Specialized Adjudication

would have taken up a disproportionate share of the generalist judges' time.

Second, concentrating cases into one or a few tribunals should produce a bench small enough to maintain the collegiality necessary to speak with a single voice. Greater consistency in court opinions would offer greater guidance to consumers of the law, reducing their need for judicial intervention. In addition, the monopoly, or near monopoly, created by funneling all cases in a field into a single tribunal should eliminate entirely disputes on issues such as forum selection and choice of law.

Third, a specialized court's judges would either be chosen for their special expertise or because new appointees could quickly acquire experience in the court's specialty. Moreover, unlike the regional circuits, the court might have enough work in a given field to justify the employment of technical assistants. If, as common experience suggests, experts are better than laymen at dealing with matters in their special areas, the specialized judiciary should handle cases more efficiently, thereby reducing the number of judge-hours required to decide any given number of cases. Each case could also be decided more rapidly, a benefit in fields where timeliness is essential.

Most important, the court's expertise should enable it to craft better opinions, especially in fields where a small number of cases are now distributed rather thinly among the regional courts. Since generalist judges are confronted with the specialty subject matter infrequently, they lack the motivation, experience, and time to develop an understanding of the law. They decide the occasional case based upon a cursory understanding of policy and receive limited feedback on how well they fared.

Thus, specialized court's sustained involvement with a field would facilitate superior decisionmaking . Such a court would be in a better position to understand when it is better to sacrifice accuracy (the "right" result in every case) for the ease with which bright-line rules can be applied and how to draw the fine distinctions necessary when accuracy is more important than administrative convenience. In addition, such a court would have the opportunity to see the practical results of its rules and the time to wait for the most appropriate vehicles for changing them. Absolute responsibility over an entire corpus of law should, at least theoretically, be exciting, for it provides a unique opportunity to oversee the development of a coherent body of doctrine, whose elements are carefully chosen to mesh effectively.

With these advantages, one might ask why the federal system has not made better use of specialized courts. Part of the answer may be that in the nation's early years, there were not many federal cases, so courts empowered to hear the entire federal docket were more efficient than courts that could only entertain actions within particular subject matter areas." Another part of the answer may, however, be that the disadvantages of specialization are as easy to list as the benefits, and that without an overwhelming need to find new efficiencies, Congress was well advised to steer clear of the strategy.® Not surprisingly, the problems with specialization track the advantages. Thus, removing a field from the regional courts' purview is a benefit from the standpoint of those courts' dockets, but it also means that the thinking of generalists no longer contributes to the field's development. Cross-pollination among legal theories is a significant source of change in the law since important patterns of reasoning sometimes emerge rather naturally in one field, yet can be meaningfully applied to other areas.' Furthermore, it makes sense for the adjudicatory system to function in a manner that promotes the like treatment of similar issues, even when they arise in different contexts.

Similarly, while concentration of cases in a collegial bench might stabilize the law, it also makes the tribunal more vulnerable to politicization than courts of general jurisdiction. When issues in a field of law are considered by courts all over the country, interest groups have limited ability to influence the direction of the law by influencing the appointment of the judges who create it. The resources-money and power-of these groups must be spread over the entire judiciary, and their 2. Cf. F. FRANKFURTER & J. LANDIS, BUSINESS OF THE SUPREME Court 147-48 (1927) (discussing the changing demands of an industrialized society).

3. See generally R. Posner, THE FEDERAL CouRTS: CRISIS AND REFORM (1985) [hereinafter FeperaL Courts]; Posner, Will the Federal Courts of Appeals Survive Until 1984? An Essay on Delegation and Specialization of the Judicial Function, 56 S. Ca. L. Rev. 761 (1983); Cavers, Law and Science: Some Points of Confrontation, in LAW AND THE SociaAL ROLE oF ScrENcE 5 (H. Jones ed. 1966); Rifkind, A Special Court for Patent Litigation? The Danger of a Specialized Judiciary, 37 AB. A. J. 425 (1951).

4. See, e.g., R. Posner, Economic Anatysis or Law (2d ed. 1977) (demonstrating expansion of economic analysis from antitrust law to other fields) [hereinafter ECONOMIC Anatysis]; Weiss, Sociobiology and the Law: The Biology of Altruism in the Courtroom of the Future, 85 Micu. L. Rev. 1041Rev. (1987) ) (use of sociobiology in trusts and estates as well as other areas).

efforts encounter interference from those organizations that are concerned with other issues on the judicial agenda. But neither resource-spreading nor influence-dilution would occur in relation to an appointment to a specialized bench that bears solely, and only, the responsibility for deciding the issues which concern an interest group. In that case, the side that is better heeled or more powerful could capture the court and create a bench more likely to issue one-sided opinions.

Even if the appointment process remained untainted, capture may occur through the court's continuous contact with the bar that practices before it. Commentators suggest that repeat players have an advantage over one-time litigants in courts of general jurisdiction.' This problem would be exacerbated on a specialized bench, where repeaters would be more likely to know all of the judges, be acquainted with the eccentricities of the court's rules and specialized law, and be positioned to find suitable vehicles for arguing the changes in the law that they desire.

The efficiency with which specialized courts operate may also sometimes be disadvantageous. Percolation of ideas cannot occur in a court that has exclusive jurisdiction over its field.® And even if some cases in the specialized field remain in the regional circuits, these courts might tend to defer to the expertise of the special bench. If conflicts fail to develop, Supreme Court activity in the specialized field will diminish. As a result, pronouncements of the specialized court will establish new law immediately, and with a fairly high degree of finality. While this is generally beneficial, it creates a greater risk of error, exacerbates the consequences of the mistakes that do occur, and reduces the chances that the public will understand or accept major changes. The only available check will be the laborious process of legislative correction, but Congress' greater attention to the court's activities raises the possibility that the court will become an extension of the legislature.' 5. See, e.g., Galanter, Why the "Haves" Come Out Ahead: Speculation on the Limits of Legal Change, 9 Law & Soc'y Rev. 95 (1974). 6. Cf. T. PLucknett, A Concise History oF THE Common Law 102-03 (4th ed. 1948) (noting that one of the prices England paid for juridic unification was that it became impossible to make a comparative study of legal rules and the common law was left "without an effective competitor").

7. See Glidden Co. v. Zdanok, 370 U.S. 530, 566 (1962) ("[I]t is probably true that Congress devotes a more lively attention to the work performed by the Court of Claims, and that it has been more prone to modify the jurisdiction assigned to that court. It remains to consider whether that circumstance suffices to render nonjudicial the decision It must also be recognized that the court's expertise is created at the expense of an isolation that jeopardizes its ability to shape the law. Because of the repetitive nature of the docket, appointments to a specialized bench might not be as highly prized as other federal judgeships. With less prestige-and presumably, the same bad pay as other federal judges-it may be harder to attract the truly talented.* To be sure, practitioners and law professors specialize too, but the nature of the adjudicative task may make repetition more boring to judges than it is to the bar or to scholars.°® If the specialist court were given exclusive jurisdiction over certain subject matter and little other responsibility, isolation would pose further risks. The court's enthusiasm for a new mode of thinking would not be tempered by consideration of cases where that reasoning was shown to be inappropriate. Lacking the full panoply of tools for furthering its policies, the court may distort the law to achieve the ends it deems appropriate. And since the court would not become engaged in the judicial dialogue that occurs when several courts consider the same issues, its judges would not have the opportunity to see how their work is received by the remainder of the judiciary or be given the incentive to write persuasively. of claims against the United States in the Court of Claims."). Cf. H. Dusrorr, THE Unirep States Tax Court 212 (1979) (noting that one reason Congress has been reluctant to convert the Tax Court from a legislative court into a constitutional court is that Congress has more power over the judges of a legislative court).

8. Prestige is a curious factor because it is also used to argue for specialized adjudication, where the claim is made that unless the judicial system is reformed, more regional judges will be needed to meet demand. If more judges are needed, than each judgeship will be less prestigious. See, e.g., H. FRIENDLY, FEDERAL JURISDICTION: A GEN-ERAL View 30 (1973) ("Prestige is a very important factor in attracting highly qualified men to the federal bench from much more lucrative pursuits."). This issue is largely disregarded in this study (except to note that equalizing specialized responsibilities among all federal judges would make the factor less relevant) for the following reasons. First, prestige is hard to measure. In some fields of law, service on a specialized bench is regarded as prestigious, see, e.g., ABA Stupy, supra note 1, at 16 (speculating that a tax court with nationwide jurisdiction would not lack in prestige); Currie & Goodman, Judicial Review of Federal Administrative Action: Quest for the Optimum Forum, 75 Coun. L. Rev. 1, 29 (1975) (noting that the NLRB is considered prestigious); in other fields a specialized bench is not considered prestigious. Second, since the ratio of judges to lawyers has declined over time, it is now less likely that an attorney will be elevated to the bench, which should make each appointment more desirable despite the increase in the total number of judgeships. Third, the candidate pool has expanded from the days that Judge Friendly worried about whether "young men [would] be willing to make the sacrifice." H. FRIENDLY, supra, at 30. 9. FEDERAL Courts, supra note 3, at 150.

Nor is it necessarily true that specialization will lead to a decrease in the demand for judicial resources. Although more stable, better crafted law offers greater guidance, the court's suc- Each court would, at a minimum, need to create "boundary law:" criteria for determining when a case is within its jurisdiction. Appeals from jurisdictional decisions and transfers between specialized and regional courts would increase the burden on the system as a whole.

Most likely, additional refinements would also be found necessary or desirable. For example, each court (or Congress) might develop its own method for handling nonspecialty issues embedded in the cases before it; each might alter its scope of review to reflect the relative competence of the trial and appellate benches; each might be tempted to adopt pleading, discovery or trial methods that function particularly well within the specialized field. The quantity of special law might then grow to the point where the enterprise was wholly unproductive. Phenomena such as dual supervision of regional district courts by regional and specialized appellate courts, a wholly specialized bar, and a return to the equivalent of the writ system are probably not in the nation's interest.

These arguments for and against specialization are well re-_hearsed. Unfortunately, most debates over specialization make their cases in contexts that provide little information about whether they are in fact valid. Some discussions are purely abstract and phrased to argue that specialization is either generally good or generally bad. Other discussions refer to specific proposals and make little effort to relate arguments to other contexts.

The questions that need to be addressed then are: when does specialization make sense, and how should Congress go about fashioning a specialized court in any particular context? The first step in answering these questions is to examine the federal judiciary's past experiences with specialization to determine whether there are specific criteria for deciding when and why this strategy is appropriate.' Several conclusions emerge from this examination. First, specialization is neither always good nor always bad. Second, there are useful criteria to identify the situations in which specialization is fruitful. Third, these criteria can also be used in structuring specialized courts so that, in the particular context, the benefits of specialization can be maximized and the disadvantages minimumized.

As noted above, Congress has never pursued a systematic strategy of satisfying the nation's demand for adjudication with specialized courts. On occasion, however, Congress has experimented with the idea, using a variety of formats. Since theoretical considerations point in so many directions, and because past experience is often cited selectively in the debate over specialization," it is helpful to examine past and current experience with the strategy before attempting to identify criteria for deciding when specialization will work. This section describes the jurisdiction and structure of those past and present sitting'? constitutional'? courts whose jurisdictional grant turns on the subject matter of the case.'* The section's organization is roughly chron-10. Experience in the states and in foreign countries also sheds light on the costs and benefits of specialization. Unfortunately, the experience in these jurisdictions cannot easily be transferred to the federal context because these courts operate against a different set of background rules. In the states, for example, uniformity is less of an issue and the right to an interlocutory appeal makes boundary disputes less costly. In civil law countries, judges receive special training and play a significantly different role in the adjudicatory process. 11. See, e.g., ABA Stupy, supra note 1, at 44 (dissenting statement). 12. By "sitting," I mean to distinguish these courts from such bodies as the Multidistrict Litigation Panel, which handles only certain procedural aspects of the cases within its purview and three-judge courts, which are assembled on a one-time basis. I do not use the word "temporary" because some of the courts discussed below sat for many years, but were either never designed to be permanent courts-the Temporary Emergency Court of Appeals is one example-or were, in fact, transitory-the Commerce Court, for instance.

13. Congress has also created several specialized legislative courts, such as the Tax Court and the bankruptcy courts. For brevity, these are not described in the empirical section of this paper, except insofar as their experience uniquely sheds light on the specialization issue.

14. By this, I mean to remove from discussion courts such as the United States Court of Military Appeals and the Choctaw and Chickasaw Citizenship Court, whose jurisdiction is determined by the identity of the litigants (other than the United States), and territorial courts, such as the local courts of the District of Columbia, whose jurisdiction is defined geographically. Also not discussed is the Court of Appeals for the District of Columbia, which has specialized adjudicatory authority in several administrative ar-ological, but departs from that mode to follow the evolution of each court over time.

Before beginning, one caveat is in order. One of the most disturbing aspects of the specialization question is the difficulty in deciding how effectively a specialized court operates. It is hard to evaluate the work product of a specialized tribunal with-.

out an intimate knowledge of the law in the court's area of authority. And even if one is comfortable examining the court's work and can comment with confidence on the ways in which the court has altered the law, there remains the problem of deciding whether the observed changes occurred because of the court's expertise, experience and deep appreciation of the issues at stake, or because it has been captured by special interests, or has succumbed to another one of the problems outlined above. In some sense, success is in the eye of the beholder. If, for example, one believes that a special interest group has identified a policy correctly, then the fact that the group has convinced the court to adopt its position counts as a benefit. Similarly, greater accountability to Congress is an advantage or a problem, depending on one's view of the roles of the judiciary and the legislature in the constitutional scheme.

