City of Southfield v. Advance Auto Parts, Inc. - Securities Fraud Appeal

PUBLISHED UNITED STATES CO URT OF APPEALS FOR THE FOURTH C IRCUIT No. 25-1188 CITY OF SOUT HFIELD GENERAL E MPLOYEES ’ RETIREME NT SYSTEM, Plaintiff – Appell a nt, and MIGUEL SUAREZ; ROBERT BASEL; HANY MAGOUR; BRIAN M. WATSON, Plaintiffs, v. ADVANCE AUTO PARTS, INC.; THOMAS R. GRECO; JEFFREY W. SHEPHERD; WILLIA M J. PELLICCIOTTI, JR., Defendant s – Appel lees. Appeal from the United States District Court for the Eastern District of North Carolina, at Raleigh. James C. Dev er, III, District Judg e. (5:23−cv−00 563−D− BM) Argued: December 10, 2025 Decided: February 17, 2026 Before DIAZ, Chief Judge, GREGORY, Circuit Judge, and Gina M. GROH, United States District Judge for the N orthern District of West Virginia, sitting by designation. Affirmed by published opinion. Chief Jud ge Diaz wrote t he opinion, in which Judge Gregory and Judge Gro h joined.

2 ARGUED: Dana Ly dia Kaersvang, DEUTSCH HUNT PL LC, Washingto n, D.C., for Appellant. William Michael Regan, Allison Michele Wuertz, H OGAN LO VELLS US LLP, New York, New York, for Appellees. O N BRIEF: Christopher M. Wood, Nashville, Tennessee, Ashley M. Price, ROBBINS GELLER RUDMAN & DOWD LLP, San Diego, California; Robert N. Hunter, Jr., HIGGIN S BENJAMIN, PLLC, Greensboro, North Carolina; Hyland Hunt, Ruthanne M. Deutsch, DEUT SCH HUNT PLLC, Washingt on, D.C., for Ap pellant. Jacey Lara Gottlieb, Ch ristine Jha, HOGAN LOVELLS US LLP, New York, New York; Clifton L. Brinson, SMITH, ANDERSON, BLOUNT, DORSETT, MITCHELL & JERNI GAN, LLP, Raleigh, N orth Carolina, for Appe llees.

3 DIAZ, Chief Judge: Advance Auto Parts, I nc., is a public comp any that sells autom obile parts and accessories. In early 2023, it a nnounced aggressive finan cial goals that exceeded invest ors’ expectations, wh ich increase d the company’ s stock price. But throughout the year, Advance Auto significantly reduced those estimates and corrected a series of accounting errors. Its stock price p lummeted after the d isclosures. The City of Southfi eld General Emp loyees’ Retirement Syste m filed suit, contending that Advance Auto and several f ormer officers committed securities fra ud by manipulating the company’ s accounting. The district court dismissed the complaint for failing to establish scienter, or wrongful intent. W e agre e that Southfield’ s allegations don’ t satisfy our scienter stan dards. So we affirm. I. W e review de novo a dismissal under Federal Rule o f Civil Procedure 12(b) (6). Employees ’ Ret. Sys. v. Macr oGenics, Inc., 61 F.4th 369, 381 (4th Ci r. 2023). W e accept the plaintif f ’ s factual allegations as true and consider the complaint in its entirety. T ellabs, Inc. v. Makor Issues & Rts., Ltd., 551 U.S. 308, 322 (2007). A. In early 202 1, Thomas Greco, Advance Aut o’s then President an d Chief Executive Officer, publicly announced a three -year plan to i ncrease t he company’s sales a nd profit margins. Over the n ext two years, Greco and Jeffrey Shep herd (Advance Auto’s then

