|
Article Excerpt
INTRODUCTION
I. EMPIRICAL STUDY OF OPINION ASSIGNMENT A. Previous Work B. Data and Methods 1. Data sources 2. Data reliability 3. Methods C. Results D. Discussion 1. Confidence in results 2. Explanations 3. Specialization trends E. Limitations 1. Data set limitations 2. Measuring "specialization" 3. Chief judges II. ASSESSMENT A. General Assessment B. Concerns About Specialized Courts 1. Politicization 2. Myopia 3. Loss of prestige 4. Other concerns C. Potential Problems 1. Erratic expertise 2. Excessive deference 3. Bias CONCLUSION: OPINION SPECIALIZATION AS REFORM APPENDIX A: SUMMARY TABLES OF SPECIALIZATION APPENDIX B: NATURE OF SUIT CODES APPENDIX C: DEFINING SPECIALIZATION INDEX OF FIGURES AND TABLES Figure 1.1. Subject-Matter Specialization, First Circuit, 1995-2005 Figure 1.2. Subject-Matter Specialization, Second Circuit, 1995-2005 Figure 1.3. Subject-Matter Specialization, Third Circuit, 1995-2005 Figure 1.4. Subject-Matter Specialization, Fourth Circuit, 1995-2005 Figure 1.5. Subject-Matter Specialization, Fifth Circuit, 1995-2005 Figure 1.6. Subject-Matter Specialization, Sixth Circuit, 1995-2005 Figure 1.7. Subject-Matter Specialization, Seventh Circuit, 1995-2005 Figure 1.8. Subject-Matter Specialization, Eighth Circuit, 1995-2005 Figure 1.9. Subject-Matter Specialization, Ninth Circuit, 1995-2005 Figure 1.10. Subject-Matter Specialization, Tenth Circuit, 1995-2005 Figure 1.11. Subject-Matter Specialization, Eleventh Circuit, 1995-2005 Figure 1.12. Subject-Matter Specialization, D.C. Circuit, 1995-2005 Figure 2. Agency Specialization, D.C. Circuit, 1995-2005 Figure 3. Frequency of Specialization (Positive and Negative) by Subject Table 1. Subject-Matter Specialization, Geographic Courts of Appeals, 1995-20 Table 2. Agency Specialization, D.C. Circuit, 1995-2005 Figure 4. False Positive Rates for Simulated Seventh Circuit Case Distributions
INTRODUCTION
Legal culture, particularly in the federal courts, celebrates the generalist judge. Indeed, the most enthusiastic celebrants are often the judges themselves. (1) Federal circuit judges, for example, frequently comment on the importance and desirability of being a generalist (2) and acknowledge the generalist's iconic status in the American legal tradition. (3) In short, many federal judges would toast to Judge Diane Wood's assertion that "we need generalist judges more than ever for the United States federal courts." (4)
The corollary to a powerful generalist ideal is a deep-seated aversion to specialization. (5) Outward support for specialization, if it exists at all, is confined to narrowly limited areas. Otherwise, judges resist specialization and distance themselves from its "spectre." (6) The aversion occasionally even crosses over to outright hostility: one outspoken judge describes the Federal Circuit as comprised of "little green men" and "people wearing propeller hats." (7)
The structure of the federal courts reflects this distaste for specialization accordingly. The system is comprised overwhelmingly of courts of general jurisdiction, with the Federal Circuit and a few other courts as the only exceptions. (8) Proposals over the years advocating for additional specialized courts have been consistently ignored, (9) whether in scientific evidence, (10) tax, (11) immigration, (12) administrative agency review, (13) patents, (14) or other areas. On the rare occasions when such proposals are implemented, most specialized courts are denied Article III status and classified as legislative (Article I) courts, such as the bankruptcy courts, the United States Tax Court, (15) and the United States Claims Court. (16)
Consistent with these attitudes, well-established rules and norms within the courts of general jurisdiction require the random assignment of cases to ensure that judges see all case types. (17) One notable former exception to random assignment in the district court context was the 1971 Bar Harbor Resolution, which allowed chief judges to reassign complex trials to specific judges. (18) In 1999, however, the Judicial Conference closed this loophole by rescinding the resolution. The Judicial Conference found that it raised specialization concerns and was "inconsistent with the concept of judicial autonomy." (19)
The romantic view of the generalist federal judge, however, is not without its costs. Obsession with the generalist deprives the federal judiciary of potential expertise, which could be extremely useful in cases involving complex doctrines and specialized knowledge. (20) To be sure, expertise is not the be-all end-all of the ideal jurist, particularly when issues require value choices rather than technical accuracy, (21) but even if expert judges cannot necessarily ensure right answers, their decisions are more likely to fall within the subset of better answers owing to their greater experience and understanding of a field.
