In Re Oracle Securities Litigation

829 F. Supp. 1176, 1993 WL 306658

• Court: District Court, N.D. California
• Decided: August 9, 1993
• Docket: C-90-0931-VRW
• Status: Published

Opinion

ORDER

WALKER, District Judge.

The settlement of these consolidated class and derivative actions must be disapproved because, although plaintiffs’ claims appear to be weak and the consideration offered by Oracle to settle the class action is relatively large, the court is unable to conclude that the Oracle directors who approved the derivative settlement acted with the independence required under Delaware law.

I

The first complaint in this securities class action litigation arrived in the clerk’s office on March 29, 1990, two days after Oracle Systems Corporation, a Delaware corporation with its principal place of business in California, announced what many analysts considered to be lower-than-expected earnings, and a day after the price of Oracle’s stock fell by thirty-one per cent. In the ensuing weeks, law firms from around the country filed eighteen more class action complaints, all of which alleged that Oracle and certain of its officers and directors (“the individual defendants”) 1 had violated the federal securities laws by failing promptly to disclose unfavorable information about Oracle’s financial health. A later amendment added Oracle’s auditor, Arthur Andersen & Co, as a defendant in the class action. Various lawyers also filed two derivative actions on behalf of Oracle against the individual defendants, alleging that they breached their fiduciary duties to the corporation and its share holders by engaging in insider trading and mismanaging Oracle in a manner that reduced Oracle’s profitability and subjected it to litigation expenses and potential liability from the class action. The derivative plaintiffs never named Arthur Andersen as a defendant in their action, but entered into a tolling agreement with Arthur Andersen on October 23,1992. The court consolidated the derivative and class actions on May 15, 1990.

Not surprisingly, in December 1992, the parties notified the court that they had reached settlements for the class and derivative actions. See J. Cooper Alexander, Do the Merits Matter ?: A Study of Settlements in Securities Class Actions, 43 StanLRev 497, 526 (1991) (“Indeed, no securities class action has [ever] gone through trial in the Northern District of California, where many such cases are filed”). 2 At a status conference on February 1, 1993, the parties outlined the terms of the proposed settlements and the court scheduled a settlement approval hearing for April 22, 1993. The salient terms of the proposed settlements are as follows:

—In exchange for dismissal of the class claims and entry of final judgment, Oracle will pay $23.25 million (in five installments to run through April 15, 1994) and Arthur Andersen will pay $1.75 million to a class settlement fund.

—Up to $200,000 of the class settlement fund may be used by class counsel to pay various expenses incurred in notifying class members and administering distribution of the fund. 3

—Up to $4.8 million of the class settlement fund may be used to pay class counsel’s fees, and $825,000 for its expenses, in accordance with the fee plans previously approved by the court.

—The derivative plaintiffs agree to dismiss their claims against the Oracle defendants and any prospective claims against Arthur Andersen without receiving any concrete, present consideration. The derivative settlement agreement notes that dropping the derivative action will reduce Oracle’s future litigation expenses and that, as a result of the prosecution of the derivative action, Oracle has made certain changes in its insider trading and revenue recognition policies.

—Oracle will pay up to $750,000 to derivative counsel for fees and costs expended in prosecuting the derivative action.

—The derivative plaintiffs agree to release the individual defendants from liability for settling the class or derivative actions.

—The class settlement agreement and the derivative settlement agreement are contingent upon one another. That is, the class settlement agreement and the derivative settlement agreement become effective only after the court has entered final judgment dismissing both the class and derivative actions.

It is this last provision which precludes court approval of either of the settlements. Although the class settlement alone appears to be reasonable, the court is not in a position to sanction the derivative settlement. Because of the interrelation of the settlements, however, the court cannot approve either one. This order will assess the class and derivative settlements in turn.

II

A

It is well known that counsel for the class who decide to settle a class action may not necessarily have the interests of their putative clients, the class members, at heart. See generally J. Macey and G. Miller, The Plaintiffs’ Attorney’s Role in Class Action and Derivative Litigation: Economic Analysis and Recommendations for Reform, 58 UChiLRev 1 (1991). As in all lawsuits, the interests of lawyers and their clients conflict; plaintiffs’ lawyers have an interest in maximizing their own fees, while their clients hope to minimize fees and maximize the recovery. In a normal lawsuit, the court rarely gets involved in this conflict since the client presumably has the final say on all major steps in the litigation. In class actions, however, class counsel run the litigation with little or, more realistically, no input from their clients since the class members generally have relatively small individual claims which give them insufficient incentive to supervise their lawyers. When the prospect of settlement arises, the unique dynamics of the lawyer/elient relationship in class actions raise particular problems: an attractive attorney fee provision in the settlement agreement may induce class counsel to settle regardless of the likelihood that further pursuit of the litigation might substantially increase the total class recovery. In recognition of this problem, FRCP 23(e) provides that “A class action shall not be dismissed or compromised without the approval of the court * *

In order to protect the interests of the class members, before approving a settlement proposal, the court must determine that the proposal is “fair, reasonable and adequate.” Galdi Securities Corp. v. Propp, 87 F.R.D. 6, 10 (S.D.N.Y.1979). Courts have identified a number of factors which are relevant in making this assessment: (1) the complexity, expense and likely duration of the litigation, (2) the reaction of the class to the settlement, (3) the stage of the proceedings and the amount of discovery completed, (4) the risks of establishing liability, (5) the risks of establishing damages, (6) the risks of maintaining the class action throughout the trial, (7) the ability of the defendants to withstand a greater judgment, (8) the range of the reasonableness of the settlement in light of the best possible recovery, (9) the range of the reasonableness of the settlement fund to a possible recovery in light of all the attendant risks of litigation. Girsh v. Jepson, 521 F.2d 153 (3d Cir.1975); City of Detroit v. Grinnell Corporation, 495 F.2d 448, 463 (2d Cir.1974).