4 Executive Vice Presid ent and Chief Finan cial Officer) assured investors that the company was on track to achieve its goals. For example, Advance Auto to uted increased profit margins in the third quart er of 2022. And when the company reported its fourth-quarter performance that year, Greco told investors tha t Adv ance Auto had “positi ve momentum.” Join t A ppendix (“ J.A. ”) 26. Still, he acknowledged that 2022 had been a “challenging yea r” and that the “overall results did not meet expectations.” J.A. 258. He promised that the company would “tak[e] decisive actions to imp rove performance in 20 23.” J.A. 26. In the same report, Advance Auto issued its 2023 guidance. The company estimated that it would see net sales between $11.4 billion and $11.6 billion; operating income between $889 millio n and $95 1 mil lion; and an operating income m argin between 7.8% and 8.2%. Greco and Shepherd both publicly emphasiz ed their focus on e xpanding operating margins. Greco was “confident” Advance A uto would continue to grow its margins in 2023, and analysts thought the company’s forecast w as “better t han” expected. J.A. 32 – 33. Greco also announced t hat he would retire at the end of 2023 because the com pany was “in the final year o f [its] three - year strate gic business plan.” J.A. 26. But he assure d investors that “ by planning for retirement in advance,” he would be able t o “faci litate a smooth transition” for his successor. J.A. 26 – 27. The market reacted po sitively to these announcements, and Advanc e Auto’s stock price increased.

5 B. Advance Auto held its annual shareholder meeting in May 2023. In a non-binding vote, shareho lders ratified Greco ’s and Shepherd’s fiscal year 2022 compensati on. Both executives received a b ase salary and a bon us for meeting key fina ncial targets. The compan y allotted about $8.4 million to Greco, citing his “[s]trong multi - year performance ” and the fact that his “ salary had not been adjusted since joining the [c] ompany in 2016.” J.A 60. It awarded Shepherd about $2.6 million, which also included a raise for “[p]erformance [that] exc eeded individual obj ectives for 2 021. ” J.A. 61. C. But the positive momentum didn’t last. With in a we ek of the shareh older meeting, Advance Auto issued first -quarter results fo r 2023. Its performan ce was “well below expectations.” J.A. 3 25. The com pany also dis closed the first of several accoun ting e rrors: it identified a pproximately $17 million in costs that were “incurred in prior years but not expensed in the corresponding peri ods.” J.A. 29. Greco and Shepherd blamed the company’s “ poor performance on external factors. ” 1 J.A. 33. Greco warned i nvestors that he “expect[ed] the competit ive environment. . . to remain very challenging” throughout the year. J. A. 294. So Ad vance Auto amended its 20 23 guidance to reduce est imates for net sales, op erating income, an d operating margin. 1 In part, they attributed the company’s difficulties to an increa singly competitive market. But some anal ysts felt that this expla nation wa s “out of line ” with what Advance Auto’s competitors wer e reporting at the ti me. J.A. 34 – 35.

6 At this poi nt, t he comp any’s C hairman, Eugene Lee, took on an “ex panded role as interim executive chai r” to “pro vid[e] additional operational oversight and support to [Advance Auto ’s ] man agement team.” J.A. 62. When Advance Auto announced second-quart er results, it again revised its guidance to forecast steeper decline s in operating income and operating margin. On an investor call, Lee attribute d the decl ine in mar gins to the company ’s low as set productivity and t he structure of its distribut ion centers. Advance Auto then a nnounced that Grec o would r esign as President and Chi ef Executive Officer in September 2023, a few months earlier than planned. The compan y also reported that Shepherd had “separated f rom the company. ” J.A. 63. And William Pellicciotti, Jr., Advance Auto’s Controller and Chief Accounting Officer, resigned shortly after. D. Advance Auto had new managers in place when it published it s 202 3 third -quarter results. But the compa ny continued to underperform, which cause d it to again revise its 2023 guidance. Advance Auto also “identified additi onal errors impacting cost of sal es and selling, general and administra tive costs. ” J.A. 374. It determined tha t the se errors affected i ts 2022 financial re sults, so it issued a public re statement. But Advan ce Auto emphasized that it had “evaluat ed the errors and deter mined that the related impa cts were not mater ial to the previously issue d consolidated fina ncial statement s for any prior period.” J.A. 374.

