SECURITIES FRAUD DAMAGES UNDER THE PSLRA

Transcription

1 Southeastern Oklahoma State University From the SelectedWorks of Mohammed A Misbah February 4, 2015 SECURITIES FRAUD DAMAGES UNDER THE PSLRA Mohammed A Misbah Available at:

2 SECURITIES FRAUD DAMAGES UNDER THE PSLRA By Mohammed A. Misbah INTRODUCTION The United States Private Securities Litigation Reform Act of 1995, ("PSLRA") implemented several substantive changes affecting certain cases brought under the federal securities laws. It was designed to reduce the number of frivolous securities lawsuits filed in federal courts. Prior to the PSLRA, a case could proceed with minimal evidence and use pre-trial discovery to search for more evidence that strongly suggested a deliberate fraud. Under the PSLRA plaintiffs need such evidence in order to commence an action. BODY I. FRAUD ON THE MARKET AND LOSS CAUSATION The PSLRA makes clear that a plaintiff in a Rule 10b-5 case has the burden of proving that the fraud perpetrated by the defendant caused the loss for which the plaintiff seeks to recover damages. 1 After the Supreme Court's decision in Dura Pharmaceuticals v. Broudo, 2 it is now clear that a plaintiff must allege loss causation in its complaint. In the case of In re Williams Securities Litigation, 3 the Tenth Circuit stated that in order to satisfy the requirements of Dura and prove damages, plaintiffs in security cases need to prove that the defendant (1) made a material misrepresentation or omission, (2) with scienter, (3) in connection with the purchase or sale of a security, (4) upon which the plaintiff relied, that (5) the plaintiff suffered economic loss, and (6) the material misrepresentation was the cause of that loss. A. Efficiency of Markets Due to the efficiency of markets Plaintiffs need not establish that they specifically relied on the fraudulent misinformation or omission when purchasing a company s stock. Under the generally accepted 1 15 U.S.C. 78u-4(b) (4) U.S. 336 (2005) F.3d 1130, (10th Cir. 2009) 1

3 Efficiency of Markets theory the market acts as the unpaid agent of the investor informing them that, given all the information available to it, the value of the stock is worth the market price. When a fraud is committed and false information about a company is released to the market, the price of that company s stock floats freely as influenced by the information. Investors are, therefore, justified in treating the market price as if it reflects the true value of the company s stock. However, perfect efficiency is not possible. No securities market will be completely efficient to the point that information will not need to be analyzed. If investors did not analyze public information there would be no mechanism by which such information would be reflected in market prices. The question that arises from the fraud on the market theory is: how efficient must a market be for a court to conclude that investors were justified in relying on the market price? For purposes of stock inflation due to fraud, the market is most likely efficient enough. For purposes of stock deflation due to corrective disclosures the answer is not so clear. 1. Effect of Fraud Stock Prices are usually a reflection of a company s cash flows. When cash flows are greater, stock prices are higher. This makes a company s financial reports very important, as well as any other information that might affect the company s cash flows. When a company makes a fraudulent statement that, if true, would enhance the company s cash flows, the stock price becomes artificially inflated. Economic loss causation is only concerned with the portion of an investment loss that would not have occurred had the fraud not been perpetrated. Inflationary Loss is the inflation in the stock price at the time of purchase minus the inflation in the share price at the time of sale. That inflation portion of the loss attributable to fraud is recoverable, but the entire investment loss is not recoverable. 2. Corrective Disclosures Corrective disclosures constitute any disclosure, communication, or event that informs the market that a company has committed fraud. Corrective disclosures must consist of new material information that is 2

