Cook v. State

In the Supreme Court of Georgia Decided: March 15, 2022 S21A1270. COOK v. THE STATE. WARREN, Justice. When a convicted criminal defendant is unconstitutionally deprived of an appeal of right, which typically occurs because her counsel provided ineffective assistance in failing to file a timely appeal, she is entitled to an untimely or “out-of-time” appeal. In this case, we examine the difficult question of whether this Court should overrule our precedent allowing a criminal defendant who alleges that she was unconstitutionally deprived of her appeal as of right to file a motion for out-of-time appeal in the trial court, as opposed to seeking a writ of habeas corpus as an exclusive remedy. After explaining the underpinnings of our precedent and engaging in an exhaustive stare decisis analysis, we conclude that the trial court out-of-time appeal procedure is not a legally cognizable vehicle for a convicted defendant to seek relief for alleged constitutional violations. *** In the Habeas Corpus Act of 1967, now codified as OCGA § 9- 14-40 et seq., the General Assembly created a post-conviction procedure for defendants to raise that—and any other— constitutional claim. Defendants in Georgia began doing so, and in 1974 this Court held in Neal v. State, 232 Ga. 96 (205 SE2d 284) (1974), that a defendant could not seek an out-of-time appeal from his conviction by motion in the trial court, explaining that he must file a petition for a writ of habeas corpus to seek relief for the “denial of the right of appeal or of the effective assistance of counsel on appeal” and affirming the trial court’s order dismissing the defendant’s motion. Id. at 96. Nevertheless, the following year, without mention of Neal or the Habeas Corpus Act, this Court began to review appeals of trial court orders denying motions for out-of-time appeals on the merits and appeals following orders by trial courts granting out-of-time 2 appeals. See King v. State, 233 Ga. 630, 630-631 (212 SE2d 807) (1975); Furgerson v. State, 234 Ga. 594, 595-596 (216 SE2d 845) (1975). Those cases did not constitute precedents on the proper way to seek an out-of-time appeal, but two decades later, in Rowland v. State, 264 Ga. 872 (452 SE2d 756) (1995), the Court for the first time held—without citing any applicable legal authority and without acknowledging or overruling the contrary holding in Neal—that a convicted defendant could seek an out-of-time appeal either in the trial court or in habeas corpus. See id. at 875. Trial courts thus continued to entertain motions for out-of-time appeal, and appellate courts continued to decide appeals following the rulings on such motions for many more years. This Court did not examine how the trial court out-of-time appeal procedure had been created and evolved; the inconsistencies that had developed between that procedure and other areas of established Georgia law; or how the procedure had become an unwitting breeding ground for legal errors made by both appellate and trial courts until we decided Collier v. State, 307 Ga. 363 (834 3 SE2d 769) (2019), in 2019. Collier raised the profile of the trial court out-of-time appeal process, and given that the only way rules have been established for this judicially created procedure is by judges making them on a case-by-case basis, we have since seen additional cases that have called upon this Court to create the rules and parameters of the trial court out-of-time appeal procedure. In Collier, and in a case that soon followed, Schoicket v. State, 312 Ga. 825 (865 SE2d 170) (2021), this Court determined that our precedent endorsing the trial court out-of-time appeal procedure as an alternative to habeas corpus had no valid legal foundation. See Collier, 307 Ga. 371-373, 376; id. at 379-382 (Peterson, J., concurring specially); Schoicket, 312 Ga. at 825. The question that follows is whether to maintain that precedent as a matter of stare decisis. We asked the parties in this case, as well as amicus curiae for major participants in the criminal justice system, to address that question. Based on their input and our extensive consideration of the issue, we conclude that stare decisis considerations do not weigh against overruling our precedent that created the trial court out-of-time 4 appeal procedure. This Court has consistently held that the most important stare decisis factor is the soundness of the reasoning of the precedent at issue, and everyone involved in this case agrees that the reasoning supporting our precedent allowing motions for out-of-time appeal in trial courts—to the extent there has been any reasoning at all—is wholly unsound. There also is agreement that the precedent is not ancient and that it does not implicate traditional reliance interests. With respect to the fourth stare decisis factor that this Court typically considers—workability—we conclude that our precedent creating the trial court out-of-time appeal procedure has a fundamental and insurmountable workability problem, because it will perpetually require this Court to fill in the details of the procedure we created. Judicial administration of habeas corpus, which everyone agrees is an appropriate procedure for seeking an out-of-time appeal when a convicted criminal defendant is unconstitutionally deprived of an appeal of right, does not require judges to make up the rules that regulate that process because the 5 General Assembly has established those rules by statute. But as Collier and Schoicket demonstrate, when it comes to the trial court out-of-time appeal procedural vehicle this Court invented, we are called upon to make up the rules that govern that procedure. And each time we do, we are required to step out of our proper judicial role and assume the role of law-makers—which is the work of the General Assembly. The dissent has no real answer to this problem. It also undervalues the unsoundness of our precedent and overstates the ways in which the procedure we created may work better than the habeas procedure that is legally proper. In the end, the dissent seeks to overcome our customary stare decisis analysis with a focus on the “entrenchment” of our precedent—a concept we have considered before in stare decisis analyses, but never to outweigh all other stare decisis factors and perpetuate an unworkable and wholly unsound precedent, as the dissent proposes here. For these reasons, which are explained more fully below, we conclude that principles of stare decisis do not require us to maintain 6 our unsound precedent creating or endorsing the trial court out-of- time appeal procedure, and we therefore overrule Rowland and its handful of progeny. We also disapprove King, Furgerson, and other decisions to the extent that they allowed out-of-time appeal claims to be litigated in trial courts without addressing the propriety of that procedure. Applying our holding to this case, we vacate the trial court’s order denying Cadedra Lynn Cook’s motion for an out-of-time appeal, and we remand the case to the trial court with direction that the motion be dismissed. 1. Procedural Posture of Cook’s Appeal In November 2013, Cook entered a negotiated plea of guilty to charges of felony murder and armed robbery, based on which the trial court entered a judgment of conviction and sentenced her to life in prison with the possibility of parole for felony murder and a concurrent 20-year term in prison for armed robbery. 1 She did not 1 Cook’s co-indictee, Eddie Clark, pleaded guilty to felony murder, armed robbery, and obstruction of a law enforcement officer in February 2014. In 2020, we vacated a trial court order denying his motion for out-of-time appeal and remanded the case for an evidentiary hearing to “determine whether plea 7 file a timely appeal, but more than six years later, she filed a motion for out-of-time appeal in the trial court, contending that she was deprived of her right to appeal because of her plea counsel’s ineffective assistance. 2 After a hearing, the trial court denied Cook’s motion for out-of-time appeal on the merits, and she timely appealed that decision. Cook and the State filed appellate briefs making arguments related to the merits of Cook’s motion. But in light of this Court’s recent examinations of the propriety of motions for out- of-time appeals in trial courts and related issues in Collier, Kelly v. State, 311 Ga. 827 (860 SE2d 740) (2021), and Schoicket, we requested and received supplemental briefing on the following two questions3: counsel’s ineffective assistance was responsible for Clark’s failure to pursue a timely appeal.” Clark v. State, 310 Ga. 489, 491 (852 SE2d 522) (2020) (citation and punctuation omitted). 2As part of her motion for out-of-time appeal, Cook also sought an out- of-time motion to withdraw her guilty plea. But we recently held that “a granted motion for out-of-time appeal does not confer a right to file an otherwise-untimely motion to withdraw a guilty plea.” Schoicket v. State, 312 Ga. 825, 833 (865 SE2d 170) (2021). 3 We appreciate the supplemental briefs filed on behalf of the parties by 8 Should this Court reconsider whether a criminal defendant who alleges that she was deprived of her right to appeal because of her counsel’s alleged ineffective assistance under Strickland v. Washington, 466 U.S. 668 (104 SCt 2052, 80 LE2d 674) (1984), be permitted to seek a remedy for that alleged constitutional violation by filing a motion for out-of-time appeal in the trial court, as opposed to filing, as her exclusive remedy, a petition for writ of habeas corpus? How do considerations of stare decisis apply in this analysis? 2. Legal Background (a) The Judicial Creation and Propagation of the Motion for Out-of-Time Appeal in Georgia Trial Courts. In 1965, the United States Supreme Court granted certiorari to consider “whether the Fourteenth Amendment [to the United States Constitution] requires that the States afford state prisoners some adequate corrective process for the hearing and determination of claims of violation of federal constitutional guarantees.” Case v. Nebraska, 381 U.S. 336, 337 (85 SCt 1486, 14 LE2d 422) (1965). But appellant’s counsel, the District Attorney, and the Attorney General. We are also grateful for the thoughtful briefs filed by amici curiae the Georgia Association of Criminal Defense Lawyers (“GACDL”) and the Prosecuting Attorneys’ Council of Georgia (“PAC”). 9 because the Nebraska legislature enacted, during the pendency of that case, a statutory post-conviction procedure that “provide[d] for a hearing of petitions . . . alleging denial of federal constitutional rights,” the Supreme Court did not address the merits of the question presented in Case and instead vacated the judgment below. Id. However, in the wake of two concurring opinions in Case that expressed approval of states providing statutory post-conviction procedures and hope that such procedures would be broadly adopted in other states, see id. at 337-340 (Clark, J., concurring), 340-347 (Brennan, J., concurring), many states enacted statutory post- conviction procedures, including but not limited to state habeas corpus procedures. See 7 Wayne R. LaFave et al., Criminal Procedure § 28.11 (a) (4th ed., Nov. 2021 Update) (the states “responded en masse to the urgings of” the concurring Justices “and others, and today ‘in each of the 50 states, the principal postconviction remedy may be used, at a minimum, to raise claims that the conviction is void for lack of jurisdiction or was obtained in violation of a constitutional right.’”) (quoting 1 DONALD E. WILKES, 10 FEDERAL AND STATE POSTCONVICTION REMEDIES AND RELIEF HANDBOOK WITH FORMS 7 (2014–2015 ed.)). Among those states was Georgia, which in 1967 enacted the Habeas Corpus Act. See Ga. L. 1967, p. 835 (codified as amended at OCGA § 9-14-40 et seq.). The Act, which is codified as Article 2 of Chapter 14 of Title 9, applies to convicted defendants and says that “this article provides the exclusive procedure for seeking a writ of habeas corpus for persons whose liberty is being restrained by virtue of a sentence imposed against them by a state court of record.” OCGA § 9-14-41. OCGA § 9-14-42 (a) further specifies that any such person “who asserts that in the proceedings which resulted in his conviction there was a substantial denial of his rights under the Constitution of the United States or of this state may institute a proceeding under this article.” About two years after the enactment of Georgia’s Habeas Corpus Act, the United States Supreme Court held, in a federal case in which counsel failed to file a notice of appeal as requested by the defendant and thereby deprived the defendant of an appeal as of 11 right, that the defendant “should be resentenced so that he may perfect an appeal in the manner prescribed by the applicable rules.” Rodriquez v. United States, 395 U.S. 327, 332 (89 SCt 1715, 23 LE2d 340) (1969). See also Douglas v. California, 372 U.S. 353, 356-358 (83 SCt 814, 9 LE2d 811) (1963) (indigent defendants have a constitutional right to appointed counsel for their first appeal as of right from a criminal conviction); Evitts v. Lucey, 469 U.S. 387, 396 (105 SCt 830, 83 LE2d 821) (1985) (“A first appeal as of right . . . is not adjudicated in accord with due process of law if the appellant does not have the effective assistance of an attorney.”). In the years that followed, this Court began to hold that convicted defendants who were unconstitutionally deprived of their right to appeal as a result of a denial of counsel or as a result of trial counsel’s constitutionally ineffective assistance could obtain an “out- of-time appeal” as relief in Georgia habeas corpus proceedings. See Roberts v. Caldwell, 230 Ga. 223, 224 (196 SE2d 444) (1973) (reversing the denial of habeas relief because the petitioner “was denied appellate counsel on his first appeal,” and remanding the 12 case “with direction to the habeas corpus court to enter an order providing for the appointment of counsel to determine if there is any justifiable ground for an appeal from the original convictions, and if such determination is in the affirmative, then an appeal may be filed and prosecuted with benefit of counsel even at this late date”) (emphasis supplied); McAuliffe v. Rutledge, 231 Ga. 745, 746 (204 SE2d 141) (1974) (reversing the judgment of the habeas court because the petitioner “was indeed denied effective assistance of counsel in attempting to appeal his conviction,” and directing that the petitioner “be allowed, if he so desires, to file an out of time appeal to the proper appellate court within 30 days from the date the remittitur from this court is filed in the trial court”) (emphasis supplied). See also Collier, 307 Ga. at 371 (observing that a “few years after Rodriquez, the ‘out of time appeal’ remedy began to appear in Georgia’s habeas corpus jurisprudence”); id. at 373 (“A request for an out-of-time appeal based on a deprivation of constitutional rights clearly may be brought in a petition for a writ 13 of habeas corpus.”).4 Accordingly, when this Court first encountered a motion for out-of-time appeal filed in a trial court (rather than in a habeas court) shortly after the Habeas Corpus Act’s enactment, we deemed the motion invalid. In Neal, we affirmed a trial court’s order dismissing an inmate’s motion seeking an untimely appeal from his conviction on a guilty plea entered seven years earlier, explaining that a petition for a writ of habeas corpus was “an adequate post- conviction remedy” for “denial of the right of appeal or of the effective assistance of counsel on appeal.” 232 Ga. at 96 (citing McAuliffe, 231 Ga. at 745). In so doing, we specifically held that the 4 There is no dispute that an out-of-time appeal may still be sought as a remedy in a habeas corpus proceeding. See, e.g., Hall v. Jackson, 310 Ga. 714, 724 (854 SE2d 539) (2021) (the appropriate remedy when a habeas court determines that appellate counsel provided ineffective assistance due to a conflict of interest “is to grant [the petitioner an] out-of-time appeal, which will allow him to start the post-conviction process anew with the assistance of conflict-free counsel”); Trauth v. State, 295 Ga. 874, 876 (763 SE2d 854) (2014) (“[W]here, as here, a pro se defendant has been improperly denied counsel for his first appeal, he is entitled to [habeas] relief in the form of having counsel appointed ‘to determine if there is any justifiable ground for an appeal from the original convictions, and if such determination is in the affirmative, file and prosecute a new direct appeal with the benefit of counsel.’”) (quoting Roberts, 230 Ga. at 224; punctuation omitted). 14 motion “should have been filed [as a habeas petition] in the superior court wherein the petitioner [was] being detained, not in the convicting court.” Id. at 96-97. Nevertheless, a year after Neal, without acknowledging our holding in that case or the existence of the Habeas Corpus Act, this Court began to review appeals from trial court denials of motions for out-of-time appeal, as well as appeals from convictions pursuant to the grants of such motions by trial courts, without addressing whether the out-of-time appeal procedure in the trial court was authorized in the first place. In King, 233 Ga. at 630-631, we considered on the merits an appeal from the trial court’s denial of a motion for out-of-time appeal, and in Furgerson, 234 Ga. at 595-596, we considered on the merits an appeal that followed the trial court’s grant of a motion for out-of-time appeal. Notably, neither King nor Furgerson constituted precedent on the trial court out-of-time appeal procedure, because in those cases we merely considered the merits of an appeal following a trial court ruling on a motion for out-of-time appeal; we did not address whether the trial court procedure was 15 valid. See Collier, 307 Ga. at 372 (“Notwithstanding [Neal], this Court began reviewing trial court rulings on out-of-time appeal motions without any discussion of the propriety of the out-of-time appeal process in the trial court.”) (citing King and Furgerson). See also Seals v. State, 311 Ga. 739, 745 (860 SE2d 419) (2021) (“[I]n none of those cases was there any discussion of jurisdiction at all, much less analysis and a holding on the issue. Decisions of this Court and of the Court of Appeals ‘do not stand for points that were neither raised by the parties nor actually decided in the resulting opinion,’ and ‘questions which merely lurk in the record, neither brought to the attention of the court nor ruled upon, are not to be considered as having been so decided as to constitute ‘precedents.’”) (citation omitted). 