In Re Continental Illinois Securities Litigation, 813 F. Supp. 633 (N.D. Ill. 1993)

Lowell E. Sachnoff, Sachnoff & Weaver, Chicago, IL, for petitioners.

GRADY, District Judge.

This class action was concluded on the merits in 1987, but the question of fees for class counsel has still not been resolved. The most recent development is an order of the Seventh Circuit Court of Appeals issued on January 21, 1993, directing me "to issue within 30 days of today an order setting forth a procedure and timetable for resolving this matter by the entry of a final fee award within 120 days of today." In the Matter of Continental Illinois Securities Litigation, 985 F.2d 867, 869 (7th Cir. 1993). The purpose of this opinion is to comply with that direction. In order to understand what follows, some background is necessary.

The case arose out of the collapse of the Penn Square Bank of Oklahoma City in 1982. The Continental Illinois National Bank and Trust Company of Chicago had purchased over $1 billion worth of the loans Penn Square had made to oil producers during the 1970's and early 1980's. *634 There was a drastic decline in world oil prices in 1982. Most of the borrowers defaulted at that time, and it was immediately evident that the loans had been inadequately secured. As a result, Penn Square Bank was closed, and Continental Bank itself became insolvent and was taken over by the FDIC. Four different class actions were filed in July and August of 1982, shortly after the closing of Penn Square Bank, by various Continental shareholders who alleged that in purchasing their stock they had been misled by misrepresentations of Continental and its officers concerning the quality of Continental's energy loan portfolio. Also joined as a defendant was Continental's auditing firm, Ernst & Whinney.

The four cases were consolidated, and it was apparent that there was no need for the large number of plaintiffs' lawyers who had filed appearances in the four separate cases. Several disputed matters had to be decided before I could certify a class, but pending class certification, I entered an order on June 21, 1983, establishing guidelines that would govern any eventual award of attorneys fees. I also suggested to counsel that they attempt to agree among themselves who would be designated class counsel. See In re Continental Illinois Securities Litigation, 572 F. Supp. 931 (N.D.Ill.1983). At that time, there was no doubt in anyone's mind that the "lodestar" method would be used to determine fees. I knew the time spent on the case would be substantial and that accurate and descriptive records of the work done by the attorneys would be essential to a lodestar determination. The order specified a number of techniques for avoiding duplication of effort and for keeping records of activity in a way that would facilitate a determination of whether the particular work was necessary and the time claimed was reasonable. See id. at 933-35. Two of these procedures, relating to legal research and lawyer conferences, are of particular relevance for present purposes:

The description of the work done should be sufficient to demonstrate that it benefitted the class or contributed to the recovery of the common fund. Notations such as "research re class action" will not suffice. The particular question researched should be described. Much of the narrative in most fee petitions consists of entries like "conference with GBS re motion to compel." As indicated above, such "conferences" should be held only when necessary, which should not be very often. But in no event would that kind of entry be sufficient to show the conference was necessary and productive. There should be a statement, albeit very brief, of specifically what was discussed and what conclusion was reached. Should such a statement necessarily include privileged information (which seems unlikely, since it will be submitted at the end of the case), it may be submitted in camera.

Id. at 934-35. I did not consider this an academic exercise. I had seen enough "legal research ... 2.5 hours" type entries in lawyers' time records to know how common they are, and I knew from experience with fee petitions how useless such entries are as an indication of what was done and how long it should have taken. My purpose in entering the order early in the case was to give fair notice to counsel as to what would be needed, so that they could keep their time records accordingly. No complaint was made about the order, and I had no reason to think counsel were not complying with it.

Counsel did not agree on who should represent the class and spent considerable time accusing each other of greed, incompetence, and disloyalty to the class. On November 27, 1984, I appointed Lawrence H. Eiger (now deceased), James F. Fornari, Nicholas E. Chimicles, and Lawrence Walner to represent the class and authorized them to enlist the services of lawyers and paralegals from their firms when necessary. In a notice to the class, counsel agreed that in no event would their fees exceed $9 million.