The evaluation problem makes it difficult to use past experience to predict how specialization will work in the future. It does not, however, make analysis of past experience useless. If scholars, practitioners, legislators, and litigants are generally pleased with a specialized court, that court was "successful" in an important sense, regardless of the evaluation of any particular expert. By the same token, if these groups distrust the court, then no matter how expert, uniform, or stable the resulting law, specialization will not be a productive direction in which to move the judiciary. This analysis therefore assumes that public perception of how well these courts performed is more important to evaluating the successes of specialization than a scholarly analysis of any one court's decisions.

The Court of Claims Created in 1855, the Court of Claims represents Congress' earliest experiment with subject-matter specialization as a eas. This court is omitted because its unique jurisdictional areas are created through specialized venue provisions; in general, the court behaves more like a regional court than a specialized tribunal.

method of relieving workload pressure. However, unlike the courts currently under discussion, which are designed to alleviate problems in the judiciary, this court was meant to reduce the burdens of Congress by handling issues formerly addressed through private bills. The Court of Claims adjudicated claims against the United States; claims that, without a waiver of sovereign immunity, were not cognizable in federal courts."® One must speculate about the reason Congress chose to put this power in a special court rather than vest it in the existing federal bench. It is most likely, however, that Congress adopted this strategy because it had special concern for cases that questioned the legality of governmental action.'® The original Act conferred upon the court two sources of jurisdiction: (1) claims against the United States founded upon its laws, regulations, or contractual obligations; and (2) cases referred it by the House of Representatives or the Senate. '' In many respects, the Court of Claims functioned like other federal courts. Its judges were appointed by the President with the advice and consent of the Senate and held office contingent upon good behavior.'* Both claimants and the United States were represented by attorneys in proceedings before the court, and decisions could be appealed to the Supreme Court as of right. There was, however, no right to a trial by jury. Since the court sat in Washington, D.C., a prohibitive distance for some nineteenth century litigants, it was given authority to use com- 16. Thus, the initial legislation did not make the court's judgments final. Instead, the court made recommendations to Congress, which had to formally adopt them to put them into effect. Although this cumbersome process was eliminated in 1863, see Act of Mar. 3, 1863, ch. 92, § 7, 12 Stat. 765, 766, the Secretary of the Treasury retained some discretion over payment of judgments until 1866, Act of Mar. 17, 1866, ch. 19, 14 Stat. 9. See also Glidden Co. v. Zdanok, 370 U.S. 530, 557 (1962) ("there is reason to believe that the Court of Claims has been constituted as it is precisely to the end that there may be a tribunal specially qualified to hold the Government to strict legal accounting"). 17. Act of Feb. 24, 1855, ch. 122, 10 Stat. 612. 18. The 1855 Congress was apparently sensitive to the need to establish public respect for the court. Two of its first judges previously served on the highest courts of their states (John James Gilchrist had been the Chief Justice in New Hampshire and Isaac Blackford, a justice on the Indiana Supreme Court). The third judge, George P. Scarburgh, was an eminent lawyer in Virginia. Court of Claims History, supra note 15, at 19. Significantly, their initial salary was higher than that of the regional judges, id. at 17. missioners-appointed officials who traveled about the country to take evidence and compile records.

Despite the superficial similarities between the Court of Claims and other federal benches, there were significant questions about the court's status in the judiciary. At first, the court lacked authority to render binding judgments. Its findings were reported to Congress, together with a recommended disposition, but final decisions were made by the Legislature. Although the increase in claims created by the Civil War led Congress to reduce its own workload further by making the judgments of the Court of Claims final,?° the Supreme Court continued to treat the court as legislative until Glidden Co. v. Zdanok was decided in 1962.7! That case, which presented the question whether a regional circuit court could render a binding judgment if one of the judges on the panel came from the Court of Claims,"? gave the Supreme Court the opportunity to explicitly state that specialized fora could indeed be constitutional courts. In finding the Court of Claims a constitutional court within the meaning of article III, the Court held that so long as the legislation establishing a court complies with the limitations of article I1J-that the court's "business is the federal business there specified and its judges and judgments are allowed the independence there expressly or impliedly made requisite"'?*--Congress could vest it with less than the full judicial power of the United States.2* Additionally, the Court held that Congress could give the Court of Claims jurisdiction over matters (such as claims against the 19. For a full account of procedures before this court, see id. at 33-34. 20. Act of Mar. 3, 1863, ch. 92, § 7, 12 Stat. 765, 766; Court of Claims History, supra note 15, at 21. This legislation also created a statute of limitation for claims against the United States, added two judges to the court, and expanded its jurisdiction to include counterclaims and setoffs of the United States against claimants, but left the Secretary of the Treasury with some discretion over appropriating money to pay judgments, Gordon v. United States, 69 U.S. 561 (1864). In 1866, this provision was repealed. Act of Mar. 17, 1866, ch. 19, 14 Stat. 9. See also Gordon v. United States, 74 U.S. 188 (1868).

Id. at 567-68. There had long been confusion as to whether Congress' power to create lower federal courts allowed it to create lower courts vested with less than the full judicial authority of the United States. See Warren, New Light on the History of the Federal Judiciary Act of 1789, 37 Harv. L. Rev. 49, 65-70 (1923).

United States) that could equally have been handled based upon the Constitution.?" Further business came when patentees were given the right to sue the United States for compensation for unauthorized use of their inventions."® By 1925, the court's workload had increased so dramatically that significant reorganization was required."® Congress divided the court in two, creating a trial-level legislative court, consisting of the former commissioners, and an appellate division, composed of the judges.*° At about the same time (and also for 25. Glidden, 370 U.S. at 550-51. In so doing, the Court recognized that Congress regarded the Court as an article III court. See Act of July 28, 1953, ch. 253, 67 Stat. 226 (giving the court article III status).

26. See e.g., Act of Mar. 3, 1883, ch. 116, 22 Stat. 485 (Civil War claims); Act of Jan. 20, 1885, ch. 25, 23 Stat. 283 (French Spoliation claims); Act of Mar. 3, 1891, ch. 538, 26 Stat. 851 (Indian depredation claims); Act of May 25, 1908, Pub. Res. 29, 35 Stat. 577 (Boxer Rebellion claims); Act of Mar. 21, 1918, ch. 25, 40 Stat. 451;Act of Mar. 2, 1919, ch. 94, 40 Stat. 1272 (World War I claims).

27. Act of Mar. 3, 1887, ch. 359, 24 Stat. 505. 28. Act of June 25, 1910, ch. 428, 36 Stat. 851. 29. Claims stemming from World War I increased filings by 500% between 1922 and 1923. As of April 23, 1924, there were 2,500 cases on the docket and new filings averaged 100 per week. In 1925, the court issued 247 opinions of which 30 were reviewed by the Supreme Court. Court of Claims History, supra note 15, at 87. 30. Act of Feb. 24, 1925, ch. 301, 43 Stat. 964. Soon thereafter, its judges received formal recognition of their status as appellate judges when their salaries were made equal to those of the remainder of the appellate federal bench. Act of Dec. 13, 1926, ch. 6, 44 Stat. 919. A convincing case can be made for the proposition that it was the fact of its tremendous activity which turned this tribunal into an article III court. Thus, the court's jurisdiction over congressional reference cases, which are not cases or controversies within the meaning of the Constitution, would have barred it from being considered a constitutional court. By restructuring, however, the commissioners-now called trial judges-were positioned to receive plenary power over that portion of the court's docket, leaving the Court of Claims judges to tend to purely judicial functions. Accordingly, the Glidden Court was willing to overlook the jurisdictional anomaly. 370 U.S. at 582. In 1966, Congress formally recognized the need to assign these cases to the trial division in order to preserve the constitutional character of the Court of Claims. Act of Oct. 15, 1966, Pub. L. No. 89-681, 80 Stat. 958. See Court of Claims History, supra note 15, at 96-97;Colaianni, Patent Litigation Before the New Claims Court, 32 Cuev. St. L. Rev. 25, 31 (1983-84).

docket reasons), Congress enacted the "Judges Bill," which made many appeals from the regional circuits to the Supreme Court discretionary.** Since the Court of Claims was included in this legislation, review of its decisions declined dramatically, and it became for all practical purposes the court of last resort for non-tort claims against the United States.

After 1925, the court's jurisdiction continued to grow, mainly through the accretion of narrow grants of jurisdiction.

Although its integrity and impartiality were never seriously 31. Act of Feb. 13, 1925, ch. 229, § 3, 43 Stat. 936, 939. 32, See Act of May 29, 1928, ch. 852, § 617, 45 Stat. 791, 877 (tax claims); Act of Aug. 30, 1935, ch. 831, § 18, 49 Stat. 1028, 1049 (oyster damage claims); Act of May 24, 1938, ch. 266, 52 Stat. 438 (unjust convictions); Act of July 1, 1944, ch. 358, 58 Stat. 649 (contract settlements); Act of Aug. 7, 1946, ch. 864, 60 Stat. 902;Act of June 25, 1948, ch. 646, § 37, 62 Stat. 869, 992 (World War II equitable contract claims); Act of Aug. 13, 1946, ch. 959, 60 Stat. 1049 (appellate jurisdiction over the Indian Claims Commission); Act of May 11, 1954, ch. 199, 68 Stat. 81 (public contract claims); Act of Sep. 8, 1960, Pub. L. No. 86-726, 74 Stat. 855 (copyright infringement actions); Act of Sep. 6, 1966, Pub. L. No. 89-554, § 8715, 80 Stat. 379, 599 (federal employees group life insurance claims); id. § 8912, at 607 (federal employees health benefits claims); Act of Oct. 2, 1968, Pub. L. No. 90-545, § 3, 82 Stat. 931, 932 (Redwood claims); Act of Apr. 3, 1970, Pub. L. No. 91-224, § 11, 84 Stat. 91, 96 (claims regarding oil spills); Act of July 23, 1970, Pub. L. No. 91-350, 84 Stat. 449 (nonappropriated funds jurisdiction); Act of July 1, 1971, Pub. L. No. 92-41, § 3, 85 Stat. 97, 98 (renegotiation jurisdiction); Act of Oct. 29, 1974, Pub. L. No. 93-498, § 11, 88 Stat. 1535, 1543 (firefighter claims);Act of Oct. 4, 1976, Pub. L. No. 94-955, § 1306(b), 90 Stat. 1520, 1720 (declaratory judgments for tax exempt charitable organizations); id. § 1201, at 1665 (certain privacy claims); Act of Oct. 8, 1976, Pub. L. No. 94-465, 90 Stat. 1990 (transfer of the docket of the Indian Claims Comm'n); Act of Mar. 13, 1978, Pub. L. No. 95-243, 92 Stat. 153 34. Pub. L. No. 97-164, 96 Stat. 25 (1982).

35. The Act retained the Claims Court's status as a legislative court, but provided it with stronger protection to its independence and greater authority. Thus, members of the new bench are appointed by the President with the advice and consent of the Senate for a term of 15 years; they can only be removed for cause (incompetency, misconduct, neglect of duty, and for practicing law), and then only with the concurrence of a majority of the judges of the Court of Appeals for the Federal Circuit. The court's powers were augmented to include the ability to enter final judgments, to rule on certain motions that would formerly have been addressed to the appellate division of the Court of Claims, and to adopt its own rules of procedure. Like its predecessor, the Claims Court's jurisdiction

The Court of Customs Appeals owes its origins to the Customs Administrative Act of 1890,2* which created the Board of | General Appraisers to adjudicate appraisement and classification issues arising in connection with tariffs and duties on goods imported into the United States.*7 By giving a single tribunal a monopoly on these cases, Congress sought to create uniformity in an area where horizontal equity was considered beneficial.*® In addition, by removing cases from the regional district courts, Congress hoped to reduce congestion in those courts and to speed up consideration of customs cases.*® Use of a special court was also considered important because the cases within its jurisdiction were likely to raise questions about the exercise of governmental power.*®

The Act was only partially successful. Board determinations were often reviewed de novo by regional courts whose conflicting decisions frustrated the goal of uniformity.*' Delay was rampant because of strategic manipulation by litigants and because review by the regional circuits and appeals to the Supreme Court took time.*? Indeed, the Act actually increased the number of cases by allowing importers to challenge assessments without paying first. In 1909, Congress acted to remedy these problems by creating the Court of Customs Appeals and vesting it with the exclusive power to dispose of appeals from the decisions of the Board of General Appraisers.** is confined to non-tort claims against the United States in which the government has waived its sovereign immunity. 43. Payne-Aldrich Tariff Act, ch. 6, § 28, 36 Stat. 11, 105 (1909). Beforehand, Congress had restricted appeals to the circuit courts of appeals. Act of May 27, 1908, ch. 205, For the most part, this system lasted for twenty years,** until Congress recognized that the patent and trademark cases that were then heard in the D.C. Circuit shared with customs and tariff cases the need for special competence, uniformity and quick resolution. Accordingly, the court's jurisdiction was enlarged in 1929 to include review of decisions of the Patent Office.

Consistent with this new authority, the court's name was changed to the Court of Customs and Patent Appeals (CCPA).*°I n recognition of its special scientific and engineering needs, the CCPA was subsequently given authority to hire technical advisors.*¢

The constitutional status of this tribunal remained unsettled for many years. Although the Board of General Appraisers was Clearly within the executive branch, the status of the CCPA was not mentioned in the acts that created it. While its judges may have signaled their understanding of their judicial status by accepting a pay decrease in 1932, this decision was characterized as a voluntary waiver of their article III salary rights rather than imposed through an exercise of legitimate congressional power.*7 § 2, 35 Stat. 403,404. 44. One significant change was made in 1914, when Congress provided both litigation parties with the right to petition the Supreme Court for a writ of certiorari to review constitutional questions and to interpret treaties, and gave the Attorney General the right to certify other cases for review. Act of Aug. 22, 1914, ch. 267, 38 Stat. 703. 45. Act of March 2, 1929, ch. 488, 45 Stat. 1475. In the next year, Congress made the right to a writ of certiorari the same for all litigants. Tariff Act of 1930, ch. 497, § 647, 46 Stat. 590, 762 (1930).