5 5 None of our early cases allowing the out-of-time appeal procedure in trial courts cited any authority, with one exception: in 1978, this Court noted that the out-of-time appeal procedure in Georgia “has no codical basis” but “appears to have had its genesis in Byrd v. Smith, 407 F2d 363 (5th Cir. 1969).” Lay v. State, 242 Ga. 225, 225 n.1 (248 SE2d 611) (1978). See also Collier, 307 Ga. at 371 n.9 (quoting Lay). But Byrd was a federal habeas decision requiring that Georgia “either allow an appeal at this time or permit an out-of-time appeal by whatever procedure is necessary.” 407 F2d at 366. Byrd did not purport to require Georgia to establish a state procedure other than habeas 16 Then, two decades later in Rowland, this Court for the first time held squarely that an out-of-time appeal could properly be obtained in the trial court as well as in habeas corpus. Specifically, we concluded that the out-of-time appeal “granted where [counsel’s] deficiency involves not the trial but the denial of the right to appeal . . . serves as a remedy” not only “for a habeas corpus petitioner who suffered a constitutional deprivation,” but also for “the criminal defendant who has shown ‘good and sufficient reason’ to a trial court.” Rowland, 264 Ga. at 875 (citations and punctuation omitted). We reached that conclusion without citing any applicable legal authority and without acknowledging or overruling our contrary holding in Neal. See id. at 875-876. However unreasoned it was, Rowland, rather than Neal, became the governing precedent on the trial court out-of-time appeal procedure in Georgia courts. See White v. State, 305 Ga. 111, 122 n.10 (823 SE2d 794) (2019) (explaining that when discord exists between older and newer corpus for considering alleged unconstitutional deprivations of the right to appeal. 17 precedents of a jurisdiction’s highest court, the more recent decision controls). See also Schoicket, 312 Ga. at 829 n.5 (citing White). In sum: even though the General Assembly statutorily established habeas corpus as the exclusive procedure for vindicating a convicted defendant’s constitutional rights, including the deprivation of the right to appeal, and established the contours of the procedure to seek such relief, this Court allowed and then expressly endorsed a procedure parallel to, but distinct from, habeas corpus for convicted defendants to seek vindication of alleged constitutional violations that frustrated their right to appeal. And allowing convicted defendants to do so in turn allowed them to circumvent the requirements and restrictions imposed by the Habeas Corpus Act. Our error was significant. By judicially creating this trial court out-of-time appeal procedure—a procedure that is neither authorized by our common law 6 nor established by statute—this 6 At common law, courts could only modify their judgments in the same term of court, which in Georgia ranges from two to eight months long. See 18 Court acted not as a body of judges, but as a body of lawmakers. Cf. Duke v. State, 306 Ga. 171, 174, 186 (829 SE2d 348) (2019) (unanimously overruling precedent that “created a judicial exception to the statutory requirements for bringing an interlocutory appeal” and recognizing that the “scheme for appellate interlocutory review is legislative in nature”) (citation and punctuation omitted). In so doing, we usurped the power of the Legislative Branch, which implicates separation-of-powers concerns under the Georgia Constitution. See Ga. Const. of 1983, Art. I, Sec. II, Par. III (“The legislative, judicial, and executive powers shall forever remain separate and distinct; and no person discharging the duties of one shall at the same time exercise the functions of either of the others . . . .”). See also Duke, 306 Ga. at 186 (“We reiterate this core separation of powers principle today. . . . [T]he General Gray v. State, 310 Ga. 259, 262 (850 SE2d 36) (2020) (“Georgia courts have long applied the common-law rule that the trial court has the inherent authority to modify a judgment within the term of court and that a motion made during the term serves to extend the power to modify.”) (citation and punctuation omitted); OCGA § 15-6-3 (setting out the “terms of court for the superior courts for each of the judicial circuits”). We are not aware of any procedure at common law by which a trial court may allow a party to challenge a conviction outside the term of court. 19 Assembly is free to change or abolish th[e] requirement [that a trial court must issue a timely certificate of immediate review before an interlocutory appeal may be pursued]. But this Court lacks that authority, and we should never have claimed it.”). Suffice it to say that we did not identify that concern at the time we decided Rowland or for many years thereafter. Instead, we continued for more than two decades to decide cases following trial court rulings on motions for out-of-time appeals—and apparently did so without consideration of the statutory and decisional authorities Rowland ignored in reaching its holding. (b) Two Decades of Misinterpretation and More Recent Course-Correction With Respect to the Availability of Out- of-Time Appeals After Guilty Pleas. As it turns out, our unsupported and seemingly unthinking creation of the trial court out-of-time procedural vehicle was matched by our long-standing erroneous application of the substantive standard for granting an out-of-time appeal following a conviction by guilty plea. Indeed, over the course of many years, “this Court and the Court of Appeals [ ] in hundreds of cases” 20 erroneously “required defendants whose convictions came by guilty pleas to show that they could actually prevail in an appeal before allowing them that appeal out-of-time.” Ringold v. State, 304 Ga. 875, 883 (823 SE2d 342) (2019) (Nahmias, P.J., concurring). That requirement was flatly inconsistent with the United States Supreme Court’s holding in Roe v. Flores-Ortega, 528 U.S. 470 (120 SCt 1029, 145 LE2d 985) (2000)—controlling precedent on Sixth Amendment ineffective-assistance-of-counsel claims involving the failure to file a timely appeal from a criminal conviction, whether that conviction was based on a guilty plea or a guilty verdict after a trial. 7 See 7 In Flores-Ortega, which involved a lawyer’s failure to file a notice of appeal following the entry of his client’s guilty plea, the Supreme Court held, among other things, “that when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken, the defendant has made out a successful ineffective assistance of counsel claim entitling him to an appeal” and that “it is unfair to require an indigent, perhaps pro se, defendant to demonstrate that his hypothetical appeal might have had merit before any advocate has ever reviewed the record in his case in search of potentially meritorious grounds for appeal.” 528 U.S. at 484, 486 (emphasis in original). See also Garza v. Idaho, ___ U.S. ___ (139 SCt 738, 746-747, 203 LE2d 77) (2019) (explaining and reaffirming Flores- Ortega). Notably, we had recognized Flores-Ortega and properly applied its holding to the analysis of requests for out-of-time appeals from convictions after trials. See White v. State, 277 Ga. 647, 648 (594 SE2d 329) (2004). See also Collier, 307 Ga. at 366 (citing White). 21 Ringold, 304 Ga. at 883 (Nahmias, P.J., concurring). As a result, in 2019 we overruled two decades’ worth of cases from this Court and the Court of Appeals involving analysis of out-of-time appeals that conflicted with Flores-Ortega. See Collier, 307 Ga. at 376-378.8 A review of those overruled cases—40 from this Court and 46 from the Court of Appeals—reveals that, for the 20-plus years that Georgia’s appellate courts endorsed an incorrect ineffective- assistance-of-counsel standard for the analysis of motions for out-of- time appeals from guilty-plea convictions, trial courts routinely denied such motions without a hearing and appellate courts affirmed those denials in short order. Rejecting those motions was, in most cases, an unsurprising result, given that our courts were applying a standard that required a defendant seeking an out-of- time appeal from her guilty-plea conviction—a defendant who almost always was not represented by counsel—not only to allege 8 Among the cases Collier overruled was Morrow v. State, 266 Ga. 3 (463 SE2d 472) (1995), and its progeny—which held that “an appeal will lie from a judgment entered on a guilty plea only if the issue on appeal can be resolved by facts appearing in the record,” id. at 3—which was also inconsistent with Flores-Ortega. See Collier, 307 Ga. at 367-369. 22 and prove that her right to appeal was frustrated by her plea counsel’s unconstitutionally deficient performance, but also to prove, based on the limited record, that her appeal would have succeeded. Since Ringold and Collier, we have observed an increasing number of appeals docketed in this Court that stem from the denial of motions for out-of-time appeal, and many of those cases involve motions for out-of-time appeal—like the one in this case— that were filed in the trial court many years after the judgment was entered on the defendant’s guilty plea. Indeed, we have identified at least five such appeals to this Court in the past year alone. See, e.g., Boone v. State, ___ Ga. ___, ___, 2022 WL 162789, at *1 (Case No. S21A1065, decided Jan. 19, 2022); Mobuary v. State, 312 Ga. 337, 338 (862 SE2d 553) (2021); Harvey v. State, 312 Ga. 263, 263 n.1 (862 SE2d 120) (2021); Terry-Hall v. State, 312 Ga. 250, 251 (862 SE2d 110) (2021); McDaniel v. State, 311 Ga. 367, 368 (857 SE2d 479) (2021). See also Collier, 307 Ga. at 374 & n.14 (recognizing that “some of our out-of-time appeal cases have involved long delays after conviction” and citing two 2011 cases with delays of 15 and 26 years). 23 (c) Habeas Corpus Reaffirmed as the “Comprehensive Statutory Means” For Vindicating Alleged Constitutional Violations After a Final Conviction. In 2019—the same year we decided Collier—this Court also reaffirmed that “‘habeas corpus is the exclusive post-appeal procedure available to a criminal defendant who asserts the denial of a constitutional right.’” Mitchum v. State, 306 Ga. 878, 883 (834 SE2d 65) (2019) (quoting State v. Smith, 276 Ga. 14, 15 (573 SE2d 64) (2002); emphasis supplied in Mitchum). In explaining our holding in Mitchum—that an extraordinary motion for new trial “was not the appropriate vehicle” for a convicted defendant to pursue alleged constitutional deprivations because “habeas corpus provided an adequate remedy”—we explained the foundational principle that “when the General Assembly in 1967 expanded the scope of matters that could be addressed through habeas corpus to include constitutional deprivation claims, an adequate statutory remedy was created to address those constitutional claims.” Id. at 884. We concluded that the Habeas Corpus Act amounted to the “creation of a comprehensive statutory means through which constitutional (and 24 only constitutional) claims could be pursued.” Id. See also Valenzuela v. Newsome, 253 Ga. 793, 794-795 (325 SE2d 370) (1985) (explaining that OCGA § 9-14-42 (a) also allowed claims for alleged violations of state statutes until a 1982 amendment to the Habeas Corpus Act). Mitchum’s holding was consistent with prior cases in which this Court rejected attempts to use post-conviction procedures other than habeas corpus as a vehicle to vindicate a convicted defendant’s constitutional rights. See, e.g., Daker v. Ray, 275 Ga. 205, 206 (563 SE2d 429) (2002) (rejecting the writ of mandamus as a means to challenge a conviction and sentence as void and identifying habeas corpus as the exclusive remedy under the circumstances); Davis v. State, 274 Ga. 865 (561 SE2d 119) (2002) (rejecting untimely motion to withdraw guilty plea and noting that the only means available to challenge the convicted defendant’s guilty plea was habeas corpus). Nevertheless, the motion for out-of- time appeal procedure remained available for convicted defendants to seek vindication of one (but only one) of their constitutional rights—the right to an appeal as of right—in the court of their 25 conviction, as opposed to only in a habeas court after filing a habeas petition. (d) Recent Skepticism About the Trial Court Out-of-Time Appeal Procedure. It was not until Collier in 2019 that this Court first examined in a published opinion how the trial court out-of-time appeal procedure had been created, how it evolved, and how it had persisted as a parallel procedure for convicted defendants who could also use habeas corpus proceedings to vindicate their constitutional rights. Although we did not determine in Collier “whether the out-of-time appeal process in the trial court should be maintained,” we began expressing skepticism in that regard, recognizing that “the trial court process is certainly an exception to the general rule that a trial court’s jurisdiction ends following a final conviction and the end of the term of court.” Collier, 307 Ga. at 376. Our review of the history of that process made apparent that the relevant precedents were conspicuously short on reasoning and authority; Collier explained that “our out-of-time appeal jurisprudence . . . focused more on the 26 remedial purpose served by an out-of-time appeal and less on the nature of the remedy or the appropriate process for obtaining it.” Id. at 371-373. Moreover, given the concern the State raised in Collier about “long-delayed out-of-time appeal motions filed in the trial court,” id. at 370, we acknowledged that motions for out-of-time appeal in trial courts were “not directly barred by the application of any statute of limitation” and considered whether the common-law doctrine of laches (also known as “prejudicial delay”) could apply to the proceedings to bar a convicted defendant from obtaining relief, id. at 375. The majority opinion concluded that the State could “raise the defense of ‘prejudicial delay’ to out-of-time appeal motions filed in the trial court,” id. at 370, and further noted that the State could argue “and the trial court [could] consider the time periods, factors, and other criteria set out in the most analogous limitation and laches provisions—those found in the Habeas Corpus Act—in determining whether the State’s defense has merit and the 27 defendant’s motion should be dismissed,” id. at 375. 9 Four Justices did not join that portion of the Collier majority, however, characterizing the majority opinion’s endorsement of a prejudicial delay defense as “tinker[ing] at the margins of the mess we have made.” Id. at 379 (Peterson, J., concurring specially, joined by Blackwell, Boggs, and Bethel, JJ.). Not long after Collier, in Kelly v. State, we addressed another outgrowth of our out-of-time-appeal jurisprudence: our prior holding in Maxwell v. State, 262 Ga. 541, 542-543 (422 SE2d 543) (1992), that if an out-of-time appeal is granted in the trial court, the defendant “is permitted to file a second motion for new trial in order to raise claims of trial counsel ineffectiveness that could not have been raised in the initial motion for new trial.” Kelly, 311 Ga. at 829. Maxwell stated “that the grant of an out-of-time appeal permits 9 It is notable that the absence of a time limitation for the trial court out- of-time appeal procedure stands in stark contrast with other post-conviction filings in trial courts, such as a motion for new trial—which must be filed “within 30 days of the entry of the judgment,” OCGA § 5-5-40 (a), and a motion to withdraw a guilty plea—which under the common law must be filed within the same term of court, see Schoicket, 312 Ga. at 827. 28 a defendant, by the grace of the court, to start the post-conviction process anew.” 262 Ga. at 542-543. We determined that this statement in Maxwell (and later cases) was overbroad, disapproved that broad reading, and held that a defendant who had been granted a motion for out-of-time appeal in the trial court was not authorized to file a second motion for new trial that did not involve ineffective- assistance-of-counsel claims that she was “unable to assert in her initial motion for new trial.” Kelly, 311 Ga. at 829, 831. We dismissed Kelly’s appeal as a result. See id. at 831. Most recently, in Schoicket, we were faced with the question of whether to implement a “logical extension” of our precedents to allow a defendant who had obtained an out-of-time appeal in the trial court also to file an otherwise-untimely motion to withdraw her guilty plea, or whether to draw the line and stop “trail-blazing” with our “invented remedy.” 312 Ga. at 825-826, 832. In declining to extend our out-of-time appeal jurisprudence to motions to withdraw guilty pleas, Justice Peterson, writing for the majority, acknowledged that “this Court ignored contrary precedent and 29 statutes in creating out of whole cloth the motion for out-of-time appeal in the trial court.” Id. at 825. The majority opinion recognized that we “should not have invented” that post-conviction procedure, id. at 826, and that “[f]or decades now, our post- conviction jurisprudence has been described as a ‘tangle’ of ‘confusing’ procedural rules,” id. at 830. Regarding the “competing concerns involved in post-conviction relief procedures—addressing violations of a defendant’s constitutional rights on the one hand, and ensuring finality on the other,” the majority opinion pointed out that “the General Assembly has enacted habeas statutes balancing these competing concerns” and that “[w]e lack the authority to substitute our policy preferences for those of the General Assembly and thereby allow a defendant to skirt the legislatively established process.” Id. at 831. Even one of the dissents in Schoicket recognized that [a]t multiple points, we could have retracted our recognition of the out-of-time appeal procedural vehicle and once again adhered to our conclusion in Neal that the General Assembly in the Habeas Corpus Act provided an adequate remedy—and the exclusive one—for the right of appeal frustrated by the ineffective assistance of counsel. Schoicket, 312 Ga. at 838 (Ellington, J., dissenting in part). 