The loss of expertise also undermines efficiency, a goal that is difficult to dismiss in an era of crowded dockets and overworked jurists. (22) Most current responses to the caseload crisis, including increasing the number of judgeships and staff positions, dispensing with oral argument, and using unpublished opinions, have now been stretched to the breaking point. (23) Any further expansion of these mechanisms risks serious harm to both uniformity (24) and accountability. (25) Meanwhile, specialization options remain neglected and underutilized because of the generalist ideal. (26)
The ideal of the generalist judge thus holds the federal courts captive. (27) Rhetorically, it discourages judges from developing specialized expertise even on an informal basis. Doctrinally, it spawns rules and structures that prevent specialization, and, at the broadest conceptual level, it prevents federal court reformers from seriously considering specialized courts and other subject-matter-based schemes. Its powerful influence is all the more extraordinary given the extent to which specialization pervades nearly every other aspect of modern society. The medical and legal professions, which for years grappled with specialization, are today remarkably specialized, particularly at the most elite levels of practice. (28) Even state courts have increasingly turned to specialized courts or a subject-matter rotation system. (29) Yet, the federal judiciary, the last bastion, remains steadfast and committed.
Is it really? Despite the rhetoric and the structural obstacles against specialization, do federal judges truly practice the generalist ideal? When presented with a chance to specialize, do they actually remain generalists?
This Article sheds light on these questions by looking empirically at the process of opinion assignment in the federal courts of appeals. Opinion assignment is one of the few instances in which judges can still specialize in certain subject areas, and it thus provides a unique opportunity to observe judicial attitudes toward specialization. Part I analyzes a newly compiled data set on opinion assignments from 1995-2005. It reveals opinion specialization to be an unmistakable part of everyday judicial practice, suggesting that the generalist judge is largely a myth.
Part II examines how one should react to this discovery of specialization in the federal courts. Proponents of the generalist judge should be outraged, as it represents a subversion of long-cherished judicial values. Part II, however, approaches opinion specialization with an open mind and shows that it actually captures many of the benefits of specialized courts without incurring their costs. Opinion specialization is a desirable if not welcome development in federal judicial practice, one that can increase expertise while staving off problems such as politicization and tunnel vision.
Finally, the Conclusion offers opinion specialization as an exciting and more viable alternative to traditional proposals for specialized courts. For those seeking to increase specialization and expertise in the federal courts, opinion specialization is far easier to implement, because it can develop through gradual accretion and requires no formal restructuring of the federal system.
A final introductory note: It may be appropriate at this point to be more precise about the term "specialization," as it has been the matter of some academic controversy. Specifically, Daniel Meador has carefully distinguished subject-matter organization from specialization. (30) For example, as consistently noted by Judge Jay Plager, the Federal Circuit is technically not a specialized patent court, because it has other types of cases on its docket and is not exclusively limited to patent law. (31) This precise distinction is concededly true, but the Federal Circuit is of course not a generalist court either. Ultimately, there are many points on the spectrum ranging from generalist to narrow specialist, but in common parlance, the operative dichotomy is between a true generalist court that hears all cases (or a close approximation) and everything else. (32) With apologies to Professor Meador and Judge Plager, this Article therefore uses the term "specialized" to denote any court or judge that deviates from the generalist ideal. (33)
I. EMPIRICAL STUDY OF OPINION ASSIGNMENT
One way of measuring judicial attitudes toward specialization is to observe how opinions are distributed among judges in the federal courts of appeals. Given the various structural impediments to specialization, including general dockets and random panel assignments, opinion assignment provides a rare instance in which judges can specialize in certain subjects. Within the confines of judicial norms about equal distribution of workload, (34) the assigning judge may distribute opinions based on the panel members' special expertise or interest. (35) Alternatively, on courts that operate by consensus, panel members may request or express preference for particular topics. Regardless of how it occurs, specialization will manifest itself in the resulting assignment patterns.
