BERZON, Circuit Judge:

ORDER

The opinion filed February 20, 2014, and published at 743 F.3d 666, is withdrawn. The superseding opinion shall be filed concurrently with this order.

Further petitions for rehearing or petitions for rehearing en banc shall be allowed in the above-captioned matter. See G.O. 5.3(a).

OPINION

The government appeals the dismissal of Xochitl Garcia-Santana's indictment for unlawful reentry in violation of 8 U.S.C. § 1326. The district court determined that Garcia's prior removal order was constitutionally inadequate because Garcia was denied her right to seek discretionary relief from removal. We affirm. In doing so, we hold that the generic definition of "conspiracy" under the Immigration and Nationality Act ("INA"), 8 U.S.C. § 1101(a)(43)(U), includes proof of an overt act in furtherance of the conspiracy.

I.

In 2002, Garcia pleaded guilty to "conspiracy to commit the crime of burglary" in violation of Nev.Rev.Stat. §§ 199.480, 205.060(1). A Nevada court found her guilty and sentenced her to a suspended twelve-month term in county jail.

Just over two weeks later, a Deciding Service Officer of the Immigration and Naturalization Service, proceeding under the summary removal procedures codified at 8 U.S.C. § 1228(b), ordered Garcia removed as an undocumented alien "convicted of an aggravated felony pursuant to ... 8 U.S.C. [§] 1227(a)(2)(A)(iii)." The Deciding Service Officer determined that Garcia was subject to "a final conviction of an aggravated felony as defined in ... 8 U.S.C. 1101(a)(43), and [was] ineligible for any relief from removal that the Attorney General may grant in an exercise of discretion." She was removed.

In 2009, Garcia unlawfully reentered the United States. Some years later, Nevada law enforcement officials notified U.S. Immigration and Customs Enforcement ("ICE") that they had booked Garcia, a previously removed alien, into a local detention center. ICE officials subsequently took Garcia into custody at her home.

A grand jury indicted Garcia on the charge that she was a previously removed alien found unlawfully in the United States, in violation of 8 U.S.C. § 1326. She moved to dismiss the indictment, arguing that her previous removal order was fundamentally unfair. The Deciding Service Officer erred, she asserted, in finding that her previous conviction qualified as an "aggravated felony" that rendered her ineligible for all discretionary relief. Denying her an opportunity to seek such relief, she concluded, constituted a violation of due process.

The district court denied Garcia's motion, ruling that conspiracy to commit the crime of burglary under Nevada law constituted an aggravated felony, so she did not qualify for any discretionary relief. Upon reconsideration, however, the court struck its order denying Garcia's motion to dismiss for the constitutional inadequacy of her previous removal order. Instead, the court granted Garcia's previous request "upon the grounds contained in Defendant['s] motion."

This appeal followed.

II.

The Due Process Clause guarantees an individual charged with illegal reentry, 8 U.S.C. § 1326, the opportunity to challenge "a prior [removal] that underlies [the] criminal charge, where the prior [removal] proceeding effectively eliminated the right of the alien to obtain judicial review." United States v. Arias-Ordonez, 597 F.3d 972, 976 (9th Cir.2010) (citing United States v. Mendoza-Lopez, 481 U.S. 828, 107 S.Ct. 2148, 95 L.Ed.2d 772 (1987)). Section 1326(d) codifies this principle. See id. It authorizes collateral attack on three conditions: (1) that the defendant exhausted available administrative remedies; (2) that the removal proceedings "deprived the alien of the opportunity for judicial review"; and (3) that the removal order "was fundamentally unfair." 8 U.S.C. § 1326(d). Removal is "fundamentally unfair," in turn, if "`(1) [a defendant's] due process rights were violated by defects in his underlying [removal] proceeding, and (2) he suffered prejudice as a result of the defects.'" United States v. Ubaldo-Figueroa, 364 F.3d 1042, 1048 (9th Cir.2004) (first alteration in original) (quoting United States v. Zarate-Martinez, 133 F.3d 1194, 1197 (9th Cir.1998), overruled on other grounds as recognized in United States v. Ballesteros-Ruiz, 319 F.3d 1101, 1105 (9th Cir.2003)).

