Distorted Choice in Corporate Bankruptcy
abstract. We ordinarily assume that a central objective of every voting process is ensuring an undistorted vote. Recent developments in corporate bankruptcy, which culminates with an elaborate vote, are quite puzzling from this perspective. Two strategies now routinely used in big cases are intended to distort, and clearly do distort, the voting process. Restructuring support agreements (RSAs) and “deathtrap” provisions remove creditors’ ability to vote for or against a proposed reorganization simply on the merits.
This Article offers the first comprehensive analysis of these new distortive techniques. One possible solution is simply to ban distortive techniques, as several scholars advocate with RSAs that offer joinder bonuses. Although an antidistortion rule would be straightforward to implement, I argue this would be a mistake. The distortive techniques respond to developments that have made reorganization difficult, such as claims trading and a greater need for speed. Further, Chapter 11’s baseline was never intended to be neutral: it nudges the parties toward confirming a reorganization plan. There also are independent justifications for some distortive techniques, and the alternative to using them might be even worse—possibly leading to more fire sales of debtors’ assets.
How can legitimate use of the new distortive techniques be distinguished from more pernicious practices? To answer this question, I outline four rules of thumb to assist the scrutiny. Courts should consider whether holdouts are a serious threat, the magnitude of the coercion, the significance of any independent justifications, and whether the holdout threat is an intentional feature of the parties’ contracts. I then apply the rules of thumb to a few prominent recent cases. I conclude by considering two obvious extensions of the analysis, so-called “gifting” transactions in Chapter 11 and bond-exchange offers outside of bankruptcy.
author. S. Samuel Arsht Professor of Corporate Law, University of Pennsylvania Carey Law School. I am grateful to Judge Thomas Ambro, Martin Bienenstock, Bill Bratton, Vince Buccola, Michael Carrier, Jill Fisch, Christine Jolls, Judge Kent Jordan, Michael Knoll, Seth Kreimer, Yair Listokin, Minor Myers, Samir Parikh, Elizabeth Pollman, Mark Roe, Roberta Romano, Robert Rasmussen, Natasha Sarin, David Schleicher, Alan Schwartz, Judge Brendan Shannon, James Sprayregen, Chief Justice Leo Strine, Jr. (retired), Kate Waldock, and David Zaring; to Melody Wang and the editors of the Yale Law Journal; to participants in faculty workshops at the University of Pennsylvania Carey Law School, Yale Law School, and Vanderbilt Law School for helpful comments; to Chris Neumann, Annie Wheat, and Jinlin Ye for superb research assistance; and to the University of Pennsylvania Carey Law School for generous summer funding. I am a member of the Financial Oversight and Management Board for Puerto Rico, and some of the issues discussed in this Article relate to that involvement. The Article reflects my personal views, not the views of the Board, its other members, or any of the people thanked in this acknowledgment.
We ordinarily assume—or at least pretend to assume—that a central objective of every voting process is ensuring an undistorted vote. Protecting the integrity of a vote is sometimes difficult, and the best way to achieve this may be contested, as reflected in the controversies over the Supreme Court’s voting-rights jurisprudence and in debates over possible foreign interference in the last presidential election.1 But nearly everyone agrees that the goal should be to remove distortions so that voters can resolve the question at hand on the merits.
Recent developments in corporate bankruptcy are quite puzzling from this perspective. Chapter 11 is organized around an elaborate vote. The debtor or other proponent of a reorganization plan divides the creditors and shareholders of the company into different classes,2 and the creditors or shareholders in each class vote to approve or reject the proposed plan.3 If every class of creditors and shareholders votes in favor of the proposed plan, and it satisfies a number of other requirements, the bankruptcy judge will confirm the plan.4 Bankruptcy law gives the bankruptcy judge ample tools to police any distortions. Voting cannot begin until the court approves a disclosure statement giving the creditors and shareholders “adequate information” about the proposed plan,5 for instance, and the judge can disqualify problematic votes.6 The law on the books is intended to produce a simple undistorted vote.
Yet the law as it plays out in practice looks radically different. Two of the most important developments in recent bankruptcy practice are intended to distort, and clearly do distort, the voting process. They remove creditors’ ability to vote simply on the merits—that is, to vote based on the plan’s proposed payout for their class.
