An Empirical investigation into appellate structure and the perceived quality of appellate review

Vanderbilt Law Review

Volumn 61, Issue 6, 2008, Pages 1745-1822

  Nash, Juonathan Remy ^a   Pardo, Rafael I ^b  

^a EMORY UNIVERSITY   (United States)

^b SEATTLE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 58149392968     PISSN: 00422533     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (300)

1. See Jonathan Remy Nash, Resuscitating Deference to Lower Federal Court Judges' Interpretations of State Law, 77 S. cal. L. rev. 975, 1022-23 (2004) (describing the Condorcet Jury Theorem).
2. See Jonathan Remy Nash, A Context-Sensitive Voting Protocol Paradigm for Multimember Courts, 56 stan. L. rev. 75, 112-13 & 112 nn.130-31 (2003) (questioning the applicability of the Condorcet Jury Theorem in the context of appellate judicial decisionmaking).
3. See, e.g., Harry T. Edwards, The Effects of Collegiality on Judicial Decisionmaking, 151 U. pa. L. rev. 1639, 1649 (2003) (arguing that collegiality contributes to sound decisionmaking that focuses on the legal issues at hand); Lewis A. Kornhauser & Lawrence G. Sager, Unpacking the Court, 96 yale L.J. 82, 100-02 (1986).
4. But see Richard L. Revesz, Environmental Regulation, Ideology, and the D.C. Circuit, 83 Va. L. Rev. 1717, 1719 (1997) (finding empirical evidence that judges on an appellate panel of the same political party are more likely to vote ideologically); Cass R. Sunstein, David Schkade & Lisa Michelle Ellman, Ideological Voting on Federal Courts of Appeals: A Preliminary Investigation, 90 Va. L. rev. 301, 316-25 (2004) (finding some evidence of ideological voting on federal courts of appeals).
5. See Erwin Chemerinsky, Decision-Makers: In Defense of Courts, 71 AM. banke. L.J. 109, 115 (1997) ("Specialization offers two major advantages: expertise and uniformity."). For an argument that it might benefit the legal system to have some judges with expertise in areas other than law, see Adrian Vermeule, Should We Have Lay Justices?, 59 stan. L. rev. 1569, 1587-98 (2007).
6. See, e.g., Rochelle Cooper Dreyfuss, The Federal Circuit: A Case Study in Specialized Courts, 64 N.Y.U. L. rev. 1, 7 (1989) (citing predictability and efficiency as reasons for creating a specialized patent court to relieve the burden of technical patent cases on generalist judges); R. Polk Wagner & Lee Petherbridge, Is the Federal Circuit Succeeding? An Empirical Assessment of Judicial Performance, 152 u. Pa. L. rev. 1105, 1114-17 (2004) (giving an account of the establishment of the Federal Circuit's exclusive appellate jurisdiction over patent law).
7. See Nash, supra note 1, at 1022.
8. Salve Regina Coll. v. Russell, 499 U.S. 225, 232 (1991).
9. Id. at 231 (noting, with a negative connotation from the perspective of lawfinding ability, that district judges "preside alone over fast-paced trials").
10. Id. at 232.
11. Id. at 231-32.
12. See, e.g., Robert B. Ahdieh, Between Dialogue and Decree: International Review of National Courts, 79 N.Y.U. L. rev. 2029, 2047 (2004) (suggesting that appellate review and appellate hierarchy are integrally related by noting that "the various characteristics and functions of appellate review suggest that some gradation of judicial authority is central to the nature of appellate review," and that "[a]n appellate system of review is one defined by hierarchy");
13. John A. Ferejohn & Larry D. Kramer, Independent Judges, Dependent Judiciary: Institutionalizing Judicial Restraint, 77 N.Y.U. L. rev. 962, 998 (2002) ("[T]he development of an appellate hierarchy with collegial courts at the higher levels and stringent rules of vertical stare decisis operates structurally to ensure that no individual judge can, by his or her actions alone, inflict too much damage on the judiciary by making aberrant or overly ambitious decisions.");
14. Mortimer N. S. Sellers, The Doctrine of Precedent in the United States of America, 54 am. J. comp. L. 67, 68 (2006) ("The essence of the American system of precedent as experienced in practice resides in the great authority and hierarchical arrangement of the courts.").
15. But cf. Pauline T. Kim, Lower Court Discretion, 82 N.Y.U. L. rev. 383, 387-88 (2007) (arguing that the common principal-agent model for analyzing lower court efforts to fulfill appellate court mandates ignores the allocation of discretion to lower courts).
16. See, e.g., Lewis A. Kornhauser, Adjudication by a Resource-Constrained Team: Hierarchy and Precedent in a Judicial System, 68 S. cal. L. rev. 1605, 1607-08 (1995) (elucidating the traditional appellate hierarchy).
17. Absent en banc review, courts of appeals are bound by prior decisions issued by the court (independent of panel composition). E.g., United States v. Myers, 200 F.3d 715, 720 (10th Cir. 2000). In general, horizontal stare decisis does not extend beyond the court that issued an opinion to sibling courts of the same hierarchical level. While intermediate appellate courts will follow decisions issued by earlier panels of the same court-notwithstanding that the composition of the judges on the panels may vary-intermediate appellate courts generally are under no precedential obligation to follow decisions issued by sibling intermediate appellate courts of similar hierarchical rank. Thus, for example, a Ninth Circuit panel may find First Circuit precedent to be persuasive and choose to follow it, but stare decisis does not demand that the Ninth Circuit so act; rather, stare decisis leaves the Ninth Circuit free to disagree with and to disregard the First Circuit precedent. See, e.g., Evan H. Caminker, Why Must Inferior Courts Obey Superior Court Precedents?, 46 stan. L. rev. 817, 824-25 (1994). Also, the rule of horizontal precedent does not extend to trial courts, as discussed below.
18. See Amy Coney Barrett, Stare Decisis and Due Process, 74 U. colo. L. rev. 1011, 1015 (2003); Caminker, supra, at 825 ("[A] district court judge may ignore the decisions of 'foreign' courts of appeals as well as other district court judges, even within the same district." (footnote omitted)); Kornhauser, supra note 12, at 1609.
19. But see Daniel J. Bussel, Power, Authority, and Precedent in Interpreting the Bankruptcy Code, 41 UCLA L. rev. 1063, 1095 (1994) (noting a "long tradition" of district judges deviating from prior precedent in the same district only in extraordinary circumstances); infra note 71 and accompanying text.
20. See, e.g., Chemerinsky, supra note 4, at 111 ("[C]ourts generally issue written decisions that, when published, have precedential effect on future rulings involving different parties."); Susan B. Haire, Stefanie Lindquist & Donald R. Songer, Appellate Court Supervision in the Federal Judiciary: A Hierarchical Perspective, 37 LAW & SocY rev. 143, 145 (2003) ("Appellate oversight in the lower tiers of the federal judicial hierarchy provides a process through which circuit judges are expected to promote legal rules that will guide decision making in subsequent cases.");
21. Kornhauser, supra note 12, at 1609.
22. For example, civil law systems do not rely upon as stringent a hierarchy, or upon rules of precedent as stringent. See, e.g., Caminker, supra note 13, at 826; Kornhauser, supra note 12, at 1608; Thomas Lundmark, Interpreting Precedents: A Comparative Study, 46 AM. J. comp. L. 211, 214 (1998) (reviewing INTERPRETIVE PRECEDENTS: A COMPARATIVE STUDY (D. Neil MacCormick & Robert s. Summers eds., 1997)) ("One of the classic differences between civil-law and common-law jurisdictions is that the former do not recognize judicial precedent as an independent source of law." (footnote omitted)). For an exposition, and critique, of the necessity and desirability of stare decisis,
23. see Caminker, supra note 13, at 865-67.
24. Compare, e.g., Caminker, supra note 13, at 828-34 (arguing that the constitutional case for the binding nature of Supreme Court precedent on lower federal courts is "quite powerful"),
25. and Richard H. Fallon, Jr., Stare Decisis and the Constitution: An Essay on Constitutional Methodology, 76 N.Y.U. L. REV. 570, 577-85 (2001) (arguing in favor of the constitutional status of stare decisis),
26. with John Harrison, The Power of Congress over the Rules of Precedent, 50 DUKE L.J. 503, 513 (2000) (arguing to the contrary), Thomas Healy, Stare Decisis as a Constitutional Requirement, 104 W. va. L. rev. 43, 92 (2001) (same),
27. Gary Lawson, Controlling Precedent: Congressional Regulation of Judicial Decision-Making, 18 const. comment. 191, 204-07 (2001) (same),
28. and Michael Stokes Paulsen, Abrogating Stare Decisis by Statute: May Congress Remove the Precedential Effect of Roe and Casey?, 109 yale L.J. 1535, 1546-47 (2000) (same).
29. See Kornhauser, supra note 12, at 1606 (taking as a baseline assumption in developing economic theory of stare decisis that "the 'judicial team' seeks to answer the expected number of 'correct' answers subject to its resource constraint"); cf. Owen M. Fiss, Objectivity and Interpretation, 34 STAN. L. REV. 739, 746-47 (1982) (discussing how judges belong to an "interpretive community" that subscribes to the rule of law).
30. Even if goals other than arriving at the correct outcome motivate judges, see, e.g., Erin O'Hara, Social Constraint or Implicit Collusion? Toward a Game Theoretic Analysis of Stare Decisis, 24 SETON HALL L. REV. 736, 738-42 (1993) (arguing that judges' self-interest-including judges' interest in expanding their influence-explains the development of horizontal stare decisis); infra notes 100-102 and accompanying text, the fact remains that, to the extent that the U.S. judicial system substantially relies on the traditional hierarchical form and rules, the extent to which a court comports with that norm will increase the perception that it is reaching correct decisions.
31. See Kornhauser, supra note 12, at 1623 ("In a completely decentralized system each judge would have to attend to the caseload of every other judge in order to identify appropriate cases for review; in a hierarchical system, only the appellate judges need have a systemic perspective on caseload.");
32. cf. id. at 1620 (noting that, absent horizontal precedent, "each judge is more likely to give each case intensive consideration" (emphasis added)); id. at 1624 (arguing in favor of "strict vertical precedent because the hierarchical structure creates a division of labor between levels of the hierarchy");
33. id. at 1625-27 (arguing in favor of horizontal precedent at the appellate, but not the trial, level).
