WALLACE, Circuit Judge:
Ernst & Whinney appeals from a general jury verdict in favor of the plaintiff class (Knapp). Knapp cross-appeals from the district court’s denial of prejudgment interest and the dismissal of various state-law causes of action. The district court had jurisdiction pursuant to 15 U.S.C. § 78aa and 28 U.S.C. § 1331. We have jurisdiction over this timely appeal pursuant to 28 U.S.C. § 1291. We affirm.
I
ATV Systems, Inc. (ATV) was founded by Philip Gomez, Douglas O’Connor, and Frank Gleason in 1981. The company specialized in point-of-sale microprocessing systems designed for retail applications and also dealt in more generalized data processing and communications equipment for office use. The principals had been involved in the high-tech [1434]*1434field and considered themselves well-versed in “turn-around” techniques. They were experienced in taking troubled companies, slashing waste, increasing productivity, improving the bottom line, and turning the companies into profitable enterprises. There were many businesses in need of such redirection in the early 1980s because interest rates were high and credit was tight. ATVs business plan was to acquire bankrupt companies at bargain prices, liquidate unprofitable product lines, and merge profitable lines into its own operations.
Although this mode of operation allowed ATV to acquire research and technology at wholesale prices that would have required years to develop independently, the approach also forced ATV to inherit the management, marketing, production, and service woes that brought the companies down in the first place. In other words, ATV acquired assets that might have become valuable in the future as research and development ripened into product lines. But at the same time, ATV acquired significant current liabilities and start-up costs as operations were restructured and redirected.
Because of this business plan, ATV was a cash-hungry venture and the founders’ personal resources were fast becoming inadequate to meet the company’s needs. The founders first considered taking ATV public in late 1982, but because the company needed immediate financial assistance, ATVs founders sought alternate sources of cash. Help arrived in the form of a $10 million injection of funding in early 1988 from various venture capital investment groups. ATV in return issued convertible debentures.
The debenture holders were not interested in a long-term investment. The game plan was to boost ATVs respectability in the industry to the point where it could be taken public. The debenture holders would then exercise their conversion rights and sell their shares on the open market. If all went well, the venture capitalists would recoup their $10 million plus a tidy profit, ATV would be out from under the burden of the debt and positioned to pursue its aims even more aggressively, and the founders, through their stock holdings, would be rewarded accordingly.
All did not go well. As a practical matter, taking the company public was now no longer an option, but a necessity. A failure to go public would not only trigger an obligation to retire most of the debenture debt, but also would increase that debt by $2 million. In retrospect, ATV was not well-positioned to make a public offering. Nevertheless, ATV retained a prestigious securities law firm, a large accounting firm, and E.F. Hutton as lead underwriter during the spring and summer of 1983.
As originally planned, the initial public offering (IPO) was to be 2.2 million shares with a price range of $15 — $18 per share. As time passed, the number of shares and the price per share declined. When the company filed the preliminary prospectus with the Securities and Exchange Commission in August 1983, it still proposed to issue 2.2 million shares, but the price had dropped to $13 — $15 per share. E.F. Hutton further reduced the price to $10 and the number of shares to only 600,000. By that time, high-tech IPOs had fallen out of favor with the market and ATVs “road shows” had not generated the interest expected. Thus, instead of the $30 million ATV hoped to generate, the gross would only be $6 million. In the end, most of the 600,-000 shares were sold by the debenture holders and after deducting substantial legal and accounting fees and other costs, ATV realized no net cash. In fact, Gomez drafted an internal memorandum indicating that the IPO cost ATV money. Taking the company public accomplished two things: ATV was relieved of some debenture debt, and the stage was set for this action.