It is interesting to note that Congress' actions were apparently influenced by the Frankfurter and Landis work. See Glidden Co. v. Zdanok, 370 U.S. 530, 559-60 (1962).

These authors had noted that in the 1890s, suggestions were made for specialized courts in many areas, including admiralty, land and pensions. F. FRANKFURTER & J. LANDIS, supra note 2, at 148. After reviewing past experience with specialization, they concluded, "The need for a coherent system of administrative law, for uniformity and despatch in adjudication, for the subtle skill required in judges called upon to synthesize the public and private claims peculiarly involved in administrative litigation" might lead Congress to reconsider specialization as a method of review of administrative action. Id. at 186. Despite Congress' interest in creating a bench with experience in patent law, none of the original judges on the court were patent experts. See, e.g., Address by W. Graham delivered before the Federal Bar Association (Oct. 3, 1932), reprinted in 14 J. Par. Orr. Soc'y 932, 940 (1932) The Supreme Court had held that the Court of Claims was an article I court in Ex Parte Bakelite Corp.,** but matters were again thrown into confusion when Congress created the CCPA later that same year. As with the Court of Claims, the CCPA's | status was ultimately settled in Glidden Co. v. Zdanok,** in which the Supreme Court held the CCPA to be a constitutional court.

The CCPA was initially unpopular. Many legislators had opposed specialization, and although the 1890 Act passed the Senate by a substantial majority, the measure was greeted with enough criticism to halt the appointment process for a session."°L ater, the court's decision to accept a reduction in pay apparently reflected a concern that it might otherwise be abolished."

In some respects, the fear was not ill founded: several commentators have since used the experience of this court to raise questions about the propriety of the specialization enterprise." Nonetheless, Congress has never been persuaded to eliminate specialization in the fields under this court's jurisdiction. In 1982, the CCPA was merged with the Court of Claims to create the Court of Appeals for the Federal Circuit, which retained its specialized authority.

Reacting to public dismay over regulation of the railroad industry and impressed by the record of the Court of Customs Appeals, President William Howard Taft convinced Congress that a special court was necessary to expedite adjudication of railway disputes, achieve uniformity, and bring expertise to bear on national railroad policy.5? Enacted over the objections of those who feared the court would be captured by special interests, the Mann-Elkins Act of 1910 created a court with exclusive authority to review the decisions of the Interstate Commerce Commission and to adjudicate certain other disputes arising under railway legislation.** Based in Washington, D.C., the new tribunal was authorized to hear cases around the country and was given all the powers of the regional circuit courts, including the power to issue stays and injunctions pending its review, and to hire a staff. Its decisions were reviewable by the Supreme Court, in the Same manner as other circuit courts of appeal.

The article III status of the Commerce Court was never certain. According to its enabling legislation, the tribunal was a constitutional court; it was to be composed of five judges appointed by the Chief Justice of the Supreme Court from the circuit courts of appeals. The picture is clouded, however. Controversy over possible bias had led to a compromise whereby five additional appointments were made by the President with the advice and consent of the Senate.®* The status of these judges was ambiguous: the statute provided that they would hold office contingent upon good behavior, would serve on the Commerce Court for "one, two, three, four, and five years, respectively," and would be available "for service in the circuit court for any district, or the court of appeals for any circuit," as the Chief Justice found necessary. When the Commerce Court was disbanded, however, there was extensive debate over the fate of these judges. Some legislators had assumed the judges' tenure was protected by the Constitution, while others thought the judgeships could be abolished along with the court.®* In the end, Congress chose to retain the judges, but for reasons not expressly stated.®'

The dispute over the status of the judges reflects a larger point about the court itself: it was a highly controversial institution. From its inception, the public perceived the Commerce Court as "owned" by the railways. The ICC objected because it was not given the power to represent itself in litigation reviewing R.R. v. ICC, 200 U.S. 361 (1906); citing also Message of President William H. Taft to Congress (Jan. 7, 1910) 45 Conc. Rec. 378, 379). June 18, 1910, ch. 309, 36 Stat. 539. See also F. FRANKFURTER & J. LANDIS, supra note 2, at 157-160.

55. See F. FRanKFuRTER & J. LANDIS, supra note 2, at 160 n.74. 56. See id. at 168-69. The debate may, in fact, have been not only about the status of these particular judges, but about the status of all article III judges and courts.

57. The judges sat by designation as needed by the regional courts. See id. the court's determinations. Railway owners feared that they would have less influence over the court than they enjoyed with Congress, while shippers believed the court was biased in favor of the owners.*® To make matters worse, the court was quick to interpret railway legislation in a manner that expanded its jurisdiction and the scope of its review,"® leading to rapid reversals by the Supreme Court.®° Congress, which had never solidly supported the Commerce Court, voted to abolish it in 1913, only three years after its creation.®

The Emergency Court of Appeals (ECA) was created by the Emergency Price Control Act of 1942 (EPCA).*? This measure, enacted when the United States entered World War II and in response to rising domestic prices, established an elaborate mechanism for controlling the cost of living throughout the country.® Although the legislation relied on a variety of special agencies for implementation,** Congress recognized that the bur-58. See Dix, supra note 53, at 241-48. 59. See, e.g., Procter & Gamble Co. v. United States, 188 Fed. 221 (1911), rev'd, 225 U.S. 282 (1912). , 60. Within a year after its creation, five of its decisions were reviewed by the Supreme Court; only one was affirmed, United States v. Baltimore & Ohio R.R., 225 U. S. 306 (1912). The others were reversed : ICC v. Goodrich Transit Co., 224 U.S. 194 (1912);Procter & Gamble, 225 U.S. 282 (1912);Hooker v. Knapp, 225 U.S. 302 (1912); ICC v.

Baltimore & Ohio R.R., 225 U. S. 326 (1912). 61. Act of Oct. 22, 1913, ch. 32, 38 Stat. 208, 219 (abolishing the court as of Dec. 31, 1913). Jan. 30, 1942, ch. 26, 56 Stat. 23, amended by, The Inflation Control Act of 1942, ch. 578, 56 Stat. 765. For commentary The ECA's jurisdiction was exclusive: no other court had the authority to pass upon the determinations of OPA, although the ECA's decisions could be reviewed by the Supreme Court. The ECA's authority was also quite narrow. Since Congress intended to rely primarily on the expertise of OPA, the ECA was instructed to review its decisions on the deferential "arbitrary and capricious" standard, rather than the "substantial evidence standard" that federal courts normally applied to the quasi-judicial decisions of administrative agencies.*' To further free OPA's hand, the court was bound by a truncated record. Although it could order supplementation of the record and could remand proceedings if not satisfied with what the OPA Administrator had done, the statute permitted OPA to take "official notice" of materials that would not normally be considered the subject of judicial notice, such as economic data and facts generated by OPA itself. Finally, to ensure that OPA's decisions would go into effect quickly, Congress did not give the court power to enjoin or stay enforcement of OPA regulations."

the Office of Defense Mobilization, the Defense Supplies Corporation, the Temporary Controls Administrator, the Reconstruction Finance Corporation, and the General Services Administration. See Transcript of Proceedings, supra note 62, at 4, 14.

65. It is instructive to note that Congress had actually considered the possibility of dispensing with judicial review entirely. However, it was persuaded that constitutional guarantees and the war effort could coexist with a court structured in the manner de- The ECA is of especial interest because it represents an attempt to establish a constitutional court that could nonetheless be disbanded when the emergency that led to its founding terminated. To this end, Congress utilized the concept of the parttime judge. Instead of relying on the appointment of new judges to serve on this court, Congress had three existing article III judges designated to divide their time between their usual circuits and the EPCA. The initial plan, though presumably in accordance with article III, soon proved impractical. The power to appoint the judges, which was to be placed in the hands of the President,"° quickly passed to the Chief Justice of the Supreme Court.7! Furthermore, by war's end the court's business had grown to the point where five judges spent most of their time tending the docket and holding hearings around the country."

As the Court of Claims and the CCPA provided the Supreme Court with the opportunity to determine the constitutionality of a specialized court,"* the ECA gave the Court the vehicle for passing on the question of exclusivity.7* Relying on Congress' plenary power to ordain and establish inferior courts and to prescribe the authority of state courts to hear federal questions, the Court in Lockerty v. Phillips'® held that Congress also had the power to withdraw jurisdiction over particular is-'sues from some federal courts and put it in others. In Yakus v. United States," the Court decided that such an allocation was constitutional even when its effect required litigants to bifurcate their claims between a regional court-in Yakus's case, a court sitting on a criminal prosecution-and the specialized tribunal.

The record of the ECA was somewhat mixed. When established, it was accepted as the best compromise between the dictates of due process and the need for uniformity, expertise, and rapid implementation."" Indeed, it was partially because of the unique requirements of a national emergency that the Supreme 70. Wilson, supra note 65, at 58. 71. EPCA,supra note 63,at § 204(c),56 Stat. 23,32. 72. See Transcript of Proceedings, supra note 62, at 13-15; The Emergency Court of Appeals, supra note 62, at 5. 73. See Glidden Co. v. Zdanok, 370 U.S. 530 (1962).

74. The Supreme Court had previously passed on the constitutionality of withdrawing power from the lower federal courts. See, e.g., Lauf v. E.G. Shinner & Co., 303 U.S. 323, 330 (1938) (upholding power to limit district court's authority to issue injunctions). 75. 319 U.S. 182, 187-88 (1943). 76. 321 U.S. 414, 437, 442-48 (1944). Court approved the court in Yakus."* But that opinion drew two dissents which cited the more troublesome aspects of specialization. Noting that an overwhelming majority of ECA cases were decided in favor of the government, Justice Roberts objected to the OPA Administrator's significant advantage over complainants in litigating before the ECA." At the same time, Justice Rutledge objected that by putting enforcement of EPCA in the regional courts, while constitutional defenses regarding the validity of the price control was isolated to the ECA, the statutory scheme inevitably required litigants to bifurcate their claims.®°M ost observers, however, considered the ECA a successful innovation. These included the ECA's judges,*? commentators,®2 and practioners.** Congress apparently concurred in this judgment. It gave the court some additional jurisdiction after the War.** While the court was terminated in 1961 when its docket was exhausted, the ECA became the template on which several subsequent courts were modeled.

Created by the Economic Stabilization Act Amendments of 1971 (ESA),®° the Temporary Emergency Court of Appeals (TECA) is the first of the courts patterned directly on the 78. See 321 U.S. at 437-48. Lower courts-and even the ECA itself-approved various features of the statutory scheme on the basis of Congress' emergency war powers, thus leaving open the question whether a similar scheme could be used when the country was at peace. See, e.g., United States v. C. Thomas Stores, 49 F. Supp. 111, 115 (D. Minn. 1943) Act of 1948, ch. 161, § 202(d), 62 Stat. 93, 97 and the Defense Production Act of 1950, ch. 932, § 408, 64 Stat. 798, 808. See Transcript of Proceedings, supra note 62, at 15. 85. Pub. L. No. 92-210, § 2, 85 Stat. 743, 749 (1971) ECA.*¢ Like the earlier court, the TECA was established as part of a legislative scheme aimed at regulating prices, and the specialization strategy was chosen to achieve rapid implementation and uniformity.®? With exclusive jurisdiction to review cases and controversies arising under emergency stabilization legislation and to hear certified constitutional issues from the lower courts, the TECA is composed of judges from the regular federal bench, chosen by the Chief Justice of the Supreme Court. It is based in Washington, D.C., but is authorized to sit in other places, generally at the location where the case arose. Since the court can be expanded by the Chief Justice to meet demand,** the duties of serving on the TECA have not greatly burdened its judges, although loss of their services to their own benches has been noted as a cost.*°T he TECA differs from the ECA in several important respects.*° First, its jurisdiction has altered over time. Rather than disband the court when the conditions leading to its establishment lapsed, Congress has instead chosen to keep the court's docket active by giving the tribunal new assignments.®' More important, while the ECA reviewed the disposition of complaints from OPA under a deferential standard based upon a truncated record, the TECA functions much like an ordinary federal appellate court. It sits in panels of three, and it reviews determina-86. See S. Rep. No. 507,92d Cong.,Ist Sess. 10,reprinted REDDEN, FEDERAL SPECIAL Court LitiGATION 427-43 (1982). 87. See S. Rep. No. 507, 92d Cong., 1st Sess. 10, reprinted in 1971U.S. CopE Cone. & ApMIN. News, 2283, 2298. 88. The court was to be composed of "three or more judges," designated from the district and circuit courts. ESA, supra note 85, § 2, 85 Stat. at 749. In 1982, there were 20 judges on this court, K. REppEN, supra note 86, at 430. In 1989, there were 12. JUDICIAL Starr Directory 4 (A. Brown ed. 1989).

89. Burger, Report to the American Bar Association, 63 AB.A. J. 504, 505 (1977); Note, The Appellate Jurisdiction of the Temporary Emergency Court of Appeals, 64 Minn. L. Rev. 1247, 1268(1980).

90. In addition to the procedural differences described in the text, there are substantive dissimilarities that arise because the ESA was enacted in response to peacetime inflation, while the EPCA was designed to handle the economic dislocations of war. See generally Note, Administration and Judicial Review of Economic Controls, 39 U. Cut L. Rev. 566 (1972). Act of 1973, Pub. L. No. 93-159, 87 Stat. 627;Energy Policy and Conservation Act of 1975, Pub. L. No. 93-163, 89 Stat. 871;Emergency Natural Gas Act of 1977, Pub. L. No. 95-2, 91 Stat. 4. The Emergency Stabilization Act creating the Temporary Emergency Court of Appeals expired in 1974.

tions of the regional district courts.*? Its decisions are appealable by writ of certiorari to the Supreme Court.

A comparison of the TECA to the regional circuits reveals, however, that there are important differences between these courts. Because uniform implementation of economic measures was a major concern of Congress, the legislation creating the TECA provided for both exclusive jurisdiction over cases arising under the ESA and also for the removal from state courts of defenses based upon the constitutionality of the ESA or upon actions taken under its authority. To achieve speedy resolution of disputes, the time limitations for appealing decisions to the TECA are short®* and the court is authorized to promulgate local rules that further accelerate its determinations.

