Banks v. Commissioner of Correction

339 Conn. 1

• Court: Supreme Court of Connecticut
• Decided: May 12, 2021
• Docket: SC20222
• Status: Published

Opinion

October 26, 2021 CONNECTICUT LAW JOURNAL Page 3

CASES ARGUED AND DETERMINED

IN THE

SUPREME COURT OF THE

STATE OF CONNECTICUT

MARK BANKS v. COMMISSIONER OF CORRECTION (SC 20222) Robinson, C. J., and Palmer, McDonald, D’Auria, Mullins, Kahn and Ecker, Js.*

Syllabus

In accordance with this court’s decision in State v. Salamon (287 Conn. 509), when a criminal defendant is charged with kidnapping in conjunc- tion with another underlying crime, such as robbery, the jury must be instructed that the defendant cannot be convicted of kidnapping if the restraint imposed on the victim was merely incidental to the commission of that underlying crime. The petitioner, who had been convicted of multiple counts of kidnapping in the first degree and robbery in the first degree in connection with armed robberies at two separate retail stores, sought a writ of habeas corpus, claiming that his due process right to a fair trial under the fifth and fourteenth amendments to the United States constitution was violated. In each armed robbery, after the petitioner obtained money from his victims, he forced them at gunpoint into the store’s bathroom and attempted to jam the bathroom door shut. The victims remained inside of the bathroom for only a few minutes, exiting once they believed that the petitioner left the store. Following this court’s determination that Salamon, which had been decided more than ten years after the petitioner’s trial, applied retroactively in habeas actions, the petitioner challenged his kidnapping convictions on the ground that the jury instructions at his criminal trial were not in accordance with the require-

* The listing of justices reflects their seniority status on this court as of the date of oral argument.

1 Page 4 CONNECTICUT LAW JOURNAL October 26, 2021

2 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction ments set forth in Salamon. The habeas court denied the petitioner’s habeas petition, concluding that the respondent, the Commissioner of Correction, had demonstrated that the absence of a Salamon instruction was harmless error. On the granting of certification, the petitioner appealed to the Appellate Court, which reversed the habeas court’s judgment. The Appellate Court concluded that a jury reasonably could have found that the petitioner’s movement and restraint of the victims were part of a continuous, uninterrupted course of conduct related to the robberies. The Appellate Court applied the harmless error standard set forth in Neder v. United States (527 U.S. 1) in determining that the absence of a Salamon instruction was not harmless beyond a reasonable doubt. On the granting of certification, the respondent appealed to this court. Held: 1. The standard articulated in Brecht v. Abrahamson (507 U.S. 619), which requires a new trial only if the instructional error had a substantial and injurious effect or influence in determining the jury’s verdict, applies to Salamon claims raised in habeas proceedings: contrary to the petition- er’s assertion that stare decisis required the application of the Neder standard, this court had not previously resolved the question of which standard applied to Salamon errors on collateral review; moreover, the Brecht standard provided the proper harmless error standard for Salamon errors in habeas actions, as it was consistent with the handling of other claims of error in habeas proceedings by both this court and the federal courts, a number of sister state courts had adopted that standard for the collateral review of constitutional errors, and it afforded a habeas petitioner significant protection, requiring a new trial unless the reviewing court has confidence that a properly instructed jury would have found the petitioner guilty beyond a reasonable doubt; furthermore, two of the principal rationales for applying a different harm standard (the Brecht standard) to constitutional errors in habeas actions than the harm standard applied to constitutional errors raised on direct appeal, namely, the finality of judgments and the extraordinary nature of the habeas remedy, applied equally to state and federal habeas proceedings, the United States Supreme Court previously had rejected the petitioner’s claim that the rule preventing a trial court from directing a guilty verdict prohibited a reviewing court from finding that a Salamon error was harmless when the evidence presented at trial compelled such a conclu- sion, the Brecht standard was not so vague as to be difficult to apply, and the application of the Brecht standard would not be unfair to the petitioner but, rather, would strike a balance between bestowing a wind- fall on the petitioner and penalizing him for failing to anticipate this court’s reinterpretation of this state’s kidnapping statutes. 2. The habeas court correctly determined that the trial court’s failure to instruct the jury at the petitioner’s criminal trial in accordance with Salamon was harmless because it did not give rise to a risk of prejudice sufficient to undermine confidence in the verdict, and, accordingly, the October 26, 2021 CONNECTICUT LAW JOURNAL Page 5

339 Conn. 1 OCTOBER, 2021 3 Banks v. Commissioner of Correction Appellate Court’s judgment was reversed: under Salamon, the jury, having found abduction, restraint, and the criminal intent associated therewith, necessarily had to find the petitioner guilty of kidnapping under the applicable statute (§ 53a-92 (a) (2) (B)) unless it found that the restraint and associated criminal intent were limited to that inherent in the robberies; moreover, even though the petitioner did not move his victims a great distance or restrain them for a long period of time, the jury could not reasonably have found that the asportation and restraint were limited to that which was necessary to carry out the robberies, as the actions occurred after the objective of each robbery had been completed, were conducted in order to make it more difficult for the victims to summon assistance and to reduce the petitioner’s risk of detection, and subjected the victims to unique risks and harms, both physical and psychological, beyond those inherent in the robberies them- selves. (Three justices concurring separately in two opinions)

Argued December 16, 2019—officially released May 12, 2021**

Procedural History

Amended petition for a writ of habeas corpus, brought to the Superior Court in the judicial district of Tolland and tried to the court, Sferrazza, J.; judgment denying the petition, from which the petitioner, on the granting of certification, appealed to the Appellate Court, DiPentima, C. J., and Prescott, J., with Keller, J., dissenting, which reversed the judgment of the habeas court and remanded the case with direction to grant the petition, and the respondent, on the granting of certification, appealed to this court. Reversed; judg- ment directed. Laurie N. Feldman, special deputy assistant state’s attorney, with whom, on the brief, were Gail P. Hardy, former state’s attorney, and Jo Anne Sulik, supervisory assistant state’s attorney, for the appellant (respondent). Pamela S. Nagy, assistant public defender, for the appellee (petitioner). ** May 12, 2021, the date that this decision was released as a slip opinion, is the operative date for all substantive and procedural purposes. Page 6 CONNECTICUT LAW JOURNAL October 26, 2021

4 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Opinion

PALMER, J. In this certified appeal and the compan- ion case decided herewith; see Bell v. Commissioner of Correction, 339 Conn. 79, A.3d (2021); we again revisit our decision in State v. Salamon, 287 Conn. 509, 949 A.2d 1092 (2008), in which we overruled our long-standing interpretation of this state’s kidnapping statutes and held that, when a criminal defendant is charged with kidnapping in conjunction with another underlying crime, such as rape or assault, the jury must be instructed that the defendant cannot be convicted of kidnapping if the restraint imposed on the victim was merely incidental to that underlying crime. See id., 542–50. We now must resolve two questions left open by Salamon and its progeny. First, when a petitioner seeking habeas relief establishes a Salamon error, does the habeas court assess the harm of that error according to the legal standard that the United States Supreme Court articulated in Brecht v. Abrahamson, 507 U.S. 619, 623, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993) (new trial is mandated if instructional error ‘‘had [a] substan- tial and injurious effect or influence in determining the jury’s verdict’’ (internal quotation marks omitted)), or the more petitioner friendly standard that the high court adopted in Neder v. United States, 527 U.S. 1, 18, 119 S. Ct. 1827, 144 L. Ed. 2d 35 (1999) (new trial is required unless it is ‘‘clear beyond a reasonable doubt that a rational jury would have found the defendant guilty absent the [instructional] error’’)? Second, did the habeas court in the present case, in denying the habeas petition of the petitioner, Mark Banks, and ruling in favor of the respondent, the Commissioner of Correc- tion, correctly conclude, as a matter of law, that a Sala- mon error is harmless when a perpetrator forcibly removes his victims from the scene of a robbery after having taken their property and then restrains them in order to facilitate his escape? Or, in the alternative, did October 26, 2021 CONNECTICUT LAW JOURNAL Page 7

339 Conn. 1 OCTOBER, 2021 5 Banks v. Commissioner of Correction

the Appellate Court, in reversing the judgment of the habeas court, correctly conclude that the petitioner was entitled to a new trial because a jury reasonably could have found that the petitioner’s postrobbery movement and restraint of his victims was merely incidental to the underlying crimes and bore no independent criminal significance? See Banks v. Commissioner of Correc- tion, 184 Conn. App. 101, 131–32, 194 A.3d 780 (2018). We conclude that the Brecht standard, which governs federal habeas actions, applies in state habeas proceed- ings as well.1 We further conclude that the habeas court correctly determined that the trial court’s failure to instruct the petitioner’s jury in accordance with Sala- mon was harmless. Accordingly, we reverse the judg- ment of the Appellate Court. I In 1995, the petitioner was arrested and charged in connection with the armed robberies of two Bedding Barn stores, the first in Newington and the second in Southington. The two cases were consolidated and tried jointly before a jury in October, 1997. The facts that the jury reasonably could have found with respect to both robberies are set forth by Judge (now Justice) Keller in her dissent from the opinion of the Appellate Court majority in the present case. See Banks v. Com- missioner of Correction, supra, 184 Conn. App. 140–43 (Keller, J., dissenting). ‘‘With respect to the earlier of the two robberies . . . Michael Kozlowski testified that he was working at the Newington Bedding Barn on August 30, 1995, at about 9 p.m. As Kozlowski prepared to close 1 As we explain more fully hereinafter, the Brecht harmless error standard, as we apply it in the context of Salamon errors, requires essentially the same showing as that required for constitutional claims alleging ineffective assistance of counsel under Strickland v. Washington, 466 U.S. 668, 687, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984), and suppression of material, exculpatory evidence under Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963). Page 8 CONNECTICUT LAW JOURNAL October 26, 2021

6 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

the store, the petitioner entered. Kozlowski testified that he approached the petitioner with the belief that the petitioner was a customer. When Kozlowski showed the petitioner a king-size bed, the petitioner said, ‘let me count my money,’ and reached into his bag and produced a gun. Kozlowski testified that the petitioner said, ‘[d]on’t try anything, I’ll bust you one, just walk over to the register.’ The petitioner then told him to get behind the counter and pointed his gun at Kozlowski’s chest. Kozlowski testified that, after the petitioner took the money from the cash register and a wallet from his coworker, Howard Silk, ‘[the petitioner] moved [Koz- lowski and Silk] . . . down to the hallway into the bathroom and . . . he then put [them] into the bath- room and put a mop handle or something behind the door.’ Kozlowski testified that the petitioner, as they walked down the hallway to the bathroom, said, ‘[d]on’t try anything; I’ll blow your head off . . . .’ Kozlowski indicated that, after the petitioner closed the bathroom door and locked Kozlowski and Silk in there, ‘[they] ducked down thinking he was going to shoot through the door because it was only a piece of plywood, basi- cally, and, [a] couple of minutes after, [they] heard a bell, which [was] on the front door, [and which rings whenever someone enters or leaves the store] . . . . [They] then . . . kicked the door, basically, and went downstairs.’ ‘‘Silk testified that he also was working at the Newing- ton Bedding Barn during the evening of August 30, 1995. Silk stated that, as he was in the process of closing the store, he noticed the petitioner following Kozlowski toward the counter. As the petitioner and Kozlowski approached, Silk realized that the petitioner was point- ing a gun at Kozlowski’s back. Silk testified that the petitioner told Kozlowski and Silk that he wanted the money, so Kozlowski took the money from the register as the petitioner pointed the gun at Silk’s chest. After October 26, 2021 CONNECTICUT LAW JOURNAL Page 9

339 Conn. 1 OCTOBER, 2021 7 Banks v. Commissioner of Correction

Silk told the petitioner that there was no safe inside the store, the petitioner led Silk and Kozlowski toward the back of the store at gunpoint. Silk testified that he handed the petitioner the $17 in his wallet and then the petitioner ‘proceeded to put [them] into the bathroom area’ and attempted to jam the door with a mop handle. Silk testified that he believed that the petitioner put them in the bathroom so that he could escape and that, after less than two minutes [possibly less than one minute], he heard the bell ring that ‘goes off when [the door] opens and . . . [he] hoped that [the bell had rung] when [the petitioner] left.’ After waiting for thirty seconds after hearing the doorbell ring, Silk and Koz- lowski were easily able to open the bathroom door. Silk testified that they went downstairs into the basement of the building to the warehouse . . . to call 911 and wait for the police to arrive in the event that the peti- tioner was still on the first floor. ‘‘In the second case, Kelly Wright testified that she was working at the Southington Bedding Barn on Sep- tember 13, 1995. She recalled that, at 8:55 p.m., five minutes before the store was set to close, while Wright’s roommate, Idelle Feltman, was waiting in the store to take her home, the petitioner and an unidentified female entered the store. Wright testified that the petitioner and the female split up and appeared to be shopping for king-size beds. Wright testified that she was sitting behind the store counter when the petitioner arrived and that she rose in order to greet him because it was store policy to do so whenever a potential customer arrived. Before Wright could make it around the counter, however, the petitioner told her to get on the floor. Wright testified that she noticed that the peti- tioner had a gun in his hand and was holding it out parallel to the floor. The petitioner told Feltman to get the money from the register. Feltman gave the peti- tioner the money in the register in a bank bag. Wright Page 10 CONNECTICUT LAW JOURNAL October 26, 2021

8 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

testified that the petitioner then inquired if there was a basement in the store, and Feltman responded by telling the petitioner that there was no basement . . . but [that] there was a bathroom. Wright testified that the petitioner led her and Feltman to the bathroom at gunpoint and told them to enter the bathroom, lock the door, and ‘not to be a hero, let the cops do their jobs.’ Wright stated that she heard a buzzer go off, which indicated that the door to the store had been opened. She and Feltman waited for a ‘little bit,’ unlocked the door, and left the bathroom to call 911. Wright estimated that about five to six minutes elapsed between the time the petitioner entered the store [and] the time she and Wright were able to contact the police. ‘‘Feltman testified that she went to the Southington Bedding Barn to pick up Wright from work because the two planned to go out to dinner. During her testimony, she recalled that two people, the petitioner and a woman, entered the store right before closing and that the pair split up after they entered the store. Feltman testified that the petitioner approached the counter and removed a gun from his bag. He waved the gun and told her to give him the money in the register. Feltman emptied the register, which contained less than $100, and handed the money to the petitioner. Feltman testi- fied that, after he obtained the money, the petitioner inquired whether there was a basement in the store and that Feltman and Wright replied that there was no basement . . . but [that] there was a bathroom. Felt- man stated that the petitioner led her and Wright in a single-file line to the bathroom and then instructed them to enter, while aiming the gun at them and causing them to be scared. Feltman and Wright entered the bathroom and waited a minute or two after they heard the door buzzer that indicated someone had entered or left the store. At this point, the two left the bathroom and found a mattress that had been placed in the narrow hallway October 26, 2021 CONNECTICUT LAW JOURNAL Page 11

339 Conn. 1 OCTOBER, 2021 9 Banks v. Commissioner of Correction

leading to the bathroom as a ‘barricade . . . .’ Feltman testified that she pushed it off to the side and ‘walked right through.’ ’’ (Footnote omitted.) Id. The following additional facts and procedural history also are relevant to the present appeal. At trial, the petitioner did not seriously contest that the charged crimes occurred as alleged. Rather, his primary defense was that the various victims had misidentified him as the perpetrator and that no evidence other than the victims’ eyewitness testimony linked him to the crimes. Therefore, the vast majority of defense counsel’s ques- tions during cross-examination pertained to the identity issue, although counsel did inquire whether the perpe- trator had physically pushed the victims toward the bathrooms, how far the bathrooms were situated from the cash registers where the robberies occurred, and the ease with which the victims exited the bathrooms. There was testimony that the bathroom in the Newing- ton store was located approximately twenty-four feet from the register. During closing arguments, the prosecutor acknowl- edged in his introduction that Connecticut’s definition of kidnapping was counterintuitive, and he argued that ‘‘what we have here is clearly kidnapping in the state of Connecticut.’’ His argument as to the alleged kidnap- ping focused entirely on the petitioner’s conduct after having completed the robberies: ‘‘[T]he use of the gun to herd the people into the back room, I’d argue to you, is that type of restraint. Obviously, they are not free to leave . . . even that particular area, the path [toward] the bathroom. They are . . . obviously not free to leave the building. They are not free to go call the police. They are restricted in their movements, again, and the intent of the [petitioner] or again since we are right now discussing what’s not in dispute, the intent of the person who committed this crime was to accomplish the commission of a felony, that is, this was part of Page 12 CONNECTICUT LAW JOURNAL October 26, 2021

10 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

the plan to allow him to escape from committing this robbery.’’ The prosecutor thus argued the case as if (1) it was undisputed that the perpetrator had forced the victims to enter and remain in the bathroom at gunpoint in order to facilitate his escape from the robbery scene, and (2) the kidnapping allegedly was predicated on the restraint of the victims only after the perpetrator had taken the money from the cash registers. In his closing argument, defense counsel conceded that the robberies took place as alleged, contesting only that the petitioner was the perpetrator. He also essen- tially agreed with the prosecutor that the underlying facts were not in dispute: ‘‘As [the prosecutor] pointed out to you, this case begins and ends with identification, and I have conceded essentially that to you.’’ Defense counsel also conceded that the conduct that the state had identified as kidnapping transpired after the robber- ies had occurred: ‘‘You and I, we don’t want to be on the other side of what we think might be a real, live gun with a person who appears to be robbing us and then asking us to go to a back room and saying, don’t be a hero.’’ (Emphasis added.) The jury found the petitioner guilty of four counts of kidnapping in the first degree in violation of General Statutes § 53a-92 (a) (2) (B),2 four counts of robbery in the first degree in violation of General Statutes § 53a- 134 (a) (4), and two counts of criminal possession of a pistol or revolver in violation of General Statutes (Rev. to 1995) § 53a-217c. The trial court rendered judgment in accordance with the jury verdict and sentenced the petitioner to a total effective sentence of twenty-five years incarceration consecutive to any sentence the petitioner was presently serving. The Appellate Court rejected the petitioner’s claims on appeal, and this court denied his petition for certification to appeal to this 2 The relevant statutory text is discussed in part III B of this opinion. October 26, 2021 CONNECTICUT LAW JOURNAL Page 13

