Harvard Law Review Harvard Law Review Harvard Law Review

Civil Procedure

Jordan v. Commissioner, Mississippi Department of Corrections

Eleventh Circuit Denies Mississippi Death Row Prisoners Discovery by Creating a Federal Lethal Injection Secrecy Privilege.

In 2015, the Supreme Court held that death row prisoners challenging a method of execution must produce a “known and available” alternative.1×1. See Glossip v. Gross, 135 S. Ct. 2726, 2731 (2015) (upholding the denial of a preliminary injunction because “prisoners failed to identify a known and available alternative method of execution that entails a lesser risk of pain, a requirement of all Eighth Amendment method-of-execution claims”). To that end, prisoners have subpoenaed state corrections departments seeking information on their execution protocols. In response, states have passed lethal injection secrecy acts2×2. See, e.g., Ariz. Rev. Stat. Ann. § 13-757(C) (2010); Ark. Code Ann. § 5-4-617 (2013); Fla. Stat. § 945.10 (2018); Ind. Code Ann. § 35-38-6-1(f) (LexisNexis 2012); La. Stat. Ann. § 15:570(e)(3)(G) (2012); Ohio Rev. Code Ann. § 2949.221 (LexisNexis Supp. 2019); Okla. Stat. Ann. tit. 22, § 1015 (West Supp. 2018); Tenn. Code Ann. § 10-7-504(h)(1) (2012); Tex. Code Crim. Proc. Ann. art. 43.14 (West 2018); Va. Code Ann. § 53.1-234 (2013); Wyo. Stat. Ann. § 7-13-916 (2019). for fear that publicizing their execution methods would prompt backlash from activists and make the death penalty impossible to administer.3×3. Mary D. Fan, The Supply-Side Attack on Lethal Injection and the Rise of Execution Secrecy, 95 B.U. L. Rev. 427, 429 (2015) (detailing how efforts to publicize the manufacturers of execution drugs have led to boycotts that interfere with the drug supply). Recently, in Jordan v. Commissioner, Mississippi Department of Corrections,4×4. 908 F.3d 1259 (11th Cir. 2018). the Eleventh Circuit affirmed a district court judgment quashing a subpoena that two Mississippi prisoners had issued to the State of Georgia.5×5. Id. at 1261. By allowing Georgia’s Lethal Injection Secrecy Act6×6. Ga. Code Ann. § 42-5-36 (2014). — a state law — to shape the scope of discovery in a federal court on a federal question, the appeals court took an unusual step not required by civil procedure doctrine. The court’s deference to a state secrecy law has a troubling implication: it suggests that states may avoid constitutional challenges by passing laws that block evidence from civil rights plaintiffs. As a result, Jordan risks frustrating efforts to vindicate constitutional rights and undermining the purpose of § 1983.

In 1976, Richard Jordan was sentenced to death for the kidnapping and murder of Edwina Marter.7×7. Jordan v. State, 786 So. 2d 987, 997–98 (Miss. 2001) (en banc). Despite procedural errors requiring the vacation of his sentence and a short-lived agreement to life imprisonment without parole,8×8. Id. at 998–99. Jordan has remained Mississippi’s longest-serving death row prisoner.9×9. Jordan v. State, 224 So. 3d 1252, 1253 (Miss. 2017) (en banc) (“Jordan has been on death row for over forty years . . . .”), cert. denied, 138 S. Ct. 2567 (2018). Ricky Chase was sentenced to death in 1990 for the murder and robbery of Elmer Hart.10×10. Chase v. State, 873 So. 2d 1013, 1015–16 (Miss. 2004) (en banc). Despite evidence of an intellectual disability, his death penalty conviction still stands.11×11. See Chase v. State, 171 So. 3d 463, 466, 471 (Miss. 2015) (en banc). Both prisoners are due to be executed via Mississippi’s three-drug protocol: first, a sedative to anesthetize; second, a paralytic to incapacitate; and third, potassium chloride to stop the heart.12×12. Jordan, 908 F.3d at 1261.

