The defendant accounting firm appeals from a judgment entered in the United States District Court for the Southern District of New York, Richard Owen, Judge, after a jury returned a verdict against it. The jury awarded plaintiff $17 million, to which the district judge added pre-judgment interest, post-judgment interest and costs, in a civil action seeking damages for losses that Manufacturers Hanover Trust Company ("Manufacturers" or "MHT") claimed to have suffered as a result of certain alleged misrepresentations that Arthur Andersen & Co. ("Andersen") made on behalf of Andersen's client, Drysdale Securities Corporation ("DSC") and its successor, Drysdale Government Securities, Inc. ("DGSI"). The district judge submitted seven separate theories of liability to the jury: misrepresentation or material omission "in connection with" the purchase or sale of securities, in violation of section 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. Sec. 78j(b), and Rule 10b-5, 17 C.F.R. Sec. 240.10b-5, promulgated thereunder; misrepresentation or material omission "in" the purchase or sale of securities, in violation of section 17(a) of the Securities Act of 1933, 15 U.S.C. Sec. 77q(a)(2); misrepresentation or material omission in a "prospectus" or "oral communication," in violation of section 12(2) of the Securities Act of 1933, 15 U.S.C. Sec. 77l(2); conspiracy to violate the above federal securities laws; aiding and abetting in the violation of the above federal securities laws; common law negligence; and common law fraud. The jury returned a general verdict, following which the judge requested that the jury state the cause or causes of action on which the verdict was premised, which it did.

On appeal, Andersen argues principally that the district court lacked federal subject matter jurisdiction; that the requisite loss causation standard for liability for securities fraud was not met; that plaintiff MHT caused its loss by its own recklessness; that it was error for the district judge to submit a negligence theory to the jury; that in selecting and instructing the jury the district judge deprived Andersen of a fundamentally fair trial; and that reversal of any of MHT's claims requires reversal of the entire verdict. MHT cross-appeals for a mini-trial exclusively on the issue of punitive damages. We affirm the judgment in favor of plaintiff in the amount of $17 million and deny the relief requested on cross-appeal. We also remand for further proceedings on the issue of pre-judgment interest only.

In this civil action MHT sought damages against the Andersen accounting firm, which it asserts made misrepresentations as to the financial status of DGSI, a company created by DSC to transact business through trading in repurchase agreements ("repos") or resale agreements ("reverse repos") involving government securities.1 DSC, a brokerage house since 1889, engaged in a scheme to purchase and sell government securities through repos and reverse repos beginning in May 1980 and ending in May 1982, just three months after DSC had transfered the repo business to a separate, newly capitalized, corporation called DGSI. The appellees presented evidence to the jury that Warren Essner, a senior Andersen audit partner, misrepresented DGSI's net worth as $20.8 million when it actually was negative $190 million.

DSC had transferred the repo business to DGSI in February 1982 for two related reasons. First, it wanted to satisfy banks that had been serving as DSC agents (acting for an undisclosed principal) and that had begun to demand that DSC provide adequate assurances of sufficient capital to absorb the risk of insolvency in the repo market. Evidence was introduced showing that beginning in December 1981, DSC received requests for an audited financial statement from Chase Manhattan Bank, Chemical Bank, U.S. Trust Co., and MHT. Second, DSC wanted to avoid a New York Stock Exchange audit of DSC repo capitalization. The concerns for DSC's financial stability developed as DSC began to lose money because of financial losses in its securities trading and alleged misappropriation of monies by DSC Chairman Joseph Ossorio and DSC trader David Heuwetter.2

The mechanics of DGSI's repo business are not disputed. Ossorio and Heuwetter created a so-called "Ponzi" scheme that profited from the use of coupon interest on securities sold. The essence of the scheme was DGSI's exploitation of an important difference between government securities transactions in (1) the "securities" or "cash" market, in which securities are straight-forwardly purchased and sold at market prices, and (2) the "repo market," in which government securities are purchased and sold pursuant to repo or reverse repo transactions. In the securities market, the price of a government security, such as a United States Treasury note, includes the market price of a particular issue and the accrued "coupon interest" on the security (i.e., the value of government payments due on the security at the time of the sale). In the repo market, the accrued coupon interest is paid only on the repurchase (or resale) transaction; the initial "loan" of the security is made at a price that includes only the market value of the security. Before the security is repurchased, its price will be "marked to market" periodically to reflect changed value. By borrowing increasing volumes of government bonds through reverse repos, selling them in the cash market and utilizing the cash and temporarily obtained accrued coupon interest to meet obligations on previously borrowed bonds and to conduct other trades, DGSI managed to stay solvent between February 1, 1982 (when DGSI was created, with the liabilities it had inherited from DSC) and May 17, 1982, when DGSI's ultimate collapse occurred and investors lost some $300 million.

Unlike Ossorio and Heuwetter, Warren Essner, a partner at Andersen, and Andersen itself, were not principals in this scheme. There is no evidence that they profited from it, or that they stood to gain anything from DGSI's precarious situation. Rather, Essner and Andersen had a limited role relating to the creation of DGSI. The fundamental issue in this case involves the scope of liability that flowed from this limited role.

Andersen had audited DSC in 1977 and 1978, but had no business relationship with it again until January 8, 1982. On that date, Ossorio contacted Essner at the Andersen firm regarding tax and accounting concerns in the contemplated creation of DGSI. Essner and Ossorio drafted a January 31, 1982 letter announcing DGSI's formation and discussing its capitalization. The letter was intended for the benefit of potential DGSI clients and indicated that DSC would transfer $5 million in net assets and liabilities to DGSI. (These assets and liabilities constituted repo and reverse repo positions in DSC's portfolio.) In addition, it stated that Heuwetter would invest $12.8 million and Ossorio $2.7 million, bringing the total capitalization to almost $21 million.