30 3. Stare Decisis Analysis In recognition of these concerns, in this case we asked the parties and amici curiae for briefs and oral arguments addressing whether the procedural vehicle of a motion for out-of-time appeal in a trial court should continue to be available to a convicted defendant who alleges that she was deprived of her right to appeal because of her counsel’s alleged ineffective assistance, as opposed to a petition for writ of habeas corpus serving as the exclusive procedural vehicle through which to seek that remedy. That question is a difficult one: given that the judicially created motion for out-of-time appeal procedure in the trial court has existed for many years now, do principles of stare decisis require that it be perpetuated? See Nalls v. State, 304 Ga. 168, 179 (815 SE2d 38) (2018) (“Before we overrule our incorrectly decided case law on this point, we must consider whether stare decisis counsels us not to.”); Ramos v. Louisiana, ___ U.S. ___, ___ (140 SCt 1390, 1412, 206 LE2d 583) (2020) (Kavanaugh, J., concurring in part) (“[A]pplying the doctrine of stare decisis, this Court ordinarily adheres to precedent, but sometimes 31 overrules precedent. The difficult question, then, is when to overrule an erroneous precedent.”). As explained more below, we conclude that stare decisis principles do not weigh in favor of adhering to our trial court out-of-time appeal precedent and that we should revert to the law properly established by the Habeas Corpus Act and our decision in Neal. (a) Background Principles Stare decisis is a “principle of policy,” State v. Jackson, 287 Ga. 646, 658 (697 SE2d 757) (2010) (citation and punctuation omitted), under which courts “generally stand by their prior decisions, because it promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process,” Pounds v. State, 309 Ga. 376, 382 (845 SE2d 48) (2020) (citation and punctuation omitted). This Court has a longstanding tradition of considering stare decisis when evaluating whether to overrule its precedents, and we invoked such an analysis even before we articulated in Jackson a more consistent set of factors 32 we would consider in doing so. See, e.g., Leary v. Durham, 4 Ga. 593, 601 (1848); Robison v. Beall, 26 Ga. 17, 60 (1858), disapproved in part on other grounds, Fitzpatrick v. McGregor, 133 Ga. 332, 339 (65 SE 859) (1909); Adams v. Brooks, 35 Ga. 63, 66 (1866); City of Atlanta v. First Presbyterian Church, 86 Ga. 730, 732-733 (13 SE 252) (1891); Ellison v. Ga. R.R. & Banking Co., 87 Ga. 691, 696 (13 SE 809) (1891); Rogers v. Carmichael, 184 Ga. 496, 510-512 (192 SE 39) (1937); Davis v. Penn Mut. Life Ins. Co., 198 Ga. 550, 563-564 (32 SE2d 180) (1944); Humthlett v. Reeves, 211 Ga. 210, 215 (85 SE2d 25) (1954); Sharpe v. Dept. of Transp., 267 Ga. 267, 270-271 (476 SE2d 722) (1996); Etkind v. Suarez, 271 Ga. 352, 356-358 (519 SE2d 210) (1999); Harper v. State, 286 Ga. 216, 218 (686 SE2d 786) (2009). Since Jackson, we have regularly considered in our stare decisis analyses “a number of factors, including the age of the precedent, the reliance interests involved, the workability of the prior decision, and most importantly, the soundness of its reasoning.” Pounds, 309 Ga. at 382 (citation and punctuation omitted). And although the soundness, age, reliance, and 33 workability factors have provided a useful framework for such analyses, the list of factors we have considered has never purported to be exclusive. See, e.g., Jackson, 287 Ga. at 658 (“[W]e consider factors such as . . . .”) (citing Montejo v. Louisiana, 556 U.S. 778 (129 SCt 2079, 173 LE2d 955) (2009)) (emphasis supplied). “But stare decisis is not an inexorable command.” Pounds, 309 Ga. at 382 (citation and punctuation omitted). See also Woodard v. State, 296 Ga. 803, 812 (771 SE2d 362) (2015) (stare decisis is not a “mechanical formula of adherence to the latest decision”) (citation and punctuation omitted). “[I]n reconsidering our prior decisions, we must balance the importance of having the question decided against the importance of having it decided right.” Gilliam v. State, 312 Ga. 60, 62 (860 SE2d 543) (2021) (citation and punctuation omitted; emphasis in original). See also Harper, 286 Ga. at 218 (“While ‘the rule of stare decisis is a wholesome one, it should not be used to sanctify and perpetuate error. Courts, like individuals, but with more caution and deliberation, must sometimes reconsider what has been already carefully considered, and rectify their own 34 mistakes.’”) (quoting City of Atlanta, 86 Ga. at 732-733) (punctuation omitted). To that end, we have explained that the source of the precedent we are examining plays an important role in that balance: “stare decisis carries less weight when our prior precedent involved the interpretation of the Constitution,” Olevik v. State, 302 Ga. 228, 245 (806 SE2d 505) (2017), whereas the force of stare decisis is generally greater with respect to an erroneous interpretation of statutory law, see Cooper Tire & Rubber Co. v. McCall, 312 Ga. 422, 435 (863 SE2d 81) (2021) (“Considerations of stare decisis have greater weight with regard to precedents interpreting statutes than precedents regarding constitutional issues.”). We have supported this distinction by reasoning that “it is much harder for the democratic process to correct or alter our interpretation of the Constitution than our interpretation of a statute or regulation.” Gilliam, 312 Ga. at 62. Here, because the few actual precedents we are reconsidering are cases in which this Court ignored or contravened the statutory scheme established by the Habeas Corpus Act, we apply greater 35 weight to them than we would if they were rooted in our interpretation of the federal or state Constitutions. See Cooper Tire & Rubber Co., 312 Ga. at 435. However, even within the context of statutory stare decisis, when “we have misinterpreted a statute by failing to consider the statute’s language at all, stare decisis applies with less force.” Nalls, 304 Ga. at 179-180 (emphasis supplied). Thus, considerations of stare decisis apply with less force to Rowland and its progeny than they otherwise would to precedents rooted in statutory interpretation. That is because Rowland and its pre-Collier progeny utterly ignored the statutory text and context of the Habeas Corpus Act and indeed contradicted Neal’s earlier and definitive holding that correctly applied the statutory scheme set forth in the Act “without engaging in any analysis of stare decisis.” Willis v. State, 304 Ga. 686, 706 (820 SE2d 640) (2018). Keeping these principles in mind, we turn to the stare decisis analysis in this case. (b) Soundness of the Reasoning We have consistently said that the soundness of the reasoning 36 of the relevant precedent is the most important factor in the stare decisis analysis. See, e.g., Pounds, 309 Ga. at 382; Olevik, 302 Ga. at 245; Jackson, 287 Ga. at 658. As we explained above and will explain more below, our creation and endorsement of the trial court out-of-time-appeal procedural vehicle, and our precedents that flow from that creation, are unequivocally unsound. As is apparent from our review of Georgia law in Division 2 (a) above, the separate procedure that this Court created for obtaining an out-of-time appeal in the trial court has neither a statutory nor a common-law basis. The initial cases in which we considered the merits of an appeal following a trial court ruling on a motion for an out-of-time appeal, King and Furgerson, did not acknowledge the existence of the Habeas Corpus Act, let alone distinguish between motions for out-of-time appeals sought in a trial court and similar filings made in a habeas court. Rowland, on the other hand, acknowledged that the defendant in that case could seek relief in habeas, but offered no analysis of the import of that point, and was otherwise devoid of reasoning showing why a trial court out-of-time- 37 appeal procedure could or should exist. And Rowland’s progeny wholly lacked acknowledgment and analysis of this state’s post- conviction statutory scheme. Most importantly for these purposes, Rowland and its progeny not only failed to cite precedent to support our Court’s creation of a trial court out-of-time appeal procedural vehicle; they also ignored precedent from this Court (Neal) that rejected the very action that Rowland and its progeny endorsed. As Justice Peterson’s special concurrence in Collier recognized, we “created out of whole cloth” the trial court out-of-time-appeal procedure; we did so “without any analysis whatsoever”; and “there can be no doubt that our reasons—to the extent that we’ve had any— have been purely ‘policy.’” 307 Ga. at 379-381 (Peterson, J., concurring specially). 10 10 By the time we decided Schoicket two years later, eight Justices expressly agreed that we had “creat[ed] out of whole cloth” the out-of-time appeal procedure. See Schoicket, 312 Ga. at 825; id. at 841-842 (Colvin, J., dissenting in part but agreeing that we created the out-of-time-appeal procedure “out of whole cloth” and proposing that we “follow Neal,” “eliminat[e] the stand-alone out-of-time appeal procedure,” and “require prisoners seeking an out-of-time appeal and associated remedies to use the habeas procedures that the General Assembly has afforded”). 38 We have also outlined above the inconsistency of the judicially created trial court out-of-time-appeal procedure with the legal conclusion articulated in other cases, including in Neal and Mitchum, that the Habeas Corpus Act serves as the exclusive procedural vehicle in Georgia law for convicted defendants to seek relief for alleged constitutional violations relating to their convictions. See Pounds, 309 Ga. at 382 (overruling cases holding that a late-filed motion for new trial was void and an order denying it must be affirmed, in part because the reasoning of the prior precedent at issue was “inconsistent with applicable legal principles articulated in our other case law in this area”). Our cases creating the motion for out-of-time appeal in the trial court are thus “unsound and contrary to the body of our law,” Jackson, 287 Ga. at 658, and that weighs most heavily in our stare decisis analysis. 11 11 In his dissent in this case, Justice Peterson argues—as he did in his dissent in Frett—that “once we’ve determined that a decision was unsound, the other Jackson factors never seem to be particularly meaningful.” See Frett v. State Farm Workers’ Compensation, 309 Ga. 44, 63-64 (844 SE2d 749) (2020) (Peterson, J., dissenting). We disagree, and our answer today is the same one the Court provided in Frett: 39 Moreover, an unusual aspect of this case is that neither the State nor any of the amici curiae—which represent the views of both prosecutors (PAC) and criminal defense lawyers (GACDL)— contends that our judicial decisions creating or endorsing the trial court out-of-time appeal procedural vehicle are supported by sound reasoning or relevant authority. 12 In light of all of these [I]t is unsurprising that the Court would expend the judicial time and resources to write an extensive analysis of stare decisis mostly in cases in which the Court decides to depart from the usual rule that we adhere to our precedents, and it is equally unsurprising that we would say little about stare decisis in the numerous cases in which we stand by our precedents. Indeed, there are plenty of recent cases in which we have adhered to and applied our statutory precedents without any discussion of—or any express reference to—the doctrine of stare decisis, even when a party has asked us to revisit those precedents. In a perfect world in which judicial time and resources were not so limited, perhaps it would be better to write about stare decisis whenever a court unremarkably adheres to its precedents. But that is not reality for very busy courts like ours. Id. at 62 n.15 (majority opinion) (citations omitted). 12 All of the parties and amici who have weighed in on the stare decisis analysis of this Court’s cases creating the out-of-time appeal procedural vehicle agree that those precedents are unsound. The Attorney General contends that “this Court’s decisions holding that motions for out-of-time appeals are available to raise ineffective-assistance claims are plainly wrong. The rule conflicts with Georgia’s statutory scheme for post-conviction relief, as well as this Court’s other cases.” The District Attorney contends that “the soundness of the reasoning for these cases . . . has certainly been ‘confusing’ at best, if not, altogether unsound, regardless of how well-intended they are.” Amicus PAC 40 considerations, such “significant unsoundness cuts heavily in favor of overruling” our prior precedent. Duke, 306 Ga. at 184 (citation and punctuation omitted). (c) Age of Precedent, Traditional Reliance Interests, and “Entrenchment” Two other familiar stare decisis factors—age and reliance interests—are nominally distinct, but can also be intertwined. This makes sense: the older a precedent is, the more opportunity it has to become part of the legal landscape on which the public, the bench and bar, and others rely. But cf., e.g., Frett v. State Farm Employee Workers’ Compensation, 309 Ga. 44, 60 (844 SE2d 749) (2020) (pointing out, in the course of overruling an 85-year-old precedent, the opposite phenomenon and noting that we had “never cited [the states that “the basis for the creation of an out-of-time appeal is unwieldy, contradictory, and unsound, as it creates two paths to relief, contradicts statute, and does not lead to finality in judgment.” And even GACDL concedes that “[t]he out-of-time appeal is a jerry-rigged remedy, illegitimate but practical,” and that “[t]here can be little doubt at this point that [this] Court was without authority to craft the out-of-time-appeal [procedure] in the first instance.” We note that in her supplemental brief, Cook appears to focus her stare-decisis analysis exclusively on Schoicket instead of on the earlier out-of- time-appeal precedent discussed above in Division 2 (a), and as a result does not offer her views on that precedent. 41 precedent] for its specific holding . . . and ha[d] not cited [the precedent] for any proposition at all in the past 60 years”) (emphasis in original). Rarely, if ever, is the age of a precedent itself dispositive in a stare decisis analysis, and that is so here. See Southall v. State, 300 Ga. 462, 468 (796 SE2d 261) (2017) (“[W]ithout more, that we have been wrong for many years is no reason to persist in the error.”) (citation and punctuation omitted). As discussed above in Division 2 (a), the first actual precedent approving and applying the trial court out-of-time appeal procedure was Rowland, a 27-year-old case that we can hardly call “ancient.” Willis, 304 Ga. at 705 (citation and punctuation omitted). We have overruled statutory precedents of comparable, and sometimes older, ages. See, e.g., Duke, 306 Ga. at 184 (overruling 19-year-old precedent); City of Cumming v. Flowers, 300 Ga. 820, 832 (846 SE2d 48) (2020) (overruling 21-year- old precedent); Woodard, 296 Ga. at 812 (overruling 24-year-old precedent); State v. Burns, 306 Ga. 117, 123-124 (829 SE2d 367) (2019) (overruling 30-year-old precedent); Jackson, 287 Ga. at 659 42 (overruling precedent that was “nearly three decades old”); Southall, 300 Ga. at 468 (overruling 45-year-old precedent). We thus move to an examination of the broader reliance interests related to our trial court out-of-time-appeal precedents. When it comes to reliance interests, we have traditionally looked to whether the precedent at issue affects property or contract issues, and whether it establishes a substantive right; reliance interests are at their apex when they involve these types of interests. See, e.g., Savage v. State, 297 Ga. 627, 641-642 (774 SE2d 624) (2015) (“[S]tare decisis is especially important where judicial decisions create substantial reliance interests, as is more common with rulings involving contract and property rights.”) (citing cases). See also Olevik, 302 Ga. at 245 (“Substantial reliance interests are an important consideration for precedents involving contract and property rights, ‘where parties may have acted in conformance with existing legal rules in order to conduct transactions.’”) (citation omitted). Compare Jackson, 287 Ga. at 658 (explaining that the precedent at issue “affects no property or contract issues and 43 establishes no substantive rights, so it creates no meaningful reliance interests”); Nalls, 304 Ga. at 180 (same). No such reliance interests are at stake here. We have repeatedly held that precedents—like the ones at issue here— relating to procedures, including post-conviction procedures, create no substantive rights “in which anyone has a significant reliance interest.” Duke, 306 Ga. at 184-185 (precedent that “disregard[ed]” the statutory requirement for a certificate of immediate review before pursuing an interlocutory appeal did “not involve substantial reliance interests”). See also, e.g., Pounds, 309 Ga. at 382 (“[T]he issues involved are ones of appellate procedure, not . . . substantive rights in which anyone has a significant reliance interest.”) (citation and punctuation omitted); Willis, 304 Ga. at 706 (“As to the reliance issues potentially at stake, we note that the holdings in [our prior precedents] are procedural in nature and establish no substantive rights.”) (citation and punctuation omitted). Moreover, Rowland and its progeny did not establish any substantive rights in the first instance; they merely created an alternative procedural vehicle— 44 albeit a legally unauthorized one—for alleging the violation of a preexisting constitutional right (i.e., the effective assistance of counsel in filing a timely appeal as of right) and for obtaining a constitutionally necessary remedy if that right was violated. 