339 Conn. 1 OCTOBER, 2021 11 Banks v. Commissioner of Correction

court. State v. Banks, 59 Conn. App. 112, 113–14, 755 A.2d 951, cert. denied, 254 Conn. 950, 762 A.2d 904 (2000). At no time on appeal did the petitioner challenge the propriety of the trial court’s jury instructions on kid- napping. Thereafter, in 2008, more than one decade after the petitioner’s trial, we decided Salamon, in which ‘‘we reconsidered our long-standing interpretation of our kidnapping statutes, General Statutes §§ 53a-91 through 53a-94a. . . . [In that case] [t]he defendant [Scott Sala- mon] had assaulted the victim at a train station late at night . . . and ultimately was charged with kidnapping in the second degree in violation of [General Statutes] § 53a-94, unlawful restraint in the first degree, and risk of injury to a child. . . . At trial, [Salamon] requested a jury instruction that, if the jury found that the restraint had been incidental to the assault, then the jury must [find him not guilty] of the charge of kidnapping. . . . [Consistent with established precedent of this court] [t]he trial court declined to give that instruction [and Salamon was convicted of second degree kidnapping in addition to the two other crimes]. . . . ‘‘[On appeal, Salamon requested that we reexamine] our long-standing interpretation of the kidnapping stat- utes to encompass even restraints that merely were incidental to and necessary for the commission of another substantive offense, such as robbery or sexual assault. . . . We [did so and] ultimately concluded that [o]ur legislature . . . intended to exclude from the scope of the more serious crime of kidnapping and its accompanying severe penalties those confinements or movements of a victim that are merely incidental to and necessary for the commission of another crime against that victim. Stated otherwise, to commit a kid- napping in conjunction with another crime, a defendant must intend to prevent the victim’s liberation for a longer period of time or to a greater degree than that Page 14 CONNECTICUT LAW JOURNAL October 26, 2021

12 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

which is necessary to commit the other crime. State v. Salamon, supra, 287 Conn. 542. ‘‘We explained in Salamon that a defendant may be convicted of both kidnapping and another substantive crime if, at any time prior to, during or after the commis- sion of that other crime, the victim is moved or confined in a way that had independent criminal significance, that is, the victim was restrained to an extent exceeding that which was necessary to accomplish or complete the other crime. Whether the movement or confinement of the victim is merely incidental to and necessary for another crime will depend on the particular facts and circumstances of each case. Consequently, when the evidence reasonably supports a finding that the restraint was not merely incidental to the commission of some other, separate crime, the ultimate factual determination must be made by the jury. For purposes of making that determination, the jury should be instructed to consider the various . . . factors [rele- vant thereto] . . . . Id., 547–48.’’ (Citations omitted; internal quotation marks omitted.) State v. Hampton, 293 Conn. 435, 459–60, 988 A.2d 167 (2009). Three years later, in Luurtsema v. Commissioner of Correction, 299 Conn. 740, 12 A.3d 817 (2011), we held that Salamon applies retroactively in habeas actions. See id., 751, 760 (plurality opinion). That is, an individ- ual, such as the petitioner in the present case, who was convicted before we decided Salamon can nevertheless bring a habeas action challenging his conviction on the ground that his or her jury was not properly instructed as to the meaning of the kidnapping statutes, as clarified in Salamon.3 3 After Luurtsema, we held, in Hinds v. Commissioner of Correction, 321 Conn. 56, 136 A.3d 596 (2016), that the procedural default rule does not bar postconviction claims that the trial court failed to instruct the jury as required under Salamon in cases rendered final before our decision in Salamon. Id., 61. We discuss our decision in Hinds in greater detail in parts II B and III D of this opinion. October 26, 2021 CONNECTICUT LAW JOURNAL Page 15

339 Conn. 1 OCTOBER, 2021 13 Banks v. Commissioner of Correction

Because, however, Luurtsema involved two ques- tions reserved by the habeas court for the advice of this court limited to the issue of retroactivity; see id., 743 (plurality opinion); we did not have occasion in that case to decide what harmless error standard applies to Salamon errors in the habeas context. Subsequently, in 2014, the petitioner in the present case filed the petition for a writ of habeas corpus that is the basis of this appeal. In his amended petition, the petitioner alleged a violation of his due process right to a fair trial under the fifth and fourteenth amendments to the United States constitution, challenging his four kidnap- ping convictions on the ground that the instructions given to the jury were not in accordance with Salamon.

The habeas court denied the petition. The court con- cluded that the respondent had demonstrated that the absence of a Salamon instruction at the petitioner’s criminal trial was harmless error because the ‘‘move- ments and confinements [of the victims] were perpe- trated after the crimes of robbery were committed and [could not] conceivably be regarded as coincidental with or necessary to complete the substantive crimes of robbery. Depriving someone of their freedom of movement by imprisoning them in a bathroom subse- quent to acquiring their money, although convenient for the robber, is not inherent in the crime of robbery. It is crystal clear that the petitioner’s intent and purpose for locking up his robbery victims [were] to postpone their summoning of assistance and reporting of the crime to [the] police, thus facilitating the petitioner’s escape from the scene and delaying detection of his crime, identity, and/or whereabouts. Also, the petitioner extended the period of infliction of duress and distress for the victims by restraining them beyond the time of fulfillment of his quest, [that is, his] seizure of cash.’’ (Emphasis in original.) Page 16 CONNECTICUT LAW JOURNAL October 26, 2021

14 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

The habeas court granted the petitioner’s petition for certification to appeal, and a divided panel of the Appellate Court reversed. Banks v. Commissioner of Correction, supra, 184 Conn. App. 132. The Appellate Court majority, applying the Neder harmless error stan- dard, determined that the absence of a Salamon instruc- tion was not harmless beyond a reasonable doubt. Id., 104, 132; see id., 112–13 n.7. Specifically, the majority, applying the six factor test that we set forth in State v. Salamon, supra, 287 Conn. 548, concluded that three of the factors tipped in the petitioner’s favor; see Banks v. Commissioner of Correction, supra, 130; and that a jury reasonably could have found that his movement and restraint of the four victims were part of ‘‘a continu- ous, uninterrupted course of conduct related to the robberies . . . .’’ Id., 132. Judge Keller penned a dissenting opinion. See id., 132–50 (Keller J., dissenting). She argued that five of the six Salamon factors pointed in the respondent’s favor and that no reasonable jury could conclude that the petitioner’s forcing his victims into an isolated bath- room at gunpoint, after having taken their valuables and emptied the cash registers, was merely incidental to and necessary for the commission of the robberies. Id., 143–50 (Keller, J., dissenting). We granted the respondent’s petition for certification to appeal, limited to the following two issues: (1) ‘‘When a habeas petitioner claims that the criminal trial court erred by omitting jury instructions on the intent element of kidnapping pursuant to [Salamon], is harm measured in accordance with [Brecht] or [Neder]?’’ And (2) ‘‘[d]id the Appellate Court [correctly] conclude that the absence of a Salamon instruction at the petitioner’s criminal trial was not harmless error?’’ Banks v. Com- missioner of Correction, 330 Conn. 950, 197 A.3d 391 (2018). Additional facts will be set forth as necessary. October 26, 2021 CONNECTICUT LAW JOURNAL Page 17

339 Conn. 1 OCTOBER, 2021 15 Banks v. Commissioner of Correction

II We begin by addressing the first certified question, that is, whether the Brecht standard or the Neder stan- dard applies to Salamon claims raised in habeas pro- ceedings. We agree with the respondent that the Brecht standard, as defined herein, is applicable. A Under Neder, which adopted the Chapman standard; see Chapman v. California, 386 U.S. 18, 24, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967); the state must demonstrate that a trial error was harmless beyond a reasonable doubt. See Neder v. United States, supra, 527 U.S. 15. In the context of a criminal conviction in which the trial court failed to instruct the jury on an element of the charged offense, the Neder standard often has been framed as also requiring that the element at issue have been uncontested at trial, and that the evidence tending to establish that element be overwhelming, before an error may be deemed harmless. See, e.g., State v. Vel- asco, 253 Conn. 210, 232, 751 A.2d 800 (2000). The fed- eral courts apply Neder to assess the harmlessness of most constitutional errors on direct review. See, e.g., Neder v. United States, supra, 7. By contrast, under Brecht, the harmlessness of consti- tutional errors is assessed according to whether the error ‘‘had [a] substantial and injurious effect or influ- ence in determining the jury’s verdict.’’ (Internal quota- tion marks omitted.) Brecht v. Abrahamson, supra, 507 U.S. 637. This standard originated as the Kotteakos test. See Kotteakos v. United States, 328 U.S. 750, 776, 66 S. Ct. 1239, 90 L. Ed. 1557 (1946). It has been adopted as the harmless error standard for nonconstitutional errors by the federal courts and, more recently, by this court in State v. Sawyer, 279 Conn. 331, 357, 904 A.2d 101 (2006), overruled on other grounds by State v. DeJesus, 288 Conn. 418, 953 A.2d 45 (2008). See State Page 18 CONNECTICUT LAW JOURNAL October 26, 2021

16 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

v. Sawyer, supra, 357 (nonconstitutional evidentiary error is harmless if ‘‘an appellate court has a fair assur- ance that the error did not substantially affect the ver- dict’’ (internal quotation marks omitted)). Brecht also is the standard by which the harmfulness of most consti- tutional errors is assessed in federal habeas actions.

The Brecht standard reserves the remedy of a new trial for errors resulting in ‘‘ ‘actual prejudice,’ ’’ as dis- tinguished from errors giving rise to a mere possibility of harm. Brecht v. Abrahamson, supra, 507 U.S. 637. We previously have likened the substantial prejudice necessary for relief from nonconstitutional error to error that is sufficiently prejudicial ‘‘to undermine confi- dence in the fairness of the verdict.’’ (Internal quotation marks omitted.) State v. Sawyer, supra, 279 Conn. 353; see also id., 352–54 (citing cases in which this court has applied ‘‘undermine confidence’’ test for purposes of determining harmfulness of nonconstitutional error). Notably, this is the same showing—characterized as a showing of a reasonable probability of a different result—required for constitutional claims alleging inef- fective assistance of counsel under Strickland v. Wash- ington, 466 U.S. 668, 687, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984), and the suppression of material, exculpatory evidence under Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963). The United States Court of Appeals for the Second Circuit has explained that, when Brecht is applied to a trial error in which the jury is not properly instructed as to an essential element of the charged crime, the reviewing court must undertake a careful, de novo review of the entire record and order a new trial unless the court is persuaded ‘‘that a properly instructed, rational jury would have found the [required element of the crime proven] beyond a reasonable doubt.’’ Peck v. United States, 106 F.3d 450, 456–57 (2d Cir. 1997). October 26, 2021 CONNECTICUT LAW JOURNAL Page 19

339 Conn. 1 OCTOBER, 2021 17 Banks v. Commissioner of Correction

Although some courts expressly place the burden of demonstrating harmlessness under Brecht on the state, the United States Supreme Court has expressed the view that it is ‘‘conceptually clearer’’ simply to place the onus on the reviewing court to determine whether an error substantially influenced the jury’s decision. O’Neal v. McAninch, 513 U.S. 432, 436, 115 S. Ct. 992, 130 L. Ed. 2d 947 (1995); see id., 436–37. We agree with the high court, however, that, when the reviewing court is in equipoise as to the question, the error must be deemed to have affected the verdict. See id., 435. For all intents and purposes, then, once a habeas petitioner has established a Salamon violation, the respondent bears the burden of demonstrating that the failure to instruct the jury in accordance with Salamon was harmless. B Although the petitioner argues that stare decisis requires us to apply Neder, in reality, this court never has resolved the question of whether Salamon errors should be assessed under Neder or Brecht on collateral review. Most of the Salamon cases that we have decided have reached us on direct review, where it is undisputed that Neder is the proper standard. Only twice have we had occasion to apply Salamon in the habeas context. In the first case, Luurtsema, as we noted, the issue reached us in the context of reserved questions from the habeas court regarding the retroactive applicability of Salamon. Luurtsema v. Commissioner of Correc- tion, supra, 299 Conn. 742 (plurality opinion). Accord- ingly, we did not have cause to address the harmless error issue. In dictum, however, we strongly suggested that Brecht—or at least something short of Neder— would be the appropriate legal standard. See id., 769–70 (plurality opinion). In Luurtsema, we explained why, in holding that Salamon applies retroactively on collateral Page 20 CONNECTICUT LAW JOURNAL October 26, 2021

18 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

review, we were not persuaded by the argument that ‘‘a finding of retroactivity would flood the court system with habeas petitioners seeking to overturn kidnapping convictions . . . .’’ (Internal quotation marks omitted.) Id., 769 (plurality opinion). We explained that, of those potential habeas cases that ‘‘fall within the ambit of Salamon . . . we expect that courts will be able to dispose summarily of many cases [when] it is suffi- ciently clear from the evidence presented at trial that the petitioner was guilty of kidnapping, as properly defined, [and] that any error arising from a failure to instruct the jury in accordance with the rule in Salamon was harmless.’’ (Footnote omitted.) Id., 769–70 (plural- ity opinion). The fact that we thought that many Sala- mon cases could be disposed of summarily on the ground of harmless error strongly suggests that we did not envision that such claims would be evaluated under the stringent Neder standard. The second case—the only one in which we had the opportunity to review a final judgment from a habeas action finding a Salamon violation—was Hinds v. Com- missioner of Correction, 321 Conn. 56, 136 A.3d 596 (2016). In that case, we expressly left open the question of whether Neder or Brecht applies to Salamon claims on collateral review. Specifically, after first having found that the Salamon error was not harmless beyond a reasonable doubt under Neder; id., 81; the majority in Hinds stated that ‘‘this court has not had the occasion to consider whether . . . a more stringent standard of harm should apply in collateral proceedings . . . . Brecht v. Abrahamson, supra, 507 U.S. 623 . . . .’’ Hinds v. Commissioner of Correction, supra, 81. We concluded, however, that ‘‘[w]e need not decide in [Hinds] whether to enter the fray by adopting the stan- dard in Brecht and the uncertainties that accompany it . . . because . . . the petitioner would prevail even under the more stringent standard . . . .’’ Id., 83. We October 26, 2021 CONNECTICUT LAW JOURNAL Page 21

339 Conn. 1 OCTOBER, 2021 19 Banks v. Commissioner of Correction

then proceeded to explain at some length why the Sala- mon error in that case would qualify as harmful, even if we were to employ a standard more favorable to the respondent. See id., 83–94; see also Epps v. Commis- sioner of Correction, 323 Conn. 901, 150 A.3d 679 (2016) (granting certification to appeal to resolve question, ‘‘unresolved by Hinds,’’ of ‘‘[w]hether . . . in a collat- eral proceeding, [when] the petitioner claims that the trial court erred by omitting an element of the criminal charge in its final instructions to the jury . . . harm [is] measured in accordance with Brecht . . . or . . . Neder’’ (emphasis added)).4 Accordingly, we disagree with the petitioner’s contention that precedent compels us to apply Neder. If anything, our prior case law sug- gests the opposite. C We conclude that Brecht, as characterized herein, provides the proper harmless error standard in state habeas actions, at least with respect to Salamon errors. We reach this conclusion for at least the following three reasons. 1 First, evaluating Salamon claims according to the Brecht standard is more consistent with how we handle other claims of error in habeas actions. The parties disagree as to which harmlessness standard is more consistent with our broader habeas jurisprudence, with each contending that adoption of the other party’s pre- ferred standard would create a ‘‘ ‘confused patchwork’ ’’ or draw ‘‘arbitrary distinctions . . . .’’ Because harm- lessness standards vary depending on the stage of review and the type of error at issue, some appearance 4 After granting certification to appeal, this court ultimately determined in Epps that certification had been improvidently granted and dismissed the respondent’s appeal. See Epps v. Commissioner of Correction, 327 Conn. 482, 485, 175 A.3d 558 (2018). Page 22 CONNECTICUT LAW JOURNAL October 26, 2021

20 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

of arbitrariness, from some vantage point, is, perhaps, inevitable, regardless of which standard we adopt. See United States v. Dominguez Benitez, 542 U.S. 74, 86, 124 S. Ct. 2333, 159 L. Ed. 2d 157 (2004) (Scalia, J., concurring in the judgment) (‘‘[b]y my count, [the United States Supreme] Court has adopted no fewer than four assertedly different standards of probability relating to the assessment of whether the outcome of trial would have been different if [the] error had not occurred, or if omitted evidence had been included’’ (emphasis omitted)); Brecht v. Abrahamson, supra, 507 U.S. 649 (White, J., dissenting) (‘‘[o]ur habeas jurispru- dence is taking on the appearance of a confused patchwork’’). It bears noting, however, that the vast majority of habeas cases that we review are subject to harm- lessness or prejudice review under some standard that is more onerous, from the petitioner’s standpoint, than Chapman/Neder. For example, many habeas cases present Strickland or Brady claims that require the petitioner to establish a reasonable probability that, but for the error, the result of the proceeding would have been different. See, e.g., Michael T. v. Commis- sioner of Correction, 307 Conn. 84, 91–92, 52 A.3d 655 (2012). The primary context in which we have required a showing of harmlessness beyond a reasonable doubt in habeas cases is when the prosecution knowingly has relied on or failed to correct false testimony at trial. See, e.g., Adams v. Commissioner of Correction, 309 Conn. 359, 370–73, 71 A.3d 512 (2013). In those cases, we reasoned that a ‘‘strict standard of materiality is appropriate . . . not just because [the constitutional violations] involve prosecutorial [impropriety], but more importantly because they involve a corruption of the truth-seeking function of the trial process.’’ (Internal quotation marks omitted.) Id., 372. Neither of those October 26, 2021 CONNECTICUT LAW JOURNAL Page 23

339 Conn. 1 OCTOBER, 2021 21 Banks v. Commissioner of Correction

rationales applies in the Salamon context, in which the state tried cases appropriately, according to our then established interpretation of the kidnapping statutes. The first concurrence ‘‘see[s] no need for this stan- dard to be ‘consistent with how we handle other claims of error in habeas actions.’ ’’ Instead, the first concur- rence contends that alleged Salamon errors are unique among habeas claims in that ‘‘[a] Salamon violation . . . is a determination that the state, in prosecuting a defendant, was unconstitutionally relieved of proving an essential element of the crime of kidnapping.’’ Although we are not bound by the federal courts in this regard, we find it noteworthy that they have roundly rejected the argument that we should carve out this particular category of constitutional error for review according to a different standard. In California v. Roy, 519 U.S. 2, 117 S. Ct. 337, 136 L. Ed. 2d 266 (1996), the United States Supreme Court confronted a scenario strikingly similar to that presented by our Salamon jurisprudence. One year after a California jury found the petitioner, Kenneth Roy, guilty of first degree murder on a theory of felony murder, the California Supreme Court clarified the state’s felony murder law, holding that a stricter instruction on the intent element was required than the one that had been given in Roy. See id., 3, citing People v. Beeman, 35 Cal. 3d 547, 561, 674 P.2d 1318, 199 Cal. Rptr. 60 (1984). The United States Supreme Court agreed with the lower courts that, in light of Beeman, Roy had been convicted on the basis of a ‘‘misdescription of an element of the crime . . . .’’ (Internal quotation marks omitted.) California v. Roy, supra, 5. Nevertheless, the high court held that, not only was such an error not a structural error that defies analysis by harmless error standards; id.; but, on collat- eral review, it was subject to harmless error analysis under the Brecht standard rather than the Neder stan- dard. See id., 5–6. Page 24 CONNECTICUT LAW JOURNAL October 26, 2021