The IPO was effective on October 18,1983. Beginning in January 1984, ATV released a series of negative earnings reports and the stock reacted sharply. Far from being a growth company poised for continued expansion, as the prospectus suggested, it became clear to the market that ATV was not doing well. In the quarter in which the company went public, ATV recorded a net loss of almost $2 million. ATV would ultimately post a loss of more than $31 million for the fiscal year ending March 31,1984. The stock was delisted by the National Association of Securities Dealers Automated Quotation [1435]*1435Market System. ATV eventually returned to a position of modest profitability, but closed its doors in 1988.
Investors who had purchased ATV stock at $10 per share and sold it for a quarter of that amount several months later wondered if there was more to the story than unexpected business reversals. This class action followed against Gomez, O’Connor, Gleason, E.F. Hutton, Ernst & Whinney (now Ernst & Young), the debenture holders who exercised their conversion rights, two ATV board members, and several others. Knapp asserted liability on a variety of theories, including federal racketeering and securities laws violations and numerous state-law claims.
All defendants settled pretrial, or were otherwise removed from the ease, except for Gomez (ATV’s chief executive officer and chairman of the board), O’Connor (ATV’s executive vice-president), and Ernst & Whin-ney (ATV’s outside auditor). At the conclusion of the evidence, the trial court dismissed the RICO and pendent state-law claims.
At Knapp’s request, the case was bifurcated into a trial of class-wide issues to be followed, if necessary, by a second-stage proceeding to adjudicate individual issues of reliance and damages. After a four-week trial, the court submitted to the jury only the securities claim, based on section 10(b) of the Securities Exchange Act of 1934, and Rule 10b-5 promulgated thereunder.
At trial, Knapp attempted to show that Ernst & Whinney should have included a “going concern” statement in its audited financial statements. Knapp also presented some evidence suggesting that Ernst & Whinney should have recognized the many problems within ATV and that, given these problems, it should have further investigated and distanced itself from ATV’s unaudited financial statements. The jury returned a general verdict against all defendants and set forth what it found to be the “true value” of ATV stock for each trading day of the class period. In addition, the jury found the proportionate fault of the settling defendants to be zero. Thereafter, the district court denied Ernst & Whinney’s motions for judgment as a matter of law or, in the alternative, for a new trial. The court also denied Knapp’s motions for prejudgment interest.
After the commencement of the second-stage proceedings, Knapp submitted claims for damages from each of the class members. In motions for summary judgment, Ernst & Whinney contested the claims of some of those class members, including four institutional investors. The district court denied Ernst & Whinney’s motion and granted Knapp’s cross-motions for summary judgment on the claims. This opinion addresses only Ernst & Whinney’s arguments. Gomez and O’Connor’s contentions are resolved in a separate, unpublished decision.
II
Prior to trial, Knapp settled with various defendants, receiving a total amount that was 25 percent greater than what the jury ultimately awarded. Ernst & Whinney contends that the damages awarded at trial should be offset by the amount recovered from the settling defendants, effectively reducing the jury’s award to zero. Whether plaintiffs’ recovery against nonsettling defendants must be offset by pretrial recovery from settling defendants is a question of law that we review de novo. See generally United Commercial Ins. Serv. v. Paymaster Corp., 962 F.2d 853, 856 (9th Cir.), cert. denied, 506 U.S. 1022, 113 S.Ct. 660, 121 L.Ed.2d 585 (1992).
In Franklin v. Kaypro Corp., 884 F.2d 1222 (9th Cir.1989) (Kaypro), cert. denied, 498 U.S. 890, 111 S.Ct. 232, 112 L.Ed.2d 192 (1990), we set out the analysis a trial court should use when determining the liability of nonsettling defendants where plaintiffs have already settled with other defendants. Id. at 1231. We explained as -follows:
Nonsettling defendants are [ ] barred from further rights of contribution from the settling defendants. At trial, the jury is asked not only to determine the total dollar damage amount, but also the percentage of culpability of each of the nonsettling defendants as well as that of the settling defendants. Nonsettling defendants as a whole will then be required to pay the percentage of the total amount for which they are responsible. The nonsettling de[1436]*1436fendants will be jointly and severally liable for that percentage, and will continue to have rights of contribution against one another.