7 When Advance Auto filed its Form 10-Q for the third quarter of 2023, it elaborated on the accounting errors. I n addition to the $17 milli on error d iscovered back in May, the company identified $1 0.2 million in pre viously undisclosed cos ts. Advan ce Aut o said the errors “primarily related to produc t returns and vendor credits.” J.A. 18. 2 By the end of 2023, Advance Auto’s operating income and margins were a fraction of its initial g uidance: its operating income totaled $114 mi llion inst ead of the originally estimated $889-951 million, and its operating margin was 1% instead of 7.8-8.2%. And in its 2023 Form 1 0- K, the compa ny corrected accounting errors in its financial state ments for 2021, 2022, and the first three quarters of 2023. The se errors a lso “primarily related to product costs and vend or credits. ” J.A. 89. In total, Advance A uto had understated its ex penses by about $ 100 million. The company promised to “devote [] significa nt time and resources to com plete its remediation of the material weak ness,” including by hirin g a ne w C hief F inancial Officer and Chief Accounting Officer. J. A. 69. 2 Advance Auto earns “vendor credits” when it returns a product to a vendor. Some vendors also offer “vendor incentives,” which act as rebates so that the more pr oduct Advance Auto purchas es, the less it pays for e ach individual unit. Although Advanc e Aut o attribut es its e rrors to vendor credits, Southfield claims that the company “improperly account[ed] for vendor incentives. ” Appellant’s Br. at 8. Advance Auto respond s that this claim mistakenly conflates the two concepts. Southfield, though, alleges that b oth credits and incentives serve as “ reductio ns to amounts owed and/or payments fro m vendors. ” J.A. 15. And at oral argument, South field’s counsel stated that Advance Auto receives vendor incentives as credits. Given the posture of the appeal, we view the al legations in the light most favorable to Southfield a nd accept that vendor incentives are linked to vendor credits. See In re Willi s Towers Wats on PLC Proxy Litig., 937 F.3d 297, 302 (4th Cir. 2019).

8 The corrections affecte d the company’s financial results only slightly: from the first quarter of 2021 through the third q uarter of 2023, Advance Auto ’s operating income decreased b y 4.5%, its cost of s ales increas ed by 0.3 9%, and its selling, general, and administrative expenses increased by 0.09%. Still, Advance A uto’s stock price droppe d in response to each disclosure. It fell by over 50% between Ma y 30, 2023, and N ovember 21, 2023. E. After multiple investors sued Advance Auto, the district court consolid ated the cases and designat ed Southfield as lead plaintiff. Southfield filed a class action complaint on behalf of itself and other s who purchased Ad vance Auto’s stock bet ween November 15, 2022 (when the company announced improve d margins), and November 20, 2023 (the day before the company disclosed additional acc ounting errors). The complaint alleged violations of SEC Rul e 10b-5, 17 C.F.R. § 240.10b-5, and Sections 10(b) and 20(a) of the Securitie s Ex change Act of 1934, 1 5 U.S.C. §§ 78j(b), 78t(a). It claimed that the defenda nts (Advan ce Auto, Greco, Shepherd, and Pellicciotti) intentionally or recklessly misre presented Advance Aut o ’s fina ncial results an d foreca sts. D efendants moved to d ismiss the complaint fo r failure to state a claim. The distric t court found that Southfield “plausibly allege d a connection between def endants’ al leged misrepresenta tion or omission and the purchase or sale of a security; (2) plaintiff’s reliance upon the misrepresentation or omission; (3) plaintiff’s economic loss; (4) and loss causation.” Suare z v. Advance Auto Part s, Inc., No. 5:23- CV - 563-D, 2025 WL 283690, at *6 (E.D.N.C. Jan. 23, 2025). Thus, Southfield plausibly