Its name notwithstanding, the TECA has endured for nineteen years. But despite its longevity, it has not been popular with litigants or commentators. The attempt to provide an ordinary trial in the district court with specialization only at the appellate level has created difficult problems for the parties. Because the courts have considered the TECA's jurisdiction limited to issues arising under the relevant legislation,** appeals may be bifurcated, causing delay, confusion, and occasionally the loss of the right to appellate review.®* To further complicate matters, the TECA has held that its jurisdiction is limited to cases in which the district court has actually adjudicated an issue within its authority, a determination that is sometimes difficult to make.®* Thus, one TECA judge has suggested that one third of the court's cases turn on jurisdictional issues.®? On the substantive level, the court has been criticized for being overly 92. See generally Nathanson, Price-Control Standards and Judicial Review-An Historical Perspective, 18 Prac. Law. 59 (Feb. 1972).

93. Compare ESA, supra note 85, § 2 at 749 (30-day time limit to file all notices of appeal) with Frep. R. App. P. 4 (allowing 60 days in some cases). 94. See Coastal States Mktg. v. New England Petroleum Corp., 604 F.2d 179 (2d Cir. 1979). See also Bray v. United States, 423 U.S. 73 (1975)(violation of a district court order appealable to the regional circuit, notwithstanding the fact that the order itself was based on the ESA)(per curiam). 95. See United States v. Cooper, 482 F.2d 1393(Temp. Emer. Ct. App. 1973); Note, supra note 89, at 1256-57. 96. See, e.g., Texaco Inc. v. Department of Energy, 616 F.2d 1193, 1198 (Temp. Emer. Ct. App. 1979), rev'd, 663 F.2d 158 (D.C. Cir. 1980). 97. Id. (Hoffman, J., dissenting). See also Note, supra note 89, at 1258. deferential to the agencies whose activity it reviews,°* or hamstrung by the record that the agencies have compiled." 6. The Special Court, Regional Rail Reorganization Act of Chosen from the existing federal bench by the Judicial Panel on Multidistrict Litigation, three judges sit on the Special Court. It is based in Washington, D.C., but may sit anywhere within the region covered by the Acts and has no formal term. The court has adopted its own rules, which follow the Federal Rules of Civil Procedure but are tailored to its unique requirements. It interprets its jurisdiction narrowly, confining its authority to those rail reorganization issues "where the critical na-98. See, e.g., Elkins, supra note 86, at 114-19 (noting that special deference is particularly inappropriate in statutes whose delegation of authority is upheld because of the checks provided by judicial review and citing Amalgamated Meat Cutters & Butcher 99. Aman, Institutionalizing the Energy Crisis: Some Structural and Procedural Lessons, 65 CorNELL L. Rev. 491, 536 (1980). See also Jaroslovsky, Chalk One Up for the Permanent Government, Wasu. Monrtuty, Oct. 1987, at 33, 37 (noting that the Government has won 90% of the cases that reach TECA). 100. See generally K. REDDEN, supra note 86, at 404-19. 101. Pub. L. No. 93-236, 87 Stat. 985 (1974). 102. Pub. L. No. 94-210, 90 Stat. 31 (1976). ture of the determination demands the consistent interpretation possible only when review is concentrated in a single court."!TM The Special Court has not been controversial, partly because its highly specialized and narrow jurisdiction keeps it out of the public eye, and partly because of the immense distinction and the careful work of its members.! In addition, a scheme to salvage bankrupt public utilities and make them solvent is unlikely to be criticized,'°* especially since the Supreme Court has held that uncompensated takings arising out of these statutes can be remedied by an action in the then-Court of Claims.!°7 7.

The It is probably safe to say that, of all the specialized tribunals, the FISCs have been the most poorly received. While the courts' constitutionally has been upheld as a balance between important competing interests,° classification of the courts' work means that it is impossible to evaluate its performance.*** Members of the legal profession may gain some comfort from the fine reputation of the jurists who have served these courts,"'* and perhaps it is for this reason that the Senate Intelligence supra note 110, at 93 (additional views of Sen. Malcolm Wallop). See generally Note, The Foreign Intelligence Surveillance Act: Legislating a Judicial Role in National Security Surveillance, 78 Micu. L. Rev. 1116Rev. (1980)).

112. The term of an appointment is set at seven years, although the original appoin- Jurists' reactions to the court have been mixed. Then-Judge Robert Bork was quoted as claiming that the secrecy of the court makes it difficult to know whether he would be willing to serve on it, Ricks, supra, while at the announcement of his resigna- 127. See, e.g., Lempert, Proceedings on Steel Reflect Changes in Rules of Trade Game, Legal Times, Aug. 9, 1982, at 1. Lay coverage of this court is mostly in trade journals focused on groups engaged in international trade. In these, the court receives rather neutral reviews, perhaps because readers include both importers and exporters. See, e.g., Sfiligoj, Ruling Seen Aiding Import Relief, 14 MeTaLworkinG News, April 6, 1987, at 4;Hess, Trade Court Rule Seen Beneficial to Importers, 41 Footwear News, Oct. 21, 1985, at 2; Wightman, 149 Women's Wear Daily, Jan. 9, 1985, at 32. 195, 236 (1975) terminated. New judges are chosen by the same process as the regional circuit court judges. The court generally sits in Wash-136. The court's jurisdiction is not plenary even though the statute speaks of it as "exclusive," 28 U.S.C. § 1295(a) (1982). This is because patent questions sometimes arise as defenses or counterclaims in other actions. If the original action is filed in state court, it is not removable to federal court because of the well-pleaded complaint rule, which interprets federal jurisdictional statutes as looking only to the claims that appear on the face of the plaintiff's well pleaded complaint. See, e.g., Consolidated World Housewares, Inc. v. Finkle, 831 F.2d 261 (Fed. Cir. 1987). Actions with patent defenses that can nonetheless be maintained in federal court on other bases will also not be heard on appeal by the Court of Appeals for the Federal Circuit because that court has, with the approval of the Supreme Court, interpreted its jurisdiction as similarly limited by the well-pleaded complaint rule. Accordingly, the appeal will go to the regional circuit. Cf. Christianson v. Colt Indus. Operating Corp., 486 U.S. 800, 809 (1988) (no jurisdiction in CAFC when patent issue raised defensively). But see Aerojet-General Corp. v. Machine Tool Works, Derlikon-Buehrle Ltd., 13 U.S.P.Q.2d 1670 (Fed. Cir. 1990)(assuming jurisdiction over appeal when patent issue is asserted in a compulsory counterclaim) (en banc).

137. 28 U.S.C. § 1295(a) (1982). The court reviews the decisions of the Court of International Trade and final orders of the International Trade Commission. Jd: at § 1295(a)(5), (6). It does not, however, hear appeals from district court trademark actions. Id. at § 1295(a)(1), (4).

138. The Court of Appeals for the Federal Circuit has exclusive jurisdiction over all appeals from the Claims Court (including claims against the United States for more than $10,000 under the Tucker Act) and over appeals from final decisions of agency boards of contract appeals under the Contract Disputes Act of 1978. Id. at § 1295(a) (2), (3), & (10). In addition, the court has exclusive jurisdiction of appeals from federal district court actions for all non-tax and non-tort claims against the United States for $10,000 or less (the "little Tucker Act"), id., and exclusive jurisdiction over most final orders and decisions of the Merit Systems Protection Board. . 195 (1985); Bender, Griffen & Lipskey, supra note "141, at 997 (noting that the CAFC upholds more patents valid than were upheld under prior law). Moskowitz, supra, cites as evidence that the court favors patentees the fact that the CAFC affirms 86% of the district court decisions finding a patent valid, but only 60% of the decisions invalidating a patent, and that 70% of the time, district courts uphold validity, as compared to 30% in the 1970s. Moskowitz also notes that pro-innovation bias is found in the manner in which the Justice Department administers the anti- This study of the nation's past experience demonstrates that specialization has proved to be something of a mixed bag, acceptable in certain contexts, but poorly received in others. As such, the historical material provides a useful basis for deciding when specialization is an effective strategy for satisfying the increasing demand for adjudication. This section attempts to abstract from the record outlined above, and from the author's previous study of the patent jurisdiction of the CAFC,**° the factors that contribute to the success of a specialized tribunal. These criteria fall into three clusters: the field of specialization, the identity of the participants, and the strategy chosen for implementation. For convenience, each is discussed individually, but because of their interrelation, they cannot be discussed in isolation. Pending proposals for specialized adjudication are used to illustrate how these criteria would influence the success of the enterprise. |

The caveat that began the previous section is relevant here as well. There, we noted the intrinsic difficulties in analyzing the workproduct of a specialized court to decide how effectively it operates. Accordingly, objective indicia of public acceptance were used in that section to evaluate past experience with specialized adjudication. The consequence of this observation for this section is to underscore the importance of building into the design of a specialized court systemic guarantees of impartiality. Factors that may not be dispositive of the success of other legal enterprises are significant here if their presence (or absence) influences the adjudicatory process in a discernably systematic Theoretically, specialized adjudication was the ideal way to 'handle the problems posed by the treaties. Adjudication would benefit from the expertise of specialists because the claims depended on law written in Spanish that was conceptually highly dis-

Because this article has focused on courts whose specialty is defined by subject matter, it should come as no surprise that the choice of field is critical to the success of the specialization ensimilar from American law. Yet the need for such a court was temporary, making it senseless to require the entire judiciary to master the complexities involved. Furthermore, speedy adjudication was required to promote efficient use of the land in the treaties. See generally Rightmire, Special Federal Courts, 13 Inu. L. Rev. 15 (1918). Econ. Hist. 498, 513 (1962). Its jurisdiction was highly limited. The court had authority to confirm only claims derived from Spain or Mexico, and then only claims that either were perfect on the date the lands were ceded, or could have been perfected had the land not been ceded; no claim could overthrow an Indian title; no confirmation could confer mineral rights unless the grant upon which confirmation was based expressly included such rights; no claim could be confirmed for land Congress had previously acted upon;

and no imperfect claim could be confirmed for more than eleven square leagues to any one original claimant. Most important, a judgment confirming a grant was essentially no more than an order that the United States quitclaim its rights; it did not reach conflicting claims by third parties. These limitations, which prevented the court from according complete relief, may have discouraged many claimants, and especially poor ones, from using the court.

To make matters worse, the court's composition may have had a systematic influence on its adjudication. It was composed of five judges appointed for a period of four years by the President with the advice and consent of the Senate from "citizens and residents of some of the States of the United States," Act of Mar. 3, supra, § 1. But since the court dealt with land located in Arizona, New Mexico, Colorado, Utah, Wyoming and Nevada, territories that had not all been admitted to statehood, this requirement had the effect of excluding from the candidate pool people who lived in most of the territories in which claimed land was located. In fact, only one judge, Wilbur F. Stone, a former Assistant United States District Attorney during the Civil War, came from a place where granted land was located. R. BRADFUTE, supra, at 27, 219-21. The other judges, Henry Sluss of Kansas, Joseph Reed of Iowa, William Murray of Tennessee and Thomas Fuller, later replaced by Frank Osborne, both of North Carolina, were from the east, where the land grant problem was not well appreciated. As a result, the court may have been especially disinclined to respect Spanish and Mexican grants and unsympathetic to the incomplete sorts of documentation presented. Id. Indeed, wealthy litigants were thought to be more successful in the court than poor ones because they could hire counsel sophisticated enough to help the court with the Spanish language and with Mexican and Spanish law. terprise. Four factors appear to be important in this regard: (1) the complexity of the law and facts raised by cases in the field;

(2) the extent to which cases in the field can be segregated from cases involving other fields of law; (3) the degree of consensus on the objectives of the law in the specialized field; and (4) the way in which cases in the field are distributed among adjudicatory bodies. a. Complexity. Complexity is the starting point not because of its central role in determining the success of specialization, but because the majority of proposals cite, as the major justification for resorting to this strategy, the need for adjudicators with technical expertise in mastering complex subject matter. It is this characteristic that led Learned Hand to call for reform in patent litigation,'** and it is, at least in part, this factor that motivates proposals for special tribunals in such fields as product liability,*® environmental law,' tax, and antitrust law."" Reference to complexity is, of course, not surprising. The more intricate the law, the more likely it is that a generalist will get things wrong, confuse matters, and encourage additional litigation. The more complicated the facts of a case, the more the judge must master before the case can be decided at all. Thus, transferring complex cases to courts staffed by experts should produce three of the major benefits ascribed to specialization, namely improved decisionmaking, a reduction in the size of the docket and a diminution in the number of judge-hours required to clear it. As Chief Judge Markey put it: "if I am doing brain surgery every day, day in and day out, chances are very good that I will do your brain surgery much quicker, or a number of them, than someone who does brain surgery once every couple of cannot stop without calling attention to the extraordinary condition of the law which makes it possible for a man without any knowledge of even the rudiments of chemistry to pass upon such questions as [the patentability of adrenalin]. . . . How long we shall continue to blunder along without the aid of unpartisan and authoritative scientific assistance in the administration of justice, no one knows; but all fair persons not conventionalized by provincial legal habits of mind ought, I should think, unite to effect some such advance.") aff 'd in part, rev'd in part, 196 F. 496 (2d Cir. 1912) First, some of the benefits of specialization may not be measurable by these criteria. For example, complex statutory schemes are often administered piecemeal by a variety of agencies. The court that obtains authority over this caseload may need more time-at least in its early years-to sort out the effects of poor administration . Similarly, when problems result from many courts deciding cases in a poorly understood area, the advantages of consolidating adjudication into a single forum might not translate into fewer filings. While the predictability that the court produces will make settlements more likely,'** if the tribunal also decides each case more rapidly, it will attract new business as adjudication becomes competitive with alternate means of dispute resolution. Moreover, turning a chaotic legal environment into one that is stable may encourage more of the activity that the court regulates, which inflates the base that produces lawsuits. For example, in the years following the establishment of the CAFC, patent applications increased.)