The first is the emergence of restructuring support agreements (RSAs).7 In the simplest type of RSA, the debtor negotiates the terms of a potential reorganization plan with a subset of its creditors—often focusing on multiple classes of creditors but sometimes targeting a single class. The RSA commits its signatories to support a future reorganization plan that conforms to the terms of the RSA, including the proposed payout to each creditor class. A creditor that signs the RSA relinquishes its ability to decide independently whether to support a reorganization plan subsequently proposed by the debtor. It does this before—often long before—a disclosure statement is approved and the proposed reorganization is submitted to creditors for a vote.
Many recent RSAs further distort the decisionmaking process by offering to pay a “support fee” to creditors who sign the RSA. Such “signing-fee RSAs” offer compensation that may reimburse creditors for the professional fees they incurred while negotiating the RSA. A signing-fee RSA may also include a fee for supporting the reorganization plan when it is proposed and waiving the right to object, as in agreements involving Puerto Rico’s electricity company and Peabody Energy.8
Alternatively, the RSA may provide a benefit to signatories, such as the right to provide debtor-in-possession financing during the case or to participate in a rights offering after the debtor’s reorganization plan is confirmed.9 These inducements, which are available only to those who sign the RSA, look like a form of vote buying, since they compensate signatories who commit to supporting an upcoming plan.
The second recent development is the use of “deathtrap” provisions in proposed reorganization plans. In a traditional deathtrap provision, the debtor proposes to give a creditor class some form of compensation if it votes “yes,” but cuts it off altogether if it votes “no.” The reorganization plan in the Trident Holding Company bankruptcy said, for instance, that if the first lien classes and the second lien classes “are Accepting Classes, each Holder of an Allowed Second Lien Claim shall receive its Pro Rata share and interest in 1% of the Warrants,” but if the First Lien Classes or the Second Lien Classes “are not Accepting Classes, Holders of Allowed Second Lien Claims shall not receive any distributions on account of such Allowed Second Lien Claims.”10
A more elaborate version—the “individually targeted deathtrap”—may offer one form of compensation to individual creditors who vote “yes” and a different compensation to individual creditors who vote “no.”11 In each case, the point is to apply pressure, using both a carrot (the compensation for a “yes” vote) and a stick (worse treatment of “no” votes) to nudge the creditors or shareholders to vote in favor of the plan.12
RSAs and deathtraps can distort the voting process in at least three ways.13 First, if they include supplemental payments, the additional compensation pressures creditors to vote for the reorganization plan even if creditors believe the payout is too low. These payments require a creditor to forgo compensation if she or the class votes “no,” and thus make a “no” vote more expensive and less attractive for reasons unrelated to the creditor’s views on the underlying merits of the plan. Second, distortive techniques can procedurally warp the voting process by binding creditors before the plan has been formally proposed and using exploding offers to induce creditors to commit early on. Third, RSAs and deathtraps may even distort the voting process by decreasing creditors’ likelihood of success in challenging the plan in the event the class of creditors votes “no.”14 By making the alternative to voting “yes” less attractive, these techniques can coerce creditors to vote for the plan.15
Despite the ubiquity of the new distortive techniques, they are just beginning to attract attention in the scholarly literature. Several scholars have written about RSAs,16 but I am not aware of any articles that devote meaningful attention to the use of deathtrap provisions or to the increase in voting distortions more generally. This Article is the first to attempt a more comprehensive analysis of the new landscape of distorted voting.
One obvious solution to bankruptcy’s voting distortions might be to prohibit or sharply restrict their use. This is the usual strategy elsewhere, and in corporate law, Delaware courts have in fact banned a somewhat analogous distortion that featured in freeze-out mergers.17 In the past, a corporate parent could use subtle forms of coercion when freezing out the minority shares of a subsidiary. If the parent made a tender offer for the minority shares, for instance, it could hint that any untendered shares would be neglected after the tender offer,18 thus diminishing the value of untendered shares. Delaware courts have cracked down on these practices by imposing a stringent antidistortion rule.19
The same approach could easily be employed in Chapter 11. In its strong form, an antidistortion rule would ban the use of RSAs and deathtrap provisions altogether, since each distorts the voting process. Under a weaker antidistortion rule, courts might prohibit only the variations of RSAs and deathtraps that introduce the most significant distortions, such as signing-fee RSAs and individually targeted deathtraps.