Certain of these factors are quite easy to assess here. First of all, the court is aware of only one objection to the proposed settlement. The paucity of objections is of little moment, however, because agency costs often discourage meaningful objection in securities class actions. See J. Macey and G. Miller, 58 UChiLRev at 19. More importantly, it is clear that this litigation is complex and that its continuation would be time-consuming and expensive. On the other hand, it is unlikely that the magnitude of the proposed settlement approaches Oracle’s maximum ability to pay. See Wall St J, Mar 24, 1993, at B4 (“Oracle Systems Corp says third quarter profit rose 74% to $29.2 million or 20 cents a share from $16.8 million or 12 cents a share a year ago despite a $24 million charge to settle certain shareholder lawsuits; revenue rose 28% to $370.1 million from $289.6 million”). The court’s inquiry must focus, then, on the reasonableness of the settlement in light of the likelihood of reaching a larger or smaller recovery should the litigation proceed further. See, e.g., Barkan v. Amsted Industries, Inc., 567 A.2d 1279, 1285 (Del. 1989).

B

Plaintiffs’ theory of the class action is that during the period July 11, 1989, through September 26, 1990 (“the Class Period”) 4 , Oracle and the individual defendants intentionally perpetrated a series of frauds which violated §§ 10(b) and 20(a) of the Securities Exchange Act of 1934. Generally, plaintiffs allege that the defendants fraudulently altered the financial picture of Oracle by improperly recognizing revenues for contingent receivables and by giving unreasonably optimistic projections of future results. These practices, plaintiffs allege, had the effect of inflating the price of Oracle stock throughout the Class Period. 5 Plaintiffs do not allege that they relied directly on these alleged misrepresentations, but instead premise liability on the “fraud-on-the-market theory.” See Basic, Inc v. Levinson, 485 U.S. 224, 108 S.Ct. 978, 99 L.Ed.2d 194 (1988). Plaintiffs allege that Arthur Andersen contributed to the stock price inflation by issuing materially false audit reports for Oracle and by helping Oracle to issue its own materially false reports.

Plaintiffs’ attacks on Oracle’s revenue recognition policy focus on its practice of immediately recognizing revenue on licensing contracts even when payment was due up to one year in the future. The policy also allowed Oracle to recognize revenue from contracts which were signed but not received at Oracle headquarters before the end of a quarter, so long as headquarters received the contract within four business days after the close of the quarter. According to plaintiffs, Oracle’s revenue recognition policy facilitated defendants’ manipulation of Oracle’s financial reports. Moreover, plaintiffs’ accounting expert, Robert Berliner, claims that Oracle’s policy for recognition of long term licensing contracts violated Generally Accepted Accounting Principles (“GAAP”) and that investors were unaware of Oracle’s accounting practices.

On July 11, 1989, Oracle announced revenue and income growth for the fourth quarter of its 1989 fiscal year (“4Q 1989”), which ended May 30, 1989. On July 12, 1989, Oracle held a teleconference with financial analysts. Plaintiffs allege that the information Oracle released on these two dates understated 4Q 1989 revenue by $21.5 million, which enabled defendants to overstate revenue projections for IQ 1990 and artificially inflated the value of Oracle stock. According to plaintiffs’ accounting expert, the stock was also inflated because Oracle had not disclosed to investors its aggressive revenue recognition practices for long term contracts. Oracle continued this alleged fraud on September 26, 1989, when it announced its IQ 1990 earnings without disclosing that these figures were based in part on revenues that should have been recognized earlier and on contracts where payment was not actually due until much later. Plaintiffs allege that this inflation of IQ 1990 revenue painted an unduly rosy portrait of Oracle’s financial state, and enabled Oracle to inflate revenue reports for the remainder of the fiscal year.

Oracle allegedly perpetrated a separate fraud on January 2, 1990, when it issued its 2Q 1990 financial report. Plaintiffs allege that the 2Q 1990 report overstated net income by 30 per cent because the announced revenues improperly included approximately $15 million in revenue from twelve contracts, which were either signed after the quarter ended or contained contingencies which should have precluded immediate recognition.

Oracle’s alleged series of frauds began to unravel on March 27, 1990, when Oracle announced that its 3Q 1990 revenues were substantially below what it had earlier projected and what analysts were expecting. According to plaintiffs, the fraud nonetheless continued, because on the same day Oracle announced an optimistic forecast of future revenues and earnings. Oracle made some disclosure of its true financial state on August 27, 1990, when it finally acknowledged that the original January 2, 1990, report for 2Q 1990 overstated revenues for that quarter by $15 million. According to plaintiffs, Oracle did not fully disclose all of its frauds until September 26, 1990, when Oracle felt compelled to restate its September 6, 1990, report of IQ 1991 revenue, thus revealing to analysts the extent of Oracle’s revenue recognition problems.

C

Plaintiffs profess to calculate their damages according to the theory outlined by Judge Sneed in Green v. Occidental Petroleum Corp., 541 F.2d 1335, 1341-46 (9th Cir. 1976) (concurring opinion). Class members who purchased Oracle common stock during the Class Period would be entitled to the lesser of (a) the mean inflation per share caused by the misrepresentations multiplied by the number of shares purchased, or (b) the amount paid for such Oracle shares less the amount realized from any sale of such shares on or before September 26, 1990, when Oracle disclosed the last of the alleged frauds. Plaintiffs base their calculation of the mean inflation per share on the declaration of their damages expert, John Hammerslough. He estimates total damages during the Class Period of $275.2 million.

The record before the court suggests that this damage estimate is highly suspect. The methodology used by Mr. Hammerslough does not conform to that suggested by Judge Sneed or accepted by the Ninth Circuit. See Green, 541 F.2d at 1342 (requiring elimination of the portion of a price decline if it is “the result of market forces unrelated to the wrong”).

The generally accepted measure of damages in securities litigation is the so-called out-of-pocket measure, which looks to the plaintiffs loss and hence the elements of damages noted above. See, e.g., Arrington v. Merrill, Lynch, Pierce, Fenner & Smith, Inc., 651 F.2d 615, 621 (9th Cir.1981). Judge Sneed’s concurring opinion in Green describes this as the difference between a “price line” reflecting the effect of the fraudulent information put into the market and the “value line” which traces what the prices would have been in the absence of fraud. See B. Cornell and R. Gregory Morgan, Using Finance Theory to Measure Damages in Fraud on the Market Cases, 37 UCLA L Rev 883, 887 fig 1 (1990).