Jordan and Chase brought a § 1983 action in federal court in the Southern District of Mississippi, challenging the use of the three-drug protocol.13×13. Id.; see 42 U.S.C. § 1983 (2012). They alleged that there was a substantial risk that the proposed sedative would not “sufficiently anesthetize” them, leaving them “conscious and fully sensate” by the time the paralytic took effect.14×14. Jordan, 908 F.3d at 1261. Thus, the prisoners could find themselves unable to breathe, slowly suffocating to death but entirely powerless to indicate their suffering.15×15. Id. Defendants, various officials within the Mississippi Department of Corrections, “denied that a single-drug procedure . . . was a feasible alternative to Mississippi’s three-drug protocol.”16×16. Id. at 1262. Specifically, defendants alleged that pentobarbital — the chemical used in the single-drug protocol — was unavailable.17×17. Id. To counter this claim, Jordan and Chase served a nonparty subpoena on the Georgia Department of Corrections (GDC) in federal court in the Northern District of Georgia.18×18. Id. They asked Georgia, which uses pentobarbital in its executions,19×19. Id. to produce documents concerning “the feasibility of a one-drug lethal injection protocol using pentobarbital, including specific details about [its] source and manner of acquiring” the drug.20×20. Id. The GDC moved to quash the subpoena, arguing that the information sought “was irrelevant . . . and, in any event, protected from disclosure by Georgia’s Lethal Injection Secrecy Act.”21×21. Id.

The district court referred the question to a magistrate judge, who granted the motion to quash.22×22. Id. The magistrate’s order rejected the defendant’s relevance argument and focused instead on Georgia’s Secrecy Act.23×23. Ga. Dep’t of Corr. v. Jordan, No. 16-cv-02582, 2016 WL 9776069,,(iv) na, 523 U.S. al. 1987) federal court int between state privilege law and federal discovery law noted hemical “porting pro at *1–2 (N.D. Ga. Oct. 20, 2016). The text of the Act has a broad preclusive effect, shielding from disclosure any potentially identifying information on the manufacturers, suppliers, and administrators of execution drugs.24×24. See Ga. Code Ann. § 42-5-36(d) (2014). F More critically, the Act had been afforded an “expansive reading” by the Eleventh Circuit in the past.25×25. Jordan, 2016 WL 9776069, at *3. For that reason, the magistrate judge not only granted the motion to quash the subpoena, but also declined to require a privilege log.26×26. Id. at *4. A privilege log is a document that may be used to meet the requirement that parties asserting privilege “[d]escribe the nature of the documents . . . not produced or disclosed . . . in a manner that, without revealing information itself privileged or protected, will enable other parties to assess the claim.” Fed. R. Civ. P. 26(b)(5). ell within that protected scope,he Georgia ect in viewing the on was non-dispositive. d the subpoena litigation in Georgia it leFedF

The district court affirmed the magistrate judge’s grant of the motion to quash over two objections from the plaintiffs.27×27. Jordan, 908 F.3d at 1262–63. First, plaintiffs challenged what they characterized as an overbroad assertion of privilege.28×28. Memorandum of Respondents in Opposition to Motion to Quash and Objections at 19–20, Jordan, 2016 WL 9776069 (No. 16-cv-02582). Second, they argued that the magistrate judge had erred by failing to require a privilege log that would specify how the Secrecy Act applied to each requested document.29×29. Id. at 17–19. The district court found these objections unavailing. It determined that the motion to quash a subpoena was nondispositive, since such motions are routine pretrial matters.30×30. Jordan, 908 F.3d at 1263. Accordingly, the magistrate’s ruling would be overturned only if it was “clearly erroneous” or “contrary to law.”31×31. Id. Because the requested documents seemed to fit within the scope of information protected by the Act, the magistrate judge’s ruling was not clearly erroneous.32×32. Id. And because the Eleventh Circuit had deemed the Act constitutional, the grant of a motion to quash was not contrary to law.33×33. Id. Therefore, the district court determined that the magistrate’s ruling should remain undisturbed.34×34. Id. at 1262–63.