The antidistortion approach has considerable appeal, especially for those of us who already are attracted to rule-of-law moralism. It would ensure a much more unbiased Chapter 11 vote than the vote in most current reorganization cases, while also reducing the risk of windfalls to favored creditors. It also would be relatively simple to implement. As already noted, existing bankruptcy law gives judges powerful tools to curb voting distortions. Before the debtor can solicit votes on a proposed reorganization plan during the case, for instance, the bankruptcy court must find that the disclosure statement provides “adequate information” to creditors.20 A court could easily hold that an RSA violates this provision—and on rare occasions, courts have done precisely this.21 RSAs and deathtrap provisions also could be struck down as inconsistent with the obligation that the plan be “proposed in good faith and not by any means forbidden by law.”22 If antidistortion is the best solution, it lies readily at hand.
The first clue that banning the new distortive techniques may not be the optimal solution comes from the response of the bankruptcy courts. With only a few exceptions, most of them in the early 2000s, bankruptcy judges have upheld both RSAs and deathtrap provisions. To be sure, bankruptcy judges’ endorsement does not necessarily mean there is no reason to worry about the distortive techniques. It is possible that judges have not yet fully recognized the distortive effects of the new techniques, or that bankruptcy judges are too quick to approve the use of provisions that make successful reorganization more likely. But courts’ acquiescence to the new distortive techniques suggests these strategies may be more justified than they initially appear.
It turns out they are. The justification for permitting at least some use of these distortive techniques begins to emerge if we take a closer look at the environment in which the new distortive techniques emerged. In the early years of Chapter 11, large debtors had the option of devising reorganization plans at a leisurely pace—they had a long “runway,” in current jargon.23 This is no longer the case. Financial distress must now be resolved much more quickly, both because the value of many troubled companies is evanescent and because lenders and other creditors use debtors’ need for liquidity as leverage to compress the timeline of the case.24 Whereas the typical Chapter 11 case lasted more than two years prior to 2000,25 the duration is now roughly one year.26
Achieving a speedy reorganization would be challenging even if the debtor were dealing with a stable group of creditors. But because claims trading is now ubiquitous, creditors’ interests are highly unstable.27 A potential deal hammered out today may fall apart tomorrow after some claims are sold to buyers who do not believe the proposed deal is a good one.
It also is easier than ever before for a distressed-debt investor to assemble a blocking position and thereby veto the debtor’s proposed reorganization plan.28 The explosion of claims trading appears to have begun roughly a decade after the current bankruptcy laws were adopted,29 and the market for distressed debt has grown exponentially since then.30
The ease with which distressed-debt traders can now acquire veto power can be both helpful and harmful. If the debtor proposes a problematic reorganization plan, a creditor’s efforts to block the plan may benefit other creditors as well. But the veto facilitated by claims trading may also enable a creditor who has a conflict of interest—such as a competitor of the debtor—or other perverse incentive to thwart confirmation of a reorganization plan that serves the interests of the debtor and other creditors. Moreover, even if a creditor does not have problematic incentives, it may seek to use its leverage to obtain a disproportionate recovery for itself.31 Given that distressed-debt traders often have a short-term focus and little reputational stake in the consequences of their intervention, the potential for problematic holdouts is significant.
Of course, the emergence of new obstacles to a successful reorganization does not justify voting distortions by itself. All else equal, the appropriate response might be something along the lines of “tough luck.” But all else is not equal in bankruptcy. Perhaps surprisingly for those who assume votes should be undistorted, the Chapter 11 vote is not intended to be neutral and uncoerced. The voting rules already include features that are designed to nudge the parties toward confirmation of a reorganization plan. If one or more classes vote against a proposed reorganization plan, for instance, the plan can nevertheless be “crammed down” if, among other things, the proposed plan “does not discriminate unfairly, and is fair and equitable.”32 There also is an implicit threat that, if the parties fail to devise a confirmable reorganization plan, the case will be converted to Chapter 7, and the debtor’s assets will be sold off in pieces by a court-appointed trustee.33
If the goal is to tilt the playing field slightly toward reorganization, and recent developments have made reorganization more difficult, distortive techniques that soften the effects of these developments might not be inherently bad. Distortive techniques may sometimes be appropriate to counteract destructive holdout activity.34
There also are independent justifications for some distortive techniques. The signing-fee RSA, for instance—which some commentators treat as per se disqualifying35—may compensate the signatories for the cost of negotiating a plan that benefits all creditors and for committing themselves and any successors to support the proposed reorganization plan even if a better alternative emerges. A deathtrap provision may resolve—or at least postpone until after confirmation—a high-stakes dispute that could otherwise derail the reorganization process by consuming the debtor in time-consuming litigation.36 To be sure, the purported benefits of a distortive technique may be exaggerated or outweighed by the potentially pernicious effects of the technique. But there often are legitimate justifications for using a technique even if it incidentally distorts the voting process. In this sense, the new distortive techniques are quite similar to lockups in corporate merger-and-acquisition transactions, where they also can be both problematic and beneficial.37
Finally, we need to consider how the parties might respond to a partial or complete ban of distortive techniques. One obvious possibility is that some distressed debtors that might otherwise reorganize under Chapter 11 would now be unable to do so and would be forced to resolve their distress through a sale of assets instead. To be sure, asset sales can be an effective solution to the debtor’s financial distress.38 But no one has a vote in an asset sale,39 and an antidistortion rule could misdirect debtors away from the traditional Chapter 11 process when Chapter 11 would be the best solution to the debtor’s financial distress.
