Printer Friendly



Rote use of a standard-form contract term can erode its meaning, a phenomenon made worse when the process of encrustation introduces various formulations of the term. When they occur, rote usage and encrustation weaken the communicative properties of boilerplate terms, leading some terms to lose much, if not all, meaning. In theory, if a clause is emptied of meaning, it can create a contractual black hole in which, as the term loses meaning, random variations in language appear and persist. What, then, are the consequences if parties exploit these variations in language by success fully advancing an interpretation the market disavows? Traditional doctrine holds that even if the court errs in the meaning it gives to a clause, parties have an incentive promptly to revise the standard language to exclude the aberrant interpretation. But what if the assumptions about the costs and motivations to revise this type of boilerplate are wrong? We seek purchase on this question with u study of the pari passu clause, a standard provision in sovereign debt contracts that almost no one seems to understand. This clause gained fame in 2011 because of a series of court decisions in New York arguably misinterpreting a particular variation of the clause. Even though the courts' interpretation put at risk a multitrillion dollar debt market, meaningful revisions to the language of the boilerplate term did not begin to appear until late 2014. In the interim, trillions of dollars in bonds were issued with an uncorrected version of the term. Market forces, in other words, worked slowly to remedy a systemic problem that caused substantial costs. We ask whether the state could do more to avoid the problem at the front end rather than depend on market forces to correct court error at the back end.
Table of Contents

  I. When Standardized Contract Terms Become Rote and Encrusted
 II. A Prototypical Example: The Pari Passu Clause
     A. The Inquiry: Does the Market Repair Black Holes?
     B. The History of Pari Passu Litigation
III. The Post-Litigation Data
     A. Data Sources and Coding
     B. Major Changes to the Core Language of Pari Passu
     C. Minor Revisions: Modifications That Might Influence
           Interpretation or Scope
        1. The Evidentiary Patch
        2. Scope of the Clause
     D. The Empirical Results of the Post-Litigation Study
        1. All Issuers
        2. Pure Sovereign Issuers
        3. The October Meetings
        4. The Pattern of Changes in Boilerplate: Rational
           Design or Random Mutation
     E. Interviews with Market Players
        1. Reasons for the Lack of Revision as of October 2014
        2. The Courts Failed to Solve a Systemic Problem
     F. Reports from the Clients
        1. The Reasons That Might Induce Revision: Standard
           Practice and Investor Preference
        2. Interviews with Fund Managers and Bankers
 IV. Normative Implications
     A. More on Collective Action with a Heavy Dose of
           Agency Costs
        1. The Coordination Problem: Private Versus Collective
        2. The Incentives of the Elite Lawyers
        3. The Sovereigns' Incentives
        4. The Investors' Incentives
        5. The Role of the Public Sector and Industry
     B. What Should Courts Do with Contractual Black Holes?


A question that courts face whenever they are asked to interpret a standard provision in a commercial contract is how to determine what the parties understood that provision to mean when they contracted. The interpretive goal in contract cases is to recover and then enforce the parties' apparent intentions, as they existed at the time of contracting. This goal implies that courts will attempt to interpret even ambiguous terms in a manner consistent with the ex ante intentions of the contracting parties in so far as a court can recover those intentions from the contract or the surrounding context. (1) But standardized terms in boilerplate contracts between sophisticated parties are vulnerable to misinterpretation. At the limit, a boilerplate term that is reused for decades and without reflection merely because it is part of a standard-form package of terms, can be emptied of any recoverable meaning: this creates a contractual black hole. (2) More commonly, terms that have lost much meaning still may provoke litigation over essentially meaningless variations in the boilerplate language. In this latter case of contractual grey holes, (3) courts may be functionally incapable of devising a plausible meaning that was attached to the linguistic variations at the time the contract was drafted. Thus, regardless of whether a boilerplate term has lost all or only almost all meaning, courts will face an interpretation conundrum that we collectively term the "black hole" problem. (4)

The dilemma that courts sometimes face when interpreting boilerplate is an inherent cost of the reliance on standardized contract terms in commercial contracts. Boilerplate terms are ubiquitous in commercial contracting because they offer the efficiency advantages of standardization. Those advantages include the development of a uniform system of communication that is independent of any particular contractual context. (5) Thus, parties in heterogeneous environments who wish to communicate a shared intent can embody that intent in a fixed and reliable formulation whose meaning does not vary with the nature of the contract or its context. (6) Unfortunately, the very elements of fixed and unchanging meaning that make boilerplate terms attractive are the same elements that can lead to the process of repetition without reflection that contributes to the erosion of that meaning over time.

In addition to the ordinary risks of obsolescence, the repetitious use of boilerplate has two pernicious effects that can undermine the utility of some boilerplate terms. The first effect is "rote usage": some standardized terms may get used by rote so consistently that they lose a shared meaning and become a ritualized legal incantation. (7) "Encrustation" is a second cost of too much repetition: the intelligibility of language deteriorates significantly as legal jargon is added to standard formulations, leading to linguistic variations of the same clause. (8) This process further weakens the communicative properties of boilerplate terms, reducing even more their reliability as signals of what the parties really meant. (9) In combination, terms that develop linguistic variations and thereafter are repeated by rote, even after the original meaning has been largely lost, can become contractual black holes.

Contractual black holes present a heightened risk that courts may be persuaded to adopt an interpretation of the term at issue that is antithetical to the functioning of a market that relies on the standard contract to regulate the rights and duties of the participating parties. The market may have disregarded the term before a court gives the black hole a contemporary interpretation. Nonetheless, the market may have an understanding of what the term does not mean, often because of the high cost to market participants of an aberrant interpretation. Unfortunately, correcting the courts' interpretation in such a case can be a slow and difficult process, even when the judicial error imposes high costs. This resistance to revision is a function of what we will call the "evidentiary vacuum" that results when a term loses its original meaning and lacks an evidentiary record of its contemporary meaning. Here the chance of a court issuing an aberrant interpretation is heightened as compared with a boilerplate term with a well-known meaning.

These conditions appear to describe the case of the pari passu clause, a boilerplate formulation common to sovereign debt contracts for nearly 200 years whose contemporary meaning was hopelessly unclear. (10) The recent history of judicial interpretation of this clause began in Brussels in a case against the Republic of Peru in September 2000, in which a court issued the first interpretation of the clause in at least a half century. (11) The same interpretation of pari passu was affirmed by a federal court in New York in a case against the Republic of Argentina in December 2011, (12) and affirmed again on appeal in that same case in October 2012 (13) and August 2013. (14) In each of these cases, the courts endorsed an interpretation of a particular variation of pari passu that required holdout creditors to be paid in full as a condition to the sovereigns paying consenting creditors under a restructuring agreement. Even though this interpretation effectively undermined efforts by sovereigns to restructure their bonds, and even though the courts' interpretation was widely vilified in the market, meaningful revisions to the language of the boilerplate term did not even begin to appear until late 2014. (15)

The extent of rote usage and encrustation in commonly used boilerplate remains an open question because the issue is only now beginning to be examined by legal and economic scholars. (16) But preliminary evidence from other markets where standard form contracts are ubiquitous suggests that the pari passu saga is representative of a larger phenomenon, extending beyond sovereign bond contracts to other contexts where boilerplate is commonly used, including insurance and merger and acquisition agreements. (17) To the extent this problem exists more broadly, it argues for a shift in contract doctrine away from the futile and ultimately costly effort to discover a shared meaning that no longer exists. (18) In this Article, we explore whether courts, or other state actors, might better deal with contractual black holes in boilerplate contracts. Our purpose is to begin the scholarly focus on the effects of this phenomenon by using the pari passu story as a prototypical exemplar. We use both qualitative and quantitative data to support the claim that courts searching for shared intent in the case of black holes in standardized contracting can result in substantial social costs.

The Article proceeds as follows. In Part I, we focus on identifying the causes and effects of black holes in standard form commercial boilerplate. We show how both agency and coordination costs peculiar to the inertia that results from black holes can undermine the standard assumption that commercial parties will promptly revise an unanticipated interpretation that generates market inefficiencies. Part II frames the inquiry: Does the celebrated dispute over the meaning of the pari passu clause provide evidence that the current means of revising black holes impose large and uncompensated social costs? In Part III, we evaluate a dataset on pari passu clauses assembled from over 1500 sovereign and quasi-sovereign issuances in the period from June 1, 2011 to May 30, 2016, together with interviews with the participants in the market. Part IV discusses the implications of the finding that, at least in certain markets, commercial actors face extraordinary difficulties in repairing black hole boilerplate following what the market perceives as an aberrant judicial interpretation of the term's contemporary meaning. These difficulties present a collective action problem whose solution can be daunting owing to the combination of coordination and agency costs that may impair the efficiency of standardized commercial contracting in large-scale markets.


We earlier introduced the concept of black holes in terms of two related but independent factors that contribute to the unintelligibility of standardized language. The first, rote usage, results when, despite many years of repetitive usage, no serious legal challenges or other methods of validation affirm the meaning of a routinely invoked term. (19) The absence of disputes over a widely used but poorly understood term, or of a collective process of updating the meaning of such standard terms, (20) disables courts or other authoritative bodies from providing parties with a tested interpretation of the boilerplate term that minimizes the risks of unintended effects. Rote usage may also develop as a species of contractual overkill, for example, as in forms that designate already enforceable agreements as "signed and sealed." (21) In effect, these rote terms become platitudes. (22) Nonetheless, the linguistic formulation continues to be retained and repeated because "parties see no reason to eliminate a term they view as costless and thus incur a risk, however small, of jeopardizing" the market's understanding of their agreement. (23) Encrustation is a second cost of too much repetition: the intelligibility of language deteriorates significantly as legal jargon and other linguistic variations are overlaid on standard formulations. Rote usage and encrustation are related phenomena although they may affect some boilerplate terms independently. (24) When combined in a particular clause or phrase, a term becomes linguistically uncertain, as no particular meaning can be uncovered that is more probable than any other meaning.