9 alleged that “defendants’ public statements amount to material misrepresentations or omissions.” Id. Even so, the district court granted d efendants’ motion to dismiss because Southfield hadn’t “create [d] the required strong inference of scienter against any individual defendant or Advance [Auto].” Id. at *8. It found the more plausible inference was that “defendants acted in good faith and corrected pu blicly available information each quarter.” Id. at *7. This appeal followed. II. We start — and finish — with scienter. 3 A. Section 10(b) of the Securities Exchange Act of 1934 and SEC Rule 10b -5 prohibit deception in connecti on with securities transactions. See 17 C.F.R. § 240.10b-5(b); 15 U.S.C. § 78j(b). This includes “ mak [ing] any untrue statement of a material fact ” or “ omit[ting] to state a material fac t necessary in order to make the state ments made . .. not misleading.” 17 C.F.R. § 240.10b-5(b). 4 3 Defendants also argue that their statements about the 2023 guida nce are either protected forwar d-looking stat ements or otherwise n ot false or misleading. The district court didn’t ad dress th ese argument s, and since we affirm for lack of scienter, we needn’t reach them. 4 Section 20(a) of t he Exchange Act “ provid es that ‘ controlling persons ’ can be vicariously lia ble ” for underlying securities violations. Karp v. First C onnecticut Ban corp, Inc., 69 F.4th 223, 236 (4th Cir. 2023) (citing 15 U.S.C. § 78t). So for § 20(a) to a pply, Southfield must first es tablish its § 10(b) claim. See id.

10 To state a § 10(b) claim, the plaintiff must plead “(1) a material misrepresentatio n or omission by the defendant; (2) scienter; (3) a connection between the misrepresentation or omission and the purchase or sale of a security; (4) reliance upon the misrepresent ation or omission; (5) ec onomic l oss; and (6) loss c ausation.” Sto neridge Inv. Partners, LLC v. Sci.-Atlanta, 552 U.S. 148, 157 (2008). For scient er, the plaintiff must show t hat the defendant acted with “a mental state embracing intent to de ceive, man ipulate, or defraud.” Tellab s, 551 U.S. at 319 (citation omitted). Recklessne ss can s uffice if the defendant’s behavior was “so highly unreas onable and such an extreme departure from the standard of ordinary care as to present a danger of misleading the plaint iff [,] to the extent that th e danger was either kn own to the defendant or so obvious that the defendant must have been aware of it.” Pub. Employees’ R et. Ass’n v. Deloitte & Touche LLP, 551 F.3d 305, 313 (4th Cir. 2 009) (citatio n modified). The Private Securitie s Litigation Reform Act imposes a heighte ned pleadin g standard “ [a]s a check against abusive litigation by private parties” in securities fraud case s. Tellabs, 551 U.S. at 313. A complaint must “state with particularity facts giving rise to a strong inference that t he defendan t acted with t he required state of mind.” 15 U.S.C. § 78u – 4(b)(2). “Evaluating the strengt h of an inference is necessarily a com parative inquiry.” Yates v. Mun. Mortg. & Equ ity, LLC, 744 F.3d 874, 885 (4 th Cir. 2014). “ [A ]n i nference of scienter can only be strong. . . when it is weig hed against the opposing inferences that may be drawn from the facts in their entirety.” Cozzarelli v. Inspire Phar m. Inc., 549 F.3d 618, 624 (4th Cir. 2008).