There was evidence that during a meeting on January 31, Heuwetter had cautioned Essner about an $11 billion "matched book"3 of repos and reverse repos that DSC controller Dennis Ruppert (who later pleaded guilty to state securities fraud charges) had fictitiously manufactured to create in part the purported $5 million transfer of DSC positions to DGSI. The true positions concealed by this false "matched book," Heuwetter testified, could not be disclosed, for fear that "if the dealer community found out the size of the positions that I was playing with ... we would be out of business the next day...."

On the evening of January 31, 1982, Essner prepared a "Statement of Subordinated Debt and Equity" to support the January 31 letter. This document did not disclose the size of the government securities positions transferred from DSC to DGSI. It reflected a purported capitalization of $20.8 million ($5 million net assets and liabilities transferred from DSC plus $15.8 million cash). There was evidence that Essner prepared the statement without consulting DSC's books and records, which, in any event, allegedly had been in incomprehensible disarray.

On and after February 1, Heuwetter delivered copies of the letter and statement to financial institutions including MHT. However, Chase and MHT pressed for an audited financial statement prepared by an independent accounting firm, even though both had been doing business with DGSI since its inception on February 1.

The parties dispute whether Ossorio asked Essner to prepare an audited statement by January 31 or during the week of February 8. In either event, on behalf of Arthur Andersen, Essner prepared a "report on specified elements of a financial statement," which purported to constitute an unqualified opinion prepared in accordance with Generally Accepted Auditing Standards ("GAAS"). There is disputed evidence, however, that Essner never had conducted an audit, and that Essner later manufactured work papers in an apparent effort to legitimize his previous reports. MHT presented evidence at trial that Essner's and Andersen's work violated several procedures required by GAAS4 as well as many of Andersen's own Audit Objectives and Procedures ("AOP").5 For preparing the letter, statement and report, Andersen billed and received from DSC $14,400. There is no evidence of any other payment to Essner or Andersen.

Andersen introduced evidence that at no time did MHT request a DGSI balance sheet as of February 1, 1982, or a DGSI balance sheet or income statement as of any later date. Andersen also introduced evidence that MHT inadequately monitored DGSI's creditworthiness and MHT's own risk exposure, in contrast to other financial institutions, several of which extricated themselves from business dealings with DGSI in time to avoid the kind of loss that MHT ultimately suffered. MHT countered with evidence that its internal controls over relevant economic risks associated with its business with DGSI was reasonable when viewed in light of prevailing industry practice.

The judge instructed the jury on seven causes of action: the three substantive securities laws, conspiracy and aiding and abetting in the violation thereof, and common law fraud and negligence. He further instructed the jury to find "either in favor of the plaintiff or the defendant." After deliberation, the jury foreman announced that "[w]e find for Manufacturers Hanover." The judge polled the jury, and then asked "which cause of action or causes of action you base this award on." Andersen did not object to this procedure. After returning to the jury room, and after some further instruction from the judge, the jury announced that "10B5 [sic], Section 12(2), 17 A [sic] and the two state statutes" had been violated. The jury was then discharged without another poll or further requests.

This case requires us to consider the role of the accountant, and the scope of his liability, in presenting to the financial community information about a financial institution seeking to attract or maintain business in transactions involving agreements to repurchase (or resell) government securities. In an SEC enforcement action arising from many of the same facts herein, we earlier held that the allegation that DSC and three of its officers, and Essner, violated section 10(b) of the 1934 Act and Rule 10b-5 thereunder and section 17(a) of the 1933 Act stated a valid federal cause of action. SEC v. Drysdale Securities Corp., 785 F.2d 38 (2d Cir.), cert. denied, --- U.S. ----, 106 S. Ct. 2894, 90 L. Ed. 2d 981 (1986). In this case, we are called upon to review jurisdictional, substantive and procedural issues arising from MHT's private action against Andersen for allegedly misrepresenting DGSI's financial status.

I. Federal Subject Matter Jurisdiction.

Andersen argues that repos are not securities, and hence that the district court lacked federal subject matter jurisdiction over this case. Whether or not repos are securities, it is clear that the district court had jurisdiction as to the claims under section 10(b) of the 1934 Act and section 17(a) of the 1933 Act. In the related litigation of SEC v. Drysdale, we assumed that repos and reverse repos are not securities, but held that it would suffice if the alleged fraud by DSC and Andersen were "in connection with the purchase or sale" of securities under section 10(b) and "in the offer or sale" of securities under section 17(a). 785 F.2d at 40-42 & n. 2.

Of course, Judge Winter's reasoning in Drysdale may not apply to MHT's claim under section 12(2) of the 1933 Act, since the underlying government securities are exempt from the proscriptions of that section. See 15 U.S.C. Sec. 77c(a)(2) (exempting securities "issued or guaranteed by the United States"); id. Sec. 77l(2) (expressly exempting securities defined in subsection 77c(a)(2)). Thus, MHT's claim under section 12(2) may be cognizable only if the repos themselves are securities. However, since we need not decide whether repos are securities to affirm as we do on the sole basis of section 10(b) and Rule 10b-5, see infra, we express no view as to whether the submission of section 12(2) to the jury was proper.

II. Section 10(b).

A. Fraud "In Connection With" Repos.

Andersen argues at the outset that the district court committed reversible error in instructing the jury that repos are securities. We disagree. Although several courts have defined repos, compare, e.g., SEC v. Miller, 495 F. Supp. 465, 467 (S.D.N.Y.1980) (characterizing repos as essentially short-term collateralized loans); In re Legel, Braswell Government Securities Corp., 648 F.2d 321, 324 n. 5 (5th Cir.1981) (same); United States v. Erickson, 601 F.2d 296, 300 n. 4 (7th Cir.) ("in substance a secured loan"), cert. denied, 444 U.S. 979, 100 S. Ct. 480, 62 L. Ed. 2d 406 (1979); Ehrlich-Bober & Co. v. University of Houston, 49 N.Y.2d 574, 577, 404 N.E.2d 726, 728, 427 N.Y.S.2d 604, 606 (1980) ("in essence, a loan transaction"), with City of Harrisburg v. Bradford Trust Co., 621 F. Supp. 463, 469-70 (M.D.Pa.1985) (repos are securities for purpose of the antifraud provisions of the 1933 and 1934 Acts); SEC v. Gomez, [Current Transfer Binder] Fed.Sec.L.Rep. (CCH) p 92,013 (S.D.Fla.1985) (same), the question of whether repos are securities remains unresolved in this Circuit. See Drysdale, 785 F.2d at 41 n. 2 (assuming without deciding that repos are not securities, but holding that fraud "in connection with" repos is cognizable under section 10(b)). We need not resolve this question to hold, as we do today, that even assuming repos are not securities, the district court's instruction that they are securities, if erroneous, constitutes mere harmless error.

