Brandt v. Wand Partners

BOUDIN, Circuit Judge. This case arises out of the failure and chapter 7 bankruptcy of Healthco International, Inc. (“Healthco”), a major global distributor of dental products and services. Following this debacle, the chapter 7 trustee brought the present case on behalf of the estate against numerous parties alleged to have been responsible for, or beneficiaries of, the leveraged buyout that precipitated the collapse of Healthco. We begin with a short history of the transactions and proceedings, and then address the claims on appeal made by the bankruptcy trustee, William Brandt. 1 *10 I. Factual Background In the late spring of 1990, Gemini Partners, L.P., a Delaware limited partnership that owned 9.96% of Healthco’s common shares, formed the Committee for Maximizing Shareholder Value of Healthco International (“the Committee”) and began a proxy contest to remove Healthco’s incumbent directors. In response, Healthco engaged Lazard Fréres & Co. LLC as its financial advisor and sought to arrange the company’s sale to another buyer. On September 4, 1990, Healthco entered into a merger agreement with affiliates of Hicks, Muse & Co. (“Hicks, Muse”), a Dallas-based investment firm. Under the agreement, a company formed by Hicks, Muse would merge with Healthco after acquiring its stock at a price of $19.25 per share. After reaching this agreement, in mid-September Healthco negotiated a separate settlement agreement with Gemini and the Committee, under which three Committee nominees became members of Healthco’s seven-member board. The settlement agreement provided that, if the merger agreement was terminated or sufficient progress toward a sale of the company was not subsequently made, the Committee could increase its share of the board from three out of seven to five out of nine. As a further spur to a merger or sale, Gemini promised each Committee director $24,000, less director compensation, if Gemini sold its shares at a profit. In February 1991, Hicks, Muse’s initial plan for a leveraged buyout 2 (“LBO”) of Healthco fell apart after Healthco’s annual physical inventory indicated that the company’s unaudited 1990 earnings were several million dollars lower than expected. Healthco’s auditors, Coopers & Lybrand L.L.P., later certified financial statements that revealed a 1990 net loss of just over $5 million and 1990 earnings of less than $22 million. Given such figures, Hicks, Muse determined the $19.25 share price was too high, and the parties set to work drawing up a new plan. On March 26, 1991, Healthco’s board voted 5-2 to approve a new merger plan involving Hicks, Muse affiliates. Marvin Cyker, Healthco’s chief executive officer and board chairman, who held stock options but no outstanding shares in Health-co, was one of the two board members who voted against the transaction. The proposal was for a tender offer for Healthco stock at $15 per share to be made under Hicks, Muse’s auspices, with financing by other parties, after which Healthco would merge with a netv entity controlled by the new investors. Lazard Fréres advised that the transaction was fair to Healthco stockholders. On April 2, a tender offer for Healthco stock was made by HMD Acquisition Corp., a wholly-owned subsidiary of Healthco Holding Co.; Healthco Holding was itself a company set up by Hicks, Muse to be the recipient of $55 million of the Hicks, Muse investors’ funds. Additional funds for the merger were to be supplied by a bank group that would provide a $50 million tender facility an available loan) in exchange for perfected first priority liens on HMD Acquisition’s shares in Healthco. Another group of in *11 vestment entities were to provide $45 million in cash, in exchange for subordinated debt. The tender offer was successful. HMD Acquisition acquired more than 90% of Healthco’s stock in the tender offer. Among the stockholders who tendered shares or options in the merger were Marvin Cyker, who received over $1 million for his stock options, and J.P. Morgan & Co., an investment firm that had held 17.3% of Healthco’s shares (13.0% on a fully diluted basis, ie., after the exercise of options) as a record shareholder for its clients. The buyout of Healthco was completed on May 22, 1991, through a short-form, cash-out merger, Del.Code Ann. tit. 8, § 253 (1999), in which HMD Acquisition Corp. was merged into Healthco. Health-co’s remaining original stockholders received $15 per share in exchange for their holdings. The Hicks, Muse investors, by contrast, were now largely dependent on Healthco’s fate. As a result of the merger, Healthco, the surviving company, inherited all of HMD Acquisition’s debts — namely, the multimillion-dollar debts owed to non-equity investors (e.g., the banks) who helped to finance the Healthco buyout. After the merger, Healthco’s financial situation steadily deteriorated. (It is unclear to what extent this was due to preexisting problems and to what extent the situation was aggravated by new debt.) In the spring of 1992, Healthco was placed on credit hold by a large European supplier, and by June 1992 more than forty of Healthco’s suppliers were refusing to ship it goods until they were paid for past receivables. After defaulting on several loan covenants, Healthco faced an inereas-ingly hostile relationship with the bank group that had financed the tender facility for the buyout. On June 9, 1993, Healthco filed a petition for chapter 11 bankruptcy in the federal bankruptcy court in Massachusetts. 11 U.S.C. § 301 (1994). That September, after declining to approve a new borrowing arrangement, the bankruptcy court granted Healthco’s motion for conversion to chapter 7 bankruptcy. Id. § 1112(a). The buyout of Healthco, which had possessed assets of greater than $300 million at the time of the merger, had ended in a liquidation proceeding that yielded less than $60 million, far less than what was needed to pay off Healthco’s creditors. On June 8, 1995, Brandt, as Healthco’s chapter 7 trustee, began the present adversary proceeding in the bankruptcy court, implicating almost all of those involved in the merger transaction and ultimately claiming around $300 million in damages. Brandt’s 22-eount complaint made 12 claims for fraudulent transfers (counts I-XII). 3 Brandt also alleged four counts of breach of fiduciary duty, or of aiding and abetting the same (counts XIII-XVI). 