The case proceeded to a $25 million settlement with Continental, a $13 million settlement with various insurance carriers on behalf of officers of Continental, and an unsuccessful 18 week jury trial of the claim *635 against Ernst & Whinney. The $38 million settlement fund had the benefit of the high interest rates that were available at the time, so that the total amount on deposit by the end of 1989 was $45 million.

Class counsel then presented their fee petition. It sought the full $9 million based on 41,955 hours in attorney and paralegal time. The attorney time was organized into 284 categories of work. Unfortunately, the time was not broken down or described in the manner I had specified in my earlier order, and the bulk of the entries consisted of the kind of vague references I had tried to avoid by that order. In my ruling on the fee petition I described in detail the difficulties I had with the attorney time and will not repeat the description here. See In re Continental Illinois Securities Litigation, 750 F. Supp. 868, 878-85, 896-900 (N.D.Ill.1990). Some categories of attorney work stood out as excessive; the time entries in most categories defied analysis because of their vagueness. The thousands of hours in paralegal time were mostly described as "organizing files," or "reviewing," "indexing," or "filing" materials that were not specifically identified. Id. at 891.

At that point, I had to devise a method for determining a fee despite the insufficiencies of the petition. The easiest thing to have done would have been to disallow all time that petitioners had not demonstrated to be reasonable. It is clear that the applicant for a fee has the burden of establishing the reasonableness of the number of hours and the hourly rate claimed. See Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S. Ct. 1933, 1941, 76 L. Ed. 2d 40 (1982); see also Blum v. Stenson, 465 U.S. 886, 897, 104 S. Ct. 1541, 1548, 79 L. Ed. 2d 891 (1984). The Seventh Circuit has noted that a judge "may eliminate hours that are not documented in sufficient detail." Tomazzoli v. Sheedy, 804 F.2d 93, 98 n. 5 (7th Cir. 1986). On the other hand, it was obvious that counsel had spent considerable time on the case and that an order disallowing all hours not adequately documented would deprive them of compensation for most of that time. I could think of no satisfactory solution but adopted an approach that was highly favorable to the petitioners:

In an effort to be as fair as possible to petitioning attorneys, I will disallow hours only where I am convinced the time is excessive or that the time did not significantly benefit the class. I will resolve moderate doubts in favor of petitioners, which means that an enormous amount of time evidenced only by vague entries will be approved. Most categories will survive intact as far as the number of hours is concerned, and I will discuss only those categories where I order a reduction. The reductions are in time spent on work I warned counsel would be subject to particular scrutiny: excessive time spent on legal research; too much time spent on "reviewing" the work of others; and too much time spent by counsel "conferring" with each other. In re Continental Illinois Securities Litigation, 572 F. Supp. at 933, 934-935. I have not disturbed the paralegal and law clerk time but have concentrated on the attorney time, which I find more susceptible to analysis.

In re Continental Illinois Securities Litigation, 750 F. Supp. at 880 (footnote omitted).

The next question was how to arrive at the appropriate reductions in those categories I could identify as clearly excessive. Several Seventh Circuit cases had approved percentage reductions, recognizing the difficulty of making a specific ruling on every individual time entry. I adopted the percentage approach, In re Continental Illinois Securities Litigation, 750 F. Supp. at 880, and made percentage reductions in 21 categories of work. Id. at 881-85, 896-900. I also reduced the rates claimed for paralegals. Id. at 889-93. A more recent Seventh Circuit case has reiterated the court's approval of percentage reductions. See Dutchak v. Central States Pension Fund, 932 F.2d 591, 597 (7th Cir.1991) (percentage reduction affirmed on petition seeking lodestar of over $3 million).

The reductions I made in the 21 categories did not have a major effect on the fee award. What did make a difference was *636 my reduction of the hourly rates requested and my rejection of any multiplier. In re Continental Illinois Securities Litigation, 750 F. Supp. at 885-89, 893-96. This resulted in an award that was roughly half the $9 million petitioners requested. (I also allowed reimbursement for most of the claimed expenses.)

