Circuit Judge Sidney R. Thomas

U.S. Court of Appeals for the Ninth Circuit

October 26, 2005

Senate Judiciary Committee
Subcommittee on Administrative Oversight and the Courts
Hearing on Revisiting Proposals to Split the Ninth Circuit
Testimony of Judge Sidney R. Thomas
October 26, 2005

Mr. Chairman and Committee members. My name is Sidney R. Thomas. I serve as a Circuit Judge on the Ninth Circuit Court of Appeals, with chambers in Billings, Montana. I presently serve as En Banc Coordinator and Death Penalty Coordinator for the Circuit. I also serve on the Executive Committee of the Circuit. I thank the Subcommittee for the opportunity to testify on the various proposals to divide or restructure the Ninth Circuit Court of Appeals. The views I express are my own.

I oppose restructuring the present Ninth Circuit to create two or more circuit courts. Division of the Ninth Circuit at this time would have a devastating effect on our Court. It would increase delay in case processing substantially and would decrease access to justice. It would create unnecessary and expensive duplication of core functions, while substantially reducing vital services. The most effective means of administering justice in the federal courts in the states comprising the Ninth Circuit is a centralized administration and staff support.

The impact of a division of the circuit would be especially harmful at the present time, when the United States Courts are faced with static or decreasing budgetary resources and when the Ninth Circuit has been forced to face an unprecedented, but probably temporary growth in its appellate caseload.

To explain my reasoning fully, I would like first to address the real world administrative impact of any split, then address some of the underlying concerns expressed by those promoting a structural division of the circuit.

Budgetary and Administrative Impact

1. Caseload trends.

The Ninth Circuit's appellate caseload from the district courts of the Circuit has stabilized during the past few years. In 2001, we faced a backlog of cases that developed from 1994-1998, during a period when the Court was operating with only eighteen of its twenty-eight judgeships filled. To address this, we adopted an aggressive case management plan. The plan was successful. At the end of 2001, the Administrative Office reported a median time from Notice of Appeal to disposition in the Ninth Circuit of 16.1 months. At the end of 2003, the median time was 13.7 months, a 14% decrease in two years. Our internal statistics showed an approximate 50% decrease in the time between the filing of the last brief and oral argument hearing during the same period. This statistic is important to us because it provides a good measure of how fast attorneys are able to get their case heard.

The Circuit would be current in its workload, except for an unusual and unanticipated circumstance: the unprecedented growth in immigration administrative petitions for review during the last several years.

The following graph shows the progress the Ninth Circuit made from 2001-2004, despite the increase in immigration cases:

Changes in Filing and Delay (2001-2004)

Circuit % Caseload % Delay
Change Change

11th Circuit - 6.2% - 16.2%
9th Circuit + 38.0% - 11.4%
5th Circuit - 1.5% - 10.5%
8th Circuit + 2.2% - 8.4%
10th Circuit - 4.1% 0.0%
3rd Circuit + .3% + .9%
4th Circuit - 6.5% + 4.2%
7th Circuit - 2.3% + 6.2%
1st Circuit - 2.2% + 6.7%
6th Circuit - .2% + 9.8%
2nd Circuit + 55.1% + 29.9%
D.C. Circuit - .8% + 38.2%

Given the increased immigration caseload, we would expect the delay figures to increase over previous years. However, the point is that application of Ninth Circuit case management techniques has been quite successful despite the increase in volume.

The present caseload challenge is the enormous increase in immigration petitions for review. The increase in immigration caseload stemmed from a decision of the Attorney General to eliminate the backlog of 56,000 cases that existed in the Board of Immigration Appeals. That decision resulted in the resolution of tens of thousands of cases by the BIA in a matter of months. Over half of the petitions for judicial review from those cases were venued in the Ninth Circuit.

The statistics available through June 30, 2005, indicate a net decrease of 1% of the total appellate caseload from the district courts and non-immigration administrative petitions for review during the period 2001-2005. In contrast, petitions for review from administrative immigration decisions increased 570% during the same period.

The following numbers illustrate the point:

Immigration Non-Immigration
Appeals Appeals
Year

2001 954 9,388
2002 2,670 8,751
2003 4,206 8,666
2004 5,368 8,906
2005 (thru 6/30/05) 6,390 9,327

Whether the increase in caseload is temporary or not remains open to question. The BIA has reported that it is has reduced its backlog to 29,000 cases, indicating that, while the courts can expect continued volume for the next several years, the volume of immigration cases should decrease as the BIA becomes current in its case processing. Recent statistics bear this out. The rate of filing of petitions for review of the decisions of the Board of Immigration Appeals has been declining. The Second Circuit, the other circuit most affected by an increase in immigration caseload (experiencing a 1312% increase in its immigration caseload from 2001-2005), has noticed a 9% decrease in petitions for review this year. The Real ID case may also have an effect in decreasing immigration caseload. The fact that immigration caseload is likely to decrease seems probable. The question is the level to which immigration caseload will ultimately stabilize.