34. See Kornhauser, supra note 12, at 1622-24;
35. Caminker, supra note 13, at 839-43. Of course, a cost in such a system is that the first court may resolve the legal question incorrectly, and then bind future courts to that rule.
36. See O'Hara, supra note 17, at 736 n.3 (identifying the "primary social cost of stare decisis" as "the entrenchment of bad decisions");
37. see also Lewis A. Kornhauser, An Economic Perspective on Stare Decisis, 65 CHI.-KENT L. REV. 63, 67-73 (1989) (discussing reliance by a court on earlier decisions by that court, even if wrongly decided, as an optimization problem and as varying depending upon institutional structure).
38. There are other social benefits that rules of stare decisis provide-certainty, predictability, fairness, and consistency. See Caminker, supra note 13, at 843-56 (discussing the desire to avoid "delayed justice," the greater decisionmaking proficiency of superior courts, and uniform interpretation and application of law as consequentialist justifications for stare decisis); Kornhauser, supra note 12, at 74-78 (discussing fairness, competence, and certainty as justifications for stare decisis). These benefits, however, are not the result of the courts necessarily reaching correct conclusions. Indeed, these benefits would inhere if courts uniformly reached bad decisions. See Kornhauser & Sager, supra note 3, at 105 (contrasting consistency, soundness, and coherence).
39. See, e.g., Daniel Berkowitz & Karen Clay, The Effect of Judicial Independence on Courts: Evidence from the American States, 35 J. LEG. STUD. 399, 422-24 (2006) (finding a strong correlation between judicial independence and court quality);
40. Jonathan Remy Nash, Prejudging Judges, 106 COLUM. L. REV. 2168, 2171 (2006) (characterizing judicial independence and judicial accountability as "competing demands upon the judiciary").
41. But see Daniel M. Herman, Legal Infrastructure, Judicial Independence, and Economic Development 1 (Univ. S. Cal. Law Sch. Legal Studies Research Paper Series, Paper No. C06-1, 2006), available at http://ssrn.com/abstract=877490 ("There is some evidence that judicial independence is associated with economic growth, but the evidence is mixed and causation is unclear.").
42. See THE FEDERALIST NOS. 78, 79, 81 (Alexander Hamilton);
43. id. NOS. 47, 48, 51 (James Madison).
44. U.S. CONST, art. III, § 1.
45. See Eric A. Posner, The Political Economy of the Bankruptcy Reform Act of 1978, 96 MICH. L. REV. 47, 123 (1997) (noting that the "double appeal system was a concession to the federal judges, a symbol of the subordination of the bankruptcy court to the district court").
46. See Act of July 1, 1898, ch. 541, 30 Stat. 544 (repealed 1978).
47. Posner, supra note 23, at 61-62.
48. See Geraldine Mund, Appointed or Anointed: Judges, Congress and the Passage of the Bankruptcy Act of 1978: Part One: Outside Looking In, 81 AM. BANKR. L.J. 1, 3-6 (2007).
49. Bankr. R. 901(7) (1973) (repealed 1978).
50. Vern Countryman, Scrambling to Define Bankruptcy Jurisdiction: The Chief Justice, the Judicial Conference, and the Legislative Process, 22 HARV. J. ON IEGIS. 1, 2 (1985). Hearsay evidence suggests that at least one Article III judge viewed bankruptcy judges as occupying the professional status equivalent to a janitor.
51. See Mund, supra note 26, at 12 n.34.
52. Posner, supra note 23, at 61 & n.25;
53. cf. Mund, supra note 26, at 12 n.34 ("As late as 1978, even though Judge James Browning, then chief judge of the Ninth Circuit, specifically invited five bankruptcy judges to attend the circuit conference, Senior District Judge Lloyd George (formerly a bankruptcy judge) reports that 'they wouldn't call me "judge." They called me mister.'" (quoting Interview with Lloyd George (Dec. 20, 2004))).
54. Pub. L. No. 91-354, 84 Stat. 468 (1970).
55. Mund, supra note 26, at 7.
56. Id. at 8.
57. Id.
58. See Posner, supra note 23, at 75.
59. Bankruptcy Reform Act of 1978, Pub. L. No. 95-598, 92 Stat. 2549 (codified as amended primarily at 11 U.S.C. §§ 101-1532 and in scattered sections of 28 U.S.C.).
60. See Countryman, supra note 28, at 7-8.
61. See id. at 8-9; Posner, supra note 23, at 77 ("The federal judges opposed the creation of more independent bankruptcy courts, because (1) they would lose their appointment power over bankruptcy judges, and thus one of their main patronage opportunities, and (2) their status would be diluted through the vast increase in the number of federal judicial positions."). Interestingly, as Congress considered various proposals for reorganizing the court structure of the bankruptcy system in its reform efforts from the 1970s that led to enactment of the Bankruptcy Code, bankruptcy judges did not seek Article III status. Instead, they lobbied Congress for appointment by the judicial council, rather than the president, for two reasons: First, they believed their merit would be properly recognized in a nonpolitical judicial appointment process; and, second, they feared that sitting judges would lack the political connections necessary for presidential appointment. See Mund, supra note 26, at 20-21, 24-25, 29. For a political economic analysis of the 1978 Act's treatment of bankruptcy judges, see Posner, supra note 23, at 74-94.
62. See 28 U.S.C. § 1471(b), (c) (Supp. II 1978), invalidated by N. Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U.S. 50 (1982).
63. 458 U.S. at 87.
64. Countryman, supra note 28, at 31.
65. 28 U.S.C. § 152(a)(1) (Supp. V 2005).
66. Id. § 151 (2000);
67. see also id. § 152(a)(1) (Supp. V 2005) ("Bankruptcy judges shall serve as judicial officers of the United States district court established under Article III of the Constitution.") .
68. Id. § 157(a) (2000).
69. 9 AM. JUR. 2D Bankruptcy § 731 (2008);
70. Bussel, supra note 13, at 1066 & n.12.
71. Disputes in bankruptcy cases generally assume one of two forms: (1) an adversary proceeding, or (2) a contested matter. Adversary proceedings include, for example, a proceeding to recover money or property; a proceeding to determine the validity, priority, or extent of a lien; a proceeding to object to or revoke a discharge; and a proceeding to determine the dischargeability of a debt. FED. R. BANKR. P. 7001. Such proceedings are initiated and advance much as any other federal lawsuit, insofar as Part VII of the Federal Rules of Bankruptcy Procedure, which governs such proceedings, virtually incorporates the Federal Rules of Civil Procedure (occasionally with modification). See, e.g., id. 7003 (FED. R. CIV. P. 3);
72. id. 7004(a) (portions of FED. R. CRV. P. 4);
73. id. 7005 (FED. R. CIV. P. 5);
74. id. 7012(b) (FED. R. CIV. P. 12(b)-(h));
75. id. 7013 (FED. R. CIV. P. 13);
76. id. 7014 (FED. R. CIV. P. 14);
77. id. 7056 (FED. R. CIV. P. 56). Disputes between parties that are not adversary proceedings are called "contested matters," and they proceed according to less complex procedures than adversary proceedings-including request for relief by motion rather than the filing of a complaint. FED. R. BANKR. P. 9014;
78. see also Khachikyan v. Hahn (In re Khachikyan), 335 B.R. 121, 125 (B.A.P. 9th Cir. 2005) ("In a contested matter, there is no summons and complaint, pleading rules are relaxed, counterclaims and third-party practice do not apply, and much pre-trial procedure is either foreshortened or dispensed with in the interest of time .") .
79. The Judicial Code describes a non-core proceeding as "a proceeding that is not a core proceeding but is otherwise related to a case under title 11." 28 U.S.C. § 157(c)(1) (2000 & Supp. V 2005).
80. 28 U.S.C. §§ 1291, 1292 (2000).
81. Id. § 1254(1).
82. Section 157(b)(1) of the Judicial Code speaks of "core proceedings arising under title 11, or arising in a case under title 11." Id. § 157(b)(1). In turn, section 157(b)(2) lists examples of core proceedings, which include matters concerning (1) administration of the estate, (2) the allowance of claims, (3) objections to discharge, and (4) plan confirmation. Id. § 157(b)(2) (2000 & Supp. V 2005).
83. Id. § 157(b)(1) (2000). Unless, that is, the district court withdraws the reference to the bankruptcy court. Id. § 157(d). In that case, the district court hears the matter in the first instance, with appeals in the ordinary course lying to the court of appeals and then the Supreme Court. See supra notes 47-48.
84. 28 U.S.C. § 158(b)(1). The statute also authorizes the creation of intercircuit BAPs, id. § 158(b)(4), but none has yet been created. Much as the bankruptcy court is a unit of the district court, the bankruptcy appellate panel may be seen as "a unit of the federal courts of appeals." Admin. Office of the U.S. Courts, The Federal Judiciary-United States Courts of Appeals, Bankruptcy Appellate Panels, http://www.uscourts.gov/courtsofappeals/bap.html (last visited Oct. 11, 2008);
85. see also 28 U.S.C. § 158(b)(1) (requiring BAPs to be established and BAP judges to be appointed by the circuit judicial council); B.A.P. 8TH CLR. R. 8016A(a)(l) ("The Clerk of the United States Court of Appeals for the Eighth Circuit shall serve as the Clerk of the United States Bankruptcy Appellate Panel for the Eighth Circuit."). Compare Coyne v. Westinghouse Credit Corp. (In re Globe Illumination Co.), 149 B.R. 614, 620-21 (Bankr. CD. Cal. 1993) (describing BAP as unit of the circuit court), with Kathleen P. March & Rigoberto V. Obregon, Are BAP Decisions Binding on Any Court?, 18 CAL. BANKR. J. 189, 197 (1990) (describing BAP as unit of the district court) .
86. The 1994 amendments to the Bankruptcy Code were designed to encourage circuit courts to create BAPs by directing that each circuit "shall establish" a BAP unless the circuit judicial council finds that existing judicial resources are insufficient to establish one or that its establishment would result in undue delay or increased cost to parties in cases under the Bankruptcy Code. 28 U.S.C. § 158(b)(1). The six regional circuits that voted against establishing BAPs "concluded that the appellate process was functioning well as already constituted and that BAPs would create undue delay or increase the cost of appeals." Henry J. Boroff, The Precedential Effect of Bankruptcy Appellate Panel Decisions, 103 COM. L.J. 212, 214 n.10 (1998) (citing Elizabeth Abbott, Bankruptcy Review Panel Makes Debut, NAT'L L.J., Mar. 3, 1997, at B1).