My fee order was filed on September 20, 1990. Two weeks earlier, in an unreported opinion of which I had not been aware, Judge James B. Zagel of this court had described an innovative procedure he uses to evaluate fee petitions:

In order to determine whether these hours are excessive I have closely examined, at the request of defendants, two specific areas of counsel's work the preparation of findings of fact and the summary judgment motion. It is my practice to conduct a close examination of two or three particular tasks performed by counsel claiming fees and to apply the findings made there to the remaining hours claimed. Counsel are informed of this practice and counsel opposing the fee is permitted to suggest the specific work that ought to be scrutinized. This sampling procedure operates on the reasonable premise that a lawyer's billing and work habits and practices are, in fact, habits and practices which will uniformly apply to all of the lawyer's work.

Evans v. City of Evanston, 1990 WL 129603, at 1 (N.D.Ill. Sept. 5, 1990). Judge Zagel's order was affirmed on appeal with an express approval of the "sampling" approach:

Fee calculation is an imprecise science, and the numbers arrived at by a district court, figured to the last penny, may be seen by counsel and others to convey an "impression of exactness" that is illusory, even "delusive." ...

Likewise the methodology used to calculate those figures, which represent no more than the learned guesses of district court judges, based on their extensive experience overseeing litigation and observing lawyers, as to how to go about determining what compensation an attorney deserves for her efforts in a given case. "There is no precise rule or formula for making these determinations," and a district court "necessarily has discretion" in choosing among the various alternative methods. Hensley v. Eckerhart, 461 U.S. 424, 436-37, 103 S. Ct. 1933, 1941, 76 L. Ed. 2d 40 (1983). So long as the method selected is not arbitrary and is likely to arrive at a fair fee, it will not be disturbed on appeal.

We are unable to conclude that the sampling technique employed by the district court in this case falls below this standard. As the district court correctly observed, a lawyer's work habits, while not immutable, are likely to be consistent for the duration of a case litigated over the course of a few years.

Evans v. City of Evanston, 941 F.2d 473, 476-77 (7th Cir.1991) (citations omitted).

Meanwhile, the fee petitioners in this case took the fees I allowed but appealed my denial of the full amount they had claimed. The Seventh Circuit reversed my order in its entirety. See Matter of Continental Illinois Securities Litigation, 962 F.2d 566 (7th Cir.1992). The portion of the opinion that is relevant to my compliance with the recent direction from the court concerns the determination of the number of compensable hours. Holding that I erred in reducing the claimed time by percentages, the court stated that

rather than disallowing specific items of unreasonable activity the judge slashed broad categories of activity by arbitrary percentages. That won't do, especially when millions of dollars are at stake. In re Fine Paper Antitrust Litigation, 751 F.2d 562, 594-95 (3d Cir.1984). The qualification is important: in small cases, the amount at issue in the request for lawyers' fees may be too slight to justify cutting it with laser precision. The meat-axe approach (we called it "trimming fat from a fee application" in Tomazzoli v. Sheedy, supra, 804 F.2d at 98) may be acceptable in such a case. (The judge awarded only $6,000 in attorney's fees in Tomazzoli.) See also Ustrak v. Fairman, supra, 851 F.2d [983] at 989 [(7th Cir.1991)] ($30,000). It is true *637 that in a large case, where the stakes in the attorney's fee determination are greater, so is the burden on the judge because the fee application will be longer and more complex. But in Evans v. City of Evanston, 941 F.2d 473, 476-77 (7th Cir.1991), we upheld the district judge's use of a sampling procedure to determine the reasonableness of a fee request so that he wouldn't have to scrutinize every entry on the lawyers' billing sheets. A judge is not permitted to destroy substantial entitlements to attorneys' fees on the basis of his inarticulable and unsubstantiated dissatisfaction with the lawyers' efforts to economize on their time and expenses.

