Opinion
3:21-cr-00353-CRB-1
01-25-2024
UNITED STATES OF AMERICA Plaintiff, v. MANISH LACHWANI, Defendant.
ORDER ON APPLICATION OF SENTENCING GUIDELINES
CHARLES R. BREYER, UNITED STATES DISTRICT JUDGE
On April 27, 2023, Defendant Manish Lachwani pled guilty to two counts of wire fraud in violation of 18 U.S.C. § 1343 and one count of securities fraud in violation of 15 U.S.C. §§ 78j(b) and 78ff and 17 C.F.R. § 240.10b-5.
In connection with sentencing, Lachwani has objected to most of the key findings and conclusions in the Presentence Investigation Report (“PSR”), including the probation officer's application of the U.S. Sentencing Guidelines (“U.S.S.G.”). See PSR at 29-34.(dkt. 96). The central issue of dispute is the application of the “loss” enhancement under U.S.S.G. § 2B1.1(b)(1). The Government and the probation officer calculate the total loss as more than $65 million, which would result in a 24-level enhancement. Lachwani argues that there was no loss at all, or in the alternative, that the Government has failed to carry its burden to show that any reasonable estimate of loss that resulted from Lachwani's criminal conduct can be made. His position is therefore that there should be no enhancement for loss. Because the loss question will have such a substantial effect on Lachwani's Guidelines calculation, the Court has postponed sentencing pending its resolution. In addition to Lachwani's own sentencing memorandum, the Court allowed Lachwani to file a response to the Government's sentencing memorandum that specifically addressed the prosecutors' arguments about loss and related issues. Def.'s Response to United States' Sentencing Mem. Re: Loss (“Def.'s Response”) (dkt. 113). The Court also held a hearing on January 11, 2024, at which it heard argument specifically on the question of loss. See Hearing Tr. (dkt. 125).
Throughout this Order, the Court cites to the document page numbers assigned by ECF, not any page numbers in the original documents.
Having considered the parties' sentencing memoranda (dkts. 103, 105), Lachwani's response memorandum, the PSR, the evidence submitted by both parties in connection with sentencing, and the arguments at the loss hearing, the Court finds that a reasonable estimate of the total loss amount under U.S.S.G. § 2B1.1(b)(1) is $3,149,056. A 16-level loss enhancement therefore applies. In addition, a 2-level enhancement for offenses involving 10 or more victims will apply under U.S.S.G. § 2B1.1(b)(2)(A). This Order memorializes the Court's reasoning on these questions, leaving the remainder of the Guidelines and sentencing issues to be resolved at sentencing. Where the Court makes statements of fact about matters that are the subject of Lachwani's objections to the PSR, the Court has resolved those objections under the applicable evidentiary standard.
I. BACKGROUND
Lachwani committed fraud between 2017 and 2020 in connection with his efforts to raise funds for HeadSpin, a software company he founded and for which he served as CEO from 2015 until the discovery of his fraud in 2020. Plea Agmt. (dkt. 88). In his Plea Agreement, Lachwani admits that he sent or directed HeadSpin employees to send financial information to potential investors that he knew overstated the company's actual revenue and annual recurring revenue (“ARR”). Id. at 5. Lachwani knew the overall numbers were inflated at least in part because he “maintained and controlled” the spreadsheet where HeadSpin's ARR was tracked, and he knew the spreadsheet both contained overstated entries for “certain customers” and “included customers that had not yet agreed to use the service by signing a contract or issuing a purchase order.” Id. at 5-7. The Plea Agreement does not itself reflect which specific customers or entries were misrepresented by Lachwani, or by how much Lachwani overstated given metrics. But Lachwani does admit that he knowingly shared the overstated revenue and ARR figures with investors through a slide deck and the spreadsheet that he controlled. Id. at 7.
“ARR is a measure of a company's subscription revenue rate at a particular point in time, annualized to show revenue the company would expect to make, at that rate, over an entire year.” Plea Agmt. at 5. Lachwani knew that ARR was “one metric to assess the past success and forecasted growth for a SaaS [software-as-a-service] company like Headspin,” and he also “knew the overstated information about ARR was significant to certain investors and their decisions to invest in HeadSpin.” Id. at 5, 6.
Lachwani's fraud overlapped with three rounds of investment in HeadSpin. Id. at 5; PSR ¶¶ 23-26. Between December 2017 and May 2018, HeadSpin raised about $10.25 million in a series of issuances of promissory notes that were convertible into HeadSpin stock. Between September and October of 2018, HeadSpin undertook a “Series B” financing round in which it raised about $20 million. Plea Agmt. at 5. And between November 2019 and April 2020, HeadSpin undertook a “Series C” financing round that yielded $60 million from investors. Id. Throughout this Order, when the Court refers to “the investors,” it refers to the Series B and Series C investors collectively.
Lachwani's fraud was discovered in 2020 by another HeadSpin employee, Stefanos Loukakos, who noticed that HeadSpin had been representing to investors that certain entities were customers when in fact they were not. PSR ¶ 27; see also Loukakos Decl. (dkt. 103-2). Loukakos brought this information to HeadSpin's independent directors, who initiated an investigation by outside legal and forensic firms. PSR ¶ 27; see also Loukakos Decl. (dkt. 103-2); Arora Decl. ¶¶ 7-9 (dkt. 103-4); Faris Decl. ¶ 11 (dkt. 103-5). Lachwani was removed as CEO of HeadSpin in May 2020. PSR ¶ 27; Arora Decl. ¶ 8. Thereafter HeadSpin retained accounting firm KPMG to reconstruct HeadSpin's actual revenues and ARR during the relevant period. Id.; see also White Decl. (dkt. 103-6); Jolley Decl. (dkt. 103-7). The KPMG investigation, which the parties refer to as the “Recast,” determined that HeadSpin's actual revenues and ARR were much less than the figures Lachwani had been reporting to potential investors. HeadSpin's directors then undertook a complex negotiation with the company's shareholders that, broadly speaking, was designed to head off the threat of litigation and keep HeadSpin afloat following the discovery of the fraud and other attendant revelations about poor management at the company. PSR ¶ 28; see also Arora Decl. ¶¶ 10-12; Background for the Recapitalization of HeadSpin, Inc., Tomeno Decl. Ex. U (“Recap Background Mem.”) (dkt. 103-8). In the resulting set of transactions, which the parties refer to as the “Recap,” Series B and Series C investors were given the opportunity to sell back up to 70% of their preferred stock at the original purchase price and exchange the remaining 30% for promissory notes. Id. The face value of the promissory notes equaled the original share purchase price, but the notes had a 5-year term and accrued 1% annual interest. Id. The investors could also opt to retain their shares in the company but have them repriced at much lower rates. Id. Some investors took the former option, some the latter, and some a combination of the two. HeadSpin survived and continues to operate today.