Even assuming that repos are not securities, they are subject to section 10(b) and Rule 10b-5. See Drysdale, supra at 38; see also Securities Act Release No. 33-6351, 1 Fed.Sec.L.Rep. (CCH) p 2024, at 2559-2 (Sept. 25, 1981) ("The antifraud provisions [of the securities laws] ... apply to the offer, sale and purchase of U.S. government securities occurring in connection with traditional repurchase agreements"); Board of Governors of the Federal Reserve System, Division of Banking Supervision and Regulation, Letter SR-82-25 (FIS), 3 Fed.Banking L.Rep. (CCH) p 60,798.35, at 38,858 (May 13, 1981) (same).6 Viewing repos, as we must, from the perspective of their economic significance, see, e.g., SEC v. W.J. Howey Co., 328 U.S. 293, 66 S. Ct. 1100, 90 L. Ed. 1244 (1946), we think that the repos at issue herein fit squarely within the statutory language in the 1934 Act describing "contract[s] to buy, purchase or otherwise acquire securities." 15 U.S.C. Sec. 78c(a)(13). As such, and consistent with our holding in Drysdale, repos are subject to the antifraud provisions of the 1934 Act. See Blue Chip Stamps v. Manor Drug Stores, 421 U.S. 723, 750-51, 95 S. Ct. 1917, 1932, 44 L. Ed. 2d 539 (1975).

Since repos involve the purchase and sale of securities, it follows that the jury's finding of fraud "in connection with" repos was a finding of fraud "in connection with" the purchase and sale of securities. See id.; see also Abrams v. Oppenheimer Government Securities, Inc., 737 F.2d 582, 586-89 (7th Cir.1984). Indeed, even if one focuses exclusively on the "loan," as opposed to the "purchase and sale," characteristics of repos, for purposes of the antifraud provisions of the 1934 Act, "[t]he terms 'sale' or 'sell' each include any contract to sell or otherwise dispose of a security." 15 U.S.C. Sec. 78c(a)(14). The repos herein certainly satisfy that broad definition. Cf. Yoder v. Orthomolecular Nutrition Institute, Inc., 751 F.2d 555, 559 (2d Cir.1985) (Friendly, J.) (holding cognizable under antifraud provisions of 1933 and 1934 Acts fraud in connection with a "contract for the issuance or transfer of a security").

In our view, the preferred instruction would have been that repos may or may not themselves constitute securities, but that, in either event, fraud "in connection with" repos satisfies the language in section 10(b) and Rule 10b-5 relating to "fraud in connection with the purchase or sale of securities." Nonetheless, the district judge's additional comment that repos themselves are securities, if inaccurate, amounts to mere harmless error in this case.

B. Causation and Andersen's Recklessness Defense.

As for the alleged violation of section 10(b), for which accountants may be held primarily liable, see, e.g., Herman & MacLean v. Huddleston, 459 U.S. 375, 103 S. Ct. 683, 74 L. Ed. 2d 548 (1983); Competitive Associates, Inc. v. Laventhol, Krekstein, Horwath & Horwath, 516 F.2d 811, 815 (2d Cir.1975); Fund of Funds, Ltd. v. Arthur Andersen & Co., 545 F. Supp. 1314, 1351-53 (S.D.N.Y.1982), Andersen argues principally that MHT failed to demonstrate "loss causation," and that the jury was not properly instructed on this element of the section 10(b) theory. The standard for liability in a civil action under section 10(b) is causation not merely in inducing the plaintiff to enter into a transaction or series of transactions, but causation of the actual loss suffered. Chemical Bank v. Arthur Andersen & Co., 726 F.2d 930, 943 & n. 23 (2d Cir.) (Friendly, J.), cert. denied, 469 U.S. 884, 105 S. Ct. 253, 83 L. Ed. 2d 190 (1984); Marbury Management, Inc. v. Kohn, 629 F.2d 705 (2d Cir.), cert. denied, 449 U.S. 1011, 101 S. Ct. 566, 66 L. Ed. 2d 469 (1980); Schlick v. Penn-Dixie Cement Corp., 507 F.2d 374, 380 (2d Cir.1974), cert. denied, 421 U.S. 976, 95 S. Ct. 1976, 44 L. Ed. 2d 467 (1975). It is clear that on the record evidence a fact-finder could conclude that Andersen's misrepresentations as to DGSI's solvency induced MHT (and other financial institutions) to do business with the newly-formed DGSI. There was evidence that the financial community, including MHT, sought assurance of adequate capitalization of DSC amidst growing concerns of insufficient funding; that Andersen materially misstated DGSI's capitalization in its audited financial statement; and that copies of the Andersen statement were distributed in Andersen envelopes with Andersen's knowledge to various financial institutions including MHT. There was evidence that MHT Senior Vice President Stephen Goodhue, head of MHT's Wall Street Department, relied primarily, though not exclusively, on the Andersen statement in approving DGSI for its government securities business. In light of this evidence, Andersen challenges the adequacy of the charge and the sufficiency of the evidence not as to "transaction causation" but only as to "loss causation."

The requirement of "loss causation" derives from the common law tort concept of "proximate causation." See Marbury Management, 629 F.2d at 708 (citing Restatement (Second) of Torts Sec. 548A (1977)). In addition to this requirement as to the significance of the misrepresentation in the chain of causation, "loss causation" "in effect requires that the damage complained of be one of the foreseeable consequences of the misrepresentation." Id. (citing Oleck v. Fischer, [1979 Transfer Binder] Fed.Sec.L.Rep. (CCH) p 96,898, at 95,702-03 (S.D.N.Y.1979), aff'd, 623 F.2d 791 (2d Cir.1980)).