87. For a historical discussion of BAPs, see Bryan T. Camp, Bound by the BAP: The Stare Decisis Effects of BAP Decisions, 34 SAN DIEGO L. REV. 1643, 1648-60 (1997);
88. infra note 75.
89. 28 U.S.C. § 158(b)(6). In the mid-1990s, when a Second Circuit BAP was in existence, "only three districts participate[d]-and these together typically receive[d] less than a third of all bankruptcy petitions filed in the Second Circuit." Camp, supra note 52, at 1660. These facts, presumably, played a large role in the ultimate decision to disband the Second Circuit BAP.
90. 28 U.S.C. § 158(c)(1) (2000 & Supp. V 2005).
91. Id. § 158(d)(1) (Supp. V 2005).
92. Id. § 1254(1) (2000).
93. Id. § 158(c)(1) (2000 & Supp. V 2005).
94. Id. § 158(d)(1) (Supp. V 2005).
95. Id. § 1254(1) (2000).
96. See generally Bernard Trujillo, Self-Organizing Legal Systems: Precedent and Variation in Bankruptcy, 2004 UTAH L. REV. 483, 490-500 (elucidating the differences between the standard federal judicial hierarchy and the bankruptcy appellate system).
97. We should note that a third possible appellate path not yet discussed-that of direct appeal from the bankruptcy court to the court of appeals-exists for a limited set of circumstances. By virtue of amendment to the Judicial Code by the Bankruptcy Abuse Prevention and Consumer Protection Act of 2005, Pub. L. No. 109-8, 119 Stat. 23, appeal may proceed directly to the court of appeals pursuant to a certification procedure if one of the following circumstances exists: (1) the appeal involves a question of law unresolved by the court of appeals for the circuit or by the Supreme Court; (2) the appeal involves a matter of public importance; (3) the appeal involves a question of law requiring resolution of conflicting decisions; or (4) the appeal may materially advance the progress of the case or proceeding in which the appeal is taken. 28 U.S.C. § 158(d)(2)(A) (Supp. V 2005).
98. Ralph R. Mabey, The Evolving Bankruptcy Bench: How Are the "Units"Faring?, 47 B.C. L. REV. 105, 123 (2005).
99. See, e.g., id. at 107 ("Most of the bankruptcy judges were bankruptcy practitioners in their prior careers.");
100. see also id. at 123 (noting that, of a random survey of bankruptcy judges in 2005, "[a]bout 83% . were bankruptcy practitioners before taking the bankruptcy bench," and that, "[o]f the 17% . who were not bankruptcy practitioners, almost all came from a business law background, as commercial litigators or corporate transactional lawyers," and further noting that the surveyed bankruptcy judges felt that their prior experience was very helpful on the bench);
101. cf. id. at 113-16 (discussing the trend among bankruptcy judges to hire more permanent, as opposed to term, law clerks, and noting that those bankruptcy judges who preferred permanent clerks often hired clerks with legal experience, and in particular practice experience in bankruptcy law).
102. See Chemerinsky, supra note 4, at 128 ("[T]he BAP is desirable because it allows specialist bankruptcy judges to replace nonspecialist federal district court judges.");
103. see also Jeffrey J. Rachlinski, Chris Guthrie & Andrew J. Wistrich, Inside the Bankruptcy Judge's Mind, 86 B.U. L. REV. 1227, 1230-31 (2006) (reporting empirical finding that bankruptcy judges as specialized actors perform "at least as well" as generalist judges in terms of not exhibiting typical biases often reflected in judgments).
104. One might argue that even district judges with no experience in bankruptcy before ascending to the bench gain some experience by virtue of hearing a steady stream of bankruptcy cases. A study by the Federal Judicial Center of the bankruptcy appellate structure, however, reached the opposite conclusion, observing that "[t]he number of first-level reviewers greatly exceeds the number of bankruptcy judges producing the judgments reviewed, and appellate caseloads are spread thinly among district judges, giving few judges much opportunity to develop bankruptcy expertise." Judith A. McKenna & Elizabeth C. Wiggins, Alternative Structures for Bankruptcy Appeals, 76 AM. BANKR. L.J. 625, 627 (2002).
105. We employ the modifier "other" because, as noted above, the Court suggested that the use of multi-judge panels is "[p]erhaps most important" in assessing lawfinding ability. Supra note 7 and accompanying text.
106. It is this factor that, presumably, vests district judges with lawfinding ability when they sit by designation on court of appeals panels. See Nash, supra note 1, at 1031 (explaining that the better term is lawfinding "ability" and not lawfinding "expertise"). One might argue that lawfinding ability is enhanced to the extent that the judge (whether district or bankruptcy) enjoys relief from her other responsibilities while hearing appeals. This seems not to be the case, however, at least for bankruptcy judges: When asked how BAP service affects their service as a bankruptcy judge, several of the [surveyed bankruptcy judges] indicated that it required adjustments to their bankruptcy court trial and hearing schedule and that it substantially added to their workload. Some of the Survey Participants suggested that those bankruptcy judges who serve full-time on the BAP should have the option of employing an additional law clerk. One Survey Participant indicated that service on the BAP was "like having a second job." Mabey, supra note 62, at 122 (footnote omitted);
107. see also Stephen A. Stripp, An Analysis of the Role of the Bankruptcy Judge and the Use of Judicial Time, 23 SETON HALL L. REV. 1329, 1330 (1993) ("The fundamental truth which is the basis for this article is that the bankruptcy caseload in many districts in this country is so overwhelming that the bankruptcy judges are sorely pressed in the struggle to cope with it.").
108. BAPs in three circuits-the Eighth, Ninth, and Tenth-have reached this conclusion. E.g., Concannon v. Imperial Cap. Bank (In re Concannon), 338 B.R. 90, 95 (B.A.P. 9th Cir. 2006) (reaffirming that the BAP will not overrule its prior rulings unless an intervening circuit court or Supreme Court decision, or subsequent legislation, undermines those rulings); Salomon N. Am. v. Knupfer (In re Wind N' Wave), 328 B.R. 176, 181 (B.A.P. 9th Cir. 2005) (same);
109. Blagg v. Miller (In re Blagg), 223 B.R. 795, 804 (B.A.P. 10th Cir. 1998) ("Our decision is dictated by the principle that we are bound by prior panel decisions. A panel cannot overrule the judgment of another panel of the court."), appeal dismissed, 198 F.3d 257 (10th Cir. 1999);
110. Smolen v. Hatley (In re Hatley), 227 B.R. 757, 761 (B.A.P. 10th Cir. 1998) (same), aff'd, 194 F.3d 1320 (10th Cir. 1999);
111. Luedtke v. Nationsbanc Mortgage Co. (In re Luedtke), 215 B.R. 390, 391 (B.A.P. 8th Cir. 1997) (relying on circuit court precedent that circuit court panel decisions bind subsequent circuit court panels to announce rule that BAP decisions bind subsequent BAP panels); Ball v. Payco-Gen. Am. Credits, Inc. (In re Ball), 185 B.R. 595, 597 (B.A.P. 9th Cir. 1995) ("We will not overrule our prior rulings unless a Ninth Circuit Court of Appeals decision, Supreme Court decision or subsequent legislation has undermined those rulings.").
112. Philadelphia Life Ins. Co. v. Proudfoot (In re Proudfoot), 144 B.R. 876, 879 (B.A.P. 9th Cir. 1992) ("[B]AP decisions originating in any district in the Ninth Circuit are binding precedent on all bankruptcy courts within the Ninth Circuit in the absence of contrary authority from the district court for the district in which the bankruptcy court sits."); In re Windmill Farms, Inc., 70 B.R. 618, 622 (B.A.P. 9th Cir. 1987), rev'd on other grounds, 841 F.2d 1467 (9th Cir. 1988).
113. Compare, e.g., Ore. Higher Educ. Assistance Found, v. Selden (In re Selden), 121 B.R. 59, 62 (D. Ore. 1990) (stating that BAP decisions bind only those bankruptcy courts sitting in the district out of which the appeal arose),
114. with Daly v. Deptula (In re Carrozzella & Richardson), 255 B.R. 267, 273 (Bankr. D. Conn. 2000) (rejecting argument that substantial motivation of Congress in creating BAPs was to generate a uniform body of bankruptcy law within the circuits, concluding that there is no principled reason why decisions of a BAP should have more precedential authority than those of district courts, and finding it odd and unseemly, if not unconstitutional, for a BAP-comprised of three non-Article III judges-to be generating for bankruptcy judges, and perhaps also for district judges, the law of the circuit until the circuit court had spoken), In re Virden, 279 B.R. 401, 409 n.12 (Bankr. D. Mass. 2002) (quoting In re Carrozzella, 255 B.R. at 272-73),
115. and Life Ins. Co. of Va. v. Barakat (In re Barakat), 173 B.R. 672, 676-80 (Bankr. CD. Cal. 1994) (concluding that BAPs bind bankruptcy courts on matters arising in core proceedings even though district courts do not), aff'd on other grounds, 99 F.3d 1520 (9th Cir. 1996). For further discussion regarding the precedential effect of BAP decisions, see Salomon N. Am., 328 B.R. at 181 n.2 (noting the Ninth Circuit BAP's prior holding that its decisions bind bankruptcy courts within the circuit, but also recognizing that some bankruptcy courts have rejected that holding);
116. Zimmer v. PSB Lending Corp. (In re Zimmer), 313 F.3d 1220, 1225 n.3 (9th Cir. 2002) (describing "binding nature of Bankruptcy Appellate Panel decisions" as "an open question," and "join[ing] Judge O'Scannlain's call for the [Ninth Circuit] Judicial Council to consider an order clarifying whether the bankruptcy courts must follow the BAP"); Bank of Maui v. Estate Analysis, Inc., 904 F.2d 470, 472 (9th Cir. 1989) ("BAP decisions cannot bind the district courts themselves. As article III courts, the district courts must always be free to decline to follow BAP decisions and to formulate their own rules within their jurisdiction.");
117. id. at 472 (O'Scannlain, J., concurring) (writing "separately to propose that the Judicial Council of this Circuit consider adoption of an order requiring that Bankruptcy Appellate Panel. decisions shall bind all of the bankruptcy courts of the circuit, subject to the restrictions imposed by article III so well discussed in the [court's] opinion");
118. Paul M. Baisier & David G. Epstein, Resolving Still Unresolved Issues of Bankruptcy Law: A Fence or an Ambulance, 69 AM. BANKR. L.J. 525, 531 (1995) ("Even stronger arguments can be made against any stare decisis effect at all for the opinion of a bankruptcy appellate panel.");
119. Chemerinsky, supra note 4, at 129-30 ("I would argue that district courts should be bound by BAP decisions. The view that an Article I court can never bind an Article III court is an overstatement.");
120. Trujillo, supra note 60, at 494 n.23 (arguing that BAPs function as district courts, and accordingly cannot issue binding opinions).
