Plaintiff Harold Scott brought suit under 42 U.S.C. § 1983, claiming (chiefly) deprivation of due process in a prison disciplinary hearing. He appeals from a judgment of the United States District Court for the Southern District of New York (Ward, /.), granting summary judgment to defendants and dismissing the complaint on the ground that Scott’s disciplinary sentence of 60 days in a special housing unit (“SHU”) did not implicate a liberty interest under Sandin v. Conner, 515 U.S. 472, 115 S.Ct. 2293, 132 L.Ed.2d 418 (1995), and that Scott therefore could not assert a section 1983 claim for denial of due process in the hearing that resulted in that sentence.
On appeal, Scott claims that the district court erred in determining whether a liberty interest was implicated under Sandin. Specifically, Scott argues that the district court: (i) improperly focused on the actual punishment imposed on and served by Scott, rather than the potential punishment that could have been imposed in the disciplinary hearing; (ii) relied on New York State regulations that were promulgated after the disciplinary hearing at issue; and (iii) failed to accord Scott an opportunity to conduct discovery on the conditions and duration of disciplinary confinement as compared to other types of confinement. We agree with the district court that in performing the Sandin analysis, a court should consider the actual penalty imposed in the disciplinary hearing rather than the potential penalty. However, because the district court’s analysis depended in part on New York State regulations, and because the court used the current regulations rather than the 1987 regulations, we vacate and remand for further proceedings.
BACKGROUND
Scott is serving a life sentence at Green Haven prison. On November 7,1986, Officer Novak searched his cell during a block-wide frisk. That evening, Scott was served with a misbehavior report, in which Novak alleged that he had found a six-inch sharpened shank in one of the legs of Scott’s bed. Scott denied the charge and after a hearing, was *285sentenced to 45 days keeplock confinement, 30 of which were deferred.
On January 22,1987, Scott was ordered to report to the Beekman Town Justice Court (which is apparently attached to the prison) for arraignment on criminal charges arising from the alleged possession of the shank. Scott refused to enter the courtroom voluntarily and the officers eventually had to force him to enter using handcuffs.
On January 28, 1987, Correctional Captain Joseph Tanner presided over a “Tier III” hearing (the “Tanner hearing”),1 at which Scott was charged with disobeying a direct order, interference and verbal harassment. Scott requested four witnesses for the hearing: Richard Schisler, the lawyer assigned to represent him at the January 22, 1987 arraignment; the Beekman town judge who presided at the arraignment; Officer Gordon LaBonte, who wrote the misbehavior report regarding the January 22,1987 incident; and Officer Stephen Albury, who gave Scott the direct order to enter the courtroom.
At the start of the hearing, Officer La-Bonte testified that Scott had violated a direct order from Officer Albury to enter the courtroom. The hearing was then adjourned until Albury was available to testify. When the hearing reconvened on February 2, 1987, Tanner denied Scott’s request to hear testimony from the judge and Schisler, ruling that these witnesses were irrelevant because neither had “witnessed] the misbehavior or had knowledge of state rules.” Albury then testified that Scott disobeyed his direct order to enter the courtroom, after which Albury handcuffed Scott and handed him over to an Officer Dunleavy. Scott then requested that Officer Dunleavy be called to testify. Tanner denied this request, noting that anything that Dunleavy might have observed occurred after the relevant conduct, i.e., Scott’s refusal to obey a direct order.
Tanner found Scott guilty of disobeying a direct order and of interference, and sen-fenced him to 60 days in keeplock confinement, and loss of privileges. Tanner also recommended a loss of 60 days of “good time” credit, a sanction that had no effect on Scott’s life sentence. Scott was then confined to Green Haven’s A-2 SHU from February 2 to March 22.
