The Commission’s proposal for regionally based adjudicative divisions reflects a conscientious attempt to respond to criticisms of the Ninth Circuit Court of Appeals "while preserving [an] administrative structure that no one has seriously challenged." Unfortunately, the plan is flawed both in conception and in execution. It is unlikely to accomplish its goals, and it has the capacity to produce much mischief. I urge the Commission to reconsider its approach.

The difficulties begin with the Commission's abandonment of circuit-wide stare decisis. The Commission agrees that it is important "to maintain a capacity, short of review by the Supreme Court, for a common body of federal law on issues that matter to the whole Pacific rim and the far western part of the United States." But this policy would be frustrated by a regime under which "[d]ecisions made in one division would not bind any other division." To be sure, the report contemplates that decisions of other divisions would "be accorded substantial weight as the judges endeavor to keep circuit law consistent." And I am confident that in most instances the judges will do that, just as circuit judges today generally respect the decisions of other circuits. But there is a difference between respecting precedent and being obliged to follow it, and that difference will come into play when it matters most.

Admittedly, panels today sometimes distinguish precedents on grounds that are not persuasive to other judges or lawyers. But my research indicates that the phenomenon is not a common one. Further, I have no doubt that judges often follow precedents they do not like, simply because it is their obligation to do so. If stare decisis did not operate circuit-wide, judges would find it too easy simply to reject precedent from another division. The Commission plan would thus authorize, if not encourage, the creation of intracircuit conflicts.

The Commission's proposal for a "Circuit Division" to resolve conflicts between divisions is not a satisfactory solution. Contrary to the Commission's assertion, the proposed change would indeed "create a new tier in the judiciary." This is made clear by the Commission's proposed statute. Under section 3(a), the Circuit Division would have jurisdiction to review "final decisions rendered in any of the court's divisions." This is not "merely substitut[ing] the Circuit Division for the existing en banc process." The en banc court does not review the decision of the panel; it reviews the decision of the district court.

Even more problematic is the Commission’s delineation of the Circuit Division's jurisdiction. The Circuit Division would be limited to resolving "square inter-divisional conflicts." Does this mean that the Circuit Division could review a case only if a panel in one division explicitly refused to follow another division's ruling? If so, the Circuit Division would lack authority to resolve tensions in circuit law caused by less blatant inconsistencies in doctrines or outcomes. See, e.g., Hale v. Arizona, 993 F.2d 1387, 1389 (9th Cir. 1993) (en banc) ("We consider these questions en banc to resolve the tension between [two panel decisions]."); Townsend v. Holman Consulting Corp., 929 F.2d 1358, 1365 (9th Cir. 1990) (en banc) (overruling panel decision, thus obviating need to maintain "unstable and awkward" distinction drawn by later case). But if something less than explicit rejection would suffice, that is an invitation to tiresome wrangling over whether two decisions really are in conflict.

In this regard, it is instructive to consider the experience of the Florida Supreme Court. That court is vested with jurisdiction to review "any decision of a district court of appeal … that expressly and directly conflicts with a decision of another district court of appeal … on the same question of law." Commentators describe the jurisdiction as "disputatious" and note that "the existence of conflict often is not so certain, meaning that a brief [seeking review] must engage in a lengthier and more convoluted argument to establish the Court’s discretion to hear the case." See Gerald Kogan & Robert Craig Waters, The Operation and Jurisdiction of the Florida Supreme Court, 18 Nova L. Rev. 1151, 1225, 1238 (1994). That is hardly a model to be emulated.

The Commission’s formulation also raises questions about the scope of the Circuit Division’s authority over the cases that it hears. Would the jurisdiction of the Circuit Division be limited to the issue that generated the conflict? If so, the panel would often be handicapped by having to consider that issue in isolation from the other issues in the case. But if the panel could consider all issues, the effect would be to give seven judges broad power to control circuit law in matters that were not the subject of an intracircuit conflict.

In this light, the Commission’s proposed method for determining the composition of the Circuit Division is particularly troubling. Four of the seven members – the chief judge of the circuit and the chief judges of the three divisions – would be the senior active judges in their respective groups. This means that a majority of the Circuit Division will often be drawn from a small and unrepresentative cohort of the court’s membership. That in turn raises the stakes for every dispute over selecting the other members of the Division or defining the scope of its authority.

A third weakness of the plan is the provision for en banc hearings within each of the three divisions. This will delay the resolution of cases, especially in light of the requirement that litigants seek en banc rehearing by the regional division before asking for review by the Circuit Division. In addition, by giving decisions the imprimatur of all of the judges of the division, the procedure may solidify the position of the division on a particular issue, thus making it more difficult to reestablish consistency throughout the circuit.

This analysis suggests several modifications in the Commission’s plan. At the very least, circuit-wide stare decisis should be retained; divisional en bancs should be eliminated; and the "Circuit Division" should be replaced with some form of circuit-wide en banc review utilizing the criteria of FRAP 35. These modifications, of course, would cut back substantially on the autonomy of the regional divisions. That would be cause for concern if there were a serious problem in the court of appeals that could be solved only through reliance on smaller adjudicative units. But on the evidence of the Commission report, no such problem exists.

The Commission’s principal argument is that "large appellate units have difficulty developing and maintaining consistent and coherent law." To support this crucial proposition, the Commission adduces only "perceptions" (unspecified), its "own experience" (unelaborated), and an assertion that the judges of a large court cannot "read all the court’s published opinions when they are issued." On the last point, I agree with Chief Judge Hug and his colleagues: the assumption that judges cannot keep sufficiently abreast of circuit law without reviewing opinions as they come out "is a relic of the pre-computer era." As for the rest, the only systematic studies of the operation of precedent in the Ninth Circuit – my own – contradict the Commission’s premise. The Commission is of course entitled to reject the methods or conclusions of those studies, but vague perceptions are no substitute for empirical data.

The Commission’s criticisms of the limited en banc in the Ninth Circuit are equally unsupported. By the Commission’s own account, no more than a minority of the Ninth Circuit’s judges are dissatisfied with the court’s use of the en banc process to perform the various en banc functions. Beyond this, it is difficult to evaluate the Commission’s criticisms, because the report lumps together functions with very different implications for structural reform, notably the resolution of conflicts and the correction of "wrong" panel decisions.

In sum, the Commission has not adequately justified its recommendation for legislation that would mandate semi-autonomous adjudicative divisions in the Ninth Circuit and authorize them in other circuits. I hope that the Commission will reconsider its approach and will, at a minimum, make the changes suggested above. I also urge the Commission to cast its recommendation as an experimental program rather than a mandatory restructuring. Even if the Commission is correct in its assessment of large appellate courts, the circuits should be given wide leeway to find solutions. It is unlikely that any single approach will work equally well in all courts or that the preferable arrangement will emerge without some false starts.