An immigration official's failure to advise an alien of his apparent eligibility for relief from removal, including voluntary departure, violates his due process rights. See, e.g., United States v. Melendez-Castro, 671 F.3d 950, 954 (9th Cir. 2012) (per curiam); United States v. Lopez-Velasquez, 629 F.3d 894, 897 (9th Cir. 2010) (en banc). An alien who has been convicted of an aggravated felony is not eligible for voluntary departure in lieu of removal. See 8 U.S.C. § 1229c(a)(1); United States v. Vidal-Mendoza, 705 F.3d 1012, 1014 n. 2 (9th Cir.2013). Garcia's prior removal order stated that she was "ineligible for any relief," because she had previously been convicted of an aggravated felony. The government challenges the grant of collateral relief only on the ground that Garcia's conviction for burglary conspiracy qualifies as an aggravated felony, contrary to Garcia's contention and the district court's ruling.1 "Aggravated felony" is defined to include "a theft offense... or burglary offense for which the term of imprisonment [is] at least one year," 8 U.S.C. § 1101(a)(43)(G), or a "conspiracy to commit an offense described in" § 1101(a)(43), 8 U.S.C. § 1101(a)(43)(U), which includes a "theft offense ... or burglary."

III.

To determine whether an offense is an aggravated felony, we "use the categorical and modified categorical approaches of Taylor v. United States, 495 U.S. 575, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990), and Shepard v. United States, 544 U.S. 13, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005)." Hernandez-Cruz v. Holder, 651 F.3d 1094, 1100 (9th Cir.2011). Under the categorical approach, "we look `not to the facts of the particular prior case,' but instead to whether `the state statute defining the crime of conviction' categorically fits within the `generic' federal definition of a corresponding aggravated felony." Moncrieffe v. Holder, ___ U.S.___, 133 S.Ct. 1678, 1684, 185 L.Ed.2d 727 (2013) (quoting Gonzales v. Duenas-Alvarez, 549 U.S. 183, 186, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007)). The "generic" definition of an offense is determined by "the contemporary usage of the term." Taylor, 495 U.S. at 592, 110 S.Ct. 2143. "[A] state offense is a categorical match [with a generic federal offense] only if a conviction of the state offense "`necessarily" involved ... facts equating to [the] generic [federal offense].'" Moncrieffe, 133 S.Ct. at 1684 (some alterations in original) (quoting Shepard v. United States, 544 U.S. 13, 24, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005) (plurality opinion)). That is, "an offense is an aggravated felony if `the full range of conduct covered by the [state criminal statute] falls within the meaning' of the relevant definition of an aggravated felony." Ngaeth v. Mukasey, 545 F.3d 796, 800 (9th Cir.2008) (per curiam) (quoting Penuliar v. Mukasey, 528 F.3d 603, 608 (9th Cir.2008), abrogated on other grounds as recognized in United States v. Martinez, 771 F.3d 672, 677-78 (9th Cir.2014)). By contrast, where the state statute of conviction "sweeps more broadly than the generic crime, a conviction under the law cannot count as an [aggravated felony], even if the defendant actually committed the offense in its generic form." Descamps v. United States, ___ U.S. ___, 133 S.Ct. 2276, 2283, 186 L.Ed.2d 438 (2013).

"Nevada law defines a conspiracy as `an agreement between two or more persons for an unlawful purpose.'" Bolden v. State, 121 Nev. 908, 124 P.3d 191, 194 (2005) (quoting Doyle v. State, 112 Nev. 879, 921 P.2d 901, 911 (1996), overruled on other grounds by Kaczmarek v. State, 120 Nev. 314, 91 P.3d 16 (2005)); see also Nev.Rev.Stat. § 199.480. Conviction of a conspiracy in Nevada requires no proof "that any overt act was done in pursuance of such unlawful conspiracy or combination." Nev.Rev.Stat. § 199.490.

For reasons we shall explain shortly, we are convinced that, applying the methodology prescribed by the Supreme Court for defining generic offenses for categorical purposes, the generic federal definition of conspiracy, codified at 8 U.S.C. § 1101(a)(43)(U), conditions conviction on performance of an overt act in pursuit of the conspiratorial objective.2 Because Nevada's conspiracy statute criminalizes a broader range of conduct than the properly determined generic definition of conspiracy, Garcia's conviction does not qualify as an aggravated felony.3

IV.