Temporary spikes in caseload are nothing new. From time to time, our Circuit and others have experienced temporary increases due to particular circumstances. A recent example is the caseload spike experienced by the border states. The California energy crisis causes the filing of a large number of petitions for review from decisions of the Federal Energy Regulatory Commission. Temporary increases in caseload due to unique circumstances are to be expected. However, the clear trend over the past five years in the aggregate is that the non-immigration caseload of the Ninth Circuit has stabilized.

The key question is how best to administer this specialized caseload. A careful examination of the cases indicates that the best method is through intensive staff review, prior to judicial involvement. The reason is that immigration relief is procedurally complex. Many petitioners fail to comply with procedural requirements; and many others file petitions over which the court of appeals lacks jurisdiction. Our current statistics indicate that well over 80% of the immigration petitions for review are resolved through centralized staff review. Less than 20% of the cases are ultimately presented to judges during the normal oral argument calendars. This statistic underscores the critical function of court staff in handling these cases.

Division of the circuit during this time would seriously disrupt the Court's ability to resolve this temporary spike in administrative immigration cases. Because of unnecessarily duplicated functions and reduced budgets, the net available staff would be significantly decreased. The marginal increase in judges would not, even in the most remote sense, compensate for this loss.

Perhaps more importantly, the disruption to judges and court staff of managing a transition to a divided circuit would slow case processing or bring it to a temporary halt. Delays in adjudicating cases would increase substantially, and it would be quite difficult to eliminate the increased backlog caused by the transition in addition to the existing backlog.

In sum, the best method of addressing the spike in immigration filings is to maintain the centralized administrative structure of the Ninth Circuit. The rate of appeal in every other category has stabilized or is declining. When the present bulge in immigration cases has been resolved, the Ninth Circuit should be current in its case processing.

2. Present budgetary posture.

At the present time, as the Subcommittee members are undoubtedly aware, the federal courts are facing a budgetary crisis. In 2004, Ralph Mecham, Executive Director of the Administrative Office of the United States Courts wrote a memorandum to all federal judges observing that: "The entire judicial branch of government faces the most serious funding challenge that I have seen during my 19 years as Director of the Administrative Office." Last year, all federal courts - including the courts of the Ninth Circuit - prepared contingency plans involving significant personnel layoffs and other cost-saving measures. Fortunately, most of those measures did not have to be implemented. However, we face a similar potential this year. Given recent budgetary history, it would be unrealistic for the Courts to plan for substantial budgetary increases, especially given the other important budgetary demands given the tragic loss of life and property due to Hurricane Katrina, and the budget demands due to the military operations in Iraq.

In short, unless there is some unforseen change in the near term, the Courts must plan to live within their present budgetary means, and to administer justice in the most efficient manner possible within those means.

Merely increasing the judiciary budget to add operating revenue will not solve the problem, for two reasons. First, as the Subcommittee is also undoubtedly aware, the Judiciary budget is prepared and allocated based on formulas that are, in great measure, caseload driven. Thus, circuit division will not necessarily mean greater funding for the federal courts in the reconfigured Ninth Circuit; it will essentially take existing funding and divide it. Any additional funding will be allocated to all circuits based on the formula. Therefore, it would take a substantial multiple of any dollars added to the judiciary budget to produce an equal amount to the bottom line of any circuit's budget. The alternative would be to take money from other circuits. This remedy might be required on the basis of the revised formulas for new circuits, but it would have an unfair and disastrous effect on other circuits that are currently experiencing severe budget crises of their own.

Second, new circuits created would have to replicate essential core functions. There will be multiple Clerks of Court and Circuit Executives, along with other top administrative staff positions. Even if funding were equalized, unnecessary duplication would cause a net budgetary effect of reduced available staff and other resources. Circuit division would reorganize the current staff resources into a more administratively top-heavy organization, less able to deliver needed services.

3. Administrative impacts of Circuit division.

Division of the Ninth Circuit would be costly, disruptive and would create enormous inefficiencies. The present structure of is designed to efficiently resolve questions that need not be decided by judges, and to present questions that require judicial resolution in the most effective manner. Division would deprive the remaining circuit courts of these resources, resulting in judges wasting time on matters that could be resolved without spending valuable judge time.

These administrative efficiencies are unique to the Ninth Circuit and only available because we have been able to aggregate our resources. To take a few examples:

? Appellate Commissioner. The position of Appellate Commissioner is unique to the Ninth Circuit. Last year, the Appellate Commissioner resolved 1,125 Criminal Justice Act fee vouchers that otherwise would have been handled by judges. He resolved 4, 062 substantive motions previously heard by judges. This position likely be eliminated in any division and most certainly would not be available to smaller units.

? Circuit Mediator. The Circuit Mediator's office settled 881 appeals last year out of 977, a 90% success rate. In contrast, the Sixth Circuit had a success rate of 42% and resolved far fewer cases on a comparative basis. The Mediator's office has also been successful in resolving highly complex cases, in which settlement depends on the participation of non-parties, such as CERCLA cases. The Mediator's office has also had success in organizing complex litigation, such as the recent high number of petitions for review filed in connection with the California energy crisis. The Mediator's office would be significantly reduced with a circuit division. Most small circuits have only one mediator and settle relatively few cases. The settlement success rates are also lower. A mediator's office needs critical mass to achieve success.