13 Eliminating this alternate procedure would not extinguish a convicted defendant’s ability to vindicate an alleged violation of her right to appeal due to constitutionally ineffective assistance of counsel, because the General Assembly long ago established a habeas corpus procedure through which a convicted defendant can vindicate that very same constitutional claim. 14 13 While GACDL concedes that “[p]rocedural rules do not normally create the kinds of reliance interests that stare decisis is concerned with because they affect no property or contract issues and establish no substantive rights,” (citations and punctuation omitted), it nonetheless points to a number of non- traditional concerns that it characterizes as reliance interests in the trial court out-of-time-appeal system. Those include “the incalculable value of the time lost waiting for a habeas court that is a stranger to the parties . . . to decide in a formal proceeding what a sentencing court can decide promptly on a motion” and the fact that defendants are not required to pay a filing fee for out-of-time appeals, whereas petitions for habeas corpus require filing fees and other costs. Compare OCGA § 15-6-77 (h), (k) with OCGA § 24-13-25. These and other factors are considered below in Division 3 (d) as part of our workability analysis. 14To the extent GACDL postulates that, if the trial court out-of-time appeal vehicle were eliminated, some number of convicted defendants would 45 We have also considered, however, a different type of reliance interest in some of our stare decisis analyses: the “entrenchment” of the precedent in the legal system. See, e.g., Frett, 309 Ga. at 60 (noting that the precedent at issue had not “become deeply entrenched in our jurisprudence”). See also Williams v. Harvey, 311 Ga. 439, 451 (858 SE2d 479) (2021) (the holdings in prior precedents are “neither ancient nor entrenched within our judicial system”) (citation and punctuation omitted); Flowers, 300 Ga. at 831-832 (the prior precedent is “neither ancient nor entrenched”) (citation and punctuation omitted). Whether entrenchment constitutes a species of reliance or whether it is instead a separate factor in the stare decisis analysis is of no moment; we see no reason we cannot consider it here. The dissent seizes on the concept of entrenchment and focuses almost myopically on it in its stare decisis analysis. Characterizing lose their appeal as of right and might also have missed the relevant statute of limitation in habeas corpus, we note that the General Assembly can grant relief from the habeas statute of limitation if it deems any such potential result unjust. 46 entrenchment “in a narrow sense [as whether a precedent’s] relevant holding has been applied frequently,” and more broadly as “potential disruption to the legal system that might be caused by suddenly jettisoning a particular precedent,” it contends that “the motion for out-of-time appeal has become deeply entrenched.” GACDL shares the disruption concern, characterizing the trial court out-of-time appeal procedure as “one thread in Georgia’s Gordian knot of criminal-appellate and post-conviction practice.” We cannot say, however, that entrenchment of the trial court out-of-time-appeal procedure in Georgia weighs so heavily in the stare decisis analysis that we should retain our erroneous precedents. First, the actual number of precedents on the trial court out-of-time appeal procedure in Georgia—i.e., cases holding that a motion for an out-of-time appeal in the trial court was a proper procedure such that the case needs to be overruled if the trial court out-of-time appeal procedure were eliminated—are few. That limited universe of cases includes Rowland—our first real precedent addressing and approving the trial court out-of-time appeal 47 procedure—and a handful of progeny cases, a fact the dissent glosses over by shifting focus to the 14 cases that were docketed in our Court last year after a trial court granted a motion for out-of-time appeal. 15 The dissent points to these cases, which reference trial court grants of motions for out-of-time-appeals in the first footnote of this Court’s opinions recounting the procedural history of murder cases, and contends that our trial court out-of-time appeal precedent “has become deeply entrenched” because “it is regularly applied in a significant number of cases.” We acknowledge the value—at least in a case like this one—in 15 We have identified only five cases—the last one decided more than 15 years ago—that appear expressly to endorse the filing of an out-of-time appeal in the trial court and thus constitute precedent for purposes of a stare decisis analysis. See Carr v. State, 281 Ga. 43, 43-44 (635 SE2d 767) (2006) (extensively quoting Rowland and directing Carr that, “should she wish to appeal her convictions, she must file a request for an out-of-time appeal in the trial court”); Cody v. State, 277 Ga. 553, 554 (592 SE2d 419) (2004) (stating that Cody “has the option of applying for an out-of-time appeal in the court of conviction”); Fulton v. State, 277 Ga. 126 (587 SE2d 20) (2003) (“Fulton may seek an out-of-time appeal in the trial court.”); Wicks v. State, 277 Ga. 121, 122 (587 SE2d 21) (2003) (“If Wicks wishes to pursue appellate relief, he can seek an out-of-time appeal in the trial court.”), disapproved on other grounds, Pounds, 309 Ga. at 378 n.4; Porter v. State, 271 Ga. 498, 500 (521 SE2d 566) (1999) (“To obtain an out-of-time appeal, Porter must apply for that relief in the trial court . . . . ”), disapproved on other grounds, Pounds, 309 Ga. at 378 n.4. 48 identifying potential disruption to the legal system that might result if a precedent is overruled. We are not unmindful of the practice that has built up around convicted defendants seeking motions for out-of-time appeal in trial courts and of appellate courts applying Rowland and its progeny to decide cases following trial-court rulings on motions for out-of-time appeal. As GACDL casts it, “the criminal legal system has arranged itself around the availability of the direct, out-of-time remedy.” But we are not willing to weight this type of entrenchment over all other factors, because the regular application of a procedural precedent—even in a “significant number of cases,” as the dissent suggests—is not dispositive of a stare decisis analysis, particularly where another procedure—the one we have explained is exclusive—is available. Nor is the second entrenchment-related point the dissent makes dispositive: that when our out-of-time appeal precedents are applied, “it often makes a substantial difference,” because when such a motion is granted “it permits an appeal that would otherwise be barred without the years-long delay of habeas.” To be sure, when 49 a trial court grants an out-of-time appeal, it allows a defendant to proceed with an appeal that would otherwise be time-barred. But that does not support the dissent’s entrenchment argument, because a convicted defendant may obtain the same result by filing a petition in a habeas court using the statutorily authorized procedure. That procedure—habeas corpus—exists now and will continue to exist even if we eliminate the trial court out-of-time appeal procedure. The dissent’s stronger claim—which we address more fully below as part of our workability analysis—is that motions for out-of- time appeals in trial courts are quicker and more efficient than habeas corpus proceedings. That may well be true in many cases. But habeas corpus is the procedure the General Assembly has established to provide the remedy for the unconstitutional deprivation of the right to appeal; this Court has determined that “habeas corpus is the exclusive post-appeal procedure available to a criminal defendant who asserts the denial of a constitutional right,” Mitchum, 306 Ga. at 883 (citation and punctuation omitted; emphasis in original); and even if the alternate procedural vehicle 50 we created may result in requests for out-of-time appeals being decided faster and more efficiently, that alone is not a reason to retain that procedure, even if it has become “entrenched.” (d) Workability That brings us to workability. As the dissent notes and as we discuss below, some aspects of the trial court out-of-time appeal procedure we created may be more workable or even preferable (at least in the view of some Justices) to the statutorily authorized habeas corpus process. But there is a fundamental—and in our view, insurmountable—workability problem with our precedents: there is no end in sight to our Court being asked to fill in the details of the trial court out-of-time appeal procedure we created. See Harper, 286 Ga. at 217-218 (overruling prior precedent in part because of workability problems stemming from the lack of “rules or precedents guiding individuals in the filing of, or courts in their consideration of,” the post-appeal procedure that our Court had created in a prior case). Administration of the Habeas Corpus Act, by contrast, does not 51 require judges to establish such rules in the first instance; the General Assembly has already established them by statute. On that point, the Habeas Corpus Act—the General Assembly’s chosen procedure for remedying alleged violations of convicted defendants’ constitutional rights, including claims that a right to appeal was frustrated because of a constitutional violation—establishes that the superior court in the county of the defendant’s detention has “exclusive jurisdiction” over the defendant’s habeas petition, OCGA § 9-14-43; contains requirements for what petitions must include, OCGA § 9-14-44, and how they must be served, OCGA § 9-14-45; establishes deadlines for answering a petition, OCGA § 9-14-47; and lays out how hearings must operate, OCGA § 9-14-48, what the habeas court must put in writing to support its judgment, OCGA § 9- 14-49, and how that judgment must be appealed, OCGA § 9-14-52. The Act imposes definite time limits within which petitions must be brought: it requires a defendant to challenge a felony conviction not involving a death sentence within four years of “the conclusion of direct review or the expiration of the time for seeking such review” 52 and to challenge a misdemeanor within one year, OCGA § 9-14-42 (c)—statutory limitations periods that are not subject to equitable tolling, see Stubbs v. Hall, 308 Ga. 354, 369 (840 SE2d 407) (2020). It provides for a statutory defense of laches, OCGA § 9-14-48 (e), and contains a bar on successive habeas petitions, OCGA § 9-14-51. Moreover, the Civil Practice Act OCGA § 9-11-1 et seq., generally applies to habeas corpus proceedings. See OCGA § 9-11-81; Mitchell v. Forrester, 247 Ga. 622, 623 (278 SE2d 368) (1981) (“[T]he CPA now applies to habeas corpus applications.”).16 Yet no corresponding requirements for motions for out-of-time appeals in trial courts are clearly applicable—at least not from the common law, statutes, or court rules, which do not establish or recognize such motions. The only rules that govern the operation of 16 Unlike in habeas proceedings, where the Civil Practice Act applies, we have not clearly answered what, if any, statutory authority related to court proceedings applies to trial court out-of-time appeal proceedings. To the contrary, we have stated that courts and parties should look to the Habeas Corpus Act for guidance—at least with respect to issues related to laches—but even then we have not directly applied the Act’s statutory limitations to the trial court out-of-time appeal procedure. See Collier, 307 Ga. at 374-375. This unanswered question makes it more likely that legal questions about the limitations of the trial court out-of-time appeal procedure will be raised again in our Court. 53 the trial court out-of-time appeal procedure are rules that trial judges formulate and appellate judges approve, modify, or reject if challenged. Every time that is done, judges again overstep their authority and re-engage in a policy-making exercise that is typically reserved for legislators. Faced with questions like whether to try to mirror the requirements of the Habeas Corpus Act, follow requirements for other types of motions filed in trial courts, or make up rules we think are the best as a matter of policy, we have over the years created or endorsed a hodge-podge of guidelines for the trial court out-of-time appeal procedure on a case-by-case basis. See, e.g., Collier, 307 Ga. at 373 (noting that “the body of case law governing procedures applicable to . . . motions [for out-of-time appeal in trial courts] is far less developed than” the Habeas Corpus Act and habeas case law and that “we have . . . addressed defenses to such motions [for out-of-time appeal] as the State has raised them”). See also id. at 373 n.12 (citing prior cases applying the doctrines of res judicata and collateral estoppel to motions for out- 54 of-time appeal in the trial court).17 And the reality is that we have little idea what rules trial courts are applying, or if they are applying whatever rules they have established with any degree of uniformity or consistency. As Justice Ellington has observed, “by allowing out- of-time appeals, we have, repeatedly, already substituted our policy preferences for those of the General Assembly and allowed defendants to skirt the legislatively-established habeas process.” Schoicket, 312 Ga. at 839 (Ellington, J., dissenting in part). For example, because no deadline or statute of limitation expressly restricts a convicted defendant’s ability to file a motion for out-of-time appeal in a trial court, in Collier the State argued that the Court should “abolish the practice of allowing defendants to file a motion for an out-of-time appeal in the trial court” to prevent defendants from filing motions for out-of-time appeal years or 17 Unless we, as a matter of policy, import all of the requirements of the Habeas Corpus Act into the trial court out-of-time appeal procedure, that procedure may allow a convicted defendant who seeks relief through a motion for out-of-time appeal in a trial court to avail herself of a completely different— and potentially less stringent—set of requirements than if she filed the very same claim as part of a petition for habeas corpus. 55 decades after their convictions, when the State may have lost the evidence needed to defend against the motion. 307 Ga. at 369. The Collier majority noted that “determining whether the out-of-time appeal process in the trial court should be maintained would involve a complex stare decisis analysis” and that those “issues ha[d] not been fully briefed” in that case. Id. at 376. The majority instead concluded that a laches-like “prejudicial delay” defense could potentially provide the State with a defense against unreasonably delayed motions for out-of-time appeal filed in trial courts. Id. at 370 (“[W]e hold that the State may raise the defense of ‘prejudicial delay’ to out-of-time appeal motions filed in the trial court.”). See also id. at 374 (“When a defendant files a motion for an out-of-time appeal in the trial court, the State may argue that the defendant’s delay in doing so has unduly prejudiced the State’s ability to respond to the motion.”); id. at 375 (“[T]he trial court may consider the time periods, factors, and other criteria set out in the most analogous limitation and laches provisions—those found in the Habeas Corpus Act”). Four Justices expressed doubt about that conclusion, 56 however, and concurred in the judgment of the prejudicial-delay portion of the majority opinion only “to the extent that it correctly observe[d] that none of our cases have held that the State cannot assert a defense of prejudicial delay.” Id. at 381 (Peterson, J., concurring specially, joined by Blackwell, Boggs, and Bethel, JJ.). How exactly a prejudicial-delay defense would work in these cases is an open question, and the answer is anyone’s guess. 18 What is more certain is that this Court ultimately (and repeatedly) will have to make up the details if we retain the trial court out-of-time appeal procedure, which requires these sorts of questions to be answered by judges. The questions did not stop after Collier. If anything, Collier begged even more questions, which Schoicket illustrates: the defendant in that case asked us to decide whether we would 18 Justice Peterson goes a step further today in his dissent by suggesting that, because the General Assembly added a laches defense to the Habeas Corpus Act in 2004 and did nothing to modify motions for out-of-time appeal in trial courts, laches does not apply to our judicially created procedure— meaning that a motion for out-of-time appeal may be filed decades after a final conviction. And Justice Ellington joins the dissent, even though he authored Collier. 