A.

"[C]ontemporary usage of [a] term" governs its generic definition under the categorical approach. Taylor, 495 U.S. at 592, 110 S.Ct. 2143. To identify that "contemporary usage," we survey the definitions codified in state and federal statutes, adopted by the Model Penal Code ("MPC"), and endorsed by scholarly commentary. See, e.g., United States v. Esparza-Herrera, 557 F.3d 1019, 1023 (9th Cir.2009) (per curiam).

i. The generic definition of an offense "roughly correspond[s] to the definitions of [the offense] in a majority of the States' criminal codes." Taylor, 495 U.S. at 589, 110 S.Ct. 2143. A survey of state conspiracy statutes reveals that the vast majority demand an overt act to sustain conviction. By our count, thirty-six states do so; if the District of Columbia, Guam, Puerto Rico, and the Virgin Islands are included, then the tally rises to forty of fifty-four jurisdictions.4

Such a great predominance of jurisdictions is more than sufficient to establish the generic federal definition of a crime. We have held the agreement of thirty-three jurisdictions qualifies as sufficient consensus to establish the generic definition of a crime. See Esparza-Herrera, 557 F.3d at 1025. Here, the even more widespread agreement among jurisdictions on an overt act requirement for the general crime of conspiracy indicates that conviction for generic conspiracy requires an overt act.

The federal government's general conspiracy statute, which criminalizes conspiracies "to commit any offense against the United States, or to defraud the United States," also requires an overt act. 18 U.S.C. § 371. Parallel federal crimes are probative, but not independently determinative, of the contemporary, generic definition of an offense. See United States v. Medina-Villa, 567 F.3d 507, 515-16 (9th Cir.2009).

ii. Taylor, which first established the proper mode of analysis in this area of law, used both the MPC and a scholarly treatise — Wayne R. LaFave & Austin W. Scott, Substantive Criminal Law (1st ed.1986) — as aids in its survey of generic "burglary." Taylor, 495 U.S. at 598, 110 S.Ct. 2143. These two sources agree that "conspiracy" to commit an offense now requires proof of an overt act, and so confirm the results of our survey of contemporary state and federal statutes.

The MPC conditions conviction for general conspiracy on proof of "an overt act in pursuance of [the] conspiracy ... done by [the defendant] or by a person with whom he conspired," unless the conspiracy concerns the commission of a first or second degree felony. Model Penal Code § 5.03(5). Because the MPC defines burglary as a felony in the third degree unless particular, narrow conditions are met,5 conviction for conspiracy to commit burglary typically requires proof of an overt act.6

The oft-cited treatise, Substantive Criminal Law, also supports an overt act requirement.7 As that treatise observes, "most of the states now require that an overt act in furtherance of the plan be proven for all or specified conspiratorial objectives." 2 LaFave, supra, § 12.2. The treatise goes on to observe that the overt-act requirement is in some instances treated as "part of the offense" and in others as "merely an element of proof." Id. Under the categorical approach, this distinction does not matter; "`a "constituent part" of the offense [that] must be proved by the prosecution in every case to sustain a conviction under a given statute[,]'" is an element of the crime for purposes of categorical analysis. United States v. Beltran-Munguia, 489 F.3d 1042, 1045 (9th Cir.2007) (emphasis and alteration in original) (quoting United States v. Hasan, 983 F.2d 150, 151 (9th Cir.1992) (per curiam)). The scholarly assessment thus confirms that generic conspiracy requires proof of an overt act.