It is worthwhile to reemphasize that the generalist ideal should make such specialization taboo. Under the ideal, judges are not supposed to specialize, whether procedures permit it or not. Indeed, the internal operating procedures for the Fifth and Eleventh Circuits emphatically declare: "Judges do not specialize. Assignments are made to equalize the workload of the entire session." (36)
Along these lines, this Part tests whether opinion assignment practice is consistent with the generalist rhetoric. It examines assignments in the United States Courts of Appeals from 1995-2005 to ascertain whether and how much de facto specialization occurs.
A. Previous Work
A review of the literature reveals only two previous studies on opinion specialization in the courts of appeals. (37) Both of these studies, however, examined court practices from nearly half a century ago. The earliest study, published by Burton Atkins, (38) examined opinion assignments in selected courts of appeals over short periods (typically two or three years) during the late 1950s and 1960s. (39) Atkins concluded that circuit opinion assignments during that period were "not random" and that "judges tend[ed] to specialize in certain substantive areas." (40) For example, on the Second Circuit, Atkins found that Judge Hays wrote 56.2% of labor opinions in which he was in the majority, whereas Judge Waterman wrote none. (41) Similarly, on the Fourth Circuit, Chief Judge Sobeloff wrote 54.1% of racial discrimination cases in which he was in the majority, whereas the next highest judge had a rate of 15%. (42)
The second study, published by J. Woodford Howard in 1981, analyzed opinion assignments in the Second, Fifth, and D.C. Circuits from 1965-67. (43) Overall, Howard found only weak opinion specialization, with no observable specialization in the Second Circuit, (44) and only a few patches in the Fifth Circuit including civil rights, economic issues, and "marine personal injuries." (45) Perhaps more interestingly, despite the weak empirical evidence of specialization, some of the circuit judges Howard interviewed expressed positive attitudes toward opinion specialization. (46) For example, some judges defended it as "bringing their best minds forward, an intelligent allocation of resources in the appeals explosion," and "[o]thers regarded it as too fluid to compromise seriously the norm of appellate review by generalists." (47)
The political science literature also has a well-developed thread on opinion assignment in the Supreme Court. In particular, studies over the years have found issue specialization on the Warren, (48) Burger, (49) and Rehnquist Courts. (50) Supreme Court studies, however, have only limited value for the broader specialization debate. Few would dispute that the Supreme Court occupies a unique position in the federal system. Due to its discretionary review power, the Supreme Court handles only a small fraction of federal cases, and focuses on more policy-oriented, constitutional, and controversial cases. (51) More importantly, because the Supreme Court has no panel system--every Justice hears every case--over eighty percent of opinion assignments are made by the Chief Justice, (52) resulting in an arguably more top-down assignment process. Opinion specialization in the Supreme Court is therefore more a function of the attitude of the Chief Justice than those of the Justices generally. (53)
B. Data and Methods
1. Data sources
Two foundational data sets provided the data for this study. The first data set was the Federal Judicial Center's (FJC) well-known Federal Court Cases database publicly available from the Inter-university Consortium for Political and Social Research (ICPSR). (54) The FJC data set is an excellent source because it contains all federal appellate cases and includes a great deal of information on each case. The primary problem with the FJC data set, however, is that it contains no information on judge assignments. The judge-specific data apparently exists in the system operated by the Administrative Office of the U.S. Courts, but the Judicial Conference has stripped it from publicly available data sets. (55)
To address this deficiency in judge-specific information, I merged the FJC data set with a database extract generously provided by Thomson West. Among other things, the West data set contained the case names, citations, docket numbers, and associated judicial authors for all opinions available on Westlaw. After I culled per curiam and other unsigned opinions, I matched the West data set to the FJC data set using docket numbers.