Id. (footnote omitted). Pursuant to the rule in Kaypro, the district court instructed the jury to apportion fault among the settling and nonsettling defendants. The jury found the proportionate fault of the settling defendants to be zero percent. Following Kaypro, the district court refused to offset the jury’s award because the jury found the settling defendants to be blameless.
Ernst & Whinney asks us to alter the Kaypro proportionate fault rule, arguing that the rule is inconsistent with 15 U.S.C. § 78bb(a) and that the rule in Kaypro is actually dicta. In the district court, Ernst & Whinney “recognize[d] that [the district court] is bound by” Kaypro. We will not consider Ernst & Whinney’s arguments concerning the applicability of Kaypro, which are being made for the first time on appeal.
Ill
After the first phase of the trial, the jury effectively found that the plaintiff class was entitled to a presumption of reliance. Upon commencement of the second phase, Ernst & Whinney moved for summary judgment on the ground that it had rebutted the presumption of reliance as to certain sophisticated investors. The district court rejected Ernst & Whinney’s argument and instead entered summary judgment for Knapp. We review a summary judgment de novo. First Pacific Bank v. Gilleran, 40 F.3d 1023, 1024 (9th Cir.1994), cert. denied, - U.S. -, 116 S.Ct. 55, 133 L.Ed.2d 19 (1995). We must determine (1) whether there is a genuine issue of material fact when the evidence is viewed in the light most favorable to the nonmoving party, and (2) whether the district court correctly applied the law. Id.
In Basic Inc. v. Levinson, 485 U.S. 224, 247, 108 S.Ct. 978, 991-92, 99 L.Ed.2d 194 (1988), the Court held that in a securities fraud action premised on a fraud on the market, purchasers and sellers of securities in the marketplace are entitled to a presumption of reliance. The presumption of reliance is premised on a belief that, at the very least, “market professionals generally consider most publicly announced material statements about companies, thereby affecting stock market prices.” Id. at 247 n. 24, 108 S.Ct. at 991 n. 24. Because a material misrepresentation or omission is generally embedded in the market price of the security, an investor who purchases or sells the security will have presumably relied on the misrepresentation.
A defendant may rebut the presumption of reliance in one of several ways. A defendant might “show that the ‘market makers’ were privy to the truth ... and [ ] that the market price would not have been affected by [ ] misrepresentations, ... [so] the basis for finding that the fraud had been transmitted through market price would be gone.” Id. at 248, 108 S.Ct. at 992. A defendant might also show that the plaintiff was aware of the misrepresentation and purchased or sold his shares anyway. Or, a defendant might establish that had the plaintiff been aware of the misrepresentation, he still would have engaged in the same market transaction. Id. at 248-49, 108 S.Ct. at 992-93; see also Blackie v. Barrack, 524 F.2d 891, 906 (9th Cir.1975) (Blackie), cert. denied, 429 U.S. 816, 97 S.Ct. 57, 50 L.Ed.2d 75 (1976).
Ernst & Whinney argues that as to four class plaintiffs, it had effectively rebutted the presumption by showing (a) that the investors were sophisticated institutional investors, and (b) that experts testified that sophisticated investors should have been aware of the information that was omitted from Ernst & Whinney’s audits, including a going concern qualification and disclosures concerning ATVs liquidity and the adequacy of collateral for an ATV loan. Ernst & Whinney relies on the following testimony from Knapp’s expert witness:
Sir, I don’t believe the average person reading [the ATV prospectus] will understand the nature of the market and the uncertainties that surround a new offering and volatility of the market, [ ] the enormous potential for downturns in the market of this kind. [ ] [I]t’s only people who have a substantial degree of experience [1437]*1437and interest in the market that will understand that.
