State of New York v. U.S. Dep't of Commerce

OPINION AND ORDER In these consolidated cases, Plaintiffs bring claims under the Administrative Procedure Act (“APA”) and the Due Process Clause of the Fifth Amendment challenging the decision of Secretary of Commerce Wilbur L. Ross, Jr. to reinstate a question concerning citizenship status on the 2020 census questionnaire. See generally New York v. U.S. Dep’t of Commerce, 315 F. Supp. 3d 766 (S.D.N.Y. 2018). In an oral ruling on July 3, 2018, the Court found that Plaintiffs had made a “strong showing” of pretext or bad faith on the part of agency decision-makers and, applying well-established precedent, thus authorized discovery beyond the administrative record. (Docket No. 207 (“July 3rd Tr.”), at 76-89). Significantly, however, the Court did not rule, and has not yet ruled, on whether or to what extent any such extra-record materials can or should be considered in making a final ruling on Plaintiffs’ claims. That is largely because the parties have not yet asked the Court to do so. Defendants were given the opportunity to file a summary judgment motion arguing that the Court’s review should be limited to the administrative record and that trial was therefore unnecessary. (See Docket No. 363). But they elected not to file such a motion — thereby conceding, as a procedural matter, that a trial is appropriate. That trial is scheduled to begin in six business days, on November 5, 2018 — a date that the Court set, in no small part, because Defendants themselves insist that resolution of Plaintiffs’ claims “is a matter of some urgency” given the need to finalize the census preparations. (Docket No. 397 (“Gov’t Stay Mot.”), at 4).Remarkably, despite the foregoing, Defendants now seek a stay of the trial and related pre-trial submissions (most of which are due today and therefore presumably done already) pending resolution of a forthcoming petition to the Supreme Court for writs of mandamus and certiorari. (See id.). Even more remarkably, although they filed their motion for a stay only three nights ago and this Court made clear less than two days ago that it would issue a written ruling in short order (Oct. 24, 2018 Pretrial Conf. Tr. (“Oct. 24th Tr.”) 19), Defendants are already seeking the very same relief from the Second Circuit. (Docket No. 402). Their request is based primarily on an October 22, 2018 Order from the Supreme Court, denying Defendants’ application to stay two of this Court’s prior Orders (namely, its July 3, 2018 Order authorizing extra-record discovery, (see July 3rd Tr. 76-89) and its August 17, 2018 Order authorizing a deposition of Acting Assistant Attorney General John Gore (see Docket No. 261)) and staying, at least temporarily, a third Order (namely, the Court’s September 21, 2018 Order authorizing a deposition of Secretary Ross, see New York v. United States Dep’t of Commerce, — F. Supp. 3d — , No. 18-CV-2921 (JMF), 2018 WL 4539659 (S.D.N.Y. Sept. 21, 2018)). See In re Dep’t of Commerce, No. 18A375, 2018 WL 5259090 (U.S. Oct. 22, 2018). “Any order granting the government’s petition,” Defendants argue, “would substantially affect the further proceedings in this Court, including whether extra-record discovery would be permissible or whether review would take place on the administrative record.” (Gov’t Stay Mot. 2).In other circumstances, the Court might well agree — albeit, only as an exercise of its discretion over case management — that the Supreme Court’s Order warrants hitting the pause button and postponing trial, as the Supreme Court’s resolution of Defendants’ forthcoming petition could bear on this Court’s analysis of the merits. But Defendants’ own “urgen[t]” need for finality calls for sticking with the trial date. (Gov’t Stay Mot. 4). And, in light of the all-toofamiliar factors relevant to the question whether a stay should be granted pending mandamus, see New York v. United States Dep’t of Commerce, — F. Supp. 3d — , No. 18-CV-2921 (JMF), 2018 WL 4279467, at *1 (S.D.N.Y. Sept. 7, 2018), Defendants are certainly not entitled to a stay.A. Defendants Fail to Show the Likelihood of Irreparable HarmFirst and foremost, Defendants fall far short of establishing a “likelihood that irreparable harm will result from the denial of a stay.” Hollingsworth v. Perry, 558 U.S. 183, 190 (2010) (per curiam). Significantly, Defendants do not claim harm here from the Court’s decision to allow extra-record discovery, and for good reason: Putting aside the possible deposition of Secretary Ross, discovery will end before Defendants file their petition with the Supreme Court. (Docket No. 401).