In generating the “value line,” Mr. Hammerslough appears to have taken the effect which defendants’ misstatements would have had on Oracle’s reported earnings and for most, but inexplicably not for all, periods multiplied that change by the price-earnings ratio then prevailing for the stock. Yet under the efficient capital market hypothesis endorsed by the plurality in Basic, a security’s value does not fluctuate with reported earnings, but varies instead with the discounted value of future cash flows which are expected to accrue to the security. R. Brealey and S. Myers, Principles of Corporate Finance (4th ed 1991) at 47-67. For a variety of reasons, these may not reflect reported earnings. Id. at 58 (“Be careful not to equate firm performance with the growth in earnings per share”). Underscoring doubts about Mr. Hammerslough’s approach in that period for which Mr. Hammerslough did not use the price-earnings method, use of that method consistent with its use in other periods would have reduced the per share damage estimate almost three dollars and aggregate damages by tens of millions. Ross Suppl Decl, May 21, 1993, at 6-7.

Mr. Hammerslough’s analysis fails to distinguish between the fraud-related and non-fraud related influences on the stock’s price behavior. Compare Wool v. Tandem Computers, Inc, 818 F.2d 1433, 1437 (9th Cir. 1987) (“the increment of artificial inflation caused by misrepresentations may fluctuate ‘as a result of market forces operating on the misrepresentations,’ ” quoting Green, 541 F.2d at 1345). Use of an event study or similar analysis is necessary more accurately to isolate the influences of information specific to Oracle which defendants allegedly have distorted. Note, Estimating Aggregate Damages in Class Action Litigation under Rule 10b-5 for Purposes of Settlement, 59 FordhamLRev 811, 822 (1991).

As a result of his failure to employ such a study, the results reached by Mr. Hammers-lough cannot be evaluated by standard measures of statistical significance. Hence, the reliability of the magnitude and direction of his value estimates are incapable of verification. The result is that defendants’ damages expert, Daniel R. Fischel, was able to point out numerous instances in which Mr. Hammerslough’s estimates reflected, not the influence of fraud, but purely random occurrences.

Mr. Hammerslough’s damage estimates are further drawn into question because he appears inadequately to have accounted for the shares actually affected by the misstatements which plaintiffs claim. A fraud on the market affects only shares traded during the Class Period. Because trading records do not generally track individual shares, it is necessary for damage experts in cases of this kind to estimate the number of shares damaged. In calculating aggregate damages, Mr. Hammerslough used a so-called “proportional decay” model to estimate the number of shares traded during the Class Period for which the class may recover damages. This model appears to assume that all investors are equally likely to trade, so that a “proportional” number of shares are assumed to come from shareholders who are long-term holders and from those who are “in-and-out” traders. Yet a share traded may have a much greater than proportional probability of being re-traded during the Class Period due to the disproportionate influence on trading of short-term traders, arbitrageurs, and similar market participants. Failure to weight the likelihood of trading to reflect the characteristics of trading peculiar to Oracle would likely result in a serious overestimation of aggregate damages. 59 Fordham LRev at 834-840. Unfortunately, defendants failed to offer a substantial alternative analysis to enable the court to weigh the seriousness of this problem.

While settlement approval does not require aggregate damages to be determined with mathematical precision, the court has numerous questions about the methodology adopted by plaintiffs’ damages expert. These questions need not be answered at the moment, however, because of other, and perhaps more serious, difficulties with plaintiffs’ ease.

D

In order to prevail on their claims, plaintiffs must, of course, prove that defendants acted with scienter. Ernst & Ernst v. Hochfelder, 425 U.S. 185, 208-210, 96 S.Ct. 1375, 1388-89, 47 L.Ed.2d 668 (1976). The parties’ submissions reveal little evidentiary support for plaintiffs’ allegations about how and why defendants inflated the price of Oracle stock. Though there is evidence which suggests that defendant Jeffrey Walker, Oracle’s chief financial officer, was concerned about Oracle’s prospects for IQ 1990 and that he might have wanted to shift contract revenue from 4Q 1989 to IQ 1990, there is no evidence that Oracle’s 4Q 1989 report on July 11, 1989, actually understated 4Q 1989 revenues. For example, though Oracle had hoped to finalize a large contract with Bechtel Software, Inc, before May 31, 1989 (the end of 4Q 1989), Oracle properly chose not to recognize revenue from the Bechtel contract until IQ 1990 because the contract was not signed until June 1990. In fact, plaintiffs have presented no evidence to the court that Oracle improperly moved any contracts from 4Q 1989 to IQ 1990.

Plaintiffs’ allegations concerning Oracle’s premature recognition of revenue in 2Q 1990 are similarly unsubstantiated. According to plaintiffs, one middle manager from Oracle claims that defendant Walker instructed him “to manage the P & L results as a first priority in order to support a planned securities offering.” Plaintiffs’ Memorandum in Support of the Settlement at 22. The offering, however, never took place, and there is no concrete evidence to show that defendants knew about any of the alleged revenue recognition problems. Consequently, even if Oracle did improperly recognize some contracts in 2Q 1990 as a result of employee mistakes or misconduct, plaintiffs would have great difficulty proving that the defendants themselves acted with scienter.

Plaintiffs’ general allegations about Oracle’s and Arthur Andersen’s accounting practices also do not seem to hold much water. Arthur Andersen’s expert, Joseph Lhotka, effectively rebuts Mr. Berliner’s contention that Oracle’s revenue recognition policy violated GAAP. Both accounting experts agree that Oracle and Arthur Andersen should have been guided by Accounting Principles Statement No. 4, which provides that revenue is generally recognized when “(1) the earnings process is complete or virtually complete, and (2) an exchange has taken place.” As Mr. Lhotka points out, Oracle complied with this principle because it recognized revenue only after a customer committed to purchase software that Oracle had already developed. When a one-year licensing agreement was signed, the earnings process was virtually complete because Oracle had the software on hand; an exchange took place because the customer agreed to pay the licensing fee in exchange for Oracle’s commitment to supply the software. Though defendants concede that Oracle employees did improperly recognize some contracts where software development was not com píete or a client’s commitment to pay was not legally binding, these instances are few and hardly reflect an intent to manipulate revenue reports.