Ernst & Whinney essentially argues that because a sophisticated investor would not have been misled by ATVs financial statements, then those investors must have purchased ATVs stock knowing full well what the true value of the stock was.
Ernst & Whinney’s argument proves too much. If sophisticated investors would have read the ATV financial statements in a way that rendered the omissions or misstatements immaterial, then the market price would not have been artificially inflated. Ernst & Whinney is arguing that not a single sophisticated investor could have been misled. Extending Ernst & Whinney’s analysis, these investors’ knowledge of the “truth” would have prevented the price from becoming artificially inflated.
This is not what the jury found, however. In the first phase of the trial, the jury found that the market price for ATV stock was inflated due to defendants’ misstatements or omissions. In so finding, the jury necessarily rejected the testimony of Knapp’s expert, insofar as he was testifying that sophisticated investors would not have been fooled. Ernst & Whinney does not now contend that the jury’s verdict was not supported by substantial evidence, and we thus will not disturb its findings. The district court therefore properly denied Ernst & Whinney’s motion for summary judgment and properly entered summary judgment for Knapp.
IV
Ernst & Whinney argues that the district court’s formulation of the instruction on loss causation was erroneous. Jury instructions must be formulated so that they fairly and adequately cover the issues presented, correctly state the law, and are not misleading. Fikes v. Cleghorn, 47 F.3d 1011, 1013 (9th Cir.1995). When the alleged error is in the formulation of the instructions, the instructions are to be considered as a whole and an abuse of discretion standard is applied to determine if they are misleading or inadequate. Gizoni v. Southwest Marine Inc., 56 F.3d 1138, 1142 n. 5 (9th Cir.), cert. denied, — U.S. -, 116 S.Ct. 381, 133 L.Ed.2d 304 (1995). However, when the argument is that the trial court misstated the elements that must be proved at trial, we review de novo. Id. at 1142.
The parties characterize the nature of the alleged error differently. Ernst & Whinney argues that because the district court misstated the loss-causation element, review is de novo. Knapp argues that because the district court did instruct on loss causation, the question is one of formulation, and should be reviewed for an abuse of discretion. We need not resolve the issue because we arrive at the same result under either standard of review.
Ernst & Whinney makes two related arguments. First, it argues that the court’s loss-causation instruction was erroneous. Second, it argues that the court should have entered a judgment as a matter of law for Ernst & Whinney as to those plaintiffs who bought and sold ATV stock before any corrective statements had been made.
We first address whether the district court’s loss-causation instruction was erroneous. Knapp argues that because Ernst & Whinney failed to object to the instruction as given, it waived its ability to challenge the instruction on appeal. However, after submitting proposed jury instructions and their respective objections to the district court, the parties met with the court at a conference held off the record. On the record, the parties stipulated that this procedure would not be cited by any party as a failure to satisfy the requirements of Federal Rule of Civil Procedure 51, which would otherwise require further objection on the record. Knapp’s attempt to avoid the impact of this stipulation is baffling, and fails to persuade us that the alleged error has been waived.
The district court rejected Ernst & Whinney’s proposed instruction, which would have directed the jury to find loss causation if Ernst & Whinney’s “misrepresentations or misleading omissions ... were the reason for the [stock’s] decline in price.” Instead of using Ernst & Whinney’s instruction, the district court instructed the jury to find loss causation if it found,
[1438]*1438that there were material misrepresentations and/or omissions which caused the market price of ATV stock purchased by the class members to be higher than it would have been if all the true facts were known or if you find that ATVs stock would not have been issued.