¹ Nor do they claim that, absent a stay, the argument they seek to press before the Supreme Court — that Plaintiffs’ claims should be resolved on the administrative record alone — would become moot. That too is for good reason, as Defendants remain free to argue at trial that the Court should disregard all evidence outside the administrative record and, if unsuccessful, can argue on appeal that the Court erred in considering extra-record evidence. Moreover, the Court has directed the parties to differentiate in their pre-and post-trial briefing between arguments based solely on the administrative record and arguments based on materials outside the record. (Oct. 24th Tr. 16). The Court anticipates differentiating along similar lines in any findings of fact and conclusions of law that it enters. It follows that, if the Court rules against Defendants on the basis of extra-record materials and a higher court holds that the Court should not have considered those materials, Defendants would be able to get complete relief. Put simply, a stay is not necessary “to protect” Supreme Court review. In re Dep’t of Commerce, 2018 WL 5259090 at *2 (Gorsuch, J., concurring in part and dissenting in part). The Supreme Court can conduct that review, as in the usual case, after final judgment.So what do Defendants cite as their irreparable harm in the absence of a stay? They complain that, without a stay, they “will be forced to expend enormous resources engaging in pretrial and trial activities that may ultimately prove to be unnecessary.” (Gov’t Stay Mot. 3).² But it is black-letter law that “[m]ere litigation expense, even substantial and unrecoupable cost, does not constitute irreparable injury.” Renegotiation Bd. v. Bannercraft Clothing Co., 415 U.S. 1, 24 (1974); see New York, 2018 WL 4279467, at *2 (collecting cases). Throughout the nation, litigants in federal district courts understand that, with certain well-established and narrow exceptions not applicable here, see, e.g., Gelboim v. Bank of Am. Corp., 135 S. Ct. 897, 905 n.5 (2015) (discussing the “narrow scope” of the collateral-order doctrine), everything that happens in those courts — up to and including trial — “retains its interlocutory character as simply a step along the route to final judgment,” Ortiz v. Jordan, 562 U.S. 180, 184 (2011) (citing Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546 (1949)). In other words, spending resources on trial first and seeking appellate review later is the overwhelming norm, not the exception — “even though the entry of an erroneous order may require additional expense and effort on the part of both litigants and the district court.” Parkinson v. Apr. Indus., Inc., 520 F.2d 650, 654 n.3 (2d Cir. 1975). Far from a nationwide epidemic of irreparable harm, that is precisely how the federal court system is supposed to work. See, e.g., Cunningham v. Hamilton Cty., 527 U.S. 198, 203-04 (1999) (describing the “several salutatory purposes” of the “final judgment rule”).³When pressed on that point at oral argument, Defendants asserted a new theory of harm not advanced in their written motion: some sort of dignitary harm flowing from the Court’s “scrutiny” of “an executive branch agency.” (Oct. 24th Tr. 12-14). But that novel theory of harm fails for several reasons. First, the decisions of executive branch agencies are not immune from scrutiny by the federal courts; indeed, the APA expressly invites such scrutiny. See 5 U.S.C §§702, 705; Sackett v. EPA, 566 U.S. 120, 128 (2012); see also, e.g., Mach Mining, LLC v. EEOC, 135 S. Ct. 1645, 1651 (2015) (discussing the “‘strong presumption’ favoring judicial review of administrative action” and collecting cases); United States v. Nourse, 34 U.S. (9 Pet.) 8, 28-29 (1835) (Marshall, C.J.) (“It would excite some surprise if, in a government of laws and of principle, furnished with a department whose appropriate duty it is to decide questions of right, not only between individuals, but between the government and individuals; a ministerial officer might, at his discretion, issue this powerful process…leaving to [the citizen] no remedy, no appeal to the laws of his country, if he should believe the claim to be unjust. But this anomaly does not exist….”). Second, whether these cases proceed to trial or not, there is no dispute that Defendants’ decision to add a citizenship question to the 2020 census will be subject to “scrutiny” by this Court and others; the only disputes between the parties concern the scope of evidence the Court may consider in applying that scrutiny and the degree of deference owed by the Court to Defendants’ decision.And third, although trials in APA cases are — as Defendants emphasize — “unusual” (Oct. 