It is important to distinguish the linguistic uncertainty that creates black holes from the more familiar interpretive challenges courts face when interpreting ambiguous terms. A term is ambiguous when it is "capable of more than one sensible and reasonable interpretation." (25) Terms that are linguistically uncertain in the sense we use here are not ambiguous but rather are acontextual. The term in question can apply to an infinitely wide spectrum of referents, as opposed to one or more reasonable alternatives, because there is no basis in the relevant context to determine what, if any, shared meaning exists.

What, then, is the mechanism that produces an encrustation that is then repeated by rote in standard boilerplate? Those running the deals in standard markets describe the process in the following terms: Lawyers hired to do a deal for clients--such as a bond offering--are instructed to use market-standard forms so far as possible. But the deals have to be tailored to the client's needs as appropriate. Names, dates, locations for payment, currencies, and other items have to be changed from whatever prior deal document is being used as a template. The client's assumption is that the lawyers possess the expertise to make the necessary marginal modifications to the standard forms to insure that they both fit the client's preferences and do not depart significantly from what the market would consider the standard package. But lawyers working with standard form language repeated for many years by rote are unlikely to have much, if any, understanding of the purpose served by these terms. The combination, then, of making marginal modifications to the contract terms to suit the needs of the transaction, coupled with ignorance of the terms' function, can result in the insertion of legal language that attempts to, but ultimately fails at, adding clarity to the terms. These insertions may occur with greater frequency when the attorneys involved have less experience with the particular boilerplate term. (26)

Encrustation thus results from adapting standard language whose contemporary meaning is unclear to the drafter. To be sure, emendations may occur in other contexts as well. But there are more error-correction mechanisms for those boilerplate terms that do have well understood meanings and frequent usage. Subsequent drafters that see a variation in an understood usage will be less motivated to repeat the variation if the earlier amendments have changed this meaning and usage in undesirable ways. Where a term has lost a shared meaning, however, and is repeated by rote simply as part of a standard package, these error-correcting mechanisms will not apply. This then leads to increasing uncertainty in the meaning of the variations in the boilerplate term.

Indeed, the very popularity of a clause as part of the standard package of terms in a given indenture can increase even as the term becomes more encrusted and less understood. (27) This phenomenon is propelled by information cascades. In an information cascade, followers might imitate the variation regardless of what they think of the merits of the choice as long as they believe there is some chance the earlier (or even only the first) movers might have had better information about the term's intended meaning. (28) Thus, one person's tinkering can become contagious even when nobody else knows what the revision is supposed to mean.

Ironically, because of rote usage, some clauses become enshrined on the standard checklist expected by market participants to be found in all contracts of a particular type or in a specific market. (29) These checklists function as templates that reduce the learning costs for potential contracting parties: a list of essential elements for every standard deal facilitates comparisons across contract documents. Paradoxically, however, habitual reliance on standard templates also exacerbates the black hole problem. Once a term becomes an essential part of a package that signals a standard set of contractual rights and obligations, rote usage will increase, thereby accelerating loss of meaning. (30) Rather than consider the underlying substance of a standard checklist, contracting parties may simply reproduce the terms reflexively. In this evidentiary vacuum, the chances of a court making an aberrant interpretation are increased as compared with a boilerplate term with well-known meaning. The market may have overlooked the function of a clause prior to litigation over its meaning. However, the market will thereafter develop views on how the court's interpretation affects the parties' contractual distribution of rights and obligations. And because a black hole had no meaning or function prior to the interpretation, the court's interpretation will likely upset this distribution in unexpected and negative ways.

The standard assumption among both commercial lawyers and legal academics is that the social costs of rote usage and encrustation are small because the costs of judicial error will be limited to the isolated case of an aberrant interpretation. This belief is supported by the reasonable assumption that sophisticated commercial parties can and are motivated to readily correct a court's interpretive mistakes. (31) Indeed, given the important role that standardization plays in replicating boilerplate terms in tens of thousands of commercial contracts and the nontrivial possibility that a court may err in interpreting those terms that are infected with rote usage and/or encrustation, commercial parties have incentives to revise their standardized contract terms promptly to ensure that a common meaning is preserved. Thus, the theory predicts that "harmful heuristics, like harmful mutations in nature, will die out." (32) Leaving inefficient interpretations of encrusted boilerplate unrevised produces unacceptable levels of uncertainty and reduces the gains from contracting.

But what if the assumption that markets evolve rapidly to repair inefficient interpretations of encrusted boilerplate is incorrect? It is, after all, an assumption based on little in the way of empirical or theoretical work on the production process of contracts. (33) Indeed, there are well-documented circumstances where sophisticated commercial parties fail to react promptly to inefficient judicial interpretations. Marcel Kahan and Michael Klausner have catalogued many sources of inertia that can delay revisions to standardized contracts. (34) But encrusted boilerplate presents an even greater challenge to the standard assumption: here commercial parties may face unique inertia costs that explain why they are unable easily to convert the encrusted boilerplate into a new and intelligible formulation. In short, boilerplate that has devolved into a black hole may motivate inaction that lasts well beyond the ordinary time for market adjustments.

Why might inertia costs be higher in the case of contractual black holes? Rote usage leads to a loss of shared meaning, both between the contracting parties and across time in the market. Over time, we theorize, encrustations introduce essentially random variations in the language of the boilerplate term: these variations are stripped of any context and thus they offer no means of distinguishing legal rights and duties among the different formulations. (35) Because these random variations are harmless for long periods of time, they persist--at least until an adverse legal interpretation poses a systemic threat to the market. (36) As the pressure then mounts to repair the damage caused by the black hole, parties confront a significant increase in the level of uncertainty across all the dimensions of inertia. Individual parties are reluctant to change the language of the disputed term because they cannot offer a plausible alternative to the aberrant interpretation. In addition, they are unsure how courts will respond to the changes they do make. Until the revised term is tested in litigation, there is uncertainty over how courts will interpret what had been an essentially "empty" term. Individual parties also may be reluctant to draft new contractual language in the evidentiary vacuum of a black hole because the new clause may be used to show that the unrevised clauses in contracts that are still outstanding mean something different from the new clause. For example, change to the contract language that closes off holdouts in a newly issued bond might put unrevised clauses in prior bonds of that sovereign at greater risk of enabling holdouts. These "legacy" costs increase as does market uncertainty: changing a term thus poses the further risk that the standard indenture will be viewed as idiosyncratic, thereby increasing learning costs. (37)

The reluctance to act in an evidentiary vacuum is greater when a contracting party moves unilaterally while other parties continue to use the existing boilerplate clause, even after a disfavored court interpretation. The uncertainty over the meaning of a black hole term coupled with the continued use of the term by others in the market heightens the risk that subsequent courts will view a unilateral revision of the clause as confirmation of the first court's interpretation. In contrast, if the market as a whole moves promptly to change the contract language, it sends a clear signal rejecting the court's interpretation of the black hole clause and affirming the market's preference for an alternative interpretation. The black hole hypothesis predicts that change will not occur in a significant way until market participants, confronting a black hole term that the market historically has disregarded, are able to solve a vexing collective action problem and coalesce around a new industry standard. Once the collective forms, the inertia costs for individual actors are significantly diminished. But as we discuss below, the very uncertainty that deters unilateral efforts to revise black hole boilerplate also increases the coordination costs that make collective action, including the efforts of public institutions to assist the market in clarifying the term's meaning, difficult to achieve. (38)

The pari passu saga that we describe below illustrates the substantial social costs caused by aberrant interpretations of contractual black holes. These costs include the long-term consequences of failed efforts to restructure bonds owing to the large number of inefficient contracts containing a now-disfavored pari passu clause, and the cost of public efforts to overcome collective action problems and induce change in response to the judicial error. Given these costs, it may be misguided for contract doctrine to adhere stubbornly to the standard interpretive command to courts to find what the parties to the litigation meant by the black hole term. Instead, social welfare considerations may support interpreting black hole boilerplate terms infected by rote usage or encrustation in ways that reduce ex ante the ex post costs of an inefficient court interpretation. (39)


A. The Inquiry: Does the Market Repair Black Holes?

Our normative claim that, in some cases, courts should abandon a search for party intent when interpreting certain boilerplate terms is often difficult to support empirically. Consider the case where a single court in one jurisdiction interprets a boilerplate term and, despite public outcry, the market continues to use the same unchanged term for years after the interpretation. Notwithstanding the public reaction, it will be challenging to discern whether, in fact, the market views the court as erring in its interpretation. If it is indeed the case that market actors do not react quickly to judicial error, all that an outsider observer sees is that the boilerplate terms in the contract have not changed. While the absence of revision may mean that the market is constrained in some way that impedes a prompt response to inefficient interpretations, the absence of revision is also consistent with the possibility that the court's interpretation reduced hitherto unacknowledged inefficiencies in the standard form. In that case, the absence of a market-wide revision to the term in question would be a confirmation of the welfare benefits of the novel interpretation. (40)

What evidence would be sufficient to support the claim that the market is peculiarly disabled from promptly revising or deleting a contractual black hole following an erroneous interpretation? Ideally, one would hope to see multiple courts across different jurisdictions interpreting an encrusted boilerplate term in a particular way, followed by numerous contracts written with the same unrevised term. To be sure, there are many sources of inertia that delay revisions to standardized contracts, (41) but inaction that lasts well beyond the typical time for market adjustments (42) justifies the inference that either contractual black holes are associated with extraordinary inertia costs (43) or the market is endorsing the prevalent court interpretation.