Mr. Berliner also contends that Oracle and Arthur Andersen should have disclosed the magnitude of Oracle’s receivables on long term licensing contracts, but Mr. Berliner cannot point to any accounting principle or law which mandates such disclosure. In the absence of such a requirement, Oracle’s failure to disclose its receivables cannot be a basis for liability.

Finally, the evidence shows that Oracle did not hide its revenue recognition practices from the market. Oracle’s 1989 10-K report (filed with the Securities and Exchange Commission on August 25, 1989) showed that it recognized revenue from the licensing of its software

either upon shipment of the product or at the time such agreements are effective (which in most instances is the date of the agreement) if the customer is creditworthy and the terms of the agreement are such that the amounts due within one year are noncancellable. The Company recognizes revenue at such time as it has substantially performed all of its contractual obligations.

Defendants have also submitted copies of investment analysts’ reports which show that the investment community was aware of Oracle’s revenue recognition practices even before the Class Period. Thus, even if plaintiffs could show that Oracle’s revenue recognition practices violated GAAP, plaintiffs would have difficulty proving that they suffered any injury from these practices. See In re Apple Computer Securities Litigation, 886 F.2d 1109, 1115 (9th Cir.1989) (“in a fraud on the market case, the defendant’s failure to disclose material information may be excused where that information has been made credibly available to the market by other sources”).

As for the claims against Arthur Andersen, plaintiffs’ evidence, at best, shows that the firm may have been lax in its auditing of Oracle. For instance, perhaps Arthur Andersen should not have permitted Oracle to recognize the $15 million in contracts in 2Q 1990, or should have forced Oracle to disclose earlier its August 27, 1990, financial restatement. There is no proof, however, that Arthur Andersen deliberately or recklessly violated any duty it owed to the plaintiffs. In the absence of such proof, it is unlikely that plaintiffs could sustain their claims against Arthur Andersen. Ernst & Ernst v. Hochfelder, 425 U.S. 185, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976). See also DiLeo v. Ernst & Young, 901 F.2d 624 (7th Cir.1990).

Given the apparent dearth of evidence to support plaintiffs’ claims, class counsel’s decision to settle cannot be said to be unreasonable. Although Mr. Hammerslough opined that damages could exceed $200 million, the parties have presented virtually no evidence to support the liability theories upon which his estimate is based and plaintiffs’ method of calculating damages appears wildly unreliable. Of course, a basic problem in any class action settlement approval process is that both sides are trying to support the settlement. Consequently, courts are frequently faced with the situation, such as this, in which one or both sides have little incentive to put forth their best case. Class counsel’s memoranda in support of the settlement, for example, are replete with concessions about the weakness of their case. Given that the lawyers have taken such a non-adversarial posture, it is difficult for the court to perceive the true merits of each side’s potential arguments. As there is no manifest indication that the parties have not dealt with the court in good faith, however, the court has no choice but to evaluate the evidence as it has actually been presented. Given the record before the court, the court can only find that the settlement is a reasonable, if not generous, one.

Ill

The derivative settlement is much more problematic. This stems in large measure from the inherent conflict between the interests supposedly protected by the class action and those benefitted by the derivative action. Beneficiaries of the former are persons who bought Oracle stock or call options during the Class Period, while beneficiaries of the latter are the current shareholders of Oracle. Because many of those who bought Oracle stock or options during the Class Period may very well have sold some or all of their Oracle shares, the two groups are by no means identical and their interests not coextensive. Inasmuch as almost all of the class settlement is to be funded by Oracle, it is the current shareholders who are paying to settle the class action; most of them will likely receive none of the proceeds of the class settlement and those who do will indirectly pay far more than they receive because of the legal fees and litigation and administrative expenses of the class action. In short, settlement of the class action results in a redistribution of wealth from Oracle’s current shareholders to certain buyers of Oracle stock and options and, most especially, the lawyers for both the plaintiffs and defendants in the class action. In theory, settlement of the derivative suit should benefit the current shareholders. It is necessary, therefore, to inquire whether the current shareholders benefit from these interrelated settlements.

The sole benefit which the derivative settlement confers upon Oracle’s current shareholders is a termination of the expenses associated with these lawsuits. As noted above, the settlement itself also mentions various administrative changes which the derivative litigation has supposedly spurred Oracle to undertake:

Oracle, the Derivative Plaintiffs, and Derivative Counsel agree that the Derivative Actions have conferred a real and substantial benefit upon Oracle. Since March 1990, when the Derivative Actions were first filed, Oracle has implemented important enhancements to its internal policies, systems, and controls in part because of the prosecution of the Derivative Actions, which enhancements address specific issues raised in the Derivative Actions. Among the enhancements are improvements to Oracle’s insider trading policy and improvements to Oracle’s revenue recognition policies and practices.

Proposed Derivative Settlement Agreement, § 5.1. At the settlement approval hearing, the court commented that this description, standing alone as it did, provided no explanation of the nature of these “improvements” and no basis to determine whether these changes conferred any benefit upon Oracle. Accordingly, in their supplemental memorandum in support of the settlement, derivative plaintiffs’ counsel provided the court with evidence that in 1991, Oracle amended its insider trading policy to discourage trading in Oracle stock by Oracle employees’ family members, as well as by employees themselves. Furthermore, on more than one occasion since the derivative litigation began, Oracle has altered its revenue recognition policies. Principal changes that Oracle has adopted include the following rules which apparently did not exist previously: (1) recognition is permittéd only if contracts are received at Oracle headquarters before a quarter ends, rather than within four days of the end of a quarter; (2) shipment is required before revenue from long term software licenses can be recognized; and (3) contracts may only be recognized if they are free of contingencies or significant uncertainties. Derivative plaintiffs’ counsel also retained an investment banker, Edward Elliott, who opined that, while the benefit created by the change in the insider trading policy is negligible, the accounting changes increased investor confidence in Oracle and conferred on Oracle and its shareholders an economic benefit in terms of increased share price which “vastly exceeds” the $750,000 fee request submitted by derivative plaintiffs’ counsel.

