United States v. Michael Gluk

831 F.3d 608, 101 Fed. R. Serv. 85, 2016 U.S. App. LEXIS 14343, 2016 WL 4150901

Opinion

ON PETITION FOR REHEARING

E. GRADY JOLLY, Circuit Judge:

The petition for panel rehearing is GRANTED, the original panel opinion (presently available at 811 F.3d 738 (5th Cir.2016)) is hereby withdrawn, and this opinion is substituted therefor. 1

Michael Baker and Michael Gluk appeal their convictions for securities fraud. Because we agree with their evidentiary challenges, we vacate their convictions and remand for a new trial. 2

I.

Michael Baker and Michael Gluk were, respectively, the CEO and CFO of Arthro-Care, a medical device company. Under their tenure (and, allegedly, with their knowledge) ArthroCare practiced “channel stuffing” with a related entity, DiscoCare.

“Channel stuffing” is a fraudulent scheme companies sometimes attempt, in an effort to smooth out uneven earnings— typically to meet Wall Street earnings expectations. Specifically, a company that anticipates missing its earnings goals will agree to sell products to a coconspirator. The company will book those sales as revenue for the current quarter, increasing reported earnings. In the following quarter, the coconspirator returns the products, decreasing the company’s reported earnings in that quarter. Effectively, the company fraudulently “borrows” earnings from the future quarter to meet earnings expectations in the present. Thus, in the second quarter, the company must have enough genuine revenue to make up for the “borrowed” earnings and to meet that quarter’s earnings expectations. If the company does not meet expectations in the second quarter, it might “borrow” ever-larger amounts of money from future quarters, until the amounts become so large that they can no longer be hidden and the fraud is revealed.

ArthroCare carried out exactly this fraud, with DiscoCare playing the role of coconspirator. Over several years, Arthro-Care fraudulently “borrowed” around $26 million from DiscoCare. This “borrowing” occurred by directing DiscoCare to buy products from ArthroCare on credit, with the agreement that ArthroCare would be paid only when DiscoCare could sell those products. Although this can be a legitimate sales strategy, it was fraudulent here because DiscoCare purchased medical devices that it knew it could not sell reasonably soon for the sole purpose of propping up ArthroCare’s quarterly earnings. This fraud was carried out under the day-to-day supervision of John Raffle, the Vice President of Strategic Business Units, and of David Applegate, another DiscoCare executive.

DiscoCare’s business model (apart from the accounting fraud) was potentially wrongful, though no charges were brought. DiscoCare provided a medical device for which most insurers refused reimbursement. To sell its device, DiscoCare reached agreements with plaintiffs’ attorneys in civil actions for personal injuries. These agreements resulted in the majority of DiscoCare’s sales. Under this agreement, DiscoCare would treat clients of the attorneys. The plaintiffs’ attorneys would then cite the expense of their clients’ treatment as a reason for defendants to settle personal injury lawsuits. DiscoCare also allegedly illegally coached doctors on which billing codes to use, in an effort to increase insurance reimbursements. This practice allegedly went as far as instructing doctors to perform an unnecessary surgical incision to classify the treatment as a surgery. No charges were filed on any of this conduct.

ArthroCare subsequently purchased DiscoCare for $25 million, a price that far exceeded its true value (DiscoCare had no employees at the time). During this purchase, the fraud began to unravel, with media reports alleging accounting improprieties. To reassure investors, Gluk and Baker made several false statements during a series of conference calls. As evidence mounted, the audit committee of ArthroCare’s board of directors commissioned an independent investigation by forensic accountants and the law firm La-tham & Watkins. As a result of this investigation, the board determined that Raffle and Applegate had committed fraud and that Gluk and Baker had not adequately supervised them. The board restated earnings, resulting in a significant drop in the value of ArthroCare stock. The board fired Raffle and Apple-gate for their roles in the fraud. The board also fired Gluk, determining that he had been remiss in not detecting the fraud earlier. Finally, the board fired Baker, determining that he should have implemented better internal controls.