121. See Baisier & Epstein, supra note 70, at 529 (noting that "[n]one of the district judges is bound by a bankruptcy appeals decision of a district judge from one of the other 93 district courts," and that "district judges in multi-judge districts are not even bound by the bankruptcy appeals decisions of other judges from that same district").
122. But see Bussel, supra note 13, at 1095-96;
123. id. at 1096 n.116 ("I am aware of only a handful of cases where district judges in the same district adopt differing views of the same
124. See, e.g., In re Romano, 350 B.R. 276, 281 (Bankr. E.D. La. 2005) ("[A] single decision of a district court in this multi-judge district is not binding upon this court.");
125. id. at 277-81 (summarizing authority for both sides); Paul Steven Singerman & Paul A. Avron, Of Precedents and Bankruptcy Court Independence, 22 AM. BANKR. INST. J. 1, 56-57 (2003) (noting conflict, gathering authorities, and finding that a majority of bankruptcy courts have held that they are not bound by the decision of a single district court judge in a multi-judge district);
126. Trujillo, supra note 60, at 494 (arguing that a bankruptcy decision by one bankruptcy judge cannot bind other bankruptcy judges in the same district, and that a bankruptcy decision by one district judge cannot bind other district judges or bankruptcy judges in the same district).
127. But see Chemerinsky, supra note 4, at 129 ("While a district court exercising original jurisdiction cannot bind other district courts, its decisions should be binding on bankruptcy courts when the district court is serving as an appeals court.").
128. Our point here is simply that BAPs seem to fit more cleanly into the standard hierarchical appellate model than do district courts sitting on appeal, not that that is necessarily mandated under the current statutory scheme or normatively desirable. The latter two points are debatable. With respect to the current statutory scheme, there are statements in the legislative history indicating that Congress created the BAPs to help foster greater uniformity in bankruptcy law. See, e.g., 140 CONG. REC. S14.463 (daily ed. Oct. 6, 1994) (statement of Sen. Heflin) ("It should be recognized that the creation of a bankruptcy appellate panel service can help to establish a dependable body of bankruptcy case law.").
129. But see Daly, 255 B.R. at 273: Any suggestion that Congress' authorization of the creation of BAP Services was motivated substantially by its desire to create a uniform body of bankruptcy law within the circuits is not supported by the BAP Service's history, which instead suggests that BAPs were conceived primarily as a tool for relieving district court judges of an ofttimes undesirable and burdensome aspect of their workload. At the same time, one can point to the certification procedure in section 158(d)(2) of the Judicial Code-under which courts of appeals may decide interlocutory appeals when, among other circumstances, the question raised is one "as to which there is no controlling decision of the court of appeals for the circuit or of the Supreme Court of the United States," 28 U.S.C. § 158(d)(2)(A)(i) (Supp. V 2005)-as evidence that Congress chose other, explicit means of increasing bankruptcy law uniformity.
130. See H.R. REP. No. 109-31, at 148 (2005), reprinted in 2005 U.S.C.C.A.N. 88, 206.
131. Commentators are divided over whether BAP decisions bind bankruptcy courts. Compare, e.g., Bussel, supra note 13, at 1098 (arguing that bankruptcy courts should consider both BAP and district court decisions as binding precedent),
132. Chemerinsky, supra note 4, at 128 ("From [a] functional perspective, I think that BAP decisions clearly should be binding on bankruptcy courts."),
133. and Camp, supra note 52, at 1676-84 (arguing that BAPs should bind both bankruptcy and district courts within a circuit),
134. with Trujillo, supra note 60, at 492 ("[O]nly opinions of the U.S. courts of appeals and the U.S. Supreme Court bind bankruptcy courts by reason of formal hierarchy."),
135. and Caminker, supra note 13, at 870-72 (arguing that theoretical considerations argue in favor of bankruptcy courts being bound by district court decisions). Moreover, strict application of vertical stare decisis is difficult, insofar as it is not certain until after the bankruptcy court has issued judgment into which appellate path the case will proceed.
136. Cf. Camp, supra note 52, at 1682: Since bankruptcy judges do not know at the time they make a decision whether it will be a BAP or a district court that will hear any appeal, and since no district court has so far considered itself bound by a BAP, it is no surprise that many bankruptcy judges feel free to disregard BAP decisions. Compare this to the United States Tax Court, which considers itself bound by its own precedent, except insofar as it has also held that it is bound "to follow a Court of Appeals decision which is squarely in point where appeal from [the] decision lies to that Court of Appeals." Golsen v. Comm'r, 54 T.C. 742, 756-57 (1970),
137. aff'd, 445 F.2d 985 (10th Cir. 1971). Because the court of appeals to which a taxpayer will appeal is determined by his state of residence, 26 U.S.C. § 7482(b)(1) (2000), it is always clear at the time of decision which circuit's precedent is binding. As to the normative question, there are those who argue that an increase in application of stare decisis would be normatively desirable.
138. See, e.g., Boroff, supra note 52, at 215, 221 (arguing that the current dual track appellate system makes it difficult to generate binding precedent, and that the system be changed to allow for development of binding precedent);
139. Bussel, supra note 13, at 1095 n.114 ("[L]ogically district courts . as well as bankruptcy courts might be bound by prior BAP decisions."). There also are strong arguments, however, that a structure other than the standard appellate hierarchy might be desirable. First, one of the bases on which the pyramidal appellate hierarchy functions is the notion that issues "percolate" up from the lower courts to the higher courts. It is the desire for percolation that, commentators argue, restricts (and properly so) application of horizontal stare decisis to the same court and not to sibling courts of equal hierarchical stature.
140. See Samuel Estreicher & Richard L. Revesz, The Uneasy Case Against Intracircuit Nonacquiescence: A Reply, 99 YALE L.J. 831, 834 (1990) ("The rejection of intercircuit stare decisis is premised upon-and given the obvious costs in deferring uniformity, is explainable only in terms of-the benefits of dialogue among the circuits.");
141. see also Maxwell Stearns, Standing Back from the Forest: Justiciability and Social Choice, 83 CAL. L. REV. 1309, 1351-52 (1995) (arguing, based upon social choice theory, that the Supreme Court "would desire intra- but not inter-circuit stare decisis," since such a regime "avoids the irrationality that would result from cyclical preferences within particular circuits, while, at the same time, reducing the likelihood that legal doctrine that results from path manipulation in a given circuit will be replicated across the circuits.").
142. But cf. O'Hara, supra note 17, at 772 (arguing that the absence of stare decisis across circuits is justified on the ground that "an agreement to follow another circuit's precedents will not save the judges in a particular circuit much time"). In the case of appeals of core bankruptcy matters, there are, anomalously, two levels of intermediate appeals. Perhaps, then, in order for issues properly to percolate up to the courts of appeals, there ought to be no horizontal stare decisis at the first intermediate level- that is, at the level of the BAPs and district courts. Second, given that the BAPs and district courts lie at the same hierarchical level, it might not make sense for horizontal stare decisis rules to apply to BAPs but not district courts. Perhaps, once again, horizontal stare decisis should not apply at all. One might argue, to the contrary, that horizontal stare decisis should apply to both courts.
143. See Chemerinsky, supra note 4, at 129. Third, perhaps bankruptcy law and society would be better served by a system other than the traditional appellate hierarchy, at the lower levels of appeals of core bankruptcy matters. Civil law systems rely far less on precedent than does the common law system dominant in the United States.
144. See supra note 15 and accompanying text. Civil law judiciaries decide cases based largely upon the proper interpretation of the governing "code." Insofar as bankruptcy turns upon the content of a code-the "Bankruptcy Code"-bankruptcy seems to provide an ideal setting for application of such judicial review.
145. Cf. Lawrence Ponoroff, The Dubious Role of Precedent in the Quest for First Principles in the Reform of the Bankruptcy Code: Some Lessons from the Civil Law and Realist Traditions, 74 AM. BANKR. L.J. 173, 216 (2000) (arguing for "a softer, more nimble, rule of precedent [that] would improve the quality of outcomes in particular bankruptcy cases"). Interestingly, while Ponoroff facially argues in favor of increased reliance on a civil law jurisprudential approach in the bankruptcy context, his arguments do not seem to accord so well with the principles underlying the structure of judicial review in civil law systems. Dean Ponoroff laments: The opportunity for two levels of appeal as a matter of right has contributed to the crush of reported decisions, a phenomenon that, in my view, has hampered pragmatic and considered decisionmaking in the bankruptcy courts. That problem is compounded by the disturbing rise in adherence to textual or plain meaning methods of interpretation in bankruptcy cases, particularly in the decisions of the circuit courts of appeal [s].
146. Id. at 181 (footnotes omitted). Ponoroff thus seems more concerned with allowing different interpretations of the Bankruptcy Code to percolate up through the judiciary. He also seems to embrace more of a realist conception of bankruptcy law than a civil law conception, explaining that "[a] more forward-looking, and less technical and "busy,' code would abate the pressure to decide and review cases on the kind of formal, textualist grounds that typically prove the most difficult to distinguish in subsequent cases."
147. Id. at 216. Indeed, Ponoroff acknowledges that he endorses "a different style of judging, one that eschews a strict adherence to precedent, but not by any means civilian, to the extent that style is perceived as the unimaginative and rote application of positive legal rules to particular fact situations performed by a cadre of mid-level bureaucrats."
148. Id. at 223. "Rather," he endorses "a style that actually places greater responsibility on the decisionmaker to reason analogically from code principles, as well as from subsidiary sources such as custom, usages, settled jurisprudential doctrine, and equity."