4 company specific and fraud related. They do not have to be issued by the company and there does not have to be one complete disclosure. After a corrective disclosure has been made the market corrects the price of the company s stock by subtracting the inflation caused by the fraud As will be discussed below, the market is usually efficient for purposes of corrective disclosures. However, exceptions to market efficiency do occur. A company can dissipate the impact of corrective information by revealing a fraud slowly through multiple corrective disclosures. When measured individually, such disclosures may have only a marginal impact on the price of the company s stock. However, when considered in sum, the market will realize the disclosures collective importance and the price of the company s stock will decline. This is known as the Leakage Theory. Conversely, a company s stock price may initially decline too much after a corresponding corrective disclosure due to the fear of possible adverse legal actions or business repercussions. However, after the possibility of such occurrences has been dissipated the company s stock price may adjust back upwards in the following days. This is known as Market Overreaction. It should be noted that corrective disclosures are not always necessary to realize loss causation. For instance, consider a company that commits fraud and inflates their stock price by 100 percent from $50 to $100. If the company subsequently goes bankrupt, due to reasons unrelated to the fraud, investors will suffer a 100 percent loss of the $100 share price. Since the bankruptcy would have occurred regardless of the fraud, investors would have lost 100 percent of the stock price anyways. However, if no fraud had occurred, their loss would only be $50 per share not $100 per share. Thus while no corrective disclosure occurred, investors should still be able to recover the $50 per share portion of their loss caused by the fraudulent inflation. B. Calculating Recoverable Damages There are multiple ways to calculate damages in a fraud case. The Dollar Drop Method assumes that the dollar decline in the price of a company s stock, associated with a corrective disclosure, is exactly equal to the dollar decline that would have occurred had the truth been disclosed at the time of purchase. However, the 3

5 value that the market assigns to fraudulent information can vary over time and often depends upon multiple economic variables. The dollar drop method systematically underestimates the inflation per share when industry and company share prices have generally been declining and it systematically overestimates inflation when industry and company share prices have been generally rising. For example, if a company overstated the size of its bond portfolio, its stock price may initially become inflated. However, the inflation would fall with an increase in interest rates and the corresponding reduction in bond prices. The problem with the dollar drop method is that it does not take such decreases or increases in inflation into account. Additionally, as an earnings statement recedes into the past, the misrepresentations that it contains become less relevant. Inflation can decline due to market forces to the point where the fraud has zero effect. Conversely, the inflation in a stock can grow over time due to market forces. Suppose the true value of an energy company s stock is initially $7. After the company fraudulently claims that it has found an oil reserve, the company s stock price becomes inflated by $3 and the total price of the stock becomes $10. Over time the market value of non-existent oil increases, while the market value of the company s actual energy sources decreases. The price of the company s stock may stay at $10, but the true value decreases to $2 and inflationary value increases to $8. After a corrective disclosure reveals that there is no oil, the stock price drops down to $2. The damages per share will not be the $8 decrease at the time of the disclosure but only the $3 of inflation at the time of the purchase. The Percentage Drop Method, also known as equivalent disclosure, requires that percentage increases due to fraud and percentage decreases due to corrective disclosures be translated into percentage changes that would have occurred if the truth had been disclosed in a timely manner. This method takes the growth or decline in inflation into account, as it factors out the change in inflation caused by non-fraud related events. 4

6 II. ADDRESSING THE THREE ERRORS MADE BY PLAINTIFFS In PSLRA cases, defendants often attempt to get the cases dismissed, during summary judgment, by arguing that the plaintiffs cannot sufficiently prove that they were harmed by the defendant s fraudulent behavior. Defendants usually attempt to make this argument in three parts. First, in attacking a plaintiff s estimation of damages per share, defendants argue that plaintiffs overstate corrective disclosures. Second, defendants argue that plaintiffs ignore the effect of, non-fraud related, confounding information that could also have lowered the price of the stock. Finally in attacking a class of plaintiffs estimation of aggregate damages, defendants argue that plaintiffs rely on unreliable statistical models to make their calculations. While these arguments may sometimes have merit, they are often used by defendants against ill prepared plaintiffs to stifle litigation. Various PSLRA cases have addressed the issues presented by these arguments and as a result the requirements for proving damages under the PSLRA have become more defined. By adhering to the criteria established in these cases, plaintiffs should be able to overcome these arguments. A. Overstating Corrective Disclosures The first argument that defendants make is that plaintiffs overstate the number of corrective disclosures. In In re Williams Securities Litigation, 4 the Tenth Circuit seemed to acknowledge the leakage theory and held that a full corrective disclosure by the company is not required to prove loss causation. A company can also dissipate the impact of information by continuing to perpetuate the fraud by denying the truth of corrective disclosures made by others or announcing the fraud along with other negative news and attributing the stock price decline solely to the non fraud related news. However, the inability of a plaintiff to point to a single corrective disclosure does not relieve a plaintiff from showing how the fraud was revealed. A plaintiff cannot simply state that the market learned the truth by a certain date. 5 Plaintiffs must show how and when the fraud was revealed. This can be done by stating 1) when 4 Supra, note 3 at Id. at