The SEC investigated ArthroCare (both informally and formally) to determine the extent of the fraud. During this investigation, Raffle and Applegate exercised their Fifth Amendment right against self-incrimination to decline to answer questions. After its investigation, the SEC sued Arth-roCare, Raffle, and Applegate for securities fraud; it did not sue Gluk or Baker. It did file a “clawback” complaint against Gluk and Baker; this complaint stated that the SEC “does not allege that Baker and Gluk participated in the wrongful conduct” but instead determined that Raffle and Applegate “intentionally withheld” information from Gluk and ArthroCare.

The government subsequently brought criminal charges, initially only against Raffle and Applegate. Raffle and Applegate pled guilty and agreed to testify against Gluk and Baker; the government then indicted Gluk and Baker for the channel stuffing. At trial, Raffle and Applegate testified that Gluk and Baker knew of the fraud; Gluk and Baker testified that they did not. The key question for the jury was whether to believe Gluk and Baker or to believe the government.

The district court made several significant evidentiary rulings challenged on appeal. First, the defendants sought to introduce the Latham report, the SEC’s clawback complaint against Baker and Gluk, and two memos regarding the SEC investigation that the SEC had prepared for the DOJ. As discussed in more detail below, these memos both summarized the SEC investigation. The 2010 SEC memo stated that “Raffle and Applegate ... misled [Gluk] about whether certain Dis-coCare sales satisfied the company’s revenue recognition criteria.... Raffle also misled the company and its external auditor about the true [fraudulent] reason for certain product exchanges by DiscoCare and another distributor.” The 2011 memo expanded on the contents of the earlier memo and provided a somewhat more detailed summary of the investigation; this memo stated that Raffle and Applegate “orchestrated a scheme to materially misstate ArthroCare’s publicly reported revenue and earnings.”

According to Gluk and Baker, these documents would have corroborated their claim that they did not know of the fraud. Specifically, the documents would have shown that independent, neutral investigators determined that Raffle and Apple-gate — and not Gluk and Baker — had carried but and concealed the fraud. Because no other independent testimony corroborated the defendant’s version of events, they argued that this evidence was essential to their defense. The district court disagreed, and excluded all these documents as more prejudicial than probative.

The district court’s second important ev-identiary ruling concerned evidence of uncharged misconduct. Specifically, the government sought to introduce testimony about the uncharged medical fraud that allegedly took place at DiscoCare. The district court allowed this testimony, over objection.

The jury returned a guilty verdict. At sentencing, the court determined that Baker must forfeit his net proceeds (a different amount than the proceeds directly traceable to the fraud, see note 11 below) from selling ArthroCare stock during the period of the fraud, an amount equal to $22,165,030.78. This appeal followed.

II.

Gluk and Baker argue that the district court’s evidentiary rulings were incorrect in two ways: the rulings kept evidence out that should have been let in, and it let in evidence that should have been kept out. We agree on both counts, and accordingly reverse the defendants’ convictions.

We review the district court’s evidentiary rulings for abuse of discretion, subject to harmless error review. United States v. El-Mezain, 664 F.3d 467, 494 (5th Cir. 2011).

A.

We first consider the district court’s exclusion of the SEC documents. First, were the documents hearsay? The parties agree that the SEC documents fit the definition of hearsay. See Fed. R. Evid. 801. The defendants, however, argue that the documents are nonetheless admissible for their truth because of the 803(8)(iii) hearsay exclusion. Rule 803(8) provides that “[a] record or statement of a public office [is admissible] if: (A) it sets out: ... (iii) in a civil case or against the government in a criminal case, factual findings from a legally authorized investigation; and (B) the opponent does not show that the source of information or other circumstances indicate a lack of trustworthiness.” Fed. R. Evid. 803 (emphasis added). 3