But how can one then distinguish between these two plausible reasons for the lack of a market response? Suppose after an extended delay we do observe an eventual and wholesale revision in the language of the boilerplate that explicitly rejects the courts' interpretation of the meaning of the black hole term. With multiple instances of judicial interpretation, particularly if the courts are taken as authoritative, researchers can more credibly dismiss the explanation that the interpretation from an aberrant court does not reflect the prevailing judicial view. And when the eventual changes to the clause expressly reject the prevailing judicial interpretation, researchers can dismiss the explanation that the market had endorsed the efficiency of the newly minted boilerplate. The foregoing conditions are, unsurprisingly perhaps, hard to find. However, the decade-long litigation between NML Capital, the hedge fund of the U.S. billionaire, Paul Singer, and the Republic of Argentina, has provided a natural experimental setting that allows us to test the market response to court decisions interpreting a boilerplate term: Is the delay in the market's response more consistent with a rational meaning for the term or with a contractual black hole that impedes the ability of the market to react to a disfavored court interpretation? (44)

B. The History of Pari Passu Litigation (45)

To understand why an examination of the history of the pari passu clause offers a valuable natural experiment, one needs to return to litigation that occurred in 2000 involving NML's predecessor, Elliott Associates, another hedge fund run by Mr. Singer. In that instance, Singer's fund was pursuing the Republic of Peru on debt claims his fund had purchased at a deep discount. (46) Obtaining a judgment against Peru for nonpayment of the debt was straightforward, as there was no dispute that Peru had not paid. Enforcement of the judgment proved much more difficult, as it always is with sovereign debtors, which is why defaulted sovereign debt trades at a deep discount on the secondary market. Elliott Associates had purchased discounted debt and sought to use its legal expertise, along with its unusually deep pockets, to recover in full from recalcitrant debtors so as to make a profit worthy of a hedge fund. As part of that endeavor, Elliott was chasing Peruvian assets in a variety of jurisdictions around the world, including Brussels. There, Elliott struck pay dirt: The commercial court in Brussels ruled, on an ex parte motion, that the version of the pari passu clause used in the Peruvian debt contracts (47) meant that Peru could not pay any other creditors without paying Elliott a pro rata share. (48) Further, the judge ruled that Elliott was entitled to an injunction against Euroclear, the Brussels-based financial clearinghouse, which would bar payments to the holders of restructured bonds (who were being paid in full on their restructured amounts) unless Elliott got its full payment on its unrestructured amount. (49)

The international financial community reacted with alarm. The luminaries in the field uniformly condemned the decision as inconsistent with long-held market understandings of the meaning of pari passu. Whatever the clause meant, and no one was certain of its contemporary meaning, the experts asserted that everyone knew that it did not mean that all creditors must be paid pro rata, despite a restructuring agreement giving preference to consenting creditors. (50) But notwithstanding the dismay expressed in dozens of academic and policy articles, (51) and even after a proposal for a new international bankruptcy court for sovereign debtors led by the International Monetary Fund (IMF), (52) for over a decade virtually no pari passu provisions in sovereign debt contracts were modified to clarify this understanding. (53) The same version of the pari passu clause that had led to the supposedly unsupportable decision in Brussels was used over and over again for a decade in literally hundreds of contracts. (54)

What explains this disjunction? The same law firms whose senior lawyers were declaiming the clear error of the Brussels interpretation were simultaneously continuing to use the same problematic version of the clause in their own sovereign clients' documents. (55) To understand why this "patent judicial error" was uncorrected in subsequent contracts, two of us conducted nearly a hundred interviews with leading sovereign debt lawyers in New York, London, Paris, and Frankfurt. (56) Their answers varied, but the typical response was: "Why should we change the clause? No court in New York or London would ever make such an error. This was an aberrant decision from an obscure court in Brussels." Moreover, the debt lawyers argued, this was an ex parte decision (57) : If Peru's lawyers had been given an opportunity to argue their position, the case would never have come down that way. (58)

Ten years later, in December 2011, a federal district court in New York shattered the assumptions of the sovereign debt lawyers who had heaped contempt on the Brussels court. This time, the hedge fund was NML Capital, holding defaulted debt owed by the Republic of Argentina. NML requested an injunction from a federal judge in New York based on a claimed violation of essentially that same version of the pari passu clause. (59) And this time the litigation was not ex parte, as Cleary Gottlieb Steen & Hamilton, the law firm on the sovereign side, had ample opportunity to make its arguments in full. Moreover, the judge had handled major sovereign debt disputes before (60) and had access to a decade's worth of academic research studying the impact of the Elliott v. Peru cases. (61) Nevertheless, after hearing extensive argument from both sides, the court held that this version of the pari passu clause required a pro rata payment to all creditors, including holdouts, in essentially the same way as had the Brussels court. (62)

The financial community again expressed profound dismay at the court's interpretation of the pari passu boilerplate language. (63) The almost universal assumption of the sovereign debt community of lawyers, academics, and government officials was that the United States Court of Appeals for the Second Circuit--traditionally, the preeminent court in the country on business law matters--would overrule the district court and repudiate the pro rata sharing interpretation of pari passu. Numerous amicus briefs were filed, including briefs by the U.S. Department of Justice, (64) the Clearing House Association, (65) and the American Bankers Association, (66) most of which excoriated the trial court's interpretation of pari passu.

Yet, in October of 2012, the Second Circuit's three-judge panel unanimously affirmed the trial judge's interpretation of Argentina's pari passu clause. (67) Again there were expressions of surprise and alarm by market insiders and more briefs filed, including by a Nobel laureate expert in sovereign debt, and three separate countries (adding their voices to that of the United States in the lower court), (68) asking the U.S. Supreme Court to take the case and repair the damage the pro rata interpretation of pari passu was inflicting on the market for sovereign bonds. Nevertheless, in June 2014, the Supreme Court declined to hear the case, seeing no substantial reason to disturb the decision of the lower court. (69)

In April 2013, as a result of the failure to get relief in the courts and because the market seemed unable or unwilling to fix the pari passu problem on its own, an effort to solve the problem began at the Spring IMF/World Bank Meetings. (70) The primary conveners were the U.S. Treasury Department and the IMF, with the support of the French Finance Ministry's Paris Club, the International Capital Market Association (ICMA), the Bank of England, and representatives from major ministries of finance and industry groups around the world. (71) This group of experts met again at the Fall IMF/World Bank meetings in October 2013. (72) In December 2013, after extensive discussions with the foregoing committee of experts, ICMA, the leading industry group, issued new proposed versions of the pari passu clause in draft form. (73) The IMF, which had been actively involved in the ICMA process, followed a few months later with an extensive report recommending that sovereign issuers revise their pari passu clauses. (74) Finally, in September 2015, after their meeting in Ankara, the G-20 Finance Ministers and Central Bank Governors included an endorsement of these reform efforts on the part of the IMF in their communique. (75)

As of early 2017, the IMF's latest report on revisions to pari passu reveals that a majority of pure sovereign issuers have now clarified or modified outright their pari passu clauses. (76) What the foregoing provides, now that widespread modifications have appeared, is the opportunity to unpack the mechanics of the process by which boilerplate terms are revised. By using a combination of the data on the changing language of pari passu and interviews with market participants involved in the revision process, one can begin to answer the questions how and when contractual black holes, infected both with rote usage and encrustation, are remedied by revisions to contract language. In the ten years between the Brussels decision in September 2000 and the New York decision in December 2011, virtually no sovereign issuer amended its pari passu clause to explicitly foreclose the interpretation in Elliott v. Peru. (77) But as of the IMF's report from mid-2015, the data show an extensive movement--albeit among only one distinctive subset of sovereign issuers--to reject explicitly the ratable payment interpretation advanced by the Brussels and New York courts. (78) In the following Part, we examine the dynamics--and associated costs--of the revision process.

Understanding those dynamics is critical to answering the initial question: How accurate is the assumption that sophisticated commercial actors will overcome inertia costs and reject a judicial interpretation of a black hole in a standard form boilerplate contract that regulates a major industry? The empirical study described below suggests that the standard assumption is false, at least in the context of one specific multi-trillion-dollar market. Changing a boilerplate term whose meaning has largely been forgotten and that has developed numerous encrustations can take years, and the process can prove enormously costly, particularly in terms of the hundreds of billions of dollars' worth of bonds issued with suboptimal terms in the interim period. (79) Those costs support the claim advanced in Part IV that a search for party intent is both futile and counterproductive when boilerplate with these characteristics is included in industry contracts.


A. Data Sources and Coding

To unpack the boilerplate revision process, we assembled a dataset of all of the available sovereign and quasi-sovereign bonds for the five-year period between June 1, 2011 and May 30, 2016. As sources, we used Thomson One Banker and Perfect Information, the two primary public data sources of offering circulars and prospectuses for debt offerings. For data on prices and maturities, we supplemented the foregoing with information from DCM Analytics. The datasets provide documentation on what the industry designates as "managed deals," where bankers and lawyers are involved in preparing the contract documentation, setting the initial prices, and finding the initial customers. The strongest AAA issuers, the United States, United Kingdom, Netherlands, Japan, Germany and France, are not included: owing to their strong credit, they do not use traditional contracts with pari passu clauses to sell their bonds. (80)

The foregoing cut of the available sovereign bond data produced a set of 1583 bonds from 90 different issuers. The bond issuances were governed either under local law or a dozen different governing foreign law regimes. For each of these bonds, we coded two basic bond characteristics. (81) These were (1) the type of issuer divided into two sub categories--pure sovereign bonds issued by nations, and quasi-sovereign issuances, such as sovereign guarantees, sub-sovereign bonds of cities, states and provinces and supranational, (82) all of whom have a degree of sovereign immunity and, most importantly, are either not subject to a bankruptcy regime or where the applicability of bankruptcy is unclear; (83) and (2) the governing law variable as either local law--the sovereign's contracts are governed by its own law--or some foreign law that the sovereign cannot change at its discretion, typically English or New York law.