Unfortunately, the record contains no evidence to support derivative plaintiffs’ counsel’s argument, and Mr. Elliott’s assumption, that their lawsuit, rather than some other stimulus, actually caused Oracle to make these changes. In fact, the investment analysts’ reports submitted by the parties show that the need to improve investor confidence in Oracle was itself a primary motivator behind Oracle’s decision to change its revenue recognition policies. Fear of further class actions, such as the potentially costly one involved here, must also have provided Oracle with good reason to change its business practices.

Notably, the corporation itself has not offered an independent analysis of these issues. Consequently, the court still has only a murky understanding of the benefit which the derivative litigation has conferred upon Oracle. It is clear, however, that, at the same time that it denies Oracle any financial award, the settlement calls for Oracle to pay derivative counsel up to $750,000 and releases the individual defendants from any claims it might have against them. The derivative settlement thus confers a substantial benefit on the individual defendants and derivative plaintiffs’ counsel. As with the class settlement, the brunt of the derivative settlement falls upon Oracle and indirectly, therefore, on its current shareholders.

The parties argue that the settlement is reasonable because derivative plaintiffs’ claims are virtually meritless. 6 The record before the court supports the parties’ contentions, as it contains little evidence to support plaintiffs’ claims. Like those of the class plaintiffs, the derivative plaintiffs’ mismanagement claims against the individual defendants seem dubious. The parties have also identified a number of faults with derivative plaintiffs’ insider trading claims. First, while some of the individual defendants did sell their stock during the Class Period, the foregoing assessment of the class settlement shows that it is unlikely that Oracle’s shares were in fact fraudulently inflated during this time. Second, none of the inside sales occurred when Oracle stock neared a peak price. Third, none of the sales amounted to more than a small fraction of each inside seller’s total Oracle holdings. Based on this presentation of the evidence, the derivative plaintiffs appear extremely unlikely to prevail on their claims against the individual defendants, and the proposed settlement and dismissal of the derivative action would seem appropriate.

The court’s analysis of the proposed settlement cannot end here, though, for the court is well aware that it has not received an unbiased portrayal of the merits of the derivative claims. The presentation of plaintiffs’ counsel is biased because the settlement offers them a substantial fee, regardless of the benefit conferred on the corporation and its shareholders. See R. Clark, Corporate Lavo, § 15.7 at 657-659 (1986) (explaining that because settlement assures them a fee, derivative plaintiffs’ attorneys generally have a strong incentive to settle even where the settlement is not in the corporation’s best interest); Note, Independent Representation for Corporate Defendants in Derivative Suits, 74 Yale LJ 524, 532 (1965) (“at settlement the probability that [derivative] plaintiffs counsel will be interested only in obtaining a lucrative fee may undermine his capacity to represent adequately the corporation’s shareholders”).

The presentation on behalf of the individual defendants is similarly biased in favor of settlement, because the settlement absolves them of liability without requiring them to part with any return consideration. See W. Haudek, The Settlement and Dismissal of Stockholders’ Actions — Part II: The Settlement, 23 Sw LJ 765, 770 (1969) (in a stockholder’s derivative action, “[t]he directors will * * *, as a rule, be hostile to the suit and cannot be effective champions of the corporate interests in the settlement talks”). The pressure on the individual defendants to settle is particularly strong because settlement ensures that the corporation may indemnify their litigation expenses, whereas a final adjudication of malfeasance would likely preclude indemnification. 8 Del.C. § 145(b); see J. Bishop, Sitting Ducks and Decoy Ducks: New Trends in the Indemnification of Corporate Directors and Officers, 77 Yale LJ 1078, 1082-1088 (1968).

Under FRCP 23.1, the court has a duty to determine whether the settlement reasonably protects the interests of the corporation and its shareholders. Mathes v. Roberts, 85 F.R.D. 710, 713 (S.D.N.Y.1980). Because of the absence of a disinterested presentation on behalf of the corporation, it is impossible for the court to perform this duty. As the record now stands, the court has before it only presentations from parties which have a vested interest in settlement.

In fact, when the settlement was first proposed to the court, the parties did not even mention Oracle’s participation in the settlement approval process. After an inquiry from the court on June 16, 1993, Oracle’s general counsel and the law firm of Morrison & Foerster, who together represent Oracle and the four insider defendants, 7 provided copies of a December 22, 1992, written consent to the derivative settlement. The four members of Oracle’s seven member board who are not named defendants in the derivative action apparently approved the settlement. In response to the June 16 order, Morrison & Foerster and the general counsel also informed the court that Oracle’s directors had decided to appoint the four non-defendant directors to a special settlement committee to decide whether to approve the settlements. On June 24, 1993, Morrison & Foerster and the general counsel forwarded to the court copies of a new written consent to the derivative settlement, dated June 23, 1993, and signed by the same four non-defendant directors on behalf of the special settlement committee.

In essence, then, Oracle’s board terminated the derivative litigation in accordance with procedures similar to those discussed by the Delaware Supreme Court in Zapata Corp. v. Maldonado, 430 A.2d 779 (Del.1981), by establishing a committee of purportedly disinterested directors to ratify the settlement. In Zapata, the court sanctioned the assignment of disinterested board members to a special litigation committee, which has the power “to dismiss derivative litigation that is believed to be detrimental to the corporation’s best interest.” Id. at 786. Unlike Zapata, Oracle’s special committee has not presented the court with a motion to dismiss, and the proposed derivative settlement is not an attempt by the board to deprive derivative plaintiffs of an action that they believe is worth carrying forward. But the proposed settlement does take the derivative claims away from Oracle, which is the real party in interest. As the Zapata court recognized, a special litigation committee’s motion to dismiss is “analogous] to a settlement in that there is a request to terminate litigation without a judicial determination of the merits.” 430 A.2d at 787. See also Kaplan v. Wyatt, 484 A.2d 501, 506-507 (Del.Ch.1984), aff'd, 499 A.2d 1184 (Del.1985) (a special litigation committee’s motion to dismiss “is a hybrid one, derived by analogy to a motion to dismiss a derivative suit based upon a voluntary settlement”). Oracle’s special settlement committee, like the special litigation committee in Zapata, has decided to withdraw Oracle’s claims against the individual defendants without receiving any return consideration. Although derivative plaintiffs’ counsel have assented to this settlement, this fact does not materially distinguish the settlement from a Zapata-style termination because, as noted above, the fee offered to derivative plaintiffs’ counsel renders their support of the settlement virtually meaningless for analytical purposes.

