PLEASE DON'T CITE THIS!

Why We Don't Allow Citation to Unpublished Opinions

By Alex Kozinski and Stephen Reinhardt

Like other courts of appeals, the Ninth Circuit issues two types of merits decisions: opinions and memorandum dispositions, the latter affectionately known as in memdispos. Opinions contain a full-blown discussion of legal issues and are certified for publication in the Federal Reporter. Once final, they are binding an all federal judges in the circuit- district, bankruptcy, magistrate, administrative, and appellate. Until superseded by an en banc or Supreme Court opinion, they are law of the circuit and may be cited freely; indeed, if they are directly on point, they must be cited.

The rule is different for memdispos. Pursuant to Ninth Circuit Rule 36-3, memdispos are not published in the Federal Reporter, nor do they have precedential value. Although memdispos can be found on Westlaw and Lexis, they may not be cited. So far as Ninth Circuit law is concerned, memdispos are a nullity.

Few procedural rules have generated as much can control as the rule prohibiting citation of memdispos. At bench and bar meetings, lawyers complain at length about being denied this fertile source of authority. Our Advisory Committee on Rules of Practice and Procedure, which is composed mostly of lawyers who practice before the court, regularly proposes that memdispos be citable. When we refuse, lawyers grumble that we just don't understand their problems.

In fact, it's the lawyers who don't understand our problems. Court of appeals judges perform two related but separate tasks. The first is error-correction: We review several thousand cases every year to ensure that the law is applied correctly by the lower courts, as well as by the many administrative agencies whose decisions we review. The second is development of the circuit's law: We write opinions that announce new rules of law or extensions of existing rules.

Writing a memdispo is straightforward. After carefully reviewing the briefs and record, we can succinctly explain who won, who lost, and why. We need not state the facts, as the parties already know them; nor need we announce a rule general enough to apply to future cases. This can often be accomplished in a few sentences with citations to two or three key cases.

Writing an opinion is much harder. The facts must be set forth in sufficient detail so lawyers and judges unfamiliar with the case can understand the question presented. At the same time, it is important to omit irrelevant facts that could form a spurious ground for distinguishing the opinion. The legal discussion must be focused enough to dispose of the case before us yet broad enough to provide useful guidance in future cases. Because we normally write opinions where the law is unclear, we must explain why we are adopting one rule and rejecting others. We most also make sure that the new rule does not conflict with precedent or sweep beyond the questions fairly presented.

While a memdispo can often be prepared in a few hours, an opinion generally takes many days (often weeks, sometimes months) of drafting, editing, polishing, revising. Frequently, this process brings to light new issues, calling for further research, which, in turn, may send the author back to square one. In short, writing an opinion is a tough, delicate, exacting, time-consuming process. Circuit judges devote something like half their time, and their clerks' time to cases in which they write opinions, dissents, or concurrences.

Once an opinion is circulated, the judges on the panel and their clerks scrutinize it very closely. Often they suggest modifications, deletions, or additions. It is quite common for judges to exchange lengthy memoranda about a proposed opinion. Sometimes, differences can't be ironed out, precipitating a concurrence or dissent. By contrast, the phrasing (as opposed to the result) of a memdispo is given relatively little scrutiny by the other chambers; dissents and concurrences are rare.

Opinions take up a disproportionate share of the court's time even after they are filed. Slip opinions are circulated to all chambers, and many judges and law clerks review them for conflicts and errors. Petitions for rehearing en banc are filed in about three-quarters of the published cases. Based on the petition and an independent review of the case, off-panel judges frequently point out problems with opinions, such as conflicts with circuit or Supreme Court authority. A panel may modify its opinion; if it does not, the objecting judge may call for a vote to take the case en banc. In 1999 there were 44 en banc calls, 21 of which were successful.

Successful or not, an en banc call consumes substantial court resources. The judge making the call circulates one or more memos criticizing the opinion, and the panel must respond. Frequently, other judges circulate memoranda supporting or opposing the en banc call. Many of these memos are as complex and extensive as the opinion itself. Before the vote, every active judge must consider all of these memos, along with the panel's opinion, any separate opinions, the petition for rehearing, and the response thereto. The process can take months to complete.