7 the corrective disclosures first began, 2) what subsequent events or disclosures contributed to the revelation, and 3) when the market was fully aware of the fraud. A fraud can begin to be revealed in a number of ways such as company press releases, announcements or investigations by regulatory agencies, analyst reports, news articles, and in some circumstances the occurrence of an actual harm that was concealed by the fraud. In In re Omnicom Group, Inc. Securities Litigation 6, the Second Circuit held that corrective disclosures not only must reveal some aspect of the fraud, they must also disclose new information to the market. Clever fraud defendants might attempt to abuse Omnicom in order to escape liability. When the initial corrective disclosure is identified by plaintiffs the defendants will argue that the stock price did not significantly react and therefore there are no damages. However, when stock price reacts to subsequent disclosures defendants will argue that the subsequent disclosures were not new information and cannot be considered corrective. In order to prevent such abuse by defendants, the facts in each case must be analyzed and the types of corrective disclosures involved must be considered. 1. Analyst Reports In Omnicom, the court further stated that a negative journalistic characterization of previously disclosed facts does not constitute a corrective disclosure. 7 However, in In re Apollo Group Inc. Securities Litigation, 8 the U.S. District Court for the District of Arizona rejected the rigid facts only approach of Omnicom. The court held that additional information in conjunction with the previously released information may be necessary to finally and fully reveal the fraud. 9 The court noted that the Ninth Circuit has also declined to adopt a bright line rule assuming an immediate market reaction and instead focuses on the facts of each case. 10 As previously noted market efficiency is not perfect. While most information can be said to be immediately absorbed by the market and reflected in a company s stock price, some information requires more F.3d 501, 514 (2d Cir. 2010) 7 Id. at F.Supp.2d 837 (D. Ariz. 2007) 9 Id. at Id. 6

8 time to be fully absorbed by the market. The court admitted that it is a very rare type of securities-fraud case in which a corrective disclosure will require such clarification. 11 Basically what the holding of Apollo means is that analyst reports can be considered corrective disclosures when the information they are based on was of such complexity that it required sophisticated analysis to explain to the market or the information they are based on was released to the market in an altered way designed to avoid market reactions. An analyst s report must be factually accurate and the first to clarify the information. 12 An example can be found in the case of In re Executive Telecard. 13 In Telecard, the U.S. District Court for the Southern District of New York held that the disclosure of a company s fraud began with a judge s opinion in an earlier case, but did not become widely dispersed to the market until the publication of a Barron s article several days later Investigations In In re Hansen, 15 the defendants were being investigated by the SEC for stock option backdating and improper accounting practices. After the investigation was announced stock prices plummeted. The court held that the mere existence of an investigation cannot support any inference of wrongdoing or fraudulent scienter. However, in In re Take Two Securities and Teamsters Local 617, 16 the defendant company revealed that it had commenced an internal investigation of its past options grants. The disclosure clearly revealed that Take- Two's options-granting practices were the subject of an investigation carried out by a governmental entity charged with ensuring compliance with the federal securities laws. Take-Two's share price dropped by 7.5% in response to the disclosure. The U.S. District Court for the District of Arizona stated that other courts have 11 In re Apollo Group Inc. Securities Litigation, 2008 WL (D. Ariz.) 12 Id. at F.Supp (S.D.N.Y. 1997) 14 Id. at F.Supp.2d 1142 (C.D. California 2007) F.Supp.2d 763 (D.Ariz., 2009) 7