149. Id. at 223-24. To the possible objection that the fact that the higher levels of bankruptcy judicial review-courts of appeals and the Supreme Court-rely upon the standard appellate hierarchy, one can point to the coexistence of Louisiana's civil law system within the United States judicial system as an example of how such a system can function.
150. See, e.g., Shelp v. Natl Sur. Corp., 333 F.2d 431, 439 (5th Cir. 1964) (in determining Louisiana law under Erie, federal courts should apply precedential rules that Louisiana's highest court would apply);
151. Alvin Rubin, Hazards of a Civilian Venturer in Federal Court: Travel and Travail on the Erie Railroad, 48 LA. L. REV. 1369 (1988).
152. But see John Burritt McArthur, Good Intentions Gone Bad: The Special No-Deference Erie Rule for Louisiana State Court Decisions, 66 LA. L. REV. 313 (2006). Indeed, the notion that bankruptcy courts do not consider themselves bound by rulings of single district judges in multi-judge districts-and therefore presumably do at some point consider themselves bound once a number of district judges in the same district reach the same conclusion-resembles the "jurisprudence constante" under which precedent develops in Louisiana and other civil law systems.
153. See, e.g., Kathryn Venturatos Lorio, The Louisiana Civil Law Tradition: Archaic or Prophetic in the Twenty-First Century, 63 LA. L. REV. 1, 6 (2002) (describing jurisprudence constante as a doctrine under which "a case may be used to discern a pattern [of decisions] that may aid in interpretation");
154. Stearns, supra, at 1357 n.143 (discussing jurisprudence constante);
155. cf. RICHARD A. POSNER, THE FEDERAL COURTS: CRISIS AND REFORM 257 (1985) (proposing, "as a special rule of stare decisis, the practice that when the first three circuits to decide an issue have decided it the same way, the remaining circuits defer to that decision"). Any potential difficulties in integrating a civil law precedential model into the larger common law-based federal court system would be mitigated by the fact that the vast majority of bankruptcy cases are not appealed beyond the first level of intermediate appellate review.
156. See Bussel, supra note 13, at 1091;
157. cf. Chemerinsky, supra note 4, at 122 (noting that "[bankruptcy law matters seem to fit in between two poles" in that "bankruptcy statutes are filled with ambiguities that require court interpretation," while there also "probably exist particular types of disputes where the law-giving function of the court is less important and alternative dispute resolution would be potentially more efficient").
158. But see Bussel, supra note 13, at 1097 ("I would have difficulty understanding why Congress would intend BAPs and district courts to serve merely as rest-stops on the road to real appellate review.").
159. See supra notes 36-38 and accompanying text.
160. 28 U.S.C. § 152(a)(1) (2000 & Supp. V 2005).
161. See Federal Courts Improvement Act of 1996, Pub. L. No. 104-317, § 303, 110 Stat. 3847, 3852 (providing that, "[w]hen filling vacancies, the court of appeals may consider reappointing incumbent bankruptcy judges under procedures prescribed by regulations issued by the Judicial Conference of the United States").
162. See Mabey, supra note 62, at 107 (noting that, of the 115 bankruptcy judges who left the bench in the decade prior to 2005, only 10 did so as a result of not being reappointed);
163. see also U.S. Court of Appeals for the Ninth Circuit, Bankruptcy Judge Reappointment Regulations § 1(e) (2001), available at http://207.41.19.15/Web/OCELibra.nsf (follow "Bankruptcy" hyperlink; then follow "Bankruptcy Judge Reappointment Regulations" hyperlink) (last visited Oct. 25, 2008) (providing that "[t]he court of appeals shall decide whether or not to reappoint the incumbent [bankruptcy] judge before considering other potentially qualified candidates" (emphasis added)). To the contrary, one might argue that the fact that bankruptcy judges must seek, and generally receive, reappointment, demonstrates the absence of judicial independence.
164. See In re Grabill Corp., 976 F.2d 1126, 1129 (7th Cir. 1992) (Easterbrook, J., dissenting) ("Rhetoric about life tenure notwithstanding, there is no substantial difference between the 14year term to which bankruptcy judges are appointed and service 'during good Behavior' for Article III judges."). Article III judges (other than Supreme Court Justices) whose service on the federal bench terminated between 1983 and 2003 served, on average, twenty-four years.
165. Judith Resnik, Judicial Selection and Democratic Theory: Demand, Supply, and Life Tenure, 26 CARDOZO L. REV. 579, 618 chart 4 (2005).
166. While Congress may reduce the duration of the fixed-term appointment for bankruptcy judges at any point via statute, the constitutionally guaranteed life tenure granted to Article III judges could only be stripped away via constitutional amendment (an exponentially more difficult proposition).
167. See COMM'N ON THE BANKR. LAWS OF THE U.S., REPORT OF THE COMMISSION ON THE BANKRUPTCY LAWS OF THE UNITED STATES, H.R. Doc. No. 93-137, pt. 1, at 95 (1973) (proposing various reforms "to enhance the real and apparent judicial independence of bankruptcy judges," including "[e]xtension of the term of the bankruptcy judges from the present six years to the proposed fifteen years");
168. cf. Nash, supra note 20, at 2196 (observing that "one can question the degree to which life tenure in fact secures for judges a larger measure of judicial independence").
169. 28 U.S.C. § 152(e) (2000). 82. U.S. CONST, art. III, § 1.
170. N. Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U.S. 50, 59 (1982) (Brennan, J., plurality opinion) ("The 'good Behaviour' Clause guarantees that Art. III judges shall enjoy life tenure, subject only to removal by impeachment.").
171. But see Saikrishna Prakash & Steven D. Smith, How to Remove a Federal Judge, 116 YALE. L.J. 72 (2006) (arguing that federal judges may be removed from office by means other than impeachment).
172. 28 U.S.C. § 152(e).
173. See Mabey, supra note 62, at 107 (listing reasons for departure from the bench for the 115 bankruptcy judges who did so in the decade prior to 2005, but not mentioning removal as one of those reasons). On the other hand, one might argue that the low rate of removal of bankruptcy judges reflects the absence of judicial independence: bankruptcy judges have behaved in a way so as to avoid removal.
174. N. Pipeline, 458 U.S. 50 at 59 (Brennan, J., plurality opinion).
175. See Mabey, supra note 62, app. A.
176. 28 U.S.C. § 153(a). Congress amended the Judicial Code in 1987 to provide for the current salary structure for bankruptcy judges. Pub. L. No. 100-202, § 408(a), 101 Stat. 1329, 1329-26 (1987).
177. See Nash, supra note 20, at 2182-92.
178. 28 U.S.C § 152(a)(1) (Supp. V 2005);
179. id. § 152(a)(3) (2000).
180. The possibility exists, however, that the judicial appointment of judges may substitute judicial patronage for political patronage and thus compromise judicial independence. See Judith Resnik, "Uncle Sam Modernizes His Justice": Inventing the Federal District Courts of the Twentieth Century for the District of Columbia and the Nation, 90 GEO. L.J. 607, 673 (2002).
181. But see Posner, supra note 23, at 81-82 ("Appointments by the judicial branch are not as controversial, because judges belong to different parties."). Furthermore, one may argue that, insofar as the circuit judges are a product of a politicized appointment process, they themselves may be politicized and thus infuse politics into the appointment process for bankruptcy judges. The merit-selection process for appointing bankruptcy judges, however, seems to have provided little opportunity for such politicization to take root. A quick look at the manner in which the Ninth Circuit conducts this process (one that seems representative of the process conducted in other circuits) suggests why this has been the case. Interested candidates must submit applications for the position.
182. See Judicial Council of the Ninth Circuit, Regulations Governing the Appointment of U.S. Bankruptcy Judges § 2.02 (2001), available at http://207.41.19.15/Web/OCELibra.nsf (follow "Bankruptcy" hyperlink; then follow "Regulations Governing the Appointment of U.S. Bankruptcy Judges" hyperlink) Oast visited Oct. 25, 2008). The Circuit advertises nationally and encourages the federal judicial districts within the circuit to advertise intensely and locally.
183. Id. § 2.01. A local merit screening committee, which generally consists of (1) the chief judge of the district in which the bankruptcy judge is to be appointed, (2) the president of the state bar association, (3) the president of one or more local bar associations within the district, (4) the dean of a law school located within the district, (5) the administrative circuit judge of the circuit geographical unit in which the bankruptcy judge is to be appointed, and (6) the chief bankruptcy judge of the district in which the bankruptcy judge is to be appointed.
184. Id. § 3.02(a). The committee recommends five applicants to the Court-Council Committee on Bankruptcy Appointments, whose membership includes three circuit judges who serve as voting members
185. . See id. §§ 3.03(c)(1), 3.04(b). The Court-Council Committee circulates a report to the Ninth Circuit Judicial Council recommending a candidate for appointment, and that report will be deemed to be the Judicial Council's recommendation to the Court of Appeals (unless the Council determines that the Court-Council Committee should reconsider its recommendation).
186. Id. §§ 3.04(c)(5), 3.05(a). The recommended candidate is appointed upon a majority vote of the members of the Court of Appeals. 28 U.S.C § 152(a)(3) (2000). For the argument that the nonpolitical nature of the bankruptcy bench may be attributable to the opacity of the process for selecting bankruptcy judges,
187. see Rafael I. Pardo, The Utility of Opacity in Judicial Selection, 64 N.Y.U. ANN. SUEV. AM. L. (forthcoming 2008), available at http://ssrn.com/abstract=1205002.
188. See Mabey, supra note 62, at 107 ("Most of the bankruptcy judges were bankruptcy practitioners in their prior careers.").
189. Cf. Resnik, supra note 91, at 670 ("Turn first to the advantages of judicial appointment of judges. As a few details of current practices illustrate, the judiciary has selected a high-quality and relatively nonpolitical corps of judges ").
190. This state of affairs can be traced to congressional efforts in the 1970s to elevate the status of bankruptcy judges. Congress established in 1970 the Commission on the Bankruptcy Laws of the United States to evaluate the then-existing bankruptcy system and to suggest recommendations for its reform. Pub. L. No. 91-354, 84 Stat. 468 (1970). In its report, the Commission envisioned that improvements in the appointment, tenure, and compensation of bankruptcy judges would enhance their "real and apparent judicial independence." COMM'N ON THE BANKR. LAWS OF THE U.S., REPORT OF THE COMMISSION ON THE BANKRUPTCY LAWS OF THE UNITED STATES, H.R. DOC. NO. 93-137, pt.1, at 95 (1973).