Another reason that it is difficult to measure the benefits of specialization is that the variables affected do not always lend themselves to scrutiny. For example, cases involving numerous parties may be complex enough to benefit from expert adjudication. Thus, Judge Jack Weinstein's experience with the Agent Orange controversy'®* no doubt helped him (and Special Master Kenneth Feinberg) handle the Shoreham litigation more efficiently.'*7 But the benefits in such cases may lie in improving the quality of decisionmaking, something not subject to quantification. Furthermore, these cases are so time consuming that the economies produced by having experts decide them may be below the margin of measuring error. By the same token, there may be no direct way to assess the advantages of removing complex cases from the dockets of the regional circuits.

But sometimes the lack of discernible effects from specialization may indicate that something has gone wrong at the planning stage. For instance, in the past, the specialized tribunal's place in the judicial hierarchy was largely ignored. Yet the effectiveness of specialization may strongly depend on this issue. Thus, if by complexity one means that the factual underpinnings of a case are highly technical, then specialization will have its greatest impact at the trial court level, where the bench's technical proficiency is helpful in evaluating complex evidence and explaining it to a jury. If, however, the complexity that motivates special treatment is in the legal regime, then specialization will be most efficient if it is the appellate tribunal that is expert, for that court's decisions on law will bind and instruct the lower, generalist fora. The difficulties encountered by the CAFC, an appellate court that reviews factually complex subject matter, and by the Tax Court, a trial court that adjudicates complex legal issues,'*® may indicate that Congress mismatched the level of specialization with the problem it sought to which permits taxpayers to have asserted tax deficiencies redetermined by the Commissioner of Internal Revenue. Its judges are appointed by the President with the advice and consent of the Senate for 15-year terms. They may be removed only for inefficiency, neglect of duty, or malfeasance in office. The court's principal office is in Washington, D.C., although sessions are heard around the country; each judge is usually on the road about one week per month in each ten-month session. See K. REDDEN, supra note 86, at resolve.'®® b. Segregability. A substantive factor that does demonstrably affect the success of a specialized bench is the segregability of the issue targeted for special treatment, that is, the extent to which that issue presents itself to the judicial system in cases raising other questions. The experience of the Temporary Emergency Court of Appeals (TECA) is a good example. This court was authorized to review district court decisions on questions arising under a succession of regulatory schemes, with the hope that its small docket would promote rapid and uniform implementation of important policies. To further that goal, the court's authority runs only to the TECA issues in its cases, all other issues in the same case must be appealed to the regional circuit.'®°T his division might work if TECA questions usually arose in isolation, but that has not been the experience. Instead, they arise in a variety of circumstances, including criminal prosecutions. As a result, parties find themselves saddled with the burdens of a bifurcated appeal, and, indeed, bear the risk of losing the right to any appeal. Moreover, the court must devote considerable effort to resolving jurisdictional issues that are unrelated to the.expertise it was expected to develop.?® To avoid this problem, the configuration of cases arising in the specialized field should be considered carefully. If the specialty issues are segregable, specialization will work effectively.

Thus, a Court of Tax Appeals is feasible because litigation between a taxpayer and the government typically raises only taxation issues.'® In contrast, if the specialized issues are generally embedded in larger cases, the strategy may prove inefficient, at least as a method for controlling the federal docket. A Constitutional Court, for example, would almost certainly be unproductive. Since constitutional questions usually appear in cases raising other issues, border disputes would cancel out the advantages of specialization. At the same time, however, segregability can be something of a mixed blessing. An isolated bench may lose touch with the generalized judiciary, fail to perceive developing trends in the law, and overutilize the techniques at its disposal. As adjudication in the specialized area passes out of the common experience, other courts cannot be influenced by the specialized tribunal or by developments in the law it adjudicates. Isolation may even reduce the incentive of the specialized judges to write persuasive opinions.

But these negative effects can to some extent be countered when the specialized issue arises in broader contexts and the specialized bench has "case" jurisdiction-the power to decide all the issues that arise in its cases. The areas in which the specialized court's jurisdiction overlaps with that of the regional courts creates avenues for interchange between these courts while preserving exclusivity in the areas where it is needed. A similar effect could be achieved by adding unrelated matters to the jurisdiction of specialized courts, as was done with the Federal Circuit, but the specialized court may gain more if its exposure to law outside its area of expertise comes from deciding non-specialty issues in the cases included in its specialized jurisdiction. Issues that arise together often raise related policy considerations. A court that hears an entire case gets an opportunity to see the broader universe in which its specialty operates and an appreciation of competing interests as well as other ways in which to further its goals.

The experience of the CAFC is informative. Patent law uniformity and stability is a prime mission of the CAFC. Nevertheless the CAFC's patent jurisdiction has bred frequent litigation over jurisdictional matters because patent claims are often nonsegregable. Although the CAFC has not adopted the TECA's "issue" jurisdiction rule, it has, with the encouragement of the Supreme Court, interpreted its authority narrowly to avoid nonpatent issues.' The court has further tied its hands by adopting a choice-of-law rule that requires it to follow the law of the regional courts in the non-patent aspects of the cases it adjudicates.®* These actions complicate the court's decisionmaking and increase the costs to litigants. They also leave the court grounds). open to accusations of pro-patentee bias, which can be said to result from a failure to fully appreciate market competition policy and a tendency to overvalue patent law as a means of fostering innovation.'®* Were the court freer to consider antitrust issues for itself, this problem-or the perception that there is a problem-might disappear.

Were Congress to act favorably on proposals for an Antitrust Court,'®? the CAFC's experience should be kept in mind.

The legislature ought to exploit the natural lines of antitrust litigation to give the special court a broad enough view of competition issues to allow it to write persuasive opinions in its field of involve voluminous evidence, they can be burdensome on trial courts. However, he would oppose a specialized court because of the inconvenience to suitors (or circuit-riding judges) and because the expertise of the court would be eroded in jury trials. Some of these objections would be less potent if the antitrust court were an appellate bench, although a fact implicit in Judge Friendly's analysis is certainly true; in antitrust law, it is finding the facts that is problematic, and therefore a trial-level court would, in some respects, be more helpful than a specialized appellate forum. Jd. important sector of the economy without arousing significant public concern.

The distinguishing factor among these courts is the extent to which there was public agreement on their missions. With the unsuccessful courts, this consensus was lacking. The Commerce Court was established during the infancy of the regulatory state, at a time when the role of the federal government in utility regulation was hotly debated. Even those who concurred in the need for regulation had little sense of how the interests of the eastern states should be balanced against those of the west, or how the needs of shippers should be weighed against the demands of railway owners.!®* Similarly, the FISCs magnify the longstanding debate between the sacrifice of basic guarantees of personal freedom and national security interests. In contrast, the ECA was created in direct response to the nation's entry into World War II, when the nation was fairly united in its willingness to make sacrifices to win the war and there was general opposition to the notion that scarce goods should be allocated solely on the basis of purchasing power. The Special Court's mission-to save railways from destruction-was, similarly, a popular goal.'

Upon reflection, it is not surprising that consensus on the goals of the legislation administered by the specialized court influences its success. A new generalist court does not need public agreement to gain public trust. The heterogeneity of its docket makes it likely that there will be some issues that it can handle in a manner that validates the bench; indeed, the court need only receive and follow precedent to gain a measure of public acceptance.'7° When the court then addresses controversial questions, its resolution may not be accepted, but its reputation will not be impugned. In contrast, when a specialized court is established in an area where there is no consensus, there is nothing for the public, practitioners, or other courts to measure its rulings against, and it becomes an easy target for those who disagree with its decisions.

For this reason, certain proposals for specialized courts are 168. F. FRANKFURTER & J. LANDIS, supra note 2, at 153. 169. See also Jordan, Specialized Courts: A Choice?, 76 Nw. U.L. Rev. 745, 760-62, 765 (1981). 170. See, e.g., Mintz, Patent Appeals Now Centralized in New Circuit Court, Legal Times, Nov. 29, 1982, at 12 (commending the CAFC for its decision in South Corp. v. United States, 690 F.2d 1368(Fed. Cir. 1982) unlikely to succeed. Consider, as an example, the suggestion for a Science Court.'"! The proposal is defended on the ground that generalist judges lack the expertise and experience to decide scientific issues. But much of the discontent with the manner in which scientific disputes are adjudicated arises because issues that pass as scientific questions often involve policy choices as to which there is no public agreement. The Delaney Clause,!" which prohibits the use of food additives that induce cancer, is illustrative. Many of the disputes generated by this prohibition sound like scientific issues: Are animal tests for carcinogenicity appropriate? How should high-dose data be extrapolated to human dosage levels? But some of the core questions at issue here are whether prohibitory rules make sense in the face of so much uncertainty, whether significant resources should be expended on quantifying one risk to human health when so many riskier activities are undertaken on a daily basis, and who should decide which risks are acceptable. A Science Court could decide these issues, but without public consensus on its basic premises, there is no assurance that the public would accept its conclusions.

Since specialized tribunals take less time than other courts to resolve open controversies definitively, and on a nationwide basis, consensus is even more important. During the time that generalist courts debate an issue, the public may grow to understand both sides of the question better. More important, experience under disparate circuit rules generates data that can be taken into account when the issue is finally resolved, be it by Congress, the Supreme Court, or defacto agreement among the circuits.'7* Once achieved, the ultimate resolution is more likely to enjoy public understanding, if not widespread acceptance. Specialized courts, like the ECA and the Special Court, must, in short, begin with consensus. The experiences of the FISCs and _ the Commerce Court show that it cannot be developed by the courts themselves.

This has important consequences for some proposals. Specialized adjudication is not a substitute for forging public agree-171. See, e.g., Kantrowitz, Proposal for an Institution for Scientific Judgment, 156 Sci. 763 (1967);Martin, The Proposed "Science Court," 75 Micu. L. Rev. 1058(1977). ment on a legislative agenda; it cannot be used, for instance, as a substitute for the legislative veto-as a check on agency authority when Congress cannot manage to make controversial decisions. Thus, an Environmental Court!" will be unacceptable if it is animated by a desire to curb freewheeling agencies while avoiding the political consequences of balancing risks to future generations against present profit maximization. A court established in such a legislative atmosphere would decide the tough questions when presented in litigation, but its decisions may not end controversies in a helpful way. Those who disagreed with its decisions would enumerate the many dangers of specialization, and it will not be possible to decide whether the court properly used its expertise, or was captured by the side that won.'"

The proposal for an Administrative Court stands on a more secure footing.!7* Although such a tribunal may be charged with the responsibility of reviewing agencies whose activities are controversial, its docket could be made broad enough to include jurisdiction over matters on which there is broad public consensus.

Decisions from that portion of its docket could be evaluated against agreed principles, and these decisions would allow public confidence to build. A court with a solid public reputation and the necessary expertise could abate much of the controversy over judicial review of agency action.

Public consensus may also be important to the judges who staff a specialized court. The repetitive nature of such a court's docket will always make appointment somewhat undesirable to the nation's best legal minds. Experience with past specialized courts demonstrates, however, that when the tribunal is organized to further an important public objective, extraordinary people do agree to serve. Indeed, anecdotal evidence suggests that appointment to a specialized bench composed of handpicked ju-174. See, e.g., Whitney, The Case for Creating a Special Environmental Court System, 14 Wm. & Mary L. Rev. 473 (1973).

175. See also Jordan, supra note 169, at 765 (noting that lack of national consensus in environmental law and health and safety regulation makes it unlikely that specialized courts will work in those areas).

176. See, e.g., Nathanson, The Administrative Court Proposal, 57 Va. L. Rev. 996 (1971) (suggesting an article I court to review administrative agencies with direct review in the Supreme Court); Cooper, The Proposed United States Administrative Court, 35 Micu. L. REv. 193 (1936) (suggesting article III court combining Court of Claims, CCPA, Customs Court, and Board of Tax Appeals into specialized court with a trial and appellate division with jurisdiction over matters handled by these courts and other agency action); H. FrieNDLY, supra note 8, at 177-90 (discussing several models).

rists can be a satisfying experience. The challenge in devising a common law to implement national policy is attractive, especially when there is general agreement on the policy. The ECA's judges, for example, spent their war years hearing cases in sixtyfive different cities while balancing the demands of their own circuits. Undoubtedly, one of the reasons these jurists were willing to devote so much of themselves to the enterprise was that their cause was generally appreciated. d. Distribution. Another indicator of the success of specialization is how the specialized court alters the distribution of cases within the judicial system. One criterion that bears scrutiny in this regard is how the specialized cases are currently dispersed. If the cases are distributed among the regional courts such that each court hears only a small number each year, the advantages of channelling all the cases into a single court will be substantial. The regional judges' lack of familiarity with the issues is likely to make adjudication time consuming and prone to error.'"" Concentrating the cases into a single court should produce a bench with the motivation, expertise, and opportunity to create a better crafted, more stable body of law. Where, however, existing courts already hear enough cases in a field to possess these advantages, specialization will not be as effective.

A second distribution factor to consider is the extent to which administration of the law is currently fragmented, or would become fragmented through specialization. There are two interrelated phenomena here. The first is disuniformity, a problem that arises when several fora independently interpret and articulate the law. If these courts have no power to overrule each other, dissimilar-and perhaps contradictory-rules will be put into force in different locations. This may create a perception of unfairness as litigants around the country are treated differently. More important, it makes it difficult for intercircuit actors to predict what law will apply to them. The second phenomenon, incoherence, is slightly different. It occurs when authority over different aspects of the law is spread among non-hierarchically related courts, each seeing too little of the picture to know which overall objectives to fulfill, or how to achieve them. If the field is sufficiently complex, the law becomes internally inconsistent because various doctrines within the law work at cross purposes.