57 “logically extend” our holding in Collier, as Justice Peterson’s special concurrence had hinted we might. In Schoicket, we were forced to grapple with whether a granted out-of-time appeal not only authorizes a defendant to file what would otherwise be an untimely appeal, but also to file an untimely motion to withdraw a guilty plea—which would also allow the defendant to expand the record that she could use to assert previously unraised claims. See Schoicket, 312 Ga. at 825; Collier, 307 Ga. at 380 (Peterson, J., concurring specially). We had held in Collier that whether a defendant “seeks an out- of-time appeal from a final judgment of conviction entered following a trial or following a guilty plea” did not matter, and we overruled cases that held otherwise. See Collier, 307 Ga. at 366-367. Nevertheless, in Schoicket we decided that the very same distinction—whether a defendant was adjudicated guilty by trial versus by entering a plea—actually did matter, because we did not allow the defendant in that case, who had been granted an out-of- time appeal after judgment of conviction was entered on her guilty 58 plea, the relief she sought: a motion to withdraw her guilty plea. See Schoicket, 312 Ga. at 832; see also Boone, ___ Ga. at ___ (relying on Schoicket to hold that “a grant of an out-of-time appeal would not entitle Boone to pursue an otherwise-untimely motion to withdraw a guilty plea”). Even though we had acknowledged that “permitting such a motion would be a logical extension of our precedent that invented certain post-conviction remedies,” and that the Court of Appeals had done just that in Dawson v. State, 302 Ga. App. 842, 843 (691 SE2d 886) (2010), and Sosa v. State, 352 Ga. App. 637, 639 & n.1 (835 SE2d 695) (2019), we rejected the extension of our precedent and overruled Dawson and Sosa. See Schoicket, 312 Ga. at 833 n.9. We concluded that “we should not have invented those remedies in the first place” and “decline[d] to invent additional remedies that might further complicate our post-conviction jurisprudence.” Id. at 826. See also id. at 837, 839 (Ellington, J., dissenting in part) (arguing that there is no “principled reason to deny guilty-plea defendants access to the procedural tool we created” and contending that we “should ensure that access to the 59 misbegotten procedure is provided in an even-handed manner” by allowing defendants who are granted an out-of-time appeal to withdraw their guilty pleas); Collier, 307 Ga. at 380 (Peterson, J., concurring specially) (noting that allowing defendants to move to withdraw guilty pleas after being granted out-of-time appeals would “appear to be merely a logical extension” of our out-of-time appeal precedents). These examples of the issues we have faced point to another aspect of this workability problem: the lines we have drawn in our trial court out-of-time appeal precedents are often based on considerations of policy rather than law. The line we drew in Schoicket—where we acknowledged that the defendant’s requested extension of our out-of-time appeal precedents was logical and anticipated by the Court, but nonetheless rejected it to avoid “further complicat[ing] our post-conviction jurisprudence,” 312 Ga. at 826—amply illustrates this concern. Indeed, we have already admitted that “our inventions” in this area “have never purported to be even-handed.” Id. at 832 (emphasizing that “a motion for out-of- 60 time appeal can be granted only if one particular kind of ineffective assistance of counsel claim succeeds, and all sorts of other ineffectiveness claims not raised on direct appeal can be brought only in habeas”). The longer our trial court out-of-time appeal precedents persist—and the more these issues are highlighted by cases like Collier, Kelly, and Schoicket—the more likely it is that we will have to continue to draw lines to define the procedure’s boundaries, including to “retreat from ‘broad statements’” we have made in some of our precedents and to “avoid dispensing unwarranted windfalls.” Schoicket, 312 Ga. at 825 (acknowledging that “following our decision in Collier, we have retreated from broad statements about the effect of a granted out-of-time appeal in order to avoid dispensing unwarranted windfalls”). In this regard, we have already identified two additional issues that arise in the trial court out-of-time appeal procedure—how claims of ineffective assistance of counsel may be presented, and when defendants are legally entitled to appointed counsel—that will likely require this Court at some point to provide answers about 61 apparent conflicts between existing precedent and practices that appear to be common in the trial court out-of-time appeal procedure. As the dissent points out, in some cases in which a convicted defendant’s lawyer misses an appellate filing deadline inadvertently and notices the oversight not long thereafter, it appears that the lawyer files a motion for out-of-time appeal in the trial court admitting his error, the State concedes that the motion should be granted, and the trial court grants the motion. See, e.g., Waller v. State, 311 Ga. 517, 518 n.1 (858 SE2d 683) (2021) (“Appellant filed a motion for out-of-time appeal through trial counsel.”); Swan v. State, 276 Ga. App. 827, 829 (625 SE2d 97) (2005) (“Trial counsel filed a motion for an out-of-time appeal . . . .”); Brown v. State, 199 Ga. App. 856, 856 (406 SE2d 516) (1991) (“The record shows defendant’s trial counsel filed a motion for out-of-time appeal . . . .”). Asserting a claim of ineffective assistance in that way may be a relatively fast and efficient means of getting the appellate process moving again. But that process appears to conflict with this Court’s precedent holding that a lawyer cannot assert his own ineffective 62 assistance. See, e.g., Hood v. State, 282 Ga. 462, 463 (651 SE2d 88) (2007) (“Because a lawyer may not ethically present a claim that he/she provided a client with ineffective assistance of counsel, a claim of ineffective assistance of trial counsel cannot be pursued unless trial counsel is no longer representing the convicted defendant.”) (citation omitted; emphasis supplied). See also Garland v. State, 283 Ga. 201, 203 & n.2 (657 SE2d 842) (2008) (explaining that “[c]ounsel prosecuting an ineffective assistance claim must be free to operate independently of the attorney whose performance is in question” and that “Georgia law has thus decisively rejected the position . . . that trial counsel is not only competent to evaluate the ineffectiveness of his/her own performance, but is ‘actually in a superior position to do so’”); Delevan v. State, 345 Ga. App. 46, 49-51 (811 SE2d 71) (2018) (vacating the trial court’s denial of defendant’s motion for an out-of- time appeal filed by the lawyer who was allegedly ineffective, concluding that “[t]he Supreme Court of Georgia has repeatedly held that an attorney may not ethically present a claim that [he] provided 63 a client with ineffective assistance of counsel. It necessarily follows that a claim of ineffective assistance of counsel may not be pursued unless the counsel at issue is no longer representing the defendant and, instead, the defendant either is represented by conflict-free counsel or represents himself pro se.”) (citation and punctuation omitted). And there’s more. When questions arise about the propriety of counsel raising an ineffective assistance claim, the trial court can attempt to ensure that a new lawyer—one who does not have a conflict of interest that precludes him from asserting his own ineffectiveness—is appointed to represent the defendant. See, e.g., Garland, 283 Ga. at 203. But that creates yet another issue in the out-of-time appeal context, because our precedent squarely holds that a motion for out-of-time appeal is a proceeding as to which a defendant is not entitled to the appointment of counsel. See, e.g., Davis v. State, 310 Ga. 547, 548 (852 SE2d 517) (2020) (‘“[B]ecause a motion for an out-of-time appeal cannot be construed as part of a criminal defendant’s first appeal of right, [the defendant] was not 64 entitled to the assistance of appointed counsel.’”) (quoting Pierce v. State, 289 Ga. 893, 894 (717 SE2d 202) (2011)).19 And that leads to the question of whether trial courts and public defenders abiding by our precedent on a convicted defendant’s right to counsel should appoint new counsel to handle motions for out-of-time appeal asserting that the defendant was deprived of an appeal of right by her prior counsel’s ineffective assistance. 20 Despite this tension, motions for out-of-time appeal appear to be granted routinely in this situation. See, e.g., Clark v. State, 309 Ga. 566, 566 n.1 (847 SE2d 160) (2020) (“Clark filed a motion for an out-of-time appeal through 19As we discuss below, a defendant also has no right to appointed counsel in a habeas corpus proceeding. We add that in describing the current state of the law, we do not mean to suggest that, as a matter of policy, the General Assembly cannot or should not provide funding for lawyers to be appointed for indigent defendants to pursue apparent violations of those defendants’ constitutional rights, even when such funding is not constitutionally required. But policy decisions such as those must be left to the Legislative Branch, and not made by judges. 20 Likewise, if a new lawyer is appointed to take over a convicted defendant’s appeal as of right and discovers that previous counsel missed the jurisdictional deadline to file a motion for new trial or a notice of appeal, the public defender arguably can decline to extend representation on the basis that the defendant is not entitled to it, and the trial court would have a basis to uphold that decision. 65 new counsel, which the trial court granted.”); Kidd v. State, 304 Ga. 543, 543 n.1 (820 SE2d 46) (2018) (“Kidd filed another motion for an out-of-time appeal, this time with appointed counsel.”). As with the other examples of unanswered questions mentioned above, we do not know with certainty what the resolution of these issues would be if presented properly in a case before this Court. But it seems likely that at some point, based on our precedent, a trial court will dismiss a motion for out-of-time appeal in which a lawyer asserts his own ineffectiveness, or uphold a public defender’s decision not to appoint new counsel for a convicted defendant whose trial counsel was constitutionally ineffective in failing to file a notice of appeal; the defendant will appeal that ruling; and this Court will have to decide what to do. We could apply our precedent as it would seem to apply in other contexts, and thereby undermine many of the benefits of the trial court out-of- time-appeal procedure on which the dissent relies heavily, or we could endorse additional deviations from precedent to maintain what we consider to be preferable policy results for the procedure we 66 created. What this Court cannot do, as long as the motion for out- of-time appeal procedure is maintained, is avoid making decisions about these sorts of issues. The entire Court agrees that we erred, and that we overstepped our limited constitutional role, by creating the trial court out-of-time appeal procedure. But we cannot simply wash our hands of that error; the need to determine the details of the procedure we created will require us (and trial judges) to perpetually overstep our judicial bounds. That makes our precedent unworkable, as this Court has repeatedly held in similar situations. See Duke, 306 Ga. at 186 n.4 (that the precedent at issue was “in need of refining only emphasizes the unworkable nature of [the precedent] as it was decided”); Harper, 286 Ga. at 217-218 (prior precedent that created the motion to vacate a criminal conviction, which was “a new post-appeal procedure for challenging a criminal conviction, . . . proved unworkable inasmuch as Georgia law is silent as to the procedural framework and rules applicable to this newly created remedy,” especially given that “[u]nlike the myriad rules governing previously 67 recognized and statutorily created procedures for challenging a criminal conviction, there are no rules or precedents guiding individuals in the filing of, or courts in their consideration of, post- appeal motions to vacate a criminal conviction”). The dissent pushes back, contending that we should not focus only on the workability of the precedents we are considering overruling, and that we should instead compare the workability of the current, erroneous precedents with the workability of the alternative: here, the statutory habeas scheme that will function as the exclusive procedure for convicted defendants to assert the denial of constitutional rights if the trial court out-of-time appeal procedure is eliminated. 21 21 We note that, although this Court has never held that this type of comparison is a required part of our stare decisis analysis, such a comparison is not necessarily inconsistent with our usual stare decisis analysis; if we examine the workability of one precedent, we at least implicitly comment on the workability of its alternatives. And we have even drawn explicit comparisons in some prior stare-decisis analyses. See, e.g., State v. Burns, 306 Ga. 117, 124 (829 SE2d 367) (2019) (“[W]e cannot say that any ‘workability’ of [the existing precedent] is sufficient to preserve the precedent. Although [that precedent’s] bright-line test is not ‘unworkable,’ neither is the alternative— applying the familiar and usual rules of evidence, which trial courts routinely do every day.”); Flowers, 300 Ga. at 833 (“[A]lthough the local-ordinance 68 In that vein, the dissent compares the current out-of-time appeal system to a world in which only habeas exists and argues that the current out-of-time appeal system is more workable. These arguments, many of which GACDL also highlights, can be organized into three primary claims: (1) that the current out-of-time appeal system is more efficient for resolving a criminal defendant’s claim that ineffective assistance of trial counsel resulted in a loss of the defendant’s appeal as of right; (2) it is more likely under the current system that an indigent defendant will have access to counsel when filing the defendant’s claim of ineffective assistance of counsel; and (3) eliminating the out-of-time appeal procedure will result in a significant shift in workload among government entities—especially with respect to district attorney offices, the Attorney General’s office, public defenders, and judges. None of those arguments is requirement is not unworkable, it is not as workable as the correct rule.”); Lejeune, 296 Ga. at 298 (“[T]he usual rule in habeas cases—that the petitioner bears the burden of proof—is more workable than the rule of Purvis and its progeny.”); Ga. Dept. of Nat. Resources v. Ctr. for a Sustainable Coast, 294 Ga. 593, 602 (755 SE2d 184) (2014) (“[A] bright line rule that only the Constitution itself or a specific waiver by the General Assembly can abrogate sovereign immunity is more workable than IBM v. Evans’ scheme allowing judicially created exceptions.”). 69 availing. With respect to its first set of arguments, the dissent contends that the out-of-time appeal procedure is more efficient because, among other reasons, the same judge who presided over the defendant’s trial can decide the motion for out-of-time appeal, which in turn avoids having to transfer the record between a trial court and a habeas court and also reduces travel costs for everyone involved. These practical observations may be correct, but the differences between the Habeas Corpus Act and our out-of-time appeal procedural vehicle reflect policy choices the General Assembly made in enacting the former. To the extent the Habeas Corpus Act can or should be improved to make post-conviction proceedings more efficient, or a trial court out-of-time appeal process for certain cases should be established, the General Assembly can make those changes legislatively. See Duke, 306 Ga. at 186 (recognizing that if “the General Assembly determines that the established framework does not adequately safeguard the interests of litigants in particular classes of cases, it is for that body to change 70 it”). Cf. OCGA § 9-14-43 (amending the Habeas Corpus Act in 2004 to require petitioners not in custody or in federal or foreign custody to file habeas petitions in the superior court of the county in which they were convicted). The dissent similarly argues that the trial court out-of-time appeal procedure is faster than habeas—particularly when counsel has missed a jurisdictional filing deadline by only a few days or weeks, the parties involved in the case are willing to consent to an out-of-time appeal, and the defendant is able to proceed to her motion for new trial or appeal as of right expeditiously. 22 That, too, may be true, although—as discussed above—it appears to be inconsistent with our precedent on counsel raising her own alleged ineffectiveness. But putting aside the anomaly of trial courts permitting counsel to raise their own ineffective assistance against 22 We note that, to the extent that the filing deadline is missed by a short period of time and the term of court in which the judgment of conviction was entered still has not expired, the trial court may set aside and reenter the judgment to allow a timely post-trial motion or appeal to be filed. See Gray, 310 Ga. at 262. The availability of this historical practice, which is rooted in the common law, should alleviate some of the dissent’s concern about ineffective counsel missing jurisdictional deadlines by just a few days or weeks. 71 themselves, there are also many instances in which out-of-time appeals are sought in trial courts not to quickly correct oversights, but rather to appeal years- or decades-old convictions. See, e.g., Collier, 307 Ga. at 374 (“[S]ome of our out-of-time appeal cases have involved long delays after conviction.”); id. at 374 n.14 (citing two cases where the lapses of time were 15 and 26 years); Sims v. State, 312 Ga. 303, 304 (862 SE2d 507) (2021) (“[A]lmost seven years later, Sims filed a pro se motion for an out-of-time appeal.”); Harvey, 312 Ga. at 265 (“Around 15 years later, Harvey filed a motion for an out- of-time appeal through new counsel.”); Davis, 310 Ga. at 548 (“Davis filed a pro se motion for an out-of-time appeal” over 20 years after his conviction). These cases not only ignore the statutory time limits the Habeas Corpus Act imposes on the ability to seek the same relief in habeas—a legislative choice the General Assembly made to promote the finality of convictions; they also raise additional logistical concerns, such as the availability of witnesses and evidence to resolve the ineffective-assistance-of-counsel claim underlying the motion. See, e.g., Schoicket, 312 Ga. at 837 72 (Ellington, J., dissenting in part) (acknowledging the State’s delay- related concerns such as “lost or destroyed records and dead or forgetful witnesses” in the context of motions to withdraw guilty pleas being filed pursuant to granted out-of-time appeals); Collier, 307 Ga. at 375 (“[E]liminating post-conviction delay before appeal is an important interest in our criminal justice system.”). The dissent also contends that defendants often have more “access to counsel” in the trial court out-of-time appeal procedure than they do in habeas. But that concern is largely undermined by the dissent’s acknowledgement that indigent defendants do not have a right to counsel in either scenario. See Davis, 310 Ga. at 548 (“‘Because a motion for an out-of-time appeal cannot be construed as part of a criminal defendant’s first appeal of right, defendant was not entitled to the assistance of appointed counsel.’”) (quoting Pierce v. State, 289 Ga. 893, 894 (717 SE2d 202) (2011)) (punctuation omitted); Gibson v. Turpin, 270 Ga. 855, 857 (513 SE2d 186) (1999) (“It is well settled that there is no federal or state constitutional 73 right to appointed counsel in Georgia habeas corpus proceedings.”). 