22 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Although Roy involved a federal collateral attack on a state conviction and, therefore, arguably was predi- cated in part on considerations of comity and federal- ism, its holding has been applied in habeas actions challenging federal convictions, where those concerns do not weigh in the balance. Peck v. United States, supra, 106 F.3d 450, a decision of the United States Court of Appeals for the Second Circuit, provides an excellent case in point. In Peck, as in Roy and Salamon, an intervening court decision—this time by the United States Supreme Court itself—meant that the habeas petitioner in that case had been found guilty by a jury that had not been properly instructed as to the intent element of the charged crime, namely, structuring cash transactions to evade bank reporting requirements in violation of various provisions of title 31 of the United States Code. Id., 451–53. The Court of Appeals, after reviewing its harmless error jurisprudence; id., 453–55; concluded that ‘‘Brecht sets forth the correct methodol- ogy for determining if an instructional error of the type present in Roy is harmless.’’ Id., 456. As we discussed, the court emphasized that Brecht requires the reviewing court to undertake a careful, de novo review of the entire record; see id., 456–57; and to order a new trial unless the court is able to conclude ‘‘that a properly instructed, rational jury would have found the [required element of the crime proven] beyond a reasonable doubt.’’ Id., 457. Although we are free to diverge from the federal courts in defining the legal standards that govern appellate review of claims raised in state habeas petitions, we find Roy and Peck to be persuasive authority. As the respondent has argued, it is a close question whether the failure to give a Salamon instruction even consti- tutes error, let alone reversible error, in a case such as this, in which the conduct on which the kidnapping convictions were predicated did not occur until after October 26, 2021 CONNECTICUT LAW JOURNAL Page 25

339 Conn. 1 OCTOBER, 2021 23 Banks v. Commissioner of Correction

the underlying robberies had been completed. By con- trast, as the companion case demonstrates, the Brecht standard provides substantial protection and mandates a new trial when a reviewing court does not retain full confidence in the fairness of the conviction. See Bell v. Commissioner of Correction, supra, 339 Conn. 82, 83. Other types of constitutional errors likewise run the gamut, ranging from more minor, technical violations to serious deprivations of fundamental constitutional rights. We see no reason to add epicycle upon epicycle by carving out a special standard of review solely for one subset of habeas claims. 2 Second, a number of our sister states have followed the federal courts in adopting Brecht as the proper harmlessness standard for collateral review of constitu- tional errors. See, e.g., White v. State, 729 So. 2d 909, 915 (Fla. 1999); State v. Thomas, 750 So. 2d 1114, 1126 (La. App. 1999), writ denied, 795 So. 2d 1203 (La. 2001); Sanchez v. State, 272 Or. App. 226, 241 n.10, 355 P.3d 172, review denied, 358 Or. 449, 366 P.3d 719 (2015); Ex parte Fierro, 934 S.W.2d 370, 372 (Tex. Crim. App. 1996), cert. denied sub nom. Fierro v. Texas, 521 U.S. 1122, 117 S. Ct. 2517, 138 L. Ed. 2d 1019 (1997); see also Hittson v. Humphrey, Docket No. 5:01-CV-384 MTT, 2012 WL 5497808, *36 (M.D. Ga. November 13, 2012) (Georgia habeas court applied Brecht), rev’d in part on other grounds sub nom. Hittson v. GDCP War- den, 759 F.3d 1210 (11th Cir. 2014), cert. denied sub nom. Hittson v. Chatman, 576 U.S. 1028, 135 S. Ct. 2126, 192 L. Ed. 2d 887 (2015). We do not mean to suggest that Brecht is the prevailing approach among our sister states. A number of jurisdictions apply Neder on collateral as well as on direct review of constitutional errors. See, e.g., In re Martinez, 3 Cal. 5th 1216, 1224–25, 407 P.3d 1, 226 Cal. Rptr. 3d 315 (2017); Guam v. Ojeda, Docket No. CRA10-011, 2011 WL 6937376, *13 (Guam Page 26 CONNECTICUT LAW JOURNAL October 26, 2021

24 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

December 23, 2011); Hill v. State, 615 N.W.2d 135, 140–41 (N.D. 2000). Still, the fact that other states have adopted and successfully applied Brecht goes a long way toward addressing the potential counterarguments discussed hereinafter. 3 Third, and most important, insofar as Brecht requires that a new trial be granted due to the omission of a Salamon instruction unless, despite the omission, a reviewing court retains confidence in the fairness of the kidnapping conviction, the Brecht test affords a habeas petitioner significant protection. Indeed, under Brecht, a petitioner seeking a new trial because of such an omission is afforded no less protection than a peti- tioner who has established, under Strickland, that coun- sel’s performance fell below constitutional standards, or, under Brady, that the state failed to turn over excul- patory information. In fact, a petitioner who, upon establishing a Salamon violation, is entitled to applica- tion of the Brecht standard actually will receive more protection than a petitioner seeking relief under Strick- land or Brady because, as we have explained, the state bears the burden of disproving harm or prejudice. As in cases involving Strickland and Brady claims, if the appeals court, upon review of the petitioner’s collateral attack on his conviction, is satisfied that the error—in this case the omission of a Salamon instruction—does not call into question the fairness of that conviction, then it seems clear that a new trial is not constitutionally required. Consequently, for the foregoing reasons, we conclude that, when a habeas petitioner convicted of kidnapping has demonstrated that the jury was not properly instructed in accordance with Salamon, the state meets its burden of establishing harmlessness only if the reviewing court, following a thorough, de novo review of the record, has confidence that a properly October 26, 2021 CONNECTICUT LAW JOURNAL Page 27

339 Conn. 1 OCTOBER, 2021 25 Banks v. Commissioner of Correction

instructed jury would have found the defendant guilty beyond a reasonable doubt. D Several additional counterarguments have been asserted as to why we should assess the harmlessness of Sala- mon errors under Neder, even in habeas actions. For the following reasons, we ultimately do not find those arguments to be persuasive. 1 First, the petitioner and the first concurrence argue that to adopt the Brecht standard would be inconsistent with the rationales that animated our decisions in Sala- mon and Luurtsema. In Brecht, the United States Supreme Court offered three principal rationales for why a different harm standard should govern constitu- tional error in habeas actions than on direct appeal: (1) the presumption of finality following direct review of a conviction and the practical challenges that the state faces in potentially having to retry a petitioner years or—as in this case—decades after a crime; see Brecht v. Abrahamson, supra, 507 U.S. 633; (2) the extraordinary nature of the habeas remedy, which should be granted only to those petitioners who have been ‘‘grievously wronged and for whom belated liberation is little enough compensation’’ (internal quotation marks omit- ted) id., 633–34; and (3) concerns of federalism and comity, which counsel that federal courts defer to state courts that already have fully reviewed a petitioner’s case and have found no reversible error. Id., 635; see also Fry v. Pliler, 551 U.S. 112, 117, 127 S. Ct. 2321, 168 L. Ed. 2d 16 (2007) (discussing ‘‘primary reasons’’ for Brecht decision). The petitioner argues that, as a general matter, the Brecht standard is not suitable for state habeas actions because the federalism and comity concerns on which Page 28 CONNECTICUT LAW JOURNAL October 26, 2021

26 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Brecht was in part predicated do not apply when a state court is deciding a state habeas claim in the first instance. He further argues that Brecht is particularly ill-suited to the review of Salamon claims. He contends that, in Salamon and Luurtsema, we assumed that, although the finality of judgments is an important con- sideration, ‘‘the interests of finality must give way to the demands of liberty and a proper respect for the intent of the legislative branch.’’ Luurtsema v. Commis- sioner of Correction, supra, 299 Conn. 766 (plurality opinion). Accordingly, he posits, we already have decided that the finality considerations addressed in Brecht are trumped by other considerations in the unique Salamon context. The first concurrence echoes these points. This argument, while perhaps facially appealing, ulti- mately is unpersuasive. As we noted, two of the three principal rationales on which Brecht relied—the finality of judgments and the extraordinary nature of the habeas remedy—apply to state court habeas proceedings no less than to their federal counterparts. Moreover, subse- quent to Brecht, the United States Supreme Court has downplayed the importance of the federalism and com- ity considerations discussed in that decision. Most nota- bly, in Fry, the high court, resolving a split among the federal courts of appeals, held that Brecht is the proper standard for assessing harm in a federal habeas action, even when the state courts failed to recognize the con- stitutional error and did not review it for harmlessness under Neder. See Fry v. Pliler, supra, 551 U.S. 114, 121–22; see also Davis v. Ayala, 576 U.S. 257, 268, 135 S. Ct. 2187, 192 L. Ed. 2d 323 (2015) (‘‘[t]he Brecht standard reflects the view that a [s]tate is not to be put to th[e] arduous task [of retrying a defendant] based on mere speculation that the defendant was prejudiced by trial error; the court must find that the defendant was actually prejudiced by the error’’ (internal quota- October 26, 2021 CONNECTICUT LAW JOURNAL Page 29

339 Conn. 1 OCTOBER, 2021 27 Banks v. Commissioner of Correction

tion marks omitted)); Calderon v. Coleman, 525 U.S. 141, 146, 119 S. Ct. 500, 142 L. Ed. 2d 521 (1998) (finality concerns on which Brecht was predicated were espe- cially compelling when seventeen years had passed since petitioner’s sentencing). Indeed, in the wake of Fry, courts and commentators have opined that the finality and habeas specific concerns were the high court’s principal rationales for adopting a less onerous harmless error standard in Brecht. See, e.g., Hittson v. GDCP Warden, supra, 759 F.3d 1276 (Carnes, C. J., concurring) (‘‘Nothing in Brecht implies, let alone clearly establishes, that state courts must apply [Neder] on collateral review. If anything, Brecht’s principal rationale—that ‘collateral review is different from direct review’ and that the ‘substantial and injurious effect’ standard is ‘better tailored to the nature and purpose of collateral review than the [Neder] stan- dard’—implies that state courts, like federal courts, are not bound to apply the [Neder] standard when conduct- ing collateral review. The . . . Fry decision took Brecht one step further away from [that] position . . . .’’); 7 W. LaFave et al., Criminal Procedure (4th Ed. 2015) § 28.9 (b), pp. 382, 398 (most federal courts of appeals apply Brecht in habeas actions challenging federal convictions, despite inapplicability of federal- ism and comity concerns); J. Sullivan, ‘‘Danforth, Retro- activity, and Federalism,’’ 61 Okla. L. Rev. 425, 497 (2008) (‘‘in Brecht . . . the [c]ourt recognized that competing interests, including finality, warrant applica- tion of a different standard for proof of harm in evaluat- ing claims of constitutional error asserted in federal habeas proceedings’’ (footnote omitted)). As we noted, a number of our sister state courts have adopted Brecht as the proper standard for assessing harmless error in state habeas actions, despite the inapplicability of the federalism and comity considerations discussed in that decision. Page 30 CONNECTICUT LAW JOURNAL October 26, 2021

28 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

We also reject the petitioner’s contention that our cases already have struck the balance against the state’s inter- ests in preserving the finality of convictions. His argu- ment confuses the existence of the right with the proper remedy for its violation. See United States v. Cappas, 29 F.3d 1187, 1193 (7th Cir. 1994) (‘‘the [Brecht] standard does not define the scope of underlying substantive constitutional rights . . . but is rather about what rem- edy is required [when] it is agreed that constitutional rights have been violated’’ (emphasis omitted)). In Sala- mon, we explained why our prior cases had not accu- rately construed Connecticut’s kidnapping laws and how, upon a closer examination, it was clear that the legislature did not intend to criminalize as kidnapping conduct that was merely incidental to the commission of another crime, that is, conduct with no real indepen- dent criminal significance separate and apart from the underlying crime. See State v. Salamon, supra, 287 Conn. 542. In Luurtsema, we further explained why due process does not permit the continued incarceration of someone who has been convicted of a crime that he or she did not commit, as properly defined. See Luurt- sema v. Commissioner of Correction, supra, 299 Conn. 758–59 (plurality opinion). In that case, however, we also made clear that a habeas petitioner may not prevail on a Salamon claim when ‘‘continued incarceration would not represent a gross miscarriage of justice, such as [when] it is clear that the legislature did intend to criminalize the conduct at issue, if perhaps not under the precise label charged. In situations [in which] the criminal justice system has relied on a prior interpreta- tion of the law so that providing retroactive relief would give the petitioner an undeserved windfall, the tradi- tional rationales underling the writ of habeas corpus may not favor full retroactivity. See Guzman v. Greene, [425 F. Supp. 2d 298, 315 (E.D.N.Y. 2006), aff’d 337 Fed. Appx. 27 (2d Cir. 2009)] (‘it is certainly not unjust, let October 26, 2021 CONNECTICUT LAW JOURNAL Page 31

339 Conn. 1 OCTOBER, 2021 29 Banks v. Commissioner of Correction

alone manifestly unjust, to keep a murderer in jail’).’’ Luurtsema v. Commissioner of Correction, supra, 764 (plurality opinion). As we noted, we further emphasized in Luurtsema that reviewing courts will be able to sum- marily dispose of many Salamon errors as harmless. Id., 769–70 (plurality opinion). Accordingly, although we certainly have not treated finality as the be all and end all, we also have never said that the mere fact that a petitioner was found guilty of kidnapping by an improperly instructed jury necessarily amounts to reversible error. 2 A related argument advanced by the petitioner is that for us to apply Brecht in the present case, and to find the errors harmless because we conclude that the evidence presented at trial supports a kidnapping conviction under the proper definition of that crime, would be tantamount to directing a verdict for the state. The United States Supreme Court has expressly rejected this argument. Only a small share of constitutional errors are struc- tural, that is, so presumptively harmful that they require automatic reversal. See, e.g., Washington v. Recuenco, 548 U.S. 212, 218, 126 S. Ct. 2546, 165 L. Ed. 2d 466 (2006). Most, rather, are subject to harmless error review. See, e.g., id. (‘‘We have repeatedly recognized that the commission of a constitutional error at trial alone does not entitle a defendant to automatic reversal. Instead, most constitutional errors can be harmless.’’ (Internal quotation marks omitted.)). This includes errors in instructing the jury as to the elements of a crime. We have said this not only in Luurtsema and other Salamon cases, but also in numerous other con- texts. See, e.g., State v. Thompson, 305 Conn. 806, 817– 18, 48 A.3d 640 (2012) (erroneous instruction concern- ing circumstances under which jury properly could Page 32 CONNECTICUT LAW JOURNAL October 26, 2021

30 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

aggregate value of stolen property for purpose of determining whether state had proven element of offense of first degree larceny requiring theft of prop- erty worth more than $10,000); Small v. Commissioner of Correction, 286 Conn. 707, 726–29, 946 A.2d 1203 (court improperly failed to instruct jury on definition of ‘‘attempt,’’ even though petitioner was charged with felony murder predicated on, inter alia, attempted rob- bery), cert. denied sub nom. Small v. Lantz, 555 U.S. 975, 129 S. Ct. 481, 172 L. Ed. 2d 336 (2008); State v. McDonough, 205 Conn. 352, 354–62, 533 A.2d 857 (1987) (erroneous instruction concerning inferences jury rea- sonably might draw from circumstantial evidence), cert. denied, 485 U.S. 906, 108 S. Ct. 1079, 99 L. Ed. 2d 238 (1988). Indeed, Neder itself stands for the proposition that the omission of a single element of a crime from the jury charge is not a structural constitutional error that is exempt from harmless error analysis. Washing- ton v. Recuenco, supra, 218–19; Lanier v. United States, 220 F.3d 833, 838 (7th Cir.), cert. denied, 531 U.S. 930, 121 S. Ct. 312, 148 L. Ed. 2d 250 (2000). Accordingly, the well established rule that a trial court may not direct a verdict of guilty in a criminal trial; State v. Ubaldi, 190 Conn. 559, 573, 462 A.2d 1001, cert. denied, 464 U.S. 916, 104 S. Ct. 280, 78 L. Ed. 2d 259 (1983); see General Statutes § 54-89; in no way implies that a reviewing court cannot adjudge a Salamon error harm- less when the evidence presented at trial, assessed under the appropriate legal standard, compels such a conclusion. See Washington v. Recuenco, supra, 220–22 (rejecting argument that subjecting trial court’s failure to submit sentencing factor to jury to harmless error analysis amounted to improper directed verdict of guilty); Neder v. United States, supra, 527 U.S. 17–19 (same, with respect to failure to properly instruct jury as to elements of crime charged). October 26, 2021 CONNECTICUT LAW JOURNAL Page 33

339 Conn. 1 OCTOBER, 2021 31 Banks v. Commissioner of Correction

3 The petitioner’s next argument against adopting Brecht as the standard for assessing the harmlessness of Salamon errors on habeas review is that the Brecht standard is vague and not easily applied. Most of the cases on which the petitioner relies for this proposition, however, were decided in the mid-1990s, in the immedi- ate aftermath of the Brecht decision. Since that time, the United States Supreme Court has provided additional guidance, clarifying several of the key ambiguities left open in Brecht itself.5 See, e.g., O’Neal v. McAninch, supra, 513 U.S. 435–37 (resolving burden of proof ques- tions); see also Fry v. Pliler, supra, 551 U.S. 121–22 (clarifying that Brecht applies on federal habeas review, regardless of whether state courts recognized error and reviewed it for harmlessness beyond reasonable doubt). Perhaps most significant, the Brecht test, as character- ized herein, is the same familiar standard—namely, whether the error undermines confidence in the convic- tion—used for ascertaining prejudice under Strickland and Brady, a standard that we are called on to apply regularly. We do not expect that it will be unduly diffi- cult to apply that standard in ascertaining the harm caused by an omitted Salamon instruction when that task is undertaken with due regard for the broader context of our holding in Salamon, that is, to ensure that a defendant is not convicted of kidnapping unless the conduct at issue has criminal significance indepen- dent of the underlying offense. 5 As Justice Scalia noted in his concurring opinion in Dominguez Benitez, we also would observe that the very same critique can be leveled with respect to the other standards of prejudice/harmless error that we regularly apply. See United States v. Dominguez Benitez, supra, 542 U.S. 86–87 (Scalia, J., concurring in the judgment) (‘‘Such ineffable gradations of probability seem to me quite beyond the ability of the judicial mind (or any mind) to grasp, and thus harmful rather than helpful to the consistency and rationality of judicial [decision making]. That is especially so when they are applied to the hypothesizing of events that never in fact occurred. Such an enterprise is not [fact-finding], but closer to divination.’’). Page 34 CONNECTICUT LAW JOURNAL October 26, 2021