The influence of these factors is demonstrated by the CAFC's experience in its patent docket. Formerly, patent disputes were spread among the regional courts, the CCPA and the Court of Claims. No regional circuit was likely to hear more thantwenty patent cases in a single year;'"* the CCPA mainly handled issues like patentability, which surfaced in patent issuance litigation, while the regional courts handled enforcement questions, such as patent misuse. Gathering all these cases into a single court produced a critical mass that improved the capacity to make the law uniform and to knit the conceptual strands of patent law into a coherent whole without Supreme Court involvement.

But not every branch of the law suffers from the distribution problems that formerly afflicted patent cases. Furthermore, in not every field is uniformity an important a value. Since consumers of patent law are intercircuit actors who base their decisions to invest in innovation upon their expectations of achieving patent protection, the uniformity, predictability and coherence achieved by the CAFC is especially prized. When these factors are not so compelling, the benefits of creating a specialized court may not outweigh the costs associated with isolating its field from the mainstream of:adjudication. It is, for example, difficult to perceive the advantages of a Social Security Court to hear appeals from the decisions of the Social Security Administration (SSA).!"° The agency, which has traditionally enjoyed far greater deference than that accorded to the Patent and Trademark Office, synthesizes the law in its field. This makes a specialized appellate bench unnecessary for coherence purposes.!*° Although uniformity is endangered by having the regional courts entertain appeals of SSA decisions, uniformity-albeit important to create horizontal equity!*!-does not have the same claim as it does in patent litigation. People do not decide to become disabled in reliance on the acquisition of benefits, thus predictability is less significant.!® Moreover, claimants are not intercircuit actors. Accordingly, equivalent treatment across the country is not a relevant consideration.'** Finally, because social security cases exist in profusion on all circuits, there is no need to funnel cases into a single court in order to achieve a critical mass.1* 2.

There are three constituencies whose interests are directly implicated by the establishment of a specialized bench: the parties whose rights are adjudicated by the new tribunal, the bar that practices before it, and the jurists who serve on it. Characteristics of each group influence the extent to which specialization can be an effective and acceptable means for resolving disputes.

a. The parties. The conditions that make specialization beneficial to the parties have been discussed above. It was noted that the extent to which the parties require predictability in the law determines the significance of the stability produced by a specialized bench. Intercircuit actors who make decisions in reliance on legal rules benefit more from specialization than do localized actors who look primarily to the law's compensatory aspects.

The cost side of the equation is determined in part by wealth, or more precisely, the parties' relative wealth. It may seem surprising that relative wealth is a factor that should influence the success of specialization, for under a legal regime in which each party pays its own way, all adjudication is arguably tainted when resources are distributed asymmetrically. Indeed, it might be expected that the expertise of the judges on a spe-182. See also Friendly, supra note 162, at 639 (arguing that specialization would be more valuable to consumers of patent law than to criminals because the latter "do not plan their activity with an eye fixed on the Bill of Rights, the Federal Penal Code, or the rules of evidence applicable in criminal trials"). 184. See also Ogilvy, The Social Security Court Proposal: A Critique, 9 J. Lecis. cialized court would compensate for the inability of poorer litigants to hire superior counsel.'*® At the same time, however, a move from allocating cases among courts by geography to an allocation based on subject | matter is burdensome to litigants, and the rich can cope better than the poor. Under a geographic allocation, a tribunal is usually near the parties' home where witnesses and evidence are close at hand; when this result is not obtained, due process arguments and doctrines such as forum nonconveniens are available to transfer venue relatively inexpensively.'** In contrast, proposals for subject-matter courts often contemplate establishment of a single tribunal that sits permanently in one location-usually Washington, D.C.'*' The farther the parties must go to reach the specialized court and the more the resolution of their dispute turns on geographic factors, the greater the disparity in the litigants' resources will matter.'%* Consider in this regard the proposal for a centralized Environmental Court. In some respects, it makes sense to channel challenges to actions of federal agencies that affect the environment, EPA enforcement suits, and civil litigation raising environmental issues into a specialized trial-level court.'®® The factual questions arising in such case$ are often technically abstruse, making the expertise of the bench a valuable asset.

185. Note, however, that this was not the experience of the Court of Private Land Claims, where richer litigants were able to hire counsel to help the bench negotiate its way through Spanish and Mexican law. See supra note 147. Perhaps this court was not in existence long enough for its judges to become expert. If so, the lesson is that relative wealth is particularly problematic in the early years of courts specifically established to deal with intricate bodies of specialized law. 186. See also 28 U.S.C. § § 1404, 1406 (1982). 187. For example, in seven years, the CAFC has held sessions outside Washington on only 13 occasions. See supra note 140.

188. Such a requirement may even raise due process concerns. Cf. Yakus v. United States, 321 U. S. 414, 437 n.5 (1944). This case upheld the Emergency Price Control Act against a challenge based on the inconvenience of having to protest actions under the Act in Washington, D.C. It should be noted that although the argument was rejected, it was not dismissed out-of-hand. Rather, the Court relied upon the national emergency that necessitated "a centralized, unitary scheme of review," on the fact that most protests were on a paper record, and on the willingness of ECA to hold sessions in other cities. In addition, the travel problem is more severe when the specialized court is the tribunal that finds the facts, for it is at this stage that proximity matters most. In Yakus, the Office of Price Administration developed the record presumably in a manner that did not create undue cost or inconvenience to the defendant. 189. Civil litigation would also include diversity cases such as, Zahn v. International Paper Co., 414 U. S. 291 (1973), and cases in which the state is a party. See, e.g., Illinois v. City of Milwaukee, 406 U.S. 91 (1972).

Since these cases are complicated, and are sometimes presented as class actions, their removal from the dockets of the district courts would help diminish caseload pressures. In addition, coalescing these cases into a single tribunal would create a critical mass and bring together a group of issues that could profitably be heard together. Having these questions decided definitively by a single tribunal would be especially helpful to multicircuit environmental defendants.

Environmental disputes are, however, the prototype of controversies that benefit from being litigated where they arise because geographic features such as wind, landscape, and water currents can play a significant role in their adjudication. The cost of bringing experts and other witnesses to the bench would be great, and since the defendants in such disputes are more often large enterprises and the plaintiffs mostly residents of the locality where the claim arose, the difference between their financial resources will systematically affect the ability of the plaintiff, or the public's perception of their ability, to litigate.1®°T he problem could be avoided by having the court hear cases around the country. If, however, the bench is kept small enough to ensure collegiality and uniformity, travel by the judges may be required. This solution has proved unpopular in the past.?®?

Another attribute of the parties that contributes to the costs of specialization is their relative sophistication in legal matters. Parties who litigate often, or who belong to tightly allied groups with parallel interests, have an advantage over infrequent users of the judicial system. Repeaters can more easily capture a specialized court because their experience teaches 190. Currie & Goodman, supra note 8, at 8-9, argue that proximity is overvalued in the modern era because the cost of travel is now low. It should be noted, however, that these authors compare the cost of traveling to a district court to the cost of traveling to an appellate court in the same circuit. Since the-distance between a district court and a centralized court will often be greater than the distance to the circuit court, travel will be more expensive. If the specialized tribunal is located in the District of Columbia, the costs of overnight accommodations are also likely to be higher than these authors anticipate. them how the court operates, how its judges think, and the kinds of arguments and issues that appeal to them. Repeaters also possess certain strategic advantages. They can coordinate their policy arguments and may be able to wait for a sympathetic case to push the positions they favor. Infrequent litigants cannot pick their cases so strategically, and therefore may have fewer opportunities to frame the issues to their advantage.

Of course, not all specialized courts would suffer from an imbalance in the sophistication of the parties. For example, the CAFC has not had this problem. Many of its litigants are vertically integrated firms that both manufacture and conduct research and development. The same parties may therefore appear on either side of infringement actions, and therefore lack the motivation to sway the court in any particular direction. Other proposals for specialization, however, involve subject matter where imbalance is likely to occur.1®? The experience of the Commerce Court shows that when it does-or when the public perceives that it has-the specialized tribunal cannot function effectively.

b. The bar. To some extent, the composition of the bar may compensate for imbalance among the litigants. If the bar practicing in a particular field is well organized and sophisticated, the fact that some litigants are one-timers may not matter. One example (not derived from a specialization proposal) is the death penalty bar; another, practice within some of the jurisdictions of the Claims Court (formerly the Court of Claims). These groups can strategize and orchestrate litigation in ways that compensate for their clients' lack of organization.

Unfortunately, reliance on the bar is not a complete answer to the problem of imbalance. First, it works only if the clients have access to these sophisticated legal services. This access often turns on the type of case involved. In the death penalty cases, the severity of the clients' predicament has generated a substantial volunteer effort; in the Claims Court area, litigants are mainly commercial enterprises embroiled in contract'®* and tax disputes with the government, or well organized government 192. The Social Security court is a prime example. See infra text accompanying notes 195-96.

193. The American Bar Association, for example, has a regular section on Public Contract Law, which prepares seminars, identifies current issues for its members, and develops litigation strategies. See, e.g., ABA, SECTION or PusLic Contract Law, New DEVELOPMENTS IN Data Ricuts (1985).

employees.' In contrast, claimants before the proposed Social Security Court are not so fortunate. They are mainly one-timers with no internal association, are unlikely to attract an organized voluntary corps, and may be unable to pay for equivalent representation.'® These claimants meet a formidable opponent in the SSA, whose appearance in every case would give it advantages in a specialized court that are far greater than those of a mere repeater. Even if the court managed to avoid capture by the SSA, the appearance of prejudice may condemn the enterprise.!**

A second problem with counting on the bar to compensate for the weaknesses of the parties is that the sort of mastery this reliance contemplates may increase other risks of specialization.

Isolation of the tribunal can be partially countered if those who appear before the court have practices outside the special field and can present arguments that draw upon developments in other areas. The more the bar organizes and specializes, the more internally focused it becomes, and the less likely it is that practitioners will keep abreast of developments elsewhere.

c. The bench. One criticism frequently leveled at proposals for specialization is that these tribunals will not be able to attract high quality jurists. The usual reason given is that the court's business will be too boring or too repetitive to hold the interest of a creative mind for long.'®? To this may be added another reason: specialized judges may be called upon to ride circuit as a way to avoid the access problems noted above. Yet the reputation of the judges who serve a court can be vital to its perceived success.

The personnel problem has long been recognized. Eminent figures were chosen for both the Court of Claims and the Court of Customs Appeals, and while it is unknown why these people chose to serve, it is noteworthy that Congress set their salary above the regional benches.'®* Since salary differentials may no longer be an acceptable way to attract good people to new specialized benches, proponents of specialization must consider other ways to improve the candidate pool.

The authors of Justice on Appeal' recommend solving this problem by staffing specialized courts with judges from the regional fora, who would serve the special bench on a part-time basis.129 Used in connection with the ECA, the TECA, the FISCs, and the Special Court for Rail Reorganization, this approach has many advantages. The judges' service on regional courts prevents them from becoming isolated or bored; a court composed in this manner is flexible because the size of the bench can be adjusted to reflect demand. The specialized assignment brings together, on a sustained basis, judges who would not normally sit together. Anecdotal evidence suggests that this is personally rewarding and it possibly facilitates cross-fertilization 197. Prestige is another factor often cited to advance the proposition that certain courts will not attract the best minds. See supra note 8. For reasons discussed there, I do not address that concern, except to note that it provides a justification for distributing specialized responsibilities evenly among all federal judges. within the federal system, promoting national uniformity in areas outside the field of specialization.?© If the entire federal bench is eligible for service, appointment may come to be viewed as an honor-a nonmonetary form of compensation.2"

The challenge in such a process lies in assuring that the method of selection does not (perhaps inadvertently) systematically bias the court.?°? Although this issue has never been raised in connection with the courts currently utilizing regional appointees, the criticism could be made, for in many cases the Chief Justice of the Supreme Court has a free hand at appointments. Lack of statutory guidance could result in appointments of jurists who share the Chief Justice's judicial philosophy, or at least the Chief's view of the legislation that the court administers. The statute that established the Special Court for Rail Reorganization is somewhat of an improvement over this procedure because it relies on the Multidistrict Litigation Panel. However, the problem is only once-removed since the Panel is itself chosen by the Chief Justice.? Random selection would be much better.? Although this would diminish the honorific dimension to an assignment, it would safeguard the court's independence. Indeed, if several new specialized courts are established and assignments to them are made temporary (as, for example, the seven-year assignments to the FISCs), the duty of special service would eventually be spread over the entire federal judiciary. The shared responsibility for special dockets would effectively counter the argument that special benches are inferior.

An alternative to the part-time judge is the procedure Congress used in creating the CAFC. The judges of that court have near exclusive jurisdiction over patent law, but isolation is avoided because of their responsibilities in other areas, such as contract and labor law. There are several advantages to this method. The judges gain the benefit of a more varied docket but are not torn between the conflicting demands of two sets of colleagues. Although they might be required to travel to accommodate litigants, they do not commute between assignments, as do some part-time judges. Furthermore, the bench's long-term commitment should inspire greater collegiality and encourage the judges to hone their expertise.

The propensity toward isolation and capture could be countered more effectively than it has been for the CAFC. There, even nonpatent matters tend to keep the judges out of the jurisprudential mainstream because most of the court's responsibilities are in highly specialized areas, and many are within the court's exclusive power. Instead of widening the docket in that way, specialized courts could be given authority in fields where a fruitful interchange between the special court and the regional circuits is especially likely, and where the court's view of the policies furthered within its special jurisdiction is challenged.?" This interchange would broaden its vision and, equally important, dilute the clout of interest groups in the appointment process. If Congress were to treat many subject areas as specialized, it could allocate to each regional court one (or more) specialty." By spreading the burden of specialization, all federal judgeships would become equally desirable.