23 Indeed, the dissent’s contention that indigent defendants, “as a practical matter,” have counsel more frequently in the trial court out-of-time appeal procedure than in habeas “when a motion is filed to correct a missed deadline” seems to be a veiled reference to the issues we have discussed above: counsel asserting claims of ineffective assistance against themselves or being appointed in proceedings where the defendants have no entitlement to appointed counsel. These practices may be challenged now that they have been highlighted. Finally, the dissent expresses concern that eliminating the alternative trial court out-of-time appeal procedure and relying 23 We share the dissent’s concern that, if motions for out-of-time appeals are eliminated in trial courts, some pro se defendants—or even attorneys who are familiar with the trial court out-of-time appeal procedure—may not initially appreciate the risk of waivers and bars to successive habeas claims they may face if they file a habeas petition that complains only of the ineffective assistance of counsel that resulted in the defendant’s frustrated appeal, rather than all of the defendant’s constitutional claims. But the requirements of habeas corpus are well-established, both in statute and in case law interpreting it, so it rings hollow to suggest that being required to use the exclusive statutory remedy that the General Assembly established decades ago is somehow a trap for the unwary. 74 exclusively on habeas corpus would necessitate a significant shift in resources within the criminal justice system. The dissent notes, for example, that motions for out-of-time appeals that are typically heard by trial courts and handled by attorneys from district attorney offices would instead be heard by habeas courts and handled by attorneys from the Attorney General’s office, and projects that those changes will result in staffing and workload (and thus financial) implications. We are certainly mindful of the resources required for government entities—including courts—to get their work done. But resource considerations such as these must be directed to the General Assembly, whose constitutional duty includes appropriating funds for the operations of our State’s government. See Ga. Const. of 1983, Art. III, Sec. IX, Par. II, III. See also Schoicket, 312 Ga. at 839 (Ellington, J., dissenting in part) (“The General Assembly is fully capable of resolving how to allocate funds to make the system work. Indeed, the General Assembly would be freed from having to appropriate, in addition to funds to make the 75 post-conviction system work as it intended when it adopted the Habeas Corpus Act over 50 years ago, additional substantial funds to make our judicially-created parallel system work.”) (emphasis in original). Moreover, representatives of two of the primary stakeholders who would be affected by any anticipated shift in workload—the Attorney General and the District Attorney—have asked this Court to overrule our trial court out-of-time appeal precedents, specifically arguing that those precedents are unworkable. We should credit the views of the constitutional officers who have responsibility for both the trial court and habeas out-of-time appeal processes over our own views of their abilities to discharge their duties in a world where motions for out-of-time appeals in trial courts do not exist. The dissent greatly underestimates the unworkability of the current trial court out-of-time-appeal procedure while overvaluing its own speculation about the disruption to the legal system that will ensue if motions for out-of-time appeals in trial courts are eliminated. The workability factor therefore weighs in favor of 76 overturning our trial court out-of-time appeal precedents, and for doing so now. (e) We Overrule our Precedents and Eliminate the Judicially Created Motion for Out-of-Time Appeal Procedure in Trial Courts. In light of the analysis conducted above, we conclude that stare decisis does not preclude overruling our precedents that created or endorsed the trial court out-of-time appeal procedural vehicle. Accordingly, we overrule Rowland v. State, 264 Ga. 872, 874-875 (452 SE2d 756) (1995), and any other decisions that approved the judicially created motion for out-of-time appeal in trial courts, to the extent that they endorsed this procedure. We also disapprove King v. State, 233 Ga. 630, 630 (212 SE2d 807) (1975), Furgerson v. State, 234 Ga. 594, 595 (216 SE2d 845) (1975), and any other decisions that have allowed out-of-time appeal claims to be litigated in trial courts without addressing the propriety of that procedure. We note that in overruling and disapproving these cases, we do not undo what has been done with respect to out-of-time appeals that already have been granted where the ensuing appeal has concluded. And that 77 important point leads us to an analysis of how today’s holding applies to this case and others going forward. 4. Our Holding Applies to this Case, Cases in the Appellate “Pipeline,” and Future Cases. Amicus curiae GACDL contends that if this Court concludes it must eliminate out-of-time appeal motions in trial courts, it should announce its intention to do so “well ahead of the change” and “set a date certain after which no motion for out-of-time appeal may be filed.” In short, GACDL asks us to apply any new rule we announce prospectively. 24 But we decline that proposal because prospective application would run afoul of the “pipeline” approach Georgia has long followed for the application of new rules of criminal procedure to criminal cases that are pending on direct review or not yet final. In Taylor v. State, 262 Ga. 584 (422 SE2d 430) (1992), this Court held that To support its request, GACDL cites only Davenport v. State, 309 Ga. 24 385, 399 (846 SE2d 83) (2020). But our decision in Davenport pertained to this Court’s decision to change a court practice—sua sponte review of evidentiary sufficiency in murder appeals—and did not endorse or otherwise authorize limiting this Court’s holdings in criminal cases to prospective application. See id. 78 [i]n order to ensure that similarly situated defendants are treated similarly and to maintain the integrity of the judicial process while still providing finality, . . . it [is] . . . appropriate to adopt the “pipeline” approach, that is, that a new [state] rule of criminal procedure . . . will be applied to all cases then on direct review or not yet final. Id. at 586.25 We have consistently recognized or followed this holding regardless of whether the judicial decision setting forth a new state rule of criminal procedure was based on a statute or on decisional law. See, e.g., Mobley v. State, 265 Ga. 292, 294 (455 SE2d 61) (1995); Smith v. State, 268 Ga. 196, 201 (486 SE2d 819) (1997); Smith v. State, 268 Ga. 860, 861 & n.10 (494 SE2d 322) (1998); Harris v. State, 273 Ga. 608, 610 (543 SE2d 716) (2001); Green v. State, 279 Ga. 455, 456 (614 SE2d 751) (2005); Stubbs, 308 Ga. at 25 We note that a different rule applies to cases in habeas corpus. See Harris, 273 Ga. at 610 (clarifying that a new state rule of criminal procedure “w[ould] not be applied to convictions challenged on habeas corpus”); Chatman v. Brown, 291 Ga. 785, 788 (733 SE2d 712) (2012); Turpin v. Todd, 268 Ga. 820, 830-831 (493 SE2d 900) (1997). We also note that the Georgia rule regarding retroactive application of new holdings in civil cases is less settled. Compare Findley v. Findley, 280 Ga. 454, 460 (629 SE2d 222) (2006) (“[W]e decline to adopt a rule of universal retroactivity in all civil cases.”) with Atlanta Oculoplastic Surgery, P.C. v. Nestlehutt, 286 Ga. 731, 739-744 (691 SE2d 218) (2010) (Nahmias, J., concurring specially, joined by Carley, P.J., and Hines, J.). 79 362 n.11.26 Neither Cook nor GACDL argues that we should overrule Taylor and its progeny, let alone engages in a stare decisis analysis of that precedent. Indeed, neither Cook nor GACDL argues that our holding in this case is somehow not the sort of judicial decision that is governed by Taylor or offers a compelling reason to reconsider Taylor or its progeny. See Collins v. State, 312 Ga. 727, 735 (864 SE2d 85) (2021) (appellant “does not offer any compelling reason to abandon precedent that has been a settled part of our law for 35 years, is straightforward in its application, and [is] not obviously unsound”). We therefore decline the invitation to apply today’s holding only prospectively. Instead, pending and future motions for 26 A recent case provides another good example: in Seals v. State, we held that a trial court’s dead-docketing a count of an indictment left that count “pending in the court below” for purposes of OCGA § 5-6-34 (a) (1), thus preventing an appeal of the case unless the defendant followed interlocutory appeal procedures. See Seals v. State, 311 Ga. 739, 739 (860 SE2d 419) (2021). That holding was applied to Seals’s case, and we immediately began applying it to dismiss criminal appeals that were filed in our Court when one or more of the counts in the case had been dead-docketed—even when the notice of appeal in the case was filed before Seals was decided. See, e.g., Favors v. State, S21A0328 (June 29, 2021); Scott v. State, S21A0507 (June 29, 2021); Drennon v. State, S21A1139 (Aug. 6, 2021). 80 out-of-time appeals in trial courts should be dismissed, and trial court orders that have decided such motions on the merits—like the one in this case—should be vacated if direct review of the case remains pending or if the case is otherwise not final. See Brooks v. State, 301 Ga. 748, 752 (804 SE2d 1) (2017) (“Because the trial court decided the merits of a motion it lacked jurisdiction to decide, we vacate the trial court’s order and remand with instructions to dismiss.”). 5. Conclusion We are faced with two choices in this case: We can overrule our trial court out-of-time appeal precedents, return to the habeas corpus process the General Assembly established for seeking post- conviction relief of this sort, and—to the extent that procedure is problematic—allow the General Assembly to fix any flaws by statute. Or we can retain our erroneous trial court out-of-time appeal precedents, maintain an alternative procedure for obtaining post-conviction relief for this one type of constitutional claim, and perpetuate our roles of judges-as-legislators who must continue to 81 establish by judicial opinion the rules for the procedural vehicle we created. In light of the analysis conducted above, we choose the former. We hold that there was and is no legal authority for motions for out-of-time appeal in trial courts and that the out-of-time appeal procedure allowed in King and Furgerson, approved in Rowland, and followed in other cases, is not a legally cognizable vehicle for a convicted defendant to seek relief from alleged constitutional violations. Our holding applies to this case and to all cases that are currently on direct review or otherwise not yet final. Accordingly, the trial court was without jurisdiction to decide Cook’s motion for out-of-time appeal in this case, and “because the trial court’s order plainly shows that it denied” the motion “on the merits,” Bonner v. State, 310 Ga. 426, 428 (851 SE2d 578) (2020), the trial court’s order must be vacated and the case remanded to the trial court with direction that the motion be dismissed. See McDaniel v. State, 311 Ga. 367, 373 (857 SE2d 479) (2021) (“[W]hen a trial court is presented with a motion that it lacks jurisdiction to 82 decide and denies the motion solely on the merits, we vacate the trial court’s order and remand with instructions to dismiss the motion.”); Brooks, 301 Ga. at 752. Cook’s remedy, if any, lies in habeas corpus. Judgment vacated and case remanded with direction. All the Justices concur, except for Boggs, P.J., who concurs specially in Division 3 (c) and (d), and Peterson, Bethel, and Ellington, JJ., who dissent. 83 S21A1270. COOK v. THE STATE. LAGRUA, Justice, concurring. I write separately to emphasize that this matter presents a challenging question of law with a fairly straightforward answer. The question is whether a criminal defendant, who has alleged that she has been deprived of an appeal of right due to ineffective assistance of counsel, may seek an out-of-time appeal as her remedy in the trial court instead of pursuing her remedy through habeas. The answer is no. The majority opinion and the dissent agree that this Court’s creation of the out-of-time appeal procedural vehicle in the trial court was unsupported by sound reasoning or relevant authority. 27 Where we part ways is on the question of whether stare decisis 27In addition, as noted by the majority opinion: An unusual aspect of this case is that neither the State nor any of the amici curiae—which represent the views of both prosecutors (PAC) and criminal defense lawyers (GACDL)—contends that our judicial decisions creating or endorsing the trial court out-of-time appeal procedural vehicle are supported by sound reasoning or relevant authority. 84 favors the retention of the out-of-time appeal procedure in the trial court, and I concur with the majority opinion that it does not. Under the doctrine of stare decisis, “courts generally stand by their prior decisions, because it promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process.” Pounds v. State, 309 Ga. 376, 382 (3) (846 SE2d 48) (2020) (citation and punctuation omitted). “Nevertheless, when governing decisions are unworkable or are badly reasoned, this Court has never felt constrained to follow precedent. Stare decisis is not an inexorable command; rather, it is a principle of policy and not a mechanical formula of adherence to the latest decision.” Woodard v. State, 296 Ga. 803, 812 (3) (b) (771 SE2d 362) (2015) (citation and punctuation omitted). As explained in the majority opinion, the stare decisis factors include: (1) the soundness of our precedents, which we all agree are entirely unsound; (2) the age of the precedents, which are not especially old; (3) the reliance interests involved, including that our precedents 85 have established a procedural right rather than a substantive one; and (4) the workability of the procedure, including that the parameters of the procedure need continual judicial revision. I agree with the majority opinion that these factors do not weigh in favor of keeping a procedure that we “created out of whole cloth” and which helped lead to “a tangled mess of post-conviction jurisprudence.” Collier v. State, 307 Ga. 363, 379 (834 SE2d 769) (2019) (Peterson, J., concurring specially). As this Court initially held correctly in Neal v. State, 232 Ga. 96 (205 SE2d 284) (1974), in the Habeas Corpus Act of 1967, the General Assembly established habeas corpus as the procedure for a criminal defendant who alleges that she was deprived of her right to appeal because of her counsel’s alleged ineffective assistance. And “it is not the job of judges to usurp that [legislative] power . . . by rewriting laws enacted by the people’s democratically elected representatives.” Barrow v. Raffensperger, 308 Ga. 660, 692 (842 SE2d 884) (2020) (Melton, C. J., concurring). Accordingly, I agree with the majority opinion that Cook’s remedy, if any, lies in habeas 86 corpus. If the General Assembly takes issue with the exclusiveness of the procedure it has created, it is incumbent upon the legislature to fashion a new procedure. I am authorized to state that Justice McMillian joins in this concurrence. 87 BOGGS, Presiding Justice, concurring specially in part. I agree with most of what is said in the majority opinion and concur fully in the judgment and in Divisions 1, 2, 3 (a), (b), and (e), 4, and 5. I do not necessarily agree with all that is said in Division 3 (c) and 3 (d), however, so I concur specially in those parts of the opinion. 88 S21A1270. COOK v. THE STATE. PETERSON, Justice, dissenting. The majority today overrules decades of nonconstitutional precedent recognizing a trial court power that is entrenched within our system, it does so without a clear sense of the likely consequences, and it does so while overreading some of our precedents in a way that would prohibit the General Assembly from fixing at least some adverse consequences. If stare decisis means anything, it should preserve longstanding and oft-applied nonconstitutional precedent at least until we know the effect of overruling. I respectfully dissent from the majority’s refusal to wait for that day. As an initial matter, I should note that I agree with much of the majority’s opinion. I agree that our approval of the motion for out-of-time appeal happened without analysis or the articulation of a proper legal basis. I don’t dispute that we held in Neal v. State, 232 Ga. 96 (205 SE2d 284) (1974), that the habeas statute provided the exclusive remedy for claims that a right to appeal was frustrated 89 by ineffective assistance of counsel.28 I agree that the motion for out- of-time appeal has contributed to the complexities of our tangled mess of a post-conviction litigation process. And I agree that we did something we should not have nearly 50 years ago when we ignored our decision in Neal in acknowledging a stand-alone motion for out- of-time appeal. Indeed, I’ve already said as much. See Schoicket v. State, 312 Ga. 825, 825 (865 SE2d 170) (2021); Collier v. State, 307 Ga. 363, 379-82 (834 SE2d 769) (2019) (Peterson, J., concurring specially). But concluding that our decades-old precedent was misguided is not sufficient to reject it. Given the importance of stare decisis, I 28 Whether Neal was rightly decided is a closer question. Neal did no statutory construction in support of its holding that the habeas corpus statute precludes a motion for out-of-time appeal in the trial court, and the habeas statute itself contains no text supporting such a proposition. The closest the statute gets is a provision that states that “this article provides the exclusive procedure for seeking a writ of habeas corpus for persons whose liberty is being restrained by virtue of a sentence imposed against them by a state court of record,” OCGA § 9-14-41 (emphasis added). Neal did not identify the similar text in effect at the time as supporting its conclusion, and a motion for out-of- time appeal does not seek a writ of habeas corpus. And the only other cases the majority cites for this proposition, Smith and Mitchum, hold only that habeas is the exclusive remedy for post-appeal ineffectiveness claims. Of course, the whole point of an out-of-time appeal motion is that the defendant has not yet had their appeal. 