The agreement of a majority of states, the federal general conspiracy statute, the MPC, and scholarly commentary reflects the importance of an overt-act requirement to contemporary criminal jurisprudence. At common law, conviction for conspiracy required no proof of an overt act. See, e.g., Whitfield v. United States, 543 U.S. 209, 213-14, 125 S.Ct. 687, 160 L.Ed.2d 611 (2005). Instead, agreement was seen as the "essence" of conspiracy, Iannelli v. United States, 420 U.S. 770, 777, 95 S.Ct. 1284, 43 L.Ed.2d 616 (1975), and "`an evil in itself, independently of any other evil [the criminal agreement] seeks to accomplish[,]'" id. at 779, 95 S.Ct. 1284 (quoting Dennis v. United States, 341 U.S. 494, 573, 71 S.Ct. 857, 95 L.Ed. 1137 (1951) (Jackson, J., concurring)). The "evil" of a conspiracy was understood to lie in the tendency of "`[c]oncerted action both [to] increase[ ] the likelihood that the criminal object will be successfully attained and [to] decrease[ ] the probability that the individuals involved will depart from their path of criminality.'" Id. at 778, 95 S.Ct. 1284 (quoting Callanan v. United States, 364 U.S. 587, 593, 81 S.Ct. 321, 5 L.Ed.2d 312 (1961)).

The move toward requiring proof of an overt act was but one manifestation of a larger shift in legal thought concerning the general crime of conspiracy, as jurists and scholars began to "view with disfavor attempts to broaden the already pervasive and wide-sweeping nets of conspiracy prosecutions." Grunewald v. United States, 353 U.S. 391, 404, 77 S.Ct. 963, 1 L.Ed.2d 931 (1957); see also 2 LaFave, supra, § 12.1 (describing some common criticisms of conspiracy). A range of concerns informed that jurisprudential disfavor, among them the observation that "the minimum of proof required to establish conspiracy is extremely low," Krulewitch v. United States, 336 U.S. 440, 452, 69 S.Ct. 716, 93 L.Ed. 790 (1949) (Jackson, J., concurring), and recognition that the procedural rules attached to conspiracy allegations make convictions easier to obtain than for substantive crimes, id. at 452-54, 69 S.Ct. 716. Proof of an overt act is often the only external evidence of a crime "predominantly mental in composition." Krulewitch, 336 U.S. at 447-48, 69 S.Ct. 716 (internal quotation marks omitted). For this reason, "[t]he function of the overt act in a conspiracy prosecution is simply to manifest that the conspiracy is at work and is neither a project still resting solely in the minds of the conspirators nor a fully completed operation no longer in existence." Yates v. United States, 354 U.S. 298, 334, 77 S.Ct. 1064, 1 L.Ed.2d 1356 (1957) (internal quotation marks and citation omitted), overruled on other grounds by Burks v. United States, 437 U.S. 1, 98 S.Ct. 2141, 57 L.Ed.2d 1 (1978); see generally Peter Buscemi, Note, Conspiracy: Statutory Reform Since the Model Penal Code, 75 Colum. L.Rev. 1122, 1153-59 (1975) (tracking legislative revision of the common law of conspiracy to include an overt-act requirement and outlining the motivations for reform). The contemporary overt act requirement thus developed to guard against the punishment of evil intent alone, and to assure that a criminal agreement actually existed.

As all the indicia we have been instructed to use under Taylor and its progeny to determine the elements of the general crime of conspiracy point toward an overt act element, we conclude that such an overt act is an element of the generic definition of conspiracy.

B.

The government maintains, however, that the reference to "conspiracy" in 8 U.S.C. § 1101(a)(43)(U) incorporates only the common-law definition of that term, without the contemporary, widely adopted overt-act safeguard. For support, the government cites a line of Supreme Court cases interpreting "conspiracy" as used in specific federal criminal statutes, rather than in the generic federal conspiracy statute. Those offense-specific cases rest on a "`settled principle of statutory construction that, absent contrary indications, Congress intends to adopt the common law definition of statutory terms,'" and hold that, absent express language to the contrary, a federal statute establishing a specific federal conspiracy offense does not require an overt act, only an agreement. Whitfield, 543 U.S. at 213, 125 S.Ct. 687 (emphasis added) (quoting United States v. Shabani, 513 U.S. 10, 13-14, 115 S.Ct. 382, 130 L.Ed.2d 225 (1994)) (conspiracy to commit money laundering in violation of 18 U.S.C. § 1956(h)); see also Shabani, 513 U.S. at 13-14, 115 S.Ct. 382 (conspiracy to distribute cocaine in violation of 21 U.S.C. § 846); Singer v. United States, 323 U.S. 338, 340, 65 S.Ct. 282, 89 L.Ed. 285 (1945) (conspiracy to aid another in evading service in the armed forces); Nash v. United States, 229 U.S. 373, 378, 33 S.Ct. 780, 57 L.Ed. 1232 (1913) (Sherman Act conspiracy).

