Loss Causation vs. Price Impact: What Each Requires, and When

Price impact asks whether the alleged misstatements affected the security’s price; loss causation asks whether the revelation of the truth caused the investors’ loss. They are answered from the same price data, usually by the same event study, at different stages of a Rule 10b-5 case and under different burdens of proof, and conflating them is the most common error in briefs and in expert reports on both sides. This note sets the two side by side: what each requires, who must prove it, when, what the economist produces for each, and where the answers diverge. The practice is on our securities class action page.

Two concepts, two stages, one data set

Price impact

The presumption of reliance in Basic rests on the premise that in an efficient market the price reflects public material information, so an investor who buys at the market price relies on the integrity of that price and, through it, on the misstatement. The presumption can be rebutted by “any showing that severs the link” between the misstatement and the price. Three later decisions set the terms on which that happens at class certification. Halliburton I held that plaintiffs need not prove loss causation to invoke the presumption. Amgen Inc. v. Connecticut Retirement Plans and Trust Funds, 568 U.S. 455 (2013), held that they need not prove materiality either, because materiality is a common question. Halliburton II held that defendants may rebut the presumption at certification with evidence that the misstatement had no price impact, since price impact is the premise on which the presumption rests. Goldman then held that the defendants bear the burden of persuasion on that question, and that the generic nature of a statement is relevant to price impact even where it overlaps with materiality: where the front-end statement is generic and the back-end disclosure is specific, the inference that the back-end decline measures the front-end statement’s effect weakens.

Front end and back end

Price impact can be shown at either end of the class period. Front-end evidence is a price reaction on the day the misstatement was made. It is often absent, because many actionable statements confirm what the market already believed rather than announcing something new, and the absence of a front-end reaction is not by itself a rebuttal. Back-end evidence is the price decline when the truth emerged, offered on the theory that the correction removed what the misstatement had maintained. The courts of appeals have accepted that theory: FindWhat Investor Group v. FindWhat.com, 658 F.3d 1282 (11th Cir. 2011), held that confirmatory misstatements that prevent a price from falling are actionable; Glickenhaus & Co. v. Household International, Inc., 787 F.3d 408 (7th Cir. 2015), and In re Allstate Corp. Securities Litigation, 966 F.3d 595 (7th Cir. 2020), applied the price-maintenance theory while insisting that the back-end evidence be examined for other explanations. Waggoner v. Barclays PLC, 875 F.3d 79 (2d Cir. 2017), held that plaintiffs need not offer direct event-study evidence of price impact to invoke the presumption for a security trading in a highly efficient market, and that defendants must rebut by a preponderance.

Rebuttal in practice

Defendants have succeeded where the evidence showed the price did not respond to the actionable statements at either end. In IBEW Local 98 Pension Fund v. Best Buy Co., 818 F.3d 775 (8th Cir. 2016), the price rose on an earlier statement that was not actionable and did not move on the later statements that were, and the Eighth Circuit held the presumption rebutted. In Arkansas Teacher Retirement System v. Goldman Sachs Group, Inc., 77 F.4th 74 (2d Cir. 2023), on remand from the Supreme Court, the Second Circuit decertified the class because the generic statements about conflicts controls did not match the specific enforcement disclosures whose price declines were the back-end evidence. On remand in Halliburton itself, the district court examined price impact disclosure by disclosure and found it for only one of the alleged corrective dates. The lesson for the economist is that price impact is assessed statement by statement and date by date, and the report has to pair each alleged misstatement with the disclosure said to correct it.

Loss causation

Loss causation is the causal connection between the misstatement and the economic loss. Dura held that an inflated purchase price is not itself a loss, because the investor who sells before the truth emerges recovers the inflation in the sale price; the loss occurs when the truth is revealed and the price falls, and only the part of the fall attributable to that revelation is recoverable. The PSLRA places the burden on the plaintiff. The courts of appeals have recognized several routes to the showing: a corrective disclosure that reveals the falsity of the statement; the materialization of a risk the misstatement concealed, which the Second Circuit described in Lentell v. Merrill Lynch & Co., 396 F.3d 161 (2d Cir. 2005), and applied in In re Vivendi, S.A. Securities Litigation, 838 F.3d 223 (2d Cir. 2016); a series of partial disclosures that together reveal the truth, as in Public Employees’ Retirement System of Mississippi v. Amedisys, Inc., 769 F.3d 313 (5th Cir. 2014); and leakage of the truth over time, which Glickenhaus accepted on a proper model and In re Williams Securities Litigation, 558 F.3d 1130 (10th Cir. 2009), rejected where no evidence tied the decline to disclosure of the fraud.

What the economist produces for loss causation is the event study’s disclosure-day analysis: the abnormal return on each alleged corrective date, its statistical significance, the catalogue of other information that reached the market that day, the apportionment of the decline among the corrective information and the confounders, and the resulting fraud-related dollar decline per share. Whether a given disclosure corrected the alleged misstatement is a legal characterization; the report establishes what was disclosed, when, whether it was new, and what the price did. The hub’s loss causation section lists what the report states for each day, and event studies in Rule 10b-5 litigation describes the method.