? Staff Attorneys. The staff attorneys were critical in the termination of a large volume of appeals - well over half the appeals filed in the Circuit. The staff attorneys presented 1,421 habeas petitioners' requests for a Certificate of Appealability. Panels denied 89% of the requests, terminating 1,265 appeals at that stage. In addition, staff presented 2,182 merits cases to screening panels, resulting in termination of another 2,029 appeals. Put in perspective, in its entirety, the First Circuit terminated 1,643 cases in 2004. The D.C. Circuit terminated a total of 1,155 cases. In addition, staff motions attorneys disposed of 10,948 motions through clerk orders. The staff attorneys office would be considerably reduced in a smaller circuit.

? Bankruptcy Appellate Panel. The BAP resolved almost 700 appeals last year. It would not exist after a circuit division. Those cases would fall back on the district courts for resolution. The loss of the BAP would come at a particularly disadvantageous time, as the courts struggle to interpret the extensive amendments to the Bankruptcy Code recently passed by Congress.

? Case tracking and batching. Because it has the resources to do it, the Ninth Circuit inventories each filed appeal for issues. The Circuit then tracks the case and the issue. Cases involving similar questions are grouped together for oral argument to promote consistent treatment. Cases are also stayed pending resolution of dispositive issues in published opinions. It is not uncommon for a published decision to result in the immediate resolution of hundreds of cases that were dependent on its outcome. This inventory and tracking system is unique to the Ninth Circuit and would not survive a circuit division given the significantly reduced staff resources.

These administrative efficiencies are especially important given the case mix of the Ninth Circuit. Over 40% of total appeals in the Ninth Circuit are filed by pro se litigants. Last year, for example, there were 5,802 pro se appeals filed in the Ninth Circuit out of 14,274 total cases. These appeals are processed by a special Pro Se Unit in the Ninth Circuit staff attorneys' office. The vast majority of these appeals are then resolved by presentation to screening panels made up of Article III judges. Very few of these cases are referred to judges' chambers for consideration by oral argument panels. The significance of this given the current case mix is multiplied with we consider that approximately half of the pro se volume consists of immigration cases.

The vast majority of immigration cases, which account almost all of the increased volume of the circuit in the last several years, are resolved through the staff process. In fact, our current statistics show that 80% of the immigration petitions for review are resolved through the staff screening process rather than on oral argument calendars.

To put this into total perspective, in an average year, approximately 50% of the filed cases are procedurally terminated through staff efforts before they reach a merits panel; of the remaining merits terminations, one-third were resolved by judicial screening panels deciding the cases based on staff presentations. Taking this all together, the Circuit staff provided the primary assistance in the resolution of approximately 80% of appeals; the remaining 20% were resolved by judges and their chambers staff on oral argument calendars.

A division of the circuit will mean far fewer staff resources available to handle the non-oral argument calendar appeals, which account for 80% of the volume of circuit work. Splitting the circuit will not create budget increases; rather it will likely take existing resources and divide them. Moreover, core functions will be replicated, and additional management positions required. Thus, there will be far less staff available for case processing.

The current case mix in the Ninth Circuit is best addressed by retaining a strong, coordinated, central staff that can perform essential case triage and resolve the vast majority of appeals. Circuit division would reduce or eliminate many of these critical personnel resources available. The inevitable result will be inefficiency, waste of judicial time, loss of services, and increased delay.

4. Loss of other cost-saving devices.

Aside from those issues that are unique to the Court of Appeals, there are other, significant cost savings that would be lost if the Ninth Circuit would be divided. For example, one of the most expensive aspects of the judiciary budget is the payment for defense of capital cases. We have been cognizant of this problem and have created a committee to review budgets for the prosecution of such cases. Chief Judge Stephen M. McNamee of the District of Arizona and Judge Barry T. Moskowitz of the Southern District of California have done remarkable work in analyzing capital case budgets. Their work has saved hundreds of thousands, if not millions, of dollars. These efforts would be significantly lost or reduced under a new division. There simply would not be enough of a critical mass of judges to serve these functions in a small circuit.

Likewise, the smaller circuits would have significantly fewer resources in space and facility planning, a division in the Ninth Circuit Executive's office which has also saved taxpayers significant sums of money and assisted in the construction of courthouses that are more efficient and less costly. An excellent example of effective planning is the new district courthouse in Seattle, which utilizes courtrooms space in an innovative and efficient manner. The planners of the Circuit Executive's office have been invaluable to smaller states like Montana and Idaho, to assist in courthouse planning given those states' very unique needs.

In short, there are enormous costs - both direct and indirect - that would be created by circuit division. Administrative duplication and waste would be substantial. Circuit division would result in a significant decrease in the services that the Circuit now provides. This effect would be compounded by our current budget situation.

5. Judicial efficiency.

The current structure of the Ninth Circuit also provides judicial economy. Administrative tasks are shared among the judges. Creation of one or more new circuits would force judges in all of the reconfigured circuits to assume greater administrative loads.