The district court’s instruction, which focused on the reason for the plaintiffs’ “alleged loss” was an accepted statement of loss causation. In a fraud-on-the-market case, plaintiffs establish loss causation if they have shown that the price on the date of purchase was inflated because of the misrepresentation. See Gray v. First Winthrop Corp., 82 F.3d 877, 886 (9th Cir.1996) (loss causation present if “disclosure of the truth would have reduced the proper valuation of [the] investment” (internal quotation omitted)); Wool v. Tandem Computers, 818 F.2d 1433, 1437 (9th Cir.1987) (Wool) (focus should be on “the difference between the purchase price and the value of the stock at the date of purchase” (internal quotation omitted)). This is precisely what the court’s instruction required the jury to find before imposing liability. Although there may be other acceptable formulations of this element, the court’s instruction was proper.
Second, Ernst & Whinney argues that, as a matter of law, plaintiffs who both purchased and sold their stock before corrective statements had been made could not have established loss causation. Ernst & Whinney in effect argues that those who had unloaded their ATV stock before the price plummeted are unable to show that the misrepresentations caused their losses. But we have held repeatedly that price changes that occur before a corrective statement might still be the “result of market forces operating on the misrepresentation.” Wool, 818 F.2d at 1437; Blackie, 524 F.2d at 909 n. 25; Green v. Occidental Petroleum Corp., 541 F.2d 1335, 1345 (9th Cir.1976) (per curiam) (Sneed, J. concurring). Thus, those who trade before corrective statements are made can suffer losses caused by the misrepresentation.
For the same reason, Ernst & Whinney’s focus on the post-disclosure price (which was significantly higher than the value the jury found) as the true value of the stock is also wrong. Although the post-disclosure price may in some cases be evidence of the value of the stock before disclosure, the post-disclosure price is not a measure of pre-disclosure value. See Wool, 818 F.2d at 1437-38 & nn. 3-4. As the district court held, “[t]he jury has already determined the stock was inflated on each day of the class period. The jury also has determined that the cause of the inflation was the material misrepresentations and omissions. [Ernst & Whinney] cannot rear-gue at this time ... that forces other than the deceptions caused the overvaluation.”
V
Ernst & Whinney contends that the district court erred by instructing the jury on aiding and abetting liability under Rule 10b-5. After the first phase of the trial, but before a final judgment, the Supreme Court held that Rule 10b-5 does not impose liability for aiding and abetting a primary violation. Central Bank v. First Interstate Bank, 511 U.S. 164, -, 114 S.Ct. 1439, 1455, 128 L.Ed.2d 119 (1994). Ernst & Whinney renewed its motion for a new trial based on the Supreme Court’s intervening decision in Central Bank. The district court denied Ernst & Whinney’s motion.
Initially, we must decide whether Ernst & Whinney properly preserved its objection to the aiding and abetting instruction. Knapp argues that Ernst & Whinney submitted an aiding and abetting instruction to the court; thus, Knapp argues, Ernst & Whinney invited the error and we cannot review it. It is true that ordinarily invited error is unreviewable. Gilchrist v. Jim Slemons Imports, 803 F.2d 1488, 1493 (9th Cir.1986). However, review of invited error in an instruction may be appropriate “when the issue arises by virtue of a change in law while the appeal is pending.” United States v. Mkhsian, 5 F.3d 1306, 1310 (9th Cir.1993). That is exactly the circumstance before us.
This situation is analogous to our discretion to review an alleged error despite a failure to object to an instruction if a “solid wall of Circuit authority” would have rendered an objection futile. Robinson v. Heil-[1439]*1439man, 563 F.2d 1304, 1307 (9th Cir.1977) (per curiam). At the time of trial, it would have been pointless for Ernst & Whinney to object to the aiding and abetting instruction. Our law was settled. See, e.g., Levine v. Diamanthuset, Inc., 950 F.2d 1478, 1483 (9th Cir.1991) (aiding and abetting liability exists under Rule 10b—S).
Exercising our discretion to review the instruction also makes sense because the district court in fact did review the issue prior to entry of final judgment. Here, the district judge was not “blind-sided,” but fully considered the question on Ernst & Whinney’s motion for a new trial and made a decision that we can review. We hold that review was not foreclosed in this appeal.