4 Finally, Brandt charged Coopers & Lybrand with accounting malpractice; accused Lazard and Valuation Research Corporation, a financial advisor of Hicks, Muse, of negligence; claimed that all the tendering shareholders were unjustly enriched and benefitted from a commercially unreasonable distribution; and alleged that the bank group was liable because of the commercially unreasonable way in which it liquidated its collateral (counts XVII-XXII). *12 Proceedings in the bankruptcy court were extensive during the balance of 1995 and throughout 1996. In addition to discovery (and discovery disputes), there were several amended complaints by Brandt, dismissal or summary judgment grants in favor of various defendants on specific claims, and efforts (generally unsuccessful) by Brandt to get interlocutory review on various rulings in the district court. Although it became clear in 1996 that a jury trial would likely be required in the district court on certain claims, the bankruptcy judge continued to oversee the matter. In early 1997, the district court began to move the remaining claims toward trial. Brandt then reached a settlement with the bank group defendants and later filed a fourth amended complaint streamlining various of the claims that remained. Shortly before trial, Brandt settled his claim with Coopers & Lybrand. Except for claims against Lazard (where jury trial had been waived), the remaining claims were tried to a jury in a 27-day trial from April 23 until June 6,1997. Brandt lost on every claim tried to the jury, and the district court found in favor of Lazard. Brandt now appeals on numerous issues. Importantly, these include the dismissal by the bankruptcy court of key fraudulent transfer claims, its grant of summary judgment for various defendants as to the unjust enrichment claims against them, the district court’s disposition of certain fiduciary duty claims, and miscellaneous claims relating to discovery and the conduct of the trial. The details, and certain concerns about our jurisdiction, are discussed in connection with each set of claims. II. Fraudulent Transfer Dismissals We start with Brandt’s effort to revive the fraudulent transfer claims that the bankruptcy judge dismissed. In essence, Brandt’s theory of fraudulent transfer is that because Healthco’s assumption of HMD Acquisition’s liabilities meant that Healthco’s assets became collateral for the debt that financed the buyout, both the tendering shareholders and the financiers obtained proceeds from a “fraudulent” transaction that deprived Healthco’s preexisting unsecured creditors of most of the value of the company’s assets. The bankruptcy court refused to see the transaction as a stripping of Healthco assets. Rejecting Brandt’s call to “collapse” the multi-step buyout into a transfer of Healthco’s assets to shareholders and buyout financiers, the bankruptcy court repeatedly held that “funds transferred by [HMD] Acquisition prior to the effectiveness of the merger [were] not transfers by [Healthco] and hence are immune from fraudulent transfer attack.” Brandt, 201 B.R. at 21. It is this refusal to “collapse” the leveraged buyout, and hence treat the payments in question as ones made in substance (although not in form) out of Healthco’s assets, which is the focus of Brandt’s challenge on appeal. Whether the transaction should have been “collapsed” appears to be a difficult issue of state law (the parties do not agree on which state or states supply the law) on which there is fairly limited precedent. 5 *13 Of course, there are similar problems in other areas (e.g., tax law, see True v. United States, 190 F.3d 1165, 1176-77 (10th Cir.1999)), and there are countless difficult arguments in policy presented by the request to collapse the buyout. Indeed, the bankruptcy court itself, in dealing with directors’ obligations of loyalty, recognized that Healthco’s assets were security for the transaction’s financing and described as “myopic” the defendants’ argument that the buyout transaction should be analyzed only in terms of its separate parts. Brandt, 208 B.R. at 302. We conclude, however, that the issue is not properly before us because our authority is limited to review of judgments by the district court and Brandt never secured a district court judgment resolving any of the fraudulent transfer claims. Abbreviating the history, the story begins with the bankruptcy court’s orders of October 27, 1995, granting motions to dismiss on the basis of a bench ruling from the prior day. The dismissals were of fraudulent transfer claims against various defendants who were for the most part tendering shareholders in the multi-step buyout: J.P. Morgan & Co., the Airlie Group defendants (a limited partnership and several individuals who owned approximately 10% of Healthco’s stock), J. Robert Casey, and Helen and Marvin Cyker. On November 6, 1995, Brandt sought leave to appeal these dismissals as interlocutory orders, 28 U.S.C. § 158(a); Fed. R. Bankr.P. 8003. On June 27, 1996, the district court denied this motion. Brandt then asked the bankruptcy court to certify the dismissals for an appeal under Federal Rule of Bankruptcy Procedure 7054(a), the bankruptcy counterpart of Federal Rule of Civil Procedure 54(b), but the bankruptcy court denied this motion. Later the bankruptcy court issued orders dismissing further claims of fraudulent transfers to Healthco shareholders, subordinated preferred shareholders, and others. Again Brandt did not secure review by the district court. At this stage, the bankruptcy court’s orders were dismissals of claims on the merits but were not final (and therefore not immediately appealable as of right). 28 U.S.C. § 158(a). The bankruptcy court has authority to deny on the merits claims that are within its core authority, and one proceeding so listed is the voiding of fraudulent conveyances. Id. § 157(b)(2)(H). Even if for some reason the claims at issue are not within this rubric (the parties have not briefed the issue and we do not decide it), Brandt did not contest the bankruptcy court’s power to dismiss on the merits, so there was also jurisdiction by consent. See In re G.S.F. Corp., 938 F.2d 1467, 1476-77 (1st Cir.1991). See generally 28 U.S.C. § 157(c)(2); Commodity Futures Trading Commission v. Schor, 478 U.S. 833, 848-50, 106 S.Ct. 3245, 92 L.Ed.2d 675 (1986). The finality issue is complicated. Although a “final judgment” rule of some kind applies to appeals ifrom the bankruptcy court to the district court (with exceptions for certification and leave of the court), the concept of finality is more flexibly applied than with regard to 'district court judgments, In re American Colonial Broad. Corp., 758 F.2d 794, 801 (1st Cir.1985); this approach recognizes that complex bankruptcies are often an umbrella for a multitude of claims between different parties and, thus, that the strict requirement of final judgment used in district court appeals — the resolution of all claims as between all parties — could delay for years district court review of matters that are essentially final as between the parties concerned in a bankruptcy. The difficulty is that despite some agreement as to which actions are final or not final, no uniform and well-developed set of rules exists and on many points there is a good deal of uncertainty. See 1 *14 Collier on Bankruptcy § 5.07 (Lawrence P. King ed., 15th ed. 2000); cf. In re Public Serv. Co. of N.H., 898 F.2d 1, 2 (1st Cir.1990) (noting “strong analogies” between adversary proceedings and ordinary district court cases, and suggesting that a bankruptcy court’s partial summary judgment order was not final). In this case, it appears that most defendants who had fraudulent transfer claims dismissed by the bankruptcy judge still