11 When weigh ing the i nferences, “the c ourt’s j ob is not to scrutinize ea ch allegation in isolation but to assess all the allegations holistically.” Tellabs, 551 U.S. at 326. We “ must compare the malicious and innocent inferences cognizable from the facts pled in the complaint, and only al low the complaint to s urvive a motion t o dis miss if the malicious inference is at least as compelling as any opposing innocent inference.” Yates, 744 F.3d at 885 (citation omitted). But i f the more compe lling inference is tha t the defendants acted i nnocently — or even negligently — we must affirm the district court. See Pub. Employees’ Ret. Ass’n., 551 F.3d at 313. B. Although we ult imately evaluate Southf ield’ s allegations holi stically, w e begin by assessing them individ ually. See KBC Asset Mgmt. NV v. DXC T ec h. Co., 19 F.4th 601, 608 (4th Cir. 202 1). 1. Southfield first argues that Greco and Shep herd were motiv ated to conceal the accounting err ors until after shareholders appr oved their “large pay packages.” Appellant’ s Br. at 28. A llegations t hat a defendant “would p ersonally benefit from a special bonus” may support scienter. Boykin v. K12, Inc., 54 F.4th 175, 186 (4th Cir. 2022). Even so, “motivations to... inc rease one’ s own c ompensation are co mmon to every compa ny and thus add little to an inference of fraud.” Cozzarelli, 549 F.3d at 627.

12 Here, the allegations don’ t show a malicious motive. Shareh olders vot ed on Greco ’s and Shepherd’ s pay pa ckages only about a week before Advance Auto disclosed the fir st round of accoun ting errors. But they did so in a non-bindin g vote t o ratify compensation that Greco and Shepherd had already earned. Southfield hasn’t alleged that the executives would have been paid any less if the shareholders had voted no. So we can’t infer that th e shareholder vote m otivated Greco and Shep herd to conceal the accounting error s. Moreover, Advance Auto’ s B oard calculated Greco ’s and Shepherd’ s pay packag es in part based on the co mpany’ s perf ormance in 2021. The Board cit ed Greco’ s “[s]trong multi -year performanc e, including development and execution of a new strategic plan that [led to] record results in 2021,” and credited Shepherd’ s “[p]erformance [that] excee ded individual objectives for 2021.” J.A. 60–61. So at least some of the compensation was for performance predatin g the start of the class period. And while “allegations of unusual stock sales are not required” to establish motive, Zak v. Chelsea Therap eutics Int’ l, Lt d., 780 F.3d 597, 607 (4th Cir. 2015), Greco and Shepherd lost money by holding onto their Advance Auto stock. That, too, militates against a strong inference of scienter. See In r e PEC Sols., Inc. Sec. Litig., 418 F.3d 379, 390 (4th Cir. 2005) (no scienter where defendants “lost over $471 million dollars in collective stock value during the class p eriod” (emphasis o mitted)). 2. Southfield next argues that defendants knew a bout the errors be cause their “laser - focus[]” on Advance Auto’ s margins meant that they were also focus ed on vendor -incentive accounting, which affected those margins. Appellant’ s Br. at 30 –32.

13 But Southfield needed to allege fa cts sho wing that defendants “knew the missing information [,]... knew that the information was relevant [,]... [an d] went ahead and left the information out anyway, with th e intent to misle ad [p] laintiffs—or at least with a reckless disregard for the risk that leaving the inf ormation ou t would make their [statements] misleading.” San Antonio Fir e & Police Pension Fund v. Syneos Health Inc., 75 F.4t h 232, 241–42 (4th Cir. 2023). Southfield hasn’t sufficiently alleged that defendants “had reason to know that the company’ s financial information was inaccur ate” when they p ublished Adva nce Au to’ s financial results. See Matrix Cap. Mgmt. Fund, LP v. Beari ngPoint, Inc., 576 F.3d 172, 188 (4th Cir. 20 09). T he complaint fails to allege facts about when, or how, defendants learned about the accounting errors. So the stronger inference is that defendants based the financial statem ents on internal data they believed was correct at the time, but which they later learned was wron g. C f. Zak, 780 F.3d a t 608–10 (finding scie nter where plaintiffs identified contradictory documents that def endants had seen before making public misstatements). And while defendants may have been focused on margins, Southfield ha sn’ t connected t hat foc us t o a know ledge o f ven dor incentives or the company’ s account ing practices. For exa mple, Southfie ld doesn’t allege that Greco, She pherd, or Pellicciotti made public statemen ts conc erning v endor incent ives, whic h might have suggested that they were reviewing that metric. And we c an’ t assume that defendan ts monitored vendor incentives simply because they were senior executives. See Syneos, 75 F.4th at 242 (“[W] e