We turn next to focus on the primary questions of whether and how the pari passu clauses changed in the wake of the NML v. Argentina decisions by the New York courts during the 2011-2013 period. Here, changes to the pari passu provisions were coded in terms of two different types observed in the data (a) major changes to the core language of the pari passu clause and (b) minor changes at the margins that might, for example, expand or contract the scope of the clause's applicability or augment the available evidence on the meaning of the term.

B. Major Changes to the Core Language of Pari Passu

First, we considered whether the clause changed in the direction of either the pro or antiratable payments interpretation. Our hypothesis was that the market would either delete the pari passu clause or revise it to reduce the risk of another court adopting the ratable payments interpretation. The baseline for each bond was the version of the clause that the relevant issuer was using before June 2011. From that baseline, we examined whether and what changes were made to the boilerplate language subsequently.

The typical and oldest version of the pari passu clause, a clause found in almost every sovereign or quasi-sovereign bond contract, states in essence:
   The notes rank and will rank without any preference among
   themselves and pari passu with all other unsubordinated public
   external indebtedness of the Republic.

We refer to this version of the clause as the "Rank" clause. Prior to the pari passu litigation, eminent practitioners in the field were openly puzzled as to why this boilerplate clause was used in a sovereign debt instrument. (84) The concept of Rank has a clear meaning in a domestic bankruptcy context, where a judge supervises the division of the debtor's limited assets among creditors of different rank. (85) All those creditors who are of equal or pari passu rank share equally once the creditors senior to them have been paid. But sovereigns do not and cannot enter a judge-supervised bankruptcy procedure. There is no procedure by which an insolvent sovereign's assets--primarily, its ability to impose taxes on its citizens--get divided. The clause, these expert sovereign debt lawyers were saying, was meaningless boilerplate, a holdover from a distant era when its inclusion in these sovereign instruments might have made more sense. (86) From the viewpoint of the contemporary market participants, the pari passu clause in sovereign debt instruments was precisely the kind of contractual black hole that had lost meaning over the years as a result of encrustation and rote usage.

Over the roughly 200 years that the clause appeared in debt instruments--steadily increasing in popularity, even as contemporary understanding diminished--encrustations began to emerge as words and phrases were added and subtracted. These variations produced several distinct linguistic departures from the original "Rank" version of the pari passu clause. For example, in 2010, immediately prior to the sample period, Italy adopted a sui generis pari passu clause in its New York law bonds providing that:
   [t]he Securities are the ... unsecured obligations of Italy and
   will rank equally with all other ... unsecured and unsubordinated
   general obligations of Italy for money borrowed.... Amounts payable
   in respect of principal of (and interest on) the Securities will be
   charged upon and be payable out of the [Treasury of Italy], equally
   and ratably with all other amounts so charged and amounts payable
   in respect of all other general loan obligations of Italy.

We term this version of the clause, that provides for explicit ratable payments, the "Pay Equally and Ratably" pari passu clause. The Pay Equally and Ratably clause poses a high risk of holdouts as it appears to promise that all creditors be paid ratably whether or not they have consented to a restructuring.

Sometime in the 1980s, other bonds began to promise investors yet a third version:
   The bonds rank, and will rank, pari passu in right of payment with
   all of the Issuer's present and future unsubordinated External
   Indebtedness. (87)

We term this version the "Rank Equally in Payment" pari passu clause. This was the version of the clause Singer's hedge fund seized upon, first in Brussels and later in New York. Singer's fund found a New York University law professor who opined that he understood the meaning of the clause. A sovereign debtor who was in arrears to creditors, Professor Andreas Lowenfeld explained, had to pay the creditors who ranked pari passu in right of payment on a pro rata basis. (88) And importantly, given that a court's order to pay is largely meaningless against a sovereign debtor, Lowenfeld explained that the clause was an intercreditor agreement that entitled a creditor who was not paid his pro rata share to an injunction against other creditors who were paid that share. (89) In the context of a debt restructuring, where some creditors have agreed to take a haircut on their bonds and others are holding out and not receiving any payment, an injunction against another creditor is a powerful remedy for the holdouts. That was precisely what the creditors in Brussels and in New York asked for and obtained.

For coding purposes, it is important to recognize that the version of the pari passu clause at issue in both the Argentine and Peruvian litigations was particularly vulnerable to Lowenfeld's pro rata or ratable payment interpretation: the word "payment" was used to modify the promise on the part of the issuer that the bonds would "rank equally." The clauses at issue in both litigations went beyond the standard "rank equally" language to include the promise that "the bonds rank, and will rank, pari passu in right of payment."

We rated the vulnerability of the different clauses based on the risk of successful litigation by holdout creditors on a scale of high risk to zero risk, with high risk being maximal vulnerability (where the contract language said explicitly that each creditor was entitled to equal ratable payments); to medium risk (the type of clause that used the word "payment" to modify "rank"); to low risk (which only articulated the concept of ranking). (90)

Finally, we coded a category of bonds in a subset of clauses primarily governed by English law as having a near zero risk. In these bonds, the standard pari passu language gets supplemented by words such as "except as subject to provisions of mandatory law." (91) On its face, this additional language appears to say that the effect of the basic clause can be negated by the passage of a local law. In other words, the debtor has the power to render the clause meaningless as a creditor weapon. We term this version of the pari passu clause the "Mandatory Law" variation. Table I summarizes the different types of pari passu clauses found in the data set in 2011 and the risk of holdouts from a debt restructuring that each posed prior to the Southern District of New York trial court opinion ("SDNY opinion"). (92)

C. Minor Revisions: Modifications That Might Influence Interpretation or Scope

Initially, we assumed that if parties disagreed with a court's interpretation of a boilerplate clause in a standard form contract, even accounting for ordinary friction costs, they ultimately would adopt what we have called a "major change": either delete the term or revise its core language to reject the now-disfavored meaning. Instead, we found in the data a number of unanticipated minor changes.

1. The Evidentiary Patch. The most common minor modification reflected in the data was the introduction of an evidentiary patch--a supplementary sentence in the offering prospectus explaining what the clause did not mean. (93) As a matter of contract doctrine, such a modification made outside the core language of the relevant contract provision is less effective than a major change to the contract language itself. (94) A court will focus first on the language of the clause in the contract: even if there is elsewhere in the offering documents a disavowal of one particular meaning, the risk of misinterpretation remains. In an analogous evidentiary patch, some bonds in the database include a further supplementary explanation regarding risk, referred to as a "risk factor" disclosure. Here, issuers report in the risk disclosure section of the offering prospectus that the pari passu clause in the contract poses a risk of an adverse interpretation for investors. (95) These evidentiary patches and risk factor disclosures may not measurably effect the interpretation of boilerplate terms when there is a large body of available evidence on the term's meaning. Issuers, however, may believe that these additions will have more influence on a court's interpretation of a contractual black hole where there is an evidentiary vacuum.

2. Scope of the Clause. Our final coding concerns the scope of the clause. Pari passu clauses vary in terms of the breadth of the promise they make. Most pari passu provisions in international bonds promise the bonds will rank pari passu with some portion of the sovereign's other unsecured and unsubordinated obligations. (96) The size of that portion ranges from the sovereign's "external indebtedness," which is usually defined as the indebtedness of the sovereign that is denominated in a foreign currency to "all unsecured and unsubordinated obligations," which is sufficiently broad to include the salaries owed by the sovereign to its domestic employees. (97) As long as the clause was seen by the contracting parties as a meaningless artifact, the breadth of the clause did not matter to the parties. However, under the court's injunction in NML v. Argentina, the scope of the clause became of paramount importance. The breadth of the court's injunction halting payments to other creditors was directly dependent on whether the holdout creditors had been promised pari passu treatment with respect to that particular set of creditors. (98) If not, then there was no right to an injunction against those creditors. (99)

Our purpose in coding for scope was to see whether, once the New York ratable payments interpretation was issued, parties became concerned about the reach of their pari passu clauses. Assuming that there was disagreement with the court's interpretation, parties should revise their contracts by narrowing the scope of pari passu in the post NML v. Argentina clauses. We coded for whether the clause was framed in terms of external indebtedness (the narrowest version); all indebtedness (the intermediate category); or all obligations (the broadest version). (100)

D. The Empirical Results of the Post-Litigation Study

1. All Issuers. We hypothesized that the parties to sovereign bond contracts--realizing that a clause whose purpose they did not understand was adding unnecessary risk--would either delete the clause or amend its core language to reject the courts' ratable payments interpretation explicitly. (101) This hypothesis seemed even more plausible after the widely condemned ratable payments interpretation was endorsed by the most important commercial courts in the United States. (102) Parties that disagree with the interpretation should either delete the clause or clarify its meaning.

Indeed, commentary at the time of the SDNY interpretation proposed deletion as a response. Buchheit and Martos, from the leading sovereign debt firm of Cleary Gottlieb, wrote in the wake of the NML v. Argentina decision:
   The clause serves no useful function in countries whose laws do not
   permit the involuntary subordination of an existing creditor, which
   is why it does not appear in the standard documentation for
   domestic debt issuances in most countries. So jettisoning the
   clause altogether will not adversely affect the position of
   creditors and will avoid the risk of further aberrant judicial
   interpretations down the road. (103)

The image accompanying the cover-page article in the New York Law Journal by a lawyer from Arnold & Porter, another leading law firm in the sovereign space, is equally illustrative. (104)

Despite this commentary from prominent sovereign debt lawyers, there is no evidence in the dataset that parties deleted the pari passu clause after the SDNY opinion. Indeed, not a single issuer deleted the pari passu clause in response to any of the court decisions or to the IMF report condemning them. Moreover, not a single bond added language to articulate clearly the rights of creditors or the obligations of sovereign debtors under the clause.