had other limited claims (e.g., unjust enrichment) pending against them (subordinated debtholders who helped finance the buyout possibly being the only significant exceptions). Furthermore, all the dismissed claims were substantially related to those that remained before the lower courts. Indeed, it is seemingly for these reasons that the bankruptcy court and district court resisted interlocutory review or certification. Brandt himself thought the dismissals were interlocutory at the time the orders were entered, and no one has disputed that view. We thus proceed on that premise, without any further effort to develop clear-cut rules in this difficult area. Eventually, the district court withdrew its reference to the bankruptcy court as to various components of the case so that it could dispose of a number of claims that required a jury trial (together with a parallel jury-waived claim against Lazard). Possibly at this time Brandt could have taken the position that the transfer of other claims as to the defendants in question rendered final the earlier orders dismissing the fraudulent transfer claims. In that event, Brandt would have had ten days following the termination of the reference to file an appeal in the district court challenging the dismissals. See Fed. R.Bankr.P. 8002(a). However, Brandt did not follow this course, nor did he alert the district court, as the court proceeded to try the remaining claims, that the bankruptcy court’s dismissal of the fraudulent transfer claims remained open to challenge in the district court. If the district court had been so alerted, it is unlikely that it would have ignored the matter; and regardless of whether the district court upheld the bankruptcy judge or reversed him and tried these claims along with the others, there would have been a resolution of the fraudulent transfer claims by the district court that could now be brought before us. Following the trial, Brandt filed new trial motions directed to the claims that had been resolved by the district court but again made no mention of the fraudulent transfer claims. Instead, after the motions were denied, Brandt filed his appeal from the district court’s judgment and then proceeded in this court to brief the dismissal of the fraudulent transfer claims as if they were encompassed by the district court’s judgment. But, of course, the district court’s judgment only resolved the claims that had been presented to the district court and decided by the judge or the jury. After various defendants objected to consideration of the fraudulent transfer claims on appeal, Brandt filed a reply brief urging that “the entry of final judgment in the District Court calls up for appellate review by this Court all interlocutory orders of which the Trustee is aggrieved, whether entered by the District Court or by the Bankruptcy Court, and this Court therefore has jurisdiction over all aspects of this appeal.” Brandt also points to his earlier efforts to appeal the dismissals as interlocutory orders and pokes fun at the notion that there should now be an appeal of those dismissals to the district court with appeals proceeding simultaneously before the district court (on the fraudulent transfer claims) and before this court (on the claims already resolved in the district court). None^of these arguments works. True, where the district court has made interlocutory decisions before entering a final judgment, an appeal from the final judgment brings up the interlocutory decisions for review by this court. John’s Insulation, Inc. v. L. Addison & *15 Assocs., Inc., 156 F.3d 101, 105 (1st Cir.1998). The difficulty is that this logic works only with respect to the interlocutory orders of the district court; the bankruptcy court, although a unit of the district court, is a distinct entity whose orders are appealable to the district court under a set of detailed restrictions and time limits. If proper and timely review is not sought in the district court, the matter never reaches that court and a fortiori does not reach this court. One could argue that the dismissal orders in question became final only when the district court dismissed the remaining claims against the same defendants following trial. If so, Brandt might then have appealed the dismissal orders to the district court, obtained a ruling and (assuming affirmance) sought to consolidate an appeal from this judgment with his previous appeal from the judgment on issues actually tried to the district court. However, Brandt did not follow this course either and cannot do so now because the time limit on an appeal to the district court expired before Brandt filed his appeal in this court. 6 The failure of Brandt’s case on this ground also spares us from considering various so-called “waiver” arguments that some of the defendants pressed based on Brandt’s failure to act earlier to raise the dismissed claims in the district court. From an equitable standpoint, one may feel some sympathy for Brandt, who was faced with poorly developed rules on finality and who made early efforts to seek district court review of the dismissals; it is much less easy to excuse the apparent failure of Brandt to call vividly to the district court’s attention the fact that, while that court was proceeding to try a set of claims properly before it, Brandt still desired to press other claims that the bankruptcy court had dismissed and which would almost certainly have been tried at the same time if the district court had overturned the bankruptcy court’s dismissals. However, our inability to address the merits does not rest on an equitable objection. Rather, it rests on the simple fact that our authority is to review judgments of the district court, and Brandt never secured a district court judgment on the fraudulent transfer claims nor is it apparent how he could do so now. Counsel for the trustee in a complicated bankruptcy case has to make its own decisions even where the law is unclear, and the course here followed did not preserve Brandt’s claims. III. The Unjust Enrichment Claims The bankruptcy court granted summary judgment rejecting claims of unjust enrichment leveled against a number of defendants. Most of the grants were never reviewed by the district court and are thus not before us, but Brandt did seek review by the district court of the summary judgments on these counts in favor of J.P. Morgan and Marvin Cyker. The district court granted Brandt leave to appeal these two summary judgments as interlocutory orders but nevertheless affirmed the bankruptcy court’s judgments on the ground that neither J.P. Morgan nor Cyker had been shown to have committed the “minimal wrongdoing” that the district court deemed required for an unjust enrichment claim under Massachusetts law. The district court’s rationale differed from those of the bankruptcy court: the bankruptcy court had ruled in favor of Cyker because he opposed the transaction, and in favor of J.P. Morgan because, as a mere recordholder, it received no direct benefit from the transaction. *16 J.P. Morgan argues that