The Eleventh Circuit affirmed. Writing for a unanimous panel, Judge Julie Carnes35×35. Judge Carnes was joined by Judge Tjoflat and Judge William Pryor. first addressed whether the district court had applied the appropriate standard of review to the magistrate judge’s determination.36×36. Jordan, 908 F.3d at 1263. A magistrate judge’s ruling on a dispositive motion requires de novo review, but if the motion is nondispositive, it may be reversed only if it is “clearly erroneous” or “contrary to law.”37×37. Id.; see Fed. R. Civ. P. 72; see also 28 U.S.C. § 636 (b)(1)(A) (2012). The appeals court observed that, while granting the motion to quash would conclusively resolve the subpoena litigation in Georgia, it would leave the underlying § 1983 action in Mississippi active.38×38. Jordan, 908 F.3d at 1264. Consequently, the Eleventh Circuit determined that the district court correctly viewed the Georgia litigation as ancillary, and the subpoena motion as nondispositive.39×39. Id. at 1263–64. Thus, the district court properly applied the “clearly erroneous” or “contrary to law” standard of review.40×40. Id. at 1264.

The appeals court then evaluated the decision to grant the motion to quash in its entirety without requiring a privilege log.41×41. Id. at 1264–67. It relied on three factors: the breadth of the shield created by the Secrecy Act; the validity of Georgia’s interest in keeping its execution procedures secret; and the legitimacy afforded to the Act by the Eleventh Circuit’s own precedent. First, the court reasoned that the Georgia legislature drafted its Secrecy Act to characterize a wide swath of execution information as “confidential state secret[s].”42×42. Id. at 1265 (quoting Ga. Code Ann. § 42-5-36(d)(2) (2014)). It explained that the subpoenaed documents most relevant to the § 1983 claims clearly fell within that protected scope.43×43. See id. at 1267. Second, given the political dynamics of death penalty opposition, any public disclosure of the information risked attracting boycotts and irrevocably jeopardizing Georgia’s supply of execution drugs.44×44. Id. at 1265. See generally Fan, supra note 3, at 436–41. Consequently, the confidentiality provided by the Secrecy Act was critical in allowing Georgia to continue administering the death penalty.45×45. See Jordan, 908 F.3d at 1265. Third, in previous cases the Eleventh Circuit had consistently upheld the Secrecy Act as constitutional.46×46. See id. at 1266. In light of these considerations, the court affirmed the grant of a motion to quash. Moreover, since the privilege created by the Secrecy Act is intentionally sweeping and the interests underlying it compelling, the court did not require the GDC to provide a privilege log.47×47. Id. at 1267 (noting that the fundamental purpose of a privilege log, which is to “enable the parties to assess [a] claim of privilege,” was not applicable where the relevant information was already shielded from disclosure under the Secrecy Act (quoting Fed. R. Civ. P. 45(e)(2)(A)(ii))).

The Eleventh Circuit interpreted Georgia’s Lethal Injection Secrecy Act as a valid limit on the scope of discovery in a federal court on a federal question, applying the state law as if that were routine. But this approach was inconsistent with established federal precedent: generally, a federal question heard in federal court triggers the federal law of discovery.48×48. See Hancock v. Hobbs, 967 F.2d 462, 466 (11th Cir. 1992) (“A claim of privilege in federal court is resolved by federal common law . . . .” (citing Fed. R. Evid. 501)). While federal courts can create new privileges, they are disfavored,49×49. New federal privileges should neither be “lightly created nor expansively construed.” United States v. Nixon, 418 U.S. 683, 710 (1974). and courts creating them are required to carefully balance the interests of the parties.50×50. An exception from the general rule disfavoring testimonial privileges is justified when the proposed privilege “promotes sufficiently important interests to outweigh the need for probative evidence.” Jaffee v. Redmond, 518 U.S. 1, 9–10 (1996) (quoting Trammel v. United States, 445 U.S. 40, 51 (1980)). In elevating Georgia’s Secrecy Act into a federal privilege, the Eleventh Circuit never considered the prisoners’ interests,51×51. The opinion paid careful attention to Georgia’s interests, considering the effects of disclosure on Georgia’s ability to acquire execution drugs. See, e.g., Jordan, 908 F.3d at 1266 (noting that “the confidentiality provided by the Act is necessary to protect Georgia’s source” of execution drugs). By contrast, it did not consider how the inability to gain information might hamper plaintiffs’ § 1983 action. which would have militated against the result in Jordan. The result suggests that states in the circuit can avoid constitutional liability by shielding the information civil rights plaintiffs would require to support their claims under § 1983. That outcome frustrates the fundamental purpose of § 1983: to create a federal forum in which plaintiffs may assert their constitutional rights, regardless of state-created barriers.