How one event study answers both, and where the answers diverge

The same back-end decline is evidence for both questions, but it is used differently. For price impact the question is whether the decline is attributable to correction of the specific statement at all; the size of the decline matters less than its existence and its fit. For loss causation the question is how much of the decline is attributable to the corrective information; the size is the answer, and it is also the first input to damages. Three situations show the divergence.

• A statement that produced no front-end reaction, followed by a significant back-end decline when it was corrected, is a price-maintenance case: price impact is supported at the back end, and loss causation turns on the apportionment of that decline.
• A significant back-end decline that is mostly explained by an earnings miss or industry news released the same day supports price impact, because some of the decline is tied to the correction, but leaves a small recoverable loss, because most of it is not.
• A generic statement paired with a specific disclosure may fail price impact under Goldman even though the disclosure day shows a large, significant, fraud-related decline that would satisfy loss causation if the class were certified.

The first situation is common and favors plaintiffs at certification; the second is the usual shape of a defense damages argument; the third is why the pairing of statements and disclosures has become the center of the price-impact fight.

What each side’s expert must show

Where each question sits in the case

1. Motion to dismiss. Loss causation must be plausibly pleaded under Dura; price impact is not at issue. A preliminary event study on the alleged corrective dates, from public data, tells counsel before filing whether the pleaded dates carry significant declines and how much competing news there was.
2. Class certification. Market efficiency and price impact, with defendants bearing the burden to rebut; loss causation and materiality are not required. The report commits to the event dates, the pairing of statements and disclosures, and the damages methodology.
3. Summary judgment and trial. Loss causation and materiality, disclosure by disclosure, on the same event study; reliance returns as a merits element.
4. Damages and settlement. The fraud-related declines become the inflation ribbon and the per-share and class-wide figures described in calculating class-wide securities damages.

A worked illustration

The figures are invented and describe no company. On the first day of the class period the issuer stated that its underwriting standards were rigorous and consistently applied; the stock closed at $60.00, and the abnormal return that day was +0.2 percent with a test statistic of 0.1. Three hundred trading days later a regulator announced an enforcement action over underwriting practices, and on the same morning the company reported quarterly earnings below guidance. The stock fell 11.0 percent; after the market and industry moves are removed, the abnormal return is −9.5 percent, with a test statistic of −4.4.

For price impact, the front end shows nothing, which is consistent with a statement that confirmed existing expectations rather than one that moved the market. The back end shows a significant decline on the day the practice was revealed, which supports price impact on a price-maintenance theory, provided the statement about underwriting standards is specific enough to be matched with an enforcement action over underwriting practices; if the statement were instead a generic assurance about compliance culture, the defense would argue the Goldman mismatch, and the pairing becomes the dispute.

For loss causation, the question is how much of the −9.5 percent belongs to the underwriting disclosure. The company’s history of reactions to earnings misses of this size, intraday prices after each release, and analyst commentary point to about −3.0 percent for the earnings news, leaving −6.5 percent, or $3.90 per share at the prior close, for the corrective disclosure. The recoverable decline is $3.90, not the $6.60 the raw 11 percent implies, and $3.90 is the first step in the inflation ribbon. The same day, read for the two questions, gives a yes on price impact and a number for loss causation, and neither answer would be available from the raw price chart.

Where to start

If a class motion is ahead, a price-impact assessment from public data before briefing tells both sides which dates will hold; start with a conflict check using party names only. If the other side has served an event study or a price-impact report, send it to us for review with its workpapers. For a worked example of the disclosure-day analysis on public data, see the Compass Diversified event study; for the stages in order, Rule 10b-5 class actions: the economist’s role at each stage; and for what counts as a misstatement or omission in the first place, our note on misstatements and omissions.

Frequently asked questions

What is price impact?

Whether an alleged misstatement affected the security’s price, either by moving it when the statement was made or by maintaining an inflated price until the truth emerged. It is the premise of the fraud-on-the-market presumption of reliance, and defendants may rebut the presumption at class certification by proving its absence.

What is loss causation?

The causal link between the misstatement and the investors’ economic loss: under Dura, the part of the price decline that was caused by the revelation of the truth rather than by anything else. Plaintiffs bear the burden of proving it on the merits.

Can the same event study prove both?

Yes, and it usually does. The disclosure-day analysis that measures the fraud-related decline for loss causation is the back-end evidence of price impact. The two questions read the same result differently: price impact asks whether a decline tied to the correction exists, loss causation asks how large it is.

What is price maintenance?

The theory that a misstatement kept an already inflated price from falling rather than raising it, so that the price impact appears when the truth emerges rather than when the statement was made. Courts of appeals have accepted it; its proof is the back-end decline, paired with the statement it corrects.

Must plaintiffs prove loss causation at class certification?

No. Halliburton I held that loss causation is not a prerequisite to invoking the presumption of reliance. Defendants may, however, offer price-impact evidence at certification that overlaps with loss causation, and after Allstate and Goldman courts must consider it for the price-impact question even though it bears on merits issues as well.