Id. at 570 (emphasis added). The court observed that

[t]he object in awarding a reasonable attorney's fee, as we have been at pains to stress, is to give the lawyer what he would have gotten in the way of a fee in an arm's length negotiation, had one been feasible. In other words the object is to simulate the market where a direct market determination is infeasible. It is infeasible in a class action because no member of the class has a sufficient stake to drive a hard or any bargain with the lawyer. So the judge has to step in and play surrogate client.

Id. at 572. The court suggested that on remand I consider setting a fee at a percentage of the recovery in the case:

The class counsel did not present and the judge did not ask for testimony or statistics concerning the fee arrangements in commercial litigation comparable to the present suit. Yet it might be quicker and easier to generate such evidence than it would be to hassle over every item or category of hours and expense and what multiple to fix and so forth. We also remind the bench and bar of our approval in the Evans case of Judge Zagel's sampling approach to the auditing of lawyers' hours.

Id. at 572-73 (emphasis added).

Following remand, Attorney Lowell E. Sachnoff appeared for the fee petitioners and urged that I adopt the percentage approach suggested by the Court of Appeals. I indicated to Mr. Sachnoff that the sampling method, approved by the Seventh Circuit in Evans and suggested by the Seventh Circuit panel in this case as a possible alternative, might be preferable. Another panel of the Seventh Circuit had upheld the sampling technique employed by Judge William T. Hart of this court in a common fund case. Harman v. Lyphomed, Inc., 945 F.2d 969, 975 (7th Cir.1991). The panel rejected class counsels' argument that the judge should have used a percentage method; rather, the court stated that "whether to use a percentage method remains in the discretion of the district court." Id. In regard to the sampling method, the court stated:

[W]e find no error in the district court's use of a "sampling" procedure to estimate the number of hours appropriate for compensation. The court selected three significant segments of the litigation that it felt reflected typical attorney work in the case. It then made findings regarding the attorney billing practices and thought this likely represented billing practices for the entire litigation.

Id.

Mr. Sachnoff filed a memorandum in support of a percentage award. The memorandum included as an exhibit a group of affidavits, letters from attorneys, and redacted copies of fee contracts entered into by various law firms in commercial litigation allegedly comparable to this case. Petitioners' memorandum argued that these materials established that "the market place provides attorneys with contingent fees ranging from a low of 20 percent to a high of 40-50 percent, plus reimbursement of out-of-pocket expenses." Memorandum of Class Counsel in Support of Award of Attorneys Fees ("Petitioners' Memorandum"), filed June 16, 1992, at 2. I studied these materials and was not persuaded that they provided a reliable guide to what fee arrangements these petitioners could have made in this case back in 1982 "in an arm's length negotiation, had one been feasible." Matter of Continental Illinois Securities Litigation, 962 F.2d at 572. The question is whether I would have authorized a percentage *638 arrangement back in 1982, when class counsel were appointed. (The inquiry is somewhat anomalous, inasmuch as there was no legal authority whatever for a district court to do such a thing in a class action at that time; the Supreme Court had made it clear that the lodestar approach was required.) Contracts entered into by lawyers at that approximate time would tend to show what the market was then. But most of the contracts referred to in the petitioners' materials were entered into in the late 1980's or 1990's, when some circuits were beginning to allow departures from the lodestar requirement. Most of the cases referred to in petitioners' materials were not class actions, which often involve the possibility of a large common fund, but instead were individual actions, where the prospect of a relatively small recovery may justify a larger percentage going to the attorney. The information provided by petitioners gives no basis for assessing the strengths of the cases, and in some instances not even the nature of the claim is apparent. Some of the contracts are not straight contingent fees but provide for a "blended" fee, consisting of hourly rates plus a percentage of the recovery. Finally, almost none of the contracts involve any of the petitioning firms in this case. The two exceptions are Mr. Sachnoff's firm and the firm of Bernstein, Litowitz, Berger and Grossman, which petitioned for relatively small amounts of money for work they did in representing the class. See In re Continental Illinois Securities Litigation, 750 F. Supp. at 896-900. Some contingent fee contracts of these two firms are referred to in petitioners' materials. However, there is no reference to any contingent fee contracts entered into at any time by any of the three petitioning firms which in combination did almost all of the work for which compensation is sought. These are the firms of Much Shelist Freed Denenberg Amend & Eiger, Greenfield & Chimicles, and Lawrence Walner & Associates. It seems fair to conclude that these firms did not believe they had been parties to any fee contracts which would tend to support their present contention that percentage fees are, even in 1993, the way they do business. In fact, if one is familiar with the decided cases throughout the country dealing with lodestar fee determinations in federal class actions, the names of these three firms appear as petitioners about as often as any.

