(Joining in part and Dissenting in part).
I join the majority in affirming the dismissals of Dimott’s .and Collamore’s § 2255 petitions as untimely. However, I cannot join in the majority’s disparate and inconsistent treatment of Casey’s petition for habeas relief, as opposed to its treatment of the other two petitions at issue, in order to avoid what this case truly calls for: a reevaluation of this Court’s opinion in Duquette in light of the Supreme Court’s decision in Mathis,
In the cases of Dimott and Collamore, the majority correctly gives “due weight” to the habeas judge’s finding that the petitioners were sentenced according to the ACCA’s enumerated clause because the habeas judge was also the sentencing judge. It is eminently reasonable that a sentencing judge is capable of determining the basis upon which he or she imposed a sentence enhancement when subsequently reviewing that sentence on a § 2255 habe-as petition. See Schriro v. Landrigan, 550 U.S. 465, 495-96, 127 S.Ct. 1933, 167 L.Ed.2d 836 (2007) (stating that a judge’s memory deserves some deference provided it is based on a complete review of the case). Here, Judge Singal had the opportunity to review Dimott arid Collamore’s cases prior to determining that he had sentenced them under the enumerated clause. Thus, that determination desérves the deference, as the panel majority recognizes.
The same deference must be given to the habeas judge who reviewed Casey’s petition, Judge Hornby, who—like Judge Singal in Dimott and Collamore’s cases— was the judge that sentenced Casey. On habeas review, Judge Hornby, also facing a silent record as to the clause under which he applied Casey’s ACCA sentencing enhancement, fourid that Casey did raise a Johrison II claim—meaning that his sentence was enhanced pursuant to the ACCA’s residual clause. See Casey, 2016 WL 6581178, at *3. Judge Hornby analyzed the habeas petition accordingly. Id., at *3-5. Yet, the majority inexplicably fails to give Judge Hornby the same deference that it gives to Judge Singal.
The majority incorrectly assumes that my “insist[ence]” that the district court found that Casey raised a timely Johnson II claim is that the court analyzed the Government’s procedural-default, argument. This is wide of the mark. Rather, I so find after according Judge Hornby’s words their clear and unambiguous meaning. See id., at *3 (“I. conclude that ... Casey’s Johnson claim is a novel constitutional claim that applies retroactively, and he has therefore shown cause for [failing to argue that the ACCA residual clause was unconstitutional at sentencing or on appeal].”), *4 n.9 (“As I have determined above, Casey’s Johnson claim is a novel constitutional claim with retroactive application. ...”). The majority rationalizes its disregard of this plain language by claiming that I “take [it] out of context.” Yet, as the majority notes, should the district court have believed that Casey had been sentenced pursuant to anything but the ACCA’s residual clause, it could have raised timeliness sua sponte. The district court was clearly aware that similar petitions had been decided on timeliness grounds—it even discussed Dimott in its decision; should it have believed such an argument appropriate, it would not have needed to reach the merits of Casey’s Johnson II claim in order to conduct a prejudice analysis. See id. at *5. But, it did not raise the issue, and after finding that Casey was sentenced pursuant to the residual clause, embarked on the more onerous procedural default analysis. “Due regard for the trial court’s processes and time investment is ... a consideration appellate courts should not overlook.” Wood, 566 U.S. at 474,132 S.Ct. 1826.
In a further departure from this Court’s guiding judicial doctrines, the majority raises sua sponte the issue of the timeliness of Casey’s habeas petition, which the Government did not argue below. In doing so, the majority ignores the advice provided by the Supreme Court in Wood that, in situations such as this, “[although a court of appeals has discretion to address, sua sponte, the timeliness of a habeas petition, appellate courts should reserve that authority for use in exceptional cases.” Id. at 473,132 S.Ct. 1826 (finding that the appellate court abused its discretion in raising the timeliness issue sua sponte); see also Cole v. Int’l Union, United Auto., Aerospace & Agric. Implement Workers of Am., 533 F.3d 932, 936 (8th Cir. 2008) (applying only a “narrow exception” to established preservation rule). The Government makes no argument that this is an exceptional case, and—especially in light of Judge Hornby’s finding that Casey raised a timely Johnson II claim—this is not the appropriate case for the Court to act on its own accord. Here, as in Wood, where the Government forewent an argument below, we should not exercise our confined discretion to save the Government’s waiver.