The government's reliance on these decisions tracks the BIA's reasoning in a precedential opinion, which relied on the same line of Supreme Court cases. See In re Richardson, 25 I. & N. Dec. 226, 228 (BIA 2010). On the basis of those cases, Richardson interpreted "conspiracy" under § 1101(a)(43)(U) as referring to the common-law definition, and thus as omitting any overt-act requirement. Id. at 230. We cannot accept this interpretation.

i. The cases cited by the government, and by Richardson, interpret specific penal statutes, each of which directly imposes criminal liability for particular acts, as well as conspiracy to carry out those same acts. Whitfield recognized this limitation, explaining that the line of cases on which the government relies established "the governing rule" to determine whether an overt act is a required element of a particular "conspiracy offense." Whitfield, 543 U.S. at 214, 125 S.Ct. 687 (emphasis added). That is, those cases explain how courts should interpret federal statutes criminalizing conspiracies.

The INA is quite different than those statutes. It is not a "conspiracy offense." It defines "aggravated felonies" for the purpose not of defining and penalizing criminal conduct, but of assigning various immigration consequences to prior convictions. See, e.g., United States v. Corona-Sanchez, 291 F.3d 1201, 1209 n. 8 (9th Cir.2002) (en banc) (listing the various uses of the "aggravated felony" concept in the INA).8 Such collateral consequences attach to convictions from all jurisdictions, not merely to federal convictions.

In applying the Taylor approach, we presume that the statute employs "uniform, categorical definitions to capture all offenses of a certain [type] ... regardless of technical definitions and labels under state law." Taylor, 495 U.S. at 590, 110 S.Ct. 2143. The definitions of aggravated felonies codified at 8 U.S.C. § 1101(a)(43) are thus descriptive, not proscriptive. They define a class of prior convictions, rather than prohibiting particular conduct. To interpret this kind of statute, Taylor instructed us to identify the "contemporary understanding of" an offense and to spurn "[t]he arcane distinctions embedded in the common-law definition." 495 U.S. at 593, 110 S.Ct. 2143 (emphasis added).

A close look at Taylor illuminates how the government's argument, and the BIA's holding in Richardson, disregard entirely the mode of analysis applicable to defining generic crimes under the categorical approach. Taylor interpreted the meaning of "burglary" within the Armed Career Criminal Act, 18 U.S.C. § 924(e). That Act imposed enhanced sentences for offenders previously convicted of, among other things, a "violent felony," which the statute defined to include "burglary." Taylor, 495 U.S. at 578, 110 S.Ct. 2143. The sentencing enhancement applied to convictions under either federal or state law, as does the aggravated felony definition at issue here. At common law, burglary was defined as "the breaking and entering of the dwelling house of another in the nighttime with the intent to commit a felony." Id. at 580 n. 3, 110 S.Ct. 2143 (internal quotation marks and citation omitted). But most contemporary state statutes, the Court noted in Taylor, had deviated from the common-law understanding of burglary, by criminalizing conduct involving "entry without a `breaking,' structures other than dwellings, offenses committed in the daytime, entry with intent to commit a crime other than a felony, etc." Id. at 593, 110 S.Ct. 2143. Stressing that "[t]he arcane distinctions embedded in the common-law definition have little relevance to modern law enforcement concerns[,]" Taylor concluded that it was "unlikely that the Members of Congress, immersed in the intensely practical concerns of controlling violent crime, would have decided to abandon their modern, generic ... definition of burglary and revert to a definition developed in the ancient English law." Id. at 593-94, 110 S.Ct. 2143.