In addition, resolution of issues in a circuit means that judges need not revisit the issues. Reconfiguring the Ninth Circuit into two or more circuits would mean that the same issue would have to be analyzed and decided in both circuits, causing a net loss of judicial efficiency.

Further, the large number of judges in the Ninth Circuit means that it is better able to handle the problems causes by persistent judicial vacancies or by judicial disability. In a circuit with a small number of circuit judges, any problems encountered by an individual judge would have far more ramifications than in a larger circuit. If a judge on a small became temporarily or permanently disabled, it would have a much greater impact than a judge experiencing problems on a larger court. Likewise, if problems developed in the confirmation of a judge who was to serve on a smaller circuit, then it would have a significant impact on the functioning of that circuit.

The Flawed Arguments for Division of the Circuit

Despite the advantages of the present structure and the significant disadvantages of imposing a circuit split at the time, given the growth of immigration cases and the budge crisis, some critics have persisted in their view that the Circuit should be divided. When the arguments are examined closely, they are not persuasive. Indeed, most of the arguments are based on faulty factual premises.

1. Caseload Growth in the Ninth Circuit

One of the major arguments justifying structural division of the Ninth Circuit is that population growth throughout the region will cause increased appellate caseloads, and that division is the only means of accommodating the uniform increase in appellate filings. This argument is based on a faulty premise. In fact, there is no correlation between population growth and federal appellate filings. Rather, increases in appellate work have been primarily based on discrete, specific circumstances that tend to be transitory. As I have indicated the non-immigration caseloads have actually decreased over time. These are the cases that emanate from the geographic regions of the Circuit, rather than external sources. If there were a correlation between population grown and caseload growth one would expect to see a growth in caseload that corresponded with population growth. However, the Circuit has not experienced such growth.

When one examples individual geographic circumstances, the point is amplified. For example, Alaska's population grew 8.5% between 1991 and 2002. However, the number of appeals filed in the Ninth Circuit from Alaska actually decreased during the same period by 88.7%. Similarly, Oregon's population increased 17% between 1991 and 2002; its federal appellate caseload decreased during the same period by 13%. Indeed, if one examines the states comprising the Northern division of the Ninth Circuit (Alaska, Washington, Oregon, Idaho, and Montana), the appellate caseload has been virtually flat for over a decade. From 1993 to 2002, while the aggregate population grew 17%, the total appellate caseload from the region decreased by 3.2%.

Creating new circuits as proposed cannot be justified based on purported growth of cases within the areas covered by the new circuits. The simple fact is that federal appellate caseload is not related to population growth. Rather, it is more influenced by other factors that tend to be transitory. For example, the federal courts in the states bordering Mexico have experienced enormous caseload growth in recent years. However, last year appeals from the Southern District of California - one of the judicial districts most affected by the problem - decreased from the previous year. The current appellate caseload challenge in the Ninth Circuit is not based on geography or population, but rather the actions of a single administrative agency.

The fact that judicial caseload emergencies tend to be transitory and driven by unique problems is also demonstrated by examining caseload in the various judicial districts. In my home state of Montana, for example, we recently experienced a judicial emergency because only one of Montana's three judgeships had been filled. To avoid dismissal of criminal cases for lack of a speedy trial, district judges were flown in from throughout the Ninth Circuit to try cases. Eventually, two more judges were confirmed and the crisis abated. This is a familiar story in our Circuit, and the judges of our Circuit have demonstrated a remarkable willingness to assist their colleagues during these critical times. That is a luxury of a larger Circuit - to be able to have the flexibility to reallocate judicial resources during times of need.

In short, if one examines the data carefully, one can quickly discern that there is no independent justification for creating new federal circuit courts in the Western United States based on population projections or the intuitive notion that caseloads are uniformly increasing throughout the region. Rather, the data indicates that caseload spikes have been driven by unique circumstances that tend to be short-lived. To address these problems, the best solution is a larger Circuit that has the flexibility to reallocate resources internally, rather than to erect structural barriers to the allocation of judicial resources.

2. Delay.

The second major faulty premise upon which the proponents of a circuit division rest their case is delay in case processing. Proponents of a split assume, without explaining, that any division of the Ninth Circuit will improve case processing time. The opposite is true. Circuit division will increase, not decrease delay.

First, as I have discussed, by use of case management techniques over the past several years, we have substantially reduced delay. The major present problem, as I have discussed, is the increase in administrative immigration petitions for judicial review. It is not only the sheer number of cases that increase the delay statistic, it is the inability of the government to file the appeal record in a timely fashion. In thousands of cases, the government has requested open-ended extensions of time - for a year or more - so that it can prepare the administrative record. Although there is virtually nothing that the Ninth Circuit can do about this, short of granting the alien summary relief - that time is charged to the Circuit in the form of an increase in the delay statistics. However, it distorts our comparative case processing time statistics. The increase in case processing time due to the increased immigration caseload cannot be regarding as reflecting on the effectiveness of judicial administration.