24th Tr. 13), they are far from unprecedented. Courts have subjected executive agencies to trials in APA cases where, as here, there are colorable claims of bad faith or pretext, see, e.g., Buffalo Cent. Terminal v. United States, 886 F. Supp. 1031, 1045-48 (W.D.N.Y. 1995), or competing expert testimony, see, e.g., Cuomo v. Baldrige, 674 F. Supp. 1089, 1090, 1093 (S.D.N.Y. 1987). In fact, it is not even unprecedented for courts to hold trials to resolve APA challenges to the administration of the census! See, e.g., City of New York v. U.S. Dep’t of Commerce, 822 F. Supp. 906, 917 (E.D.N.Y. 1993), vacated, 34 F.3d 1114 (2d Cir. 1994), rev’d sub nom. Wisconsin v. City of New York, 517 U.S. 1 (1996); Cuomo, 674 F. Supp. at 1091; Carey v. Klutznick, 508 F. Supp. 420 (S.D.N.Y. 1980), rev’d and remanded for a new trial, 653 F.2d 732 (2d Cir. 1981). Notably, Defendants cannot cite a single other instance in which the Government has sought the writ of mandamus, a form of extraordinary relief, to halt such “scrutiny.” (Oct. 24th Tr. 13-14). It is the Government’s conduct in this case, not the Court’s review, that is “highly unusual, to say the least.” In re Dep’t of Commerce, 2018 WL 5259090, at *1 (Gorsuch, J., concurring in part and dissenting in part).B. Defendants Fail to Show a Likelihood of Success on the Merits of Any Question that Would Justify a Stay of TrialDefendants’ failure to show the likelihood of irreparable harm is, by itself, fatal to their stay application, but they also fail to show that a likelihood of success on the merits warrants a stay of trial. See Hollingsworth, 558 U.S. at 190. To be sure, the Supreme Court’s October 22, 2018 Order suggests that that Court may rule that this Court erred in its September 21, 2018 7 Order authorizing a deposition of Secretary Ross.⁴ But that prospect alone does not warrant delaying the trial at Defendants’ request. If the Supreme Court vacates this Court’s September 21, 2018 Order before, during, or after trial, it will have no effect on the existing record, which presently lacks Secretary Ross’s deposition testimony. And, however unlikely it may be, but compare, e.g., Barnes v. E-Systems, Inc. Grp. Hosp. Med. & Surgical Ins. Plan, 501 U.S. 1301, 1303 (1991) (Scalia, J.) (in chambers) (granting a stay pending a petition for certiorari based in part on the prediction that “a grant of certiorari” was “probable”), with Barnes v. E-Systems, Inc. Grp. Hosp. Med. & Surgical Ins. Plan, 502 U.S. 981 (1991) (mem.) (denying certiorari), if the Supreme Court allows a deposition of Secretary Ross before this Court enters final judgment, the transcript of that deposition can presumably be added to the trial record. In any event, it is Plaintiffs who bear the burden of proof in these cases, see, e.g., Schaffer v. Weast, 546 U.S. 49, 56-57 (2005), and Plaintiffs who seek to secure Secretary Ross’s deposition to meet that burden. Despite that, Plaintiffs are content to take their chances and proceed to trial knowing that, even if the Supreme Court ultimately lifts the stay and allows a deposition of Secretary Ross, it may be too late for them to benefit in these cases. Thus, while the likelihood of success on the merits of Defendants’ challenge to this Court’s September 21, 2018 Order justifies the already existing stay of that Order, it does not justify a stay of trial.Perhaps recognizing that, Defendants confidently predict that the Supreme Court is likely to opine that this Court erred in authorizing extra-record discovery in the first place. (Gov’t Stay Mot. 2-3). But they base that prediction almost exclusively on the dissent from the Supreme Court’s Order. (See id. at 1-3). It should go without saying that the dissent did not carry the day in the Supreme Court; instead, it represents the views of only two Justices. More to the point, there is nothing in the Supreme Court’s Order itself that supports Defendants’ confident prediction. Admittedly, the Supreme Court’s Order states that “[t]he denial of the stay with respect to” the July 3, 2018 Order “does not preclude the applicants from making arguments with respect to” that Order. In re Dep’t of Commerce, 2018 WL 5259090 at *1 (emphasis added). But it is rather aggressive to read that language as an “invit[ation],” as Defendants do. (Gov’t Stay Mot. 3). After all, if one person says to another “you are not precluded from attending my party,” the latter would be hard pressed to describe the expression as an “invitation.”