9 found that similar allegations of significant stock drops in response to announced SEC investigations are sufficient to plead loss causation under the framework established by Dura and its progeny. What these rulings essentially mean is that the announcement of an investigation alone is not a corrective disclosure, as the defendant may in fact have done nothing wrong. However, if the announcement of an investigation contains content that suggests impropriety or fraud, then the announcement can be considered a corrective disclosure. 3. Zone of Risk In Lentell v. Merrill Lynch and Co. Inc, 17 the Second Circuit Court of Appeals stated that a misstatement or omission is the proximate cause of an investment loss if the risk that caused the loss was within the zone of risk concealed by the misrepresentation or omission. A risk is considered to be within the zone of risk concealed by a fraud if the truth would lead a reasonable investor to consider that risk when purchasing a company s stock. 18 In Holmes v. Grubman, 19 the Supreme Court of Georgia stated that the truth could be revealed by the actual materialization of the concealed risk rather than by a public disclosure that the risk exists. Essentially the zone of risk covers any event in which the reasonable probability of occurring was concealed by the fraud. Examples can include bankruptcy, lower (than stated expected) earnings, or negative action by a regulatory agency. It is important to note that such risks are not merely investment risks but risks specifically hidden from the market by the fraud. B. Not Recognizing Confounding Information The second argument that defendants make is that plaintiffs ignore the effect of, non-fraud related, confounding information. Defendants inevitably point the court to Judge Sneed s concurrence Green v. Occidental Petroleum Corp. 20 In Green, Judge Sneed coined the term value line and essentially stated that F.3d 161,173 (2d Cir. 2005) 18 Id S.E.2d 196, 201 (Ga. 2010) WL 367 (C.D.Cal.) 8

10 damages in a securities fraud case are limited to the difference between the price at which the stock actually sold and the price at which the stock should have sold absent the fraud. In In re Executive Telecard, 21 the U.S. District Court for the Southern District of New York stated that damages in securities fraud cases are limited to the inflation caused by the fraud at the time of purchase. This requires elimination of price decline that is the result of forces unrelated to the fraud. There are three types of confounding information that can have an effect on a company s stock price. First there is market wide information that affects the stock prices of companies across the entire market in all industries. Second there is industry wide information that affects the stock prices of companies in specific industries. Finally there is company specific information. This is other news about the company, not related to the fraud that may have an effect on the company s stock price. 1. Event Study In In re Executive Telecard, 979 F.Supp (S.D.N.Y. 1997), the U.S. District Court for the Southern District of New York chastised the plaintiffs for not distinguishing between fraud related and non-fraud related influences. Two years later in In re Oracle Securities Litigation, 22 the U.S. District Court for the Central District of California did the same. However, the court in Oracle also stated that [a]nalysis of loss causation calls on courts to perform a balancing act between allowing plaintiffs to link each and every bit of negative information about a company to an initial misrepresentation that overstated that company's chances for success and exacting such a high standard as to eliminate the possibility of 10b-5 claims altogether. 23 What plaintiffs need to do in order to establish loss causation per In re Williams Sec. Litig., is (1) show how the fraud was revealed; (2) show how the revelation of the fraud caused the decline in stock price; and (3) differentiate the effect of revelation from market, industry, and non fraud related company news. 24 The court in F.Supp. 1021,1025 (S.D.N.Y., 1997) WL , *12 (N.D. Cal. 2009) 23 Id. (citing In re Williams Sec. Lifting., 558 F.3d 1130, 1140 (10th Cir.2009)). 24 Supra, note 3 at