Brandt is seeking to appeal directly to this court from the bankruptcy court rulings, which Brandt may not do. Brandt’s arguments in his opening brief suggest that he has the same view. But given the district court’s affirmance and the lack of any limiting language in the notice of appeal to this court, we are free to treat Brandt as appealing from the district judge’s affir-mance of these orders. So viewed, these district court orders merged in the final judgment entered by the district court and are properly before us now. Cf. In re Parque Forestal, Inc., 949 F.2d 504, 508 (1st Cir.1991). Brandt appears to be right that under Massachusetts law unjust enrichment does not always require a finding of wrongdoing by the defendant. There are cases, albeit addressed to a somewhat different problem (mutual mistake), that hold that wrongdoing is not required so long as retention of the benefit would be unjust. E.g., White v. White, 346 Mass. 76, 190 N.E.2d 102, 104 (1963); National Shawmut Bank of Boston v. Fidelity Mut. Life Ins. Co., 318 Mass. 142, 61 N.E.2d 18, 22 (1945); see Keller v. O’Brien, 425 Mass. 774, 683 N.E.2d 1026, 1029-33 (1997). See generally Restatement of Restitution ch. 2, intro, note (1936). Indeed, the district court so instructed the jury on the unjust enrichment claim against Gemini, listing three elements of unjust enrichment that the plaintiff “must show”: First, a benefit or enrichment was conferred upon the defendant ...; second, the retention of that benefit or enrichment resulted in a detriment to [the plaintiff]; and, third, there are circumstances which make the retention of that benefit ... unjust. However, if, as the above suggests, the district court erred in its reason for affirming the dismissal of the claims against J.P. Morgan and Cyker, the error was harmless — and this is so even without reliance on the different reasons for those dismissals given by the bankruptcy judge. After receiving the above instruction on unjust enrichment, which did not require a showing of wrongdoing, the jury proceeded to reject on the merits the claim that Gemini was unjustly enriched by the payment made to Gemini in exchange for its Health-co shares. The counterpart claims against J.P. Morgan and Cyker were' of the same order but weaker. Gemini was the partnership that precipitated the original abortive LBO and then actively cooperated in achieving the second and successful one. As noted above, after its failed attempt to take over Healthco, Gemini entered an agreement that effectively gave it power to control the nominations of three of the seven members of Healthco’s board, and the three resulting nominees were on Healthco’s board, and voted for the buyout and merger, when it approved the merger plan by a 5-2 vote. By contrast, J.P. Morgan held its shares as recordholder for others and played no active role in the buyout, merely tendering shares in response to a public offer. And Cyker, who later sold stock options in Healthco, opposed and voted against the buyout. It is hard to see how a jury that found in Gemini’s favor could possibly have resolved in Brandt’s favor the decidedly weaker claims against the other two defendants. The jury verdict against Gemini thus entitles us to treat any error in the rationale for dismissing the claims against the other two defendants as harmless. See Fite v. Digital Equip. Corp., 232 F.3d 3, 6 (1st Cir.2000). As in Wills v. Brown University, 184 F.3d 20, 30 (1st Cir.1999), there is no practical likelihood that the dismissed claim could have succeeded where the tried claim failed. Other circuits have similarly found summary judgment orders harmless based on the implications of subsequent jury verdicts. See, e.g., Gross v. Weingarten, 217 F.3d 208, 219-20 (4th Cir.2000); Thompson v. Boggs, 33 F.3d 847, 859 (7th Cir.1994), cert. denied, 514 U.S. 1063, 115 S.Ct. 1692, 131 *17 L.Ed.2d 556 (1995); Wing v. Britton, 748 F.2d 494, 498 (8th Cir.1984). IV. The Fiduciary Duty Claims One of the claims made by Brandt charged the directors of HMD Acquisition Corp. with breaching their “fiduciary duties to Healthco and HMD Acquisition and their successors, shareholders, and creditors.” The bankruptcy judge dismissed this claim on the ground that these directors owed their duties to HMD Acquisition and not to Healthco. Even though the defendants also began to serve as directors of Healthco beginning on April 30, 1991, when the tender offer closed, the bankruptcy judge said that Healthco had by then “already committed itself to the transaction through its prior board.” The bankruptcy judge also said that although these were non-core claims, he was entitled to determine them on the merits because the parties had in effect consented to their disposition. On April 23, 1997, the first day of the jury trial, the district court announced that it was treating the bankruptcy court judgment dismissing the fiduciary duty claims as a proposed conclusion of law on a non-core matter, 28 U.S.C. § 157(c)(1), and then said that it was accepting and adopting the bankruptcy court’s recommendation. In this court, the defendants argue that Brandt forfeited any appeal when he failed to object to the bankruptcy court’s recommendation within ten days of the district court’s recharacterization. See Fed.R.Bankr.P. 9033(b). But if the bankruptcy court ruling was converted at that time to a recommendation, there was no reason for a further objection since the ruling was simultaneously resolved on the merits by the district court. The district court’s merits resolution is merged in its final judgment and properly before us. Nonetheless, all this is for naught. In his opening brief Brandt devotes only a single paragraph to the ruling on HMD Acquisition’s directors that he now seeks to reverse, saying that any claim for duty breached by the directors of HMD Acquisition “survived the merger.” Brandt’s terse argument does not attempt to address the lower courts’ ruling that there was no breach of any fiduciary duty to Healthco when the critical decision was taken. Brandt’s effort to offer new arguments in his reply brief, after the defendants filed their answering briefs, comes too late. Rivera-Muriente v. Agosto-Alicea, 959 F.2d 349, 354 (1st Cir.1992). V. Discovery Matters Brandt argues that the bankruptcy court committed reversible error in various discovery rulings. None of these rulings was formally appealed to the district court. However, during a pre-trial telephone conference on February 14, 1997, the district court judge indicated that he was aware of the bankruptcy judge’s discovery orders and was reluctant to disturb them, but would nonetheless “allow some minimum amount of further pretrial discovery.” To the extent that the district court did modify the bankruptcy court’s discovery orders, the modified orders are obviously before us for review. The jurisdictional issue is more debatable as to the discovery orders of the bankruptcy court that were not disturbed. Perhaps the district court’s statements could be regarded as an implicit affir-mance of those