⁵ In any event, even if the Supreme Court’s language could reasonably be read as an invitation, it is rank speculation to infer from that invitation that the Supreme Court is likely to hold, in the present interlocutory posture no less, that this Court erred in authorizing extra-record discovery.In fact, for several reasons, the Court concludes that the Supreme Court is unlikely to disturb the July 3, 2018 Order in advance of this Court’s consideration of the merits. First, that Order pertained to discovery, which — apart from the possible deposition of Secretary Ross — will be complete when Defendants file their petition with the Supreme Court. (See Docket No. 401).⁶ Second, Defendants’ suggestion that this Court’s July 3, 2018 Order somehow licensed a burdensome intrusion into the workings of the Executive Branch is overblown. The Court was careful to observe “that discovery in an APA action, when permitted, should not transform the litigation into one involving all the liberal discovery available under the federal rules” and should instead be limited to what is “necessary to effectuate the Court’s judicial review.” (July 3rd Tr. 85 (internal quotation marks omitted)). On that basis, the Court sharply curtailed the discovery Plaintiffs could conduct. (See id. at 85-87 (limiting Plaintiffs to ten depositions and limiting discovery, absent agreement or leave of Court, to the Departments of Commerce and Justice)).⁷ Moreover, Defendants’ cries of intrusion and burden ring hollow in light of their own conduct. Rather than seek immediate review of the Court’s July 3, 2018 Order authorizing extra-record discovery, they waited nearly two full months — until extra-record discovery was substantially complete — before seeking a stay and any form of appellate review. See New York, 2018 WL 4279467, at *2.Finally, the Court’s decision to authorize extra-record discovery was, and remains, well founded. For starters, although judicial review of agency action is generally limited to the administrative record, see, e.g., Florida Power & Light Co. v. Lorion, 470 U.S. 729, 743-44 (1985), it is well established that “an extra-record investigation by the reviewing court may be appropriate when there has been a strong showing in support of a claim of bad faith or improper behavior on the part of agency decisionmakers,” Nat’l Audubon Soc’y v. Hoffman, 132 F.3d 7, 14 (2d Cir. 1997). The “bad faith” exception “is logical because once there is a showing of bad faith by the agency, the reviewing court has lost its reason to trust the agency. There is no reason, then, to presume that the record is complete, and justice is served only by going beyond the record to ascertain the true range of information before the agency.” James N. Saul, Overly Restrictive Administrative Records and the Frustration of Judicial Review, 38 Envtl. L. 1301, 1308 (2008). More importantly, the exception was spawned by the Supreme Court itself, see Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 420 (1971), overruled on other grounds by Califano v. Sanders, 430 U.S. 99, 105 (1977), and has been adopted by every Court of Appeals in the country, see Saul, 38 Envtl. L. at 1308-09 & n.57. Indeed, Defendants do not dispute — and have never disputed — that “bad faith” can justify extra-record discovery. (See, e.g., Docket No. 194, at 4 (conceding that there is a “bad faith” exception to the “record rule”)). And nothing in the Supreme Court’s October 22, 2018 Order casts doubt on the well-established exception.Notably, even the Justices who dissented from the Supreme Court’s Order seem to accept that there is a “bad faith” exception to the record rule. See In re Dep’t of Commerce, 2018 WL 5259090 at *1-2 (Gorsuch, J., concurring in part and dissenting in part). Instead, they take issue with this Court’s conclusion that Plaintiffs made a sufficient preliminary showing to trigger that exception. See id. The Court respectfully disagrees. This Court’s conclusion that Plaintiffs had made such a showing was not based on a finding that Secretary Ross “c[ame] to office inclined to favor a different policy direction, solicit[ed] support from other agencies to bolster his views, disagree[d] with staff, or cut[] through red tape.” Id. at *1. Such circumstances, even taken together, would not be exceptional. Instead, the Court’s conclusion was based on a combination of circumstances that were, taken together, most exceptional: (1) Secretary Ross’s own admission that he had “already decided to add the citizenship question before he reached out to the Justice Department” to request the question; (2) evidence that he had “overruled senior Census Bureau career staff, who had concluded…that reinstating the citizenship question would be very costly and harm the quality of the census count”; (3) indications that the Census Bureau had “deviated