191. Professor Frank Cross has expressed a similar view in his empirical study of decisions rendered by U.S. Courts of Appeals: [T]here are typically nonfrivolous legal arguments for each side in circuit court cases, so it is impossible to code certain cases as being legally correct (or incorrect) without the researcher second-guessing and effectively overriding the judge. Such an approach offers an unreliable tool for evaluating judicial decisions because it probably reflects more about the researcher than about the judges being evaluated. Research requires a more objective tool for evaluating the law. FRANK B. CROSS, DECISION MAKING IN THE U.S. COURTS OF APPEALS 46-47 (2007).
192. But see Wagner & Petherbridge, supra note 5, at 1127-28 (noting the limits of "result-oriented statistical studies");
193. Stephen J. Choi & G. Mitu Gulati, Bias in Judicial Citations: A New Window into the Behavior of Judges? 5 (N.Y.U. Law & Econ. Research Paper Series, Paper No. 06-29, 2006), available at http://ssrn.com/ abstract=913663 (using empirical data to argue that judges of one political party are more likely to cite opinions authored by judges of the same party, especially in particular "high stakes" settings).
194. See supra note 17.
195. See, e.g., Concannon v. Imperial Capital Bank (In re Concannon), 338 B.R. 90, 93 (B.A.P. 9th CIR. 2006) ("[W]e review the bankruptcy court's conclusions of law and interpretation of the Bankruptcy Code de novo.");
196. Official Unsecured Creditors Comm. of Valley-Vulcan Mold Co. v. Ampco-Pittsburgh Corp. (In re Valley-Vulcan Mold Co.), 237 B.R. 322, 326 (B.A.P. 6th Cir. 1999) (stating that conclusions of law by bankruptcy court are reviewed by BAP de novo);
197. CHARLES ALAN WRIGHT & ARTHUR R. MILLER, FEDERAL PRACTICE AND PROCEDURE § 2588, at 470 (2006).
198. It is possible that, notwithstanding the legal standard to the contrary, appellate courts do not always reexamine legal issues de novo in practice. Perhaps, for example, courts of appeals are inclined to rely upon the expertise of BAPs (sub rosa, of course, since the law dictates otherwise) and thus are inclined to affirm BAP opinions. Or, perhaps equally, the appellate courts might more often than not affirm district court opinions on the ground that district judges enjoy Article III status and thus are independent. In either case, it would be accurate to view an appellate court affirmance as embracing the lower court opinion as correct.
199. See O'Hara, supra note 17, at 737-38 (suggesting that judges make decisions "to impose their normative views, beliefs, and mores on [society]");
200. Richard A. Posner, What Do Judges and Justices Maximize? (The Same Thing Everyone Else Does), 3 SUP. CT. ECON. REV. 1, 39 (1993) ("Judges are rational, and they pursue instrumental and consumption goals of the same general kind and in the same general way that private persons do.").
201. See Nash, supra note 20, at 2196. Indeed, Professor Resnik has identified such careerism by bankruptcy judges.
202. See Resnik, supra note 91, at 673 (observing that "[a]n increasingly well-trodden path is for a person to shift from magistrate or bankruptcy judge to district court judge"). A recent study of the bankruptcy bench, however, indicated that only eight of the 115 bankruptcy judges who left the bench from 1995 to 2005 did so as a result of appointment to an Article III judgeship.
203. Mabey, supra note 62, at 107.
204. See, e.g., Caminker, supra note 13, at 827 & n.40 and the authorities cited therein;
205. see also Nash, supra note 20, at 2197-98 (discussing the desire of Article III judges to avoid impeachment, public chastisement, and overruling by the legislature).
206. Note, however, that other motivations may explain bankruptcy judges' behavior. See, e.g., LYNN M. LOPUCKI, COURTING FAILURE: HOW COMPETITION fOR BIG CASES IS CORRUPTING THE BANKRUPTCY COURTS (2005) (arguing that bankruptcy judges in different districts compete for large corporate bankruptcy cases);
207. Marcus Cole, 'Delaware Is Not a State': Are We Witnessing Jurisdictional Competition in Bankruptcy?, 55 VAND. L. REV. 1845, 1890-93 (2002) (arguing that, in order to conform to dominant state culture favorable to corporations, Delaware bankruptcy judges compete for corporate bankruptcy filings).
208. See, e.g., William M. Landes & Richard A. Posner, Legal Precedent: A Theoretical and Empirical Analysis, 19 J.L. & ECON. 249 (1976) (arguing that citation practices are not essentially a matter of taste but rather are systematic and susceptible to empirical study);
209. John Henry Merryman, The Authority of Authority: What the California Supreme Court Cited in 1950, 6 STAN. L. REV. 613 (1954);
210. John Henry Merryman, Toward a Theory of Citations: An Empirical Study of the Citation Practice of the California Supreme Court in 1950, 1960, and 1970, 50 s. CAL. L. REV. 381 (1977) [hereinafter Merryman, Toward a Theory of Citations];
211. cf. William M. Landes, Lawrence Lessig & Michael E. Solimine, Judicial Influence: A Citation Analysis of Federal Courts of Appeals Judges, 27 J. LEGAL STUD. 271, 271-76 (1998) (noting that "[c]itations are at best a crude and rough proxy for measuring influence," and identifying potential drawbacks and limitations to empirical analyses of judicial citations).
212. See Landes & Posner, supra note 104, at 251 & n.3 (excluding from citation study "citations indicating rejection of the cited case as a precedent"). Our study also includes only positive citations.
213. But see Landes et al., supra note 104, at 273 (deciding "not [to] distinguish . between favorable, critical, or distinguishing citations" insofar as "[c]ritical citations, in particular to opinions outside the citing circuit, are also a gauge of influence since it is easier to ignore an unimportant decision than to spell out reasons for not following it").
214. See Landes & Posner, supra note 104, at 251 (excluding from citation study nonprecedential citations).
215. David J. Walsh, On the Meaning and Pattern of Legal Citations: Evidence from State Wrongful Discharge Precedent Cases, 31 LAW & SOC'YREV. 337, 340-41 (1997) (stating that cases cited for their influential nature will depend more on the quality of the decisions than the stature of the cited court).
216. See id. at 339 (suggesting that citations may "indicate intercourt communication and influence on judicial decisionmaking");
217. cf. McKenna & Wiggins, supra note 65, at 651 ("The availability of published opinions is generally thought to be an important aspect of the appellate process because written opinions provide guidance to judges and litigants by explaining the reasons for the appellate decision.").
218. Walsh, supra note 107, at 339.
219. Cf. id. at 342 (distinguishing between "strong" and "weak" citations).
220. See McKenna & Wiggins, supra note 65, at 628 ("Bankruptcy appellate panel (BAP) judges provide specialized bankruptcy expertise that their bankruptcy colleagues . value highly as a source of authority.").
221. See id. at 678 ("Circuit judges, on average, have less specialized knowledge than bankruptcy judges, particularly those selected to serve on BAPs.").
222. It is also conceivable that courts of appeals in circuits that have BAPs are somewhat favorably inclined to cite to those BAPs, to the extent that they consider the BAPs to be adjuncts of the courts of appeals. See supra note 51.
223. See supra note 37 and accompanying text.
224. See supra fig. 1.
225. Reported case law documents are those released for publication in West Federal Reporters.
226. Coverage for the FBKR-CS database begins with the year 1789.
227. The preliminary field for case law documents (i.e., decisions or orders issued by a court) in Westlaw is found at the top of such documents and generally contains the name of the court that issued the document. In its entirety, the first search query was as follows: "11 u.s.c." & pr ("district court" "bankruptcy appellate panel" % "court of appeals") & da (aft 9/30/1997 & bef 10/01/2000).
228. In its entirety, the second search query was as follows: "11 u.s.c." & pr ("court of appeals") & da (aft 9/30/1997 & bef 10/01/2000).
229. Each search query produced a numbered result list in which the opinions were listed in reverse chronological order. For the first-tier database, the results were organized by court in reverse chronological order (i.e., district court opinions were listed first in reverse chronological order followed by BAP opinions listed in reverse chronological order). We used a random number generator to select the opinions from each result list that we would analyze. For each result list, we randomly generated a set of unique numbers falling within the range of the total documents retrieved by the search query.
230. See supra notes 49-50 and accompanying text.
231. We tailored our search in this manner for two reasons. First, we wanted to facilitate comparisons of our data with official government data regarding bankruptcy appeals. Generally, such data track the government's fiscal year, which begins on October 1st and ends on September 30th, rather than the calendar year. Second, we chose the specific time period for this study in order to capture the BAP experience at its apex in terms of participating circuits. BAPs did not become a fixture of the bankruptcy judicial system until 1996. The enactment of the Bankruptcy Code in 1978 amended the Judicial Code to permit, but not require, the establishment of BAPs on a circuit-by-circuit basis. Only the First and Ninth Circuits chose to do so, establishing their BAPs in 1979 and 1980, respectively. In the wake of the Marathon decision, however, the First Circuit concluded that continued operation of a BAP would be inappropriate until Congress remedied the defects in the constitutionally infirm, bankruptcy jurisdictional scheme. Massachusetts v. Dartmouth House Nursing Home, Inc., 726 F.2d 26 (1st Cir. 1984). The Ninth Circuit reached the opposite conclusion in
232. Briney v. Burley (In re Burley), 738 F.2d 981 (9th Cir. 1984), holding that circuit court supervision of the BAP satisfied Marathon's requirement of Article III judicial review. Despite the measures taken by Congress in 1984 through the Bankruptcy Amendments and Federal Judgeship Act to address the Marathon decision, the First Circuit Judicial Council chose not to reauthorize its BAP, thus leaving the Ninth Circuit as the only circuit with an operating BAP. This state of affairs changed with the Bankruptcy Reform Act of 1994, which amended the Judicial Code to require the judicial council of each circuit to create a BAP absent a finding by the council that (1) insufficient judicial resources in the circuit would preclude its establishment, or (2) that establishment of a BAP would produce undue delay or increased cost to parties in bankruptcy cases. Pub. L. No. 103-394, § 104(c)(3), 108 Stat. 4106, 4109 (codified at 28 U.S.C. § 158(b)(1)). Prompted into action by this amendment, in 1996 the First Circuit reauthorized and the Second, Sixth, Eighth, and Tenth Circuits established BAPs. The Second Circuit BAP, however, ceased operations on July 1, 2000.