orders (or at least some of them) on interlocutory appeal; if so, the affirmance would be merged into the final judgment and properly before us. We will assume this is so arguendo since it does not alter the result. The most controversial of the bankruptcy judge’s orders is that of June 20, 1996, which limited each side to ten depositions as of right, in accordance with Federal Rule of Civil Procedure 30(a)(2), with the remaining depositions to be conducted from September through December 1996. Brandt had already taken four depositions and was therefore allowed only six more under the order. But the order also pro *18 vided that further depositions could be taken with leave of the court and in accordance with the general principles set forth in Rule 26(b)(2). Brandt, who had planned to take sixty or so additional depositions, immediately asked the bankruptcy judge to remove any limit or at least to allow dozens of depositions, and the bankruptcy judge refused. On December 17, 1996, Brandt asked permission to take additional depositions and for a one-month extension of the deposition deadline. Although Brandt identified 19 additional individuals and the subjects in question, the bankruptcy judge denied the motion, saying that it came only two weeks before the long-established deadline, a year-and-a-half after the complaint was filed, and two-and-a-half years after the trustee began investigation. The court said it was not impressed with the need for the depositions and that “[p]er-mitting the requested depositions [would] unnecessarily increase counsel’s fees and [would] more likely delay the trial scheduled to begin April 7,1997.” Brandt then sought but was denied leave by the district court for an immediate appeal. But, thereafter, in a pre-trial conference on February 14, 1997, the district court allowed each side to take an additional 20 hours of depositions before trial. At a further pre-trial hearing on March 17, 1997, Brandt asked for an adjournment of the April trial to allow for more depositions; but after learning that the 20 additional deposition hours had not yet been exhausted, the district court rejected the adjournment motion. Later, the court granted Brandt two additional depositions during the trial. Discovery decisions by the bankruptcy judge or district court are reviewed for abuse of discretion, and the discretion in this area is very broad, recognizing that an appeals court simply cannot manage the intricate process of discovery from a distance. In Modern Continental/Obayashi v. Occupational Safety & Health Review Commission, 196 F.3d 274, 281 (1st Cir.1999), this court spoke of the need for “a clear showing of manifest injustice,” saying that, to warrant reversal, the lower court’s discovery order must be “plainly wrong” and must be shown to have resulted in “substantial prejudice” to the complaining party. Although at first blush the limitations imposed by the bankruptcy judge seem severe, even when somewhat modified by the district court, there is more to the story. Brandt devotes almost ten pages of his brief to explaining that the case involves a large amount of money and many parties and that none of the defendants registered any objection to his original proposal to take sixty or more depositions. Of course, the lack of objection from other parties is not dispositive; the bankruptcy judge had an independent responsibility to manage the litigation and conserve the resources of the estate. But the size and scope of the litigation might well have provided a basis for justifying a greater number of depositions than was allowed. However, the bankruptcy judge did not say that only ten depositions were permitted. Obviously concerned with the slow pace and mounting expense of discovery, he held the plaintiffs feet to the fire to move quickly and then justify any additional requests for depositions on a case-specific basis. In fact, the order fixing the ten-deposition limit referred to Federal Rules’ provisions setting the criteria for justifying additional discovery. Thus, the bankruptcy judge’s order is not quite the arbitrary limit that Brandt suggests. The more troubling aspect is the bankruptcy court’s refusal in December 1996 to extend the deadline and allow further depositions. Brandt’s request at that time was reasonably detailed as to proposed deponents and the subject matter for questioning. It is hard to lean too heavily on the bankruptcy judge’s brief statement that he was “not impressed with the critical nature of the dispositions.” And while *19 the district judge effectively allowed another four or five depositions, this was far short of what Brandt had sought even in December. However we might otherwise feel about the severe limit on depositions — and it would take a more detailed examination of the record for us to make a final judgment — Brandt’s opening brief is virtually devoid of any showing that Brandt was prejudiced. In the entire ten-page discussion there is only a single elliptical sentence describing a specific witness. Even this discussion does not make clear why Brandt thinks the witness was so vital. Thus there is no reason to think that the outcome of the case was affected by the limit on depositions. Brandt says that this is a catch 22, since one can never be sure what further discovery might have adduced. While this (standard) argument has some force, it is not conclusive: both in justifying discovery and in explaining later why a refusal to allow it caused harm, lawyers are accustomed to showing specifically just what gaps in their claim and defense might be filled by evidence within the likely knowledge of the witness. And it is just such specifics that are absent from Brandt’s opening brief. Indeed, the trial being over, it should have been even easier than before or at trial for Brandt to explain just where he thinks that additional depositions could have filled any apparent gaps in the case presented. In his reply brief, Brandt does make a broader, but at the same time better supported, showing that he was expected to conduct too much discovery within too brief a time frame when one takes into account both depositions and the huge number of documents to be sorted and analyzed. But Brandt’s time frame may be an artificial one; there is some reason to think that he could have made more progress at an earlier stage and that he moved too slowly even after the initial discovery deadline was set in June 1996. 