significantly from standard operating procedures in adding the citizenship question”; and (4) Plaintiffs’ prima facie showing that Secretary Ross’s stated justification was pre-textual. (July 3rd Tr. 82-83 (emphasis added) (internal quotation marks and brackets omitted)). Most significant, the Court found reason to believe that Secretary Ross had provided false explanations of his reasons for, and the genesis of, the citizenship question — in both his decision memorandum and in testimony under oath before Congress. (July 3rd Tr. 79-80). If those circumstances, taken together, are not sufficient to make a preliminary finding of bad faith that would warrant extra-record investigation, it is hard to know what circumstances would — short of an agency head’s outright confession that his reasons were pretextual (in which case, of course, there would be no need for discovery). In fact, circumstances far short of those present in these cases have been found by other courts to justify discovery beyond the administrative record. See, e.g., Pub. Power Council v. Johnson, 674 F.2d 791, 794-95 (9th Cir. 1982); Batalla Vidal v. Duke, No. 16-CV-4756 (NGG), 2017 WL 4737280, at *3-5 (E.D.N.Y. Oct. 19, 2017); Tummino v. von Eschenbach, 427 F. Supp. 2d 212, 231-33 (E.D.N.Y. 2006). Thus, there is nothing unusual with this Court’s decision to allow extra-record discovery and — in light of Defendants’ election not to move for summary judgment — to adjudicate Plaintiffs’ claims of bad faith and pretext through a trial.If those circumstances, taken together, are not sufficient to make a preliminary finding of bad faith that would warrant extra-record investigation, it is hard to know what circumstances would — short of an agency head’s outright confession that his reasons were pretextual (in which case, of course, there would be no need for discovery). In fact, circumstances far short of those present in these cases have been found by other courts to justify discovery beyond the administrative record. See, e.g., Pub. Power Council v. Johnson, 674 F.2d 791, 794-95 (9th Cir. 1982); Batalla Vidal v. Duke, No. 16-CV-4756 (NGG), 2017 WL 4737280, at *3-5 (E.D.N.Y. Oct. 19, 2017); Tummino v. von Eschenbach, 427 F. Supp. 2d 212, 231-33 (E.D.N.Y. 2006). Thus, there is nothing unusual with this Court’s decision to allow extra-record discovery and — in light of Defendants’ election not to move for summary judgment — to adjudicate Plaintiffs’ claims of bad faith and pretext through a trial.C. Issuance of a Stay Would Injure Plaintiffs and Harm the Public InterestIn short, Defendants fail to carry their burden on either of the first two, and “most critical,” factors in the analysis of whether a stay is warranted. Nken v. Holder, 556 U.S. 418, 434 (2009). The Court could stop there, see id. at 435, but the third and fourth factors — “whether issuance of the stay will substantially injure the other parties interested in the proceeding” and “where the public interest lies,” U.S. S.E.C. v. Citigroup Glob. Mkts. Inc., 673 F.3d 158, 162 (2d Cir. 2012) — also weigh heavily against a stay. As noted, Defendants have repeatedly insisted, and insist even now, that the resolution of Plaintiffs’ claims “is a matter of some urgency.” (Gov’t Stay Mot. 4; see Docket No. 103, at 4-5 (noting that “the Census Bureau has indicated in its public planning documents that it intends to start printing the physical 2020 Census questionnaire by May 2019″ and that Ron Jarmin, Acting Director of the Census Bureau and a Defendant here, “testified under oath before Congress…that the Census Bureau would like to ‘have everything settled for the questionnaire this fall’” and “wants to resolve this issue ‘very quickly’”)). Awaiting prophylactic guidance from the Supreme Court — which may not come for months and may not come at all — would make it difficult, if not impossible, to meet that goal.⁸ More broadly, as the Court has noted previously, “there is a strong interest in ensuring that the census proceeds in an orderly, transparent, and fair manner — and, relatedly, that it is conducted in a manner that ‘bolsters public confidence in the integrity of the process and helps strengthen this mainstay of our democracy.’” New York, 2018 WL 4279467, at *3 (quoting Franklin v. Massachusetts, 505 U.S. 788, 818 (1992) (Stevens, J., concurring in part and concurring in the judgment)). Those interests weigh heavily against any delay and in favor of making an adequate record for this Court to render an initial decision — and for higher courts to then review that decision without any risk that those courts would conclude that a remand to develop the record would be in order.In