There is little evidence that percentage fees have caught on even now, ten years later. The fee petitioners here, with one exception, are large law firms, and all are located in big cities Chicago, Philadelphia and New York. Billing by such firms is almost invariably by the hour. Preoccupation with billable hours is such that many, if not most, large firms now establish quotas that lawyers are expected to meet, such as 2200 billable hours per year. Failure to meet these quotas is usually fatal to the lawyer's progress in the firm. One need only talk occasionally to young lawyers to know what widespread dissatisfaction there is with these quotas and the time commitments they require. The point here is not to argue whether the hourly billing regime is good or bad, but to indicate that it is still the prevailing method for billing clients in all kinds of commercial litigation. See Ellen Joan Pollock, New Methods of Legal Billing Find Few Fans, WALL ST. J., Nov. 4, 1992, at B1.

Even if the materials submitted by petitioners supported the proposition that the market for plaintiffs' attorneys in federal class action cases in the early 1980's was based on percentage fees, they would not help me arrive at a percentage figure that would fairly reflect how this case would have fared in that market. Petitioners argue for 20 percent, and that takes them up to their $9 million cap. This pragmatic approach is easy for them to take, but as the remand opinion pointed out, it is not feasible for a class member to drive a hard bargain with the lawyer "because no member of the class has a sufficient stake to drive a hard or any bargain with the lawyer. So the judge has to step in and play surrogate client." Matter of Continental Illinois Securities Litigation, 962 F.2d at 572. In playing "surrogate client," how high should I go? Counsel in the individual class actions were competing so *639 fiercely to become designated counsel for the class that I think I could have driven a very hard bargain. The lawyers were literally fighting each other for the case. See In re Continental Illinois Securities Litigation, 750 F. Supp. at 874-75. I did not ask them to "bid" for the privilege of representing the class, but that is really the only way I could have determined what the market was. That very point was made by Judge Vaughn R. Walker in the case of In re Oracle Securities Litigation, 131 F.R.D. 688, 697 (N.D.Cal.1990) (selection of class counsel will be made through competitive bidding); 132 F.R.D. 538 (class counsel selected on basis of competitive bids); 136 F.R.D. 639 (denying motion to reconsider). By opening up the selection of class counsel to competitive bidding at the outset of the case, Judge Walker was able to determine the best bargain available to the class in the market. (He prohibited consultation among the bidders.) That kind of process was not available to me in this case in 1982, nor am I able to reconstruct the situation now so as to test how small a percentage any of these petitioners would have been willing to take.

Petitioners' memorandum argued that if I were to use the lodestar method again, "excluding all hours disallowed by this Court in its September 20, 1990, opinion," (emphasis in original), and follow the mandate of the Court of Appeals in regard to hourly rates and risk multipliers, the result would be an award that would still reach $9 million. Petitioners' Memorandum at 3. In other words, in petitioners' view, whether I use a percentage or lodestar method, the result will be the same. Implicit in this argument is the assumption that the time I did not reduce originally, but left intact because it was not described by petitioners sufficiently for me to determine whether it had been reasonably expended, was exempt from reconsideration on remand of the case. I do not agree with that premise. The fee order of September 20, 1990, was reversed in its entirety. The case was remanded "for a redetermination of the fee due class counsel." Matter of Continental Illinois Securities Litigation, 962 F.2d at 572.