233. Thus, our selection criteria reduced the random sample of documents relating to the first-tier database by approximately 18%.
234. Similar to the first-tier database, see supra note 123, our selection criteria reduced the random sample of documents relating to the second-tier database by approximately 17%.
235. See infra Appendix tbl.1.
236. See, e.g., George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD, 1 (1984).
237. According to a recent empirical study, approximately 2% of federal civil lawsuits in 2002 ended in trial. Marc Galanter, The Vanishing Trial: An Examination of Trials and Related Matters in Federal and State Courts, 1 J. EMPIRICAL LEGAL STUD. 459, 463 tbl.1 (2004). The bankruptcy analogue of a federal civil lawsuit is an adversary proceeding. See supra note 45. In 2002, approximately 5% of adversary proceedings terminated during or after trial. Elizabeth Warren, Vanishing Trials: The New Age of American Law, 79 AM. BANKR. L.J. 915, 930 (2005).
238. Even if the group of appeals resolved by the circuit courts are nonrandom such that our results would not hold if the circuit courts also decided those cases that were not appealed beyond the first level of intermediate review, such theorizing is an exercise in futility. Simply put, we cannot measure the outcomes of circuit court decisions that do not exist. In other words, because we look to measure quality of appellate review that the circuit court perceives, we ought not to fret about those cases that will never see light of day in the circuit court.
239. CROSS, supra note 95, at 123-47.
240. Because the Judicial Code mandates that, in circuits with BAPs, bankruptcy appeals will be heard by the BAP unless one of the parties to the appeal elects to have the district court hear the appeal, 28 U.S.C. § 158(c)(1) (2000 & Supp. V 2005), the dynamic of any potential selection bias at work in the BAP perhaps should be understood as the product of the subset of appeals where the forum preferences of the parties to the appeal have aligned. Although there could be instances where all parties prefer to have the appeal heard by the district court, there would also be instances where only one party had such a preference. Thus, a BAP docket is unique in that all of its appeals theoretically involve litigants with a consistent forum preference. We say "theoretically" since it is conceivable that a party with an inconsistent forum preference may have failed, in a timely fashion, to elect a district court to hear the appeal.
241. Because we do not differentiate between district courts from circuits with BAPs (BAP circuits) and those from circuits without BAPs (non-BAP circuits) in our analyses, there is concern that any potential selection bias at work in BAP circuits could be masked by those observations from non-BAP circuits. Approximately 31% of the observations in the first-tier database and 36% of the observations in the second-tier database consisted of district court opinions from non-BAP circuits.
242. See infra Appendix tbl.1. We conducted bivariate statistical analyses to ascertain whether selection bias existed in the BAP circuits by focusing on those circumstances in which one would expect to see such bias have a disproportionate effectnamely, (1) the subject matter of the appeal and (2) affirmance rates by the court of appeals. For neither of these circumstances did we find evidence of selection bias. First, we examined whether a statistically significant relationship exists in BAP circuits between the subject matter of the appeal and the first-tier appellate court to hear the appeal. To do so, we classified observations according to whether the subject of the appeal fell into one of the four most frequently occurring subjects of appeal heard by first-tier appellate courts. For the first-tier database, for all observations, and for those observations from BAP circuits, the four most frequently occurring subjects were matters relating to discharge, procedure/jurisdiction, avoiding powers, and multiple subjects. For the second-tier database, for all observations and for those observations from BAP circuits, the four most frequently occurring subjects were matters relating to discharge, claims, avoiding powers, and multiple subjects. For the first-tier database, approximately 56% of the appeals heard by district courts in BAP circuits as well as all district courts combined involve one of the four most frequently occurring subjects. For the second-tier database, approximately 64% of the appeals heard by district courts in BAP circuits and 59% of the appeals heard by all district courts combined involve one of the four most frequently occurring subjects. After applying a chi-square test with one degree of freedom, we found no statistically significant relationship in BAP circuits between the subject matter of the appeal and the first-tier appellate court to hear the appeal (a p-value of 0.288 for the first-tier database and a p-value of 0.876 for the second-tier database). Second, for all observations in the first-tier database, we examine whether a statistically significant relationship exists between the subject matter of the appeal and whether there is a subsequent appeal to the circuit court. Again, we classify observations according to whether the subject of the appeal fell into one of the four most frequently occurring subject matter categories. For those observations involving subsequent appeal to the circuit court, approximately 62% involved a top subject matter category. For those observations without circuit court review, approximately 56% involved a top subject matter category. Applying a chi-square test with one degree of freedom, we found no statistically significant relationship (p = 0.475) between the subject matter of the appeal and subsequent appeal to the circuit court. Finally, we found that courts of appeals affirm district courts in BAP circuits at a rate similar to that of their counterparts in non-BAP circuits. In the first-tier database, there were 77 observations for which there was a subsequent appeal to the court of appeals. The subsequent history for 10 of those observations, however, already existed in the second-tier database (i.e., the circuit court opinion from the second-tier database reviewed an opinion from the first-tier database). When combining the 67 unique observations from the first-tier database involving a subsequent appeal to the court of appeals with the 170 observations in the second-tier database, courts of appeals partly or fully affirmed district courts in BAP circuits approximately 77% of the time and partly or fully affirmed district courts in non-BAP circuits approximately 71% of the time. Bivariate analysis confirms that no statistically significant difference exists between the rate at which courts of appeals partly or fully affirmed district courts from BAP circuits and district courts from non-BAP circuits (chi-squared = 0.9094, df = 1, p = 0.340). Furthermore, if one considers only those observations where the court of appeals fully affirmed district courts, the affirmance rates in BAP and non-BAP circuits are similar. The circuit court affirmance rate of district court dispositions in BAP circuits was approximately 67% in comparison to 65% in non-BAP circuits. Bivariate analysis confirms that this difference is not statistically significant (chi-squared = 0.1419, df = 1,p = 0.706).
243. With respect to citations, if there is a forum-selection bias, then the BAPs are not deciding some categories of cases-and, perhaps, certain issues-that the district courts are. This logically should translate into an increase in citations to district court opinions as compared to BAP opinions, since other courts facing such issues and wishing to include citations will have no opportunity to cite to any BAP opinions. Yet, as we discuss below, our data analyses generally show that BAP citations are favored. In short, if there is a selection bias, then our statistical analyses, if anything, understate the extent to which BAP citations are favored.
244. For the frequency of the dispositions rendered on appeal in first-tier and second-tier level opinions, see infra Appendix tbl.2.
245. As set forth below, we denominate this variable "Affirmance." See infra text accompanying notes 150-151.
246. As set forth below, we denominate this variable "Full Affirmance." See id.
247. KeyCite organizes citing references for a case by segregating negative citing references from positive citing references. KeyCite further organizes negative and positive citing references according to the depth of treatment given by the citing reference to the cited opinion. Four categories exist for the depth of treatment provided by the citing reference: (1) "examined," indicating that the citing reference contains an extended discussion of the cited opinion, usually more than a printed page of text; (2) "discussed," indicating that the citing reference contains a substantial discussion of the cited opinion, usually more than a paragraph but less than a printed page; (3) "cited," indicating that the citing reference contains some discussion of the cited opinion, usually less than a paragraph; and (4) "mentioned," indicating that the citing reference contains a brief reference to the cited opinion, usually in a string citation. Finally, KeyCite identifies citing references that directly quote the cited opinion.
248. For citation data for those first-tier appellate opinions with positive citing references, see infra Appendix tbl.3.
249. Of the 77 observations in the first-tier database for which there was a subsequent appeal to the court of appeals, 50 were district court dispositions and 27 were BAP dispositions. As there were a total of 162 district court and 106 BAP dispositions in the first-tier database, infra Appendix tbl.1, approximately 31% of the district court dispositions and 25% of the BAP dispositions involved subsequent appeal. As our first-tier database only includes opinions that disposed of the appeal on the merits, these figures seem to be consistent with empirical evidence that has estimated that up to a third of first-tier appellate dispositions rendered on the merits have been further appealed to the court of appeals.
250. See McKenna & Wiggins, supra note 65, at 630;
251. see also U.S. BANKRUPTCY APPELLATE PANEL FOR THE EIGHTH CIRCUIT, STATISTICAL REPORT: JANUARY 1, 2005 - DECEMBER 31, 2005 (2005), available at http://www.ca8.uscourts. gov/ newbap/stats/q4-05.pdf (documenting that approximately 30% of bankruptcy appeals in the Eighth Circuit in 2005 were taken to the U.S. Court of Appeals).
252. The 91% circuit court affirmance rate of BAP dispositions approximates the rate at which the U.S. Court of Appeals for the Tenth Circuit partly or fully affirmed merit-based BAP dispositions-that is, 90%-during the eleven-year period beginning on July 1, 1996 and ending on June 30, 2007. See U.S. BANKRUPTCY APPELLATE PANEL FOR THE TENTH CIRCUIT, ANNUAL REPORT OF BANKRUPTCY APPEALS IN PARTICIPATING BAP DISTRICTS FOR THE STATISTICAL YEAR JULY 1, 2006 - JUNE 30, 2007 (INCLUDING DISPOSITION STATISTICS FOR APPEALS DISPOSED OF SINCE JULY 1, 1996) 8 (2007), available at http://www.bapl0.uscourts.gov/stats/2007.pdf.
253. The 81% full affirmance rate of BAP dispositions by circuit courts approximates the rate at which the U.S. Court of Appeals for the Tenth Circuit fully affirmed merit-based BAP dispositions-that is, approximately 89%-during the ten-year period beginning on July 1, 1996, and ending on June 30, 2006. Id.
254. For the details of these results, see infra Appendix tb1.4.
255. The first-tier database contained 202 observations in which a federal court positively cited to the opinion issued by the first-tier appellate court. In conducting our bivariate analyses, we exclude extreme outliers (i.e., those observations involving extreme values in the tails of the distribution of the positive citing reference data). We define an extreme outlier to be any observation with a total number of positive citations that falls above the third quartile of the positive citing reference data (7 citations) by more than three times the interquartile range for such data (5 citations).