As discussed later, the Court agrees with Lachwani that the Recast does not independently support the conclusion that Lachwani's criminal conduct was responsible for the entire difference between HeadSpin's actual financial metrics and the overstated metrics reported by Lachwani. KPMG was not retained specifically to investigate the fraud.
II. LEGAL STANDARD
“All sentencing proceedings begin with the district court's calculations of the applicable Guidelines range.” United States v. Prien-Pinto, 917 F.3d 1155, 1157 (9th Cir. 2019) (citing Gall v. United States, 552 U.S. 38, 49 (2007)).
The government bears the burden of proof on the facts underlying a sentence enhancement. United States v. Zolp, 479 F.3d 715, 718 (9th Cir. 2007). At sentencing, “the district court may rely on undisputed statements in the PSR,” but if the defendant objects to those statements “the district court is obligated to resolve the factual dispute, and the government bears the burden of proof to establish the factual predicate[.]” United States v. Ameline, 409 F.3d 1073, 1085-86 (9th Cir. 2005); see also Fed. R. Crim. P. 32(i)(3)(B); United States v. Showalter, 569 F.3d 1150, 1160 (9th Cir. 2009).
Generally, the Government must carry its burden by a preponderance of the evidence, but the Government must meet a higher standard-proof by “clear and convincing evidence”-in cases where there is “an extremely disproportionate impact on the sentence.” United States v. Valle, 940 F.3d 473, 479 (9th Cir. 2019) (quoting United States v. Jordan, 256 F.3d 922, 930 (9th Cir. 2001)). Here, with respect to the loss enhancement-where the Government seeks an enhancement of 24 levels and the Base Offense Level is 7-the Government must prove the underlying facts by clear and convincing evidence. See United States v. Lonich, 23 F.4th 881, 911-912 (9th Cir. 2022) (“We have repeatedly recognized that our cases commonly turn on the last two [Valensia] factors: whether the enhanced sentence is four or more offense levels higher (factor 5) and more than double the initial sentencing range (factor 6).”) (discussing United States v. Valensia, 222 F.3d 1173 (9th Cir. 2000)). Under the clear and convincing evidence standard, “the factfinder must have an ‘abiding conviction that the truth of [the] factual contentions' at issue is ‘highly probable.'” Lonich, 23 F.4th at 916 (quoting Mondaca-Vega v. Lynch, 808 F.3d 413, 422 (9th Cir. 2015) (en banc)) (alteration in original).
The Government briefly suggests that only a preponderance of the evidence standard should apply because “[t]he Ninth Circuit caselaw requiring clear and convincing evidence for guidelines enhancements is clearly irreconcilable with modern Due Process doctrine.” United States' Sentencing Memorandum (“Gov't Mem.”) at 18 (dkt. 103). But the Ninth Circuit rule remains good law, reaffirmed as recently as 2022. See Lonich, 23 F.4th at 911-12. The Government does not argue that the enhancements it seeks would not have “an extremely disproportionate impact on the sentence.” Valle, 940 F.3d at 479. The Government also suggests that the clear and convincing evidence standard does not apply where “the loss amount is from charged conduct for which the defendant has been convicted.” Gov't Mem. at 18 (citing Lonich, 23 F.4th at 913). But here the conduct to which Lachwani has pled guilty does not by itself answer all the factual questions necessary for the Court to make a reasonable estimate of the loss.
In making factual determinations, “a sentencing judge is generally not restricted to evidence that would be admissible at trial. But ‘inadmissible evidence cannot be considered [at sentencing] if it lacks sufficient indicia of reliability to support its probable accuracy.'” United States v. Egge, 223 F.3d 1128, 1132 (9th Cir. 2000) (citation omitted, alternation in original).
III. DISCUSSION
A. Loss Enhancement Under U.S.S.G. § 2B1.1(b)(1)
Section 2B1.1(b)(1) of the Guidelines imposes an increase in the offense level according to the loss that resulted from the offense. “The court need not make its loss calculation with absolute precision; rather, it need only make a reasonable estimate of the loss based on the available information.” Zolp, 479 F.3d at 718 (citing U.S.S.G. § 2B1.1, cmt. n. 3(C); United States v. Peyton, 353 F.3d 1080, 1090 n.11 (9th Cir. 2003)). The commentary to the Guidelines defines “loss” as “the greater of actual loss or intended loss.” Id. (quoting U.S.S.G. § 2B1.1, cmt. n.3(A)). In turn, “actual loss” is defined as “the reasonably foreseeable pecuniary harm that resulted from the offense.” U.S.S.G. § 2B1.1, cmt. n. 3(A)(i). “The [G]uidelines do not present a single universal method for loss calculation under § 2B1.1-nor could they, given the fact-intensive and individualized nature of the inquiry.” Zolp, 479 F.3d at 718. Rather, “the district court's obligation is to ‘adopt a reasonable ‘realistic, economic' projection of loss based on the evidence presented.'” United States v. Holmes, No. 5:18-CR-00258-EJD-1, 2023 WL 149108, at *2 (N.D. Cal. Jan. 10, 2023) (quoting United States v. W. Coast Aluminum Heat Treating Co., 265 F.3d 986, 991 (9th Cir. 2001)).