90 cannot join the majority in ripping the motion for out-of-time appeal out of our post-conviction system’s “tangled mess.” Collier, 307 Ga. at 379) (Peterson, J., concurring specially). In a footnote in Schoicket, I observed that such a step might be appropriate. See 312 Ga. at 830 (1) n.6. But the extensive briefing and argument in this case has convinced me otherwise. The motion for out-of-time appeal is more entrenched in our legal system than I had realized, and the policy implications of reversing course now are beyond our Court’s ability even to understand fully today, much less solve. The majority dismisses my discussion of those implications as speculative. I agree; neither I nor the majority has a full sense of the effect of reversing course, and so speculation is the best we can do. But respectfully, the majority’s response does not help its case; we shouldn’t overrule decades of precedent without more than speculation about the effects of such a move. In my view, stare decisis exists for cases like this one, and I would retain our incorrect precedent as the lesser of two evils. 1. Stare decisis is a question of judicial policy, not an 91 objective formula. Whether stare decisis should preserve a legally incorrect precedent is a question of policy, not of law. See State v. Jackson, 287 Ga. 646, 658 (5) (697 SE2d 757) (2010) (“[S]tare decisis is not an inexorable command, nor a mechanical formula of adherence to the latest decision. Stare decisis is instead a principle of policy.” (citation and punctuation omitted)). That policy question weighs the value of having an issue decided against the value of deciding it right. Almost always, questions of policy are for the other two branches of government. But stare decisis is that rare kind of policy that — since the very beginning of our Court — we have consistently understood to be within the judicial power to apply. See, e.g., Leary v. Durham, 4 Ga. 593, 601 (1848) (observing in case involving property rights that “where a rule of law has been firmly established for half a century[] at least, though originally, perhaps, on mistaken or erroneous principles, and no greater evil is to be apprehended from an adherence to it, than may be expected from a departure from it, that stare decisis ought to be our motto”). 92 Questions of policy often leave more room for disagreement than questions of law. There are only so many possible meanings legal text can have, and one meaning is almost always objectively more correct than the other possibilities. But there are many different ways to approach the kinds of policy questions that stare decisis presents. And so it is here. That the members of this Court disagree over whether to retain our prior precedent creating motions for out-of-time appeals does not mean that some of us are applying the law unfaithfully. It simply means that, in our reasoned exercise of prudential judgment, we arrive at different conclusions. Understanding stare decisis as a matter of judicial policy, our primary precedent setting the framework for deciding questions of stare decisis does not limit us to an exhaustive list of factors to consider. Rather, we have framed this as a balancing of considerations in which we consider factors “such as” — not limited to — “the age of the precedent, the reliance interests at stake, the workability of the decision, and, most importantly, the soundness of its reasoning.” Jackson, 287 Ga. at 658 (5). The United States 93 Supreme Court precedent on which our framework is based is consistent with that understanding. See Montejo v. Louisiana, 556 U.S. 778, 792-97 (129 SCt 2079, 173 LE2d 955) (2009) (noting relevant factors “include” age, reliance interests, soundness of precedent, and workability before weighing the “marginal benefits” of the prior rule “against its substantial costs” to conclude that it “does not pay its way”) (citation and punctuation omitted)). Apart from the four factors identified in Jackson, we have identified other considerations that are quite germane to the question before us in this case. We have made clear that stare decisis applies with greater force to statutory precedents than to constitutional precedents, as it is more difficult for the legislature to undo a constitutional decision. Compare, e.g., Abernathy v. City of Albany, 269 Ga. 88, 90 (495 SE2d 13) (1998) (“Even those who regard ‘stare decisis’ with something less than enthusiasm recognize that the principle has even greater weight where the precedent relates to interpretation of a statute.” (citation and punctuation omitted)), with Olevik v. State, 302 Ga. 228, 245 (2) (c) (iv) (806 SE2d 505) 94 (2017) (“[S]tare decisis carries less weight when our prior precedent involved the interpretation of the Constitution, which is more difficult than statutory interpretation for the legislative process to correct. This doesn’t mean that we disregard stare decisis altogether, though; what it actually means is that the first stare decisis factor (soundness of reasoning) becomes even more critical. The more wrong a prior precedent got the Constitution, the less room there is for the other factors to preserve it.” (citation omitted)). Another important consideration in determining whether to retain prior precedent is the extent to which it has become entrenched in the legal system — meaning, in a narrow sense, that its relevant holding has been applied frequently. See, e.g., Frett v. State Farm Employee Workers’ Compensation, 309 Ga. 44, 60 (3) (c) (844 SE2d 749) (2020) (considering the extent to which a precedent is “entrenched” in our jurisprudence by examining how often and how recently it has been cited, particularly for its relevant holdings). Although this concept of entrenchment can include the age of the precedent, the extent to which others have relied on it, and its 95 workability, it is not limited to those categories. Understanding entrenchment more broadly, it is important that we consider potential disruption to the legal system that might be caused by suddenly jettisoning a particular precedent. The majority criticizes me for focusing too much on this idea. But there’s little else warranting much attention. We don’t particularly disagree on the correctness of the creation of the motion for out-of-time appeal. And the other Jackson factors either aren’t terribly meaningful (age), 29 or are encompassed within my treatment of entrenchment (reliance and workability). 2. In my view, stare decisis counsels that we retain our 29 Indeed, once we’ve determined that a decision was unsound, the other Jackson factors never seem to be particularly meaningful. Since 2010, our Court has tended to recite the Jackson “four-factor test” consistently, and just as consistently overrule precedent after woodenly ticking through those factors. See, e.g., Frett v. State Farm Employee Workers’ Comp., 309 Ga. 44, 63- 64 (844 SE2d 749) (2020) (Peterson, J., dissenting) (noting that every time an opinion explicitly applied the Jackson test even as to statutory precedent, we overruled the precedent). The majority responds by reiterating the Frett majority’s response: it’s “unsurprising” that the Court does extensive analysis of stare decisis “mostly” in cases where the Court overrules precedent. The word “mostly” is critical to that argument, and it is wrong. The Court doesn’t do stare decisis analysis “mostly” in cases overruling precedent, it does stare decisis analysis exclusively in cases overruling precedent. The majority does not identify a single case in which we have concluded a precedent of ours was unsound, but nevertheless determined that stare decisis warranted retaining that precedent. 96 current rule allowing trial courts to consider motions for out-of-time appeal. Here, we are dealing with the sort of nonconstitutional precedent to which stare decisis applies more strongly. The precedent at issue here is of the sort that the General Assembly might easily alter or eliminate, and thus eliminating that precedent ourselves should give us greater pause. Although the majority opinion posits that stare decisis applies with less force here because Rowland v. State, 264 Ga. 872 (452 SE2d 756) (1995), and its pre- Collier progeny ignored the relevant statutory text, I’ve already explained that Neal did not engage with the text of the habeas statute, either. As noted above, I don’t dispute Neal’s holding. But the majority opinion’s focus on the soundness of the reasoning (or lack thereof) in the decisions that subsequently deviated from that holding, while important, is only the beginning of the application of stare decisis. “[S]tare decisis does not even begin to apply until we doubt the correctness of a previous precedent.” Frett, 309 Ga. at 65 (Peterson, J., dissenting). If we believe a precedent to be correct, we simply 97 apply it and stop there. But if we doubt the correctness of the precedent, we do not simply overrule that precedent without consideration of other factors. “[I]f stare decisis is to mean something, we need more than that to overrule a statutory precedent.” Id. (a) The motion for out-of-time appeal plays a significant role in our legal system. Turning to those other considerations, they include the extent to which the precedent has become entrenched in our legal system. And here, nearly 50 years after it first appeared in our case law, the motion for out-of-time appeal has become deeply entrenched, at least in the senses that (1) it is regularly applied in a significant number of cases and (2) when it is applied, it often makes a substantial difference — when the motion is granted, it permits an appeal that would otherwise be barred without the years-long delay of habeas. Thus, although I suggested otherwise in Schoicket, the briefing and argument in this case has convinced me that jettisoning our precedent here could be enormously disruptive. Undoubtedly, our precedent allowing out-of-time appeals 98 affects a significant number of cases. This Court entertained appeals of murder convictions enabled by granted motions for out-of-time appeals at least 14 times last year alone.30 A similar number for the Court of Appeals is not readily ascertainable; that court does not systematically report the procedural history of each criminal case it decides in the same way that we do in our murder cases (or, for that matter, publish all of its opinions as we do). But if a similar proportion of their criminal appeals arise from granted motions for out-of-time appeal (as seems reasonable to suppose), the true annual 30 It is our practice in direct murder appeals to include the case’s procedural history in our opinion’s first footnote. At least 14 such footnotes in 2021 indicated a granted motion for out-of-time appeal. See Williams v. State, 312 Ga. 386, 386 n.1 (863 SE2d 44) (2021); Baker v. State, 312 Ga. 363, 363 n.1 (863 SE2d 55) (2021); Walker v. State, 312 Ga. 332, 332 n.1 (862 SE2d 542) (2021); Thompson v. State, 312 Ga. 254, 254 n.1 (862 SE2d 317) (2021); Walker v. State, 312 Ga. 232, 232 n.1 (862 SE2d 285) (2021); Williams v. State, 312 Ga. 195, 195 n.1 (862 SE2d 108) (2021); Sullivan v. State, 311 Ga. 835, 835 n.1 (860 SE2d 576) (2021); Holmes v. State, 311 Ga. 698, 698 n.1 (859 SE2d 475) (2021); Rogers v. State, 311 Ga. 634, 634 n.1 (859 SE2d 92) (2021); Felts v. State, 311 Ga. 547, 547 n.1 (858 SE2d 708) (2021); Waller v. State, 311 Ga. 517, 518 n.1 (858 SE2d 683) (2021); Abbott v. State, 311 Ga. 478, 478 n.1 (858 SE2d 696) (2021); Thomas v. State, 311 Ga. 280, 280 n.1 (857 SE2d 223) (2021); Kirkland v. State, 310 Ga. 738, 738 n.1 (854 SE2d 508) (2021). We do not generally include such a footnote in other criminal cases, and so this list of 14 may be underinclusive for 2021. 99 number of such appeals may well be between 50 and 100 per year, 31 and even more in years of heavier caseloads. 32 And even denied motions for out-of-time appeal (which may not always themselves be appealed) would first have resulted in full habeas proceedings under the Court’s holding today. The majority opinion first responds by dismissing the significance of these cases on the basis that they are not “precedent” for stare decisis purposes because they did not decide whether the existence of a motion for out-of-time appeal was appropriate. But that misunderstands the nature of precedent for these purposes. At least by the time of our 1995 decision in Rowland, our precedent definitively established that motions for out-of-time appeal are 31 During 2021, we disposed of 304 direct appeals. All of our murder appeals would have been contained within that number and typically make up a substantial majority of our direct appeals; this suggests that we perhaps disposed of 200 to 250 murder appeals during 2021. During the same period, the Court of Appeals disposed of 804 criminal appeals. 32 And this number does not even include the dozens of unpublished orders we have issued over the years dismissing appeals for untimeliness and citing Rowland for the proposition that the appellant may be able to file a motion for out-of-time appeal if the untimeliness was a result of counsel’s ineffective assistance. 100 proper. The 14 cases from last year in which our opinions noted that motions had been granted thus represent faithful application of that precedent and thus are precedent themselves for the purpose of considering the extent of the entrenchment of the precedent in question today.33 (b) Shifting all out-of-time appeals to habeas will have significant negative effects. The majority’s argument that the long delays of habeas are not meaningful because some motions for out-of-time appeal are filed years or decades later is unavailing. Long-belated motions are a problem (although, as the majority would presumably agree, not one beyond the power of the General Assembly to end). But they seem relatively rare; the majority cites only a handful. Most of the motions that the majority consigns to the often years-long delays of habeas are far more frequently filed shortly after deadlines are missed. In the vast majority of the 14 cases we considered last year from the 33The application of a case’s well-settled holding does not require citing that case or discussing its analysis. At the very least, every case post-Rowland that has granted or denied on the merits a motion for out-of-time appeal has applied Rowland’s holding, whether or not it cites it. 101 grant of a motion for out-of-time appeal, the motion was filed less than one year after the notice of appeal was due — indeed, often considerably less than that, indicating that the issue was one of attorney oversight rather than an attempt to attack an old conviction. In one case, the notice of appeal had been filed a mere six days late. See Abbott v. State, 311 Ga. 478, 478 n.1 (858 SE2d 696) (2021). In another, the issue was the trial court’s failure to formally vacate an initial order denying a motion for new trial in ruling on a second amended motion. See Walker v. State, 312 Ga. 332, 332 n.1 (862 SE2d 542) (2021). Particularly in those sorts of cases, where the only issue is whether counsel failed to file timely, allowing post-conviction proceedings to be resolved by granting a motion for out-of-time appeal in the trial court is much more efficient than requiring a convicted defendant to turn to the much lengthier habeas process. Granting a motion for out-of-time appeal allows the claims to be resolved promptly by the judge who presided over the trial. It avoids the need for an inmate to grapple with the procedural hurdles of 102 filing a habeas petition, avoids the need to transfer records between jurisdictions, and reduces travel costs for lawyers and prisoners. And the speed of this process is particularly important when this Court — as a matter of policy — has sought to prevent delays in the resolution of post-conviction proceedings. See Owens v. State, 303 Ga. 254, 258-60 (4) (811 SE2d 420) (2018) (directing the Council of Superior Court Judges to propose a rule to reduce post-conviction delays). Although the majority says that a convicted defendant can obtain the same result by filing in habeas, a grant of relief after years of additional litigation is not obviously the same as a grant of relief that is secured promptly. The majority responds that the strictures of habeas “reflect policy choices the General Assembly made.” But some of the most important strictures were not in place when we created the out-of- time appeal vehicle. It was not until 2004 that the General Assembly imposed a limitations period for seeking habeas corpus relief from 103 felony convictions. See Ga. L. 2004, p. 917-918, § 1. 34 And it was only that same year that the General Assembly provided that laches may be a basis for dismissal. See id., § 3. We presume that the General Assembly was aware of the availability of motions for out-of-time appeal and chose not to apply limitations periods or laches to them when it enacted these modifications to habeas. 35 See Grange Mutual Casualty Co. v. Woodard, 300 Ga. 848, 852 (2) (a) (797 SE2d 848) (2017) (“[A]ll statutes are presumed to be enacted by the legislature with full knowledge of the existing condition of the law and with reference to it.” (citation and punctuation omitted)). 34 I thank the Georgia Association of Criminal Defense Lawyers for its excellent amicus brief bringing this point to my attention. That brief also suggests that a 1986 amendment to a Title 40 statute applied limitations to seeking habeas relief from certain traffic convictions. See Ga. L. 1986 p. 444, § 1. The actual effect of the 1986 provision is not entirely clear, but even if it did create a limitation on habeas petitions, it did so only for certain traffic offenses. 