7 But we need not resolve this point, since, as we have already noted, arguments first developed in a reply brief come too late. Rivera-Muriente, 959 F.2d at 354. Brandt’s second claim of discovery error concerns the failure of Coopers & Lybrand to produce documents from Coopers’ foreign offices relating to its review of Healthco’s year-end financial statements for 1990. The unsecured creditors earlier sought to obtain these and other papers from Coopers, see Fed.R.Bankr.P.2004, and Brandt and Coopers agreed on the production of certain of these documents, but apparently Coopers failed to produce documents from its foreign offices. Yet it was not until February 20, 1997, less than two months before the scheduled trial date and six months after the bankruptcy court’s deadline for document discovery, that Brandt filed an expedited motion with the bankruptcy court to obtain the audit-related papers from Coopers. Although Brandt now offers an explanation as to why these papers were necessary, the request originally filed in the bankruptcy court merely asserted that Brandt “need[ed] to review all C & L memoranda and audit workpapers regarding Healthco’s foreign subsidiaries in order to prepare properly his case for trial.” And, not surprisingly, the bankruptcy court denied the motion without explanation about a week after it was filed. There is no indication that Brandt then brought the matter to the attention of the district court by an interlocutory appeal; nor does it appear that, as trial ap- *20 proaehed, he ever asked the district court for belated document discovery against Coopers, which might conceivably have been justified if a new need arose at the last minute. In any event, Brandt apparently never gave the bankruptcy judge the explanation he now gives us as to why the papers from Coopers’ foreign subsidiaries were necessary. Faced only with a bland and belated statement that the papers were needed, the bankruptcy court acted within its discretion in denying the requested discovery. Finally, Brandt says that the bankruptcy court erred in refusing to permit him to discover the identity of the beneficial owners of the Healthco shares that were tendered by J.P. Morgan and Chancellor. Brandt argues that this information was necessary so that Brandt could direct its unjust enrichment claims against those who actually benefitted from the $15 per share buyout of Healthco stock. This point takes on added significance because the bankruptcy court relied on the fact that J.P. Morgan and Chancellor were merely recordholders in dismissing the unjust enrichment claims against them. Brandt attempts to show that the denial of an opportunity to discover beneficial ownership was based on the bankruptcy court’s misconstrual of its own orders. However, there is no indication that a ruling on this discovery issue was ever sought from the district court. In any case, the jury rejected the unjust enrichment claims directed at defendants who were both stockholders and active in promoting the LBO; it is very hard to see how Brandt could have expected a more favorable result if he had unearthed the names of passive beneficial stockholders for whom record ownership was held in the name of J.P. Morgan or Chancellor. VI. Conduct of the Trial Brandt objects to a set of alleged errors occurring during the course of the trial and says that the errors and misconduct of defense counsel fatally tainted the verdict. Specifically, Brandt objects to references to settlements with other defendants, admission of an expert’s testimony and report, time limits imposed by the trial judge, restrictions on the “publication” of documents to the jury, and comments or evidence designed to paint the trustee or the trustee’s counsel in a bad light. We consider the claims of error in the order in which Brandt has briefed them. First, citing McInnis v. A.M.F., Inc., 765 F.2d 240 (1st Cir.1985), Brandt complains of references to settlements Brandt reached with other defendants. In Mclnnis, this court construed broadly Federal Rule of Evidence 408, which excludes settlements when offered to prove the validity or invalidity of a claim. Id. at 246-48. There, the plaintiff, the victim in a motorcycle accident, had sued the manufacturers (for making a defective product); the plaintiff had also previously obtained a settlement paid on behalf of the driver of a car that had hit the motorcycle, and the trial court admitted evidence of the settlement to show that the accident had been caused by the driver of the car rather than the faulty manufacture of the motorcycle. Id. at 241-42. Mclnnis held that using the settlement agreement to show causation amounted to using it to show the invalidity of a claim, and found that the error in admitting evidence of the settlement required a new trial. Id. at 246-rit8. Here, Brandt says that one of the defendant’s opening statements at trial mentioned Brandt’s settlements with other parties. However, the passages that Brandt identifies refer not to settlements but to the fact that Brandt had initially sued 69 people or businesses. The thrust was not that other defendants had settled (and were therefore the real perpetrators) but rather that Brandt was a plaintiff who sued everyone in sight regardless of whether the individual defendant was responsible. This was not an offer of proof of, or a reference to, a settlement, which is *21 what Rule 408 and Melnnis are concerned with. Some of Brandt’s discussion of this issue suggests that he is concerned not so much with the inference of settlement, but with the inference that a large number of parties were responsible for the transaction but the blame has been unfairly focused on the few remaining at trial. While this was a possible inference, it is not clear that, in this respect, the comments complained of were very helpful to the defendants; indeed, they might rather have suggested that the parties remaining at trial were those most responsible. In any event, the trustee makes no substantial effort to make a real showing of prejudice. Brandt also refers in his brief to a closing argument by defense counsel insinuating that the proof offered in the trial of negligence by Coopers & Lybrand “undermines the integrity of the case against the defendants in this courtroom.” Whether or not the inference is a fair one, once again it has nothing to do with settlement, there having been affirmative evidence against Coopers & Lybrand offered during the trial itself. It is worth adding that the first references to settlement were made not by defendants but by Brandt’s counsel. Cf. Willco Kuwait (Trading) S.A.K. v. de-Savary, 843 F.2d 618, 625 (1st Cir.1988). Second, Brandt says that the district court erred in permitting the defendants to call Robert W. Berliner — Brandt’s accounting expert — to examine him about portions - of a report he had prepared for Brandt. The disputed portion of the report concerns Berliner’s conclusion that Coopers had negligently performed the Healthco audit for 1990; the implication, which defendants hoped would be drawn, was that Coopers and not the defendants at trial bore responsibility for the unhappy outcome of the LBO. Brandt made a timely objection that the report was hearsay and now says that evidence regarding it was highly prejudicial and should have been excluded under Federal Rule of Evidence 403. Brandt expressly admits in his opening brief that the defendants had “a right to argue that Coopers was the cause of the failure of Healthco,” but objects that the defendants were obliged to prove this through their own evidence and expert witnesses. The latter is an overstatement: at Brandt’s behest, Berliner gave testimony arguably