Duenas-Alvarez, 549 U.S. 183, 127 S.Ct. 815, reinforces this conclusion. That case, applying Taylor, considered, as do we, the description of a generic aggravated felony. Subsection 1101(a)(43)(G), at issue in Duenas-Alvarez, lists a "theft offense" as an aggravated felony. Duenas-Alvarez turned on whether the term "theft offense" included the crime of aiding and abetting a theft offense. The common law had distinguished between first-degree principals, second-degree principals, and accessories before the fact, precluding automatic incorporation of the broad concept of "aiding and abetting" into the description of a substantive crime. Id. at 189, 127 S.Ct. 815. But, as Duenas-Alvarez explained, "criminal law now uniformly treats those who fall into th[ose] categories alike." Id. at 190, 127 S.Ct. 815. Rejecting, as in Taylor, reliance on common law concepts, Duenas-Alvarez used the prevalent, contemporary law of aiding and abetting instead, and concluded that the bare statutory reference to those convicted of "theft" included aiders and abetters, as well as principal offenders.

We must apply the same approach here, as our issue is parallel to those in Taylor and Duenas-Alvarez: What set of prior state and federal criminal convictions did Congress mean to encompass in a provision assigning consequences to such previous convictions? As the INA aggravated felony definition is used to impose collateral consequences for earlier state and federal convictions, Taylor and Duenas-Alvarez direct us to presume that Congress sought to track contemporary state criminal practice, not now-abandoned common law concepts. "In the absence of any specific indication that Congress meant to incorporate the common-law meaning of [a term], we shall not read into the statute a definition... so obviously ill suited to its purposes." Taylor, 495 U.S. at 594, 110 S.Ct. 2143; see also Corona-Sanchez, 291 F.3d at 1205 ("Although the common law definition informs us and is the starting point of our analysis, it is not the end point. Indeed, such an approach was rejected by the Supreme Court in Taylor, 495 U.S. at 592-96, 110 S.Ct. 2143....").

The government retorts that adopting the contemporary, generic definition of conspiracy — that is, requiring an overt act — is an implausible interpretation of congressional intent, because a "wide range of criminal conduct ... would fall outside this reading." Not so. As we have seen, the predominant majority of state statutes already subscribe to the generic understanding of general conspiracy, as does the general federal crime of conspiracy. Only a small subset of conspiracy convictions, emanating from that minority of jurisdictions that retain the common-law definition of conspiracy, will not trigger adverse immigration consequences.9 Even were it otherwise, some measure of underinclusiveness is inevitable under the categorical approach, as the Supreme Court has expressly noted. See Moncrieffe, 133 S.Ct. at 1693; see also Descamps, 133 S.Ct. at 2287-89.

ii. In its petition for rehearing, the government contends that United States v. Castleman, ___ U.S.___, 134 S.Ct. 1405, 188 L.Ed.2d 426 (2014), undermines our basis for concluding that Taylor, not the Whitfield line of cases, controls. It argues that, in Castleman, the Supreme Court relied on a common-law definition in interpreting a statute that, like the INA, creates subsequent liability based on convictions from all jurisdictions, rather than directly imposing federal criminal liability for the relevant conduct.

Castleman confronted a statute prohibiting possession of a firearm by anyone who had been convicted of a "misdemeanor crime of domestic violence." 18 U.S.C. § 922(g)(9). "[M]isdemeanor crime of domestic violence" was, in turn, defined as a misdemeanor offense committed within certain intimate relationships which, among other things, "has, as an element, the use or attempted use of physical force." 18 U.S.C. § 921(a)(33)(A).10 The question was how to interpret the term "physical force" within the definition in § 921(a)(33)(A). Castleman, 134 S.Ct. at 1409, 1412. The Court did not adopt its recent construction of "physical force" in the definition of "violent felony" in 18 U.S.C. § 924(e)(2)(B)(i) as requiring "violent force," and instead looked to the common-law-meaning canon of statutory construction. Castleman, 134 S.Ct. at 1409-10 (emphasis and internal quotation marks omitted) (citing Johnson v. United States, 559 U.S. 133, 140, 130 S.Ct. 1265, 176 L.Ed.2d 1 (2010)). The government argues that Castleman shows that the common-law presumption applies not only to direct penal statutes, but also to statutes that apply later consequences to earlier convictions from various jurisdictions.