We find Andersen's argument that the district court failed to instruct the jury properly on "loss causation" to be without merit. The district judge clearly defined "proximate causation" and used that term throughout his charge. In defining the term, the judge specifically noted the requirement "that the damage was either a direct result of the misleading statement or one which could reasonably have been foreseen." Further, the judge charged that the jury must determine whether any of MHT's loss was due to its own business shortcomings; whether the bank caused its own losses, and if so, to return a verdict for Andersen even if Andersen acted improperly; whether Andersen caused part and MHT caused part of MHT's loss and, if so, to determine how much loss is attributable to Andersen; whether MHT had satisfied its burden of demonstrating that its loss was not caused by its own recklessness or negligence, and if MHT's recklessness caused any or all of its damage, then Andersen is not liable for such damage; and whether any of MHT's loss was caused not by Andersen's conduct but by MHT's own recklessness or the conduct of third parties, and if so, then MHT is not entitled to recover for such loss. Indeed, we note that the court's proximate cause charge was virtually identical to the charge requested by Andersen.

When it comes to evaluating the evidence of "loss causation," we cannot accept Andersen's argument, premised in part on Bennett v. United States Trust Co., 770 F.2d 308 (2d Cir.1985), cert. denied, --- U.S. ----, 106 S. Ct. 800, 88 L. Ed. 2d 776 (1986) that "there is simply no direct or proximate relationship between the loss and the misrepresentation." Id. at 314. In Bennett, plaintiffs borrowed funds from the U.S. Trust Co. to purchase utility stocks which were then deposited with U.S. Trust as collateral. The stock declined in value, and ultimately U.S. Trust liquidated plaintiffs' account. Plaintiffs not only lost their equity in the stock, but also became liable to U.S. Trust for $1.2 million in unpaid interest and principal on their loan. Plaintiffs sued U.S. Trust, claiming that in making the loan, U.S. Trust knowingly or recklessly misrepresented to them that the Federal Reserve margin requirements do not apply to public utility stock deposited with a bank as collateral. Plaintiffs did not and could not allege, however, that U.S. Trust in any way recommended that they purchase public utility stock in general, or any particular public utility stock. Rather, "[t]he Bennetts, and the Bennetts alone, decided to invest in public utility stock." Id. at 313-14. The same cannot necessarily be said of MHT. There was certainly evidence upon which a rational trier of fact could find, as the jury apparently found, that Andersen, by its misrepresentations, induced MHT to enter into repurchase agreements with DGSI involving the particular underlying government securities. The financial community had come to mistrust DSC's solvency before the Andersen report was issued. In this context, the Andersen report portrayed a new, highly capitalized company on whose promises to repurchase or resell particular government securities, an institution such as MHT could reasonably rely. Andersen was aware that its report was intended to be and actually was circulated to institutions including MHT for the purpose of inducing them to participate in government securities repurchase agreements.

This case is thus distinguishable from Bennett and more similar to Marbury Management, supra. In the latter case, a trainee in a brokerage firm misrepresented his expertise by claiming that he was a stockbroker and a "portfolio management specialist." That misrepresentation induced the plaintiffs to purchase and retain specific securities that the trainee recommended, despite their misgivings about the stock. Similarly, by misrepresenting DGSI's financial status, Andersen may reasonably have been found to have induced MHT to enter into repurchase agreements involving DGSI government securities, despite MHT's (and the financial community's) earlier misgivings about the financial risk associated with entering into precisely these agreements with DGSI. Further, although the misrepresentation in Marbury Management did not go to the intrinsic investment characteristics of the stock, it did go to the investment quality of the stock purchases because, had the plaintiffs known that their "broker" was an inexperienced trainee, they asserted they would not have accepted his recommendations, especially given their initial reservations. Similarly, although the misrepresentation in the present case did not go to the investment characteristics of the underlying government securities--and, under our holding in Drysdale, this was not required--it did go to the investment quality of the repos since, presumably, MHT would not have contracted with DGSI to purchase and sell government securities had MHT known of the misrepresentation regarding DGSI's finances, particularly given that the Andersen statements were in part a response to the financial community's concern regarding DSC's stability.

Andersen also relies heavily on Edwards & Hanly v. Wells Fargo Securities Clearance Corp., 602 F.2d 478 (2d Cir.1979), cert. denied, 444 U.S. 1045, 100 S. Ct. 734, 62 L. Ed. 2d 731 (1980), and argues that MHT's alleged "recklessness" caused its own loss, and that the verdict in the district court therefore contravenes the established principle that " '[t]he securities laws were not enacted to protect sophisticated businessmen from their own errors of judgment.' " Id. at 486 (quoting Hirsch v. du Pont, 553 F.2d 750, 763 (2d Cir.1977)). In our view, Andersen misapplies this principle in seeking to resurrect imperfections in MHT's internal operations as a bar to recovering for losses caused by Andersen's misrepresentations. Cf. Faller Group, Inc. v. Jaffe, 564 F. Supp. 1177, 1181 (S.D.N.Y.1983) (failure to initiate certain investigative checks does not bar action for fraud) (citing Mallis v. Bankers Trust Co., 615 F.2d 68 (2d Cir.1980) (Friendly, J.), cert. denied, 449 U.S. 1123, 101 S. Ct. 938, 67 L. Ed. 2d 109 (1981); Dupuy v. Dupuy, 551 F.2d 1005 (5th Cir.) (Wisdom, J.), cert. denied, 434 U.S. 911, 98 S. Ct. 312, 54 L. Ed. 2d 197 (1977)).