their pending motion before the Second Circuit, Defendants contend that a stay of trial would not prevent resolution of Plaintiffs’ claims before the census questionnaires have to be printed. (Motion for Stay (“2d Cir. Stay Mot.”), at 9, Docket No. 68, No. 18-2856 (2d Cir. Oct. 25, 2018); see also Oct. 24th Tr. 11-12). The Court does not share their confidence. There is no telling when the Supreme Court will issue a decision on Defendants’ forthcoming petition. It could do so in days; or it could take months. If the Supreme Court’s decision does not affect this Court’s plan to proceed with a trial, the Court would then have to reschedule trial — no small task given the upcoming holidays, the parties’ schedules (including two trials in parallel cases pending in other districts scheduled in January), and the Court’s own congested calendar.⁹ If the Supreme Court’s decision makes clear that Plaintiffs’ claims should be resolved by summary judgment rather than trial, the parties will need to prepare extensive motion papers. In either case, it will take time for this Court to issue a written ruling and enter final judgment. And whatever this Court decides, the losing parties will almost certainly appeal to the Second Circuit and, in turn, to the Supreme Court. It would be hard enough for that normally lengthy process to run its course by next May or June — when the census questionnaires are apparently scheduled to be printed (see Docket No. 103, at 4-5; Oct. 24th Tr. 11) — if these cases proceed to trial on November 5, 2018. Granting a stay of indefinite duration could make a timely final decision next to impossible.****In short, as prudent as it might be under other circumstances to await further guidance from the Supreme Court, there are good reasons not to do so here and instead to proceed to trial as scheduled. Time is of the essence. At bottom, Defendants are seeking a preemptive ruling from the Supreme Court on a decision that this Court has not yet even made — namely, what evidence the Court may consider in ruling on the merits — thereby seeking to disrupt “the appropriate relationship between the respective courts.” Coopers & Lybrand v. Livesay, 437 U.S. 463, 476 (1978). Making matters worse, Defendants have not yet even formally asked the Court to make a decision on that issue. They elected not to do so in the form of a summary judgment motion, and thus conceded, as a procedural matter, that trial is appropriate. And, perhaps most importantly, Defendants suffer no substantive, cognizable harm whatsoever in proceeding to trial as scheduled. They can make, and thus preserve, any argument they want about the scope of what this Court may consider in rendering a decision. And if they are unsuccessful before this Court, they can seek review of this Court’s final judgment from the Second Circuit and, if necessary, the Supreme Court — as they could in any other case.Put simply, the pending challenge to this Court’s Order authorizing a deposition of Secretary Ross notwithstanding, Defendants provide no basis to deviate from the wellestablished and well-justified procedures that have generally been applied in federal courts for generations — whereby district courts decide cases in the first instance, followed by an appeal by the losing party, on a full record, to the court of appeals and, thereafter, a petition to the Supreme Court. Defendants may yet have their day to argue the merits in the Supreme Court. But for many salutary reasons, that day should not come before this Court has decided the merits in the first instance. See, e.g., Firestone Tire & Rubber Co. v. Risjord, 449 U.S. 368, 374 (1981) (“[The final judgment rule] emphasizes the deference that appellate courts owe to the trial judge as the individual initially called upon to decide the many questions of law and fact that occur in the course of a trial. Permitting piecemeal appeals would undermine the independence of the district judge, as well as the special role that individual plays in our judicial system. In addition, the rule is in accordance with the sensible policy of avoid[ing] the obstruction to just claims that would come from permitting the harassment and cost of a succession of separate appeals from the various rulings to which a litigation may give rise, from its initiation to entry of judgment. The rule also serves the important purpose of promoting efficient judicial administration.” (internal quotation marks and citation omitted)).Accordingly, Defendants’ motion for a stay of trial and associated deadlines is DENIED. The Clerk of Court is directed to terminate Docket No. 397.SO ORDERED.Dated: October 26, 2018New York, New York