The government responds that SEC documents did not set out “factual findings from a legally authorized investigation.” In support of this argument, the government cites Smith v. Isuzu Motors, Ltd., 137 F.3d 859 (5th Cir. 1998). In Isuzu Motors, we carved out a narrow restriction to 803(8)(iii) 4 : we held that otherwise-qualifying agency reports are not “factual findings from a legally authorized investigation” if they “embody the positions and opinions of individual staff members [of an agency], which the agency ultimately declined to accept.” Id. at 862. Thus, under Isuzu Motors, when an agency disavows (“declines to accept”) a report prepared by a staff member, that report does not qualify for the 803(8)(iii) exclusion. Conversely, if the agency takes no action, then a report prepared by a staff member in the ordinary course of duty and circulated outside the agency is exactly the sort of “factual findings from a legally authorized investigation” that 803(8)(iii) is designed to exclude from the prohibition on hearsay.

Further, the facts of Isuzu Motors shed light on what it means for an agency to “decline to accept” the findings in a document. In Isuzu Motors, a member of Congress “asked [the National Highway Traffic Safety Administration] to establish stability standards for certain types of passenger vehicles.” Id. at 862. In response to that request, NHTSA staff members prepared three initial memos that endorsed the Congressman’s position. After further consideration, however, “[t]he NHTSA ultimately rejected the ... petition.” Id.

We held that this, subsequent rejection established that the NHTSA had “declined to accept” the findings in the memos and thus that the memos were not “factual findings” that could be ascribed to the agency. Id. Thus, Isuzu Motors applies where the agency has taken some affirmative action to disavow the findings contained in a document — such as issuing a subsequent determination rejecting the findings of the earlier document.

Turning to the facts of this case, the SEC documents are covered by rule 803(8) — and are therefore admissible for the truth of their contents — unless record evidence shows that the SEC “declined to accept” the positions in those documents or otherwise disavowed the findings in any way.

We' begin with the 2010 SEC memo. 5 The heading of this memo states that it is from “SEC” and to “Main Justice,” a synonym for the DOJ. Based on this heading, the memo appears to be an official communication from the SEC to the DOJ. Thus, it seems to be a “factual finding” of the SEC and therefore to be admissible under rule 803(8)(iii).

Nevertheless, the government argues that the report is not admissible under 803(8)(iii), as interpreted by Isuzu Motors. According to the government, the 2010 memo was not “approved by the Commission” at its highest levels, and thereby is not a “factual finding” by the SEC. This argument, however, gets the Isuzu Motors analysis precisely backwards. Isuzu Motors does not allow agency reports only when those reports are endorsed at the highest level. Instead, Isuzu Motors only forbids agency reports when they have been disavowed in some way. Here, no evidence suggests that the SEC disavowed the contents of the 2010 memo; consequently, that memo represents “factual findings from a legally authorized investigation” and is admissible under rule 803(8)(iii).

The same analysis applies to the second memo from the SEC to the DOJ that Gluk and Baker seek to introduce. 6 Unlike the first memo, the heading of the second memo states that the memo is from an individual- — SEC attorney Jim Etri — rather than from the SEC as an entity. Etri, however, was the author of both memos and no evidence suggests that the SEC repudiated the second memo’s findings or that they were made outside of Etri’s capacity as an SEC attorney.

Finally, the SEC’s clawback complaint unambiguously represents an official action of the SEC. In the case of all three SEC documents, they were transmitted or filed by an SEC attorney to others outside the SEC. When an agency professional transmits a document to others outside the agency, that document is presumptively a factual finding of the agency. That presumption could be easily rebutted by evidence that others in the agency, such as superiors, had disavowed the contents of the document or otherwise “declined to accept” those contents. The government has not presented any evidence that any SEC employee “declined to accept” the contents of the SEC memos or the SEC clawback complaint.

Accordingly, we hold that all three SEC documents are admissible for the truth of their contents under rule 803(8)(iii). 7

B.

We thus turn to the question under Rule 403 of whether the district court abused its discretion in determining that that prejudicial effect of the SEC documents substantially outweighed their probative value.