Zapata bears particular relevance because Oracle is a Delaware corporation. In Burks v. Lasker, 441 U.S. 471, 486, 99 S.Ct. 1831, 1841, 60 L.Ed.2d 404 (1979), the Supreme Court held that courts, when deciding whether to allow a corporation’s independent directors to discontinue a derivative action, must apply the law of the state of incorporation. See also Johnson v. Hui, 811 F.Supp. 479, 483 (N.D.Cal.1991). As in Zapata, then, this court must assess whether the special committee’s decision to terminate the derivative action adequately protects the interests of the corporation and its shareholders.

The Zapata court established a two-part test for courts to determine whether to approve the termination of a derivative suits by a special committee:

First, the Court should inquire into the independence and good faith of the committee and the bases supporting its conclusions. Limited discovery may be ordered to facilitate such inquiries. The corporation should have the burden of proving independence, good faith and a reasonable investigation, rather than presuming independence, good faith and reasonableness. If the Court determines either that the committee is not independent or has not shown reasonable bases for its conclusions, or if the Court is not satisfied for other reasons relating to the process, including but not limited to the good faith of the committee, the Court shall deny the corporation’s motion. If, however, the Court is satisfied under Rule 56 standards that the committee was independent and showed reasonable bases for good faith findings and recommendations, the Court may proceed, in its discretion to the second step. The second step provides, we believe, the essential key in striking the balance between legitimate corporate claims as expressed in a derivative stockholder suit and a corporation’s best interests as expressed by an independent investigating committee. The Court should determine, applying its own independent business judgment, whether the motion should be granted.

430 A.2d at 788-789. The proposed derivative settlement fails at the first step of the Zapata test, because the record does not show that the directors who approved the settlement acted with the requisite independence and good faith.

Morrison & Foerster and Oracle’s general counsel argue together that the disinterested directors’ approval of the settlements meets the business judgment rule and adequately protects the interests of the corporation. As counsel point out, courts generally respect the business judgment of a disinterested majority of directors. See, e.g., Republic National Life Insurance Co v. Beasley, 73 F.R.D. 658, 668 (S.D.N.Y.1977) (“[t]he business judgment of an independent, disinterested board of directors that settlement * * * of a derivative action is in the best interests of the company and its shareholders is respected by the Courts”). Still, the “requirement of director independence inhers [sic] in the conception and rationale of the business judgment rule.” Aronson v. Lewis, 473 A.2d 805, 816 (Del.1984). Independence is not established by the fact that directors are not defendants in the derivative action. Thus, in Zapata, the court noted that:

[Notwithstanding our conviction that Delaware law entrusts the corporate power to a properly authorized committee, we must be mindful that directors are passing judgment on fellow directors in the same corporation and fellow directors, in this instance, who designated them to serve both as directors and committee members. The question naturally arises whether a “there but for the grace of God go I” empathy might not play a role.

430 A.2d at 787; see also G. Dent, The Power of Directors to Terminate Shareholder Litigation: The Death of the Derivative Suit?, 75 NW U L Rev 96,111-114 (1980) (outlining how “[t]he independence of [non-defendant] directors * * * may be compromised in many ways”).

Since “independent” directors are often beholden to the defendant directors who appointed them, the retention of independent counsel by these directors provides one of the few safeguards to ensure the legitimacy of their acts and to aid the court in assessing the reasonableness of a derivative settlement or termination. See Messing v. FDI, Inc., 439 F.Supp. 776, 782 n. 8 (D.N.J.1977). There is no indication, however, that any of the supposedly disinterested directors on Oracle’s settlement committee ever conferred with independent counsel to discuss whether the proposed derivative settlement was in Oracle’s best interests. Instead, the written consents indicate that these directors were advised by Oracle’s own general counsel.

The role played by Oracle’s general counsel in this litigation is distressingly ambiguous. Early submissions to the court, including the joint answer to the derivative complaint filed by Oracle and three of the insider defendants, indicated that the general counsel represented the insider defendants while Morrison & Foerster represented the corporation. More recently, all of the memoranda submitted on behalf of the corporation and the insider defendants in connection with the class and derivative settlements have listed Morrison & Foerster and the general counsel as jointly representing the corporation and the insider defendants. On July 19,1993, the court issued an order querying the effect of the apparent sharing of counsel among the corporation and the insider defendants. In a marked change of tune, Morrison & Foerster and the general counsel jointly responded that the general counsel has represented only the corporation in the class and derivative actions. The response explained further that Morrison & Foerster has represented the insider defendants in both actions and Oracle in the class action. Morrison & Foerster and the general counsel aver that the appearances on the answer to the derivative complaint were inadvertently reversed and that the memoranda submitted in connection with the settlements listed Morrison & Foerster and the general counsel together merely “[i]n order to shorten lawyer listings.” Thus, counsel argue, Oracle has been represented by independent counsel- — -its own general counsel — throughout the derivative litigation.