Mr. Sachnoff suggested several times that a conference in chambers might be helpful in resolving the matter, and on November 17, 1992, we did have a meeting. Mr. Sachnoff reiterated his position that a fee based on a percentage of the recovery made the most sense and would avoid additional work that would be required by the lodestar method. I explained why I did not think a percentage approach was appropriate here, and commented specifically on why the percentage fee materials he had furnished me were not persuasive. I told him that I would be willing to consider a percentage arrangement in future class actions, but only in connection with a bidding procedure such as that used by Judge Walker in Oracle Securities, so that there would be a means of obtaining the best price for the class.

I told Mr. Sachnoff that I was still of the view that sampling would be the better approach. He suggested, as he had in his earlier written submissions, that I had already done an exhaustive examination of all time entries and that there was no point in doing it again. I told him, as I had explained in the fee order, that I had not been able to do an adequate analysis of the time entries because of the format in which they were submitted. The sampling method would allow me to select a few categories of work that I was most familiar with, make a determination as to the reasonableness of the time claimed in those categories, and then apply the result to the balance of the entries. I believe we understood each other, but the meeting brought us no closer to agreement.

On December 1, 1992, I entered the following order:

The court has considered the request of class counsel that their fees be calculated on the basis of a percentage of the settlement fund. The court believes that a preferable method would be the "sampling" technique approved in Evans v. City of Evanston, 941 F.2d 473, 476-77 (7th Cir.1991), and recommended by the Court of Appeals as a possibility in this case.

*640 The court will proceed on that basis and may invite further comment by petitioners during the process. All attorney time and all paralegal time claimed in the case will be subject to the sampling.

Should petitioners have any preferences as to sampling methodology, they may file a memorandum setting forth their suggestions.

Petitioners' response was to file a mandamus petition against me in the Court of Appeals, requesting that I be directed to comply with the mandate. Mr. Sachnoff argued on behalf of petitioners that my order of December 1 indicates that I intended to "review again all of Class Counsels' hours, the vast majority of which were never subject to appeal and were awarded by Judge Grady after his previous thorough review." Petition for Writ of Mandamus, at 4. The argument was repeated throughout the petition:

Judge Grady's December 1, 1992, order reveals that he intends to review yet again all of the attorney hours and all of the paralegal hours spent in the entire case utilizing a sampling procedure.

Id. at 6.

Standing on its own, Judge Grady's intention to sample "[a]ll attorney time and all paralegal time" defies logic and would amount to a complete duplication of what he has already done.

Id. at 7.

Seven months after this Court issued its decision, Judge Grady proposed to begin to do a review of all time expended in the case, a review which took him one year the first time he did it.

Id. at 12.

Petitioners' interpretation of my December 1 order was the exact opposite of the meaning I had intended to convey. My plan was not to review all of petitioners' time entries but instead to sample a limited number of those entries and apply the result to the balance of the entries. It had not occurred to me that my use of the word "sampling" would fail to convey that meaning, especially in the context of my citation of Evans v. City of Evanston, where Judge Zagel had analyzed only two categories of the attorneys' work and applied the result to the remaining categories. In stating that all attorney time and all paralegal time "would be subject to the sampling," I meant to inform the petitioners that the time categories I did not reduce in my September order would not be exempt from the recalculation but would be governed by the result of the sample.

Petitioners' misinterpretation of my December 1 order seemed so obvious to me that I had no doubt the Court of Appeals would see their error. I also assumed that if the court had any doubt about my intentions it would ask for my response to the mandamus petition. Rule 21(b) of the Federal Rules of Appellate Procedure, governing writs of mandamus directed to judges, provides that

[i]f the court is of the opinion that the writ should not be granted, it shall deny the petition. Otherwise, it shall order that an answer to the petition be filed by the respondents within the time fixed by the order. The order shall be served by the Clerk on the judge....

No such order was entered, and I therefore made no response to the mandamus petition.