The government argues that the reports would have improperly influenced and thus prejudiced the jury in performing its duties because the SEC examined no more information than the jury. According to the government, the SEC was essentially a fact-finding body, no more capable than the jury of determining whether Gluk and Baker had committed accounting fraud. The government worried that the “jury may have [incorrectly] believed that the SEC [was] better positioned to make factual findings”; that is to say that the jury may have been intimidated into blindly adopting the SEC’s conclusions when the jury’s fair judgment should be the sole determinant of guilt or innocence. But the jury is perfectly capable of weighing the evidence contained in the SEC documents against other evidence to the contrary and making an independent decision. Weighing evidence against other evidence is a core function of the jury, and we find no reason to be concerned that a properly instructed jury would improperly defer to the SEC’s findings.

Moreover, Gluk and Baker argue that the SEC documents were highly probative precisely because the SEC was better positioned to make factual findings and that professional findings would have been highly probative of the defendants’ culpability. SEC staff are experts in’ understanding and evaluating financial fraud. Administrative findings, the defendants assert, are admissible precisely because administrative expertise might aid the jury. The defendants’ arguments may be an overstatement, but we have touched on this issue before, in Smith v. Universal Services, Inc., 454 F.2d 154 (5th Cir. 1972).

In Smith, we held that EEOC reports, though not binding, are nevertheless admissible at trial. In effect, Smith held that an EEOC report can assist the jury in the same way an expert’s testimony would. “The fact that an [EEOC] investigator, trained and experienced in the area of discriminatory practices and the various methods by which they can be secreted, has found that it is likely that such an unlawful practice has occurred, is highly probative of the ultimate issue involved in such cases.” Id. at 157 (emphasis added). 8 Gluk and Baker point out numerous cases that have followed this same reasoning, both for the EEOC and other agencies. E.g., Hodge v. Seiler, 558 F.2d 284, 288 (5th Cir. 1977) (admitting a HUD report).

The government contends that these cases should be limited to the EEOC context and should not apply to SEC investigations. The government points out that, as this case demonstrates, financial-fraud cases can turn on credibility determinations — which is indeed the sole providence of the jury. But employment discrimination cases are equally likely to turn on credibility determinations; thus, the relevance of witness credibility provides no more reason to exclude SEC documents than to exclude EEOC reports. Moreover, investigations of employment discriminar tion and investigations of accounting fraud both typically involve complex legal intricacies where expert administrative guidance may assist the jury in weighing the evidence. In short, we see no relevant distinction between the SEC memos and an EEOC report; both are “highly probative.” Smith, 454 F.2d at 157.

The government presents a slightly different argument regarding the SEC claw-back complaint. The government argues that “charging decisions” are not highly probative because “many factors unrelated to guilt may influence those decisions and their admission therefore risks misleading the jury and confusing the issues.” United States v. Reed, 641 F.3d 992, 993 (8th Cir. 2011). If Gluk and Baker sought to admit the mere fact that the SEC had brought a clawback complaint, then the government’s argument would be correct — the simple fact that the SEC brought a clawback complaint instead of some other charge, is of very limited probative value.

In fact, however, Gluk and Baker do not seek to admit the charging decision, as the defendant in Reed attempted to do. Rather, Gluk and Baker want to admit the clawback complaint, which contains factual statements helpful to their defense. These statements have the same probative value as the similar statements in the memos. 9

We accordingly hold that the SEC documents are adequately probative in helping the jury weigh the credibility issues presented in this case. The district court enjoys great discretion under rule 403 to exclude evidence. See United States v. O’Keefe, 426 F.3d 274, 280 (5th Cir. 2005). Nonetheless, this discretion is not unlimited. Viewed correctly, the probative value of the SEC report exceeded any undue prejudicial effect. The jury was entitled to know that the SEC conducted an investigation, and concluded that Raffle and Applegate “mislead” the company, its accounting staff, and the CFO (Gluk) by “withholding information” and “lfying]”; the jury should also know that, after conducting this investigation, the SEC chose to “not allege that Baker and Gluk participated in the wrongful conduct” that occurred at ArthroCare. Of course, the jury was also entitled to hear that the government conducted an independent investigation and reached a different conclusion after securing the cooperation of Raffle and Applegate. Armed with all relevant information, the jury would then appropriately weigh the evidence and decide what to believe. Accordingly, we hold that the district court abused its discretion in determining that the SEC statements should be excluded under rule 403. 10

c.