B. The Investors Suffered an “Actual Loss”
Before the Court proceeds to measure the loss, it is necessary to resolve the issue of whether there was a loss under the meaning of the Guidelines. See United States v. Berger, 587 F.3d 1038, 1045 (9th Cir. 2009). Lachwani argues that there was not. His position is that there is no evidence “that any investor lost money investing in HeadSpin” and that there are even present indications that the investors might ultimately profit over the long term. Defendant's Sentencing Memorandum (“Def's Mem.”) at 18-19 (dkt. 105). In other words, although Lachwani criminally misrepresented the state of HeadSpin to investors in various ways, those investors still received stock in a company that, as it turned out, was able to survive the discovery of the fraud and retain enough capital to offer them most of their money back in the Recap.
The position has some intuitive appeal, and Lachwani argues it forcefully. But it has two fatal problems. First, it cannot account for the nature of Lachwani's fraudulent representations. As defense counsel conceded at the hearing on loss, an investor in a startup like HeadSpin decides whether to invest based on the prospects of a large return. Hearing Tr. at 46:9-47:24; see also Investor 1 Decl. ¶¶ 2, 9-10. (dkt. 104-4) (noting that his investment firm “invests in growth-stage technology companies with the goal of obtaining a significant return on its investment” and that the firm “targets high-growth companies for investment” and “believed that Headspin was achieving significant sales and revenue growth based on the false information that Lachwani provided”); Investor 2 Decl. ¶¶ 2, 10 (dkt. 104-5) (similar). The investor does not simply expect to get her principal back, or she would just put her money in the bank. Nor does she expect to make a small profit at some distant time in the future; on the contrary, the record shows that investors cared about HeadSpin's apparently rapid growth and hoped it would continue. Investor 1 Decl. ¶¶ 9, 11; Investor 1 Internal Emails and Slide Deck, Tomeno Decl. Ex. V (dkt. 104-7); Investor 2 Slide Deck, Tomeno Decl. Ex. X; Investor 1 Investment Memo, Tomeno Decl. Ex. Y. Lachwani's fraudulent representations were designed to attract investment to his company, and as such they were designed to make it seem like HeadSpin had secured more sales and subscription revenue, was growing more quickly, and had secured more lucrative, high-profile customers than it really had. This was plainly why Lachwani inflated the ARR figures to investors, who in internal materials repeatedly remarked on the firm's ARR growth. Investor 1 Decl. ¶¶ 9, 11; Investor 1 Internal Emails and Slide Deck, Tomeno Decl. Ex. V (dkt. 104-7); Investor 2 Slide Deck, Tomeno Decl. Ex. X; Investor 1 Investment Memo, Tomeno Decl. Ex. Y; Investor 3 Emails, Tomeno Decl. Ex. AA. Creating a sense of rapid and impressive growth is also why Lachwani passed other misleading information to investors, such as the slide deck stating that “HeadSpin is Trusted by 1,000+ Companies Worldwide,” and displaying the logos of a number of prominent firms that were not, in fact, HeadSpin's customers, or another slide that reported the firm to have deployed 22,000 or 32,000 devices when in fact it had only deployed 2,000. The record amply supports the conclusion that Lachwani's misrepresentations were intended to boost the appearance that HeadSpin was growing more rapidly and impressively than it really was. Indeed, there are no other explanations for Lachwani's conduct that are even plausible, despite Lachwani's suggestion that he was merely “gilding the lily” because he is a perfectionist or for some other idiosyncratic reasons.
Lachwani argues that the customer slide was not misleading because it did not state that HeadSpin had over 1,000 paying customers, or that all the logos on the slide were current paying customers. See Def.'s Response at 20. This is an implausible and excessively literal interpretation of the slide. Just under the header is a line that says that “[a] more comprehensive list can be found at https://www.headspin.io/customers.” Id. (emphasis added). One of the firm's potential investors emphasized the information contained in that slide in internal emails, referring to the “logos” listed as the firm's “customers.” Tomeno Decl. Ex. AA at 276. Lachwani also argues that the device count figures were not inaccurate, and that the Government only believes them to be inaccurate because it misunderstands the way HeadSpin's technology worked. See Def.'s Response at 19-20. But the record shows that even HeadSpin's own senior employees, including one of its lead engineers, believed this figure was bogus. See Loukakos Decl. ¶ 7; Id. Ex. A at 7, 12.
Importantly, the record also shows that HeadSpin's investors would not have invested in HeadSpin-thus parting with millions of dollars-had they known the truth about HeadSpin's performance. See Arora Decl. ¶ 4; Investor 1 Decl. ¶¶ 9-11; Investor 2 Decl. ¶¶ 9-10; Investor 3 Decl. ¶¶ 4-5. Because they invested in HeadSpin, at least some of those investors lost out on chances to make other investments that would have actually matched the high-growth investment profile they were targeting. See Investor 1 Decl. ¶ 10; Investor 2 Decl. ¶ 10.
Lachwani objects that the investor declarations only say that the investors would not have invested had they known about HeadSpin's true financial situation, and that some of the reasons the financial situation was worse than represented to investors did not have to do with the fraud. The Court overrules this objection for two reasons. First, this argument comes very close to arguing that Lachwani's misrepresentations were not actually material to the investors. But that route is closed to Lachwani. He has pled guilty to three counts of fraud, each of which have a materiality element. Second, Lachwani's argument assumes that the investors would nevertheless have invested in HeadSpin had they known Lachwani, its CEO, was defrauding them. That cannot withstand common sense. Cf. USA v. Hussain, No. 16-CR-00462-CRB-1, 2019 WL 1995764, at *8 (N.D. Cal. May 6, 2019), afFd sub nom. United States v. Hussain, 818 Fed.Appx. 765 (9th Cir. 2020) (“The question is not . . . whether HP would have purchased Autonomy if it had known about the full extent of Autonomy's hardware sales. The question is whether HP would have purchased Autonomy if it knew about the fraud.”) (citation omitted).