Castleman is inapposite. As discussed above, in Taylor the Court interpreted the term "burglary," which was used but not defined in 18 U.S.C. § 924(e)(2)(B)(ii). It first rejected the idea that the definition was supplied by the statute of conviction, finding Congress intended a consistent standard. Taylor, 495 U.S. at 590, 110 S.Ct. 2143. It then concluded that Congress used that term not in the common-law sense, but in the "generic sense in which the term is now used in the criminal codes of most States." Id. at 594, 598, 110 S.Ct. 2143. That is, the Court found that the undefined offense of "burglary" was meant as a stand-in for a consistent set of elements, and that those elements should be discerned by reference to the consensus of contemporary sources. In this case, we have proceeded in the same way for the undefined offense of "conspiracy."

By contrast, in Castleman there was no call to look to contemporary sources for a definition of "misdemeanor crime of domestic violence," 18 U.S.C. § 922(g)(9), because the statute itself provided such a definition — namely, the use (or attempted use) of force within certain intimate relationships. 18 U.S.C. § 921(a)(33)(A). The question was how to interpret an element within a definition that was explicitly provided by Congress. Thus the interpretive question at issue in Castleman was very different from the problem of construing what Congress means when it refers to a specific crime, such as "burglary" or "conspiracy," without providing a definition.

Consistent with that difference, none of the opinions in Castleman addressed the contemporary-sources methodology. The majority adopted a common-law interpretation of the term "force," as Justice Alito, concurring in the judgment, would have also done for different reasons. Castleman, 134 S.Ct. at 1410, 1422. Neither discussed the contemporary-sources methodology. Justice Scalia, concurring in part and concurring in the judgment, argued that the Court should give "force" the same meaning as it had in Johnson, principally based on "the presumption of consistent usage," but he also made no reference to the contemporary-sources methodology. Id. at 1417. Moreover, none of the parties' briefs advocated a contemporary-sources approach.11

Thus, the contemporary-sources methodology was simply not at issue in Castleman.12 That being the case, Castleman cannot answer the core issue in this case: should the court rely on a common-law, or contemporary-sources, definition of "conspiracy?"

In short, our response to the government's argument based on Castleman is essentially the same as our response to its argument based on the Whitfield line of cases. While the presumption of common-law meaning was appropriate in considering the statutes at issue in those cases, the interpretive question at issue here is critically different. The question in this case is how to interpret an undefined offense, "conspiracy," when it refers to convictions for that crime in various jurisdictions for the purpose of determining collateral consequences. The Supreme Court's cases provide a clear answer to that question: we interpret such provisions using Taylor's contemporary sources methodology, not the common law meaning.

The government also argues in its petition for rehearing that Castleman illustrates another error in our reasoning. According to the government, it may be appropriate to adopt the contemporary definition of a given crime when, as in Taylor, that definition is broader than the common-law understanding. But when, as in Castleman, the common-law definition would capture more conduct than the alternative, the government suggests that the common-law definition is the right one.

We decline to adopt such an asymmetrical principle. Taylor instructs us to look to the contemporary definition of the offense at issue, not the broadest available definition. The cases applying Taylor have so understood its directive.

For example, courts considering the interpretation of generic "manslaughter" have concluded that the contemporary definition requires a mens rea of at least recklessness. See, e.g., Dominguez-Ochoa, 386 F.3d at 646; United States v. Roblero-Ramirez, 716 F.3d 1122, 1126 (8th Cir.2013) (collecting cases). Those cases applied the narrower contemporary definition of manslaughter rather than the broader common-law definition, which would include manslaughter convictions based on a negligence theory. See Dominguez-Ochoa, 386 F.3d at 646; 2 LaFave, supra, § 15.4. As the manslaughter cases illustrate, Taylor's methodology controls regardless of whether the contemporary definition of a given crime is broader or narrower than the common-law understanding.

iii. The BIA's contrary conclusion in Richardson, 25 I. & N. Dec. 226, is due no deference under Chevron, USA, Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). The government has not urged us to defer to the BIA's interpretation of "conspiracy," with good reason, as we now explain.