32 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

4 Finally, the first concurrence asserts that to apply Brecht in the present case would be unfair to the peti- tioner and to others similarly situated. That this court opted to revisit and revise our interpretation of the state’s kidnapping laws following his conviction is no more the fault of the petitioner than of the state. Although would-be offenders were on notice in 1995 that they could be charged with kidnapping solely on the basis of the restraint inherent in robberies or assaults, they, like the state, did not have any reason to try their cases with the Salamon distinction in mind. Moreover, it may seem discrepant to assess the impact of the instructional error according to the more forgiv- ing Brecht standard when, if we had decided Salamon one decade earlier, while the petitioner’s direct appeal was pending, the state would have borne the burden of proving that the error was harmless beyond a reason- able doubt. As we discussed, the somewhat scattershot nature of harmless error jurisprudence, with standards varying by the type of error at issue, the stage of review, and the jurisdiction in which the claim is reviewed, means that whichever standard we apply to Salamon errors in state habeas cases may appear to be unfair or incon- gruous from one vantage point or another. It certainly will not seem unjust from the respondent’s standpoint to require a showing that there is some reasonable likelihood that the failure to give the jury a Salamon instruction had a substantial and injurious effect or influence on the outcome before requiring the state to retry the petitioner for crimes that were committed more than twenty-five years ago. If the ultimate question is how to allocate the risk that an appellate court will revise its interpretation of a criminal statute such as § 53a-92 after a conviction has been obtained and affirmed on direct appeal, we think that the fairest and October 26, 2021 CONNECTICUT LAW JOURNAL Page 35

339 Conn. 1 OCTOBER, 2021 33 Banks v. Commissioner of Correction

most reasonable approach is to adopt the Brecht stan- dard, as defined herein, to tip the balance in favor of the petitioner in close cases, and to evaluate the facts of a particular case with an eye toward whether there is a reasonable likelihood that the petitioner could and would have presented a different and potentially suc- cessful defense with the benefit of Salamon’s guidance. This avoids bestowing a windfall on the petitioner but also does not penalize him for failing to anticipate our reinterpretation of the kidnapping statute.6 III A Having concluded that the harm associated with depriving the petitioner of a Salamon instruction should 6 The first concurrence also queries, first, why it is necessary to decide by which standard to assess the harmfulness of Salamon errors on collateral review and, second, whether it is fair to apply Brecht in the petitioner’s case when our trial courts and the Appellate Court, acting without the benefit of this opinion, have in the past applied Neder to other petitioners’ Salamon claims. The short answer to the first question is that it ordinarily is the duty of this court to clarify the law when called on to do so in the context of a justiciable case. In the present case, the respondent contends that the Appellate Court has applied the incorrect harmless error standard, and this court granted certification to address that question. Banks v. Com- missioner of Correction, 330 Conn. 950, 197 A.3d 391 (2018). Although it is true that, on rare occasion, we refrain, for prudential reasons, from resolving a legal question that is squarely and properly before us, that is the exception rather than the rule, and we perceive no compelling reason why we should shrink from our duty in this instance. The fact that today’s ruling potentially may impact a relatively small universe of future cases surely does not justify a failure to resolve the certified question, especially as the companion case itself demonstrates the broad range of configurations in which Salamon claims may present themselves. We are no more troubled by the first concurrence’s second charge. It is not at all uncommon for courts of last resort—the United States Supreme Court not least among them—to allow difficult legal questions to percolate in the lower courts for some time before ultimately resolving them. This, and the very nature of our judicial system, means that appellate tribunals frequently will resolve legal controversies in novel ways, such that future litigants are subject to rules different from those that bound past litigants. Far from unfair, this is the very essence of the common law. To hold ourselves yoked to a legal rule simply because the lower courts previously Page 36 CONNECTICUT LAW JOURNAL October 26, 2021

34 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

be assessed under the Brecht standard, we turn now to the second question in this appeal: whether the largely undisputed facts that were presented at trial and cred- ited by the jury satisfy the statutory definition of kidnap- ping, as clarified by this court, such that we are confident that a properly instructed jury would have found the petitioner guilty of kidnapping beyond a rea- sonable doubt notwithstanding the Salamon error. We note at the outset that the question of whether conduct bears independent criminal significance as kid- napping is one of law. It is, of course, the function of the jury to find the relevant facts, including, ultimately, whether a crime was committed with the intent neces- sary to qualify as kidnapping, namely, the specific intent to prevent the victim’s liberation and not simply to perpetrate the underlying crime. See State v. Salamon, supra, 287 Conn. 532, 547–48. At the same time, it is beyond cavil that it is the role of the judiciary to inter- pret the relevant statutes and to define, as a matter of law, what type of conduct constitutes kidnapping according to those statutes. See id., 529. As we explained in Salamon, a necessary corollary is that it falls to the courts to define the intent element of the crime of kidnapping; see id., 534–35; to delineate the ways in which kidnapping differs from coterminous crimes such as robbery and sexual assault; see id., 542; and to specify the factors that are relevant to that analy- sis; see id., 548; in light of our understanding of the legislative history of the kidnapping statutes and the policy objectives that animated their modern revision. See id., 542, 546. As we explain more fully hereinafter; see part III B of this opinion; our prior cases addressed the Salamon issue and defined the relevant factors in the context have applied it would be to turn the concept of controlling legal authority on its head. This we decline to do. October 26, 2021 CONNECTICUT LAW JOURNAL Page 37

339 Conn. 1 OCTOBER, 2021 35 Banks v. Commissioner of Correction

of the restraint involved in an ongoing sexual and/or physical assault. We now confront a distinct and novel scenario—asportation and confinement to facilitate a perpetrator’s escape following the completion of a rob- bery—in which we have not previously had cause to apply Salamon, and in which a different provision of the kidnapping statutes is at issue. Although it always will be for the jury to find the relevant facts and to determine whether the perpetrator had the requisite criminal intent, it falls to this court to define in the first instance how that criminal intent differs from the intent necessary to commit an underlying crime for this cate- gory of offenses, as well as how the factors that we have articulated and applied to assess harmless error in the sexual assault context operate in this novel arena. As we explained in Salamon, those questions ultimately are ones of legislative intent, framed by the history and policy rationales that animate the relevant statutes. See id., 529, 542. B We next consider the standards by which a reviewing court is to assess the harmfulness of a Salamon error in the context of a robbery in which the perpetrator moves and confines the victims after having forcibly taken valuables in their possession. For the reasons that follow, we are not persuaded by the reasoning of the Appellate Court majority, which concluded that a jury reasonably might find that the restraint and aspor- tation involved in the present case were undertaken as part of ongoing robberies and, therefore, that the petitioner might not have intended to restrain the vic- tims more than was necessary to carry out those robber- ies. See Banks v. Commissioner of Correction, supra, 184 Conn. App. 124–25, 131–32. 1 First, there is the matter of the statute at issue. In Salamon, the defendant was convicted of violating Page 38 CONNECTICUT LAW JOURNAL October 26, 2021

36 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

§ 53a-94 (a), which provides that ‘‘[a] person is guilty of kidnapping in the second degree when he abducts another person.’’ See also State v. Salamon, supra, 287 Conn. 529–30 (explaining distinction between kidnap- ping, which requires abduction, and lesser offense of unlawful restraint, which merely requires restraint). Most of our subsequent Salamon cases have involved violations of that statute; e.g., State v. Fields, 302 Conn. 236, 238, 24 A.3d 1243 (2011); or of § 53a-92 (a) (2) (A), which provides that ‘‘[a] person is guilty of kidnapping in the first degree when he abducts another person and . . . he restrains the person abducted with intent to . . . inflict physical injury upon him or violate or abuse him sexually . . . .’’ E.g., Hinds v. Commissioner of Correction, supra, 321 Conn. 59; State v. Ward, 306 Conn. 718, 721, 51 A.3d 970 (2012); State v. Hampton, supra, 293 Conn. 437–38; State v. DeJesus, supra, 288 Conn. 420. All of those cases involved allegations that the alleged kidnapping was intertwined with a sexual or physical assault. The present case, by contrast, requires that we con- strue General Statutes § 53a-92 (a) (2) (B), which pro- vides that ‘‘[a] person is guilty of kidnapping in the first degree when he abducts another person and . . . he restrains the person abducted with intent to . . . accomplish or advance the commission of a felony . . . .’’7 As we concluded in Salamon with respect to the underlying crime of assault; see State v. Salamon, supra, 7 We note that one of our Salamon cases, namely, State v. Flores, 301 Conn. 77, 79, 17 A.3d 1025 (2011), did involve an underlying robbery and a conviction under § 53a-92 (a) (2) (B). In that case, we concluded that the lack of a Salamon instruction was not harmless error because, among other things, the victim was released immediately after the defendant forcibly took her property and the state did not argue that the restraint occurred for any longer than was necessary to commit the robbery. Id., 85, 87. For that reason, Flores is readily distinguishable from the present case. October 26, 2021 CONNECTICUT LAW JOURNAL Page 39

339 Conn. 1 OCTOBER, 2021 37 Banks v. Commissioner of Correction

287 Conn. 542; it is clear that the legislature did not intend to criminalize as kidnapping unlawful restraint that is no greater than necessary for, and involves no wrongful intent other than that inherent in, the comple- tion of a robbery. It is equally clear, however, that the legislature, in adopting § 53a-92 (a) (2) (B), did intend that additional, gratuitous restraint used to accomplish or advance the commission of a robbery carry the added penalties associated with kidnapping. In other words, the mere fact that a perpetrator restrains a victim during the course of and in the service of a robbery does not mean that, under Salamon, the conduct does not constitute kidnapping. To so hold—or to permit a jury to so reason—would be to render § 53a-92 (a) (2) (B) a nullity, insofar as that statute criminalizes only such kidnappings. See State v. Buggs, 219 Kan. 203, 214, 547 P.2d 720 (1976) (construing similar Kansas statute). Accordingly, we are not persuaded by the reasoning of the Appellate Court majority that a jury reasonably might find that the petitioner’s conduct was not kidnap- ping merely because it occurred as part of the course of events of the robberies. See Banks v. Commissioner of Correction, supra, 184 Conn. App. 124–25, 131–32. As Judge Keller explained in her dissent and Judge Lavine in his dissent in the companion case, that is not the relevant legal inquiry. See id., 133 n.1, 145 n.6 (Keller, J., dissenting); see also Bell v. Commissioner of Correc- tion, 184 Conn. App. 150, 183 n.5, 194 A.3d 809 (2018) (Lavine, J., dissenting), aff’d, 339 Conn. 79, A.3d (2021). Rather, under Salamon, a jury, having found abduction, restraint, and the criminal intent associated therewith in the furtherance of a robbery, will necessar- ily find the petitioner guilty of kidnapping under § 53a- 92 (a) (2) (B) unless it also finds that the restraint, and the associated criminal intent, was limited to that inherent in the robbery itself. As we explain in the discussion that follows, when the question for the jury Page 40 CONNECTICUT LAW JOURNAL October 26, 2021

38 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

is properly framed in that manner, only one answer is reasonably possible in view of the facts of the pres- ent case. We find Virgin Islands v. Ventura, 775 F.2d 92 (3d Cir. 1985), to be instructive in this regard. In that case, the United States Court of Appeals for the Third Circuit construed a territorial statute that provided that ‘‘[w]ho- ever abducts, takes or carries away any person by force or threat with the intent to commit rape is guilty of kidnapping . . . .’’ Id., 96, quoting Act of June 30, 1983, No. 4838, 1983 V.I. Sess. Laws 100, 101 (codified at V.I. Code Ann. tit. 14, § 1052 (b)). As has this court; see, e.g., State v. Salamon, supra, 287 Conn. 548; the Third Circuit, in Virgin Islands v. Berry, 604 F.2d 221, 227 (3d. Cir. 1979), previously had applied a multifactor test to assess when, as a general matter, an alleged kidnapping committed in conjunction with another crime constitutes a discrete offense. In Ventura, how- ever, the Third Circuit held that the fact that the legisla- ture chose to specifically criminalize asportation incident to rape after Berry was decided meant that, with respect to cases falling under the new statute, certain factors of the test were necessarily satisfied and need not be considered on a case-by-case basis. See Virgin Islands v. Ventura, supra, 97 (‘‘[t]o apply the Berry factor [at issue in Ventura] would effectively override the will of the Virgin Islands legislature’’). Like- wise, in the present case, to instruct a jury that it could find that there was no kidnapping merely because the restraint occurred in the course of an ongoing robbery would effectively override the will of our state’s legisla- ture, which was to impose heightened penalties for precisely such conduct. 2 Second, the decision of the Appellate Court in this case not only runs afoul of the statutory language, but October 26, 2021 CONNECTICUT LAW JOURNAL Page 41

339 Conn. 1 OCTOBER, 2021 39 Banks v. Commissioner of Correction

also fails to take due account of the particular factual scenario presented by this case. This case is categori- cally distinct from all of our prior Salamon cases insofar as the petitioner indisputably had accomplished the criminal objective of his underlying crimes prior to the commencement of the alleged kidnapping. Under such circumstances, there simply is no concern that the intent of the legislature will be frustrated by prosecuting a defendant for kidnapping solely on the basis of the restraint inherent in or necessary to accomplish the underlying crime. Many if not most robbers choose to leave the scene immediately upon obtaining the fruits of their crime. Numerous sister state courts have con- cluded, as a matter of law, that a perpetrator’s choice to remain at the crime scene and further restrict a victim’s liberty after having robbed him or her manifests independent, criminal significance. See, e.g., Black v. State, 630 So. 2d 609, 619 (Fla. App. 1993) (when restraint and removal occurred after perpetrator had taken money from store, court deemed it ‘‘obvious that the movement and confinement . . . [were] not inher- ent in the nature of the robbery . . . [but] had some significance independent of the robbery’’ (internal quo- tation marks omitted)), review denied, 639 So. 2d 976 (Fla. 1994); State v. Blouvet, 965 S.W.2d 489, 492 (Tenn. Crim. App. 1997) (‘‘[h]olding [the victim] at gunpoint and moving her about the store [are] certainly not inci- dental to the already accomplished felony of aggravated robbery’’); State v. Allen, 94 Wn. 2d 860, 864, 621 P.2d 143 (1980) (concluding that brief abduction that occurred after robbery to facilitate perpetrators’ flight from scene was ‘‘a wholly separate event’’ and not inci- dental to robbery because ‘‘[n]either the flight from the scene of the robbery nor the means of flight therefrom [were] statutorily or logically . . . part of [the] rob- bery’’), overruled on other grounds by State v. Vladovic, 99 Wn. 2d 413, 662 P.2d 853 (1983); see also State v. Page 42 CONNECTICUT LAW JOURNAL October 26, 2021

40 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Golder, 127 Conn. App. 181, 190–91, 14 A.3d 399 (when defendant moved victim to different room and restrained her therein to facilitate his escape after tak- ing her jewelry, restraint bore independent criminal significance from completed burglary), cert. denied, 301 Conn. 912, 19 A.3d 180 (2011); People v. Bautista, 147 App. Div. 3d 1214, 1218, 47 N.Y.S.3d 503 (2017) (‘‘a kidnapping is generally deemed to merge with another offense . . . [when] there is minimal asportation immediately preceding the other crime or [when] the restraint and underlying crime are essentially simulta- neous’’ (internal quotation marks omitted)). Although we have not yet had occasion to expressly state this proposition, we implied it in two prior cases. In State v. Fields, supra, 302 Conn. 236, we addressed the state’s contention that the defendant was not enti- tled to a new trial on the challenged kidnapping count because the alleged kidnapping did not occur until after the assault of the victim had been completed. See id., 251. We indicated that ‘‘[w]e might agree with this con- tention’’ had the factual premises for the state’s argu- ment not been in dispute. Id. Subsequently, in Hinds v. Commissioner of Correction, supra, 321 Conn. 56, we distinguished certain cases from other jurisdictions that held the lack of a Salamon-type instruction to be harmless error because those cases involved ‘‘contin- ued restraint after completion of the nonkidnapping offenses . . . .’’ Id., 87. This makes sense. There is nothing specific to—let alone inherent in—the crime of robbery about forcing someone at gunpoint to the back of a store and restraining them in a bathroom or cooler. That conduct could just as well follow, and facilitate the offender’s escape from, a physical or sexual assault, or other crime. The purpose is to escape unhindered from a crime scene—which, presumably, is a goal of most crim- October 26, 2021 CONNECTICUT LAW JOURNAL Page 43

339 Conn. 1 OCTOBER, 2021 41 Banks v. Commissioner of Correction

inals—and the specific nature of the underlying crime is simply irrelevant. 3 The Appellate Court majority rejected the respon- dent’s argument—and the conclusion of the habeas court—that the petitioner’s asportation and restraint of his victims necessarily bore independent criminal significance because they did not commence until after the petitioner had accomplished his primary goal of obtaining wrongful possession of the cash in the store registers. See Banks v. Commissioner of Correction, supra, 184 Conn. App. 114, 120. The majority reasoned that, according to some authorities, a robbery is not necessarily completed at the time that a perpetrator obtains unlawful possession of a victim’s property and, therefore, that force exercised after the petitioner had taken the victims’ money reasonably could be consid- ered to be incidental to the robberies. See id., 120–28. Judge Keller responded, and we agree, that we need not engage in an ‘‘unduly legalistic’’ analysis of the pre- cise moment at which a robbery ends. Id., 148 (Keller, J., dissenting). Regardless of whether the robberies can be said to have been ongoing in some sense, even after the petitioner took possession of the victims’ money, the important point for present purposes is that any movement and confinement imposed after that time served a fundamentally different objective. Whether this category of restraints implicates the concerns that we addressed in Salamon and, more broadly, whether they are of the type that the legislature intended to independently criminalize are questions of law that fall to this court to resolve. C Both the Appellate Court majority, in concluding that the Salamon error in the present case prejudiced the Page 44 CONNECTICUT LAW JOURNAL October 26, 2021