Scott appealed Tanner’s ruling to Thomas Coughlin, the Commissioner of the Department of Correctional Services. Coordinator of Prison Discipline Donald Selsky, exercising authority delegated by Coughlin, denied the appeal on April 13, 1987. Scott’s subsequent Article 78 proceeding was dismissed for failure to state a claim: the state court ruled that Scott was obliged to obey even an improper order by a correction officer. Scott’s eventual trial before a jury in the Beekman court on the weapon possession charges ended in a mistrial. The State moved to dismiss the charges, and the Beek-man court did so on January 18,1989.
On January 22,1990, Scott filed this action, asserting ten causes of action. However, five of these claims were dismissed on motions to dismiss, and we have dismissed Scott’s appeal of the dismissal of these claims pursuant to 28 U.S.C. § 1915(e) as lacking an arguable basis in law or fact.
Scott asserted five claims pursuant to 42 U.S.C. § 1983 that are relevant to this appeal: three relate to the denial of due process at the Tanner hearing; one pleads retaliation for Scott’s assertion of his due process rights; and one alleges violation of his due process at the November 13, 1986 hearing and a conspiracy to falsely charge and prosecute Scott.
The district court dismissed these remaining claims on October 23, 1996, on the ground that Scott failed to demonstrate that he had been deprived of a liberty interest under Sandin v. Conner, 515 U.S. 472, 115 S.Ct. 2293, 132 L.Ed.2d 418 (1995). In ruling that Scott’s 60-day confinement in the A-2 SHU was not an “atypical or significant *286hardship” under Sandin, the court made the following findings: (i) “Scott’s disciplinary segregation [did not] adversely affeet[] the duration of his sentence”; (ii) “Scott has not demonstrated that his confinement differed materially in either duration or degree of restriction from the confinement of those inmates assigned to the A2-SHU for non-disciplinary reasons”; and (iii) “[i]n light of the fact that Scott had been incarcerated since 1984, the time he spent in the A2-SHU represents an insignificant proportion of his time under sentence, and thus cannot be characterized as ‘a major disruption in his environment.’ ”
DISCUSSION
Scott claims principally that he was deprived of due process at the Tanner hearing. But if Scott cannot establish that he had a protected liberty interest in being free from the punishment that was imposed upon him as a result of that hearing, he has no due process claim under the Fourteenth Amendment. See Frazier v. Coughlin, 81 F.3d 313, 318 (2d Cir.1996) (per curiam,). The district court concluded that Scott had not demonstrated that he had been deprived of a liberty interest under Sandin by his 60-day confinement in the A-2 SHU. Scott asserts several errors in the district court’s analysis.
I
Scott claims that in deciding whether a liberty interest was implicated in his disciplinary hearing, the district court should have considered the maximum potential penalty he faced—ie., unlimited SHU confinement—rather than the actual punishment imposed, 60-day confinement in A-2 SHU. We disagree.
This is an issue of first impression. Prior to Sandin, the Supreme Court looked to the language of state statutes and regulations to determine if the state had created a liberty interest in being free from certain types of punishment. Under this approach, the Court found a liberty interest if the state had used “mandatory language” in the statute or regulation, thereby limiting the state official’s discretion. Sandin, 515 U.S. at 479-81, 115 S.Ct. at 2298-99.
Sandin recognized that this approach had “led to the involvement of federal courts in the day-to-day management of prisons” and interfered with the “appropriate deference and flexibility” that should be accorded to state officials in managing the prison environment. Id. at 482, 115 S.Ct. at 2299. The Court held that, in the future, liberty interests “will be generally limited to freedom from restraint which ... imposes atypical and significant hardship on the inmate in relation to the ordinary incidents of prison life.” Id. at 484, 115 S.Ct. at 2300 (emphasis added). Conner’s sentence of 30 days of disciplinary segregation in SHU did not rise to this standard because “Conner’s confinement did not exceed similar, but totally discretionary confinement in either duration or degree of restriction,” and lockdown time was significant even for the general prison population. Id. at 486, 115 S.Ct. at 2301. The Court concluded that “[biased on a comparison between inmates inside and outside disciplinary segregation,” Conner’s placement in SHU did not “work a major disruption in his environment.” Id.