Case processing delay is not related to caseload, or size of circuit. The Commission on Structural Alternatives for the Federal Courts of Appeal, more popularly known as the "White Commission," studied the subject of delay thoroughly in 1998 and concluded that circuit size was not a critical factor in appellate delay. Specifically, the White Commission wrote:

We have reviewed all of the available objective data
routinely used in court administration to measure the
performance and efficiency of the federal appellate
courts, but we cannot say that the statistical criteria
tip decisively in one direction or the other. While
there are differences among the courts of appeals,
differences in judicial vacancy rates, caseload mix,
and operating procedures make it impossible to
attribute them to any single factor such as size.

Commission on Structural Alternatives for the Federal Courts of Appeal, Final Report, p. 39 (1998).

In addition to a lack of relationship between circuit size and delay, there is no statistical relationship between total caseload and disposition time.

Perhaps the real question is what the goal is in terms of case processing time, what structure best achieves it, and at what cost. For example, the difference between median case processing time for all circuits and the Ninth last year was 3.3 months. We were not the slowest circuit. If, by using management techniques, the Ninth Circuit could reduce total case processing time by 2.4 months in just two years, is there a compelling reason to cause serious disruption in the federal courts with the hope of reducing total case processing time by a few more months? Or is it better to continue to improve effectiveness and efficiency within the current structure?

If one examines the administrative structure of the Ninth Circuit and its efficiencies, I believe the only conclusion that can be drawn is that a circuit split will increase delay, rather than decrease it. A division of the circuit will cause the loss of a large number of administrative tools to reduce appellate caseload, and will place more cases and administrative tasks on judges.

Circuit division does not eliminate caseload; it merely reallocates it. The cases still need to be decided. There is no evidence that demonstrates that the present caseload could be more effectively or efficiently managed by dividing the Ninth Circuit. In terms of efficient case processing, the best model at the present time is a strong, central administrative staff to examine cases for procedural and jurisdictional defects before the cases are referred to oral argument panels. If the ability to handle 80% of the Ninth Circuit's cases is impaired, and if circuit judges are forced to spend much more time with administrative matters, then the inevitable result will be increased delay to the litigants.

The best solution to resolving case processing delay is within the existing institution. Circuit division will not eliminate delay; it will create unnecessary delay.

3. En Banc Procedure

The Ninth Circuit's limited en banc procedure has been cited as a rationale for circuit division. However, a close examination will dispel the notion that circuit division is justified in order to guarantee a full court en banc hearing.

First, this involves an extraordinarily small number of cases. Out of 5,783 cases decided in the Ninth Circuit between September 2003 and September 2004, only 13 ( or .2% ) were reheard en banc. This experience is consistent with the practices of other circuits. Of 27,438 cases decided nationally within the same period, only a total of 59 (or .002%) were heard en banc. The following chart for the time period:

En Banc Hearings: All Circuits (9/2003-9/2004)

District of Columbia 2
First Circuit 1
Second Circuit 0
Third Circuit 2
Fourth Circuit 7
Fifth Circuit 13
Sixth Circuit 6
Seventh Circuit 2
Eight Circuit 5
Ninth Circuit 13
Tenth Circuit 4
Eleventh Circuit 4

Second, very few of the decisions made by the en banc panels involved close votes. Since 1996, almost 70% of the en banc cases were decided by margins of 8-3 or more. Forty-two percent of the cases were decided unanimously. Only 15% of the decisions - 20 in the last 9 years - involved a one vote margin.

Third, the worry that a minority of the Court could determine the outcome of an en banc case has been ameliorated by the Court's recent decision to increase the size of the limited en banc court to 15. `In addition, that argument neglects two significant facts: (1) well over 99% of the cases decided by the Ninth Circuit - and all the circuit courts for that matter - are decided by three judge panels, in which the votes of two judges bind the entire Circuit and (2) the Ninth Circuit allows for a full court en banc rehearing. As yet, there has not been an occasion in which a majority of the eligible judges has voted to rehear a case before the entire court.

Fourth, although fifteen judges are ultimately drawn to serve on a Ninth Circuit en banc court, the determination whether to take a case en banc remains with the full court. By statute (28 U.S.C. § 46(c)), a vote in favor of en banc rehearing by a majority of non-recused active judges is required to take a case en banc. Moreover, any active or senior judge may call for en banc rehearing, and all may participate in the exchange of views - often extensive - that precedes the vote.

Fifth, the Court has taken concerns about the representative nature of the limited en banc panel seriously and studied the question. Prompted by issues raised during the White Commission hearings, the Ninth Circuit formed an Evaluation Committee to examine some of the issues raised more closely, including the limited en banc procedure. To answer the questions relating to en banc procedures, the Evaluation Committee consulted with a number of outside academic experts. One of the experts consulted was Professor D.H. Kaye of the College of Law, Arizona State University, a noted expert in the field of law and statistics, who conducted a statistical analysis of the size of the limited en banc court in relation to a full court of 28 judges. Professor Kaye calculated the probability that the outcome of the limited en banc court vote would be the same as that of a full court of 28. He posited a binary issue (judges would vote either to affirm or to reverse), and he considered the possible divisions among 28 judges. He found that expanding the en banc court would result in only a trivial gain in the degree by which an en banc court decision would represent the views of all judges of the court. The largest gain would occur when there were 28 active judges who divided 17 to 11 in their views as to whether the panel opinion was correct. Yet even in that situation, if the limited en banc court were expanded to 13, the gain in accuracy of "representativeness" would be only 3.5 cases per hundred, and only 7 cases per hundred if the limited en banc court were expanded to 15.