11 Oracle required the use of an Event Study to filter out the effect of non-fraud related influences. An event study is a method used to assess the impact of an event on the value of a firm. In an event study, statistics are used to quantify some variable in an objective manner that is standard and replicable. Important things to consider when conducting an event study are (1) what event windows should be used to measure the stock s abnormal and normal rates of change; (2) What type of Industry Index to use, and (3) What type of non-fraud related information to consider. a. Event Window An event window is a period of time over which the impact of an event is measured. For fraud on the market cases the event window is the period of time over which the effect of the fraudulent disclosure/omission is measured. The impact of the fraud is measured by the abnormal change in price of a company s stock. The start date for the event window should be the date when the fraudulent information was dispersed to the market. The end date for the event window should be the date when the fraud was fully revealed and absorbed by the market. The market may take several days to react to the disclosure of fraud. Conversely the market may overreact and take several days to correct itself. Event windows should take the Leakage Theory and Market Overreaction into account. For example, if Alpha Company makes a corrective disclosure and the market overreacts, Alpha s stock price may immediately decline from $10 to $1, even though the inflation caused by the fraud was only worth $5. The event study must take this overreaction into consideration. However, it can be difficult to determine when the market has corrected itself and the true value of the stock is reflected in its price. For instance, in the example above, Alpha s stock price may rise to $3 at the close of the second day after the disclosure. On the third day after the disclosure, Alpha s stock price rises to $3.75. By mid morning it is at $4.00. In the afternoon it falls back to $3.50, and at close Alpha s stock price is again only $3. If only daily data is looked at it would appear that Alpha s stock price has settled at $3. However, if 10

12 inter-day data is considered, it is apparent that Alpha s stock price is still quite volatile. Inter-day data should be considered in determining when to end the event window. b. Company Specific Information The movement of the stock price during the event window should be compared to the stock s normal rate of change. The normal rate of change can be determined by the movement of the stock price over a different but similar window of time that was unaffected by fraud. For instance, if the event window includes a date when the company issued a press release about a new product, unrelated to the fraud, then the comparison window should also cover a span of time that includes a press release about a new product. If the change/decrease in the company s stock price is much different during the event window than it is during the comparison window, then the event study can be used to show that the fraud caused an abnormal change/loss in the price of the company s stock. Event studies must also factor out the effect of company specific information. This is the other news about the company that is not related to the fraud, but may have an effect on the company s stock price. The effects of company specific information can be isolated through statistical analysis by assigning the information a variable. For instance, in the above example, the product related press release is company specific information that should be isolated as a variable in a statistical model. When we compare the stock s performance during the event window to its performance during the comparison window, we can see the portion of the change in stock price that was caused by the non-fraud related information. If this change is negative it should be subtracted from the calculation of damages per share. c. Market and Industry Indexes In addition to factoring out company specific information, the change in a company s stock price during the event window should also be compared generally to the change in the stock prices throughout the market and specifically to the change in the stock prices of other companies in the same industry. Therefore the event 11

13 study can be used to show that the company s abnormal change in stock price was due to the fraud committed and not to any market or industry wide effects. Market wide comparisons are made by comparing the performance of the company s stock, during the event window, to the performance of an overall market index, such as the Wilshire 5000 or the S&P 500, during the same time period. Industry comparisons can be made by comparing the performance of the company s stock to the performance of industry indexes, such as those used by rating agencies like Moody s or Standard and Poor s. In In re Executive Telecard, 25 the court criticized the plaintiff s industry comparison because the defendant company was historically much more volatile than the companies that made up the index. In order to avoid this, plaintiffs should use indexes consisting of companies that are similar to the defendant company. Similar companies should produce comparable products and services, be of comparable size or market share, and have comparable levels of historic volatility. 2. Collateral Damage In addition to conducting event studies to factor out the effect of non-fraud related information, plaintiffs must also consider the effects of Collateral Damage. When a corrective disclosure occurs, the market acquires two pieces of information: 1) the substantive information that it would have had if the fraud had never occurred and 2) the knowledge that the company committed fraud. 26 The first piece of information causes the market to correct the stocks inflation. 27 The second piece of information causes the market to reevaluate the company and can lead to collateral damage. 28 Examples of collateral damage are the market s loss of faith in the company s management, the fear of pending litigation, and the unwillingness of others to do business with the company. Collateral damage results in the decline of a company s stock price, due to the disclosure of fraud, in addition to the inflation caused by F.Supp. 1021, 1027 (S.D.N.Y. 1997) 26 Bradford Cornell, Collateral Damage and Securities Litigation, 3 U. Utah L. Rev (2009). 27 Id. 28 Id. 12

