation. Even so, Lopez acts as if nothing has changed. In the name of giving respect to executive agencies, the panel disrespects the Supreme Court. We can’t do that. Guided by the clear command of Loper Bright, we should have taken this case en banc and reclaimed our judicial role. I respectfully dissent from the denial of rehearing en banc.