These complicating factors suggest it would be a mistake to ban distortive techniques altogether. This doesn’t mean that distortive techniques should always be permitted, however. RSAs and deathtrap provisions sometimes do distort the voting process in indefensible ways. RSA fees sometimes appear to be little more than vote buying, for instance, and the structure of some deathtrap provisions is highly coercive.
Thus far, courts seem to have taken a “know it when they see it” approach to the new distortive techniques. When a debtor agreed to exchange 100% of the company’s post-reorganization stock for a secured creditor’s $238 million claim—without doing a market test of the transaction or negotiating with any other creditors (including nearly $1.2 billion of other secured claims)—the court balked, finding that the agreement “breeds contempt rather than fostering negotiations,” and refused to approve it.40 In most (though not all) other cases, courts have approved the distortive techniques.
My goal in this Article is to offer additional guidance for the determinations courts are now making. I start with the standard assumption that the general objective of the bankruptcy process is value maximization. Chapter 11 does not seek to achieve this objective directly, however. Instead, it provides a framework for renegotiating the parties’ entitlements that culminates with the Chapter 11 voting process. An assessment of the new distortive techniques therefore needs to consider carefully both the parties’ entitlements and the procedural integrity of the Chapter 11 process.
My analysis suggests that some distortive techniques should nearly always be permitted and others usually barred. Traditional deathtraps are an example of the former; exploding RSAs that give potential signatories only a brief period of time to decide are in the latter category. For distortive techniques that fall in the middle, I offer a handful of rules of thumb. I then apply the rules of thumb to four important recent cases. My analysis suggests that if bankruptcy judges clearly signal a willingness to strike down egregious uses of distortive techniques, the parties will adjust accordingly, eventually rendering court intervention unnecessary in most cases.41
Part I of the Article describes the new distortive techniques in more detail, using the RSA in the ResCap case42 and the deathtrap provision in the Momentive case43as my principal illustrations. In Part II, I ask why bankruptcy courts have been so willing to condone their use, particularly given courts’ hostility to distortion in other contexts. I argue that the answer lies in the dramatic recent changes that have made corporate reorganization much more difficult than in the past, coupled with the surprisingly unneutral baseline of the Chapter 11 voting process.
Parts III and IV develop and apply a framework for scrutinizing the new distortive techniques. Part III offers four rules of thumb. The first and second focus on the threat of holdout behavior and the magnitude of procedural or entitlement coercion. The third and fourth consider any independent justifications for a distortive technique that might warrant marginally more coercion and in rarer cases, any special contractual terms that might call for less. I conclude the Part by considering whether and when a creditor should be permitted to change its vote on a plan. In Part IV, I apply the rules of thumb to some complex recent cases: the elaborate RSA for Puerto Rico Electric Power Authority (PREPA), Puerto Rico’s electricity company; the convoluted deathtrap provision in the Momentive case; and the use of both RSAs and deathtraps in two prominent coal-company reorganizations.
Although my focus is on strategies that distort the Chapter 11 vote, the analysis has implications for a variety of related issues. In Part V, I consider two of the most obvious extensions: so-called gifting transactions in Chapter 11 and coercive bond exchanges outside of bankruptcy. Gifting transactions are especially interesting, because they are quite similar to RSAs and deathtrap provisions in some respects, yet courts have viewed them with considerably more suspicion. I consider in Part V why this might be so.