If the case goes en banc, eleven judges must make their way to San Francisco or Pasadena to hear oral argument and confer. Because the deliberative process is much more complicated for a panel of eleven than a panel of three, hammering out an en banc opinion is even more difficult and time-consuming than writing an ordinary panel opinion.

Now consider the numbers. During calendar year 1999, the Ninth Circuit decided some 4,500 cases on the merits, approximately 700 by opinion and 3,800 by memdispo. Each active judge heard 450 cases as part of a three-judge panel and had writing responsibility in a third of those cases. That works out to an average of 150 dispositions - 20 opinions and 130 memdispos - per judge. In addition, each of us was required to review, comment on, and eventually join or dissent from 40 opinions and 260 memdispos circulated by other judges with whom we sat.

Writing 20 opinions a year is like writing a law review article every two and a half weeks, joining 40 opinions is akin to commenting extensively once a week or so on articles written by others. Just from the numbers, it's obvious that memdispos get written a lot faster than opinions- about one every other day. It is also obvious that explaining to the parties who wins, who loses, and why takes far less time than preparing an opinion that will serve as precedent throughout the circuit and beyond. Moreover, we seldom review the memdispos of other panels or take them en banc. Not worrying about making law in 3,800 memdispos frees us to concentrate on those dispositions that affect others besides the parties to the appeal - the published opinions.

If memdispos could be cited as precedent, conscientious judges would have to pay much closer attention to their precise wording. Language that might be adequate when applied to a particular case might well be unacceptable if applied to future cases raising from different fact patterns. And, though three judges might agree on the outcome of the case before them, they might not agree on the precise reasoning or the rule to be applied to future cases. Unpublished concurrences and dissents would become much more common, as individual judges would feel obligated to clarify their differences with the majority, even when those differences had no bearing on the case before them. In short, we would have to start treating the 130 memdispos for which we were responsible, and the 260 memdispos we receive from other judges, as mini-opinions. We would also have to pay much closer attention to the memdispos written by judges on other panels - at the rate of 10 a day.

Obviously, it would be impossible to do this without neglecting our other responsibilities. We write opinions in only 15 percent of the cases already and may well have to reduce that number. Or we could write opinions that are less carefully reasoned. Or spend less time keeping the law of the circuit consistent through the en banc process. Or reduce our memdispos to one-word judgement orders, as have other circuits. None of these are palatable alternatives, yet something would have to give.

Lawyers argue that we need not change our internal practices, that we should just keep doing what we're doing but let memdispos be cited as precedent. But what does precedent mean? Surely it suggests that the three judges on a panel subscribe not merely to the result but also to the phrasing of the disposition.

With memdispos, this is simply not true. Most are drafted by law clerks with relatively few edits from the judges. Fully 40 percent of our memdispos are in screening cases, which are prepared by our central staff. Every month, three judges meet with the staff attorneys who present us with the briefs, records, and proposed memdispos in 100 to 150 screening cases. If we unanimously agree that a case can be resolved without oral argument, we make sure the results is correct, but we seldom edit the memdispo, much less rewrite it from scratch. Is it because the memdispos could not be improved by further judicial attention? No, it's because the result is what matters in those cases, not the precise wording of the disposition. Any refinements in language would cost valuable time yet make little difference to the parties. Using the language of a memdispo to predict how the court would decide a different case would be highly misleading.

We are a large court with many judges. Keeping the law of the circuit clear and consistent is a full-time job, even without having to worry about the thousands of unpublished dispositions we issue every year. Trying to extract from memdispos a precedential value that we didn't put into them may give some lawyers an undeserved advantage in a few cases, but it would also damage the court in important and permanent ways. Based on our combined three decades of experience as Ninth Circuit judges, we can say with confidence that citation of memdispos is an uncommonly bad idea. We urge lawyers to drop it once and for all.