35 The majority interprets this point as though I’m arguing that Collier was wrongly decided in authorizing a possible defense of prejudicial delay. This dissent expresses no opinion on that point. I merely point out that the General Assembly apparently did not share the majority’s policy concerns when that body enacted a limitations period and laches for habeas and chose not to extend those provisions to motions for out-of-time appeal, which had existed for decades before the General Assembly’s 2004 legislation. Nothing in Collier held otherwise. 104 Another significant effect on the criminal justice system is the workload impact shifting all these cases to habeas would have, both in terms of shifting court and counsel, and in terms of likely resulting in fewer state concessions. The current process spreads the obligation for handling such matters across all of the state’s district attorneys and judges, rather than concentrating the burden on a few judges in counties with prisons and shifting all prosecutorial responsibilities to the Office of the Attorney General. 36 During 2021, we had 174 certificates of probable cause filed in our Court, seeking to initiate an appeal from the denial of a habeas petition, which suggests that the total number of habeas petitions resolved by superior courts during that time frame is unlikely to have been much more than 200. If we were to add 50 to 100 additional habeas 36 Although this could eventually be rectified in part by the General Assembly’s reallocation of appropriations, this is easier said than done. For one thing, it’s doubtful that any particular district attorney would be relieved of enough such motions to warrant reducing staff. At the same time, concentrating all of the cases in the AG’s office as habeas petitions may well result in a substantial increase in hourly fees paid to special assistant attorneys general contracted to defend convictions against habeas petitions. This is particularly likely because habeas petitions generally require more litigation than a motion, such that the change worked by the Court today likely will result in more work, not just shifting the work to different personnel. 105 petitions per year — as the majority today may well do, although we can’t be sure — that could represent an increase in habeas caseloads of 25 to 50 percent, while concentrating those cases among the relatively few superior court judges who sit in circuits with prisons. Although the majority argues that we should “credit” the assessment of the Attorney General and the District Attorney who have asked this Court to overrule its out-of-time appeal precedents, neither of those constitutional officers has addressed the question of resources in their briefs in this case. And the largest resource impact would be on a small group of trial court judges, who are scarcely able to weigh in. Moreover, our precedent reveals anecdotally that the State often agrees that the trial court should grant a motion for out-of- time appeal, limiting the amount of litigation required to resolve such an issue. But it’s one thing for the trial prosecutor — familiar with the case and defense counsel — to agree that ineffective assistance likely occurred and should be remedied. It’s another thing altogether for the State’s habeas counsel — likely unfamiliar with 106 the case, defense counsel, and the trial court — to do so; more meritorious motions would likely have to be litigated fully before relief could be granted. And finally, our prior approach often had the effect of affording indigent defendants access to counsel for their motion for out-of-time appeal,37 something that will be — at best — less certain if we suddenly require inmates to turn to habeas. See Gibson v. Turpin, 270 Ga. 855, 860-61 (1) (513 SE2d 186) (1999).38 Moreover, they won’t be able to bring only their claim regarding a frustrated appeal without seriously risking waiving all other habeas-eligible claims. 37 The majority responds that indigent defendants do not have a right to counsel for a motion for an out-of-time appeal. But although I agree that indigent defendants do not have a right of appointed counsel for such a motion, as a practical matter they often in fact have such counsel when a motion is filed to correct a missed deadline. The majority’s observation that a public defender can decline to represent a defendant who is not entitled to representation is not responsive to this point. 38 I note that today’s step of entirely eliminating motions for out-of-time appeal as a procedure to remedy the frustration of a defendant’s right to appeal makes all the more relevant and timely Justice Ellington’s previous suggestion that the superior court rules be amended to ensure that plea colloquies more thoroughly inform defendants of their rights to withdraw their plea and appeal their conviction. See Schoicket, 312 Ga. at 839-40 (Ellington, J., dissenting in part); see also id. at 833 (1) n.10 (majority opinion noting that Justice Ellington’s “thoughtful suggestion” merited consideration). 107 See, e.g., Williamson v. State, 305 Ga. 889, 897 (4) (827 SE2d 857) (2019) (“Any claim of ineffectiveness of counsel must be made at the earliest practicable moment.”).39 And now they will likely need to do so pro se, without having had a hearing on a motion for new trial and without a right to access the record of their trial without first explaining why they need it. See Rutledge v. State, 309 Ga. 508 (847 SE2d 143) (2020) (explaining that “after the time for appeal has expired there is no due process or equal protection right to a free copy of one’s court records absent a showing of necessity or justification” (citation and punctuation omitted)). The Court’s decision today puts a defendant whose attorney missed the deadline for filing a notice of appeal in a very difficult position. (c) Neither lawyer ethical rules nor our decisions interpreting them are cause to abandon our precedent allowing motions for out-of-time appeals. 39 On the other hand, the rule that ineffectiveness claims must be raised at the earliest practicable moment appears to be another rule that we have created without a basis in any statutory authority. See Smith v. State, 255 Ga. 654, 655 (3) (341 SE2d 5) (1986) (“It is a requisite of a sound system of criminal justice, serving alike the proper ends of defendants and of the public, that any contention concerning the violation of the constitutional right of counsel should be made at the earliest practicable moment.”). Under the majority’s approach to stare decisis, the fate of that rule is uncertain as well. 108 The majority responds to much of this by arguing that our precedent appears to prohibit lawyers from asserting their own ineffectiveness. If that were true, it would give me pause. But the majority overreads our precedent. Our precedent makes clear that trial counsel cannot litigate contested ineffectiveness claims on a motion for new trial, on appeal, or on habeas. But it does not address the question of whether trial counsel may assert an uncontested ineffectiveness claim in a motion for out-of-time appeal in the trial court, which textual changes to the Georgia Rules of Professional Conduct (the “Rules”) following our only decision analyzing any Rule on this point strongly suggest is permissible. The majority’s broader reading is wrong and, if adopted as a holding in a future case, would tie the hands of the General Assembly in any later legislative effort to fix what the majority breaks today. In Castell v. Kemp, 254 Ga. 556 (331 SE2d 528) (1985), we denied an application for interlocutory appeal from a trial court order disqualifying a lawyer. See id. at 558. The only analysis in the opinion was the trial court order, which we quoted in full without 109 comment beyond denying the application. The trial court disqualified Bruce Harvey, who had served as trial counsel for the defendant, from representing the defendant in his habeas petition. The trial court cited one non-binding rule and two non-binding comments on the rules that applied at the time, each of which related to pitfalls of lawyers serving as witnesses.40 Different text of the old rule — again, in a non-binding comment — indicated that testifying on an uncontested issue would not be a problem: “It is not objectional for a lawyer who is a potential witness to be an advocate if it is unlikely that he will be called as a witness because his testimony would be merely cumulative or if his testimony will relate only to an uncontested issue.” Georgia Code of Professional Responsibility, EC 5-10 (1985). The trial court did not note this provision, no doubt because it was not at issue in the contested habeas case before it. The trial court reasoned that the habeas petition sought to 40 The trial court also cited an Eleventh Circuit case for the proposition that “there may arise issues of potentially differing interests of the lawyer and his client,” but made no conclusions about any conflict-of-interest rules. 110 raise claims of ineffectiveness regarding Mr. Harvey, and that “virtually all the evidence of ineffective assistance of counsel is within counsel Harvey’s personal knowledge,” and “some evidence can only come from Mr. Harvey.” Castell, 254 Ga. at 558. The trial court noted its concern that the challenges to Mr. Harvey’s credibility his testimony would inevitably generate, and the necessity of cross-examination, posed threats to the integrity and reliability of the judicial process. See id. at 557-58. Sixteen years later, the State Bar proposed, and we approved, a wholesale set of new rules modeled after the ABA model rules. The old rule language quoted by the trial court largely no longer exists in our current rules (which were again significantly revised in 2018). The current binding rules on lawyers serving as witnesses generally prohibit it, but expressly allow it where “the testimony relates to an uncontested issue.” Ga. R. Prof. Conduct 3.7 (a) (1); see also Cmt. 3 to Rule 3.7 (“Paragraph (a) (1) recognizes that if the testimony will be uncontested, the ambiguities in the dual role are purely theoretical.”). 111 It’s well-settled that “when statutory amendments [] materially alter text that this Court has previously interpreted, our pre-amendment precedent no longer binds lower courts to the extent the amendments change the meaning of the text.” State v. Stanford, 312 Ga. 707, 710 n.3 (864 SE2d 448) (2021). And the Georgia Rules of Professional Conduct are, like all legal text, subject to this principle. To the extent that our quotation of the trial court’s order in Castell adopted the trial court’s reasoning as a holding, the material alteration of the rules made Castell obsolete, at least as to testimony that was uncontested. Deleting a non-binding comment, and replacing it with a binding rule — even if similar — is a substantive change to the meaning of the rules. Over two decades after Castell and six years after the adoption of the new rules, we cited Castell as the sole authority for the proposition that “[b]ecause a lawyer may not ethically present a claim that he/she provided a client with ineffective assistance of counsel, a claim of ineffective assistance of trial counsel cannot be pursued unless trial counsel is no longer representing the convicted 112 defendant.” Hood v. State, 282 Ga. 462, 463 (651 SE2d 88) (2007). We did so only for the proposition that the defendant had not waived ineffectiveness claims by failing to raise them before a belated pro se motion for new trial. See id. We then cited Hood in Garland v. State, 283 Ga. 201, 202 (657 SE2d 842) (2008), in support of our conclusion that a defendant was entitled to new counsel on appeal to pursue ineffectiveness claims. Hood and Garland are the only decisions of ours the majority cites 41 for the proposition that trial counsel are ethically prohibited from raising a claim of ineffective assistance of counsel in a motion for out-of-time appeal, even when uncontested. They do not stand for that proposition. The only analysis on this point that any of these cases included was the quotation of the trial court’s order in Castell, which applied now-superseded rules regarding lawyers serving as witnesses.42 Now, the current rules generally prohibit lawyers from 41 The majority also cites a decision of our Court of Appeals. See Delevan v. State, 345 Ga. App. 46, 49-51 (811 SE2d 71) (2018). That case relies on the same authority already discussed and did not involve uncontested testimony. 42 We also have stated more broadly that an attorney cannot “reasonably 113 serving as witnesses on contested points, precisely the context in which Castell and its progeny were decided. I have no quarrel with that holding. But there is no basis to extend their holdings to cases of uncontested testimony. The majority’s only response to all of this is not to cite Castell. In short, neither the Rules nor our precedent interpreting them prohibit trial counsel from seeking an out-of-time appeal when the facts underlying the claim to which the lawyer would have to testify are uncontested (and which, if the trial lawyer is still counsel, naturally will be sought soon after the deadline has passed). Allowing this kind of motion is the key value of the system that presently exists relative to the system that will exist following the majority’s opinion. 3. The majority’s workability concerns do not warrant overruling our precedent. be expected to assert or argue his own ineffectiveness on appeal[,]” particularly in the contexts of determining whether a defendant needs new, conflict-free counsel, see Garland, 283 Ga. at 203, and determining whether a claim of ineffective assistance has been waived, see White v. Kelso, 261 Ga. 32, 32 (401 SE2d 733) (1991). But that is not the same as saying that a lawyer is categorically prohibited from asserting his own ineffectiveness in all circumstances. 114 Now, it’s important to note once again that considerations like these are not proper reasons to invent new avenues for relief that contravene statutes or otherwise misapply statutory law and equivalent legal principles in the first instance. But once we have done so — for decades, and in ways that are now entrenched in our legal system — these issues are appropriate to consider in determining whether to change course. The majority’s response to these concerns is to observe that the General Assembly could rectify any of these possible consequences of tossing aside the current system. And the majority takes the position that because we created the current situation, it is our responsibility to fix it. But this is always true when we’ve gotten a case wrong. The majority’s approach places a perpetual thumb on the scale in favor of overturning precedent, which is backwards. The idea that it doesn’t matter how much we break the legal system because the General Assembly can glue the pieces back together is simply not an argument compatible with stare decisis. Moreover, all of the problems with the existing system also can be remedied by 115 actions of the General Assembly. Stare decisis means that when we have competing workability considerations on both sides of a choice whether to retain precedent, all of which can be addressed by the legislature, we put the thumb on the scale of retaining our current precedent. To be sure, our invention of the motion for out-of-time appeal has led to problems, only some of which are identified at length in my concurrence in Collier and our decision in Schoicket (and, of course, the majority’s decision today). But we have only a binary choice of retaining or overruling precedent; stare decisis allows us to keep a wrong precedent, not to improve on it. And this binary choice is a blunt instrument for solving the problems we have recognized. The General Assembly, on the other hand, has far more options. For instance, it might limit the filing of such motions to a six-month period after the deadline for the filing of a notice of appeal. Or it might tie the ability to file such a motion to the appointment of new counsel for appeal, requiring counsel to file the motion within, say, 60 days of appointment. 116 The majority also contends that our current precedents are unworkable because they present an ongoing requirement that this Court “fill in the details of the trial court out-of-time appeal procedure we created,” requiring us to exercise legislative power. I disagree. There is no equivalent to the civil practice act for criminal cases. Many procedural vehicles are creatures of decisional law that require, from time to time, our definition and refining. But we do not do so unaided; there already are many background rules — common law, statutory, and decisional — that apply to criminal proceedings generally, and may well apply to motions for out-of-time appeal. To the extent that we occasionally may be called upon to answer questions about the application of such background rules, that is what appellate courts do. Indeed, we did just that — unanimously — less than five months ago regarding a judicially created procedural mechanism in criminal cases. See Walker v. State, 312 Ga. 640, 644 (2) (864 SE2d 398) (2021) (leaving undisturbed trial courts’ authority to dismiss criminal charges for want of prosecution 117 and defining some of the contours of that authority). 43 The majority focuses too narrowly on the effect this precedent has on us at the expense of a proper concern for the effect it has more broadly. And to the extent that parties ask us to expand those procedures, such tinkering is not ours to do. We made clear in Schoicket our intention to cease inventing new procedural mechanisms in this area. See 312 Ga. at 832 (1). But for the reasons outlined above, neither should we jettison the mechanisms that have existed ever since we erroneously invented them decades ago. We should leave the next move — if any — for the General Assembly. I am authorized to state that Justice Bethel and Justice Ellington join in this dissent. 43 In Walker, we observed that the State and its amici had pointed us to “no statute that purports to eliminate or limit this longstanding practice[.]” Id. at 644 (2). But that merely goes to the soundness of the original creation of the practice, not to the majority’s argument that being called upon to define the contours of a judicially created procedure is unworkable. 118