In determining whether the state might have conceivably created a liberty interest in avoiding the punishment Conner received, the Sandin Court analyzed the nature and duration of the penalty that was actually imposed. Although Sandin does not state in terms that actual punishment should be considered in the liberty interest analysis (in fact, Conner’s 30-day SHU confinement represented the maximum punishment available in his disciplinary proceedings), Sandin’s reasoning suggests that the proper focus is on actual punishment. Sandin limits federal due process supervision to cases where confinement “imposes atypical and significant hardships on the inmate,” id at 484,115 S.Ct. at 2300 (emphasis added), and thereby frames a backward-looking inquiry (as to the liberty interest) that measures the hardship in terms of the plaintiffs experience. Looking to the actual punishment imposed upon the plaintiff in his disciplinary hearing, rather than the punishment that could have been imposed, or something else, is more consistent with the purpose of Sandin and the retrospective nature of its mandated analysis.
*287Several of our recent cases on Sandin implicitly endorse a look to the actual punishment imposed in the hearing, rather than the potential punishment. In enumerating the errors in the district court’s analysis of the liberty interest in Brooks v. DiFasi, 112 F.3d 46 (2d Cir.1997), we noted that the district court “did not consider the length of the confinement in determining whether it imposed an atypical and significant hardship.” Id. at 48 (emphasis added). Similarly, in Wright v. Coughlin, 132 F.3d 133 (2d Cir.1998), we listed as one of the relevant factors under Sandin “the duration of the disciplinary segregation imposed compared to discretionary confinement.” Id. at 136 (emphasis added).
Scott makes several arguments in support of the potential punishment approach. First, Scott claims that Wolff v. McDonnell, 418 U.S. 539, 94 S.Ct. 2963, 41 L.Ed.2d 935 (1974), favors his position. But Wolff, a class action brought by a representative plaintiff who had been deprived of good time credit following a disciplinary hearing, id. at 553 n. 11, 94 S.Ct. at 2973 n. 11, did not need to concern itself with the terms or duration of confinement, actual or potential. Instead, the Court’s liberty interest analysis focused on whether the state had created a right to “good time” credit which could be deprived only in specific circumstances. See id. at 557, 94 S.Ct. at 2975; see also Sandin, 515 U.S. at 478, 115 S.Ct. at 2297 (noting that Wolff’s contribution to the “landscape of prisoners’ due process” did not derive from “its description of liberty interests”).
Second, Scott claims that pr e-Sandin cases from this circuit require us to look to potential punishment. See McCann v. Coughlin, 698 F.2d 112, 121 (2d Cir.1983); McKinnon v. Patterson, 568 F.2d 930, 939 (2d Cir.1977). To the extent that we so held in those cases, Sandin is to the contrary and we follow it. In fact, those pr e-Sandin cases involved potential sentences of 14 days or less of keep-lock or SHU confinement. See McCann, 698 F.2d at 121 (“We conclude, therefore, that an inmate who is or may be sentenced to a term of confinement in a Special Housing Unit has a right to the procedural protections of the Due Process Clause.”); McKinnon, 568 F.2d at 939 (holding that two weeks of keeplock confinement was a substantial deprivation covered by due process protections). And Sandin is quite clear that exposure to such periods and quality of confinement will typically be insufficient to trigger a liberty interest. See, e.g., Frazier v. Coughlin, 81 F.3d 313, 317-18 (2d Cir.1996) (per curiam) (12 days in SHU is not a “significant deprivation of a liberty interest”).
Third, Scott draws analogies to other contexts, such as the Sixth Amendment jury trial cases, in which the Supreme Court has looked to the maximum penalty that can be imposed in a trial to determine whether a jury trial is required. See, e.g., Duncan v. Louisiana, 391 U.S. 145, 159-61, 88 S.Ct. 1444, 1453-54, 20 L.Ed.2d 491 (1968). These analogies are inapt. No right to due process is implicated in the prison context unless a liberty interest has been deprived, and we read Sandin to. require that we look to actual punishment in making this determination.