The parties disagree as to the standard we should use to review the district court’s denial of Ernst & Whinney’s motion for a new trial. Ordinarily, we review the district court’s denial of a motion for a new trial for an abuse of discretion. Traver v. Meshriy, 627 F.2d 934, 940-41 (9th Cir.1980) (Traver) (“In determining whether to grant a new trial, the district court must decide whether, in its conscientious opinion, the verdict is clearly contrary to the weight of the evidence. On appeal, this court will not disturb a verdict unless, viewing the evidence in a manner most favorable to the prevailing party, we can say that the trial court abused its discretion.” (internal citations omitted)); Robins v. Harum, 773 F.2d 1004, 1006 (9th Cir.1985). However, we have also held that whether a case is submitted to the jury on incorrect legal theories is a question of law that we review de novo. Counts v. Burlington Northern R.R. Co., 952 F.2d 1136, 1139 (9th Cir.1991) (Counts). Ernst & Whinney would have us review the district court’s denial de novo; Knapp insists that our review is for an abuse of discretion.
Here, however, we are not reviewing an issue of law. In its denial of Ernst & Whinney’s motion, the district court correctly recognized that one of the theories of liability submitted to the jury was not proper. Ernst & Whinney does not challenge this aspect of the district court’s decision. Rather, Ernst & Whinney is challenging the district court’s decision to refuse a new trial despite the erroneous submission. This decision was entirely discretionary.
The jury rendered a general verdict and therefore did not state whether its findings were based on a primary theory of liability or a secondary theory of liability. “As a general rule, a general jury verdict will be upheld only if there is substantial evidence to support each and every theory of liability submitted to the jury.” Kern v. Levolor Lorentzen, Inc., 899 F.2d 772, 777 (9th Cir.1990) (Kern) (internal quotation omitted). In Traver, however, we carved out an exception to this general rule, holding:
Where more than one theory of recovery has been submitted to the jury in a civil case, and where on appeal it is claimed that as to one of the theories there was a lack of evidential support or an error of law in submitting the theory to the jmy, the reviewing court has discretion to construe a general verdict as attributable to another theory if it was supported by substantial evidence and was submitted to the jury free from error.
627 F.2d at 938. This narrow exception has been the target of vigorous and persuasive criticism. See Kern, 899 F.2d at 789-92 (Kozinski, J., dissenting); compare Anixter v. Home-Stake Prod., 77 F.3d 1215, 1229 (10th Cir.1996) (Anixter) (reversing general verdict that could have been based on either primary or secondary liability under Rule 10b-5 because it could not be said ‘“with absolute certainty’ that the jury was not influenced by the submission of the improper or erroneous instruction”). However, this exception is the law in our circuit, and the district court, which first reviewed the alleged error, had the discretion to apply the exception if it found that substantial evidence supported a primary theory of liability.
Our review of the district court’s decision to apply the exception in Traver is limited to whether the court abused its discretion. We will not reverse this exercise of discretion unless we have “a definite and firm conviction that the court below committed a clear error of judgment in the conclusion it reached upon a weighing of the relevant fac[1440]*1440tors.” Smith v. Jackson, 84 F.3d 1213, 1221 (9th Cir.1996) (Smith).
In deciding to exercise its discretion under Traver, the district court examined the two theories of liability — primary and secondary liability — and analyzed the following:
(1) the potential for confusion of the jury,
(2) whether the losing party’s defenses apply to the count upon which the verdict is being sustained, (3) the strength of the evidence supporting the count relied upon to sustain the verdict, and (4) the extent to which the same disputed issues of fact apply to the various legal theories.
citing Counts, 952 F.2d at 1140; see also Portland Feminist Women’s Health Center v. Advocates for Life, Inc., 62 F.3d 280, 285-86 (9th Cir.1994) (applying analysis in case where an intervening Supreme Court decision eliminated a theory of liability that had been submitted to the jury). It is undisputed that the district court applied the Traver analysis. Thus, the district court’s decision was based “upon a weighing of the relevant factors.” Smith, 84 F.3d at 1221. The remaining question is whether the district court “committed a clear error of judgment” in deciding that although an aiding and abetting theory of liability could not support the jury’s general verdict, sufficient evidence of primary liability under Rule 10b-5 supported the verdict. See id.