We note the applicable legal standard regarding plaintiff's conduct. Plaintiff's burden of persuasion "is simply to negate recklessness when defendant puts that in issue, not to establish due care." Mallis, 615 F.2d at 79 (footnote omitted). Thus, although Edwards & Hanly had expressly "approach[ed] the issue of [plaintiff's] conduct from the standpoint of ... due diligence," 602 F.2d at 485, the applicable standard is recklessness. We note, however, that Judge Friendly asserted in dicta that "[t]he facts in Edwards & Hanly would have led to dismissal under the Dupuy [recklessness] standard." Mallis, 615 F.2d at 79 n. 9.7

In our view, the evidence adduced at trial demonstrates that Andersen's reliance on Edwards & Hanly is misplaced, and its arguments of recklessness on MHT's part unpersuasive. We note that plaintiff Edwards & Hanly, a brokerage house, was denied recovery because, in that case, "[t]he primary cause of E & H's loss was its failure to comply with Regulation T, ... [promulgated] pursuant to Sec. 7 of the Exchange Act, 15 U.S.C. Sec. 78g." 602 F.2d at 486. Regulation T expressly requires broker-dealers, such as E & H, to ensure that a security purchased in a cash account is "held in the account" or that "the creditor accepts in good faith the customer's statement that the security is owned by the customer or the customer's principal, and that it will be promptly deposited in the account." See 12 C.F.R. Sec. 220.8(a)(2). In the present case, however, Andersen has not alleged any particular regulatory violation by MHT, which, in any event, was acting not as a broker purchasing or selling securities for its customer, but as fiscal agent for an undisclosed principal, providing for its customers short-term repo and reverse repo positions in government securities, and assuming responsibility to its customers for DGSI's obligations. Of course, MHT may be denied recovery on the basis of adequate proof of common law recklessness as assessed in light of industry practice. We do not wish to suggest that a defendant must show a violation of some statutory or regulatory proscription to show recklessness. But the lack herein of an established statutory or regulatory cognate to Regulation T, violation of which was central to our holding in Edwards & Hanly, forces us to inquire as to what, if anything, in this case stands as similarly compelling proof of recklessness.

Andersen's arguments on this critical issue rest principally on its comparisons of various aspects of MHT's DGSI transactions with those of other financial institutions that managed to escape the DGSI collapse either unscathed or considerably less harmed than was MHT. We find these arguments ultimately unpersuasive. For example, Andersen argues that MHT recklessly failed adequately to control its own volume and other risks. Volume risks relate to the amount of money that MHT invested in repo and reverse repo positions with DGSI at any one time. Although some financial institutions, such as Merrill Lynch, Dean Witter Reynolds, Salomon Brothers, First Boston, Citibank and Prudential-Bache, determined that they would transact from zero to $50 million worth of repo business with DGSI, based principally on its cash reserves, there was also evidence that other reputable firms, notably U.S. Trust Company and Chase Manhattan Bank, transacted some $2 to $2.5 billion in repo business with DGSI, as did MHT. Thus, industry practice as to volume risk was not nearly as uniform as Andersen would suggest. Further, volume risk is but one element in total risk exposure, and its significance in the calculus of total risk exposure may be diluted to the extent that the primary government securities dealer, here, DGSI, maintains matched books of repos and reverse repos, a practice that DGSI fraudulently purported to maintain. There was evidence regarding the government securities business of other primary dealers, W.E. Pollack, Prudential-Bache, and Dean Witter, suggesting that the volume of business transacted by DGSI generally and with a particular customer, such as MHT, might have comported with the practices of those firms had matched books in fact been sustained.

As to non-volume risk controls, Andersen introduced evidence that MHT failed adequately to monitor the following: the "market risk" that the cash and bonds which it received from DGSI might be worth less than the cash and bonds delivered to DGSI; the "coupon interest risk" that DGSI might fail to pay accrued coupon interest when due or to return borrowed bonds; and the "repo interest risk" that DGSI might fail to pay repo interest owed. But there was also evidence--certainly enough to allow a jury to find the claim of recklessness negated--that MHT monitored these risks reasonably when measured by prevailing industry practice. As to "market risk," there was evidence that MHT followed the common industry practice of "marking to market" at 100% of the stated par value of the securities. U.S. Trust, the only other intermediary bank that testified as to risk controls, apparently marked to market in the same way as MHT: the intermediary bank would not initiate marks, but would process the marks requested by the principals to the transactions. There was also evidence that MHT periodically checked to ensure that the DGSI transactions were being marked to market. In our view, the evidence of these internal controls by MHT was sufficient to justify a jury's negating the charge of recklessness as to market risk, whether or not MHT could have or should have actually initiated, as opposed to merely processed and periodically monitored, markings to market. Cf. Faller Group, Inc., 564 F. Supp. at 1181 (failure to initiate further investigative checks did not bar action for fraud) (citing Mallis, supra; Dupuy, supra ).

As to "coupon interest risk," there was evidence that MHT's handwritten records were transcribed from computerized risk assessments that were substantially similar to Andersen's own computer assessments of accrued coupon risk, except that MHT's program entailed a one-day lag in calculation and did not account for DGSI "fails" in tendering cash or securities due by the time the Fed wire closes at the end of a business day under reverse repos until MHT employee D'Amore corrected by hand the computer assessment each day to account for such fails. There was also evidence that MHT controlled coupon risk by demanding that DGSI process more repos than reverse repos through MHT, thereby reducing risk since, with repos, the borrower was obligated to remit coupon interest to DGSI, while with reverse repos, DGSI was obligated to remit coupon interest to the lender. Further, there was evidence that in monitoring coupon risk, MHT set up a collateral account into which DGSI was to deposit cash and securities to cover MHT's coupon interest exposure. Andersen's own expert on repo procedures, Robert Bird, conceded that prior to DGSI's collapse, it was not the industry practice to collateralize accrued coupon interest, and that MHT's coupon risk controls, including its collateral account, constituted "a fair effort to attempt to control the coupon exposure." Certainly these internal controls render MHT considerably more conscientious than E & H, which often failed even to mark the defendant's sales "long"--a failure "which alone might not even be a breach of duty." Edwards & Hanly, 602 F.2d at 486 (citing Naftalin & Co. v. Merrill Lynch, Pierce, Fenner & Smith, 469 F.2d 1166, 1175 (8th Cir.1972)). Also, even assuming, arguendo, that Andersen is correct that MHT often failed to enforce this collateral control procedure when DGSI ignored requests for additional collateral, the mere institution and partial enforcement of the collateral control procedure appears to have rendered MHT ahead of, not behind, then-prevailing industry practice in monitoring coupon interest risk.