The government next argues that, if the district court did err in excluding the SEC statement, the error was harmless. See United States v. El-Mezain, 664 F.3d 467, 525 (5th Cir. 2011). The government points out that the SEC documents barely mention Gluk and Baker. Yet this argument actually shows the harm of excluding the documents: The memos “provide[d] a high-level overview of the underlying facts” surrounding ArthroCare’s fraud. That a description of the ArthroCare fraud amounts to a chronicle of “the misdeed of John Raffle and David Applegate” is particularly relevant, as is the claim that Raffle and Applegate “orchestrated [the] scheme” and that “Baker and Gluk [were not alleged to have] participated in the wrongful conduct.” The very absence of Gluk and Baker from these documents actually supports the story they presented to the jury: that they did not know about Raffle and Applegate’s fraud. We are also unconvinced by the government’s argument that the error was harmless because the jury had access to all the evidence that formed the basis for the SEC report. That the SEC, a largely neutral investigator, had concluded that Raffle and Applegate were at the center of the fraud — and, indeed, had repeatedly lied to others at ArthroCare — could appropriately assist the jury in weighing the evidence presented at trial. To acknowledge a reasonable possibility, the jury could have ultimately accepted the SEC reports as casting a reasonable doubt on the government’s case. That is not harmless. Accordingly, we hold that the district court’s error in excluding the SEC documents was not harmless and we reverse and remand based on these errors.

D.

Finally, Gluk and Baker argue that the district court erred by admitting evidence of uncharged fraud that purportedly took place at DiscoCare. Baker argues that:

The [government’s] strategy, from evidence to argument, was clear. The government recognized an obvious truth: accounting fraud is bland. A straightforward attempt to prove an accounting fraud case would be difficult, both because the rules of accounting contain ample gray area and also because jurors might well be too bored to care. In order to convict, jurors need to be outraged, and few jurors are so moved by outsized accounts receivable and improper revenue recognition. In order to spark a sense of outrage, the prosecution went outside the charges proper. It went to the DiscoCare fraud and its lurid details of needless incisions performed at the behest of Florida ambulance chasers.

The defendants argue that this evidence was impermissible character evidence and, in any event, was more prejudicial than probative. This evidence of uncharged misconduct arguably creates the inference that Gluk and Baker were bad people involved in shady operations and thus were the sort of people who might have tolerated accounting fraud. The defendants strenuously contend that using this type of evidence to demonstrate the character of the defendant is impermissible under Federal Rule of Evidence 404.

The government, however, dismisses this argument, saying that activities at DiseoCare were intrinsic to the charges of wire fraud and were highly relevant. The government further argues that details about the activities at DiseoCare explain why Gluk and Baker would make misleading statements to investors (i.e., to hide those salacious details). Moreover, evidence that shows how involved Gluk and Baker were with the DiseoCare model demonstrates that they were involved in day-to-day operations of DiseoCare in its relationship with ArthroCare; this involvement is relevant to the credibility of their claim to have known nothing about Raffle’s and Applegate’s fraud.

While at least some evidence of the Dis-coCare conduct, is undeniably relevant to ArthroCare’s accounting fraud, the relevance of some limited evidence does not license the government to introduce the magnitude of testimony it elicited; nor does that limited relevance allow the government to emphasize the DiseoCare fraud, not chargeable to the defendants, in jury arguments. Allowing such breadth of testimony relating to salacious goings-on at DiseoCare was error.

We believe that the district court could have done more to police the line between proper and improper evidence; it could have been more careful to prevent the government from dwelling on the salacious details of DiscoCare’s business practices that could not be charged to the defendants. Because we reverse on other grounds, we need not determine whether this error independently justifies reversal or, conversely, whether it would have been harmless error in the absence of the reversible error we previously have identified.

III.