Though the court’s research revealed no cases which deal specifically with this circumstance, there is a substantial body of authority proscribing dual representation of corporate and individual defendants in a derivative action. See, e.g., Cannon v. U.S. Acoustics Corp., 398 F.Supp. 209 (N.D.Ill.1975), aff'd in part, rev’d in part, 532 F.2d 1118 (7th Cir. 1976); Lewis v. Shaffer Stores Co., 218 F.Supp. 238 (S.D.N.Y.1963); Messing v. FDI, Inc., 439 F.Supp. 776 (D.N.J.1977); Murphy v. Washington American League Base Ball Club, Inc., 324 F.2d 394 (D.C.Cir. 1963); Garlen v. Green Mansions, Inc., 9 A.D.2d 760, 193 N.Y.S.2d 116 (1959); Dukas v. Davis Aircraft Products Co., Inc., 129 Misc.2d 846, 494 N.Y.S.2d 632 (Sup.1985); Essential Enterprises Corp. v. Dorsey Corp., 40 Del.Ch. 343, 182 A.2d 647 (1962). 8 The derivative action exists, after all, to benefit the corporation, not derivative plaintiffs’ counsel or the individual defendants. See Cannon v. U.S. Acoustics Corp., 398 F.Supp. at 213. Dual representation is impermissible, particularly at the settlement stage, because:

If the same counsel represents both the corporation and the director and officer defendants, the interests of the corporation are likely to receive insufficient protection. An increased recovery for the corporation is wholly incompatible with the goal of limiting the defendants’ liability. Defendants’ counsel is thus placed in an untenable position, and more often than not he will succumb to the pressure to approve any settlement between the shareholder and his individual clients.

Note, Independent Representation for Corporate Defendants in Derivative Suits, 74 Yale LJ 524, 532 (1965). See also New York City Bar Association Committee on Professional and Judicial Ethics, Opinion 812, 15 Record of NYCBA 80 (1960).

Contrary to the assertions of defense counsel, representation of the corporation’s interests by in-house counsel does not ameliorate this conflict, for in-house attorneys are inevitably subservient to the interests of the defendant directors and officers whom they serve. See G. Dent, The Power of Directors to Terminate Shareholder Litigation: The Death of the Derivative Suit?, 75 NW U L Rev 96, 117 n 113 (1980) (“the corporation’s house counsel or regular outside counsel * * * are too financially dependent on the board to be disinterested”); Note, Developments in the Law — Conflicts of Interest in the Legal Profession, 94 Harv L Rev 1244, 1342 (1981) (“while the inhouse attorney is nominally the representative of the corporation, his personal loyalties will inevitably be to the individual executives who hired him”). Corporate representation by in-house counsel in derivative litigation involving a public corporation also creates an appearance of impropriety which casts a shadow over the entire proceeding, particularly where corporate counsel advocates a settlement that is highly favorable to the individual defendants who are his superiors. See Rowen v. Le Mars Mutual Insurance Co. of Iowa, 230 N.W.2d 905, 915-916 (Iowa 1975); Note, Derivative Actions — Rowen v. Lemars Mutual Insurance Co. — Disqualification of Corporate Counsel and Appointment of Independent Counsel, 2 J Corp L 174, 180 (1976).

The cases on dual representation recognize these problems and require corporations involved in derivative suits to retain counsel with no prior ties to the individual defendants or the corporation. Garlen v. Green Mansions, Inc., 9 A.D.2d 760, 193 N.Y.S.2d 116, 117 (1959) (appearance by a corporation in a derivative suit “must be by independent counsel whose interests will not conflict with those of the individual defendant”); Lewis v. Shaffer Stores Co., 218 F.Supp. 238, 240 (S.D.N.Y.1963) (“it would be wise for the corporation to retain independent counsel, who have had no previous connection with the corporation, to advise it as to the position which it should take in this controversy”); Messing v. FDI, Inc., 439 F.Supp. 776, 782 (D.N.J.1977) (“Independent counsel for the corporation, unshackled by any ties to the directors, would be in the unique position of having only the corporation’s interest at stake”); see also Note, Independent Representation for Corporate Defendants in Derivative Suits, 74 Yale L J 524, 534-535 (1965). Compare Schmidt v. Magnetic Head Corp., 97 A.D.2d 151, 468 N.Y.S.2d 649 (1983) (denying motion to disqualify corporate counsel in dispute over board membership where counsel had no prior connection with corporation or its directors). It is also clear that an inanimate corporate entity, which is run by directors who are themselves defendants in the derivative litigation, cannot effectively waive a conflict of interest as might an individual under applicable professional rules such as Cal R Prof Conduct 3-600(E) and 3-310. Note, Independent Representation for Corporate Defendants in Derivative Suits, 74 Yale LJ 524, 534-535 (1965); New York City Bar Association Committee on Professional and Judicial Ethics, Opinion 812, 15 Record of NYCBA 80 (1960); Note, Disqualification of Corporate Counsel in Derivative Actions: Jacuzzi and the Inadequacy of Dual Representation, 31 Hast LJ 347, 360 (1979).

In the matter before the court, the conflict of interest could not be stronger. The general counsel, of course, is an employee of Oracle. At the same time, the defendants in the derivative action include three of Oracle’s senior executive officers — Lawrence Ellison, Robert Miner, and Gary Kennedy — and three of Oracle’s current directors — Ellison, Miner, and Donald Lucas. It seems indubitable, then, that the general counsel would be reluctant to recommend that the corporation take any position adverse to these men, individual defendants for whom he works on a day-to-day basis and who control his future with the corporation.

The settlement committee’s reliance on the inherently biased advice of in-house counsel makes their approval of the settlement worthless for purposes of analyzing whether the settlement reasonably protects the interests of the corporation and its shareholders. As one court noted in the context of a union derivative action, “the organization is entitled to an evaluation and representation of its institutional interests by independent counsel, unencumbered by potentially conflicting obligations to any defendant officer.” International Brotherhood of Teamsters v. Hoffa, 242 F.Supp. 246 (D.D.C.1965). Here, because Oracle’s special settlement committee lacked independent counsel, the derivative settlement reeks of collusion between derivative plaintiffs’ counsel and the individual defendants, at the expense of the corporation. This is not to say that the terms of the proposed settlement are in fact unreasonable, but the court cannot fairly conclude otherwise based on the suspect presentations it has so far received.

The posture of this settlement stands in stark contrast to Sullivan v. Hammer, ¶ 95,-415 Fed.Sec.L.Rptr. (CCH) 97,061, 1990 WL 114223 (Del.Ch.1990), where the court approved the derivative settlement only after receiving assurances that the special committee charged with evaluating the settlement was advised by independent counsel. See also Johnson v. Hui, 811 F.Supp. 479, 487 (N.D.Cal.1991) (in assessing independence of special litigation committee which terminated derivative litigation, court notes that law firm representing special litigation committee had no prior contacts with corporation or the individual defendants). Unlike the special committee in Hammer, the Oracle special settlement committee acted without the benefit of disinterested legal advice on the merits of the derivative action. Oracle’s decision to settle and dismiss the derivative action simply does not satisfy the independence and good faith requirements of Zapata.