14 cannot impute factua l knowledge to indi viduals merel y based on their profe ssional position.”). 5 Southfield esse ntially asks us to infer that because defendants wer e focus ed on margins, they were clo sely monitorin g ot her metric s that c ould affect margins. And that we should then “ infe r f r om that infer ence that [defendants] acted wit h scienter.” Maguir e Fin., LP v. Po werSecur e Int’l, Inc., 876 F.3d 541, 54 7 (4th Cir. 2017). But “ [a] plaintiff may not stack i nference upon infer ence to sat isfy the [Reform Act’ s] pleading standard.” Id. The “strong infere nce of scienter ” must “be supported by facts, not other inferences.” Id. The complaint does allege that Advance Auto identified vendor incentives as a “Critical Accounting Polic[y].” J.A. 27 –28. I n a company statement, Advance Auto said that it “ regularly revie w[s] the receivables from vendors to ensure they are able t o meet their obligation s” and that “[p]eriodic assess ments of the [amounts accrued fro m vendors] are performed to determine the appropriateness of the estimate and are adjusted for accordingly.” J.A. 28. 5 One m ight e xpect Pellicciotti, a s Ch ief Acc ounting Officer, to ha ve been more focused on the accounting information. But Southfield do esn’t all ege that, nor does it plead what Pellicciotti knew or how he may have c ontributed t o the error s. Without more, we can’t infer scienter. See Matrix, 576 F. 3d at 190 (explainin g that a plaintiff must allege particular facts s upporting that “ each in dividual defendant ” acted with sciente r (citation modified)).

15 But this stateme nt doesn’t specify wh o par ticipates in the revie w process or otherwise ex plain how the policy implicates the individual defendan ts. These “omis sions and ambiguities count against infer ring scienter.” T ellabs, 551 U.S. at 326. Perhaps given their professed interest in mar gins, defendants should have paid more attention to t he compa ny’ s accounting. But negligence isn’t enoug h to sustain a § 10(b) claim. See Pub. Employees’ Ret. Ass’n, 551 F.3d at 314 (“ [T] o establish a strong inference of scienter, plain tif fs must do more than merely demon strate that defendants should or could have done more.”). 3. Southfield also argues that the acc ounting errors—whi ch totaled a pproximately $100 million in losses over three years— were so large that defe ndants must have kn own about them. W e may give “[i]nfere ntial weight” to the magnitude of the e rrors, b ut “only in the context of [a defe ndant’ s] financial positi on.” Matrix, 576 F.3d a t 184. “As a general matter, courts should be wary when defendant s focus on the size of revenues in an effort to minimize the materiali ty of mis statements of income or to su ggest the defendant lacke d motive.” Id. That said, it’ s reasonabl e to assume that “an individual is more likely to realize that she is missing $10 if she has $50 in her bank account than if she has $50,000 in h er bank account.” Id. So to determine whether Advan ce Auto must have known about the errors, we compare the resulti ng $100 million l oss with the company’ s revenue.