42 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

petitioner, and Judge Keller, in maintaining that it did not, contended that the six Salamon factors, on bal- ance, tipped in favor of their positions. Id., 115–30; id., 143–50 (Keller, J., dissenting). Those factors are (1) the nature and duration of the victim’s movement or confinement, (2) whether that movement or confine- ment occurred during the commission of the separate offense, (3) whether the restraint was inherent in the nature of the separate offense, (4) whether the restraint prevented the victim from summoning assistance, (5) whether the restraint reduced the perpetrator’s risk of detection, and (6) whether the restraint created a significant danger or increased the victim’s risk of harm independent of that posed by the separate offense. State v. Salamon, supra, 287 Conn. 548. We agree with Judge Keller that, in cases such as this, in which it is undis- puted that the perpetrator unlawfully restrained his vic- tims following, and to facilitate his escape from the location of, a robbery, the Salamon factors typically will tip against the petitioner’s claim. But see footnote 14 of this opinion. There is little dispute that the first Salamon factor— the extent of the victims’ asportation and confine- ment—tends to favor the petitioner in the present case, as will often be true in cases of this ilk. See Banks v. Commissioner of Correction, supra, 184 Conn. App. 115–20 (comparing cases in which this court and Appel- late Court concluded that conduct did, or did not, as matter of law, have independent criminal significance). That is to say, a jury reasonably could conclude that moving robbery victims fewer than ten yards and con- fining them for, at most, a few minutes during a perpe- trator’s escape from the crime scene does not, simply by virtue of the times and distances involved, bear inde- pendent criminal significance. In the preceding discussion, we explained why the second and third Salamon factors ordinarily will tip October 26, 2021 CONNECTICUT LAW JOURNAL Page 45

339 Conn. 1 OCTOBER, 2021 43 Banks v. Commissioner of Correction

against the petitioner in a postrobbery kidnapping sce- nario. The conduct at issue occurred after the objective of the robbery had been completed. Nor was it inherent in the crime of robbery, insofar as many, if not most, robberies end with the perpetrator’s simply fleeing the premises rather than moving and confining the victims. In the discussion that follows, we explain why the final three Salamon factors also favor the respondent in such situations. 1 The parties, and the Appellate Court majorities and dissents, in both this case and the companion case, Bell, disagree as to how the Salamon factors are to be balanced and whether any of the six factors is disposi- tive. Although the relative importance of the various factors will vary depending on the context, we have made clear that the touchstone in any Salamon case, in assessing whether conduct associated with a rape, robbery, or assault has independent criminal signifi- cance as a kidnapping, is the intent of the offender. See State v. Salamon, supra, 287 Conn. 532 (‘‘the proper inquiry for a jury evaluating a kidnapping charge is not whether the confinement or movement of the victim was minimal or incidental to another offense against the victim but, rather, whether it was accomplished with the requisite intent, that is, to prevent the victim’s liberation’’); id., 534 (intent element is what defines abduction, the sine qua non of kidnapping); id., 542 (‘‘to commit a kidnapping in conjunction with another crime, a defendant must intend to prevent the victim’s liberation’’). We have continued to emphasize this point in our subsequent Salamon cases. See, e.g., Hinds v. Commissioner of Correction, supra, 321 Conn. 90 (‘‘the ultimate question [is] the perpetrator’s intent in taking these actions’’ (emphasis in original)); State v. Fields, supra, 302 Conn. 247 (referring to Salamon instruction as ‘‘incidental intent instruction’’); State v. Winot, 294 Page 46 CONNECTICUT LAW JOURNAL October 26, 2021

44 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Conn. 753, 762, 988 A.2d 188 (2010) (‘‘we repeatedly [have] explained [that] the touchstone for determining whether the movement or confinement at issue consti- tuted kidnapping was not its extensiveness, but rather, whether it was accomplished with the requisite intent’’ (internal quotation marks omitted)). The fourth and fifth Salamon factors—whether the restraint prevented the victim from summoning assistance or reduced the perpetrator’s risk of detection—must be understood in that light. In a scenario such as this, in which the perpetrator removes the victims from the scene of the robbery and restrains them after having forcibly taken their prop- erty, we agree with the respondent that the fourth and fifth Salamon factors are the ones that speak most directly to the intent of the perpetrator. The goal of a robbery is to take possession of another’s property. Once that property has been taken by force, the purpose of leading the victims to a different, more isolated loca- tion and requiring that they remain there for some period of time is, undoubtedly, to facilitate the offend- er’s escape from the premises, undetected and unob- structed. The Kansas Supreme Court put the point most succinctly: ‘‘The forced direction of a store clerk to cross the store to open a cash register is not a kidnap- ping; locking him in a cooler to facilitate escape is.’’ State v. Buggs, supra, 219 Kan. 216. Indeed, sister state courts have found it apparent that restraining robbery victims after having taken their property bespeaks an independent criminal intention to facilitate escape with- out detection or apprehension. See, e.g., Miles v. State, 839 So. 2d 814, 820 (Fla. App. 2003). Nothing in the record of the present case suggests that a different result is warranted. Silk provided the only direct evidence regarding the petitioner’s rationale for forcing his victims into the bathrooms after robbing them. In response to the question ‘‘what was the pur- October 26, 2021 CONNECTICUT LAW JOURNAL Page 47

339 Conn. 1 OCTOBER, 2021 45 Banks v. Commissioner of Correction

pose in your going into the bathroom,’’ Silk answered: ‘‘I would say just so he could get away.’’8 The petitioner has not been able to articulate any other plausible ratio- nale for his postrobbery treatment of the victims other than to facilitate his escape, and none is apparent.9 In cases involving strikingly similar fact patterns, sister state courts have had no difficulty concluding that the purpose for secreting and restraining robbery victims after taking their property is to facilitate the perpetra- tor’s escape. See, e.g., Ferguson v. State, 533 So. 2d 763, 764 (Fla. 1988) (forced confinement of restaurant employees in restroom after robbery was ‘‘intended to make it more difficult for the victims to identify the perpetrator and immediately call for help’’); Richardson v. State, 875 So. 2d 673, 678 (Fla. App. 2004) (confining convenience store clerk in back freezer room after tak- ing money from cash register ‘‘facilitated the robbers’ flight from the crime scene . . . and substantially les- sened the risk of quick detection’’); State v. Buggs, supra, 219 Kan. 215–16 (restraint or asportation that substantially lessens risk of detection is sufficient to constitute kidnapping). We conclude, then, that the fourth and fifth Salamon factors also tip in the respon- dent’s favor. 2 Turning to the sixth Salamon factor, we also agree with the respondent that shepherding a victim at gun- point into a back room of a retail establishment after having robbed her invariably ‘‘create[s] a significant 8 There also was testimony that the petitioner originally sought to force his victims into the basement, which would have further isolated them, until he was informed that the stores had no basement. 9 Indeed, in his brief to this court, the petitioner essentially acknowledges that the victims were forced into the bathroom to facilitate his escape, stating that ‘‘it is conceivable that jurors would view the fact that [the] petitioner moved the employees into the bathrooms so that he could escape as being part and parcel of the robberies.’’ (Emphasis added.) Page 48 CONNECTICUT LAW JOURNAL October 26, 2021

46 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

danger or increase[s] the victim’s risk of harm indepen- dent of that posed by the separate offense.’’ State v. Salamon, supra, 287 Conn. 548. In Salamon, we sug- gested that the distinct danger that is relevant to the question of whether criminal conduct bears indepen- dent significance as kidnapping need not be physical danger. See id., 536. Criminal conduct that inspires dis- tinct fears or has a uniquely harmful psychological impact on the victim also qualifies. See id. (‘‘[a]mong the evils that both the common law and later statutory prohibitions against kidnapping sought to address were the isolation of a victim from the protections of society and the law and the special fear . . . inherent in such isolation’’). Other courts and commentators have reached the same conclusion. See, e.g., People v. Ngu- yen, 22 Cal. 4th 872, 886, 997 P.2d 493, 95 Cal. Rptr. 2d 178 (2000) (holding that conduct that substantially increases risk of psychological trauma to victim is legiti- mate basis for finding separate offense); 2 A.L.I., Model Penal Code and Commentaries (1980) § 212.1, p. 222 (discussing unique evil of kidnapping as means of ter- rorizing victim).

Isolating and restraining victims at gunpoint after having robbed them causes them to experience fears that are different both in degree and in kind from the fears that naturally accompany being robbed. This is especially true in a highly visible commercial setting. When a retail store or restaurant is robbed, the victim reasonably may expect that the perpetrator will release her and flee the premises as soon as he has taken the property, bringing the danger to an end. When a perpe- trator opts to restrain a victim after a commercial rob- bery rather than leaving the premises, however, the victim justifiably fears that her ordeal may be just begin- ning. See, e.g., Latimore v. Barnes, Docket No. C11-5527 (SBA), 2015 WL 1406904, *2 (N.D. Cal. March 27, 2015). October 26, 2021 CONNECTICUT LAW JOURNAL Page 49

339 Conn. 1 OCTOBER, 2021 47 Banks v. Commissioner of Correction

The case books are filled with instances in which a robber restrains, isolates, and/or moves his victims after taking their money as a prelude to committing addi- tional, more dangerous crimes. The robbery victim who is led at gunpoint away from the visibility of a commer- cial storefront understands that the offender may be isolating and restraining her not merely to facilitate his escape but as a prelude to a physical assault,10 sexual assault,11 use of the victim as a hostage or human shield,12 or even murder.13 The prospect is undeniably terrifying. See, e.g., Commonwealth v. Hughes, 264 Pa. Super. 118, 126, 399 A.2d 694 (1979); N. Kanellis, Note, ‘‘Kidnapping in Iowa: Movements Incidental to Sexual Abuse,’’ 67 Iowa L. Rev. 773, 782 (1982). It is undisputed that the victims in the present case did in fact experience such heightened and distinct psychological harm as a result of their postrobbery asportation and restraint. One of the victims, Feltman, testified that she feared for her life during the kidnap- ping. Another, Kozlowski, testified that he ducked down while imprisoned in the bathroom for fear that the peti- tioner would shoot him through the door. Defense coun- sel conceded as much in his closing argument, acknowledging that ‘‘all four victims had a very hard experience. . . . I will concede for you that . . . Wright and . . . Feltman had probably the most diffi- cult . . . harshest, the most traumatic experiences of their [lives] . . . .’’ 10 See, e.g., People v. Shay, 60 App. Div. 2d 698, 698, 400 N.Y.S.2d 383 (1977). 11 See, e.g., Lovette v. State, 636 So. 2d 1304, 1305–1306 (Fla. 1994); Mills v. State, 236 Ga. 365, 365, 223 S.E.2d 725 (1976); State v. Coleman, 865 S.W.2d 455, 456 (Tenn. 1993). 12 See, e.g., State v. Vue, Docket No. C4-92-86, 1992 WL 153093, *2 (Minn. App. July 7, 1992); People v. Addison, 151 App. Div. 2d 372, 372, 543 N.Y.S.2d 74, appeal denied, 74 N.Y.2d 946, 549 N.E.2d 483, 550 N.Y.S.2d 281 (1989), and appeal denied, 74 N.Y.2d 946, 549 N.E.2d 483, 550 N.Y.S.2d 281 (1989). 13 See, e.g., Lovette v. State, 636 So. 2d 1304, 1305–1306 (Fla. 1994); State v. Robinson, Docket No. 01C01-9207-CR-00234, 1993 WL 273953, *3 (Tenn. Crim. App. July 22, 1993). Page 50 CONNECTICUT LAW JOURNAL October 26, 2021

48 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Moreover, the harms and dangers involved in a post- robbery kidnapping are not solely psychological. Even if the perpetrator does not plan to assault, imperil, or kill his victims following a robbery, spiriting them at gunpoint to a more isolated location necessarily increases the risk that they will suffer serious physical injury or death. See, e.g., Eaglehorse v. State, 286 N.W.2d 329, 331 (S.D. 1979); see also Note, ‘‘A Rationale of the Law of Kidnapping,’’ 53 Colum. L. Rev. 540, 547–48 (1953). As the Supreme Court of California has explained, ‘‘[i]t takes but little imagination to envision the kind of violent events whose likelihood of occur- rence is great [when a kidnapping victim is forced to travel a great distance under the threat of injury by a deadly weapon]. Ready examples include not only desperate attempts by the victim to extricate himself but also unforeseen intervention by third parties.’’ Peo- ple v. Lara, 12 Cal. 3d 903, 908 n.4, 528 P.2d 365, 117 Cal. Rptr. 549 (1974); see, e.g., Latimore v. Barnes, supra, 2015 WL 1406904, *4 (in assessing whether restraint constitutes independent crime of ‘‘kidnapping to commit robbery,’’ jury considers ‘‘whether the move- ment increased a victim’s risk of harm . . . [including] the danger inherent in a victim’s foreseeable attempts to escape’’ (internal quotation marks omitted)); N. Kanellis, supra, 67 Iowa L. Rev. 782 (‘‘[t]aking a victim from familiar surroundings to an unknown location . . . may aggravate the victim’s [freedom seeking] impulses, causing the victim to attempt an escape and thus possibly incur more serious bodily harm’’ (footnote omitted)). Once again, one need not look far to find cases in which a victim, fearing that kidnapping would be a precursor to rape, assault, or murder, panicked or tried to resist, bringing about a tragic, self-fulfilling prophesy. See, e.g., People v. Laursen 8 Cal. 3d 192, 196, 501 P.2d 1145, 104 Cal. Rptr. 425 (1972) (motorist was shot when resisting kidnapping during escape from October 26, 2021 CONNECTICUT LAW JOURNAL Page 51

339 Conn. 1 OCTOBER, 2021 49 Banks v. Commissioner of Correction

robbery), appeal dismissed and cert. denied, 412 U.S. 915, 93 S. Ct. 2738, 37 L. Ed. 2d 142 (1973); Johnson v. State, 281 Ga. App. 7, 8, 635 S.E.2d 278 (2006) (when robbery victim resisted being locked in camper, defen- dant struck him in head with revolver, causing him to fade in and out of consciousness). 3 When we examine the six Salamon factors, then, five of the factors most often will resolve in favor of the respondent when the perpetrator of a robbery forces his victims to leave the relative safety of a highly visible commercial store front and sequesters them in a more isolated area after having robbed them. The asportation and confinement take place after the completion of the primary objective of the robbery. The restraint is not inherent in the nature of the robbery, insofar as many, if not most, robberies culminate with the offender simply fleeing with the fruits of the robbery; leading a victim to a back room at gunpoint is no more linked to the crime of robbery than to rape, assault, or any other crime. Such removal and restraint typically are for the purpose of, and have the effect of, making it more difficult for the victim to summon assistance and reduc- ing the offender’s risk of detection, if not a prelude to the commission of a distinct crime, such as a sexual assault. Finally, the conduct necessarily subjects the victim to unique risks and harms, both physical and psychological, beyond those inherent in the robbery itself. Under such circumstances, these five factors almost invariably will outweigh the first factor—the nature and duration of the movement and confine- ment—which is more subjective and fact based, and takes center stage in close cases such as Salamon.14 14 We note that our conclusion that removing and restraining a victim after a robbery to facilitate the perpetrator’s escape usually holds independent criminal significance does not mean that there could not be close cases in which the failure to submit the question to a jury would constitute prejudicial error. For example, we would hesitate to find harmless a trial court’s failure Page 52 CONNECTICUT LAW JOURNAL October 26, 2021

50 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

D Finally, we take this opportunity to clarify some of the language in Salamon and Hinds that has been a source of confusion among litigants and the lower courts, the present case included. Specifically, in Sala- mon we stated that ‘‘[o]ur legislature, in replacing a single, broadly worded kidnapping provision with a gra- dated scheme that distinguishes kidnappings from unlawful restraints by the presence of an intent to pre- vent a victim’s liberation, intended to exclude from the scope of the more serious crime of kidnapping and its accompanying severe penalties those confinements or movements of a victim that are merely incidental to and necessary for the commission of another crime against that victim. Stated otherwise, to commit a kid- napping in conjunction with another crime, a defendant must intend to prevent the victim’s liberation for a longer period of time or to a greater degree than that which is necessary to commit the other crime.’’ (Emphasis added.) State v. Salamon, supra, 287 Conn. 542. Subsequently, in Hinds, in response to the dis- senting justices, the majority explained that, ‘‘by focus- ing solely on whether there was any restraint or asportation beyond that necessary for the commission to give the jury a Salamon instruction in a case in which the alleged postrob- bery conduct involved no asportation and only minimal restraint, such as a parting order that the victim lie on the ground or ‘‘don’t be a hero.’’ See, e.g., Hill v. State, 642 So. 2d 796, 797–98 (Fla. App. 1994) (jury could find restraint was incidental to robbery when perpetrators merely forced victims to lie down on floor before escaping); State v. White, 362 S.W.3d 559, 562–63 (Tenn. 2012) (whether confinement was incident to robbery was question for jury when defendant, while leaving scene of robbery, removed telephones and directed victims to lie down on floor and wait eight or nine minutes). It is debatable, for example, whether the victims are placed at increased risk of harm under such circumstances. The present case, however, is funda- mentally different from those cases. The petitioner took each victim at gunpoint to a different, more isolated part of the store, where he ordered them to remain confined in an enclosed space and attempted to block their escape. October 26, 2021 CONNECTICUT LAW JOURNAL Page 53