In short, the investors were induced to part with their money based on fraudulent misrepresentations about HeadSpin's prospects for growth-and thus for future returns- and they would not have parted with their money had they known about the fraud. It is true that the investors got something in return, but what they got in return were shares in a company with materially worse prospects for future returns than they believed they were getting. In other words, at the time of the transaction, the investors got something worth less than what they paid. For that reason, they suffered an actual loss. As discussed in Part III(C) below, Lachwani is entitled to a credit against the loss for the actual value of the shares the investors received at the time of the transaction. But he is not entitled to a finding that there was no actual loss merely because, at some time after the company had absorbed the impact of the fraud, the company had sufficient assets to reimburse investors. It is the Court's task to undertake an “fact-intensive and individualized” inquiry, Zolp, 479 F.3d at 718, and to give a “realistic, economic” account of the loss, W. Coast Aluminum Heat Treating Co., 265 F.3d at 991. Adopting Lachwani's position would not account for the nature of the fraud and the actual loss it caused the investors at the time of the investment.
Lachwani's argument that the investors suffered no actual loss fails for a second reason. The fact that investors were ultimately able to recover much of their money in the Recap is the product of contingent events that occurred after the discovery of the fraud- events that took place at least in part to reduce the fallout from the discovery of Lachwani's criminal activity. The Guidelines contemplate credit against loss for “[t]he money returned, and the fair market value of the property returned and the services rendered, by the defendant or other persons acting jointly with the defendant, to the victim before the offense was detected.” U.S.S.G. § 2B1.1, cmt. n.3(E)(i) (emphasis added). After Lachwani's fraud was discovered, members of the board of directors negotiated the Recap transaction with the investors that helped to head off the threat of litigation and a possible bankruptcy. See Arora Decl. ¶ 11. Despite Lachwani's argument to the contrary, the record shows that Lachwani's cooperation was necessary to the success of the Recap. He agreed to step down as CEO, resign as a member of the Board, cancel his own shares in the company, provide a cash refund to one investor in exchange for a release of that investor's claims against him, and transfer shares of his own stock to another investor. See Recap Background Mem. at 119, 122-23, Tomeno Decl. Ex. U. Crediting Lachwani for the amount returned to investors in the Recap would amount to rewarding him for actions taken to reduce the consequences of his fraud after its discovery. This would be contrary to the guidelines and Ninth Circuit law. See U.S.S.G. § 2B1.1, cmt. n.3(E)(i); United States v. Stoddard, 150 F.3d 1140, 1146 (9th Cir. 1998) (“Repayments before detection show an untainted intent to reduce any loss. Repayments after detection may show no more than an effort to reduce accountability.”).
Lachwani argues that the money was not “returned” to the investors, but rather “the private company the investors owned decided to undertake a liquidity event in which shares were exchanged for cash, notes, reclassified shares, or a combination of the three.” Def's Response at 11. This argument ignores that the Recap was necessary because of the discovery of the fraud, and that Lachwani's cooperation was an integral part of it. Lachwani also seems to assume that the investors would have had ready access to their original funds if the Recap had not been successful. But the record shows that if the Recap had not taken place, HeadSpin would not have had sufficient funds to repay the investors, despite their liquidation preferences.
The Recap Background Memorandum notes that the directors considered a possible “wind-down” of the company either privately or through bankruptcy:
With respect to an orderly wind-down, in the absence of litigation, the Committee initially estimated that, after paying current debt, other liabilities, professional fees and salaries of essential staff during the wind-down, the Company would have a maximum of approximately $96 million remaining to distribute to its preferred stockholders, who had total liquidation preferences of $152 million (assuming no litigation or other legal disputes). That would result in the return of at most 63% of their liquidation preferences, and the holders of common stock and employees holding options would receive nothing. With hindsight, the Committee believes that legal expenses and contingencies would have caused the amount available to distribute to stockholders in a wind-down to have been substantially less. Recognizing the potential risk of litigation by stockholders, which would consume a substantial portion of the Company's cash, the Committee also considered winding the Company down in a formal bankruptcy proceeding, which would stay all litigation. After consulting bankruptcy counsel, the Committee estimated that a bankruptcy proceeding would take two or more years to complete and consume even more of the Company's cash given the disparate preferences and claims of the stockholders, creditors, customers and employees that would be contested.Recap Background Mem. at 119, Tomeno Decl. Ex. U.
Nor, finally, does United States v. Ortland, 109 F.3d 539 (9th Cir. 1997), help Lachwani's case that there was no loss. Lachwani argues that, at least in fraud cases involving investment in nonpublic companies, Ortland requires sentencing judges to find “that the actual loss to the investors [is] the total amount they did not regain.” Id. at 548. But Lachwani reads too much into Ortland, which does not stand for that broad proposition. First, and most importantly, the relevant issue raised on appeal in Ortland was whether “the district court miscalculated the amount of loss his actions caused because the court incorrectly included consequential damages in the loss calculation.” Id. at 547. The question was whether a receiver's fees-incurred during the process of winding up the partnership and returning the money to the defrauded investors-counted as consequential damages or “direct loss.” See id. at 547-48.
That case certainly contains language that describes the unchallenged method by which the district court calculated the loss amount. But the opinion does not describe that method as part of an analysis of the proper way to calculate loss in an investment fraud case. Indeed, the opinion does not even cite Guidelines commentary or case law relevant to the methods of loss calculation. Instead, the case's description of the loss calculations was necessary to its explanation of why the receiver's fees were part of the direct loss rather than consequential damages. To whatever extent the court's description of the method of loss calculation might be read as endorsement of that method in other cases, it would be unnecessary to the decision and, thus, merely dicta. See Cetacean Cmty. v. Bush, 386 F.3d 1169, 1173 (9th Cir. 2004).