16 Losing $100 million over three years is certainly significant in the abstract. But for a compa ny generating about $ 33 billion in revenue over that same p eriod, it ’ s a relatively small sum. I ndeed, the accounting correc tions increased Advance Auto ’s cost of sales by only 0.39% and its se lling, general, an d administrativ e expenses by 0.09%. V iewed holistically, the se facts don’t support a strong inference of scienter. S ee Matrix, 5 76 F.3d at 184–85 (declining to give $100 million w orth of errors inferenti al weight when the company made over $823 mil lion per q uarter); see also Y ates, 744 F.3d at 889 ($44.9 million loss was “relati vely small” compared to the company’ s adjust ed shareholder equity of $723 million). 4. Southfield next alleges that Advance A uto manipulat ed its accou nting in the past, raising the l ikelihood that defendan ts acted with scienter here. Southfield relies on an anonymous former employee who asserts that in 2021, Advance Aut o “ stockpil[ed] vendor incentives, only to later release them ” to boost product ma rgins. J.A. 15. The form er employee “believed everyone in [the company’ s] senior manage ment team had to be aware of the rebate release issue.” J.A. 31. The employee allegedly reported their concerns to their supervisor and rec eived an email from a Seni or V ice President of Finance admitting that Advance Auto was “m asking product margin through its handling of vendor incentives. ” J.A. 32. “Generally, plaintiffs are free to use the allegations of confidential witnesses to support an inference of scienter.” KBC, 19 F.4th at 609. But “such allegations will only be afforded the weight they are due given t heir indicia of re liability.” Id.

17 We give little weight to this anonymous witness. The employee left Advance Auto in 2021, well before th e start of the class peri od. Se e T eachers’ Ret. Sys. v. Hunter, 4 77 F.3d 162, 179 (4th Cir. 2007) (finding co nfidential witness unreliable given their “lack o f familiarity” with the r elevant c onduct). Nor is there an y in dication that the employe e “passed their concerns on to [the individual defendants] or that the individual [d] efendants were otherwise aware of the problems a lleged.” KBC, 19 F.4th at 609; see also Nolte v. Cap. One Fin. Corp., 390 F.3d 31 1, 316 (4th C ir. 2004) (affirming dismissal in part because plaintiffs didn’ t allege “that manag ement was ever informe d of” e mployee concern s). Southfield alleges that the employee discussed the issue with their direct supervisor and a Senior V ice President of Finance, neither of whom is a defendant here. The employee’ s mere belief that the rest of the senior management team knew about the issue is too speculati ve to support scienter. See Y ates, 744 F.3d at 887 (co nclusory state ments about a defendant’ s stat e of mind are insufficient); see also KBC, 19 F.4th at 609. 5. Southfield next asks us to infer sc ienter becau se Greco, She pherd, and Pellicc iotti left Advance Auto with in months of the first c orrective disclosure. “ But executive departures are, at best, weak evidence of scienter.” Syneos, 75 F.4th at 244. So Southfield needed to bols ter the departures with “alleg ations demonstrating [d] efendants’ contemp oraneous knowledge” of the accounting errors for us to gi ve this factor greater weight. In r e T riangle C ap. Corp. Sec. Liti g., 988 F.3d 743, 754 (4th Cir. 2021). This Southfield failed to do.

18 Advance Auto did acknowledge that hiring a new Chief Financial Officer and Chief Accounting Officer w as part of its ef fort to remediate the accounting errors. B ut while this statement reinforces “ the substantial accou nting challenges the [c] o mpany then faced, it does not compel an inference that... the... individual defendants were bent on committing fraud.” Y ates, 74 4 F. 3d at 889. That the indivi dual defendants may h ave taken the fall for failing to catch the acc ounting errors does n’ t mean that w e can infer fraudulent intent. 6. Southfield also claims t hat Advance Auto’ s financial restatements, which corrected nearly three years’ worth of financial results, are “indicative of scienter.” Appellant’ s B r. at 38. According to S outhfield, by issuing the corrections, Advance Auto acknowledge d that its or iginal statements didn’t comply with G enerally Accepted Ac counting P rinciples (“GAAP”). But “ [t] he m ere m isapplication of accountin g principles,” witho ut corroborating evidence of frau dulent intent, doesn’t establish scienter. PE C Sols., 418 F.3d at 389 (citation omitted); see also Svezzese v. Durate k, Inc., 67 F. App’x 169, 173 (4th Cir. 2003). As we’ ve explained, $100 million in undisclo sed los ses over three years is n’t particularly dramatic for a compan y that generated abo ut $33 billion in reve nue over the same perio d. Nor do the alleged GAAP violations a dd new material facts. Southfield argues that Greco and Sheph erd “ would have regu larly r eviewed” the orig inal financial statements. Appellant’ s Br. at 38. But that still doesn’t gi ve rise to a strong i nference that defendants had access to the correct accounting info rmation and deliberately or recklessly ign ored it when publishing the original statements. W ithout that missi ng link, this clai m “simply