On January 21, 1993, the Court of Appeals filed the opinion which has occasioned this discussion. In the Matter of Continental Illinois Securities Litigation, 985 F.2d 867 (7th Cir.1993). The opinion referred to a "mass of affidavits concerning contingent fees charged in comparable multimillion dollars commercial suits," id. at 868 (the identical materials discussed at 637-38 supra), and stated that my order of December 1 had "failed even to mention this evidence, let alone offer any ground for rejecting it in favor of restarting the fee inquiry, almost four years after it began, from scratch." Id. at 869.

The order of December 1 was not intended to be my final opinion on fees in this case. It was a communication to the fee petitioners that I had elected to use the sampling method and that I invited their suggestions as to the specifics. I did not anticipate that the order was going to be *641 the subject of any appellate action. The order was not intended to explain anything, let alone something that requires as much discussion as my reaction to petitioners' materials concerning percentage fees and my rejection of the percentage method in this case. I had explained my position to Mr. Sachnoff in the conference I had with him. I expected to file a detailed opinion explaining the basis for my ultimate redetermination of fees, just as I had done when I determined fees originally. I was not ready to do that on December 1 and was not aware of any reason to say more than I did in the order I entered on that date.

The Court of Appeals appears to believe that I intended a reexamination of all time entries in the fee petition. In referring to the mandamus petition, the court stated:

The petition asks us to direct the judge not to undertake, as his order of December 1 suggests he intends to, a review of the fee application from the ground up a process that could well take another year and be followed by another appeal, even though no member of the class objects to the $9 million award sought by the lawyers.

In the Matter of Continental Illinois Securities Litigation, 985 F.2d at 869. On the same page, the court remarked that

the order of December 1 implies that the judge means to recompute the fee award from the ground up on the basis of an unspecified method of sampling from the total number of hours 42,000 listed in the time sheets.

Id. I hope I have made clear that reexamining all of the time sheets is the farthest thing from my mind. The reference to "an unspecified method of sampling" is accurate in part, but it is not clear to me how my citation of Evans could fail to convey that I intended to use the method approved in that case. It is true that I have not yet decided what segments of work should serve as the sample, or how many there should be. Judge Zagel used only two categories as a sample in Evans, and Judge Hart used only three in Harman. I am limited here by the relative paucity of useful work descriptions, but I believe I can find a sufficient number to serve as a fair sample. In the order of December 1, I elicited petitioners' suggestions and would still entertain them in the event the fee determination remains my responsibility in this case.

A more serious problem is certain language of the court which suggests agreement with petitioners' argument that the time I did not disallow in my original order is insulated from reevaluation on remand:

Our opinion did mention the possibility of sampling time sheets in order to economize on a district judge's time when he is faced with a fee application that covers many thousands of hours of lawyer and paralegal time, but in context it should have been apparent that we were not suggesting that the procedure be followed on remand in a case in which the district judge had already reviewed all the time sheets. The only error the judge had made that related to the number of hours submitted by the lawyers was in arbitrarily reducing the hours for research and conferences. The correction of that error would require at most a sampling of those hours....