To justify its divergence from Wood’s guidance and find that the Government did not forfeit its timeliness argument, the majority speculates—in the Government’s favor—as to the reason that the Government did not advance this argument. I cannot subscribe to this guesswork approach. This Court religiously finds a party’s failure to raise an argument before the district court as waived on appeal. See, e.g., United States v. Román-Huertas, 848 F.3d 72, 77 (1st Cir. 2017) (“The Government did not raise [petitioner’s] untimely objection before the district court, ... and so it [is] waived....”); Sotirion v. United States, 617 F.3d 27, 32 (1st Cir. 2010) (finding the Government’s procedural default argument waived for failing to raise it as a defense in the district court to a § 2255 petition). The same waiver must apply here, and we should refrain from such “unguided speculation.” Cf. Holloway v. Arkansas, 435 U.S. 475, 491, 98 S.Ct. 1173, 55 L.Ed.2d 426 (1978) (finding a harmless-error analysis inappropriate in assessing constitutional error of joint representation); Walsh v. TelTech Systems, Inc., 821 F.3d 155, 160 (1st Cir. 2016) (stating that appellate courts draw all reasonable inferences in favor of the nonmoving party but ignore unsupported speculation when reviewing an award of summary judgment).
Moreover, I have significant qualms with the effect that the majority’s reasoning has on the waiver doctrine. The majority credits the Government for “brief[ing] [the timeliness issue] on appeal and argu[ing] that it did not waive the timeliness bar.” Yet, this is precisely what the waiver doctrine is intended to prevent. Applying the majority’s approach would allow any party that chose not to raise an argument in the district court to simply brief that issue on appeal and argue that it did not waive the issue below. In those circumstances, we would find the argument waived, as we should in this one. Further, the Government only acknowledges its failure to raise the timeliness issue in a footnote in its opening brief to this Court, providing scant explanation as to why the claim was not raised below or why it should not be treated as waived. Instead, it states that the petitioners have briefed the issue and that this Court may raise it sua sponte. Such an undeveloped address is hardly sufficient to save the argument from waiver on appeal. United States v. Zannino, 895 F.2d 1, 17 (1st Cir. 1990) (“[I]ssues adverted to in a perfunctory manner, unaccompanied by some effort at developed argumentation, are deemed waived.”).
Instead, I would find that the Government relinquished its timeliness argument in the district court. I note that this same U.S. Attorney’s Office (for the District of Maine) raised the issue of timeliness in its oppositions to both Dimott’s and Collar more’s § 2255 petitions, both filed within six weeks of its opposition to Casey’s petition. While the majority attributes the Government’s decision not to advance this argument in response to Casey’s petition as inadvertence rather than strategy, I do not so conjecture. This strikes me as an appropriate basis for finding that the Government displayed its “clear and accurate understanding of the timeliness issue” and “knew that it had an arguable statute of limitations defense,” but relinquished that argument. Wood, 566 U.S. at 474, 132 S.Ct. 1826 (internal quotation marks and citation omitted).
Finally, the majority’s finding that Casey failed to satisfy his burden of proving by a preponderance of the evidence that he was sentenced under the residual clause is equally unpersuasive. The majority pronounces that, in the face of a silent record, placing the burden on a petitioner “makes sense ... because they were certainly present at sentencing and knowledgeable about the conditions under which they were sentenced.” I fail to see what could better satisfy the majority’s evidentiary requirement that petitioner was sentenced under the residual clause than a finding by the sentencing judge, who was also “certainly present at sentencing” and far more knowledgeable of his own sentencing decisions. I have a difficult time thinking of what further evidence, in the face of a silent record, could be more convincing. The majority suggests in a footnote that Casey could have asked for a remand to the district court to prove that he was sentenced solely under the residual clause; however, such a request would have been nonsensical after the habeas judge clearly already found as much. See Casey, 2016 WL 6581178, at *3.