256. See infra Appendix tbl.3 (describing distribution of positive citing references to first-tier opinions). Accordingly, we excluded any observations with more than 22 positive citing references. Pursuant to this measure, we eliminated 2 observations from our analysis-both BAP opinions-leaving us with a total of 200 observations for analysis. Accordingly, approximately 99% of the first-tier appellate opinions in our sample that were cited positively by other federal courts are included in our bivariate analyses of the citing reference data.
257. Cf. Merryman, Toward a Theory of Citations, supra note 104, at 403 (arguing that the larger number of citations by the California Supreme Court to opinions issued by the courts of New York State may be due to the large case literature arising out of New York).
258. For a full accounting of these results, see infra Appendix tbl.5.
259. For purposes of this analysis, we once again exclude extreme outliers according to the criteria discussed in supra note 142.
260. For a full accounting of these results, see infra Appendix tbl.5.
261. For purposes of these analyses, we excluded extreme outliers according to the criteria discussed in supra note 142. For a full accounting of these results, see infra Appendix tbl.5.
262. We might assume that an opinion that comprehensively and effectively addresses an unresolved or debated issue of law that has been repeatedly litigated not only will be heavily cited, but also will be cited more quickly. Thus, for purposes of this analysis, we exclude the extreme outliers we identified with respect to total number of positive citing references.
263. See supra note 142. We further sought to identify whether there were any extreme outliers in terms of the number of days it took for the first-tier appellate opinions to be cited. In this instance, we define an extreme outlier to be any observation with a total number of days that falls above the third quartile of the immediacy data (638 days) by more than 3 times the interquartile range for such data (520.5 days). On the basis of these parameters, there were no additional extreme outliers.
264. For a full accounting of these results, see infra Appendix tbl.5.
265. See supra Part III.A.2.
266. See supra text accompanying notes 134-135.
267. The opinions in the database were issued during one of three government fiscal years-1998, 1999, or 2000.
268. See supra note 122 and accompanying text. The reference category was first-tier opinions issued during the 2000 fiscal year and was accordingly excluded from the model.
269. The four most frequently occurring subjects were matters relating to discharge (24%), multiple subjects (17%), avoiding powers (10%), and procedure/jurisdiction (9%).
270. Both the Court and Published variables are significant predictors of Affirmance and Full Affirmance. The Appellant, Appellee, and Subject variables are significant predictors of Full Affirmance (but not Affirmance). For detailed results from this regression model, see infra Appendix tbl.6.
271. When performing inference, we implement the recommended practice of conveying levels of uncertainty by using the notation [#, #] to indicate the lower and upper bounds of the 95% confidence interval around our estimates.
272. See Lee Epstein, Andrew D. Martin & Matthew M. Schneider, On the Effective Communication of the Results of Empirical Studies, Part I, 59 VAND. L. REV. 1811, 1835-37 (2006).
273. For detailed results from this regression model, see infra Appendix tbl.7.
274. The model also identifies the Published, Dispute, and Subject variables as significant predictors of whether the first-tier appellate opinion will have been cited by other federal courts.
275. There were actually 202 such observations. For purposes of our regression analyses, however, we eliminated 2 extreme outliers, which left 200 observations to be analyzed.See supra note 142.
276. A zero-truncated negative binomial regression model is appropriate here since (1) the aggregate number of positive citations is a count variable that is overdispersed, and (2) there are no zero values for this subset of observations (i.e., all opinions have at least one positive citing reference).
277. We ran the regression model four times to account for four different types of citing federal courts (i.e., bankruptcy court, district court, BAP, and extracircuit federal courts). In order to predict the number of citing references by courts of appeals, however, we used a Poisson regression model. When fitting a negative binomial regression model, the likelihood ratio test for alpha-the overdispersion parameter-was not significant (chi-squared = 2.52, df = 1, p = 0.056), thus indicating that the Poisson model is preferred. We do not use a zero-truncated model for any of these dependent variables since some of the observations do have zero values.
278. To predict the total number of positive district court citations to first-tier appellate opinions, we initially fitted a negative binomial regression model that included all of the independent variables in the negative binomial regression model used for the other types of citations (the full model). Although the Court variable was a statistically significant predictor in the full model, the model as a whole was not statistically significant (chi-squared = 19.21, df = 12, p = 0.0836). Accordingly, using a backward-selection stepwise regression, we fitted a partial model that only included the Court, Debtor, Subject, and Fiscal Year variables. This partial model was statistically significant (chi-squared = 16.70, df = 5, p - 0.0051), and the Court variable continued to be a statistically significant predictor (p = 0.032).
279. For detailed results from these regression models, see infra Appendix tbl.8.
280. In one instance, we do not use a negative binomial regression model. In order to predict the number of citing references that examined the first-tier appellate opinion (i.e., an opinion that contains an extended discussion of the cited opinion usually more than a printed page of text), we used a Poisson regression model since the values for this dependent variable were not overdispersed. When using a negative binomial model, the likelihood ratio test for alpha-the overdispersion parameter-was not significant (chi-squared = 0.0019, df = 1, p = 0.483), thus indicating that the Poisson model is preferred.
281. To predict the total number of positive citations that provided discussion of more than a paragraph but less than a printed page of the opinion, we initially fitted a negative binomial regression model that included all of the independent variables included in the negative binomial regression model used for the other types of citations (the full model). Although the Court variable was a statistically significant predictor in the full model, the model as a whole was not statistically significant (chi-squared = 15.41, df = 12, p = 0.2200). Accordingly, using a backward-selection stepwise regression, we fitted a partial model that only included the Court and Subject variables. This partial model was statistically significant (chi-squared = 9.67, df = 2, p - 0.0080), and the Court variable continued to be a statistically significant predictor (p = 0.008).
282. For detailed results from this regression model, see infra Appendix tbl.9.
283. None of the other independent variables was a statistically significant predictor of direct quotation of the first-tier appellate opinion by its citing reference. For detailed results from this regression model, see infra Appendix tbl.11.
284. The model incorporates the same independent variables and observations from the binary logistic regression model used to predict the tendency for direct quotation of first-tier appellate opinions.
285. No statistically significant relationship existed between any of the other independent variables and the number of citations that directly quoted the first-tier appellate opinion. For detailed results from this regression model, see infra Appendix tbl.9.
286. For detailed results from this regression model, see infra Appendix tbl.10.
287. See supra Part III.C.1.
288. See CROSS, supra note 95, at 39-68.
289. A study conducted by the Federal Judicial Center found that courts of appeals fully affirmed the judgments of district courts in bankruptcy appeals approximately 73% of the time, and the study further estimated that the affirmance rates for BAP judgments would be similar.
290. See McKenna & Wiggins, supra note 65, at 630. We conclude that our statistically significant evidence contravenes the prior understanding of outcomes in the bankruptcy appeals system. In response to the account we have provided regarding circuit court affirmance, one might argue that courts of appeals simply prefer to reduce their workload and that they accordingly tend to defer to BAPs (even if the legal standard calls for more exacting review). Assuming, however, that the predilection toward leisure does not outweigh the desire of courts of appeals to resolve appeals correctly, then one must assume that, to the extent that courts of appeals affirm-and, on this story, defer to-BAPs more than district courts, they do so because they believe that BAPs are structurally more likely to resolve the issues correctly than are district courts. It is true that, to the extent that the courts of appeals simply defer to BAPs to a greater extent on principle (i.e., with less review of how the BAPs actually resolve particular cases), we cannot say whether the affirmance rate measures actual quality of appellate review. Even on this story, however, the affirmance rate does measure the perceived quality of appellate review. One also might argue that courts of appeals in BAP circuits are heavily invested in the success of the BAPs given that the judicial councils in those circuits have decided to establish BAPs in the first instance,
291. see 28 U.S.C. § 158(b)(1) (2000) (setting forth procedures for establishing a BAP), and given that the judicial councils have also selected the bankruptcy judges who will serve on the BAPs,
292. see id. § 158(b)(3) (setting forth procedures for selecting BAP judges). The desire to legitimate the BAP as an institution, in other words, may motivate the greater tendency of courts of appeals to affirm BAPs. Insofar as review by a court of appeals is transparent, however, it is unlikely that the court would simply affirm BAP judgments and reasoning that were obviously wrong. Put another way, it seems unlikely that the desire of a court of appeals to legitimate a BAP would entirely outweigh the desire to resolve appeals correctly.
293. See supra notes 106-107 and accompanying text.
294. See infra Appendix tbl.8.
295. See McKenna & Wiggins, supra note 65, at 678 ("There is anecdotal evidence that circuit judges find the BAP decisions they review better reasoned and the cases better prepared for review than decisions from the district courts, and that this impression is independent of the likelihood of affirmance or reversal." (emphasis added)).
296. See infra Appendix tbl.8.
297. See McKenna & Wiggins, supra note 65, at 628.
298. See, e.g., id. at 634:[U]sers of the complex bankruptcy system probably want precedent not just settled, but settled right. If early (and in the Ninth Circuit, not so early) impressions about the quality of work by the bankruptcy appellate panels hold up, the dual needs for binding authority and substantive correctness argue for some sort of a dual or hybrid system involving bankruptcy appellate panels in some form. The benefits of establishing such tribunals may impose various costs, although they may be minimized depending on the manner in which such tribunals are integrated within existing judicial operations.
299. See, e.g., id. at 629; Bankruptcy Appeals, Lawyers Wary of New System Begun This Month, N.Y.L.J., July 11, 1996, at 5 ("Steven Flanders, the Second Circuit Executive, though, while noting that the BAPs represent a 'major institutional change' for the circuit, believes the costs will be manageable. Mr. Flanders indicated that the calendaring of appeals will be integrated into the circuit clerk's office ."). From the perspective of litigants, a cost of particular concern would be that of disposition time. In the bankruptcy appeals context, however, preliminary evidence tentatively suggests that BAPs appear to have reduced such costs. A study by the Federal Judicial Center reported that, "[i]n most circuits, the overall mean and median disposition times for BAPs are lower than the national figure for bankruptcy appeals to the district courts, but in most circuits they are based on a rather small number of cases, which limits the conclusions to be drawn from the figures."
300. McKenna & Wiggins, supra note 65, at 661. We emphasize that our findings do not speak to whether it is more desirable to have many such tribunals-as is the case with BAPs-or just one national tribunal-as is the case, for example, with the United States Court of Appeals for the Federal Circuit for patent appeals. That issue would seem to turn largely on the importance of having an intermediate appellate tribunal announce legal rules with national uniformity. See, e.g., id. at 649.