The Evaluation Committee also met with a number of other scholars to discuss this issue, including Professor Linda Cohen, Department of Economics, University of California, Irvine; Professor John Ferejohn, Hoover Institute, Stanford University; Professor Louis Kornhauser, New York University School of Law; Professor Matt McCubbins, Department of Political Science, University of California, San Diego; and Professor Roger Noll, Department of Economics, Stanford University, CA. These scholars consulted by the Committee confirmed the import of the calculations done by Professor Kaye in concluding that the current random draw is effective in providing a representative en banc court of 11 judges.

To supplement the analysis by Professor Kaye and the other consultants, the Evaluation Committee requested Professor Arthur Hellman of the University of Pittsburgh School of Law to conduct an empirical study of actual en banc outcomes. His conclusion was that the evidence strongly indicates that in a substantial majority of en banc cases the limited en banc court has reached the same result that a majority of active judges would have reached. He also concluded that in the cases in doubt, expanding the limited en banc court would have added to the judges' burdens without enhancing the "representativeness" of the outcome. He observed:

It is true that enlarging the size of the en banc court would make it more "representative" in an abstract sense. But the more important question is whether it would produce decisions, with majority, concurring and dissenting opinions, that better represent the views of the court's active judges. Probability analysis and empirical data both indicate that the gains would at best be marginal.

Sixth, none of the bills would totally eliminate the limited en banc court. Under any scenario, the circuit containing California would eventually have too many judges for a permanent full court en banc panel. So, to the extent that the procedure is view as problematic, none of the pending legislation addresses it fully.

Seventh, when all factors are considered, the limited en banc court is a valuable tool. Rehearing a case en banc uses up significant circuit resources. It is a time and energy consuming process. The limited en banc system employed by the Ninth Circuit should be analyzed as to its legitimacy, representativeness, and deliberative quality. The limited en banc panel has rarely, if ever, reversed the decision of a prior en banc panel. Indeed, it is rarely requested to do so. There is no compelling evidence that the decisions of the limited en banc panel are not accepted as the binding decisions of the Court. Our internal studies, and all external studies, have concluded that the composition of the panel is sufficiently representative. Having too many judges can interfere with the deliberative process; limiting the panel number to eleven strikes an appropriate balance between the number required for legitimacy and representativeness and the number required for effective deliberations. It also strikes the proper balance of resources needed to resolve en banc-worthy issues.

Finally, and perhaps most importantly, the question of size of the en banc panel is a matter within the administrative control of the Ninth Circuit. No legislation is required to either increase the size of the panel, or to mandate a full court en banc panel. That can be accomplished by vote of the judges of the circuit.

For all of these reasons, the limited en banc system employed by the Ninth Circuit does not justify a circuit division.

4. Case conflict.

All academic studies of the Ninth Circuit have concluded that conflict in panel decisions is not a significant problem. The Ninth Circuit Evaluation Committee studied this in detail and concluded that there was no credible evidence that the Ninth Circuit experienced conflict problems in a greater proportion thatn that of other circuits.

We have employed a number of techniques to avoid case conflicts.

First, as I have previously discussed, the Ninth Circuit uses a case tracking system that identifies issues involved in each appeal. An inventory sheet is prepared for each case prior to its transmittal to a panel listing all potential cases that might have a bearing on the case.

Second, prior to the issuance of the opinion, each judge receives a pre-publication report that describes the holding and also identifies each case that the tracking system indicates may be affected by the opinion.

Third, we have an extensive en banc process in which off-panel judges raise questions about published opinions. This process often results in the modification of the opinions without the necessity of rehearing en banc. The parties also participate in the process by filing petitions for rehearing en banc, which are reviewed by each chambers.

5. Reversal rate.

Supreme Court review affects only a handful of cases. For example, in 2003, the Ninth Circuit had 12,151 appeals filed. In the same period, there were 1,462 petitions for a writ of certiorari filed seeking Supreme Court review of Ninth Circuit decisions. The Supreme Court granted 25 of those petitions, or 1.7% of total petitions sought. The Court reversed the Ninth Circuit in 19 cases. Supreme Court reversals affect a minuscule number of cases, and cannot serve as a meaningful point of evaluation of judicial administration. Thus, the Supreme Court reversal is not particularly instructive concerning structural division of a circuit court.

Further, in recent years, the reversal rate of the Ninth Circuit has not deviated much from the rest of the Circuits. In the 2003-2004 term, the national reversal rate was 77%; the reversal rate for the Ninth Circuit was 76%. In the 2002-03 term, the national reversal rate was 73%; the Ninth Circuit's was 75%.