Second, as already noted, the Ninth Circuit recognizes that “the guidelines do not present a single universal method for loss calculation under § 2B1.1-nor could they, given the fact-intensive and individualized nature of the inquiry.” Zolp, 479 F.3d at 718. Accordingly, in a subsequent unpublished decision, the Ninth Circuit declined to read Ortland as prescribing a universally applicable principle for loss calculation. See United States v. Bickerstaff, 185 F.3d 869, 869 n.4 (9th Cir. 1999) (unpublished) (rejecting appellant's argument that the district court erred by not following the method of loss calculation described in Ortland and noting that the district judge had “sufficiently distinguished Bickerstaff's fraud from the fraud in Ortland” and that “application note 8 to guideline section 2F1.1 specifically authorizes the consideration of the nature and extent of the fraud in calculating the loss amount.”). In other words, while Ortland does show that one district court used a method of loss calculation that might be favorable to Lachwani here, it does not follow that this method is required in all investment fraud cases.
Even if Ortland did say what Lachwani suggests, it would not result in a determination that there was no loss because-as the Court has already noted-the record shows that HeadSpin would not have had sufficient assets to completely reimburse investors if the company was dissolved. See note 7 above.
In short, then, the Court finds that the investors suffered an actual loss as a result of Lachwani's fraud, and Lachwani is not entitled to credit against the loss for the Recap. Ortland is not to the contrary.
C. A “Reasonable Estimate” of the Net Loss from Lachwani's Fraud Is Possible
The Court must next endeavor to estimate the value of the loss. The Government advocates two distinct approaches for doing so. While the Court ultimately cannot adopt either of the Government's proposed loss figures, the record nevertheless provides a sufficient basis for the Court to calculate an appropriate loss figure.
The Government's preferred position, which was adopted by Probation in the PSR, is that the actual loss is about $92 million, the full value of the money obtained by HeadSpin during the period of Lachwani's fraudulent scheme. The Government argues that this figure is supported by the fact that no investor would have invested had they known the truth about HeadSpin, that at the time of the stock sales “it was not reasonably probable that the stock had any value,” and the lack of a market for HeadSpin shares made them effectively worthless. United States' Sentencing Memorandum (“Gov't Mem.”) at 18 (dkt. 103).
But neither the facts nor the law support the view that HeadSpin stock can or should be judged to have been worthless at the time of the investment. With respect to the facts, the record shows that HeadSpin was not a “sham” company. It had some intrinsic value separate and apart from the fraud-it had a real product, some real customers, and a real market, as the subsequent reconstruction of its finances by KPMG shows. See, e.g., Recap Background Mem., Tomeno Decl. Ex. U; see also Loukakos Decl. Ex. A at 6. HeadSpin was a privately held company, to be sure, and the stocks were not liquid assets. But while this certainly makes it more difficult to value them, it does not follow that the stocks were valueless. See Holmes, 2023 WL 149108 at *4.
With respect to the law, Ninth Circuit precedent requires that, in cases of fraudulently induced investment in a company that is not a complete sham, the district court “may not assume that the loss inflicted equals the full pre-disclosure value of the stock[.]” Zolp, 479 F.3d at 719; see also United States v. Laurienti, 611 F.3d 530, 558 (9th Cir. 2010) (“If the underlying company was worthless or practically worthless, it is reasonable to use the total investment cost as the actual loss. But, if the underlying company has intrinsic value, then the use of the total investment cost is erroneous.”) (citing Zolp, 479 F.3d at 719); Holmes, 2023 WL 149108 at *4 (“The fact that the investors did not have-and never will have-an opportunity to liquidate their shares is undoubtedly unfortunate, but illiquidity and resale difficulty does not deprive Defendant of credit for the inherent value of Theranos shares.”); United States v. Leonard, 529 F.3d 83, 93 (2d Cir. 2008) (finding that the “district court erred in not deducting from the purchase price the actual value of . . . illiquid securities”). The Government suggests that Zolp's holding should be limited to public companies, but-as the district court also held in Holmes- Zolp does not “provide . . . a basis to distinguish the loss calculations for public securities from those involving private securities.” Holmes, 2023 WL 149108 at *4.
The Court's task here will be the same one it would face if HeadSpin were publicly traded: namely, to “disentangle the underlying value of the stock, inflation of that value due to the fraud, and either inflation or deflation of that value due to unrelated causes.” Zolp, 479 F.3d at 719; see also Berger 587 F.3d at 1046-47 (instructing the district court to apply a loss calculation method that “attempt[s] to gauge the difference between [the] share price-as inflated through fraudulent representation-and what that price would have been absent the misrepresentation”).
The Government offers an alternative calculation designed to aid the Court in carrying out the more complicated task prescribed in Zolp. Relying on the analysis of its expert Jeffrey Hart, the Government argues that the loss is the total difference between HeadSpin's actual value at the time of the Series B and C offerings and the actual amount HeadSpin acquired in those funding rounds. Hart endeavors to calculate the actual value of HeadSpin's shares at the time of the Series B and Series C offerings using two distinct methods: one based on the share price set in the Recap and the other based on multiples of HeadSpin's actual ARR as calculated by KPMG. See Hart Report ¶¶ 23, 29-46 (dkt. 1031). Hart then subtracts the actual value of the Series B and C offerings from the amounts paid to HeadSpin by the Series B and C investors. Id. Hart's first method yields an estimate of the combined value of the Series B and Series C offerings of about $18.1 million, while the second method yields an estimate of about $20.7 million. Id. ¶¶ 32, 46 The total amount actually paid for the Series B and C offerings was $79,979,610, so the first method yields a total net loss of about $61.8 million, and the second method yields a total net loss of about $59.2 million. Id.
Lachwani objects that the Government disclosed this method of loss calculation, and the evidence underlying it, belatedly. The Government did not propose this method of calculation to the probation officer. But the Court has granted Lachwani ample time and space to respond to the Government's additional arguments based on the Hart Report, including additional briefing and a separate hearing on loss. Lachwani could have sought a further continuance if necessary, but he did not do so.