The data does show major and minor changes to the clause of the types mentioned above--revisions that appear designed to reduce the risk of future courts adopting the ratable payments interpretation. (105) We examine the pattern of adoption of these major and minor modifications in the June 2011 to May 2016 period. There was no real movement to change any of the language of the core portion of the clause in any direction for almost three years after the federal district court's first ratable payments interpretation in December 2011. In late 2014, however, revisions began to move dramatically toward the low risk, Rank version of pari passu by removing the word "payment" from the boilerplate term. This convergence on a single type of pari passu clause served to eliminate the language that supported the ratable payments interpretation in the New York and Brussels cases. (106)

We first report the percentage of issuances for all issuers in the dataset that contain at least one change, whether major or minor, to the pari passu clause. To determine a change, we use the pari passu clause in a sovereign's last offering prior to the start of the dataset in 2011 but after 2005 as the point of comparison (the "initial pari passu clause"). Figure 1 reports for all issuers in the dataset. The data is reported on a quarterly basis starting from the second quarter of 2011 to the second quarter of 2016. For each quarter we report the percentage of bonds issued containing changes to the clause relative to the initial pari passu clause. The vertical lines in Figure 1 are the points at which key events occurred such as the SDNY decision (December 2011), the two appellate court decisions (October 2012 and August 2013), and the Supreme Court denial of certiorari (June 2014).

The 2011 SDNY decision produced no immediate reaction in the data. For an entire year after that decision, only a small fraction of sovereign or quasi-sovereign issuers made any major or minor changes to their pari passu clauses. The inactivity in the market is even more vivid if one considers the dollar amount of bond debt that was issued during this period with unmodified versions of the clause. Figure 2 depicts the aggregate dollar amount of issuances by quarter for the bonds in the dataset, categorized by those with and without a change in the pari passu clause, where change is measured relative to a sovereign's initial pari passu clause. Figure 2 excludes those issuers that had a low risk Rank clause as their initial pari passu clause--the eventual market standard as we will see below. Excluding the Rank version allows us to determine the dollar amount of bonds using an unmodified clause that does not represent what the market eventually views as the standard and is thus presumably optimal. As Figure 2 shows, over $1. 5 trillion worth of sovereign and quasi-sovereign bond debt was issued with such unmodified clauses during 2012 alone.

One might wonder here whether the market thought that the SDNY decision was an outlier that was certain to be overturned and therefore did not need to be taken seriously. Yet, in October 2012, the Second Circuit Court of Appeals affirmed the trial court's decision. Though some changes in the pari passu clause then began to occur, the large majority of issuances in both numbers and aggregate dollar amount remained unchanged. During 2013, another $1. 4 trillion of bonds were issued, again without any revision to pari passu. A handful of issuers did revise the contract language in 2013, but the overwhelming majority of bonds were issued without any attempt to modify or clarify their pari passu clauses. (107) The first nine months of 2014 continued in the same fashion with another $0. 9 trillion in bonds issued without revisions in the issuers' clauses, despite a June 2014 U.S. Supreme Court's denial of appeal from the Second Circuit. (108)

Figures 1 and 2 illustrate a simple point: If we assume that continuing to use a clause presenting a significant risk of an interpretation error imposes costs on the parties to the contract, then it follows that substantial social costs were incurred because of an apparently high level of inertia impeding change during this period. Between June 1, 2011 and May 30, 2016, roughly $5. 4 trillion worth of sovereign and quasi-sovereign bonds were issued with terms that were likely suboptimal. (109) To be sure, the precise costs (in terms of the additional risk of holdouts caused by using the risky term) vary across bond issues depending on the likelihood of default and the precise wording of the clause in question. But it seems plausible to conclude that substantial costs were incurred because the market was so slow in adjusting to the decisions of the New York courts. (110)

While change in the pari passu clause came only slowly after June 1, 2011, the pace of change increased after October 1, 2014. In Figure 1, note that the percentage of issuances with pari passu revisions for those issuers not initially with the Rank version of the clause increases more than 100 percent, from 9 percent in the third quarter of 2014 to 19 percent in the fourth quarter of 2014. In Figure 2, note that bond offerings with a revised term by issuers not initially with the Rank version accounted for 19 percent of the aggregate dollar amount of all offerings in the third quarter of 2014 and then increased to 60 percent of the aggregate dollar amount of all offerings in the fourth quarter of 2014. By the second quarter of 2016, bond offerings with a revised contract clause or evidentiary patch accounted for 67 percent of the aggregate dollar amount of all offerings. Moreover, all the major and minor changes during this period were attempts to constrain or repudiate the effects of the NML v. Argentina decision.

In sum, we see evidence of an extended period of inertia followed by increasing volume of changes to the pari passu clause, particularly after October 1, 2014. Our conjecture from observing this phenomenon is that once clauses such as pari passu are encrusted with legal jargon and repeated by rote over many years, they are impervious to amendment for a considerable period. (111) The eventual shift in the market to adopt changes repudiating the ratable payment interpretation to the pari passu clause allows us to rule out a plausible competing hypothesis--that market participants in fact agreed with the SDNY opinion and were thus content with their existing pari passu terms.

To determine what caused the extended delay in effecting the revisions to pari passu, we delve into the dynamics of the changes once they began to occur. Of the bond issues in the dataset that reveal a change in the pari passu clause relative to their last issuance before June 1, 2011, 93. 6 percent involve bonds issued directly by the sovereign, or what we refer to as "pure sovereign" issuances. (112) Accordingly, we next focus on the dynamics of pari passu changes for the Pure Sovereign issuers.

2. Pure Sovereign Issuers. We start with an examination of major changes in pari passu language for the pure sovereign issuers: these major revisions involve the greatest reduction in the risk of holdouts that variations in the pari passu clauses pose. We posit that sovereigns face the greatest inertia costs in undertaking a major modification of the pari passu clause. We focus especially on pure sovereign issuers that are governed by foreign law, as sovereigns that issue under their own local law will have other means to protect themselves against holdouts besides modifying the pari passu clause. (113) We therefore expect that changes in the pari passu clause itself will occur primarily for those pure sovereigns that issue bonds under foreign law, typically English or New York law. Restricting our analysis to pure sovereign issuances under foreign law left us with 791 issuances from June 1, 2011 to May 30, 2016. (114)

In Figure 3, we report the percentage of bond issues by pure sovereign issuers under foreign law by quarter that involve a major revision in the pari passu clause relative to the sovereign's initial pari passu clause. Note the dramatic increase in major revisions to the pari passu clause after October 1, 2014. The percentage of revised clauses issued by pure sovereigns under foreign law increases from 9 percent in the third quarter of 2014 to 40 percent in the fourth quarter of 2014. We use both quantitative and qualitative data to explain below why October 2014 was so important.

Prior to October 2014, very few bonds issued by pure sovereigns under foreign law included a major change to the pari passu clause. (115) A different story emerges when we examine the minor modifications made in response to the court cases in NML v. Argentina. As described earlier, none of these minor modifications would have fully corrected for the courts' interpretation. Yet, minor modifications do appear and potentially with some significance. To provide an example, Brazil reported to investors in the risk disclosure section of its July 2014 prospectus supplement, the following:
   Recent federal court decisions in New York create uncertainty
   regarding the meaning of ranking provisions and could potentially
   reduce or hinder the ability of sovereign issuers, including
   Brazil, to restructure their debt. (116)

It is perhaps an overstatement to suggest that this statement even qualifies as a minor modification of pari passu. The disclosure here does little more than inform the investor that there is a new risk that remains in the documentation: at best, the statement conveys an undertone of disapproval of the NML v. Argentina decision to a future court. (117)

Figure 4 reports minor modifications relative to a sovereign's initial pari passu clause as a percentage of all pure sovereign issuances under foreign law for each quarter of the dataset. Figure 3 showed that there were no major shifts through 2012, and little response in 2013. When we look at minor shifts in Figure 4, the data show some changes in 2012 and increasing changes in 2013 and 2014. Market participants appeared much more willing to experiment with minor shifts prior to October 2014, as compared with major shifts. This is consistent with higher inertia costs affecting efforts to modify directly the risk level of a pari passu clause as opposed to modifications designed to provide evidence on the meaning of the pari passu clause, given the evidentiary vacuum of meaning for a black hole term. Notwithstanding the evidentiary patches and other minor changes in the offering prospectuses prior to October 2014, the overall pattern shown in Figure 3, for major shifts, remains the same for minor shifts as shown in Figure 4. After October 2014, there is a marked increase in the minor modifications to the pari passu clause. All of the changes in Figure 4 are in the direction of constraining the effects of the NML v. Argentina decision by limiting the risk of the pari passu clause.

3. The October Meetings. The timing of the cluster of changes, and the sudden increase in the rate of change in October 2014 correlates with a set of meetings at which a number of the key players in sovereign debt law gathered. The first of those gatherings was held at Columbia Law School immediately prior to the annual IMF/World Bank meetings in Washington, D.C. The second was held shortly thereafter at the New York Federal Reserve. Two of us hosted the Columbia Law School meeting: at that juncture we had collected roughly three quarters of the dataset, which showed clearly the slow response to NML v. Argentina reported above. We were puzzled both by the lack of revision to the core language of the various pari passu clauses, despite the expressed dismay over the NML v. Argentina litigation, and by the prevalence of a strategy that relied on minor modifications to the risk disclosure sections of the bonds. To better understand the data, we invited thirty of the most knowledgeable parties from both London and New York to participate in a discussion of the data. (118) We report in Part III.C the content of the conversations with these senior practitioners. Relevant here, however, is that many of the practitioners expressed dissatisfaction with the draft clauses that ICMA and the official sector had promulgated. Some were unhappy at not being consulted during the ICMA drafting process, others objected to the clause that had been proposed, and still others had clients who were uncertain whether any revisions to the ratable payments interpretation would advance their interests. At the conclusion of the Columbia meeting, we assumed that any significant movement toward wholesale revision of the clause was unlikely in the near term.