Generally, "we have held that the [BIA's] precedential orders [interpreting the INA], which bind third parties, qualify for ... deference" under Chevron. Marmolejo-Campos v. Holder, 558 F.3d 903, 909 (9th Cir.2009) (en banc). Such deference is due "regardless of whether the order under review is the precedential decision itself or a subsequent unpublished order that relies on it." Id. at 911. We have, on occasion, accorded such deference to the definition of generic offenses listed in 8 U.S.C. § 1101(a)(43). See, e.g., Renteria-Morales v. Mukasey, 551 F.3d 1076, 1081 (9th Cir.2008); Parilla v. Gonzales, 414 F.3d 1038, 1041 (9th Cir.2005).

Historically, we implemented Chevron via a two-step inquiry, asking first whether a statute was ambiguous and, if so, whether the agency's interpretation of it was reasonable. See, e.g., Ariz. Health Care Cost Containment Sys. v. McClellan, 508 F.3d 1243, 1249 (9th Cir.2007). More recently, however, the Supreme Court has authorized courts to omit evaluation of statutory ambiguity on the ground that, "if Congress has directly spoken to an issue then any agency interpretation contradicting what Congress has said would be unreasonable." Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208, 218 n. 4, 129 S.Ct. 1498, 173 L.Ed.2d 369 (2009); see also United States v. Home Concrete & Supply, LLC, ___ U.S.___, 132 S.Ct. 1836, 1846 n. 1, 182 L.Ed.2d 746 (2012) (Scalia, J., concurring in part and concurring in the judgment) ("Whether a particular statute is ambiguous makes no difference if the interpretation adopted by the agency is clearly reasonable — and it would be a waste of time to conduct that inquiry."); Matthew C. Stephenson & Adrian Vermeule, Chevron Has Only One Step, 95 Va. L.Rev. 597, 599-600 (2009).

Here, the one-step approach makes much more sense. Chevron instructs us to "employ[ ] traditional tools of statutory construction." Chevron, 467 U.S. at 843 n. 9, 104 S.Ct. 2778. In this case, it might at first glance appear that traditional tools of statutory construction point toward two different interpretations of the term "conspiracy," either one of which is seemingly reasonable. As noted, application of the methodology employed in Taylor indicates that the generic definition of "conspiracy" requires an overt act. By contrast, application of the presumption that undefined terms carry their common-law meaning indicates that the statute's bare reference to "conspiracy" does not include any such overt-act requirement. Whitfield, 543 U.S. at 213, 125 S.Ct. 687 (citing Shabani, 513 U.S. at 13-14, 115 S.Ct. 382).

We conclude, however, that the BIA's interpretation of the statute's reference to conspiracy is impermissible, as that interpretation entirely ignores the one methodology properly applicable in this context — namely, the mode of analysis derived from Taylor and its progeny, which we use to determine generic crimes for the purposes of categorical analysis of prior convictions.13 As we have seen, the Supreme Court specifically held inapplicable in the context of defining generic federal crimes for purposes of Taylor categorical analysis the principle on which Richardson rests — "that, absent contrary indications, Congress intends to adopt the common law definition of statutory terms." Shabani, 513 U.S. at 13-14, 115 S.Ct. 382. Where, as here, the Supreme Court has prescribed the mode of determining congressional intent and declared the alternative, relied on by Richardson, and by the government in this case, "ill suited to [the] purposes" of a statute establishing collateral consequences, Taylor, 495 U.S. at 594, 110 S.Ct. 2143, we cannot use it, and cannot defer to an agency decision that does. Whether we characterize this conclusion as (1) a rejection of the BIA's interpretation at Chevron step one because the only correct traditional tool of statutory construction unambiguously yields a different result, or (2) a rejection at Chevron step two on the ground that the statute is ambiguous but the BIA's interpretation unreasonable in light of its improper methodology, makes no difference.

We thus hold that "conspiracy," under 8 U.S.C. § 1101(a)(43)(U), requires proof of an overt act, and reject the BIA's contrary conclusion.

V.

The Nevada statute of conviction, Nev.Rev.Stat. § 199.480, requires no proof of an overt act, Nev.Rev.Stat. § 199.490. The generic definition of conspiracy, codified at 8 U.S.C. § 1101(a)(43)(U), does. Garcia's prior conviction, for conspiracy to commit burglary, is therefore not an aggravated felony under the INA.

AFFIRMED.