In the Matter of Continental Illinois Securities Litigation, 985 F.2d at 869 (emphasis in original). I did not read the remand opinion as directing me to employ any particular method and had the impression that the decision was being left to me. As I understand it, the court did not prohibit me from reexamining the whole petition if I thought it necessary, but suggested the sampling and percentage methods as ways to avoid that. If the court is now saying that the mandate prohibits sampling, the court obviously has that power. But if the opinion of January 21 is simply suggesting that on the facts of this case sampling makes no sense, let me state why I think it does. First, it does not appear to me that the only determinations that were reversed by the remand order were those regarding legal research and lawyer conferences. I made percentage reductions in twenty-one categories of work. See In re Continental Illinois Securities Litigation, 750 F. Supp. at 881-85. It is true that a large part of *642 the basis for the reductions was my belief that excessive time was being claimed for legal research and conferences. But I also pointed out other respects in which the claims were excessive. There was, for example, an excessive 169 hours claimed for drafting the consolidated complaint, which was simply a slightly amended version of the four original complaints filed by the competing candidates for class representative. I held that the 165 hours claimed for drafting essentially boilerplate orders in regard to discovery and organization of plaintiffs' counsel was excessive. I disallowed most of the time claimed for the efforts to certify inappropriate class representatives because those efforts were not for the benefit of the class but the benefit of attorneys who desired to represent the class. Other claims were made for time spent not on the class action but on the separate derivative action, and this was not properly chargeable to the class. The time claimed for document review by counsel was enormous and much of it appeared to be excessive. And, of course, there was the whole body of paralegal time, none of which had anything to do with the excessive time claimed by attorneys for legal research and conferring. This description of the miscellaneous nature of the excessive time in this case could be extended, but I hope the point is made that, as far as I am concerned, a redetermination of the fee is not simply a matter of reexamining the research and conference hours I reduced.

The statement of the Court of Appeals that I "had already reviewed all the time sheets" (emphasis in original) is literally true if "review" means that I examined them. I did, as I explained in the opinion, "examine" each of the hundreds of pages of time sheets, but I also explained that most of these pages gave me no basis for determining whether the described activity benefited the class or should reasonably have taken the amount of time claimed. In many, if not most instances, it was not even possible to tell how much time was being claimed for a particular item of work, since so many of the time entries simply listed a series of separate items of work, often unrelated, with a total number of hours claimed for the series, without a breakdown of how the time was allocated. See In re Continental Illinois Securities Litigation, 750 F. Supp. at 878-79.

For these reasons, I believe it would be unfair to the class and a windfall to petitioners to give petitioners automatic credit for the thousands of hours I did not reduce the first time. The impact of doing this would be magnified by the court's direction that I must allow higher hourly rates and also award a risk multiplier. Almost all of the hours used as the base of the multiplier would be hours that are not susceptible of direct evaluation. Sampling offers a way to avoid this result. Had sampling occurred to me, and had I thought it would meet with appellate approval, I would have used that method the first time. It seems to me that it is not too late to use it now.

Turning to the specific inquiry of the court, the procedure I would use to redetermine the fees is as follows:

1. I would select a small number of categories of attorney work, perhaps six, to use as a sample. I would determine whether there are "specific items of unreasonable activity," Matter of Continental Illinois Securities Litigation, 962 F.2d at 570, which should be disallowed. It has never been possible for any judge to decide precisely how long any piece of work done outside his or her presence did or should have taken. It is not possible to recreate the activity of a lawyer which is described in a few words in an entry on a time sheet. The best the judge can do is to apply his or her experience, judgment and knowledge of the particular case to the information furnished by counsel and make a decision as to whether the time claimed appears reasonable. "If the hours submitted for a particular category of attorneys or a segment of work appear high, the court should adjust them accordingly." Harman, 945 F.2d at 975 (emphasis added). Obviously, the better the work is described the less difficult the evaluation is, but in no event is it easy.

*643 2. I would apply the result of this sample to the rest of the attorney hours, in accordance with Evans and Harman.

3. I would also apply the result of the sample to all paralegal time, on the theory that there would be a correlation between the efficiency of attorneys and the efficiency of paralegals working under their direction. I realize that, this, too, is inexact, but in view of the vague descriptions of what the paralegals did, I am unable to think of a better alternative.

4. I would apply what I find to be the appropriate market rates for the attorney and paralegal time found to be compensable.

5. I would determine an appropriate multiplier and also allow for the delay in payment, in accordance with the mandate.

I am as anxious as anyone to bring this matter to a conclusion. Had the mandamus petition not been filed, I would have begun the sampling process and would probably be fairly well along by now, if not finished with it. In view of the tone of the panel opinion of January 21, however, it would be presumptuous to assume that the panel will approve of the procedure I have outlined above. I will not initiate that procedure unless and until the court does signal its approval. I believe I could complete the process within 60 days thereafter.