implying that Coopers did not bear responsibility for the failure of Healthco, so he certainly could be cross-examined and impeached on this issue. Whether the Berliner report was admissible as the admission of an opposing party, and therefore admissible not just to impeach but as proof of the facts asserted in it, is a different question which the district court resolved in favor of the defendants. The district court, supported on appeal by the defendants, viewed the report as an admission of Brandt through an agent (the expert) acting within the scope of his agency, and therefore found it admissible under Federal Rule of Evidence 801(d)(2)(D). Since the report was prepared by Berliner during his work for Brandt, it might at first blush seem to fit comfortably within this rule, assuming always that Berliner could be regarded as an agent for this purpose. Some authority points in this direction but the Third Circuit emphatically disagrees, saying that an expert is more like an independent contractor offering his own opinion and is not “controlled” by the party who employs him. Kirk v. Raymark Indus., Inc., 61 F.3d 147, 163-64 (3d Cir.1995), ce rt. denied, 516 U.S. 1145, 116 S.Ct. 1015, 134 L.Ed.2d 95 (1996) (discussed in 30B Charles Alan Wright & Michael H. Graham, Federal Practice and Procedure § 7022, at 202 n. 1 (2000)). The authorities are fairly sparse, but we need not decide the Rule 801(d) issue. Prior to introducing in evidence the pertinent portion of the report, the defendants asked Berliner questions and elicited statements from him as to Cooper’s actions that covered more or less the same ground as the report. As noted above, the defen *22 dants’ questions were permissible cross-examination. The resulting statements— not unexpected unless Berliner was prepared to contradict his report — were in-court statements not subject to a hearsay objection. Accordingly, even if the report itself were objectionable, any error in its admission is rendered harmless by the questioning of Berliner. See Texaco P.R., Inc. v. Department of Consumer Affairs, 60 F.3d 867, 886 (1st Cir.1995). As for the objection under Rule 403, it is hard to understand Brandt’s argument. Brandt agrees that whether Coopers was careless was a pertinent issue and Berliner’s testimony and report were directed to that question. No doubt the testimony had more impact because it came from Brandt’s own expert, but the expert was one whom Brandt himself had called to testify at trial and who had given testimony that might otherwise have led the jury to believe that Coopers was not at fault. Assuming a Rule 403 objection to the Berliner evidence was preserved, it was not error under Rule 403 to allow the evidence. Third, Brandt objects, in the caption of one section of his opening brief, to the district court’s placing “unreasonable pre-determined time limitations upon the trial,” i.e., 60 hours. Then — in the body of the discussion — he develops two arguments: that defense counsel manipulated the time limits to Brandt’s disadvantage (naming witnesses, forcing Brandt to reserve some of his time to cross-examine them, and then not calling those witnesses); and that the district court promised Brandt that he could use all of his otherwise unused time for his closing argument but then limited him to four and a half hours when he still had ten hours remaining. How trial time should be limited — obviously some limitations are appropriate— raises interesting problems, see Borges v. Our Lady of the Sea Corp., 935 F.2d 436, 442-43 (1st Cir.1991), but they need not be addressed here because (despite the caption in the opening brief) Brandt’s argument makes no effort to show that the 60 hours of trial time allotted to each side was unreasonable. The related suggestion that defense counsel manipulated the time limits by listing witnesses who were not called is mentioned in a single sentence, is not seriously supported, and is therefore waived. Massachusetts Sch. of Law v. American Bar Ass'n, 142 F.3d 26, 43 (1st Cir.1998). 8 We add that we have found no additional serious support for this claim in Brandt’s earlier arguments to the district court on this same issue. The bulk of Brandt’s argument is directed to a different and, as presented, more striking claim that the district court promised unlimited time for closing argument (so long as the 60-hour limit was not exceeded) and then broke this promise. Brandt describes a colloquy during the trial where the district court allegedly “prohibited the Trustee’s counsel from publishing to the jury relevant portions of voluminous documents that had been received in evidence”; and Brandt then quotes his counsel as asking the court whether it was “going to impose any limitation on the time of closing assuming I still have it available in my allotted hours.” Brandt’s brief then quotes the court as saying: “You can have any length of closing.” The trial transcript shows that the district court never made an unconditional promise to allow Brandt to use any unused time in closing argument. The district court said, “You can have any length of closing as long as” — and was then interrupted by Brandt’s counsel who said, “Then that solves a lot of my problem.” Shortly before the close of evidence, the court made clear that it did not intend to allow Brandt to make a ten-hour closing argument even though he still had ten *23 hours left on his clock and, despite a pro forma protest, Brandt then suggested four and a half hours and made no effort to show that this would prejudice him or that he could not present the substance of his case in this time frame. In the end he elected to use less in order to complete his argument within one day. Providing no specifics, Brandt intimates that he was somehow limited in his ability to publish documents or deposition transcript evidence to the jury during trial and that he hoped to use the closing argument to read portions of this evidence to the jury. In fact, the trial transcript shows that Brandt published a great deal of such evidence during the trial, and the colloquy to which he refers to show that he was limited actually appears to have been concerned with how the materials were presented, the district court having objected to Brandt’s counsel reading deposition pages at length to the jury while purporting to question the witness. Once again, Brandt’s brief points to no specific material, let alone material of vital importance, that he was effectively prevented from publishing to the jury. Fourth, Brandt argues that during the course of the trial, some defense counsel made arguments or introduced evidence besmirching the character of the trustee by suggesting that he was in the business of acting as a trustee for many bankrupt companies, that he received fees based on the amount of money he collected, that in the past he had sued many defendants on claims like the ones pressed here, that he hired his own company to provide administrative services to the estate, and so on. These charges, says Brandt, were irrelevant and (if marginally