In the final analysis, it is important not to exaggerate the potential problem with attracting qualified judges to serve on specialized courts. However persuasive these problems may seem in theory, the quality of the judges willing to serve has, in the past, been high. And Congress can take other action to enhance the professional value of a judgeship. Appellate judges on the special courts could be more regularly chosen to sit by designation on the regional circuits; specialized trial judges could be given assignments by the Multidistrict Litigation Panel. The President could also make a practice of nominating special judges to regional judgeships after a suitable tenure on the specialized court.?°' If the judges assume a more important role in the judicial process, the public would be more confident that the 205. One example is to give the CAFC power over antitrust law, which would make a nice contrast to its patent jurisdiction.

206. In the CAFC example, to make room for antitrust jurisdiction, power over matters arising under the Contract Dispute Act, or review of the decisions of the Merit Protection Review Board could be assigned to regional courts.

207. Alex Kozinski, for example, moved from the Claims Court to the Court of Ap-peals for the Ninth Circuit. Taz AMERICAN BENCH 48 (5th ed. 1989-90).

President and the Senate thought rigorously about candidates' general suitability to the judiciary.

There are many ways to structure a specialized court. Possibilities include legislative specialized courts and constitutional specialized courts; specialized courts with generalized judges, generalized courts with exclusive special jurisdiction, and panels with categorical case assignments." Even within these broad paradigms, there are variations: specialization at both the trial and appellate levels; specialized trial courts with general appellate courts; or general trial courts reviewed by specialized appellate courts. The judges can be chosen in different ways and the court can be given exclusive, nationwide jurisdiction over the special subject matter or lesser authority. In some models, litigants can be given a choice whether to use a special or general court. Since each of these strategies balances the benefits of specialization against the costs somewhat differently, the viability of specialization is a function of the manner in which the proposal is implemented. The relative advantages of many of these strategies have been discussed at several earlier points in the article; they are summarized here.

a. The locus of specialization. One of the most significant determinants of success in specialization is the level in the judicial hierarchy where the special bench is located. The choice should depend upon the problems in the field that led to the decision to adopt the strategy, with attention to those systemic problems that are exacerbated by the use of such courts.

Specialization at both trial and appellate levels is an attractive option from several perspectives. In fields where specialization is thought desirable because of the need for expertise, dual specialization may work best because the parties obtain the benefit of proficient factfinding, and the trial court's expert interpretation and application of the law is not diluted through insensitive review. Specialization at both levels is particularly advantageous in fields where normal judicial procedures must be modified to attain specific goals. The new courts can use the revised procedures in all their cases, without disturbing the operations of the remainder of the judicial system." Finally, when 208. See P. Carrinctron, D. MEapor & M. RosENBERG, supra note 1, at 167-79. 209. For instance the FISCs use special procedures to keep their docket confiden-timeliness is important, dual specialization promotes the most rapid decisionmaking.?'°D espite these benefits, this strategy has been used sparingly. The FISCs are organized in this way and in some of its specialty areas the CAFC reviews tribunals that are themselves specialized. Congress' reluctance to use dual specialization is not, however, surprising, for it carries serious disadvantages. The possibility of doctrinal deviation is significantly increased when there is no generalist input into a case. Furthermore, there is nothing in such a system to counteract the risk of capture. The FISCs arguably require dual specialization because of the security problems that would accompany review of foreign surveillance warrants in regional courts. However, careful implementation of a single-level model can often embody the best of both worlds-providing generalists to keep pronouncements in line with jurisprudential trends and specialists to contribute expertise.

The question, then, is whether specialization should be used on appeal or at the trial. Where uniform, predictable, stable law is the motivation for specialization, or where the law is so complex that expertise is required to interpret it and elaborate upon it, specialization at the appellate level is preferable. An appellate court has plenary power to reverse lower court determinations on the law. If it is given exclusive jurisdiction over a field, it can make needed doctrinal innovations and yet attain uniformity without Supreme Court involvement. Since not every case is appealed, an appellate bench can have nationwide jurisdiction without sacrificing the collegiality necessary for stability. Appellate level specialization is also better when geographic proximity is an issue because there is less need for the litigants to appear at the appellate stage of the proceedings, and the appellate court does not usually engage in fact-finding.

When the facts of the case give rise to the problems that require specialization, it is the trial court that should be specialtial. Cf. H. Duprorr, THE Unirrep States Tax Court, AN Historic ANALYSIS 207-08 (1979) (noting that the court was not converted from a legislative into a constitutional court because of the Treasury Department's unwillingness to conform to the procedure in the federal courts whereby the Justice Department represents administrative agencies). 210. See supra text accompanying notes 62-99. This was 'a consideration in establishing the ECA and the TECA. Furthermore, the choice of specialized jurisdiction over tax deficiencies was influenced by the time-value of the judgments. Cf. H. DuBrorr, supra note 209, at 395-97. ized. A specialized trial judge has the proficiency necessary to evaluate the evidence and the expertise to make the best use of the managerial tools provided by the federal rules for complex case."* In jury-tried cases, the judge's skill would be an asset in instructing the jury and determining the admissibility of esoteric evidence. Generalized review would assure that like issues are treated the same way in all contexts, and that the law in the special area is kept within the legal mainstream. b. Status. The courts discussed in the earlier portions of this article are staffed by judges protected by the guarantees of article III, but the federal system has long used other tribunals, such as legislative courts and agency-based adjudicators, to achieve many of the objectives sought through specialization.?" Thus, another point to consider in structuring a specialized court is whether a constitutional court is necessary-or constitutionally required.

In many ways, nonconstitutional courts represent the better approach. Consider, for example, a field regulated by an administrative agency in which adjudication is conducted by an independent tribunal intra-agency, subject to review by a generalist court, either district or appellate.?'* From a pragmatic point of view, agency adjudication is attractive because assignment of the judges to specific agencies means they can concentrate on their field and develop a strong understanding of its intricacies. In addition, the judges would not be diverted to other administrative cases, as could more generalized Administrative Court judges.?" 211. See,e.g.,Fep. R. Civ. P. 16,53;Fep. R. Evin. 201,706. 212. The conceptual difference between the use of free-standing legislative courts and administrative adjudication is beyond the scope of this article. For brief discussions of the ambiguity, see H. Dusrorr, supra note 209, at 113; Turner, The Tax Court of the United States, its Origin and Functions, in THe History AND PHILOSOPHY OF TAXATION 31, 38-46 (1955). -36 (special supp. Apr. 6, 1989). This is basically the one-tier model discussed extensively by Currie & Goodman, supra note 8, at 1-61. For purposes of this discussion, the reviewing role of officials such as commissioners is ignored. Such review procedures, though common, are not consistent with rigorous independence so long as the reviewing officials also possess policy-making power. Psychologically, placement of the court within the agency would remind reviewing courts to be vigilant for signs of capture. It may also diminish public suspicion of an institution called a "court," which is structured differently from the regional fora.?*® Administratively, one advantage would be that facilities such as specialized libraries could be shared.?!* More important, whenever Congress expanded the administrative scheme and gave the agency new funding, its adjudicatory division's budget would also increase. Thus the court would automatically receive the resources needed to meet the new demand for adjudication.

The flexibility offered by legislative and agency models can be especially advantageous. The current docket crisis obscures the fact that in any given field of law, the demand for adjudication varies over time.?'7 Benches that are not protected by life tenure can be made smaller if litigation activity in the specialty field diminishes. In addition, these courts may avoid the counterproductive incentive structure that is created by the establishment of a permanent article III court. This problem is demonstrated by considering the specialized court that has successfully synthesized and articulated the law in its special field. In theory, the business of such a court would decline dramatically because greater stability in the law leaves less room for dispute and greater predictability facilitates settlements. But if the docket declined enough, it may become more economical to transfer the few remaining cases back to the regional circuits. If so, the special judges would face a lifetime of sitting by designation around the country, a burdensome proposition, or reassignment-perhaps to courts and dockets they liked less.?** To avoid cial Nature of the Tax Court Be Recognized?, 24 Gro. Wasu. L. Rev. 619 (1956); H.

DusBrorFF, supra note 209, at 202. 215. Cf. H. Duprorr, supra note 209, at 168 (even after vesting Board of Tax Appeals with many of the powers of a court, it "still had a name which obscured the nature of its activities. As a result many taxpayers who were unfamiliar with the Board assumed it functioned like an administrative body rather than like a court."); id at 171 (noting that Congress hesitated to make the Board a court because its lack of juries, specialized nature, narrow jurisdiction, and jurisdiction limited to cases in which the Commissioner of Internal Revenue was a party made it different from traditional courts). Similar considerations led the United States Judicial Conference to oppose changing the status of the Tax Court from an article I to an article III court. Jd. at 212.

216. Technical assistants could also be shared, although this might compromise the appearance of independence.

217. 218. The Commerce Court judges suffered this fate; the judges of the CCPA took that fate, the court may begin to modify the law it administers, to widen substantive rights or alter procedural gatekeeping rules to admit new claimants."*® While the increased supply of adjudication might be welcome to litigants and the bar, the result would run counter to one of the objectives of specialization, namely decreasing the size of the federal docket.2?°O ne problem with relying on legislative courts or agency adjudicators is that the federal system regards life tenure and salary guarantees as the best assurance of judicial independence.

To provide similar insulation between article I courts and the legislature, Congress has relied upon long terms and a tradition of virtually automatic reappointment." Ironically, in some ways pay cuts to avoid a similar future.

219. The past history of specialization does not furnish evidence that this has occurred. However, there are some indications that the converse is true; that pressured courts tend to reduce substantive and procedural rights. See, e.g., Finley v. United States, 109 S. Ct. 2003(1989) (refusing to extend the doctrine of pendent jurisdiction to include pendent parties, even when the federal claim is within the exclusive jurisdiction of the United States); Zahn v. International Paper Co., 414 U.S. 291 (1973) (requiring every individual in diversity class actions to have a claim meeting the statutory amount in controversy). Neither of these results are constitutionally required, and, if state courts are taken into consideration, both results are inefficient since they multiply lawsuits and increase the number of times that the same evidence must be considered. Since the best that can be said of both cases is that they might reduce the workload of the federal system, the conclusion that they were motivated by docket problems is appealing.

Recent decisions in the antitrust field display a similar pattern. See, e.g., Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574 (1986) (making it more difficult to prove predatory pricing conspiracies); Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979) (narrowing monopolization offense), cert. denied, 444 U.S. 1093 (1980); Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) (shifting vertical restraint analysis from per se rules to rules of reason, thereby making it harder to prove case); Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477 (1977) New Evinence, New Learninc (L. White ed. 1988). One wonders whether Judges Bork, Easterbrook and Posner, had they been appointed to an Antitrust Court rather than to the regional circuits, would have pursued this path to its obvious destination. But see Hershey, Where "Temporary" Has Lasted 13 Years, N.Y. Times, Nov. 5, 1984, at B10, col. 4 (noting that Congress has been slow to disband specialty courts). 220. It may also be possible for Congress to staff a specialized article III court with judges ineligible for assignment on a regional court. See Northern Pipeline Constr. Co. v.

Marathon Pipe Line Co., 458 U.S. 50, 74 n.28 (1982). As noted above, this strategy would not be desirable from an acceptability standpoint; psychologically, it is equally likely to produce a bench bent on keeping itself busy. this mechanism provides less flexibility than a constitutional court established on the part-time model. After World War II, for example, the judges of the ECA simply returned to full time service on their original circuits. Furthermore, it is important to remember that the constraints of article III are sometimes constitutionally mandated.

Although the precise limit of Congress' power to steer cases to legislative tribunals has not been delineated, the Supreme Court has apparently rejected the notion that the requirements of article III can be overcome by simply establishing a need for specialized adjudication.??* And while some of the proposals for specialized courts involve cases that arguably fall within some of the exceptions that the Court has recognized to article III, these exceptions are construed narrowly."?* Even where they do apply, they may not cover the entire jurisdiction of a specialized court. This is particularly true if the court is organized according to the suggestions made here, and given power over all claims transactionally related to the issue requiring special treatment."

Nonetheless, the contours of the constitutional limitation on legislative courts have not been fully defined; since the advantages of such courts are substantial, the strategy should be fully explored when specialization is considered.?*> The Board of Tax 954, 83 Stat. 487, 730. Members of the Board of General Appraisers had life tenure. Act of May 27, 1908, ch. 205, § 3, 35 Stat. 403, 406. See also H. Dusrorr, supra note 209, at 212 n.325 (noting that since 1924, only three members of the Tax Court and its predecessor, the Board of Tax Appeals, have ever been refused reappointment).

222. See, e.g., Northern Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U.S. 50 (1982). Cf. Granfinanciera S. A. v. Nordberg, 109 S. Ct. 2782(1989) (using jurisprudence developed in connection with article III requirements to determine right to trial by jury).

223. See, e.g., Granfinanciera, 109 S. Ct. at 2796. 224. The Supreme Court gave some credence to the notion that if a specialized claim could be adjudicated legislatively, then counterclaims arising from the transaction upon which the claim is based could also be adjudicated by the legislative court. See Commody Futures Trading Comm'n v. Schor, 478 U.S. 833 (1986). However, that case may be sui generis. So far as article III protects the personal rights of litigants, the interest had been waived. Id. at 848. Furthermore, there were reasons to believe the structural interests protected by article III were not endangered. Id. at 850-54 (citing particularization of the law, review of factual determinations on a "weight of the evidence" test, right to de novo review of legal determinations, and fact that claim was based on state law).

Nor is it possible to assess the vitality of Schor after Granfinanciera.