6. Number of Opinions.

A small minority of judges on the Circuit have complained that the Circuit produces too many opinions, and that the judges of the Court cannot keep up with the state of the law. At the onset, I emphasize that the majority of the members of the Court do not share this view and are able to keep up with Circuit law. More importantly, the Ninth Circuit is not the largest producers of opinions. The latest statistics show that the Eighth Circuit produced more opinions than the Ninth Circuit last year. Other Circuits are quite close in production of opinions. If division of a circuit is justified on this basis, other circuits will have to be divided.

The following chart shows that the Ninth Circuit does not produce an inordinate amount of circuit opinions relative to other circuits and that the number of opinions produced is not a function of court size:

Number of Published Opinions/Circuit: 2004

Circuit Number of Authorized % Opinion
Opinions Judgeships per Auth. Jdshp

Eighth Circuit 701 11 63.7%
Ninth Circuit 691 28 24.7%
Seventh Circuit 608 11 55.3%
Sixth Circuit 453 16 28.3%
Second Circuit 425 13 32.7%
Fifth Circuit 412 17 24.2%
Eleventh Circuit 378 12 31.5%
First Circuit 372 6 62.0%
Tenth Circuit 326 12 27.2%
Third Circuit 320 14 22.9%
D.C. Circuit 240 12 20.0%
Fourth Circuit 222 15 14.8%

The chart suggests that there is no relationship between the number of judges in a circuit and the number or rate of opinions produced. Further, a high volume of circuit opinions is an asset to circuit administration because precedential opinions settle circuit law. This is of great assistance to district judges, as Chief Judge John Coughenour testified to this Subcommittee last year. Further, circuit division would create the need for multiple panels in each new circuit to revisit issues, creating an enormous waste of judicial resources.

7. Collegiality

Collegiality is often cited as a reason to create smaller circuits. In many cases, judges on smaller circuits have enjoyed a strong rapport. This doesn't mean, however, that judges on a larger circuit cannot achieve a similar rapport. Indeed, as most judges on our Court have testified repeatedly, we enjoy a very collegial atmosphere on our Court, despite differences of opinion. In some ways, a larger court is better able to absorb strong personality differences. When personal differences arise on a smaller court, a court may become rapidly dysfunctional. There are many examples of this. My point is not to argue that a larger circuit is more, or less, collegial than a smaller circuit; only to point out that a close working environment does not always produce collegiality.

Some proponents of a split have argued that the judges on our Court do not sit in panels as often as these observers believe they should. However, a careful look at other circuits should show that this is an exaggerated problem. For example, the Eleventh Circuit, which was touted as an example to the Committee employs a large number of visiting judges. Indeed, 66% of the published opinions of the Eleventh Circuit involved a visiting judge on the panel. In contrast, only 33% of the published opinions of the Ninth Circuit involved a visiting judge. This is not to criticize the practice of the Eleventh Circuit, by any means. However, the point is that paring the size of a Circuit does not necessarily mean that judges will be sitting with each other more often. Indeed, as caseload increases, more visiting judges will be required, and the so-called collegiality created by frequency of sitting will be diminished.

On our Court, we have daily substantive interchanges of opinions and ideas through e-mail, some of them quite spirited. We sit often together on en banc panels. We have frequent contact. One excellent measure of collegiality is the degree to which judges resolve differences. Well over 90% of the cases are decided by unanimous vote. Further, there has been an increasing trend on our Court for off-panel judges who have concerns about panel opinions being able to work out differences with the panel without proceeding to a vote on whether to rehear the case en banc. During 2003, there were thirteen en banc calls or potential en banc calls that did not result in a ballot because the panel agreed to amend its opinion. This amounted to almost a quarter of the en banc calls. Given the frequency of communication and the internal indicia of collegiality, additional panel sittings would not materially improve our understanding of each other, at least in my opinion.

Nor would a circuit division necessarily produce a closer working environment. The geography of the Ninth Circuit, regardless of how it might be divided, precludes daily person-to-person contact. A single judge located in Hawaii, Alaska, or Montana is not going to have daily in person contact with other circuit judges, regardless of circuit configuration. In any circuit, for example, my chambers would not be located within driving distance of any other chambers. The daily in person interaction between judges will not change with a circuit split. The primary contact of the judges in any circuit division would remain as it is now, primarily by e-mail and telephone. Personal contact would be limited to court meetings and oral arguments. The illusion of increasing personal contact is not a reason to divide the Circuit.

8. Connection with Community

Coming from a less populated state, I feel strongly that a court must have a strong connection with the community it serves. Part of the premise for change is that smaller circuits would promote that. However, attorneys in states like Montana are unlikely to feel a significantly more intimate connection with a Circuit whose headquarters is in Seattle or Las Vegas or Phoenix, as opposed to a Circuit whose headquarters is in San Francisco. Likewise, no circuit division would place all circuit judges in an intimate environment; they would still maintain chambers hundreds or thousands of miles apart.