relevant in some respects) far more prejudicial than is proper under Rule 403. Brandt’s complaints would have more force if he had not invited many of these “charges” by the claims that his counsel made during opening argument, claims later echoed by Brandt himself when he briefly appeared as a witness. In opening to the jury, Brandt’s counsel sought to paint a picture of the trustee as essentially a neutral party engaged in a quasi-official function: counsel said that the trustee was “supervised” by the bankruptcy judge, was “a disinterested party” and would not “get to keep any of the money [from a verdict] himself.” Later, Brandt himself told the jury that the money recovered from defendants would be “disseminated” to Health-co’s creditors. During cross-examination, Brandt conceded that the trustee would receive “a commission” from litigation proceeds based on a percentage formula. See 11 U.S.C. § 326(a). But even assuming that counsel for the trustee did not provoke defense counsel’s responses, and that one or more of the defense counsel went too far in some of their remarks, the trial judge addressed the issue appropriately. After concluding that the comments were improper, the judge rebuked counsel and directed the jury to disregard the comments. It is our practice to presume that such instructions are followed, unless the evidence is hopelessly sure to warp the jury’s judgment. Conde v. Starlight I, Inc., 103 F.3d 210, 213 (1st Cir.1997). The same conclusion, and much of the same analysis, applies to Brandt’s complaints about remarks in some defense counsel’s closing arguments that Brandt believes unfairly portrayed his attorneys in an ill light. As with the comments regarding Brandt himself, the trial judge responded to the remarks about Brandt’s attorneys by instructing the jury to disregard negative comments about their integrity. Without approving every remark or question posed by defense counsel, we find that this is not a case in which the verdict should be overturned or a new trial required. Cf. Fernandez v. Corporacion Insular De Seguros, 79 F.3d 207, 210 (1st Cir.1996); United States v. Maccini 721 F.2d 840, 846-47 (1st Cir.1983). Affirmed. Footnotes [1] Aspects of Healthco's bankruptcy are addressed in Hides, Muse & Co. v. Brandt (In re Healthco Int’l, Inc.), 136 F.3d 45 (1st Cir.1998); Brandt v. Repco Printers & Lithographies, Inc. (In re Healthco Int’l, Inc.), 132 F.3d 104 (1st Cir.1997); Brandt v. Hicks, Muse & Co., 213 B.R. 784 (D.Mass.1997); Brandt v. Hides, Muse & Co. (In re Healthco Int’l, Inc.), 208 B.R. 288 (Bankr.D.Mass.1997); Brandt v. Hickes, Muse & Co. (In re Healthco Int’l, Inc.), 203 B.R. 515 (Bankr.D.Mass.1996); Brandt v. Hickes, Muse & Co. (In re Healthco Int’l, Inc.), 201 B.R. 19 (Bankr.D.Mass.1996); Brandt v. *10 Hicks, Muse & Co. (In re Healthco Int’l, Inc.), 195 B.R. 971 (Bankr.D.Mass.1996); In re Healthco Int’l, Inc., 174 B.R. 174 (Bankr.D.Mass.1994). Two opinions in this group, Brandt, 213 B.R. 784, and Brandt, 208 B.R. 288, provide more detailed accounts of the leveraged buyout than our own summary. [2] A leveraged buyout is a transaction to acquire a corporation "in which a substantial portion of the purchase price paid for the stock of a target corporation is borrowed and where the loan is secured by the target corporation’s assets." Mellon Bank, N.A. v. Metro Communications, Inc., 945 F.2d 635, 645 (3d Cir.1991), cert. denied, 503 U.S. 937, 112 S.Ct. 1476, 117 L.Ed.2d 620 (1992); 3 Norton Bankruptcy Law and Practice 2d § 58A:1, at 58A-2 to 58A-3 (William L. Norton, Jr., ed. 1997). See generally Day, Walls & Dolak, Riding the Rapids: Financing the Leveraged Transaction Without Getting Wet, 41 Syracuse L.Rev. 661 (1990). [3] Those accused of benefitting from fraudulent transfers included (1) Hicks, Muse, the primary orchestrator of the leveraged buyout; (2) the bank group and subordinated deb-tholders who helped finance the buyout; (3) various other Healthco stockholders, whose tendering of shares allowed the buyout to proceed; and (4) various professionals who were paid for their roles in bringing about the buyout. [4] The primary targets of these counts were the directors of Healthco and HMD Acquisition, as well as Healthco’s controlling shareholders. Their alleged aiders and abettors included Hicks, Muse, Healthco Holding Co., the bank group, Healthco directors who voted for the buyout, Valuation Research which had endorsed the feasibility of the original $19 buyout plan, and Lazard which had endorsed the fairness of the $15 plan. [5] See, e.g., Kupetz v. Wolf, 845 F.2d 842, 847-48, 850 (9th Cir.1988) (respecting "the formal structure of [the] LBO,” and declining to apply a theory of constructive fraud); United States v. Tabor Court Realty Corp., 803 F.2d 1288, 1297 (3d Cir.1986) (applying Pennsylvania’s fraudulent conveyance statute to leveraged buyouts), cert. denied, 483 U.S. 1005, 107 S.Ct. 3229, 97 L.Ed.2d 735 (1987); MFS/Sun Life Trust-High Yield Series v. Van Dusen Airport Servs. Co., 910 F.Supp. 913, 933-34 (S.D.N.Y.1995) (finding collapsing an LBO appropriate where "all parties to each subsidiary transfer were aware of the overall leveraged buyout”); Wieboldt Stores, Inc. v. Schottenstein, 94 B.R. 488, 501-03 (N.D.Ill.1988) (collapsing an LBO with respect to "the controlling shareholders, the LBO lenders, and the insider shareholders,” but not with respect to shareholders who were only aware of the tender offer made to them); Murphy v. Meritor Sav. Bank (In re O’Day Corp.), 126 B.R. 370, 394 (Bankr.D.Mass.1991) (collapsing an LBO where "all parties ... were aware of the structure of the transaction and *13 participated in implementing it”); In re Reveo D.S., Inc., 118 B.R. 468, 517-18 (Bankr.N.D.Ohio 1990) (noting the competition between more traditional “anti-collapse” and more modern "pro-collapse” perspectives). [6] Fed.R.Bankr.8002(a). This discrepancy in timing also forecloses any option we otherwise might have had to treat the appeal to us as an appeal of the bankruptcy court orders filed in the wrong court and to transfer the appeal to the district court based on the transfer statute, 28 U.S.C. § 1631. Notably, Brandt has neither cited the transfer statute nor made any such transfer request. [7] Brandt says that he had insufficient time to conduct discovery between June 1996 (when the case management order was adopted) and December 1996 (the scheduled end of discovery), as well as insufficient time for additional discovery before trial in April 1997. However, Brandt became Healthco’s trustee in October 1993, filed his complaint in June 1995, and had possession of many of the documents that he complains he had insufficient time to review long before June 1996. Further, in the six months between the case management order and his motion for additional depositions, Brandt conducted only six depositions. [8] A similar lack of development marks Brandt’s suggestion, in the "Issues Presented” portion of his opening brief, that the district court erred in imposing a time penalty after Brandt made an unsuccessful motion.