The combined data set included all opinions written between 1995 and 2005 in the United States Courts of Appeals for all circuits except the Federal Circuit. For purposes of analysis, I made a number of simplifying assumptions: First, because judges presumably express subject-matter preference by writing actual opinions, I treated case opinions involving multiple docket numbers as a single data point. Second, whenever a docket number was associated with multiple majority opinions (for example, because of rehearing, en banc review, or remand), I counted each unique judge only once. To illustrate, if Judge A wrote the majority opinion and an opinion on rehearing, and Judge B wrote the en banc decision, I credited Judges A and B with one opinion each. Since the original author often writes or is presumptively entitled to write an affirming en banc decision or an opinion on remand, this rule prevents double counting. Finally, I consolidated the subject-matter codes from the FJC data set to obtain more meaningful categories of workable size. (56)
2. Data reliability
Scholars have recently raised the issue of error in the FJC data set, (57) and the Westlaw extract was of unknown quality, so I manually checked a random sample (58) against the actual opinions to verify the accuracy of the combined data set. For the purpose of matching judges with case subject matter, the accuracy rate of the data set was high (95%, margin of error = +/-4%). As might be expected, some of the subject-matter classifications were interpretative, since cases can involve multiple issues or change in character over time, but they were otherwise generally accurate, in line with previous research. (59) Most errors involved attributing the opinion to the presiding judge rather than the authoring judge, possibly because of the way Westlaw extracts header information from the opinion text. To the extent that these errors are not correlated with subject matter (and there is no reason to believe that they are), they are acceptable. At worst, the errors bias the study toward a finding of no specialization.
3. Methods
To analyze the data, I created two-way contingency tables for each circuit showing the number of opinions written by each judge in the various subject-matter categories. I used an "OTHER" category to consolidate judges who wrote fewer than thirty case opinions during the period. These judges wrote too few opinions from which to draw inferences, and they were often temporary assignments who sat by designation, so they did not reflect long-term attitudes.
The number of opinions that judges write of course cannot be directly compared. Some subjects are more common than others. At the same time, particularly over a long time span, some judges write more opinions than others, whether because of ascension or retirement from the bench, a reduced caseload due to senior status, or greater prolificacy. To detect instances of specialization, observed frequencies must be compared with the number of expected opinions that a judge should write given the judge's overall caseload and the circuit's docket patterns (in the absence of specialization). I calculated expected frequencies using median polish, (60) a well-established method for detecting instances in which the observed frequencies deviate significantly from a random distribution. (61)
C. Results
Figures 1.1 through 1.12 graphically summarize the most likely instances of specialization found in the analysis. Each horizontal line represents a particular subject matter, and each dot represents a single judge. The degree of specialization is measured using (Pearson) standardized residuals, a statistical measure for standardizing the differential between the observed number of opinions and the expected number of opinions. Residuals with an absolute value above three are typically thought to be of interest, (62) so all such instances are labeled with the judge's name. Conversely, the gray region denotes residuals between -3.0 and 3.0, instances in which specialization was not found. For clarity, Appendix A lists all of instances of specialization along with the observed and expected number of opinions.
There are two immediate caveats in reading the results in Figures 1.1 through 1.12. First, for ease of reading, the graphs cap residuals at +/-8, so a few instances of extreme specialization are truncated in the graphical display. Second, the data set unfortunately conflates judges within a circuit with the same last name. Examples include Judges Richard and Morris Arnold on the Eighth Circuit, and Judges William and Betty Fletcher on the Ninth. (63) These observations are therefore invalid and were accordingly excluded. (64) The list of excluded judges unfortunately includes several former academics, whom we might have expected to have specialization tendencies.
Finally, the D.C. Circuit offers an additional way to assess specialization given its unusual and substantial docket of agency review cases. Figure 2 breaks down the D.C. Circuit opinions by agency reviewed.
[FIGURE 1.1 OMITTED]
[FIGURE 1.2 OMITTED]
[FIGURE 1.3 OMITTED]
[FIGURE 1.4 OMITTED]
[FIGURE 1.5 OMITTED]
[FIGURE 1.6 OMITTED]
[FIGURE 1.7 OMITTED]
[FIGURE 1.8 OMITTED]
[FIGURE 1.9 OMITTED]
[FIGURE 1.10 OMITTED]
[FIGURE 1.11 OMITTED]
[FIGURE 1.12 OMITTED]
[FIGURE 2 OMITTED]
D. Discussion
The results strongly suggest that specialization is alive and well in the federal appellate judiciary. Opinion assignments are not randomly distributed, and in some instances the frequency in which certain judges write is wildly disproportionate to their colleagues. To be sure, no circuit judge appears to write in a few areas to the exclusion of all others, but that may be because random panel assignments prevent it. Notably, however, specialization manifests itself primarily in the positive direction. Judges take a disproportionate number of cases in their preferred subject areas; seldom do they shun unwanted ones. (65)
1. Confidence in results
One important preliminary question is whether one might see these matters as purely a matter of chance. With so many judge-subject matter pairings, some statistical outliers are inevitable. The...
|