A federal question heard in federal court necessarily triggers the federal common law of discovery.52×52. See Fed. R. Evid. 501. And federal common law has a presumption against creating new privileges.53×53. When determining a new privilege issue, “we start with the primary assumption that there is a general duty to give what testimony one is capable of giving, and that any exemptions which may exist are distinctly exceptional.” United States v. Bryan, 339 U.S. 323, 331 (1950) (citation omitted). Rule 501 of the Federal Rules of Evidence mandates that unless an appropriate provision determines otherwise, “[t]he common law — as interpreted by United States courts . . . — governs” federal discovery disputes.54×54. Fed. R. Evid. 501. Some lower federal courts have interpreted Rule 501 as generative, in the sense that it reflects a liberal approach to creating new privileges.55×55. See, e.g., Jaffee, 518 U.S. at 6–8 (surveying circuit court precedent identifying a novel psychotherapist-patient privilege); Riley v. City of Chester, 612 F.2d 708, 714 (3d Cir. 1979) (construing the “flexible language” of Rule 501 to “encompass a reporter’s privilege”). This interpretation treats the provision of the power to create as an indicator that Congress welcomed new privileges. The dominant approach reads Congress’s action not as encouragement but as abdication — leaving the matter to the courts. See George Fisher, Evidence 930 (3d ed. 2013). Congress adopted Rule 501, which resolves privilege disputes through common law rulemaking, rather than codify specific privileges drafted by an advisory committee. See Timothy P. Glynn, Federalizing Privilege, 52 Am. U. L. Rev. 59, 90 (2002). Scholars have attributed this choice to Congress’s sharp divisions over the “substantive social policies” advanced by privilege law. See Fisher, supra, at 931. Notably, the death penalty was in abeyance when Congress codified evidence laws. See Furman v. Georgia, 408 U.S. 238, 239–40 (1972) (holding that the death penalty as then applied was unconstitutional). But more have understood Rule 501 to counsel against creating novel privilege rules. The Supreme Court, for its part, has clarified that new federal privileges should neither be “lightly created nor expansively construed,”56×56. United States v. Nixon, 418 U.S. 683, 710 (1974). given that the Federal Rules of Evidence have a “permissive”57×57. Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589 (1993). character reflecting the view that “the public . . . has a right to every man’s evidence.”58×58. Bryan, 339 U.S. at 331 (citation omitted). To further emphasize this point, the Court explicitly stated that it is “disinclined” to use its power under Rule 501 “expansively.”59×59. Univ. of Pa. v. Equal Emp’t Opportunity Comm’n, 493 U.S. 182, 189 (1990). For these reasons, federal common law reflects a presumption against the creation of new privileges.

In Jordan, the underlying case was a § 1983 action implicating federal constitutional rights in federal court. Consistent with Rule 501, and in the absence of any relevant provision regarding privilege,60×60. 18 U.S.C. § 3596 (2012) regulates the imposition of the federal death penalty but creates no lethal injection secrecy. federal common law controlled. Federal common law includes no lethal injection privilege.