339 Conn. 1 OCTOBER, 2021 51 Banks v. Commissioner of Correction

of the sexual assault, the dissenting justices ignore[d] the ‘incidental to’ language in Salamon. . . . Restraint may be incidental to a sexual assault that is not neces- sary for its commission.’’ (Citations omitted; emphasis omitted.) Hinds v. Commissioner of Correction, supra, 321 Conn. 89–90. We did not invent the highlighted language in Sala- mon. Substantially similar articulations of the ‘‘inciden- tal and necessary’’ test appeared in the decisions of many of our sister states prior to Salamon. See, e.g., People v. Daniels, 71 Cal. 2d 1119, 1139, 459 P.2d 225, 80 Cal. Rptr. 897 (1969); Cejvanovic v. State, Docket No. 05-1226, 2006 WL 3614067, *2 (Iowa App. December 13, 2006) (decision without published opinion, 728 N.W.2d 223); State v. Robbins, 272 Kan. 158, 175, 32 P.3d 171 (2001); State v. Rogers, 17 Ohio St. 3d 174, 181, 478 N.E.2d 984, vacated on other grounds, 474 U.S. 1002, 106 S. Ct. 518, 88 L. Ed. 2d 452 (1985); State v. Atkin, 135 P.3d 894, 898–99 (Utah App.), cert. denied, 150 P.3d 58 (Utah 2006); State v. Harris, Docket No. 55561-8, 2006 WL 2246194, *2 (Wn. App. August 7, 2006) (decision without published opinion, 134 Wn. App. 1029), review denied, 160 Wn. 2d 1015, 161 P.3d 1027 (2007). Nevertheless, that language, both on its own and in conjunction with the cited language in Hinds, has led to confusion about, first, what it means for force or restraint to be necessary to the commission of a rape, robbery, or assault, and, second, whether the incidental and necessary test is to be understood as conjunctive or disjunctive. That is, does Salamon apply only when an offender’s conduct is not only incidental to but also necessary to commit the underlying crime, or is a defen- dant entitled to Salamon’s protections if either prong of the test applies (if, for example, restraint of the victim was incidental to a sexual assault but was not necessary to accomplish the assault). Both questions are front and center in the present case, in which the parties Page 54 CONNECTICUT LAW JOURNAL October 26, 2021

52 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

disagree over the proper application of the incidental and necessary test. With respect to the meaning of the word ‘‘necessary,’’ greater clarity may be achieved and ambiguities resolved by emphasizing three points often made by sister state courts. First, although restraint is not strictly necessary to accomplish a rape, robbery, or assault— restraint is not an essential element of those crimes— some degree of restraint, whether by physical force or threat and fear, almost always accompanies their commission. See Frederick v. State, 931 So. 2d 967, 970 (Fla. App. 2006); State v. White, 362 S.W.3d 559, 568 (Tenn. 2012). We have used the word ‘‘necessary’’ in that spirit. Second, our statement that kidnapping requires pre- vention of the victim’s liberation for a longer period of time or to a greater degree than that which is necessary to commit the underlying crime does not mean that an offender must commit the underlying crime in the least intrusive manner lest he be subjected to criminal liabil- ity for kidnapping. Hinds itself is an excellent illustra- tion of this point. We suggested in that case that if, for example, a jury concluded that the petitioner decided to move his victim the short distance from the parking lot where he abducted her to a nearby grassy area not to avoid detection or identification but, rather, because ‘‘he could not perform in the lit space [of the parking lot], or simply to avoid the hard paved surface while kneeling on the ground [as he assaulted her]’’; Hinds v. Commissioner of Correction, supra, 321 Conn. 80; then there was no indication that our legislature intended to criminalize his conduct as kidnapping merely because the act might have been completed less intrusively. See id., 79–80. Third, one complication of thinking about necessity in this manner is that doing so forces a reviewing court October 26, 2021 CONNECTICUT LAW JOURNAL Page 55

339 Conn. 1 OCTOBER, 2021 53 Banks v. Commissioner of Correction

to confront thorny, angels on the head of a pin questions about the degree of restraint employed in any particular crime. Could an offender have accomplished a rape, robbery, or assault using less restraint? Would an under- lying crime committed less intrusively have been the same crime? To avoid having to confront these sorts of prickly metaphysical and counterfactual questions, sister state courts have framed the issue in terms of whether the restraint involved in a crime reflects a different criminal intention or creates a different risk of harm than that necessarily present in the underlying crime.15 In other words, the jury looks not to whether each of the offend- er’s specific physical actions was strictly necessary to commit the underlying crime but, instead, to whether his actions demonstrated an animus or purpose beyond that necessarily involved in the underlying crime (e.g., 15 See, e.g., People v. Daniels, supra, 71 Cal. 2d 1139 (‘‘the intent of the [California] [l]egislature . . . was to exclude from [the] reach [of the felony kidnapping statute] not only standstill robberies . . . but also those in which the movements of the victim are merely incidental to the commission of the robbery and do not substantially increase the risk of harm over and above that necessarily present in the crime of robbery itself’’ (citation omitted; emphasis added; internal quotation marks omitted)); State v. Hart- ley, Docket No. 90AP-859, 1991 WL 132417, *6 (Ohio App. July 18, 1991) (because ‘‘the kidnapping by deception in order to lure [the victim] to a place where he could be raped involved a separate animus from the restraint necessary to effectuate the rape . . . the trial court properly sentenced the [defendant] to consecutive sentences for rape and kidnapping’’); State v. Rollins, 605 S.W.2d 828, 830 (Tenn. Crim. App. 1980) (‘‘a number of courts have held that the forced movement of a crime victim during the course of a crime [that] is merely incidental to the perpetration of the crime, as from one room of a house to another, is not kidnapping if there is no substantially increased risk of harm over and above that necessarily present in the crime of robbery itself’’ (emphasis added)); State v. Williams, Docket No. 36772- 2-I, 1997 WL 469623, *3 (Wn. App. August 18, 1997) (relevant issue to consider in determining if confinement, movement or detention is sufficient to warrant separate kidnapping conviction is ‘‘whether the defendant’s conduct substan- tially increased [the] risk of harm over and above that necessarily present in the [underlying] crime . . . itself’’ (internal quotation marks omitted)), review denied, 134 Wn. 2d 1026, 958 P.2d 315 (1998). Page 56 CONNECTICUT LAW JOURNAL October 26, 2021

54 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

to wrongfully obtain property from, sexually violate, or physically injure the victim) or whether they imposed harms or risks on the victim beyond those that necessar- ily accompany any rape, robbery, or assault. Understanding Salamon in this way also resolves the second question that divided the parties and the members of the Appellate Court in this case and the companion case, namely, whether conduct that is merely incidental to but goes beyond that necessary to commit the underlying crime satisfies the legislative definition of kidnapping. The cited language from Sala- mon is ambiguous and, arguably, consistent with either party’s interpretation. Certainly, in Hinds, we suggested that the reading favored by the petitioner is the correct one; conduct that is wholly incidental to the commis- sion of an underlying crime cannot qualify as kidnap- ping, regardless of whether it is strictly necessary to commit that crime. See id., 89–90. Understanding neces- sity in the manner that we have suggested is consistent with that approach. It also is consistent with the approach followed by many of our sister state courts, on whose decisions we relied in Salamon. See State v. Salamon, supra, 287 Conn. 518, 527 and n.16. Those courts speak of restraints or restrictions on a victim’s liberty as being merely incidental to the underlying crime, necessary to commit the underlying crime, inherent in the nature of the underlying crime, and having no independent criminal significance, often using those and related expressions more or less interchangeably. See, e.g., United States v. Santistevan, 25 M.J. 123, 126 (C.M.A. 1987); Mackerley v. State, 754 So. 2d 132, 137 (Fla. App. 2000), quashed on other grounds, 777 So. 2d 969 (Fla. 2001); State v. Martin, 222 N.C. App. 213, 221, 729 S.E.2d 717, review denied, 366 N.C. 413, 735 S.E.2d 187 (2012), and review dismissed, 372 N.C. 300, 826 S.E.2d 710 (2019); see also F. Wozniak, Annot., ‘‘Seizure or Deten- October 26, 2021 CONNECTICUT LAW JOURNAL Page 57

339 Conn. 1 OCTOBER, 2021 55 Banks v. Commissioner of Correction

tion for Purpose of Committing Rape, Robbery, or Other Offense as Constituting Separate Crime of Kidnapping,’’ 39 A.L.R.5th 283, 357, § 2[a] (1996) (describing tests used by different jurisdictions as being more or less formulaic ‘‘but not substantially different’’). Our own cases have, at times, followed the same approach. See, e.g., State v. Fields, supra, 302 Conn. 247, 252 (citing Salamon and conflating independent criminal signifi- cance, conduct that is not merely incidental to underly- ing offense, and restraint beyond that necessary to commit underlying offense). These various terms, then, are merely different ways of expressing the same concept, namely, whether the restraint imposed evidenced an independent criminal intent or subjected the victims to risks distinct from those necessarily entailed by or inherent in the underly- ing offenses. The six Salamon factors offer a useful framework for answering those questions. IV We therefore conclude that the petitioner cannot prevail on his claim under Salamon because the respon- dent has demonstrated, in light of the undisputed facts presented at trial, that the absence of the instruction mandated by Salamon gave rise to no discernible risk of prejudice, let alone a risk sufficient to undermine confidence in the verdict, such that a properly instructed jury would have found the petitioner guilty beyond a reasonable doubt. Indeed, as we explained, we agree with our sister state courts, which have con- cluded, as a matter of law, that, when a perpetrator, having taken his victims’ valuables, then leads them at gunpoint away from a highly visible commercial store- front and confines them in an isolated area of the store while he makes his escape, thereby exposing them to new and different risks, such conduct is not inherent Page 58 CONNECTICUT LAW JOURNAL October 26, 2021

56 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

in the nature of the robbery but, rather, indisputably has independent criminal significance.16 The judgment of the Appellate Court is reversed and the case is remanded to that court with direction to render judgment affirming the judgment of the habeas court. In this opinion ROBINSON, C. J., and MULLINS and KAHN, Js., concurred.

D’AURIA, J., with whom McDONALD, J., joins, con- curring. Because I agree with the majority that, in the present case, the lack of a jury instruction pursuant to State v. Salamon, 287 Conn. 509, 949 A.2d 1092 (2008), was harmless, I concur in the result. I do not agree, however, with the standard that the majority adopts for determining whether any error was harmless under these circumstances. The majority determines that, when a petitioner seeking habeas relief establishes a Salamon error, the habeas court must assess the harm of that error according to the legal standard that the United States Supreme Court articulated in Brecht v. Abrahamson, 507 U.S. 619, 623, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993) (new trial mandated if instructional error ‘‘had [a] substantial and injurious effect or influ- ence in determining the jury’s verdict’’ (internal quota- tion marks omitted)), rather than the more petitioner friendly standard adopted in Neder v. United States, 527 U.S. 1, 18, 119 S. Ct. 1827, 144 L. Ed. 2d 35 (1999) (new trial required if it is ‘‘clear beyond a reasonable doubt that a rational jury would have found the defen- dant guilty absent the [instructional] error’’). I disagree 16 Consequently, it is apparent, as Judge Keller concluded; see Banks v. Commissioner of Correction, supra, 184 Conn. App. 133 (Keller, J., dis- senting); that the petitioner could not prevail on his claim, even if we agreed with him that the respondent was required to establish harmlessness beyond a reasonable doubt. The concurring justices do not disagree with this con- clusion. October 26, 2021 CONNECTICUT LAW JOURNAL Page 59

339 Conn. 1 OCTOBER, 2021 57 Banks v. Commissioner of Correction

with the majority opinion for two reasons. First, because the petitioner, Mark Banks, cannot prevail under either standard, I do not believe that this court needs to—or should—decide which standard applies, especially as it is unclear how many, if any, future cases this standard will apply to. Second, I believe that Neder is the proper standard. Accordingly, I respectfully concur. I In Hinds v. Commissioner of Correction, 321 Conn. 56, 136 A.3d 596 (2016), we concluded that, because the petitioner would prevail under either the Brecht or Neder standards for determining the harmlessness of a Salamon violation on collateral review, we did not need to ‘‘enter the fray’’ and decide whether to adopt the more state friendly Brecht standard ‘‘and the uncer- tainties that accompany it.’’ Id., 83. In the present case, I agree with the majority that the petitioner would not prevail under either standard,1 and, in my view, the majority’s decision to apply the Brecht standard is dic- tum. I therefore do not believe that this late in the day for Salamon claims there is any greater justification to ‘‘enter the fray’’ than there was five years ago in Hinds. In fact, I believe there is less. It has been more than one decade since this court released its decision in Salamon and then held that the new rule in Salamon applies retroactively to collateral attacks on final judgments in Luurtsema v. Commis- 1 I agree fully with then Judge Keller’s well reasoned dissenting opinion in the Appellate Court that the absence of a Salamon instruction was harmless beyond a reasonable doubt under the Neder standard. Banks v. Commis- sioner of Correction, 184 Conn. App. 101, 132–33, 194 A.3d 780 (2018) (Keller, J., dissenting). The majority likewise agrees with Judge Keller that ‘‘the petitioner could not prevail on his claim even, if we agreed with him that the respondent [the Commissioner of Correction] was required to establish harmlessness beyond a reasonable doubt.’’ Footnote 16 of the majority opinion. Page 60 CONNECTICUT LAW JOURNAL October 26, 2021

58 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

sioner of Correction, 299 Conn. 740, 769, 12 A.3d 817 (2011) (plurality opinion). In Luurtsema,2 in holding that Salamon would apply retroactively to those liti- gants whose direct appeals already had been decided at the time of Salamon, we explained that, ‘‘[o]f the 1.5 percent of . . . inmates incarcerated for kidnapping or unlawful restraint, one can reasonably assume that only a small subset will fall within the ambit of Salamon.’’ Id. Although the parties have not provided any data in the present case on the number of potential Salamon claims that remain for cases that already have become final, there is a real possibility that the new standard that the majority adopts might never be applied to another case.3 I am not suggesting that, when this 2 I note that Luurtsema was a plurality decision. Nevertheless, the plurali- ty’s analysis in that case belies the majority’s contention that this case supports the adoption of the Brecht standard. Additionally, the court in Hinds, which was a majority decision, relied heavily on Luurtsema, treating it as controlling law, and, thus, I do the same. See Hinds v. Commissioner of Correction, supra, 321 Conn. 61 (‘‘we conclude that [the] retroactivity decision [in Luurtsema v. Commissioner of Correction, supra, 299 Conn. 740] compels the conclusion that challenges to kidnapping instructions in criminal proceedings rendered final before Salamon are not subject to the procedural default rule’’ (emphasis added)). 3 In Epps v. Commissioner of Correction, 327 Conn. 482, 175 A.3d 558 (2018), the petitioner’s counsel represented in the petitioner’s brief: ‘‘There are no records of cases by type of claim kept by the court or public defender, but a repeated canvass of appointed counsel turned up only three cases with potential to present Luurtsema-Salamon issues for decision by the habeas court. Counsel also have resolved by agreement a number of Salamon habeas cases without necessitating trial of the case.’’ Epps v. Commissioner of Correction, Conn. Supreme Court Briefs & Appendices, November Term, 2017, Petitioner’s Brief p. 25 n.23. Counsel for the respondent, the Commis- sioner of Correction, in Epps did not contradict this representation. Our own databases show no present cases pending in our court or the Appellate Court other than Britton v. Commissioner of Correction, 185 Conn. App. 388, 197 A.3d 895 (2018) (Salamon violation was harmless beyond reasonable doubt under Neder), petition for cert. filed (Conn. November 26, 2018) (No. 180266). The parties have provided this court with no data regarding how many cases our decision in this case will affect. With perhaps the exception of Britton, the present case and the companion case, Bell v. Commissioner of Correction, 339 Conn. 79, A.3d (2021), which we also decided today, might very well be the only cases in which this new standard will October 26, 2021 CONNECTICUT LAW JOURNAL Page 61

339 Conn. 1 OCTOBER, 2021 59 Banks v. Commissioner of Correction

court’s holding likely will not affect many cases, we should not resolve the issue. Rather, I believe that, when addressing the issue is not necessary to resolve the case, we should refrain from altering the applicable standard at the eleventh hour. Thus, as we are not required to decide this issue to adjudicate the petition- er’s claims in the present case or the claims asserted in the companion case, Bell v. Commissioner of Correc- tion, 339 Conn. 79, A.3d (2021), also decided today, I do not believe we should. II The majority having determined to reach the issue, I also disagree with its resolution of the issue. The majority properly notes that only twice have we had occasion to apply Salamon in the habeas context and contends that, in these two cases, we did not envision that such claims would be evaluated under the more petitioner friendly Neder standard. I disagree. I believe this court’s retroactive application of the Salamon rule to both pending and final cases has strongly suggested that the Neder standard applies to all Salamon viola- tions. Certainly, that has been the understanding of our Appellate Court, our habeas courts, and the respondent, the Commissioner of Correction. Further, unlike the majority, I also find there to be countervailing policy concerns that militate in favor of continuing to apply the Neder standard for claims such as those of the peti- tioner. A In Salamon, this court overruled our long-standing interpretation of this state’s kidnapping statutes and be applied. Moreover, as I discuss in greater detail in part II of this opinion, both the Appellate Court and habeas courts consistently have relied on Luurtsema and Hinds as holding that the Neder standard is the proper standard, thereby not only belying the majority’s contention that these cases ‘‘strongly [suggest]’’ that the Brecht standard is more appropriate, but also showing that all but a very few remaining Salamon claims raised on collateral review will be subject to the Brecht standard. Page 62 CONNECTICUT LAW JOURNAL October 26, 2021

60 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

held that, when a criminal defendant is charged with kidnapping in conjunction with another underlying crime, such as rape or assault, the trial court must instruct the jury that it cannot find the defendant guilty of kidnapping if the restraint imposed on the victim was merely incidental to that underlying crime. State v. Salamon, supra, 287 Conn. 550. In so holding, we relied heavily on ‘‘the common law of kidnapping, the history and circumstances surrounding the promulga- tion of our current kidnapping statutes and the policy objectives animating those statutes . . . .’’ Id., 542. A prominent concern at the time of the enactment of our current kidnapping statutes was prosecutorial over- charging: ‘‘Beginning in the 1950s . . . questions sur- faced about the propriety of [our] expansively worded kidnapping statutes. In particular, concerns were expressed that the newly adopted kidnapping statutes permitted the imposition of extremely severe sanctions for a broad and ill defined range of behavior, including relatively trivial types of restraint. . . . [E]xamples of abusive prosecution for kidnapping [were] common, [so the legislature sought] to restrict the scope of kid- napping, as an alternative or cumulative treatment of behavior whose chief significance is robbery or rape, because the broad scope of this overlapping offense has given rise to serious injustice . . . .’’ (Citations omitted; internal quotation marks omitted.) Id., 538–39. We noted in Salamon that, in drafting our current kid- napping statute, the legislature ‘‘intended to create a new statutory scheme that recognized varying degrees of unlawful restrictions on a victim’s liberty by drawing a distinction between a ‘restraint,’ which, standing alone, comprises the crime of unlawful restraint, and an ‘abduction,’ which comprises the crime of kidnapping,’’ thereby intending ‘‘to exclude from the scope of the more serious crime of kidnapping and its accompanying severe penalties those confinements or movements of October 26, 2021 CONNECTICUT LAW JOURNAL Page 63