225. Nor is the constitutionality of article III specialized courts entirely clear. For example, the temporary nature of the appointment to some of these courts, and the right of the Chief Justice to make (and presumably change) assignments gives the specialized judiciary so much less protection than that accorded regional courts that a colorable argument could be made that part-time courts are not constitutional tribunals. Although no such challenge has been raised, there is little assurance in the courts being staffed by Appeals and the United States Tax Court represent experiments with this model;??* they have both received high praise for their work?' and for the flexibility they give to Congress.?2 article III judges since it is now clear that such judges can be assigned certain kinds of nonjudicial business. Mistretta v. United States, 109 S. Ct. 647 (1989). Furthermore, just as the Supreme Court has countenanced certain article I tribunals despite acknowledging the problem that would be posed if a "phalanx of non-Article III tribunals" were created, Commodity Futures Trading Comm'n v. Schor, 478 U.S. 833, 855 (1986), it may perceive more structural problems in segregating the entire federal docket than it sees when only a small portion of the federal power is withdrawn from the district courts. Cf. Lockerty v. Phillips, 319 U. S. 182, 187-88 (1943) It should be noted, however, that the Board of Tax Appeals met a series of obstacles that would have to be overcome if agency-based adjudicatory models were widely adopted. For example, the decision to establish an adjudicatory mechanism within the Treasury led to questions about the neutrality of appointments. See, e.g., H. Dusrorr, supra note 209, at 146-50 (noting that Congress, concerned with pro-government bias, considered a requirement that the major political parties receive equal representation, and a prohibition on service by former members of the Treasury). Both proposals were, however, rejected and the President was given plenary authority to appoint members of the Board, Revenue Act of 1926, ch. 27, § 1000, 44 Stat. 9, 106). For many years, the Board's work suffered because it lacked the procedural powers necessary to affect justadjudications. See H. Dusrorr, supra note 209, at 114 (power to subpoena witnesses and evidence and order depositions); id. at 168 (power to use federal procedure and evidentiary rules, to dismiss cases for procedural defects and to impose filing fees); id. at 168, 170 (power to enforce subpoenas and punish contempt); id. at 479-81 (authority to declare provision unconstitutional, citing Independent Life Ins. Co. of America v. Commission, 17 B.T.A. 757 (1929), aff'd, 67 F.2d 470 (6th Cir. 1933), rev'd on other grounds and without reaching the propriety of the Board's action, 292 U.S. 371 (1934)); id. at 484-86 (equitable power to entertain action for recoupment); id. at 488-93 (power to apply doctrine of equitable estoppel). The Board's failure to be named a court may have made it difficult to attract high caliber appointees. Id. at 206 (noting that the Tax Court's status as an article I court made it more difficult to attract good candidates). Its status even led to difficulties in reserving hearing rooms at courthouses in the cities where it traveled.

Id. at 178 (noting that courtrooms are unsuitable to administrative hearings which attract noisy crowds). Until the decision in Dobson v. Commissioner, 320 U.S. 489 (1943), it was unclear what standard governed the review of its decisions; after enactment of the Administrative Procedure Act, there was confusion as to its applicability to the Board. See H. Dusrorr, supra note 209, at 191-92. Finally, the Board's nonjudicial status made ¢. Exclusivity. Many proposals for specialized courts assume that the new tribunal will have exclusive or near exclusive nationwide authority in the specialty field, a procedure that is clearly advantageous when the reason for specialization is to achieve uniformity. The specialized court's pronouncements will be final, except when reviewed by the Supreme Court.?*° Thus, even without Supreme Court intervention, a single rule will be in effect across the nation."*° .

Exclusivity is not, however, the only possible deployment of specialized benches. Uniformity, for example, can also be achieved by making the decisions of the specialized courts binding on fora with overlapping power. Although the federal system does not usually require courts in equivalent hierarchical positions to defer to one another, this approach has been suggested in other contexts.?3! It would enable the special bench to establish itself in a given city, while providing a forum for litigants who do not wish to travel. In addition, it would allow the system to accommodate more cases in the specialty area without expanding the specialized bench or endangering its collegiality.

When uniformity is not the reason for specialization, neither exclusivity nor deference is conceptually required. In that case, the possibility of giving litigants a meaningful choice between a specialized court and a court of general jurisdiction should be explored. Tax law is essentially administered on this model. Taxpayers who disagree with the Internal Revenue Service's assessment of liability can pay the assessment and sue for a refund it impossible for it to restrict appearances to lawyers, thereby reducing the likelihood that the judges' vision would be extended by the arguments they heard. 229. Judge Friendly suggested that the Supreme Court accord a greater degree of finality to the decisions of specialized courts. See H. FRIENDLY, supra note 8, at 159 n.29. The problem with this approach is that it may exacerbate the risks of isolation and capture. It is, however, questionable whether the Supreme Court, which accords rather infrequent review to any individual appellate court, would provide enough of a generalist perspective in any event.

230. Unless, of course, litigants take the position that they need not acquiesce in the appellate court's decision. See generally Estreicher & Revesz, supra note 183.

231. See, e.g., S. EsTREICHER & J. SEXTON, supra note 191, at 124 (one way to lighten the burden on the Supreme Court would be to require some degree of deference to the first regional circuit to decide a legal issue). Cf. Christianson v. Colt Indus. Operating Corp., 870 F.2d 1292(7th Cir.), cert. denied, 110 S. Ct. 81 (1989) ("Although we recognize that the Federal Circuit's decision does not bind us, the comprehensive nature of the decision, along with the recognition that Congress created the Federal Circuit with the goal of achieving uniformity and coherence in the patent laws, counsel us against straying far from the court's thorough analysis of the difficult [patent] issues presented by this case." Id. at 1298-99 (citation omitted).).

in federal district court or in the Claims Court. Alternatively, the deficiency can be challenged directly in the Tax Court.

Claims Court cases are reviewed by the CAFC; the others, by the regional circuits. With minor variations, the system has been extremely successful for nearly a century.?*? Having a choice of courts diffuses suspicion that the process is stacked in favor of the government because taxpayers who are wary of specialized tribunals can usually use the district courts.22 At the same time, those who would like special expertise can exploit the proficiency of the Tax Court.

Conferring jurisdiction in the specialty area on both general and specialized courts is also advantageous to the judges. It creates an audience for the special court's opinions, which should motivate that tribunal to fully explicate its decisions. Consideration of similar issues by both generalized and specialized courts would prevent intellectual isolation and give each bench the benefits of the other's thinking.?** Although Supreme Court involvement is then necessary to eliminate conflicts between the courts, the delay permits percolation of the issues (albeit on a smaller scale than within the regional courts), and gives the ultimate decisionmakers an experiential foundation for their determinations.

As commentators have pointed out, however, this approach also has disadvantages. It requires maintenance of a new tribunal without a fully compensatory reduction in the generalist courts' workload. Furthermore, forum shopping between the special and generalist court creates new sources of conflict, or at least a need to resolve definitively which litigant's choice of forum should prevail.?*°2 32. See generally Jordan, supra note 169, at 752; Worthy, The Tax Litigation Structure, 5 Ga. L. Rev. 248 (1971);Dobson v. Commissioner, 320 U.S. 489, 498-501 (1943).

. 233. There is one exception: taxpayers who cannot afford to prepay the alleged tax deficiency cannot sue for a refund and thus are forced into the Tax Court. Y. 493 (1978) (noting that one major benefit of diversity jurisdiction is that it creates avenues of interchange between state courts-which can be regarded as "specialists" in state law-and federal courts). 235. Jordan, supra note 169, at 754, notes that this has not been a problem in the tax context because establishment of the Tax Court implicitly waived the Internal Revenue Commissioner's claim to adjudicate in the district court or the Claims Court. She speculates as to how well an optional system would work in private litigation, where each side is likely to prefer a different court. One solution to this problem is, of course, the one adopted in connection with the regional courts; the side that files first usually avoids d. Jurisdiction. The importance of a specialized court's jurisdictional grant cannot be overemphasized. Thus, as stated earlier in this article, the proper choice of subject areas can offset the problems associated with doctrinal isolation and diffuse the ability of interest groups to influence the appointments process inappropriately. Another aspect of the jurisdictional question centers more closely on the parties, for one of the significant costs of specialization from the litigants' point of view is that it sometimes requires them to try part of their cases in one tribunal while other aspects are litigated elsewhere.?** While the problem has never risen to the level of implicating the constitutionality of these courts,?*" specialization is likely to be more acceptable to the public if the need to resort to more than one trial forum can be avoided. Analogous complaints have been made about bifurcated appeals.?** For trial courts, certain restrictions on subject matter jurisdiction are inevitable. The judicial power of the United States is limited. Furthermore, the objectives of trial-level specialization-rapid resolutions and expert application of law to facts-would be compromised if the courts' work were largely determined by the way the parties chose to join their claims. litigating in the adversary's choice of forum, unless doctrines like forum nonconveniens come into play. Alternatively, simultaneous litigation is ended when one of the two courts in which an action is pending renders a judgment that supports dismissal of the other action on res judicata grounds. Note that in these cases, the happenstance of filing or deciding first is not seen as objectionable because the litigant is not thought to have an independent right to have a specific forum apply the law. Cf. Piper Aircraft Co. v. Reyno, 454 U.S. 235 (1981). But since one of the attractions of specialized adjudication is the bench's expertise, it may be hard to argue that litigants are not injured when their choice of forum is not honored. See also Miller, Whipsaw Problems in Tax Cases: Remarks at the Oct. 14, 1971Court of Claims Judicial Conference, 25 Tax Law. 193 (1972) '(noting other problems with overlapping authority). 236. See, e.g., Court of Claims History, supra note 15 (describing expansion of Court of Claims jurisdiction so claims for set offs and counterclaims could be asserted).

See also text accompanying note 128 (noting that the limitations on the jurisdiction of the Court of International Trade have been found problematic). As the experience of the Court of Private Land Claims, described supra note 147, makes clear, similar problems are engendered by limitations on the types of relief the special court can accord. See also H. Dusrorr, supra note 209, at 484 (noting problem of not being able to assert recoupment claims in the Tax Court); cf. Note, supra note 194, at 537 (noting the problems posed to Indian claimants by the limitation of relief obtainable in the Court of Claims to monetary damages). 937. See, e.g., Yakus v. United States, 321 U.S. 414 (1944) (upholding statutory scheme that bifurcated criminal defense between regional court and the Emergency Court of Appeals). rights inadvertently. Because the route of review would be clear no later than the time that the pretrial order was entered, trial courts would be able to apply any distinctive procedural rules required by the specialized tribunal.**? For courts that have been given exclusive authority in order to create uniformity, abrogating the well-pleaded complaint rule would insure that every specialty issue raised in the federal system is heard by the specialized court.2** Finally, hearing and deciding the additional issues that make up a case will help keep the intellectual perspective of the specialized court fresh and broad.

The nation's past experience in specialized adjudication suggests that the technique can make a significant contribution to judicial administration; but it is not without its problems. These are most evident when one thinks about designing the ideal specialized court. If its judges are to become proficient enough to impart expertise, if the court is to be collegial enough to create stable, predictable law, and if the docket is to be small enough so that cases are decided swiftly, the jurisdiction of the court must be narrow and defined with precision. But these constraints are troublesome. They give the judges of the court a crabbed view of the legal landscape which can lead to distorted decisionmaking. Thus, while it is difficult to prove (or disprove) that specialized judges do a better, fairer, or more proficient job, it is possible to demonstrate that they take the policies that their field of law furthers very seriously indeed. The lack of interchange between the specialized court and the remainder of the judiciary can only exacerbate the problem, for the absence of peer review means that the specialized bench is not called upon to fully defend its positions.

And, of course, controversies do not present themselves to the legal system in the manner in which they are considered in the classroom. A casebook editor waves a blue pencil and issues that are not in the curriculum disappear. These questions do not vanish with similar dispatch in the course of litigation. Either the specialized court must decide them, or the parties must bring part of their case in other fora. If the first route is fol-242. See, e.g., Meador, supra note 1, at 475. lowed, the specialized court may become busier than initially predicted, especially since parties that expect good results on nonspecialty issues may plead so as to invoke the special court's jurisdiction. If the second approach is taken, the parties are burdened with litigating in multiple fora, boundary law must be written, and the decisions reached in bifurcated cases may, asa whole, look very different from any single tribunal's resolution.

Problems can be minimized through the choice of specialty fields, and by implementing the strategy with care. Subjects singled out for special treatment should be segregable from the remainder of the federal docket, yet involve the court in adjudicating a broad range of issues. Thought should be given to the level in the adjudicatory hierarchy where expertise is most helpful so room can be left in the process for meaningful input from generalists. And because it is not possible to distinguish between changes in the law that occur because a court is expert and changes that occur because it has been captured, schemes must be designed with systemic guarantees of impartiality. For similar reasons, specialization should be used only when there is some public consensus on the policies underlying the law that the special tribunal administers.

But it must be recognized that none of the schemes for specialization are likely to make a meaningful difference to the workload crises, for the number of judge-hours required to adjudicate a federal docket of a given size can be expected to remain roughly the same. Specialty judges may decide cases within their field more quickly (although this has not been demonstrated), but the savings are at least partially offset by the need to create boundary law, to write rules to prevent forum shopping between the specialty and regional courts, to correct for procedural differences between geographic and subject-matter case allocation, and to account for the differing competence of the trial and reviewing courts which is created when only one of them is specialized.

Furthermore, if specialization is a success, if specialized courts adjudicate their caseloads more effectively and efficiently than regional courts and achieve public acceptance, changes would occur in the supply and demand for adjudicative services.

On the demand side, greater stability in the law could encourage primary activity in the field administered by the specialized court. While this stability might increase the likelihood of settlement and therefore reduce the number of cases reaching the court, rapid adjudication may attract the business of those who might otherwise have sought arbitration or mediation. On the supply side, a court that clears its docket easily would not be under the pressures currently suffered by the regional courts. Accordingly, specialty courts may take a more expansive view of substantive rights and gatekeeping rules than do overworked generalist courts.

In the end, it seems clear that there are some fields of law that would benefit from consolidation and expert adjudication. Decisions to establish new specialized tribunals should be animated by a desire to capture these benefits, rather than primarily by the wish to resolve the federal docket crisis. Such courts can and should be organized in a way that avoids the pitfalls.