Given the deference owed to the habeas judge here, I would find that, under any of the standards announced by our sister circuits and discussed by the majority,9 Casey has shown that he was sentenced pursuant to the residual clause and thus brought forth a timely Johnson II claim. This Court should analyze the matter- accordingly. As the district court did below, we would accordingly need to address whether Casey’s claim is procedurally defaulted for failing to raise it at trial or on direct appeal. The district court found there to be cause for Casey not having raised the issue, but that, while believing that Mathis casts significant doubt on the vitality of Duquette, it was bound by this Circuit’s precedent to find that Maine burglary is generic and also falls under the enumerated clause. Casey, 2016 WL 6581178, at *5. Accordingly, it found that Casey did not suffer any actual prejudice. Id. at *4.
The district court was correct in its ruling' given its boundaries. However, this Court is not so constrained. See United States v. Tavares, 843 F.3d 1, 11 (1st Cir. 2016) (stating that the court may overturn prior panel decisions when controlling authority is subsequently announced or when, in light' of new authority, the panel would likely have changed its “collective mind.” (quoting United States v. Pires, 642 F.3d 1, 9 (1st Cir. 2011))). Mathis is subsequent controlling authority which calls into question the vitality of our opinion in Duquette. See United States v. Whindleton, 797 F.3d 105, 113 (1st Cir. 2015) (“An exception to the doctrine of stare decisis applies if ‘[a]n existing panel decision [is] undermined by controlling authority, subsequently announced, such as an opinion of the Supreme Court....’” (alterations in original) (citing United States v. Rodriguez-Pacheco, 475 F.3d-434, 441 (1st Cir. 2007))). In Duquette, we found that because the Maine burglary statute contains all of the elements of generic burglary, under Taylor, 495 U.S. 575, 110 S.Ct. 2143, 109 L.Ed.2d 607, it qualified as generic burglary under the ACCA’s enumerated clause. However, Mathis has undermined this analysis, instead -calling for us to determine if one (or more) of the elements of Maine burglary is- broader than the corresponding element of the generic offense. If, so, then Maine’s burglary statute, like Iowa’s burglary statute, cannot fall under the ACCA’s enumerated clause.
While we have not conducted this reanalysis of Duquette, Casey’s petition for habeas relief calls for us to do so tó determine if Casey suffered actual prejudice. Addressing this- more difficult issue— which the majority seeks to avoid—is necessary to decide this case.10
Accordingly, I join in affirming the outcome proposed by the majority in the cases of Dimott and Collamore, and respectfully dissent from the majority in regards to Casey’s petition for habeas relief.
. As the majority explains, there is an emerging split amongst the circuit courts as to the burden of proof placed on petitioners facing a silent record who, through a § 2255 petition, maintain that their sentences were enhanced pursuant to the residual clause of the ACCA. The Fifth Circuit described this split well in Taylor, 873 F.3d at 480-81 (citing Beeman, 871 F.3d at 1221-22 (finding ¿at a defendant must show that "more likely than not” he was sentenced according to the residual clause); Snyder, 871 F.3d 1122 (10th Cir. 2017) (stating that courts should look to the law at the time of sentencing and determine whether a defendant’s convictions fell within the scope of the other ACCA clauses); Geozos, 870 F.3d at 895 (holding that, "when it is unclear whether a sentencing court relied on the residual clause in finding that a defendant qualified as an armed career criminal, but it may have, the defendant’s § 2255 claim 'relies on’ the constitutional rule announced in Johnson II.” (citing Winston, 850 F.3d at 682)); Winston, 850 F.3d at 682 (finding that imposing the burden on movants would result in "selective application” of the new rule announced in Johnson II)).
. As pointed out by the district court below, Casey, 2016 WL 6581178, at *5 n.16, and the Government in its Rule 28(j) letter to the Court, numerous federal circuits have recently reviewed state burglary statutes in light of Mathis to determine whether they continue to qualify as enumerated felonies under the ACCA.