As to "repo interest risk," there was evidence only that at least two other firms had developed computer programs to monitor this risk, the importance of which must not be overstated in the total calculus of risk, and that MHT was beginning to develop its own program when the collapse occurred.

In short, applying the overall standards of industry practice, the jury reasonably could conclude that MHT endeavored to control risk sufficiently to negate Andersen's claim of MHT's recklessness.

Andersen also argues that MHT recklessly failed to heed warning signs of DGSI's impending collapse. In our view, Andersen relies improperly on Edwards & Hanly and thereby mischaracterizes MHT's immediate response to warning signals as a longer-term organizational policy of conscious avoidance. During the approximately one month preceding DGSI's collapse, DGSI often failed to return securities due. As Andersen points out, in April 1982 "fails" lasting three or more days occurred on 58 percent of DGSI's scheduled returns. But the countervailing evidence was sufficient to allow a finding that the charge of recklessness had been negated: MHT officers Martello, D'Amore and Zinns apparently worked with DGSI in April and May in an effort to reduce exposure risks; and, in May 1982, MHT was in daily communication with DGSI regarding the former's exposure. Further, there was evidence that in MHT's experience, DGSI consistently cured fails, and that, in any event, only fails that occurred in the week prior to the collapse of DGSI were still open on May 17, 1982. This contrasts with Edwards & Hanly, where, in response to fails of up to several months for securities due within several days, repeatedly over a period of approximately one year, plaintiff E & H made only "sporadic inquiries" and even then apparently contented itself with evasive answers. Edwards & Hanly, 602 F.2d at 486-87. Finally, we must recognize the sheer speed of events herein in contrast to Edwards & Hanly. While E & H had approximately one year during which it could have protected itself from ongoing fails, here MHT had at most one-quarter to one-third that duration from the time of the circulation of the Andersen statement in February 1982 until the ultimate collapse of DGSI in May 1982. Indeed, the duration must be considered shorter still when measured from the time when fails became a legitimate source of concern. In sum, as to the brief period of MHT's interaction with DGSI, the evidence supports the conclusion that MHT endeavored to protect its investments, and negated the charge of recklessness.

III. Section 17(a) of the 1933 Act.

Since we affirm the judgment of the district court on the basis of section 10(b) of the 1934 Act, we need not reach the issues raised herein as to whether there is an implied private right of action under section 17(a) of the 1933 Act, and, if so, whether liability thereunder may be predicated on the basis of mere negligence. See Zerman v. Ball, 735 F.2d 15, 23 (2d Cir.1984) (declining to decide whether a private right of action exists under section 17(a) of the 1933 Act since plaintiff might instead prevail on the basis of Rule 10b-5). However, nothing in our opinion today should be construed as commenting in any way on Judge Friendly's recent admonition that our conclusion in Kirshner v. United States, 603 F.2d 234, 241 (2d Cir.1978), cert. denied, 442 U.S. 909, 99 S. Ct. 2821, 61 L. Ed. 2d 274 (1979), that such a private right of action exists, "may be open to reexamination." See Yoder, 751 F.2d at 559 n. 3.

IV. The Verdict.

Much is made by both sides on this appeal of the import of the procedures surrounding the return of the jury verdict. In our view, the procedures employed by the district judge do not warrant reversal, as Andersen now urges, or a review of five allegedly valid verdicts, as MHT now urges. Rather, we think it appropriate to review, as we have done, only the substantive federal securities cause of action, namely, the claim, made under section 10(b), as to which we can say with certainty that a clear and explicit verdict has been rendered. Cf. Morris v. Pennsylvania R. Co., 187 F.2d 837, 841 (2d Cir.1951) (errors in verdict procedure should be "localized so that the sound portions of the verdict may be saved") (Clark, J.).

On the morning of the jury charge, the district judge informed counsel that he "did not find satisfactory any form of [special] verdict sheet." The judge charged the jury on the three federal substantive securities claims, the federal "subsidiary theories" of conspiracy and aiding and abetting and the two state common law theories of fraud and negligence. He instructed the jury to find "either in favor of the plaintiff or the defendant, and if your verdict is in favor of the plaintiff you are to state the amount of damage that you find." Later, he sent into the jury room copies of the conspiracy and aiding and abetting charge. At 5:48 p.m., the jury announced its verdict, as follows:

THE FOREMAN: We find for Manufacturers Hanover.

THE CLERK: In what amount?

THE FOREMAN: 17 million.

Judge Owen then asked the jury "to advise the court on the record of which cause of action or causes of action you base this award on." Counsel for neither side objected. The jury returned to the jury room. A few minutes later, the jury asked by note whether it was to identify by number the statutes that had been violated or to explain its reasoning. The court responded, "Just state by number." Although it is only of contextual interest, at this time the jury apparently had in the jury room that portion of the charge reciting the language of Rule 10b-5 and Secs. 10(b), 12(2) and 17(a) and the charge on aiding and abetting and conspiracy, but not on fraud and negligence. At approximately 6:17 p.m., the jury returned, and the following colloquy ensued:

THE COURT: Mr. Foreman, you are prepared to state which of the statutes were violated?

THE FOREMAN: Yes, your Honor. We feel that 10B5 [sic], Section 12(2), 17 A [sic] and the two state statutes.

THE COURT: Were all violated?

THE FOREMAN: Were all violated.

Without further questions, polling, or objection, the judge then excused the jury.