Finally, some district courts have argued that it is unworkable for states to determine whether a liberty interest has been created based on actual punishment, because states would not know in advance of the sentence whether certain procedural requirements apply in a particular ease. See, e.g., Justice v. Coughlin, 941 F.Supp. 1312, 1324 (N.D.N.Y.1996). Maybe, but one assumes that states will take the precaution of providing the required level of due process to every inmate who realistically faces a punishment that is atypical under Sandin, a precaution that would render the actual punishment rule perfectly workable.2
Accordingly, we hold that in conducting the Sandin analysis to determine whether a disciplinary sentence “imposes atypical and significant hardship on the inmate in *288relation to the ordinary incidents of prison life,” Sandin, 515 U.S. at 484, 115 S.Ct. at 2800, courts should consider the degree and duration of the sentence actually imposed in the hearing and not the maximum sentence that might have been imposed.
II
Scott claims that in determining that correction officials had extensive discretion in assigning prisoners to SHU for various nondisciplinary reasons for extended durations, the district court relied on the current state regulations, rather than those in effect at the time of the 1987 hearing. He claims that “[pjrior to 1988, Section 304 severely curtailed the correction officials’ discretion to-place inmates in restrictive confinement, and the regulations did not contain any form of generalized restrictive confinement such as the ‘administrative confinement’ introduced in 1988.”
The district court did in fact rely on the regulations current at the time of the district court decision, rather than those existing in 1987. The current regulations were adopted in large part on July 15, 1988. Because the alleged deprivation of due process occurred in 1987, the district court should have looked to the regulations as they were then. See, e.g., Wright v. Smith, 21 F.3d 496, 497 (2d Cir.1994). The 1987 regulations differ from the current version in several respects. The 1987 regulations did not contain the current catch-all provision permitting an inmate to “be admitted to a special housing unit for any other reason, with the approval of the deputy commissioner for facility operations.” N.Y. Comp.Codes R. & Regs. tit. 7, § 301.7 (1997); see N.Y. Comp.Codes R. & Regs. tit. 7, § 304.2 (1987) (“Except for automatic admission cases, no inmate shall be placed in a special housing unit other than in accordance with this section.”). Nor did the 1987 regulations provide for general “administrative segregation” as current § 301.4 permits. See N.Y. Comp.Codes R. & Regs. tit. 7, § 301.4 (1997). We by no means imply that these differences would have (or should have) altered the district court’s findings. But they bear upon an overall assessment that is for the district court to make. We therefore remand to the district court to perform the Sandin analysis in light of the conditions of prison life as they existed in 1987.3
CONCLUSION
For the reasons stated above, the judgment is vacated and remanded.
. New York conducts three types of disciplinary hearings for its inmates. Tier I hearings address the least serious infractions and have as their maximum punishment loss of privileges such as recreation. Tier II hearings address more serious infractions and may result in 30 days of confinement in an SHU. Tier III hearings concern the most serious violations and may result in unlimited SHU confinement (up to the length of the sentence) and recommended loss of "good time” credits.
. For example, the New York State regulations applicable to "Tier III" hearings appear to provide inmates with all of the incidents of due process required by the Supreme Court in Wolff. See N.Y. Comp.Codes R. & Regs. tit. 7, §§ 254.3 (advance written notice of charges); 254.5 (limited right to call witnesses); 254.6 (limited right to testify and present documentary evidence); 254.7(a)(5) (written statement of disposition of charges).
. Scott also argues that the district court improperly denied him discovery. However, he points to no discovery demand, and in fact plaintiff submitted an affidavit in which he described the specific differences between his confinement in the A-2 SHU and in the general population. Nevertheless, on remand, the district court may decide whether to rule on defendant’s summary judgment motion on the current record or to permit additional discovery.