Lachwani objects to these calculations for two reasons. He first argues that both of Hart's methods for calculating the actual values of the Series B and C offerings are unreliable. Here the Court will confine its discussion to the method involving the ARR multiple, since this is the approach the Court will ultimately adopt. ARR is a metric that is commonly used as part of the valuation of Software-as-a-Service (“SaaS”) companies like HeadSpin. Plea Agmt. at 5; Hart Report ¶¶ 34-35; Jolley Decl. ¶ 3. It is a metric that Lachwani has admitted to knowing was significant to investors' decisions. Plea Agmt. at 5. And it is a metric that several of HeadSpin's largest investors themselves say was especially significant to their decisions to invest in HeadSpin-statements that find support in their internal communications about HeadSpin. Investor 1 Decl. ¶ 11; Investor 2 Decl. ¶ 7; Investor 3 Decl. ¶ 3-4; Investor 1 Internal Emails and Slide Deck, Tomeno Decl. Ex. V (dkt. 104-7); Investor 2 Slide Deck, Tomeno Decl. Ex. X; Investor 1 Investment Memo, Tomeno Decl. Ex. Y; Investor 3 Emails, Tomeno Decl. Ex. AA. One central way that HeadSpin's investors evaluated whether the price of the Series B and Series C offerings was a reasonable valuation of the company was by using ARR multiples. This means dividing the total valuation of HeadSpin implied by the Series B and C offerings by HeadSpin's ARR to calculate the “multiple.” Hart Report ¶¶ 33-38. That multiple can then be compared with multiples from other firms. Id. HeadSpin's investors calculated and evaluated ARR multiples as part of their investment decisions. Investor 1 Decl. ¶ 11; Investor 2 Decl. ¶ 7; Investor 3 Decl. ¶ 4.
Lachwani is certainly correct that ARR and ARR growth were not the only factors considered by investors, but the evidence that they were of central significance is clear and convincing.
Lachwani, in large part through his expert John Ciancanelli, makes the case that Hart's ARR approach to reconstructing HeadSpin's value is relatively simplistic and would not likely be used, on its own, by one looking to definitively value a private firm. See Ciancanelli Report at 186-193, Hemann Decl. Ex. 7 (dkt. 107). Ciancanelli points out that HeadSpin's investors looked to a variety of metrics in determining whether HeadSpin's valuation was reasonable; he argues that Hart failed to perform additional analyses that would have made his estimates more reliable or sophisticated; and he argues that has failed to use other commonly used valuation methods. But while Ciancanelli is undoubtedly correct that the ARR multiple method does not provide a precise measure of HeadSpin's value, precision is not always equivalent to reliability. That is especially so where, as here, “absolute precision” is not required; “rather, [the Court] need only make a reasonable estimate of the loss based on the available information.” Zolp, 479 F.3d at 718 (citing U.S.S.G. § 2B1.1, cmt. n. 3(C); United States v. Peyton, 353 F.3d 1080, 1090 n.11 (9th Cir. 2003)). The Court finds that Hart's analysis can reliably serve this purpose.
Hart's approach has several indicia of reliability. First, Hart's professional experience qualifies him to offer it. Second, the analysis yields a similar reconstructed valuation as that produced by his other method, which is based on the share price agreed to in the negotiated Recap. The fact that the numbers generally agree, despite being produced in quite different ways, adds to their credibility. Third, the analysis has the benefit of reproducing one method of valuing HeadSpin that was used by the investors themselves. Investor 1 Decl. ¶ 11; Investor 2 Decl. ¶ 7; Investor 3 Decl. ¶ 4. Fourth, the ARR multiple calculation contains assumptions favorable to Lachwani, and thus it arguably overstates the actual value of the shares at the time of the Series B and C sales. Hart indicates, for instance, that the inflated figures provided to investors may have led them to value the company using a higher ARR multiple than they otherwise would have. Hart Report ¶ 38; see also Investor 1 Decl. ¶ 11 (“A fair valuation of HeadSpin at the time of [Investor 1 's] investments would have been significantly less than the valuation at which [Investor 1] invested because a lower multiple would have applied to HeadSpin's actual ARR.”). Hart nevertheless used the same high multiples to estimate the value of the company based on the true ARR figures. Finally, Ciancanelli himself does not deny that ARR multiples were used by HeadSpin's investors as one approach to valuing the company and does not contradict the well-attested view that such multiples are one customary metric used to value SaaS companies like HeadSpin. It is reasonably certain that Hart's ARR multiple method is reliable and can support a reasonable estimate of the loss based on the available evidence.
Lachwani argues that the round valuation figures used in the funding rounds ($500 million in Series B and $1.1 billion in Series C) were merely “thrown out” by HeadSpin and thus that reconstructing HeadSpin's valuation with any financial metrics is impossible. He also argues that investors disbelieved the financial metrics anyway, so the portion of those metrics that were inflated by Lachwani's conduct could not have moved the needle on their valuation of the company. These arguments fail because, again, they essentially ask the Court to find that HeadSpin's financial metrics were not material to the investors' decisions, when in fact Lachwani has pled guilty to providing material, false information regarding revenue and ARR to investors. See Plea Agmt. At any rate, the investors may not have produced the round valuation figures, but they obviously did evaluate the offering and determine that those figures were fair valuations of the company. Then they purchased equity at that price. That is a “valuation.”
Lachwani's second objection is to Hart's reliance on the KPMG “Recast” data, from which he derives the actual ARR figures used to revalue the Series B and Series C offerings. Lachwani points out, correctly, that the KPMG reconstruction of HeadSpin's finances was not carried out for the purpose of investigating Lachwani's fraud; rather, it was carried out for the purpose of reconstructing the company's finances over a specified period. See Recap Background Mem., Tomeno Decl. Ex. U at 119. In other words, KPMG never claimed that the difference between its reconstructed figures and the original, overstated figures was entirely the product of Lachwani's fraud. KPMG's analysis cannot support that claim.