Consequently, by deciding to apply state law in federal court, Jordan effectively created a new federal privilege. The Eleventh Circuit, following the example of the Fourth Circuit, could have determined that quashing the subpoena was appropriate because producing information on execution protocols would be unduly burdensome for Georgia.61×61. The Fourth Circuit relied on an undue burden analysis to quash a subpoena issued by the same plaintiffs in Virginia Department of Corrections v. Jordan, 921 F.3d 180, 188–92 (4th Cir. 2019). However, it did not embark on that type of analysis. Instead, it applied Georgia’s Lethal Injection Secrecy Act to the privilege context — an outcome that is disfavored under Rule 501 and not firmly supported by the circuit’s precedents. Though previous decisions on the Secrecy Act had affirmed its constitutionality,62×62. See, e.g., Jones v. Comm’r, Ga. Dep’t of Corr., 811 F.3d 1288, 1292–95 (11th Cir. 2016); Terrell v. Bryson, 807 F.3d 1276, 1276–77 (11th Cir. 2015); Gissendaner v. Comm’r, Ga Dep’t of Corr., 803 F.3d 565, 576 (11th Cir. 2015); Wellons v. Comm’r, Ga. Dep’t of Corr., 754 F.3d 1260, 1261 (11th Cir. 2014). the question of what, if any, effect the Act should have on the federal law of privilege had not been decided.63×63. None of the earlier cases assessing the Act’s constitutionality implicated subpoenas in federal court. See cases cited supra note 62. Thus, in Jordan, the Eleventh Circuit created a new federal privilege.

In doing so, the Court failed to consider the interests that should be accounted for in a privilege analysis. While its requirements are incredibly broad, Rule 501 is understood to prescribe a balancing methodology.64×64. See Trammel v. United States, 445 U.S. 40, 51 (1980) (assessing whether an asserted privilege “promote[d] sufficiently important interests to outweigh the need for probative evidence”). Therefore, courts considering new privileges would normally weigh the interests that the privilege tramples against the interests protected by secrecy “in . . . light of reason and experience.”65×65. See Jaffee v. Redmond, 518 U.S. 1, 6 (1996) (“‘[R]eason and experience[]’ [are] the touchstones for acceptance of a privilege under Rule 501 . . . .” (citations omitted)). Typically, “reason” refers to a policy judgment by the court. 66×66. See Elizabeth Kimberly (Kyhm) Penfil, In the Light of Reason and Experience: Should Federal Evidence Law Protect Confidential Communications Between Same-Sex Partners?, 88 Marq. L. Rev. 815, 815 (2005) (“In considering whether to create a new evidentiary privilege[,] . . . federal courts consider . . . whether federal policy supports the privilege . . . .”). “Experience” calls for an assessment of the privilege’s pedigree, analyzing whether it is analogous to a privilege traditionally recognized at common law.67×67. See, e.g., United States v. Gillock, 445 U.S. 360, 368 (1980) (declining to recognize a new privilege because it lacked “historical antecedents” and was not “ensconced in . . . common law”); see also Jaffee, 518 U.S. at 14 (considering the “uniform judgment of the states” in creating a psychotherapist privilege). Troublingly, Jordan eschewed this approach, inappropriately deferring to state law and declining to be explicit about its creation of a new federal privilege.68×68. Notably, the court declined to address the arguments raised in the briefs on whether quashing a subpoena required the creation of a new federal privilege. See, e.g., Brief of Appellants Richard Jordan and Ricky Chase at 21–26, Jordan, 908 F.3d 1259 (No. 17-12948). The opinion neither mentioned the presumption against creation nor explained why this particular privilege was warranted in light of it. Providing such an explanation would be challenging; lethal injection secrecy is not only novel but also an uneasy fit with traditional privileges, which are recognized in order to facilitate open communication (for example, the attorney-client privilege) or to uphold human dignity in intimate relationships (for example, spousal privilege).69×69. See, e.g., Upjohn Co. v. United States, 449 U.S. 383, 389 (1981) (discussing attorney-client privilege); Wolfle v. United States, 291 U.S. 7, 14 (1934) (describing spousal privilege).

As a result, the opinion gives the impression that Georgia’s law in and of itself justifies the new privilege.70×70. See, e.g., Jordan, 908 F.3d at 1264–67 (relying extensively on the Lethal Injection Secrecy Act to justify quashing the subpoena and rejecting the need for a privilege log). But state privilege law need not be considered and commands no deference under Rule 501.71×71. See Fisher, supra note 55, at 931 (explaining that when federal law provides the rule of decision, federal common law should control (citing Fed. R. Evid. 501)). By incorporating state privilege law automatically, without fully grappling with the presumption against new federal privileges, the circuit misapplied federal law to the benefit of state policy.