339 Conn. 1 OCTOBER, 2021 61 Banks v. Commissioner of Correction

a victim that are merely incidental to and necessary for the commission of another crime against that victim.’’ Id., 541–42. Nevertheless, in interpreting our kidnapping statutes prior to Salamon, this court failed to recognize this exclusion. Id., 543. ‘‘Unfortunately, that interpreta- tion has afforded prosecutors virtually unbridled discre- tion to charge the same conduct either as a kidnapping or as an unlawful restraint despite the significant differ- ences in the penalties that attach to those offenses. Similarly, our prior construction of the kidnapping stat- utes has permitted prosecutors—indeed, it has encour- aged them—to include a kidnapping charge in any case involving a sexual assault or robbery.’’ Id., 543–44. Subsequently, but with limited analysis, this court in State v. Sanseverino, 287 Conn. 608, 949 A.2d 1156 (2008) (overruled in part on other grounds by State v. DeJesus, 288 Conn. 418, 953 A.2d 45 (2008)),4 super- seded in part after reconsideration en banc, 291 Conn. 574, 969 A.2d 710 (2009), applied the new rule in Sala- mon retroactively to direct appeals pending at the time Salamon was decided. State v. Sanseverino, supra, 287 Conn. 620 n.11. Not until approximately three years after Salamon did this court address its applicability to collateral attacks on final judgments, although we did not have to address the proper standard for evaluating harm because the retroactivity issue came to this court by way of reserved question. In Luurtsema v. Commissioner of Correction, supra, 299 Conn. 740, we agreed with the petitioner that, ‘‘as a matter of state common law, Salamon should be afforded fully retroactive effect,’’ including as to collat- eral attacks on final judgments by way of habeas peti- tions. Id., 751 (plurality opinion). We explained that, 4 In DeJesus, this court overruled Sanseverino to the extent that Sansever- ino held that a judgment of acquittal, rather than a new trial, could serve as a proper remedy for a Salamon violation. Page 64 CONNECTICUT LAW JOURNAL October 26, 2021

62 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

‘‘[a]s a matter of federal constitutional law, each juris- diction is free to decide whether, and under what cir- cumstances, it will afford habeas petitioners the retroactive benefit of new judicial interpretations of the substantive criminal law issued after their convictions became final.’’ Id., 754 (plurality opinion). We recog- nized that, ‘‘in the federal system, the United States Supreme Court has adopted a per se rule that, when federal courts reinterpret congressional legislation, new interpretations of substantive criminal statutes must be applied retroactively on collateral review.’’ Id., 754–55 (plurality opinion). Our own determination of retroactivity in Luurtsema focused on the purpose of habeas relief: ‘‘The principal purpose of the writ of habeas corpus is to serve as a bulwark against convictions that violate fundamental fairness. . . . To mount a successful collateral attack on his conviction, a prisoner must demonstrate . . . a fundamental unfairness or miscarriage of justice . . . .’’ (Internal quotation marks omitted.) Id., 758 (plu- rality opinion). We reasoned that, ‘‘regardless of whether one reads Salamon to be a change or clarifica- tion of the law, the court in Salamon saw itself as discerning the original legislative meaning of [General Statutes] § 53a-92 (a) (2) (A). . . . If the legislature never intended an assault to constitute kidnapping, without evidence of the perpetrator’s independent intent to restrain the victim, then the petitioner . . . stands convicted of a crime that he arguably did not commit. This conclusion raises serious due process concerns. It is well settled that due process requires the state to prove every element of the offense charged beyond a reasonable doubt. . . . Under our system of justice, considerations of finality simply cannot justify the continued incarceration of someone who did not commit the crime of which he stands convicted.’’ (Cita- tions omitted; internal quotation marks omitted.) Id., 758–59 (plurality opinion). Thus, although we declined October 26, 2021 CONNECTICUT LAW JOURNAL Page 65

339 Conn. 1 OCTOBER, 2021 63 Banks v. Commissioner of Correction

the petitioner’s invitation to adopt a per se rule in favor of full retroactivity, we concluded that, ‘‘when an appel- late court provides a new interpretation of a substantive criminal statute, an inmate convicted under a prior, more expansive reading of the statute presumptively will be entitled to the benefit of the new interpretation on collateral attack.’’ Id., 760 (plurality opinion). This court in Luurtsema proceeded at some length to reject the ‘‘five rationales either for adopting a per se rule against retroactive relief or for denying relief in [that] case: (1) the fact that law enforcement relied on the old interpretation of the kidnapping statutes while trying the petitioner; (2) the fact that the retroac- tive application of Salamon has no deterrent value or remedial purpose; (3) the fear that our courts will be ‘flooded’ with habeas petitions from other inmates con- victed [of kidnapping]; (4) the difficulty of retrying such cases where significant time has elapsed since convic- tion; and, perhaps most importantly (5) the concern that victims will be retraumatized by again having to testify and endure another round of judicial proceed- ings.’’ Id., 765 (plurality opinion). In response to these arguments, we explained that ‘‘many of the concerns raised by the state in the habeas context apply with equal force to direct appeals, in which it is undisputed that appellants receive the benefit of retroactive appli- cation of judicial decisions that narrow the scope of liability under a criminal statute.’’ Id., 766 (plurality opinion).5 Unlike the majority, I believe that this court’s 5 We pointed out in Luurtsema that State v. Sanseverino, supra, 287 Conn. 620 n.11, provided an instructive case in point: ‘‘The crimes charged in [Sanseverino] commenced in June or July of 1998 . . . a mere two to three months after the incident for which the petitioner in [Luurtsema] was convicted. Whereas the petitioner’s conviction became final in 2003, how- ever, Sanseverino was still under review when we decided Salamon in 2008. Any concerns regarding prosecutorial reliance and the burdens associated with retrying a ten year old crime apply to Sanseverino no less than to [Luurtsema].’’ (Citation omitted.) Luurtsema v. Commissioner of Correc- tion, supra, 299 Conn. 766–67 (plurality opinion). Also, more specifically as to the respondent’s third rationale regarding the opening of the floodgates, this court acknowledged that ‘‘[t]here is little Page 66 CONNECTICUT LAW JOURNAL October 26, 2021

64 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

rejection of the state’s arguments in Luurtsema sup- ports the Neder standard, not the Brecht standard. First, although the Neder standard was not at issue in Luurtsema, we signaled that it applies on collateral review. It is not so, as the majority suggests, that ‘‘we did not envision that such claims would be evaluated under the stringent Neder standard’’ because ‘‘we thought that many Salamon cases could be disposed of summarily on the ground of harmless error . . . .’’ In Luurtsema, we emphasized that the state’s concerns over applying Salamon retroactively to cases on collat- eral review applied equally to our earlier determination to apply Salamon retroactively to cases pending on direct appeal when we decided Sanseverino. See Luurt- sema v. Commissioner of Correction, supra, 299 Conn. 769–70 (plurality opinion). We went on to indicate that error arising from the lack of a Salamon instruction in cases brought before the court on collateral review would be reviewed for harm; see id. (plurality opinion); citing to State v. Hampton, 293 Conn. 435, 463–64, 978 A.2d 1089 (2009), which had applied the Neder harmless error standard. doubt that some petitioners will come forward contending that they are serving substantially longer sentences than are prescribed by the [Penal] [C]ode, as properly construed. In [his] brief, however, the [respondent] has identified only five such petitions that have been filed in the more than two years since we decided Salamon and Sanseverino. At oral argument before this court, the [respondent] declined to provide additional information as to the number of present inmates who might have a colorable claim under Salamon. Of the 1.5 percent of . . . inmates incarcerated for kidnapping or unlawful restraint, one can reasonably assume that only a small subset will fall within the ambit of Salamon. Of those, we expect that courts will be able to dispose summarily of many cases where it is sufficiently clear from the evidence presented at trial that the petitioner was guilty of kidnapping, as properly defined, that any error arising from a failure to instruct the jury in accordance with the rule in Salamon was harmless. See, e.g., State v. Hampton, 293 Conn. 435, 463–64, 978 A.2d 1089 (2009). Likewise, we doubt the [respondent] will expend the resources to retry cases [when] it is reason- ably clear that a petitioner could not have been convicted of kidnapping under the correct interpretation of the statute.’’ (Footnote omitted.) Luurt- sema v. Commissioner of Correction, supra, 299 Conn. 769–70 (plurality opinion). October 26, 2021 CONNECTICUT LAW JOURNAL Page 67

339 Conn. 1 OCTOBER, 2021 65 Banks v. Commissioner of Correction

Subsequently, in Hinds v. Commissioner of Correc- tion, supra, 321 Conn. 56, this court conceded that, by including this citation to Hampton, we had ‘‘indicated that the proper standard to make such an assessment would be the harmless error standard applied on direct appeal,’’ i.e., Neder. See id., 77. Thus, to the extent that this court in Luurtsema ‘‘strongly suggested’’ anything about the proper harmless error standard to apply on collateral review, it has strongly suggested the Neder standard. In Hinds, this court was asked to address how the retroactivity of Salamon on collateral review interacted with our procedural default rule.6 Id., 60. We concluded that ‘‘[the] retroactivity decision [in Luurtsema v. Com- missioner of Correction, supra, 299 Conn. 740] compels the conclusion that challenges to kidnapping instruc- tions in criminal proceedings rendered final before Salamon are not subject to the procedural default rule.’’ Hinds v. Commissioner of Correction, supra, 321 Conn. 61. In reaching this conclusion, this court examined its reasoning in Luurtsema for applying the rule in Salamon retroactively and determined that this reason- ing was inconsistent with the procedural default rule: ‘‘[A]pplication of the procedural default bar to protect finality of judgments seems inconsistent with the rea- soning in [Luurtsema] that the interests of finality must give way to the demands of liberty and a proper respect for the intent of the legislative branch.’’ (Internal quota- tion marks omitted.) Id., 73. 6 ‘‘[W]e have adopted the procedural default standard prescribed in Wain- wright v. Sykes, [433 U.S. 72, 87, 97 S. Ct. 2497, 53 L. Ed. 2d 594 (1977)]. . . . Under this standard, the petitioner must demonstrate good cause for his failure to raise a claim at trial or on direct appeal and actual prejudice resulting from the impropriety claimed in the habeas petition. . . . [T]he cause and prejudice test is designed to prevent full review of issues in habeas corpus proceedings that counsel did not raise at trial or on appeal for reasons of tactics, inadvertence or ignorance . . . .’’ (Citations omitted; internal quotation marks omitted.) Hinds v. Commissioner of Correction, supra, 321 Conn. 71. Page 68 CONNECTICUT LAW JOURNAL October 26, 2021

66 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Particularly significant to the issue at hand, we specif- ically noted in Hinds that, in Luurtsema, we did not apply any heightened standard, such as the procedural default doctrine’s heightened prejudice standard, to keep any floodgates from swinging open.7 Id., 74–76. Rather, we recognized that, in Luurtsema, we cited the Neder harmless error standard for direct appeal. See id., 75. Thus, this court already has rejected the applica- tion of a heightened prejudice standard to Salamon claims on collateral review. In Hinds, this court then went on to discuss the applicable standard for determin- ing harm. Again, we recognized that, ‘‘[i]n [Luurtsema], the court indicated that the proper standard to make such an assessment would be the harmless error stan- dard applied on direct appeal [by citing to Hampton].’’ (Emphasis added.) Id., 77.

Only after deciding that ‘‘the petitioner [in Hinds was] entitled to relief under our established harmless error standard’’; id., 81; did we ‘‘note that this court has not had the occasion to consider whether, even in the absence of procedural default, a more stringent stan- dard of harm should apply in collateral proceedings,’’ 7 Specifically, in reaching our conclusion in Hinds, we explained: ‘‘Other aspects of the court’s reasoning [in Luurtsema also] bolster our conclusion that this holding was not intended to afford relief to only those petitioners who could avoid or overcome the procedural default bar. The court in [Luurtsema] extensively considered limitations on its retroactivity ruling, but did not cite procedural default as such a limitation. Availability of that doctrine and its heightened prejudice standard would have been a natural response to the [respondent’s] floodgates argument had the court intended the doctrine to apply. Instead, the court responded [by explaining that] . . . one can reasonably assume that only a small subset will fall within the ambit of Salamon. . . . One particular aspect of this response is telling. The court cited the harmless error standard for direct appeal—a standard wholly inconsistent with the actual prejudice standard for procedurally defaulted claims—as the limiting mechanism for colorable but ultimately nonmeritorious claims.’’ (Citations omitted; footnotes omitted; internal quo- tation marks omitted.) Hinds v. Commissioner of Correction, supra, 321 Conn. 74–75. October 26, 2021 CONNECTICUT LAW JOURNAL Page 69

339 Conn. 1 OCTOBER, 2021 67 Banks v. Commissioner of Correction

such as the Brecht standard.8 Id. Nevertheless, because the dissenting justices’ conclusion in Hinds that the petitioner was ‘‘not entitled to a new trial due to his failure to establish the actual prejudice to overcome a procedurally defaulted claim appear[ed] to signal a retreat from our holdings in Salamon and [Luurtsema], we . . . explain[ed] why the petitioner [in Hinds] would prevail even under the more stringent standard’’ the dissenting justices applied. Id., 83. Although we decided Luurtsema on a reservation for advice and a stipulation of facts, and our language in Hinds regarding Brecht might arguably be considered dictum,9 we noted in Hinds that Neder was ‘‘the stan- dard that was applied by the habeas court in [that] case and ha[d] been applied in several other cases. See, e.g., Eric M. v. Commissioner of Correction, 153 Conn. App. 837, 845, 108 A.3d 1128 (2014), cert. denied, 315 Conn. 915, 106 A.3d 308 (2015); St. John v. Warden, Superior Court, judicial district of Tolland, Docket No. CV-11- 4003987-S (March 7, 2013); see also Epps v. Commis- sioner of Correction, 153 Conn. App. 729, 738, 740, 104 A.3d 760 (2014) (determining that petitioner must overcome procedural default but applying direct appeal 8 ‘‘In Brecht v. Abrahamson, supra, 507 U.S. 623, a bare majority of the United States Supreme Court departed from its history of more than 200 years of parity between direct appeals and habeas corpus proceedings for constitutional claims. . . . Citing federalism, comity, finality and other pru- dential considerations, the court determined that habeas proceedings require a standard that imposes a less stringent burden on the state when the constitutional error is not structural. . . . Brecht and its progeny have raised numerous questions as to the precise standard to be applied in determining whether a particular type of error is harmless, and what degree of certainty as to whether that standard has been met.’’ (Citations omitted.) Hinds v. Commissioner of Correction, supra, 321 Conn. 81–82. 9 It is at best arguable that Hinds’ use of the Neder standard was dictum. In fact, because nothing in the majority’s opinion in Hinds suggests that Brecht is the proper standard, and because the majority in Hinds explained that it analyzed the petitioner’s claim in that case under the Brecht standard only to respond to the arguments of the dissenting justices, it is more accurate to say that any discussion of the Brecht standard was dictum. Page 70 CONNECTICUT LAW JOURNAL October 26, 2021

68 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

harmless error standard in prejudice analysis); Nogueira v. Warden, Superior Court, judicial district of Tolland, Docket No. CV-XX-XXXXXXX-S (June 10, 2015) (same); Smith v. Warden, Superior Court, judicial district of Tolland, Docket No. CV-XX-XXXXXXX-S (September 13, 2011) (same).’’ Hinds v. Commissioner of Correction, supra, 321 Conn. 77. Hinds does not indicate that a contrary standard might apply. Although limited, this court’s prior case law applying the rule in Salamon retroactively on collateral review suggests that the Neder standard is the proper standard for determining harm, although the parties did not explicitly raise, and this court did not explicitly address, this issue. The majority’s contention that our prior case law ‘‘strongly suggests’’ the adoption of the Brecht standard is belied by the plain language of these cases and sug- gests that these cases have been misunderstood by both our habeas courts and the Appellate Court, which, con- sistent with the signals this court gave in Hinds, have uniformly applied the Neder standard to Salamon viola- tions raised in a collateral proceeding. See Palmer v. Commissioner of Correction, 202 Conn. App. 902, 242 A.3d 1084 (2021) (per curiam affirmance of judgment of habeas court, which applied Neder harmless error standard to Salamon claim); John B. v. Commissioner of Correction, 194 Conn. App. 767, 774, 222 A.3d 984 (2019) (applying Neder harmless error standard to Sala- mon claim raised in habeas appeal, citing to Hinds as support), cert. denied, 334 Conn. 919, 222 A.3d 513 (2020); Britton v. Commissioner of Correction, 185 Conn. App. 388, 400, 197 A.3d 895 (2018) (same), peti- tion for cert. filed (Conn. November 26, 2018) (No. 180266); Pereira v. Commissioner of Correction, 176 Conn. App. 762, 768, 171 A.3d 105 (same), cert. denied, 327 Conn. 984, 175 A.3d 43 (2017); White v. Commis- sioner of Correction, 170 Conn. App. 415, 427, 154 A.3d 1054 (2017) (same); Nogueira v. Commissioner of Cor- rection, 168 Conn. App. 803, 814, 149 A.3d 983 (same), October 26, 2021 CONNECTICUT LAW JOURNAL Page 71

339 Conn. 1 OCTOBER, 2021 69 Banks v. Commissioner of Correction

cert. denied, 323 Conn. 949, 169 A.3d 792 (2016); Farmer v. Commissioner of Correction, 165 Conn. App. 455, 460, 139 A.3d 767 (same), cert. denied, 323 Conn. 905, 150 A.3d 685 (2016); Coltherst v. Commissioner of Cor- rection, Docket No. CV-XX-XXXXXXX-S, 2019 WL 7425147, *10–12 (Conn. Super. November 19, 2019) (applying Neder standard and finding lack of Salamon instruction harmless). I do not mean to suggest that this court should not correct errors by lower courts or not clarify ambiguous case law. But I do not believe that this is what the majority is doing in its analysis. Rather, the majority has decided for policy reasons to reverse course and to adopt a standard inconsistent with our prior case law. In fact, even as late as 2014, in Epps v. Commissioner of Correction, 153 Conn. App. 729, 738, 740, 104 A.3d 760 (2014), appeal dismissed, 327 Conn. 482, 175 A.3d 558 (2018), the respondent was arguing in habeas cases involving the failure to give a Salamon instruction that the Neder standard for harmless error applied, a stan- dard he did not prevail on in the Appellate Court in Epps. Only before this court in Epps did the respondent begin to argue that the Brecht standard should govern.10 10 After this court released its decision in Hinds, in Epps v. Commissioner of Correction, 327 Conn. 482, 175 A.3d 558 (2018), we permitted the respon- dent, whose petition for certification to appeal was pending at the time, ‘‘to file an amended petition for certification. Over the petitioner’s objection, this court granted the respondent’s amended petition, which raised the question ‘left unresolved’ by Hinds regarding the proper measurement of harm in collateral proceedings like the . . . one [in Epps] and the question of whether, irrespective of which standard applied, harm had been estab- lished in the petitioner’s criminal case.’’ Id., 484. This court, however, subse- quently dismissed the appeal on the ground that certification was improvidently granted because ‘‘[t]he respondent had squarely argued to the habeas court that the petition should be assessed under the harmless beyond a reasonable doubt standard. The respondent never argued in the alternative that a higher standard of harmfulness should apply to collateral proceedings even if the petitioner’s claim was not subject to procedural default, despite federal case law applying a higher standard since 1993. Accordingly, we conclude[d] that this [was] not the proper case in which to fairly address this consequential issue . . . .’’ Id., 485. Page 72 CONNECTICUT LAW JOURNAL October 26, 2021