The best method of establishing and maintaining a sense of community is through the use of technology and through continued contact between the Circuit and community it serves. To that end, we have made enormous strides over the past several years. Ninth Circuit opinions are immediately posted on the Circuit's website, which contains an enormous amount of useful information. Digitized audio files of Ninth Circuit arguments are available on the website the day after argument. The Clerk's office has made briefs, orders, and audiofiles of cases in which the public has expressed an interest immediately available via the internet. Video argument will soon be available to litigants who cannot afford to travel in person for oral argument. Many of these advances were hastened as a result of conferences between the bench and bar of the states in the Ninth Circuit. Technology allows the Circuit to stay in close contact with the community it serves. However, technology is not always cheap. Because the Ninth Circuit has pooled resources, it can continue to improve the service it provides to litigants and the public. However, the resources for doing so would be seriously diminished in a small circuit.

9. Travel.

The costs and personal impacts of judicial travel has also been cited as a rationale for circuit division. However, regardless of how the circuit may be reconfigured, judicial travel is unlikely to be reduced significantly. All of the bills contemplate multiple places of holding court. In some instances, depending on the proposed legislation, judges and attorneys will have to travel more than they do at present. A small handful of judges may personally benefit, but the net savings are negligible, if any.

10. Summary.

None of the critics of the Ninth Circuit have demonstrated how division would improve judicial administration. When the specific critiques are examined, none provides a justification for the radical remedy of circuit division.

Analysis of the Proposals to Divide the Ninth Circuit

In my view, there are six important criterion for the creation of a new circuit: (1) the new circuit must have sufficient critical mass; (2) the division should allocate cases in approximately equal proportions; (3) the new circuit must have geographic coherence; (4) the new circuit should have jurisprudential coherence; (5) division should increase the efficiency of judicial administration, and (6) the division should be supported by a consensus of the affected court. Unfortunately, each of the proposed structural alternatives fails to meet this criteria; by contrast, the existing structure is satisfies it.

1. Critical mass.

A circuit court of appeals must have sufficient caseload and budget to be viable. Any three-way division of the circuit would create circuits that lack a critical mass of cases and resources. Compounding the problem is the fact that any newly carved out circuit in the western United States would cover vast geographic territory. Limited resources would be stretched to the breaking point. Likewise, any two-way division of the circuit that involves a small subset of states (such as the proposed Northwest division) would suffer from the same infirmity.

2. Proportionality.

None of the current proposals would divide the circuit equally in terms of caseload. The only proposal that has been forwarded in the past that achieves rough proportionality is the Hruska Commission proposal which would divide California and place the Northern and Eastern Districts of California into a Circuit along with the Northwestern states (Alaska, Washington, Oregon, Idaho, and Montana), and place the Central and Southern Districts of California into a circuit with the Southwestern states (Nevada and Arizona) and the Pacific jurisdictions (Hawaii and the territories). That proposal, however, suffers from fatal jurisprudential flaws.

3. Geographic coherence.

The proposed Northwest split and the three-way split have geographic coherence, meaning that there is a sufficient geographic nexus to allow viable geographic governance. The "stringbean circuit" proposal (Alaska, Washington, Oregon, Idaho, Montana, Nevada, and Arizona) lacks geographic coherence. Central administration in either Phoenix or Seattle would prejudice the attorneys not located in those regions. It would also mean that many attorneys would have a greater distance to travel to the circuit headquarters. Even more deficient is the "hopscotch circuit" proposal (Alaska, Washington, Oregon, Idaho, Montana, and Arizona) that passed the United States Senate a few years ago. That configuration would leave Arizona without a border with any other state in the circuit.

4. Jurisprudential coherence.

Any division will disrupt Ninth Circuit jurisprudence. This is not only true because of the development of federal law, but because most of the states which form the Ninth Circuit have strong jurisprudential ties to California. California adopted the Field Code in 1850, followed by Oregon and Washington in 1854; Nevada in 1861; and Arizona, Idaho and Montana in 1864. In addition, all the other Ninth Circuit states have adopted significant aspects of California law, and rely on California judicial construction.

The present configuration promotes judicial coherence by developing consistent federal law in areas affecting business in the West: admiralty, timber, Native American rights, and intellectual property - just to name a few.

The worst configuration in terms of jurisprudential coherence was the Hruska Commission proposal to divide California into two circuits. Adoption of that configuration would meant that California would be subject variant interpretations of federal and state law. Challenges to state law and initiatives could be brought in either circuit, with the possibility of inconsistent results.

5. Efficient judicial administration.

As I have previously discussed, any circuit division will dramatically decrease the efficiency of judicial administration by requiring replication of core functions, and reduction of vital staff functions.

6. Consensus.

To date, Congress has never divided a Circuit unless there was a consensus of the judges on the Circuit that division was required. Not only is there no consensus among Ninth Circuit judges supporting a division, but the vast majority of judges oppose the split. In fact, only 3 of the 25 active judges of the circuit favor circuit division.

For all of these reasons, I oppose a structural division of the Ninth Circuit. The best means of addressing the present challenges is within the existing structure. Division will be costly, inefficient, ineffective, and result in the significant impairment of the administration of justice in the Western United States. I thank the Subcommittee for its consideration of my views and those of my colleagues.