Accordingly, for the reasons stated, we VACATE Baker and Gluk’s convictions, and REMAND for a new trial. 11

VACATED and REMANDED.

1 . This opinion discusses the hearsay and harmless error issues in more depth; the revisions are largely contained in Part II.A-II.C.

2 . Because we reverse based on the evidentia-ry challenges, we do not reach the defendants’ other challenges to their convictions.

3 . If the documents contain factual findings that qualify for 803(8), they are not rendered "inadmissible merely because [the documents] state a conclusion or opinion. As long as the conclusion is based on a factual investigation and satisfies the Rule’s trustworthiness requirement, it should be admissible along with other portions of the report.” Beech Aircraft Corp. v. Rainey, 488 U.S. 153, 170, 109 S.Ct. 439, 102 L.Ed.2d 445 (1988).

4 . At the time Isuzu Motors was decided, the current 803(8)(iii) was 803(8)(C).

5 . This memo provides an overview of the SEC investigation and states that "Raffle and Ap-plegate also mislead the company's CFO [i.e., Gluk] about whether certain DiscoCare sales satisfied the company’s revenue recognition criteria.... Raffle and Applegate intentionally withheld this information from ArthroCare’s CFO to prevent the revenue from being reversed. ... Raffle also mislead the company and its external auditor about the true [fraudulent] reason for certain product exchanges by DiscoCare and another distributor.”

6 '. The second memo provides a slightly longer summary of the SEC enforcement. It "summarizes the misdeeds of John Raffle and David Applegate” and explains how "[t]hey orchestrated a scheme to materially misstate ArthroCare’s publicly reported revenue and earnings.” It also makes multiple references to Raffle’s "lies” to "ArthroCare accounting staff.”

7 . This hearsay exception does not apply to the Latham report; thus, that report is hearsay. Additionally, we express no opinion regarding whether some statements contained within the SEC documents could themselves be hearsay. If so, this second level of hearsay would be inadmissible — unless, of course, some other exception or exclusion applied to the hearsay within hearsay.

8 . The government, citing out-of-circuit law, argues that the expertise of the agency preparing a report does not increase the probative value of a report prepared by that agency. See Coleman v. Home Depot, Inc., 306 F.3d 1333, 1345 (3d Cir. 2002). This argument flatly contradicts Smith and we reject it.

9 . Nor do we believe that discussion of the complaint would necessitate undue digressions into the details of the Sarbanes-Oxley clawback framework. The government is free to point out that, to be entitled to a clawback, the SEC was not required to allege that the defendants had any role in the fraud. The defendants, in turn, are free to point out that the SEC went out of its way to state that it did "not allege that Baker and Gluk participated in the wrongful conduct.” The jury can then draw whatever conclusions from these arguments that it thinks best.

10 . Gluk and Baker also argue that the district court abused its discretion by excluding the Latham report. We disagree. As noted above, the Latham report does not qualify for a hearsay exception and therefore is admissible only for impeachment purposes. The limited admissibility of the Latham report significantly diminishes the probative value of the Latham report. Further, as the government points out, Raffle and Applegate testified at length about the Latham investigation, so the additional probative value of admitting the report is limited. Nevertheless, as Gluk and Baker note, the Latham report had significant impeachment value. Given this impeachment value, we are not prepared to hold that admitting it would have been erroneous. At the same time, excluding it was not an abuse of the district court’s considerable discretion.

11 . Because we reverse the convictions, we do not reach Baker's challenge to the forfeiture calculation. We note, however, that forfeiture is not a fine (despite being subject to the same Eighth Amendment limits). See United States v. Bajakajian, 524 U.S. 321, 328, 118 S.Ct. 2028, 141 L.Ed.2d 314 (1998). As we read the relevant forfeiture statutes, they exist to require defendants to give up the proceeds of their crimes, not to punish them for those crimes. See United States v. Hatfield, 795 F.Supp.2d 219 (E.D.N.Y. 2011). Requiring forfeiture of the entire value of stock sold would require forfeiting compensation, even when that compensation is not traceable to fraud.