Andersen contends that the judge's request for the bases of the jury's verdict was improper and thus failed to convert to a special verdict the jury's initial general verdict, as to which a reversal on any one claim would necessitate a new trial. We note that Andersen failed to object to the request, and thus has not preserved this issue for appeal. See, e.g., Turchio v. D/S A/S Den Norske Africa, 509 F.2d 101, 105 (2d Cir.1974). Andersen's response to this point--that the request was a nullity, and hence required no objection--begs the very question that Andersen itself raises as to whether the request was legally a nullity. Whether a request regarding a verdict is timely is a proper legal question. Cf. Baker v. Sherwood Construction Co., 409 F.2d 194, 195 (10th Cir.1969) (per curiam) (too late to poll the jurors individually after clerk read verdict and jury was polled collectively and excused). In general, counsel must object to any procedure that it wishes to raise on appeal. See Fed.R.Civ.P. 46; Robinson v. Shapiro, 646 F.2d 734, 742 (2d Cir.1981). Further, the failure to object precludes Andersen from contesting the oral, as opposed to written, form of the inquiry. Turchio, 509 F.2d at 105.

As to the propriety of the judge's request, our standard of review is the abuse of discretion standard. See Cann v. Ford Motor Co., 658 F.2d 54, 58 (2d Cir.1981), cert. denied, 456 U.S. 960, 102 S. Ct. 2036, 72 L. Ed. 2d 484 (1982). We note first that the request was not "tantamount, in its effect, to a direction to the jury" to alter or otherwise compromise the decisiveness of its ultimate verdict. Compare McCollum v. Stahl, 579 F.2d 869, 871 (4th Cir.1978) (resubmission of interrogatories as to damages and defendant's conduct after jury found no wrongful conduct by defendant held improper), cert. denied, 440 U.S. 912, 99 S. Ct. 1225, 59 L. Ed. 2d 460 (1979). Further, the request for the legal basis for the verdict did not result in the rendering of a special verdict under Fed.R.Civ.P. 49(a). The term "special verdict" is a term of art, and properly speaking, refers only to special findings regarding factual issues that the court may ask the jury to resolve, see Griffin v. Matherne, 471 F.2d 911, 917 n. 6 (5th Cir.1973); Skidmore v. Baltimore & Ohio R. Co., 167 F.2d 54, 65-70 (2d Cir.), cert. denied, 335 U.S. 816, 69 S. Ct. 34, 93 L. Ed. 371 (1948); 9 C. Wright & A. Miller, Federal Practice and Procedure Sec. 2506 at 498-502 (1971), or perhaps mixed questions of law and fact, assuming applicable legal standards are charged. See Wright & Miller, supra at 502. By contrast, Judge Owen simply asked the jury to specify the cause or causes of action upon which it based its verdict; in short, he inquired as to the legal basis or bases of its verdict. Again, we note counsel's failure to object to the oral form of the inquiry or the answers.

In any event, we find no basis for concluding that this sequence resulted in post hoc rationalization by the jury. See Litton Systems, Inc. v. American Tel. & Tel. Co., 700 F.2d 785, 802-04 (2d Cir.1983) (upholding post-verdict resubmission of unresolved mixed questions), cert. denied, 464 U.S. 1073, 104 S. Ct. 984, 79 L. Ed. 2d 220 (1984). In our view, the potential for inaccuracy or mere rationalization might be significant with factual interrogatories, since in reaching its verdict, the jury may not collectively have resolved the particular factual questions posed. By contrast, with a post-verdict inquiry into the legal basis for a verdict, as in the present case, it is virtually inconceivable that the jury, having deliberated and returned a verdict, had not already considered the question posed and thus would provide a mere rationalization in its response to the post-verdict inquiry.

In his initial charge, Judge Owen effectively had instructed the jury to consider the very issue presented in the post-verdict inquiry--the seven possible bases for any finding of liability--and had instructed the jury on the elements of each of the seven causes of action. Juries are presumed to follow instructions. United States v. Siegel, 717 F.2d 9, 19 (2d Cir.1983) (citing Watkins v. Sowders, 449 U.S. 341, 347, 101 S. Ct. 654, 658, 66 L. Ed. 2d 549 (1981)). Andersen offers no reason to believe that the jury did not consider the seven causes of action charged in reaching its verdict. Further, Andersen certainly has shown no reason to suspect that, once Judge Owen submitted the post-verdict request to the jury, the jury did not earnestly consider, resolve, and report the results of its deliberations. See Litton Systems, 700 F.2d at 803-04 (no basis to conclude that jury did not deliberate conscientiously over post-verdict resubmission of unresolved law/fact interrogatories). Thus, the post-verdict question does not undermine the fairness or decisiveness of the verdict.8

The legal inquiry herein is also justifiable on grounds similar to those advanced in support of a separate post-verdict interrogatory on damages, see 5A J. Moore, Moore's Federal Practice p 49.06, at 2236 (1971), or a jury's sua sponte itemization of damages returned with its general verdict, see Dagnello v. Long Island Railroad Co., 193 F. Supp. 552, 554 (S.D.N.Y.1960) (Weinfeld, J.), aff'd, 289 F.2d 797 (2d Cir.1961). In Dagnello, Judge Weinfeld justified the sua sponte itemization thusly:

Since the jury's method of computing the damages is before the Court, it is desirable that it remain as of record so that if the defendant is advised to prosecute an appeal, the reviewing court will have knowledge of the basis upon which the jury reached its judgment.

Id. It was because the jury's additional finding went to the legal "basis" of the verdict that the finding was legitimate in Dagnello as in the instant case.