Senior representatives of the IMF, the U.S. Treasury and ICMA, as well as senior lawyers who had been on the drafting committee for the proposed revised clauses, attended the meeting at Columbia. Dismayed by the conversation, senior statesmen in the group convened a second meeting a few weeks later. (119) Unlike the Columbia meeting, this next session, hosted at the offices of the New York Federal Reserve, was by invitation only. The Federal Reserve meeting was comprised of a select group of elite lawyers, most of whom were at the Columbia meeting as well. Almost all of the lawyers represented the pure sovereign issuers, mostly emerging market sovereigns from Latin America issuing bonds under New York or English law. Each of the invitees understood that they were being asked by the public sector authorities to contribute to a coordinated effort to combat the danger posed by NML v. Argentina's ratable payment interpretation. (120) Sources who attended reported that the dynamics of the second meeting were remarkably different from the Columbia session: instead of the dissension at the first meeting, a consensus quickly emerged that everyone involved needed, and was willing to, cooperate in trying to solve the systemic problem caused by the rogue interpretation of the New York courts. (121) We cannot prove causation, but the data for the sovereign issuers shows a remarkable change in late October 2014, coinciding with the two New York meetings. (122)

4. The Pattern of Changes in Boilerplate: Rational Design or Random Mutation. Figures 1 through 4 tell us that changes to the pari passu clauses were slow to occur. But they do not give many clues as to why, other than that something happened in October 2014 that increased the rate of change significantly. We draw here from a companion empirical article in which we conduct an econometric analysis of that question. (123)

As described earlier, and as Figure 5 shows, we know that sovereign issuers were not all using the same pari passu clauses at the outset: some were more vulnerable than others to the risk of erroneous interpretation posed by the NML v. Argentina litigation. (124) Moreover, some sovereigns with very high credit ratings might be thought to present such a low probability of default that they were indifferent to the type of pari passu clause used in their contracts. One might wonder whether the issuers that modified their clauses earlier had the most vulnerable clauses and/or the lowest credit ratings. Affirmative answers to both those questions would suggest that the substantial social costs described earlier were exaggerated. Yet no such neat pattern appears in the data. There are some clearly identifiable subsets of issuers for whom changes occur earlier than others. But the pattern is not a clear correlation between speed of revision and strength of credit ratings or vulnerability to the interpretation in NML v. Argentina. Moreover, the initial revisions we observe do not involve major changes to the contract language but rather more evidentiary-focused minor modifications. The fact of sovereigns avoiding major changes and attempting instead to add evidence of their understanding of the existing contract language is consistent with the kind of evidentiary vacuum that is characteristic of a black hole clause. (125)

The most clearly identifiable subset of issuers who undertook either minor or major revisions are the pure sovereigns issuing bonds under either English or New York law. Almost none of the changes come from the quasi-sovereign issuers, such as government-guaranteed bonds, bonds from cities and provinces, and supranational bonds. And, importantly, these quasi-sovereign issuers are often the riskier debtors as compared to the corresponding pure sovereign issuers. In our empirical paper, we analyze the dynamics of change more specifically within the subset of pure sovereign bonds issued under New York and English laws. There, we run a statistical test of the predictions drawn from two competing models of the dynamics of change in boilerplate contracts. (126)

The first model is the traditional conception of contracting that we call the "Rational Design" model. (127) Here, contracting parties are assumed to tailor contract terms rationally and optimally to their needs. If courts appear to err in interpreting contract language, parties will respond promptly by revising the terms to clarify their joint intent. In markets using standard forms, this process of error correction may be slower owing to ordinary inertia costs: individual parties may, for example, rationally prefer the network benefits from retaining the original terms. (128) Nevertheless, variation in the terms among parties doing the same type of deal ultimately will result from the different characteristics of the contracting parties. (129) Rational Design predicts that parties will adjust their pari passu clauses to account for different perceptions of the risk of future default. For example, sovereigns who anticipate, or wish to signal, a low risk of default will be motivated to precommit not to restructure in the future by writing pari passu terms that increase the risk of holdouts. (130) In this model, contractual black holes do not arise and, even if they did, they would not persist.

In the second, "Random Mutation" model, contract language in standardized boilerplate is assumed to follow an essentially random evolutionary path. (131) Standardized contracts are rarely drafted from scratch: they are largely copied from prior deals by agents with imperfect and incomplete understanding of the prior deals and the terms that regulate them. (132) Rote usage and encrustation will occur, particularly with contract terms that are widely used but are not regularly tested or otherwise updated, and the result will be the periodic creation of contractual black holes. Attempting to fit the standard form to the particular requirements of an individual client produces variation that contributes to the formulation of contractual black holes. In this model, repairing black holes is more difficult because of uncertainty. Parties are ignorant of the terms' meaning and of any incentive effects, and in this evidentiary vacuum, coordination is required to effect a revision.

After running a series of tests pitting the two models against each other, there is little evidence in the data to support the Rational Design model. (133) A case such as NML v. Argentina should motivate a revision of the contracts under both the Rational Design and Random Mutation models, but the character of the revisions should be dramatically different in each case. The clearest differential prediction from the two models is that revisions should result in the same amount of variation across issuers as prior to the court decision shock if the Rational Design story is true--after all, variation is optimal, rational and desired. In contrast, the revisions postshock should result in a reduction in variation under the Random Mutation model, according to which variation is essentially random, unintended and creates unnecessary risk. Figure 5 provides a graphical depiction of the frequency of different sovereign offerings by type of pari passu clause. After the October 2014 meetings that led to collectivizing consensus in the marketplace, we see a dramatic reduction from 2015 to 2016 in the variation in the types of clauses in response to NML v. Argentina, (134) Prior to the case, there was a range of pari passu clauses of different risk levels. (135) After the case, all of the changes that take place push towards a single form of the clause--the low risk Rank version which is also the original version of pari passu. This movement to eliminate the variation is consistent with the Random Mutation model and not with the Rational Design model. (136) Interestingly, the movement of these revisions toward the oldest variant of pari passu does not eliminate the risk of holdouts entirely, as the adoption of a Mandatory Law clause would do (see our categorization in Table I). Instead these market-wide effects appear to reflect a conscious decision in the market to return to the original version of pari passu.

In sum, as of June 2016, the data on bonds issued by the pure sovereigns under New York and English law indicate the vast majority of new issuances are moving to the low risk Rank form of the clause and explicitly rejecting the ratable payments interpretation. And, invariably, this new language is the formulation prescribed by ICMA and endorsed by the IMF's report, thus highlighting the importance of coordinating forces in spurring the movement in the market to the original version of pari passu. (137)

In our empirical paper, we also find that the following indicators correlate with the likelihood of a sovereign modifying the pari passu clause earlier rather than later: (a) representation by a law firm that is a market leader in terms of the volume of clients it represents; (138) (b) location in Latin America, a region whose New York lawyers constituted the bulk of the participants at the two coordination meetings held in October 2014; (139) and (c) whether the issuer had defaulted and restructured sometime over the past ten years. (140) These factors, as we explain in Part IV, point towards the reasons for resistance to change being a combination of difficulties in coordinating change and the agency costs of lawyers and government debt managers being focused on the short-term goal of getting deals done. (141)

E. Interviews with Market Players

The data--showing extraordinary resistance to changes in pari passu language for the first three years of the study followed by a dramatic shift toward uniformity in pure sovereign issuances that is accelerating to the present day--reveal patterns of behavior that are confounding. To better understand this phenomenon, we turned first to the transactional lawyers who produced the contract terms and disclosure documents for the bonds in our data. As noted above, we hosted a conference at Columbia University Law School in early October 2014, shortly before the Fall IMF/World Bank meetings. We asked the gathered experts (142) if they might help answer two core questions: First, what explains the failure of the sovereign debt industry to modify or amend pari passu now that multiple New York courts have handed down a ratable payments interpretation? And, second, what was the best strategy going forward to combat this problematic interpretation?

We recorded what was said at the meeting and then had individual conversations or email exchanges with each of the attending lawyers, not all of whom responded. (143) One theme was expressed by a number of the lawyers, and echoed in the conversations after the October 2014 meeting. Sovereign clients were not yet comfortable making changes to their clauses. To pursue this theme, one of us then conducted a series of interviews with sovereign clients through senior officials at government debt offices. In Part III.E, we report on conversations with these government debt managers. Finally, we spoke to senior managers at fifteen investment firms that had specialties investing in government debt. (144) Taken together, we spoke to over a hundred different individuals actively involved in the sovereign debt markets.

1. Reasons for the Lack of Revision as of October 2014. When interviewed during 2008-2011, these same lawyers gave two primary reasons for not modifying or amending their clauses in response to the ratable payments interpretation from Brussels in 2000. First, ratable payments was a bizarre interpretation from an inexpert court in Brussels made in the context of an ex parte hearing, and no New York or London court would rule in such a fashion. Second, modifying the clause would send a negative signal to any future court faced with the same interpretive issue: it would imply that the original boilerplate was ambiguous and thus make it more susceptible to the ratable payments meaning.