19 rides around in circles on the inadequate co attails ” of Southfield’ s other allegations. PEC Sols., 418 F.3d at 390. 6 7. Finally, Southfield ar gues that Advance Auto ’ s first correction to the 2023 guidance is “suspi cious” because it cam e just months after the company i ssued the initial rosy forecast. Appellant’ s Br. at 42. “[T] emporal proximity between... allegedly false statements or omissions and the subsequent di sclosure of the truth” is someti mes “ relevant to the scienter inquiry.” KBC, 19 F.4th at 612. A “short time between the two statements [may] s uggest that the speaker must have known the truth when they uttered the falsity.” Syneos, 75 F.4th at 244 (citation modified). But that’ s only so “when the temporal proximity betwe en false and truthful statements is very clos e [,]... otherwise, the link between a n ear lier misstatement and a later revela tion is purely speculative. ” Id. (cit ation modified); see al so KBC, 19 F.4th at 613 (declining to “dra w a str ong inference of scienter from the proximity of [defend ant’ s] optimistic statements i n the summer and its m ore sobering news in th e fall”). 6 The district court fou nd that the restatement s were evidence of defendants’ good faith efforts to correct the errors. We ’re n ot willing to go that far, s ince two o f the three disclosure s we re made after Greco, Sheph erd, and Pellicciotti left Advance Auto. And the other disclosure didn’t alert investors to a ny specific problem s with the company’ s accounting, while ins isting that “the cumulative impact was not ma terial to the current period or any previously issued fina ncial statements. ” J.A. 351. D isclosures that “continue[] to represent the reliability of [the company’ s] fi nancial information notwithstanding internal control problems... do not lend themselves to any inferences one way or the other relatin g to scienter.” Mat rix, 576 F.3d at 187.

20 Here, Advance Auto issued its 2023 guidance in February and revised it for the first time in May. “[T]he months -long gap between the st atements is too wide to support a n inference of scienter. ” Syneos, 75 F.4th at 244. C. Even though none of Southfield’ s allegations support a s trong infere nce of scient er alone, we still need t o consider whether, “evaluating the complaint holistically, the combined allegations can do what their individual parts failed to do.” KBC, 19 F.4th at 613 (citation mo dified). “A llegations of scienter that would n ot indepe ndently create a s trong inference of sci enter might [complement] each other to create an inference of su f ficient strength to satisfy the [Reform Act ].” Matrix, 576 F.3 d at 187–88 (cit ation modified). Not so here. Southfie ld alleges that (1) Greco and Shepherd received large pay packages before identifying the first account ing error; (2) the two executives also publ icly emphasized improving mar gins, which the accounting err ors af fected; (3) the errors totaled about $100 milli on; (4) an anonymous witness claimed that Advance Auto had committed prior accounting mischief; (5) the individual defendants left the company within months of the first correct ive disclosure; (6) Advance Auto corrected the errors through s everal restatements, the bulk of which were issued after the indivi dual defendants left the company; and (7) Advance Auto revised it s 2023 guidance three m onths after issuing it. T aken toget her, these facts may support a plausible inference of scienter. But that won’t do. See T ellabs, 551 U.S. at 3 14. Even though the re may be “some reason to believe [d] efendants ac ted wit h the requisite intent... there is not en ough to create a str ong

21 inference of scienter.” KBC, 1 9 F.4th at 613. So the distric t court correctly dis missed the § 10(b) claim. And because Southfield’ s § 10(b) claim fail s, so does i ts claim for vicarious liability under § 20(a). See Karp, 69 F.4th at 236. AFFIRMED