Finally, in the present case, which entailed six weeks of trial time to resolve complex securities issues pleaded on seven distinct theories, the jury's post-verdict specifications play an important role in rationally maximizing the use of scarce judicial resources. See Berkey Photo Inc. v. Eastman Kodak Co., 603 F.2d 263, 279 (2d Cir.1979) (urging use of special verdicts and interrogatories in "large and complex cases" to resolve issues, reduce probability of "laborious and expensive retrial," and facilitate appellate review), cert. denied, 444 U.S. 1093, 100 S. Ct. 1061, 62 L. Ed. 2d 783 (1980). We wish to emphasize that the preferred procedure, in the interest of avoiding confusion or unnecessary use of trial time, would have been for the district judge to request a statement of the bases for the verdict in his initial instructions and to provide the jury with a written interrogatory form. However, it is clear that in some instances additional instructions may be given even after a jury begins deliberations or returns a verdict. See, e.g., Litton Systems, 700 F.2d at 802-04. The particular inquiry herein did not portend manifest injustice but rather promoted a more complete record; hence, we do not hold the procedure invoked invalid. On the contrary, the procedure was preferable to allowing a lengthy and complex trial to conclude with one general verdict without any indication as to the legal basis or bases for that verdict. Such an unexplained verdict would risk wasting precious judicial resources since, as Andersen argues, it might be overturned for, inter Alia, any serious error in the charge relating to a single claim since the reviewing court would be unable to determine the legal basis of the verdict. See, e.g., Greenbelt Cooperative Publishing Assoc. v. Bresler, 398 U.S. 6, 11, 90 S. Ct. 1537, 1540, 26 L. Ed. 2d 6 (1970).

The judgment herein must be sustained if any one of the indicated bases for the verdict is sustained. See, e.g., Turner v. United States, 396 U.S. 398, 420-21, 90 S. Ct. 642, 654, 24 L. Ed. 2d 610 (1970). Since, as discussed, we affirm on the basis of the claim under section 10(b), one of the bases on which the jury clearly returned its verdict, the verdict procedure poses no bar to our affirming the judgment of the district court.

As to the jury's reference to "state statutes" in its answer to Judge Owen's request, we are unable to say whether this inaccuracy reflects a misstatement in the word "state" or the word "statute." We find unpersuasive MHT's contention that the jury must have meant not the state common law theories but the federal conspiracy and aiding and abetting claims because the jury had the charge on those claims in the jury room and the court had charged that conspiracy and aiding and abetting are "subsidiary theories" to the substantive federal claims. That argument strikes us as little more than an educated guess. The integrity of the jury system and the concern for fairness to litigants simply cannot accommodate such guesswork in sustaining a verdict. In any event, since we affirm on the basis of section 10(b), we need not consider whether Andersen may be liable as an aider and abettor as a matter of law, notwithstanding the verdict procedure, or whether any of the other theories of liability support the jury's verdict.

The parties raise several additional issues, which we can dispose of briefly. First, Andersen argues that the empanelment of the jury was fundamentally unfair because the district judge expressly excused venire persons who could not sit for a trial likely to last several weeks. A jury selection procedure is unconstitutional if it "systematically and intentionally" discriminates or otherwise deprives a litigant of a chance to have his case heard by a cross-section of the community. See Thiel v. Southern Pacific Co., 328 U.S. 217, 224, 66 S. Ct. 984, 987, 90 L. Ed. 1181 (1946) (automatic exclusion of wage-earners who earn less than four dollars per day held unconstitutional). In the present case, since there was no per se exclusion of any group, and no systematic and intentional discrimination against any person or persons, the district judge's empanelment procedure was not unconstitutional.

Andersen also contests the award of pre-judgment interest. The district judge granted MHT's motion to amend the judgment to provide for costs and post-judgment interest, which Andersen did not contest, and for pre-judgment interest. As to pre-judgment interest, in a memorandum and order dated September 12, 1985, the court specifically based its award on the theory that, under N.Y.Civ.Prac.L. Sec. 5001, "[t]he state fraud and negligence findings mandate an award of pre-judgment interest." In our view, the post-verdict inquiry did not result in a legally ascertainable finding of liability under the state common law causes of action, and we instead affirm solely on the basis of section 10(b). Therefore, we cannot say whether the district court order of pre-judgment interest was intended to apply herein. Pre-judgment interest on federal securities claims, unlike on New York state law claims, is not mandatory. See, e.g., Blau v. Lehman, 368 U.S. 403, 414, 82 S. Ct. 451, 457, 7 L. Ed. 2d 403 (1962).

We therefore remand to the district court the issue of whether and to what extent pre-judgment interest is appropriate in this case. The purpose of pre-judgment interest is to compensate a plaintiff for loss of the use of funds ultimately awarded. See, e.g., Freschi v. Grand Cole Venture, 767 F.2d 1041, 1051 & n. 12 (2d Cir.1985). Thus, should the district court determine that pre-judgment interest is appropriate, it may further consider the extent, if any, to which MHT had pre-trial use of "deferred credit" funds or of funds received upon release to MHT on May 17, 1982 of certain collateral securities, and the extent, if any, to which MHT's pre-award tax savings on losses due to the DGSI collapse exceed post-award tax liability on the judgment award. Thus, we remand to the district court for resolution of these issues.

Finally, we deny MHT's cross-appeal for a mini-trial on punitive damages. First, the district judge was within his discretion in denying MHT's motion for an amended complaint seeking punitive damages as both untimely and unwarranted. See Fed.R.Civ.P. 16; see also 3 J. Moore, Moore's Federal Practice p 16.19, at 16-74 (1985). Second, in any event, it is well settled in this Circuit that proof of fraud or misrepresentation in a private civil action, even if such proof would constitute grounds for punitive damages under common law, generally warrants no such damages under the federal securities laws. See, e.g., Green v. Wolf Corp., 406 F.2d 291 (2d Cir.1968), cert. denied, 395 U.S. 977, 89 S. Ct. 2131, 23 L. Ed. 2d 766 (1969); Globus v. Law Research Service, Inc., 418 F.2d 1276, 1283 (2d Cir.1969); see also 3A H. Bloomenthal, Securities and Federal Corporate Law Sec. 8.28 (1985) (citing cases).

We have considered the remaining arguments raised herein, and we find them to be without merit.

In light of the foregoing, we affirm the judgment of the district court--excluding its award of pre-judgment interest--on the basis of section 10(b) and Rule 10b-5 thereunder. The cause is remanded to the district court solely for resolution of Andersen's claims as to pre-judgment interest in a manner not inconsistent with this opinion.

Judgment affirmed in part. Award of pre-judgment interest vacated and remanded.