In October 2014, neither of the prior explanations for inaction seemed valid. A New York federal district court had affirmed a ratable payments interpretation that twice was endorsed unanimously by the leading appellate court on business matters, followed by a denial of certiorari by the U.S. Supreme Court despite arguments to the court in favor of reversal by four different nations, a Nobel laureate in economics, and a former Deputy Director of the IMF. (145) So, why had the clause still not been modified?

a. The Perfect Storm. The initial response from the New York lawyers was that the New York interpretation was a function of unusual facts. "A perfect storm," stated one senior attorney. (146) Argentina had apparently angered SDNY Federal District Judge Thomas Griesa to an unusual degree by the lack of respect it had shown him and his rulings over the decade or so that the litigation had proceeded. (147) Argentina also made the strategic error of passing an explicit law that made the holdout creditors functionally junior to everyone else. (148) But neither the trial nor appellate court clarified whether its decision was based on the unusual facts at issue--the law that Argentina passed, its unusually bad behavior, etc.--or on its finding that the explicit terms of the pari passu clause required Argentina to pay the holdouts ratably. (149) If it had not been before, it now was clear that the clause lacked any plausible shared meaning, and there was a substantial risk that a court would resolve the linguistic uncertainty in favor of the ratable payments interpretation. Those two facts argued for repairing the clause promptly, regardless of any complaints about what the courts had done. That did not happen. Instead, these same lawyers devoted considerable efforts over the next two years pursuing every available avenue to have the ratable payments interpretation overturned in court.

Our conclusion from the foregoing is that the lawyers across the industry--who worked diligently to generate support for numerous amicus briefs--must have believed it would be more difficult to persuade the industry to modify the boilerplate than to organize industry efforts to persuade the appellate courts to reverse the trial court. And the difficulty the proponents had in generating the industry-wide litigation position opposing NML v. Argentina implies that it must be even more difficult to change certain boilerplate contract terms. But why?

b. British Courts Are Better. The English lawyers had a different response from their U.S. counterparts. They claimed that no British court would ever rule in the same narrowly textualist fashion as the New York courts. (150) Their confidence that, unlike the U.S. courts, a British court would look to market practice to rule out a ratable payments interpretation struck us as unwarranted, especially because there was no clear market understanding of the clause's meaning. A British court had been asked to interpret pari passu in Kensington International Ltd. v. Republic of the Congo but elected to sidestep the interpretive question. (151) And, more recently, a former President of the U.K. Supreme Court opined on behalf of the holdout creditors that Judge Griesa's ratable payments interpretation was correct under English contract law. (152) Nevertheless, the British lawyers were unmoved: This episode was the result of the flaws in U.S. contract law, and many predicted that sovereign issuers, unhappy with the U.S. courts' failure to examine market understandings, would now issue their bonds under English law rather than New York law. (153)

Our data actually show that since October 2014, lawyers at the big British law firms have been altering the pari passu clauses in pure sovereign issuances governed by English law at roughly the same rate as their U.S. counterparts. We asked a subset of the British lawyers about the apparent inconsistency between their current actions and their prior assertions. They explained that despite the trivial risk of an incorrect interpretation by a British court, revisions were necessary because the standard documentation package for sovereigns issuing under foreign laws has now been changed. And clients want the standard package of clauses that is appropriate for their type of issuer.

c. Our Clients Are Uncertain What to Do. At the public discussion at Columbia Law School in October 2014, one participating lawyer raised the matter of client preferences. This lawyer, an industry group representative rather than a senior law firm partner, said quietly that perhaps one reason the clauses were not being modified was that some of the clients were not as confident as the lawyers in the room that the New York courts' interpretation should be disavowed. A number of the participants agreed with this evaluation of their clients' interests. In effect, we were told: "We haven't been able to modify the clause because at least some of our clients are not convinced that is the best thing to do." The inference was that while they agreed with the IMF and others that the courts' decisions were incorrect, (154) clients had not seen any strong negative reaction from the market to the pari passu decisions in NML v. Argentina, other than for Argentina itself. (155)

Lee Buchheit, the dean of sovereign debt lawyers, explained why lawyers had not been able simply to delete the pari passu clause after NML v. Argentina:
   The principal drawback of a textual amputation of the clause is
   optical; it will leave a gaping hole in the term sheets, rating
   agency reports and tick-the-box summaries of the features of new
   debt issuances. Never having had a clear idea of what purpose the
   pari passu clause actually served in a cross-border debt
   instrument, underwriters and most investors will surely not have a
   clear idea of the implications of not having it. The guiding
   principle of the underwriting community in matters of documentation
   has always been that if it was good enough for my father, it's good
   enough for me. Excising the clause altogether could therefore
   entail a significant educational initiative. (156)

d. The Past Is Irrelevant. Perhaps unsurprisingly given the drama caused by the pari passu litigation, a number of scholars have embarked on investigations into the original meaning of the clause in sovereign instruments. (157) Both Sung Hui Kim and Anna Gelpern, each examining different sets of archival records, (158) found evidence that the most likely understanding of the clause in a dispute over Nazi bonds in the 1930s was that it required ratable payments. (159) Ben Chabot's archival research into the infamous Black Eagle bonds issued by Mexico a century ago suggests that ratable payments was the most likely understanding of the clause at that time as well. (160)

Given that no one currently in the industry appeared to understand the meaning of the clause, we asked whether this historical evidence affected the thinking of contemporary lawyers. But the lawyers did not find the historical research to be relevant to the question of what the clause meant in the contemporary context.

One eminent lawyer responded:
   The context today is completely different. You are dredging up
   meanings from a period where there was absolute sovereign immunity
   and where gunboats were the primary means of enforcement. We don't
   have either of those things today. (161)

That response brought us back to the original puzzle: If history was irrelevant to unearthing the meaning of the clause and there was no real contemporary understanding, why was the clause being used at all?

2. The Courts Failed to Solve a Systemic Problem. Despite the often-puzzling responses to our questions, a coherent story does emerge from the interviews with the leading lawyers in the sovereign bond industry. All of these lawyers share the strong view held by the official sector (the IMF, the U.S. Treasury, Paris Club, and others) and by ICMA that by undermining the ability of a defaulting sovereign to restructure its debt, the ratable payments interpretation was systemically harmful to the global economy in general and to the market for sovereign bonds in particular. (162) Yet, the lawyers were reluctant to act on that belief because their commitment to the industry may have conflicted with the interests of certain clients. Taking their statements at face value, the U.S. lawyers viewed the risk of future courts following the "precedent" of NML v. Argentina as small owing to the "perfect storm" that engulfed that case. Similarly, the British lawyers held the view that British courts would rely more heavily on industry "experts" than on the textualist jurisprudence of the Second Circuit. To be sure, one could interpret those beliefs as simply a justification for inaction in the face of perceived conflicts among client interests. Given those conflicts, the lawyers all agree that by applying the reigning principles of contract interpretation, the courts in NML v. Argentina failed to prevent systemic risk to the industry and left individual law firms in a quandary. How can a standard boilerplate contract adequately represent the interests of diverse clients with very different interests?

But the preceding story only explains why no efforts to reject the ratable payments gloss on pari passu occurred between June 2011 and October 2014. Yet to be explained are two remaining puzzles: What then stimulated the dramatic increase in modified clauses in pure sovereign bonds issued shortly after October 2014? And why are the revisions that reject the ratable payment interpretation confined to pure sovereign issuances yet are virtually nonexistent in other bond categories, even though the same sovereigns whose clauses have been modified are often ultimately responsible for these other bond issues as well? Put differently, the risks posed by a ratable payments interpretation are the same or worse in those other bonds.

F. Reports from the Clients

In talking to the sovereign clients in twenty-seven different countries spanning three continents, ranging from AAA issuers to emerging market issuers, we focused on the managers at government debt offices. These debt offices are the primary clients in a sovereign bond deal, even though, in theory, investment banks are on the other side of the transaction. The lawyers for the investment banks on a sovereign deal tend to be what are called "designated underwriter's counsel": the sovereign debtor is the one who picks the counsel for the investment banks. (163) Bradley and coauthors report that while the investment banks that manage a deal for the sovereign issuer tend to change from issuance to issuance, the designated underwriter's counsel tends to stay the same over long periods of time. (164) The logic is that these lawyers develop over time a deep understanding of the debtor and can more easily prepare the relevant legal documents. (165) While in some cases the sovereign debtor will have a different law firm as its primary counsel, it is often the case that the designated underwriter's counsel is the only set of external lawyers for the deal. Further, it is the issuer who usually picks up the cost of the lawyers, whether they use two sets of lawyers or just one.

What we report below is drawn largely from interviews one of us conducted for a project asking how government debt managers think about the pricing of contract provisions in sovereign debt. (166) In particular, we asked the debt managers from the different debt offices two questions: First, what was their view of the risk posed by the ratable payments interpretation in NML v. Argentina for the type of clauses they were using? And, second, to the extent they had issued both sovereign and quasi-sovereign bonds, what were their reasons for using different types of pari passu clauses in these different issuances?

1. The Reasons That Might Induce Revision: Standard Practice and Investor Preference. The sovereign debt lawyers had told us that many clients from the government debt offices were unsure of the best response to the NML v. Argentina decisions. The subset of clients interviewed uniformly confirmed that view. In every case, the debt managers were aware of the litigation by the hedge funds and that payments to the creditors holding restructured bonds had been frozen as a result. But in no case did these debt managers--over sixty senior officials across the various debt offices--seem to have a strong view of what they should do in response. We heard few statements of dismay at what the New York courts generally had done or Judge Griesa specifically had done. (167) And relatedly, there was no expressed desire to avoid New York law or New York courts in future sovereign instruments.

The debt managers for the most part thought that the outcome of the NML v. Argentina litigation was bad. But they did not seem particularly concerned that the meaning of a key provision in many of their own bonds had been challenged. The explanation for not immediately changing their clauses was largely consistent across the debt offices: They would change their pari passu clauses when and if their lawyers told them that the standard boilerplate formulations were changing. A number of the debt managers emphasized the importance of having a lawyer who was in constant contact with the IMF legal department and was current on the latest improvements in the standard forms for sovereign bonds. Their preferences appeared to be driven by what the IMF considered to be "good practice." But other than changes to the standard forms that were emanating from the IMF, the primary audience they were serving were the investors who were repeat purchasers of their bonds--specifically the dealers or bankers doing the initial placements. And those investors were not clamoring, or even murmuring, that they wanted the clauses changed one way or the other.