70 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Since the Appellate Court decided Epps, the respondent has relied on the Neder standard in multiple cases before the Appellate Court in which a Salamon violation was raised, not even mentioning Brecht in appellate briefs. See, e.g., Nogueira v. Commissioner of Correc- tion, supra, 168 Conn. App. 814; Farmer v. Commis- sioner of Correction, supra, 165 Conn. App. 460. From my research, it appears that the first habeas case in which the respondent challenged the Neder standard before the Appellate Court was White v. Commissioner of Correction, supra, 170 Conn. App. 427, which was not decided until 2017. Thus, not only the courts, but also the respondent has interpreted Hinds as applying, if not adopting, the Neder standard, which is reasonable given our analysis in that case. B Having made clear in Luurtsema that the retroactiv- ity of new judicial interpretations of substantive crimi- nal law on collateral review is a ‘‘matter of state common law,’’ and that we are not bound by federal law, the determination of the appropriate standard to apply for analyzing harm is likewise a matter of state common law and a policy question for this court to determine. I do not read the majority opinion to suggest otherwise. See, e.g., Schilberg Integrated Metals Corp. v. Continental Casualty Co., 263 Conn. 245, 255, 819 A.2d 773 (2003) (determination of proper burden of proof based on certain policy considerations); Albert Mendel & Son, Inc. v. Krogh, 4 Conn. App. 117, 124, 492 A.2d 536 (1985) (‘‘The proper allocation of the burden of proof may be distilled to a question of policy and fairness based on experience in different situations. . . . A number of variables are considered in determin- ing where the burden properly lies. One consideration is which party has readier access to knowledge about the fact in question.’’ (Citations omitted; footnote omit- October 26, 2021 CONNECTICUT LAW JOURNAL Page 73

339 Conn. 1 OCTOBER, 2021 71 Banks v. Commissioner of Correction

ted.)). I would come to a different policy determination than the majority does. Brecht concerned a federal court’s collateral review of a state court criminal judgment pursuant to 28 U.S.C. § 2254. See Brecht v. Abrahamson, supra, 507 U.S. 626. The court in Brecht and in later cases emphasized that the choice to apply a more government friendly harm- lessness standard was driven in large part by concerns unique to federal habeas review of state court judg- ments—namely, that the application of a petitioner friendly standard as in Neder would (1) invade state sovereignty over criminal matters (i.e., federalism and comity), (2) undercut the historic limitation of habeas relief to those ‘‘ ‘grievously wronged,’ ’’ (3) infringe on a state’s interest in finality, and (4) impose significant societal costs. Id., 633–38; see also Fry v. Pliler, 551 U.S. 112, 117, 127 S. Ct. 2321, 168 L. Ed. 2d 16 (2007). Notwithstanding significant differences between fed- eral habeas review of state criminal judgments and the situation we confront upon a finding of a Salamon violation, the majority’s own rationale in opting for the Brecht standard is still driven by these policies in signifi- cant part. This is where I disagree most pointedly with the court’s opinion. First,11 unlike the majority, I see no need for this standard to be ‘‘consistent with how we handle other 11 Obviously, federalism and comity do not inform our determination of the proper standard to apply. Thus, I am not persuaded by the fact that federal courts have adopted the Brecht standard as the proper harmlessness standard for collateral review of a state conviction for constitutional errors or by the fact that a number of sister state courts have followed suit. As mentioned previously, we are not bound by federal law in determining the retroactivity of new judicial interpretations of substantive criminal law on collateral review, which is a matter of state law. Also, as the majority concedes, many jurisdictions, if not the majority of jurisdictions, apply the Neder standard on collateral as well as on direct review of state convictions for constitutional errors. See, e.g., In re Martinez, 3 Cal. 5th 1216, 1225, 407 P.3d 1, 226 Cal. Rptr. 3d 315 (2017); Guam v. Ojeda, Docket No. CRA10- 011, 2011 WL 6937376, *13 (Guam December 23, 2011); Hill v. State, 615 N.W.2d 135, 140–41 (N.D. 2000). Page 74 CONNECTICUT LAW JOURNAL October 26, 2021

72 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

claims of error in habeas actions.’’ A Salamon violation is unlike many other claims of error in habeas actions: it is a determination that the state, in prosecuting a defendant, was unconstitutionally relieved of proving an essential element of the crime of kidnapping. In both Salamon and Luurtsema, this court indicated that a reinterpretation of the language of our kidnapping stat- ute was long overdue and that the legislature never intended to criminalize conduct that this court had erro- neously interpreted the statute to capture, i.e., a restraint of a victim merely incidental to the commis- sion of a separate crime. For nearly one-half century, interpreting the statute broadly had led to prosecutors overcharging and juries finding defendants guilty on allegations the legislature never intended to punish. A Salamon claim is traditionally the kind of claim that should have been raised on direct appeal and is not permitted to be raised in a habeas action unless the petitioner can satisfy the procedural default rule. See Jackson v. Commissioner of Correction, 227 Conn. 124, 131–32, 629 A.2d 413 (1993). Similar claims about the proper interpretation of our kidnapping statutes had been made and rejected several times in the thirty Moreover, the present case, involving state court collateral review of a state conviction, is more akin to federal court collateral review of a federal conviction for constitutional error, than to federal court collateral review of a state conviction for constitutional error. There is currently a split among the federal courts of appeals regarding whether the Brecht standard applies to federal habeas review of federal court convictions, thereby suggesting that Brecht is not clearly the proper standard to apply in this case. See James v. United States, 217 Fed. Appx. 431, 435 (6th Cir. 2007) (‘‘State court convictions are examined on collateral review to determine whether an error ‘had substantial and injurious effect or influence in determining the jury’s verdict.’ . . . This standard does not apply to collateral review of [federal court] convictions, as [state court] convictions are entitled to special deference.’’ (Citations omitted.)). But see United States v. Smith, 723 F.3d 510, 516 (4th Cir. 2013) (noting that United States Supreme Court has not decided this issue and applying Brecht standard because ‘‘society has the same interest in the finality of federal convictions as it does in state convic- tions’’), cert. denied, 572 U.S. 1043, 134 S. Ct. 1774, 188 L. Ed. 2d 610 (2014). October 26, 2021 CONNECTICUT LAW JOURNAL Page 75

339 Conn. 1 OCTOBER, 2021 73 Banks v. Commissioner of Correction

years preceding Salamon. See State v. Salamon, supra, 287 Conn. 531 (‘‘Since 1977, we have had numerous opportunities to examine the scope of the kidnapping statutes, generally in response to a claim that the crime of kidnapping was not intended to apply to a restraint that was merely incidental to the commission of another crime. See, e.g., State v. Luurtsema, [262 Conn. 179, 200, 811 A.2d 223 (2002)]; State v. Wilcox, [254 Conn. 441, 465–66, 758 A.2d 824 (2000)]; State v. Amarillo, [198 Conn. 285, 304–306, 503 A.2d 146 (1986)]; State v. Vass, 191 Conn. 604, 614, 469 A.2d 767 (1983); State v. Johnson, 185 Conn. 163, 177–78, 440 A.2d 858 (1981), aff’d, 460 U.S. 73, 103 S. Ct. 969, 74 L. Ed. 2d 823 (1983); State v. Briggs, 179 Conn. 328, 338–39, 426 A.2d 298 (1979), cert. denied, 447 U.S. 912, 100 S. Ct. 3000, 64 L. Ed. 2d 862 (1980); State v. DeWitt, 177 Conn. 637, 640– 41, 419 A.2d 861 (1979); State v. Lee, [177 Conn. 335, 342–43, 417 A.2d 354 (1979)]; State v. Chetcuti, [173 Conn. 165, 170, 377 A.2d 263 (1977)].’’). Bringing such an argument to the court was not for the fainthearted— as it bordered on being frivolous—and at least could be criticized as ‘‘depreciat[ing]’’ other claims. (Internal quotation marks omitted.) State v. Pelletier, 209 Conn. 564, 567, 552 A.2d 805 (1989). Until one day, this court agreed. For reasons of fairness, as discussed previously, this court in Luurtsema and Hinds held that Salamon claims raised for the first time in a habeas action are unique—the byproduct of the retroactivity exception to the general rule against these kinds of claims being raised on collateral review, along with the procedural default rule being forgiven. For this reason, I do not see why there is any need for us to apply a standard consistent with that employed in other habeas cases. Rather, if this court is trying to create consistency, the same standard should be used for Salamon claims both on direct review and collateral review, as there is no real Page 76 CONNECTICUT LAW JOURNAL October 26, 2021

74 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

distinction between the claims other than the temporal relation of the case to our decision in Salamon, which is merely a matter of good or bad fortune. Further, this court, our Appellate Court, and habeas courts consistently have applied the Neder standard to Salamon claims on collateral review. See Hinds v. Commissioner of Correction, supra, 321 Conn. 77; see also id. (citing cases). So, not only would the petitioner in this case be afforded less protection than defendants whose cases were pending or were initiated after Sala- mon, but he will be afforded less protection than other habeas petitioners who have raised Salamon claims on collateral review in the years since Luurtsema and Hinds but prior to our decision in the present case. I am not suggesting that these lower court cases bind our analysis. I am suggesting that these cases demon- strate the unfairness of adopting the Brecht standard at this late date, a clearly appropriate policy consider- ation in determining whether this court should alter the applicable standard. The majority defends this arbitrariness by rationaliz- ing that ‘‘the somewhat scattershot nature of harmless error jurisprudence, with standards varying by the type of error at issue, the stage of review, and the jurisdiction in which the claim is reviewed, means that whichever standard we apply to Salamon errors in state habeas cases may appear to be unfair or incongruous from one vantage or another.’’ Choosing the Brecht standard over the Neder standard seems to me to be little more than an attempt to counterbalance this court’s determination to apply Salamon retroactively to collateral challenges and to forgive the procedural default rule by giving the respondent a break on the harmlessness standard. The majority reasons: ‘‘It certainly will not seem unjust from the respondent’s standpoint to require a showing that there is some reasonable likelihood that the failure to give the jury a Salamon instruction had a substantial October 26, 2021 CONNECTICUT LAW JOURNAL Page 77

339 Conn. 1 OCTOBER, 2021 75 Banks v. Commissioner of Correction

and injurious effect or influence on the outcome before requiring the state to retry the petitioner for crimes that were committed more than twenty-five years ago.’’ In my view, this reasoning misses the mark and loos- ens Luurtsema and Hinds from their jurisprudential moorings. The twenty-five years since the date of the petitioner’s trial should of course play no role in this analysis but merely provide color. To the extent that the majority’s rationale reflects a concern about finality, this court has already indicated that that interest ‘‘must give way to the demands of liberty and a proper respect for the intent of the legislative branch.’’ (Internal quota- tion marks omitted.) Hinds v. Commissioner of Correc- tion, supra, 321 Conn. 73; see also Luurtsema v. Commissioner of Correction, supra, 299 Conn. 759 (plu- rality opinion) (‘‘considerations of finality simply can- not justify the continued incarceration of someone who did not commit the crime of which he stands con- victed’’). It is not the petitioner’s fault that this court awakened late to the fact that the legislature never intended for the present language of our kidnapping statute to extend as broadly as it had been construed, more than one decade after his criminal trial and eight years after his direct appeal became final. Although the majority is correct that it is not the state’s fault, either, that this court reinterpreted our kidnapping statutes when it did, the state surely benefited more from a jurisprudential regime that relieved it of having to prove an essential element than did those defendants accused of the kidnappings.12 12 It makes sense for the respondent to bear the heavier burden of proving harmlessness on collateral review, such as in the present case, as the state not only was relieved of having to bear its burden of proving an element at the underlying criminal trial, but it was more likely to establish, even pre-Salamon, the underlying facts regarding the kidnapping, such as the timing and length of the restraint. In other words, although the state may have asked additional questions of the witnesses in light of Salamon, it was more likely to have put forth all of its admissible evidence regarding the crime, including the length of the restraint and when the restraint occurred Page 78 CONNECTICUT LAW JOURNAL October 26, 2021

76 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

In holding that Salamon applies retroactively to cases that are already final, this court recognized that older cases may need to be retried, accepting this as a neces- sary consequence of correcting injustices that had occurred as a result of prosecutorial overcharging and this court’s failure to construe our statutes properly so as to curb that practice. See Luurtsema v. Commis- sioner of Correction, supra, 299 Conn. 766–67, 772–73 (plurality opinion). The fact that the state must now deal with these consequences should not come at the expense of the petitioner, especially when limiting the petitioner’s protection out of fairness to the state would contradict the purpose of our holdings in Luurtsema and Hinds. Moreover, this consequence is not unique to collateral review, as this court noted in Luurtsema when it recognized that cases pending on direct appeal at the time of Salamon might also require retrial years after the initial criminal trial. Id., 766–67 (plurality opinion). There is perhaps no better way to summarize my concerns regarding the fairness of the court’s rationale in the present case than that of the sage author of the majority opinion (whose collegiality and guidance I already miss): ‘‘[T]o apply Brecht in the present case would be unfair to the petitioner and to others similarly situated. That this court opted to revisit and revise our interpretation of the state’s kidnapping laws following his conviction is no more the fault of the petitioner than of the state. Although would-be offenders were on notice in 1995 that they could be charged with kidnap- ping solely on the basis of the restraint inherent in robberies or assaults, they, like the state, did not have in relation to other conduct, than the defendant was to have defended on grounds relevant to Salamon. The state typically attempts to provide the fact finder with a complete picture of the crime, whereas the defendant would have had no motivation, pre-Salamon, to ask questions and to offer evidence regarding the length and timing of the restraint that were unneces- sary pre-Salamon. October 26, 2021 CONNECTICUT LAW JOURNAL Page 79

339 Conn. 1 OCTOBER, 2021 77 Banks v. Commissioner of Correction

any reason to try their cases with the Salamon distinc- tion in mind. Moreover, it may seem discrepant to assess the impact of the instructional error according to the more forgiving Brecht standard when, if we had decided Salamon one decade earlier, while the petitioner’s direct appeal was pending, the state would have borne the burden of proving that the error was harmless beyond a reasonable doubt.’’ III Which brings me back to my original point: Why are we deciding this issue at this late date? By declining to ‘‘enter the fray’’ and answer this question; Hinds v. Commissioner of Correction, supra, 321 Conn. 83; we would not, in my view, be ‘‘shrink[ing] from our duty’’ any more than this court already has to date in declining to answer this unique question in Hinds. Footnote 6 of the majority opinion. This court, the Appellate Court, and habeas courts, consistent with the signals this court gave in Hinds, have been applying the Neder standard to Salamon claims raised on collateral review, and both petitioners and the respondent have managed to prevail under that standard. The parties have provided us with no data regarding how many potential cases remain to which the Brecht standard would apply, especially as it appears that some cases have been resolved by agree- ment. See footnote 3 of this opinion. The number cannot be infinite. Most evidence points to the fact that it is likely a very small number, and obviously shrinking. It is of course true that a lower court cannot bind this court and that courts of last resort, such as the United States Supreme Court, may allow issues to ‘‘percolate’’ before deciding them. But it must be rare for a court to allow an issue to percolate to the point of virtual extinction and to decide it only after giving strong sig- nals as to the correct outcome, and after the lower courts have uniformly decided the issue that way. Page 80 CONNECTICUT LAW JOURNAL October 26, 2021

78 OCTOBER, 2021 339 Conn. 1 Banks v. Commissioner of Correction

Given that the outcome of this appeal would be the same regardless of the standard applied, and that the majority’s holding will possibly have little to no effect going forward, I believe it is imprudent to address and decide this issue. In my view, if there are in fact any cases left in this shrinking universe to which today’s announcement will apply, it is unfortunate that the court is now changing its mind about the standard that will govern harmlessness. It is also ironic given that the fallout from Salamon itself has concerned the appli- cation of retroactivity rules fairly to a change of this court’s mind. Policy considerations support this view. The peti- tioner in the present case and any petitioners going forward, assuming there are any, will receive less pro- tection than those who happened to have their habeas petitions decided before today’s decision. This appears to be a matter of pure chance and perpetuates the unfairness that Salamon and its progeny have attempted to correct. Every petitioner to have brought a collateral Salamon challenge has had appointed coun- sel. We have no understanding about how the public defender’s office prioritizes inmate cases or whether and to what extent inmates might understand that they had a Salamon claim. Finally, I am concerned that, by needlessly ‘‘enter[ing] the fray’’; Hinds v. Commis- sioner of Correction, supra, 321 Conn. 83; regarding which standard to apply, this court might not realize how its decision to adopt the Brecht standard in the context of a Salamon claim might have possible ramifi- cations in contexts we have yet to anticipate. Accordingly, I respectfully concur.

ECKER, J., concurring in the judgment. I respectfully disagree with the majority opinion to the extent that it adopts and applies the harmless error analysis set forth in Brecht v. Abrahamson, 507 U.S. 619, 623, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993). See id. (new trial is mandated if instructional error ‘‘had substantial and injurious effect or influence in determining the jury’s verdict’’ (internal quotation marks omitted)). For the reasons explained in part II of Justice D’Auria’s concur- ring opinion in this case, I would instead apply the standard articulated in Neder v. United States, 527 U.S. 1, 18, 119 S. Ct. 1827, 144 L. Ed. 2d 35 (1999). See id. (new trial is required unless it is ‘‘clear beyond a reasonable doubt that a rational jury would have found the defendant guilty absent the [instructional] error’’). Accordingly, I join part II of Justice D’Auria’s concur- ring opinion. I nevertheless concur in the judgment because, applying the Neder standard, the state has met its burden of establishing harmless error on this record. See footnote 1 of Justice D’Auria’s concurring opinion.