In focusing on the possible jury confusion, the district court reasoned that Knapp did not emphasize a theory of secondary liability, and focused on Ernst & Whinney’s failure to include a going concern qualification. The district court, after reviewing the record, stated that it was “unlikely that the jury found the defendant liable on a secondary theory of liability.”
Ernst & Whinney disputes the district court’s view of Knapp’s evidentiary focus. Ernst & Whinney argues that each expert spent some time, and some experts spent the majority of time, providing evidence that supported secondary liability. Therefore, Ernst & Whinney insists that, because a larger percentage of time was spent examining experts about ATVs unaudited financial statements, the district court “clearly erred” in deciding that Knapp’s focus was on a primary theory of liability. Even if Ernst & Whinney’s comparative transcript-page count is accurate, a review of the entire transcript supports the district court’s position that the parties emphasized Ernst & Whinney’s failure to include a going concern qualification. For example, Knapp’s accounting and auditing expert spent a significant portion of her testimony explaining why Ernst & Whinney should have included a going concern qualification. As the district court explained, Knapp’s closing argument emphasized the lack of the going concern qualification, and mentioned Ernst & Whinney’s association with the financial statements as a separate issue. This separation of issues supports the district court’s determination that the jury was unlikely to be confused. Thus, the record supports the district court’s decision that the potential for jury confusion was minimal.
The district court next determined that Ernst & Whinney’s defenses applied essentially equally to both theories of liability. Ernst & Whinney does not dispute this, but instead asserts that its defenses “most certainly would not have been the same if there had been no claim for aiding and abetting.” We do not doubt that Ernst & Whinney’s trial strategy would have been different absent an aiding and abetting theory of liability. However, Traver instructed the district court to consider only “whether privileges or defenses of the losing party apply to the count upon which the verdict is being sustained so that they would have been considered by the jury with reference to the count.” Traver, 627 F.2d at 938-39. The district court properly considered this part of the Traver analysis.
The district court next discussed the strength of the evidence supporting the primary liability theory. The court determined the evidence was substantial, referring specifically to:
(1) the mix of information contained in the prospectus and its included financials materially misstated the company’s current economic posture and known future prospectus, (2) the defendant either subjectively knew of the misstatements or was willfully blind to them, and (3) the defen[1441]*1441dant was motivated by the desire to garner substantial fees. The court also found the evidence was sufficient to convince a reasonable jury that the defendant was aware the scaled back offering would dramatically impact ATV’s liquidity. Nevertheless, the defendant failed to include a going concern qualification.
We do not fault the court’s application of the third part of the Traver analysis.
Finally, the district court decided that underlying both theories of liability were similar issues of fact. “The critical element separating primary from aiding and abetting violations is the existence of a representation, either by statement or omission, made by the defendant, that is relied upon by the plaintiff.” Anixter, 77 F.3d at 1225. Here, the primary issues of fact were whether the statements made within ATV’s financial statements were deceptive and whether Ernst & Whinney acted with scienter. The district court correctly determined that there is sufficient applicability of these two issues to both theories of liability.
Therefore, because the district court properly weighed each of the Traver considerations, the court did not abuse its discretion in applying the exception and upholding the general verdict.