Nor does other evidence in the record permit the Court to attribute the entirety of the difference between the Recast figures and the original figures to Lachwani's criminal conduct. HeadSpin suffered from serious internal management issues, especially with regard to tracking its finances. Many of those issues appear to have stemmed from Lachwani's management, but not all that conduct was necessarily criminal. One key example will suffice to illustrate the point. A very large portion of the overstatement of ARR resulted from the way Lachwani and HeadSpin accounted for “revenue” it received from resellers of its services-the parties call this “Channel ARR.” In essence, HeadSpin booked the full revenues payable to the resellers from third parties as HeadSpin's own revenue. Lachwani told others at the company that this was a legitimate way of accounting for reseller revenues. See Okmyanskaya Decl. ¶¶ 7-8 (dkt. 103-3). It is at least possible that he believed this, and the Government has not carried its burden to prove otherwise. This method of accounting for Channel ARR was known to some in the company, and it was apparently approved by one of HeadSpin's accountants. See Id. ¶ 8; see Sadarangani Proffer Interview, Hemann Decl. Ex. 7 at 85 (Dkt. 107 at 85). There is also some evidence that HeadSpin may have informed one investor about this practice. See Barczak Decl. Ex. 4 at (dkt. 115). In short, Lachwani's position is that the Channel ARR accounting was at least believed to be a legitimate method of accounting-albeit one that happened to greatly inflate HeadSpin's ARR metrics with moneys that HeadSpin would never see.
The Court has doubts about whether the Channel ARR accounting was in fact innocent, but doubts are not enough; the Court must find the facts underlying its loss calculation by clear and convincing evidence. In support of its own contention that the Channel ARR entries were fraudulent, the Government mainly relies on one paragraph in a declaration by Sana Okmyanskaya, HeadSpin's former bookkeeper, where she says that she later learned “what Mr. Lachwani told me about the reseller arrangements was not true.” See Okmyanskaya Decl. ¶ 8. Okmyanskaya's meaning is somewhat ambiguous, and it does not explain away the other evidence to which Lachwani points the Court's attention. Nor did the Government during the loss hearing. The Court is therefore unable to say to a reasonable certainty that the Channel ARR accounting was fraudulent.
In short, then, a large portion of the overstatement of ARR identified by KPMG was due to Channel ARR accounting problems, and the Government has not carried its burden to show that Lachwani is criminally responsible for those problems. This means that the Court cannot assume that Lachwani is criminally responsible for the entire difference between HeadSpin's overstated ARR as reported by Lachwani and HeadSpin's actual ARR as reported by KPMG. Hart's analysis assumes this to be the case. Thus, because the Court's task is to “disentangle the underlying value of the stock, inflation of that value due to the fraud, and either inflation or deflation of that value due to unrelated causes,” Zolp, 479 F.3d at 719, the Court cannot simply adopt Hart's loss estimate. See Holmes, 2023 WL 149108, at *3 (“The Ninth Circuit applies a general loss causation principle at sentencing, ‘permitting a district court to impose sentencing enhancements only for losses that resulted from the defendant's fraud.'”) (quoting Berger, 587 F.3d at 1043).
Nevertheless, while Hart's exact numbers do not sufficiently enable the Court to do the “disentangl[ing]” required by Zolp, Hart's methods-along with other facts available to the Court-still enable a reasonable estimate of the loss. Lachwani has admitted to making several fraudulent entries in his ARR spreadsheet that were passed along to investors. These, his counsel conceded, inflated the reported ARR figures by a total of $1.5 million. These figures can be used to estimate loss using the ARR multiple method applied by Hart. Unlike the Government's expert, however, the Court will downwardly adjust the overstated ARR figures only for those ARR entries that Lachwani has conceded were overstated because of his fraud.
At the hearing, Lachwani's counsel conceded that $1.5 million in ARR was overstated because of Lachwani's misrepresentations:
THE COURT: Okay. What you are saying to me is the amount of money that you believe can be attributed to the defendant's conduct, reserving all your other arguments about loss, and so forth and so on, you said several million. What are you saying? Is it $2 million, $3 million, $5 million, $8 million? What is it?
MR. BERNSTEIN: $1.5 million, $900,000 from the Uber account and
THE COURT: It's $1.5 million?
MR. BERNSTEIN: That's the amount, and that represents about 2 percent of the overstatement.Hearing Tr. at 67:3-67:13. The $1.5 million figure also appears in Lachwani's sentencing memorandum, which attributes the $1.5 million to the overstatement of the ARR from three specific customers discussed in the PSR. Def's Mem. at 12 n.7; see PSR ¶ 17. The numbers used by the Court are not exactly equal to the numbers used by Lachwani in his memorandum, which appear to have been rounded. Instead, the Court uses the overstated values as stated in the Oct. 2019 ARR Spreadsheet and the actual values as reported in the PSR. See PSR ¶ 17; Oct. 2019 ARR Spreadsheet, “Committed Customers” tab, Rows 47, 111.
In addition, for evidentiary reasons, the Court will measure loss only for the Series C transaction. The only version of Lachwani's ARR spreadsheet that is in the record was provided to investors in connection with that transaction. See Oct. 2019 ARR Spreadsheet, Investor 2 Decl. Ex. A (dkt. 104-5). This matters because Lachwani's ARR spreadsheet carried different fraudulent entries at different times. The October 2019 version, for example, contains only two of the three admittedly fraudulent entries. See id., “Committed Customers” tab, Rows 47, 111; see also PSR ¶ 17. Thus, the Court only calculates the Series C loss using two fraudulent entries-those for Uber and American Express-which together overstated ARR reported to Series C investors by a total of $1,078,000. The absence in the record of a spreadsheet furnished in connection with the Series B transaction makes it impossible for the Court to determine with reasonable certainty which fraudulent entries it contained, so the Court must set the Series B transaction aside.
That is, $960,000 for the Uber account and $118,000 for the American Express account. PSR ¶ 17; see Oct. 2019 ARR Spreadsheet, “Committed Customers” tab, rows 47, 111.
To be clear, the record supports the finding that Lachwani committed fraud in connection with the Series B transaction, and Lachwani pled guilty to doing so. But the record simply does not contain sufficient information to allow the Court to reasonably estimate the loss for that transaction.
With these parameters, the Court calculates loss as follows. The Court's calculations are intended to mirror those performed by Hart using his ARR method. See Hart Report ¶¶ 40-46; see also id. Schedule 4.