14 the fraud. It is a well-settled principle that damages in a securities fraud case are limited to the difference between the price at which a stock sold and the price at which the stock would have sold absent the fraud. 29 Defendants use this principle to argue for a damages amount that is as small as possible. However, to disallow plaintiffs in a fraud case from recovering declines in stock price due to collateral damage is most unjust. First, collateral damage, like a stock s decrease in inflation, is directly caused by the disclosure of the fraud. If the objective of a legal rule is to deter certain undesirable behavior without simultaneously deterring beneficial behavior then disallowing collateral damages makes no sense. Allowing plaintiffs to recover collateral damages will only further deter companies from committing the undesirable behavior of fraud. It would not have any deterrent effect on beneficial behavior because collateral damage, like inflationary loss, is caused by fraud and not any legitimate business purposes. Second, collateral damage is in a way a form of inflation. A company s stock price can be inflated not only due to a specific misrepresentation but also due to the general presumption that the company s management is competent and trustworthy. Upon disclosure of the fraud, this general presumption of trustworthiness is dispelled, and the stock price s inflation accordingly decreases. C. Unreliable Statistical Models Used To Compute Aggregate Damages The third argument that defendants make is that plaintiffs use unreliable statistical models to compute aggregate damages. In fact, defendants have been successful in getting cases dismissed during summary judgment by arguing that the plaintiffs damages model does not meet the Daubert standards. In Daubert v. Merrell Dow Pharmaceuticals, 30 the United States Supreme court held that scientific or technical evidence must be reliable in that 1) the technique or model used to generate the evidence can be tested, 2) the model was subject to peer review, 3) the known or potential error rate of the model has been identified, and 4) the model has been generally accepted by the relevant scientific community. However, contrary to what defendants might 29 Supra, note U.S. 579 (1993) 13

15 argue, aggregate damages models can be very helpful in large securities fraud cases. If plaintiffs take certain precautions in creating their damages models, they should be immune to Daubert scrutiny. Aggregate damages are calculated by multiplying damages per share by the total number of damaged shares. Juries must receive evidence showing the number of damaged shares. Defendants argue that it is impossible to ascertain each class member s trades and thus find the aggregate number of damaged shares. However, plaintiffs rightfully argue that individual claims data should not be used to estimate aggregate damages because it will underestimate the total number of damaged shares. This is because the victims of fraud only have enforcement rights in proportion to their injuries. 31 It is unlikely that each damaged shareholder will file an individual claim, because it would cost too much to investigate and prosecute the offense and the litigation would not be worthwhile. 32 This is why statistical models are necessary to calculate aggregate damages. 1. Single Trader Models Defendants will inevitably point to the decision in Kaufman v. Motorola Inc., 33 in which the court held that a single trader/proportional model did not meet Daubert standards because it was never tested against reality and never accepted by professional economists. However, the results from proportional single trader models are consistent with the results of more accurate multi-trader models when appropriate assumptions and parameters are used. 34 Most studies that discredit the single trader model refer to models that use daily trading volume and total shares outstanding to calculate aggregate damages. Using total shares outstanding and daily trading volume overstates damaged shares because they include trades made by specialists and day traders Frank H. Easterbrook, Optimal Damages in Securities Cases, 52 U.Chi.L.Rev. 611, 621 (1985). 32 Id WL (N.D. Ill. 2000) 34 Michael Barclay & Frank C. Torchio, A Comparison of Trading Models Used for Calculating Aggregate Damages in Securities Litigation, 64 Law and Contem. Prob. 105, (2001). 35 Id. at