VI
After the jury returned its verdict, Knapp moved for prejudgment interest, calculated at the rate prevailing during the class period. Knapp cross-appeals, challenging the district court’s denial of this request. We review the district court’s decision for an abuse of discretion. Burgess v. Premier Corp., 727 F.2d 826, 838 (9th Cir.1984). However, the required demonstration of a clear error of judgment is phrased somewhat differently when reviewing the issue of prejudgment interest. “The court’s decision ... [is to be] upset only if it is so unfair or inequitable as to require it.” Id. “Whether interest will be awarded is a question of fairness, lying within the [trial] court’s sound discretion, to be answered by balancing the equities.” Wessel v. Buhler, 437 F.2d 279, 284 (9th Cir.1971).
Before denying Knapp prejudgment interest, the district court considered the facts and provided several bases for its decision. First, the district court stated that awarding Knapp prejudgment interest was improperly speculative. The district court correctly stated that prejudgment interest serves a compensatory function, designed to make the injured party whole. Frank Music Corp. v. Metro-Goldwyn-Mayer, Inc., 886 F.2d 1545, 1550 (9th Cir.1989), cert. denied, 494 U.S. 1017, 110 S.Ct. 1321, 108 L.Ed.2d 496 (1990). The district court also correctly recognized that speculative damages are generally not awarded. Lindy Pen Co. v. Bic Pen Corp., 982 F.2d 1400, 1407-08 (9th Cir.), cert. denied, 510 U.S. 815, 114 S.Ct. 64, 126 L.Ed.2d 34 (1993). It concluded that the speculative nature of prejudgment interest in this case cut against allowing an award.
As the district court rightly explained, ATV was an exceptionally speculative investment. Furthermore, as the district court stated, “[h]ad the class members not purchased ATV stock, they may have invested in something else with the expectation of gaining a profit.” The district court pointed out that “it is speculative to conclude that other investments would have been profitable,” and that, therefore, prejudgment interest was also impermissibly speculative.
We have doubts about this reasoning. If, as the district court postulated, this same class of plaintiffs had invested in a basket of securities as risky as ATV, the class as a whole would expect to receive from then-investments a rate of return higher than the risk-free rate prevailing during the class period. Thus, awarding Knapp interest based on the prevailing risk-free rate would not necessarily have been impermissibly speculative. See, e.g., Western Pacific Fisheries v. S.S. President Grant, 730 F.2d 1280, 1289 (9th Cir.1984) (using risk-free 52-Week Treasury Bill rate).
The district court’s second reason was an unwillingness to award prejudgment interest because “[t]he record is devoid of any support for the premise that the case involved evil intent or premeditated fraud by any defendant.” The defendants’ degree of culpability was permissible for the district [1442]*1442court to consider. See Osterneck v. Ernst & Whinney, 489 U.S. 169, 176, 109 S.Ct. 987, 991, 103 L.Ed.2d 146 (1989) (court may consider, among other things, “the degree of personal wrongdoing on the part of the defendant”). Furthermore, the district court’s findings as to the defendants’ extent of culpability were not clearly erroneous. Although each defendant’s state of mind may have triggered liability under Rule 10b-5, the district court’s analysis was consistent with the jury’s findings. The district court did not find that the defendants lacked scienter. Rather, it found that the evidence pointed to a state of mind less than a malicious intent to defraud investors. The record contains ample support for the finding.
Finally, the district court also properly considered whether prejudgment interest would amount to a windfall recovery for Knapp. See id. (court may consider “whether prejudgment interest is necessary to compensate the plaintiff fully for his injuries”). The district court held that because Knapp received an amount from settling defendants that was even greater than the jury’s verdict, prejudgment interest was not necessary for full compensation to Knapp. The district court’s finding was not clearly erroneous.
Overall, the district court balanced the equities and determined that prejudgment interest was not in the interest of justice. Although we have doubts about the district court’s argument that prejudgment interest would be overly speculative, its overall balance was not an abuse of discretion. The district court properly considered two issues and its overall analysis did not indicate its discretion was abused.
Because we affirm the jury’s verdict, we need not consider Knapp’s request to reinstate dismissed pendant state-law claims.
AFFIRMED.