In addition to avoiding direct engagement with the presumption against new privileges, the panel discussed Georgia’s interest in secrecy approvingly without any similar consideration of the prisoners’ interests. To frame its evaluation of Georgia’s interests, the court explicitly referred to the “concerted effort by death penalty opponents to make lethal injection drugs unavailable.”72×72. Jordan, 908 F.3d at 1265. It then emphasized the effect that publicizing information on execution protocols would have on Georgia’s ability to execute.73×73. Id. (“[W]ithout the confidentiality offered . . . by the statute . . . there is significant risk that persons and entities necessary to the execution would become unwilling to participate.” (first alteration in original) (citation omitted)). By contrast, the court devoted relatively little attention to the prisoners’ weighty interests in ensuring that the method of their execution is not illegally cruel. This silence is important; it shows that the panel failed not only to wrestle with the presumption against new privileges, but also to apply the balancing test that new privileges require.

Failure to consider the prisoners’ interests led the court to ignore two large-scale consequences of its decision — both of which will affect prisoners negatively. First, if execution information is barred from discovery, method-of-execution claims become virtually impossible.74×74. See Deborah W. Denno, The Lethal Injection Quandary: How Medicine Has Dismantled the Death Penalty, 76 Fordham L. Rev. 49, 96 (2007) (explaining that a “lack of information . . . makes it difficult — if not impossible — to evaluate the constitutionality of lethal injection on any level”). The opinion will hinder the progress of litigation designed to make the death penalty less cruel; if denied the capacity to compel discovery on execution procedures, prisoners will be unable to meet the Glossip requirements to enforce their Eighth Amendment rights.75×75. See Glossip v. Gross, 135 S. Ct. 2726, 2737 (2015); see also Eric Berger, Lethal Injection Secrecy and Eighth Amendment Due Process, 55 B.C. L. Rev. 1367, 1412–16 (2014) (arguing that lethal injection secrecy forecloses successful Glossip claims and might also degrade due process rights). Further, if the Eleventh Circuit’s approach proves popular, many executions will not face public scrutiny.76×76. See Denno, supra note 74, at 95–96.

Second, if a state can simply shield the evidence required to sustain a constitutional claim, then § 1983 risks becoming a useless vehicle. The purpose of § 1983 is to enforce federal constitutional rights, in part by enabling plaintiffs to avoid entanglement with state-created procedural barriers to civil rights actions.77×77. See Harry A. Blackmun, Section 1983 and Federal Protection of Individual Rights — Will the Statute Remain Alive or Fade Away?, 60 N.Y.U. L. Rev. 1, 5–7 (1985); see also, e.g., Felder v. Casey, 487 U.S. 131, 152–53 (1988) (holding that an arrestee’s failure to comply with a state’s notice-of-claim statute was not fatal to his federal civil rights action). To allow a § 1983 claim to be hindered by a state law risks building precedent whereby states can avoid constitutional strictures by legislating around them. In fact, this problem was explicitly recognized in federal court in Kelly v. City of San Jose.78×78. 114 F.R.D. 653 (N.D. Cal. 1987). In that case, the court noted that “[i]t obviously would make no sense to permit state law to determine what evidence is discoverable in cases brought pursuant to federal statutes whose central purpose is to protect citizens from abuses of power by state and local authorities.”79×79. Id. at 656. Indeed, “[i]f state law controlled, state authorities could effectively insulate themselves from constitutional norms simply by developing doctrines that made it virtually impossible for plaintiffs to discover the kind of information they need” to vindicate their constitutional rights.80×80. Id. By the same logic, the Eleventh Circuit erred by allowing a state law to hinder a § 1983 claim.

By permitting state law to determine federal privileges in the civil rights context, the Eleventh Circuit embraced a misplaced federalism. Rather than adhere to long-settled precedent that federal privilege law should reign supreme on federal questions in federal court, the court treated state law as authoritative. For civil rights plaintiffs, this presents a real conundrum. If a state can wriggle out of constitutional challenges by inventing procedural roadblocks, then § 1983 could be rendered useless. Section 1983 promises citizens an incisive tool to vindicate their constitutional rights against the state. In Jordan, the Eleventh Circuit blunted the blade.