16 Recently published research suggests that trading volume should be decreased by 10% for shares trading on the NYSE and by 15% for shares trading on the NASDAQ. 36 When single trader models use float and adjusted volume rather than daily trading volume and total shares outstanding, they are much more accurate. Float is defined as the number of shares that could have been traded on a given day. 37 Using float instead of total shares outstanding reduces the number of damaged shares calculated by a model. For example it is common for float to be less than 50% of outstanding shares for a class period of 128 days. 38 Adjusted volume is the daily trading volume adjusted to eliminate insider and intra-day specialist trades. 39 Using adjusted trading volume, single trader models can better differentiate between shares that were traded during the class period and damaged shares that were retained by investors. 2. Multi-Trader Models While single trader models are often subject to criticism multi trader models are actually looked upon quite favorably. Two-trader models divide a company s shares into two groups and daily trading volume is determined by each group s relative propensity to trade. 40 Multi-trader models go a step further and split the shares of a company s stock into multiple groups with different acceleration rates. The different acceleration rates reflect the propensity of each group s shareholders to trade the stock during the class period. Multi-trader models use multiple rates to achieve a better estimate of the number of shares that were actually damaged by the defendant s fraudulent conduct. Some experts have advocated for the use of models with no less than three classifications of investors: investors, traders, and day traders. Even the defense oriented National Economics Research Associates, has advocated for a trading model that uses at least three trading groups. 36 Id. 37 Id. at Id. at Id. at Id. at

17 In In re Frontier Ins. Group Securities Litigation, 41 the U.S. District Court for the Eastern District of New York denied the defendants Daubert motion to have expert testimony using a multi-trader model stricken. In In re WorldCom Securities Litigation, 42 a multi-trader model was used to calculate aggregate damages. The U.S. District Court for the Southern District of New York held that the permissibility or reliability of the stocktrading model was not at issue and that evidence of an aggregate damages figure is appropriate where there is actual trading data available for over fifty percent of the trades at issue. The multi trader model used in Worldcom has survived repeated Daubert challenges in other cases. 3. No Perfect Model It is probable that any attempt by the defense to exclude the expert testimony based on a properly constructed single trader model or multi trader model would fail. There will of course be shortcomings to any aggregate damages model. However, in In re RMed Int l Inc v. Sloan s Supermarkets, Inc., 43 the U.S. District Court for the Southern District of New York held that these shortcomings do not go to the model s overall admissibility but merely to its weight and credibility. This implies that a plaintiff s aggregate damages models should be subject to cross-examination and opposing expert testimony, but not to dismissal. CONCLUSION In a securities fraud case, companies perpetrate fraud on the market by making misrepresentations or omissions that cause artificial inflation in the company s stock price. Due to this, investors relying on the efficiency of markets will pay an artificially high price for the company s stock. When the fraud is inevitably revealed, the price of the company s stock will decline by the inflationary amount. This decline in inflation represents the damages that are recoverable in a securities fraud case. Under the PSLRA, plaintiffs have the burden of proving loss causation. In order to avoid the charge that they overstate corrective disclosures, plaintiffs must ensure that each corrective disclosure cited provides new or 41 No. 1:94-cv LBS, slip op. (E.D.N.Y., 2002) WL , *4 (S.D.N.Y. 2005) WL , *2 (S.D.N.Y.). 16

18 clarifying information that contributed to the revelation of the fraud. In order to avoid the charge that they ignored confounding information, plaintiffs must conduct event studies that factor out the effects of company specific information and industry and market wide occurrences. In order to avoid the charge that they rely on unacceptable statistical models in calculating aggregate damages plaintiffs must structure their models according to appropriate assumptions and parameters. If plaintiffs do these three things, they should meet the requirements of pleading damages in a securities fraud case under the PSLRA. 17