Table 1: Calculations for Fraud-Adjusted ARR Figure
Calculation
Schematic
Figure
[A]
$54,700,000
Amount of Uber Overstatement
[B]
$960,000
Amount of American Express Overstatement
[C]
$118,000
Fraud-Adjusted ARR
[D] = [A] - [B] - [C]
$53,622,000
Hart Report ¶¶ 40-46. The Oct. 2019 ARR Spreadsheet reports an unrounded figure of $54,658,095. See Oct. 2019 ARR Spreadsheet, “Forecast” tab, cell N11. The Court adopts the rounded figure used in the Hart Report-which Hart in turn took from the internal investment memorandum of one of HeadSpin's major investors-because it results in a more conservative loss estimate. Lachwani's Plea Agreement also states that he sent a slide deck to Series C investors that “stated that HeadSpin's ARR was more than $54 million as of the second quarter of 2019, when [he] knew this number was overstated.” Plea Agmt. at 7.
Table 2: Calculations for Fraud-Adjusted Value of the Series C Offering
Calculation
Schematic
Figure
[E]
19.0
Fraud-Adjusted ARR (from Table 1)
[D]
$53,622,000
Fraud-Adjusted Enterprise Value
[F] = [D] * [E]
$1,018,818,000
[G]
$58,200,000
[H]
$59,999,954
Fraud-Adjusted Post-Money Valuation
[I] = [F] + [G] + [H]
$1,137,017,954
[J]
5%
Fraud-Adjusted Value of Series C Offering
K] = [I] * [J]
$ 56,850,898
Hart Report ¶ 44, Fig. 9.
Hart Report ¶ 44, Fig. 9. The amount of cash assets HeadSpin had at the time of the Series C offering.
Hart Report ¶ 44, Fig. 9. The amount of cash acquired by HeadSpin through the Series C offering.
Hart Report ¶ 44, Fig. 9.
Table 3: Loss Calculation
Calculation
Schematic
Figure
Amount Paid by Series C Investors
[L]
$59,999,954
Fraud-Adjusted Value of Series C Offering (from Table 2)
[K]
$56,850,898
Loss
= [L] - [K]
$3,149,056
Thus, the Court finds that, based on the available evidence, a reasonable estimate of the loss resulting from Lachwani's fraud is $3,149,056. This figure reflects the estimated difference in value between what the Series C investors paid to HeadSpin and what those investors received, insofar as that difference can be attributed to Lachwani's criminal conduct.
To be sure, the Court's estimate is just that-an estimate. But it should be emphasized that this estimate is likely to understate the loss in HeadSpin's value that resulted from Lachwani's criminal conduct. First, and most obviously, it does not account for loss incurred in the Series B offering. Second, the calculations cannot account for the non-financial metrics about which Lachwani misled investors, including the number of devices deployed on HeadSpin's networks and number of customers HeadSpin had. Knowing the truth about these matters may have led investors to use lower ARR multiples to value the company, which would result in a greater loss. See Hart Report ¶¶ 38-39; see also Investor 1 Decl. ¶ 11. Third, the calculations cannot account for the loss from the fraud Lachwani committed to conceal HeadSpin's financial problems, such as the forged invoices Lachwani used to cover up the large volume of “unbilled receivables” that HeadSpin carried in its books. See Tomeno Decl. ¶¶ 4-12 (dkt. 104-7); Id. Exs. A-H; Okmyanskaya Decl. ¶¶ 6-7; see also Okmyanskaya Proffer Interview at 35, Hemann Decl. Ex. 6. Fourth, the calculations cannot account for the significant reduction in value that would have been assigned to the company by any sensible investor who was aware that Lachwani, HeadSpin's CEO, was committing fraud. In other words, in endeavoring to carry out its difficult task of making a “reasonable estimate of loss based on the available evidence,” the Court has made several assumptions favorable to Lachwani. This is because the Court's calculations do not account for several factors that had an adverse effect on HeadSpin's value, but which are not susceptible to easy quantification, at least on this record.
To the extent that Lachwani objects that this was not what his forged invoices were designed to do, the Court resolves the factual dispute against him.
Under the Guidelines, a loss of $3,149,056 results in a 16-level enhancement. U.S.S.G. § 2B1.1(b)(1).
D. Lachwani's Fraud Had More Than Ten Victims
Section 2B1.1(b)(2)(A)(i) of the Sentencing Guidelines provides for a 2-level enhancement if the offense “involved 10 or more victims.” U.S.S.G. § 2B1.1(b)(2)(A)(i). “Victim” is defined as a person, including a corporation, “who sustained any part of the actual loss determined under subsection (b)(1).” U.S.S.G. § 2B1.1(b)(1), cmt. n.1. A district court need only “make a reasonable estimate . . . based on the available information” in counting victims. United States v. George, 949 F.3d 1181, 1186 (9th Cir. 2020), cert. denied, 141 S.Ct. 605 (2020) (quoting Zolp, 479 F.3d at 719).
The PSR recommends an enhancement under Section 2B1.1(b)(2)(A), and the parties agree that there were at least 10 investors in the Series B and C funding rounds. Plea Agmt. at 10. Lachwani argues that none of these investors suffered an actual loss because no investor suffered a loss in the Series B and Series C funding, in accordance with his other arguments addressed above. But the Court has found that the investors in question did suffer an actual loss, and thus the Court finds that a 2-level enhancement under Section 2B1.1(b)(2)(A)(i) applies.
IV. CONCLUSION
For the foregoing reasons, the Court's Guidelines calculations will be, in part, as follows:
a. Base Offense Level, U.S.S.G. §2B1.1(a): 7
b. Specific Offense Characteristics, U.S.S.G. §2B1.1(b)(1)(I) (loss more than $1.5 million but less than $3.5 million): +16
c. Specific Offense Characteristics, U.S.S.G. §2B1.1(b)(2)(A) (more than 10 victims): +2
The Court will reserve ruling on the remaining Guidelines adjustments recommended in the PSR until sentencing.
The parties are directed to confer and propose a date for sentencing to the Courtroom Deputy.
IT IS SO ORDERED.