For the foregoing reasons, approval of the derivative settlement is DENIED. Because the proposed class settlement is conditioned on dismissal of the derivative action, approval of the class settlement must also be DENIED.

IV

Before examining the alternatives for resolution of the derivative suit, it should be noted, as class counsel have recently reminded the court, that the class settlement agreement permits the defendants to waive settlement of the derivative settlement as a condition to their final consent to the class settlement. Since the court is satisfied that the class settlement, with the exception of the added $200,000 expense provision, is reasonable, should defendants waive approval of the derivative settlement as a condition to then-consent, the court is prepared to approve the class settlement agreement, minus the added expense provision, forthwith.

Whether or not the defendants choose to follow this course, the corporation must, before proceeding further with the derivative action, retain independent counsel having no prior relationship with the corporation or the individual defendants. Although some courts have gone so far as to appoint corporate counsel in derivative actions, see Rowen v. LeMars Mutual Insurance Co. of Iowa, 230 N.W.2d 905 (Iowa 1975); Niedermeyer v. Niedermeyer, ¶ 94,123 Fed Sec L Rep (CCH) 94,492, 1973 WL 419 (D.Ore.1973), it seems more appropriate here to defer to the independent directors on the selection of corporate counsel. Tydings v. Berk Enterprises, 80 Md.App. 634, 565 A.2d 390 (1989); Messing v. FDI, Inc., 439 F.Supp. 776, 783 (D.N.J. 1977). “Certainly new counsel will recognize their duty to represent solely the interests of the corporate entitfy]. And should difficulties arise, the parties or counsel may apply to the court for additional relief.” Cannon v. U.S. Acoustics Corp., 398 F.Supp. 209, 220 (N.D.Ill.1975).

Once the independent directors have retained independent counsel, the corporation would seem to have three options for resolution of the derivative suit: settlement, termination, or trial. While the court’s role is not, of course, to advise the corporation on litigation strategy, if the investigation of independent counsel confirms derivative plaintiffs’ case to be as weak as the parties have so far presented it, then termination of the derivative suit would seem the most sensible course to take. In any event, under Delaware law, a final decision on these issues depends in large part on the business judgment of the independent directors, after they have persuaded the court that they earned out a good faith investigation into the merits of the derivative action in accordance with Delaware law.

The parties shall appear for a status conference on September 10, 1993, at 2 pm, to discuss these matters, and shall submit brief written reports not later than five days in advance of the conference.

SO ORDERED.

1 . The individual defendants and their positions with Oracle are: Lawrence Ellison, co-founder, president, chief executive officer, director, and chairman of the board as of May 1990; Donald Lucas, director, chairman of the audit committee of the Oracle board, and chairman of the board until May 1990; Jeffrey Walker, senior vice president and chief financial officer until 1991; Gary Kennedy, senior vice president and head of Oracle’s United States marketing organization; and Robert Miner, co-founder, vice president, and director.

2 . Following publication of Professor Alexander's article, In re Apple Computer Securities Litigation was tried in the San Jose division, but the verdict as to the only defendants found liable was set aside under FRCP 50 as "confused” and "internally inconsistent.” 1991 WL 238298, 1991 US Dist LEXIS 15608, CCH Fed Sec L Rep ¶ 96,252 (N.D.Cal.1991).

3 . As the court noted at the settlement hearing, this provision of the class settlement agreement conflicts with the fee plans previously submitted by class counsel and approved by the court. 132 F.R.D. 538 (N.D.Cal.1990). Consequently, this provision must be stricken if the parties continue to pursue a class settlement.

4 . For purposes of the proposed class settlement, the class period actually extends to March 27, 1991. Though plaintiffs contend that Oracle disclosed the last fraud by September 26, 1990, plaintiffs agreed to extend the class period to March 27, 1991, in order to broaden the scope of the liability release created by the settlement. This solely benefits defendants and serves no interest of the class.

5 . Plaintiffs initially alleged that some of the individual defendants engaged in insider trading during the Class Period, but have since conceded that there is almost no evidence to support this charge.

6 . Derivative plaintiffs' counsel also argue that even if the derivative claims have some merit, further prosecution of them would serve only to harm Oracle, since Oracle has agreed (via separate agreements and its corporate by-laws) to indemnify the individual defendants for expenses and liability incurred in the derivative action. Of course, if derivative plaintiffs’ counsel actually believed this argument, then they should never have brought their lawsuit in the first place. Derivative plaintiffs knew of Oracle’s indemnification policy when they filed their lawsuit, and were equally capable of conducting their cost-benefit analysis at that time.

Nevertheless, the indemnity agreements would likely not apply if derivative plaintiffs prevailed in this lawsuit. Neither Delaware law, 8 Del.C. § 145, nor Oracle's indemnity policy permit indemnification of an officer or director who has been adjudged to have acted in a manner which he reasonably believed to be opposed to Oracle’s interests. Since several of plaintiffs’ claims require a finding of intent or scienter, the indemnity agreements would likely have no effect in the event derivative plaintiffs pursued their claims to a favorable verdict.

7 . The fifth individual defendant, outside director Donald Lucas, is represented separately by another law firm.

8 . Morrison & Foerster and Oracle’s general counsel point out that one California court of appeal has permitted dual representation of corporate and individual defendants in a derivative action. Jacuzzi v. Jacuzzi Bros., Inc., 243 Cal. App.2d 1, 52 Cal.Rptr. 147 (1966). Because Oracle is a Delaware corporation, however, under Burks v. Lasker, 441 U.S. 471, 486, 99 S.Ct. 1831, 1841, 60 L.Ed.2d 404 (1979), the court is not bound by Jacuzzi. Moreover, logic and the weight of recent authority are against Jacuzzi. See Note, Disqualification of Corporate Counsel in Derivative Actions: Jacuzzi and the Inadequacy of Dual Representation, 31 Hast LJ 347 (1979).