We pressed: What if the investors wanted a version of the clause that was even more susceptible to the ratable payments interpretation? Wouldn't there be a negotiation over how many basis points that would cost? In their view, these questions revealed a fundamental misunderstanding on our part of the way prices for bonds were set and the relevance of contract terms such as pari passu to prices. In fact, there were no pricing decisions based on the "legal terms" of the document (the terms that the lawyers draft) in contrast to the "business terms" that impact the bonds' selling price.

Legal terms may need to be changed because the relevant standard template has changed. But that, as one senior manager told us, is to ensure that the legal terms do not impact the price. (168) There are terms that are relevant to the price--such as the currency, maturity, and governing law, the material that appears on the term sheet (169)--but legal terms such as pari passu and negative pledge are not part of that equation. The job of the lawyer, one of the first debt managers we interviewed explained, was to make sure that he or she was "irrelevant to the transaction." (170) The managers and their investors want to be able to set the price as a function of the "stuff that matters"--the economics of the sovereign debtor, not whether an obscure clause contained the word "payment." (171)

Undeterred, we asked: "But even if the investors are not sure of what they want and don't care about the pricing of the legal terms, don't you want to draft your clauses so that you can ensure a smooth restructuring in the event that you ever go into default?" We were told that what we were suggesting was simply not done. The job of the lawyers was not to draft terms to protect the sovereign in case an adverse event occurred in the future. Instead, the lawyer's sole task was to ensure that the right standard template was used. If the form is standard for a particular category of issuer, then neither side has to worry about the legal terms: both parties know that they have the standard package and neither side is trying to take advantage of the other in terms of the contract provisions. Pricing then can be done independently of the legal terms, one respondent explained, when neither side is worried about the lawyers inserting terms to help their clients deal with a future event such as a default.

a. The Variation in the Clauses. Given that the debt managers were emphasizing the importance of using "the standard" provisions, we next asked them what each meant by "standard." We knew that standard did not mean identical, since there was considerable variation across issuers in terms of the precise wording of the clauses. But in addition--and here is where we focused our questions--many sovereigns had variations in the types of pari passu clauses they were using in their own bonds. Standard, to reiterate, didn't seem particularly standard. So why the variation? Was it because the investors in one type of bond had made a different bargain with the debtors than in another bond?

Our starting premise was wrong, we were told. There was no bargaining between the issuer and the creditors over the type of pari passu clause that would be used. Instead, the clauses were the ones that were standard for that type of issuance. The standard template for pure sovereign issuances under New York and English law was changing and that was why the sovereign bond contracts under those laws were changing accordingly. The templates for the other types of issuance, such as sovereign guaranteed bonds or local law governed sovereign bonds, had not been addressed by the IMF/ICMA initiative, so those bonds were not changing. Moreover, the debt managers explained further, primary responsibility for anything but the pure sovereign issuances lay with a different set of bureaucrats and lawyers. (172) To be sure, all of these obligations--and there are many hundreds of them in the period--would probably end up on the sovereign's balance sheet if there were a sovereign default. (173) But amending or modifying the clauses in those other bonds was not something that concerned the debt managers for the sovereigns, and it was not even something that they planned to discuss with their colleagues who did the other types of issuances. (174)

None of the foregoing is meant to suggest that the managers were disinterested in learning about the clauses in their own bonds and what the variations were in the clauses within their countries' issuances. They were curious about the implications of having one variation in a clause versus another, but nothing they learned about the different levels of risk appeared to motivate them to remedy their clauses: that was just not how the process of changing standard boilerplate worked. They did, however, care about the IMF's strong support for the new market standards and about being seen as good global citizens from the IMF's perspective. In this vein, every one of the debt offices where we talked to managers who had not yet changed the clauses in their foreign law bonds told us that they would be changing soon. Their investors would want to have standard clauses, and they did not want to be on the IMF's list of nonresponsive countries. (175)

b. No Midstream Changes. The final question posed to the debt managers concerned the wisdom of changing the pari passu provisions in their older bonds, given that their plan was to revise the pari passu clauses in their future bonds. In theory, this action would be favored by both issuers and their creditors assuming both feared that the older pari passu clauses could be exploited in the future by holdout creditors. If so, then why not change the clauses with a vote of the creditors during the good times when there were no holdout creditors: a vote between 50 percent and 75 percent of the bondholders could have deleted or modified the pari passu clauses in every one of the bonds at issue.

Few of the debt managers seemed to have given this question much consideration. (176) They were willing to consider changing new bond issues because the IMF, ICMA, and their outside counsel were telling them that the standard forms were changing. Moreover, this change was important for systemic reasons and they were willing to be good global citizens. But the IMF was asking them to revise the terms of their old bonds as well, and this was not something they were willing to do. Those changes would cost money that neither the sovereign debt managers nor their investors were willing to pay. (177)

2. Interviews with Fund Managers and Bankers. We learned through the interviews with government debt managers that they looked primarily to their investor bases to discern whether there was a demand for changing boilerplate language. Thus, for our final set of interviews, we talked to fifteen of the financial firms that purchase, and sometimes litigate, sovereign debt contracts. (178) To the extent these firms are holding sovereign bonds at the time of a crisis--and the firms held bonds in every recent sovereign debt crisis, including Argentina, Greece, Ukraine and Cyprus--they could see firsthand how the legal terms of the contract can matter a great deal once default looms. Our starting premise was that these firms were making decisions about which bonds to buy and sell--and, at least indirectly, what messages to convey to the debt managers--as a function of the contract provisions in the bonds. (179)

As a general matter, the investors indicated that they did not consult their in-house lawyers for advice on which bonds to buy as a function of the legal terms. (180) Legal terms, the investors said, are not relevant until very late in the game when the sovereign is trying desperately to avoid default. (181) It is only when those efforts fail that the lawyers are asked to determine what the legal terms mean and whether they might provide an advantage or disadvantage in restructuring negotiations. (182) This practice seemed inconsistent with rational investment strategy. If contract terms such as the form of the pari passu clause mattered ex post, in a near default state, ceteris paribus, they necessarily should matter ex ante when the bonds are issued as well. (183) Many of the players at the investment firms had quantitative backgrounds and so they understood the puzzle precisely. The explanation offered was the market reality. No one paid attention to anything but whether the bond had the "standard documentation"; that is, not until the very end, when everyone scrambled to find a good lawyer to tell them what their documents mean. (184) In short, what matters to the investors at the front end of the transaction is that the boilerplate legal terms are standard. What is not relevant is how the legal terms in that standard might affect the price of the bond once the sovereign debtor approaches default. Consequently, it does not matter what precisely constitutes the standard legal terms.


We have been gathering data on the pari passu puzzle for more than a decade. At first glance, the empirical evidence collected--both quantitative and qualitative--seems puzzling and even confounding. Contract theory predicts that contract drafters will revise standard contract terms when faced with an interpretation adverse to their clients' interests. (185) That no more than a handful of corrective changes were made to the pari passu boilerplate for over three years after a federal court in New York endorsed the ratable payments interpretation, and roughly fourteen years after the Brussels decision, is perplexing. This is especially true because the drafting lawyers, and the entire sovereign bond industry, were nearly unanimous in condemning the series of judicial decisions that permitted the holdout creditors to prevail. Moreover, when clarifying revisions began to appear, they seemed to be prompted by two conferences held in New York in October 2014, rather than by the succession of adverse court decisions by the leading commercial courts in the United States. Adding to the mystery, the clarifying revisions that began to appear in many bond issues in late 2014 were limited to bonds issued directly by a sovereign. But virtually no such changes have yet been made to pari passu clauses in bonds issued by sub-sovereigns even where those bonds are guaranteed by the same sovereigns whose own pari passu clauses have been repaired. (186)

Notwithstanding the apparent contradictions, however, a story does emerge from the data. This is not the only story that explains the pari passu puzzle, but it is the only explanation that fits all the data sources collected. What follows, then, is first an effort to connect the dots by integrating both the quantitative and qualitative data. Thereafter, we address the question with which we began: What interpretive rule would better motivate courts to avoid systemic costs when asked to determine the legal meaning in encrusted boilerplate like the pari passu clause?

A. More on Collective Action with a Heavy Dose of Agency Costs

The evidentiary vacuum accompanying a black hole will deter contracting parties from simply unilaterally changing the language of the black hole after an aberrant court interpretation. First movers are unable to rely on contextual cues to discern the true meaning of the black hole clause, and the resulting uncertainty creates the risk of idiosyncrasy in a world that depends on standardization. To take just one example, where one party moves unilaterally but others continue to use the same term, the moving party faces a heightened risk that subsequent courts will view the change as confirming the first court's interpretation. At the same time, the first mover also faces uncertainty about how the market will interpret the meaning of the revised clause. These inertia costs undermine the incentives to change the actual language of the clause until market participants are able collectively to send a signal rejecting the aberrant court interpretation and adopting a new market standard.

1. The Coordination Problem: Private Versus Collective Interests. But how can parties in the market coordinate to act collectively? In certain markets, the coordination problem is made worse by the perverse incentives of critical market participants that hinder the market's ability to overcome the collective action constraint. In the case of the sovereign debt market, the private interests of key participants--the elite lawyers and their clients, the representatives of the sovereigns and the investors--are inconsistent with those parties' collective interests. The collective interests of the lawyers are to protect the industry and the market for sovereign bonds so that future issuances proceed smoothly and future business can grow. Yet, the private interests of the same set of lawyers are the mirror image of the interests of their de facto clients, the debt managers and the investment banks. And both sets of agents are subject to hyperbolic discounting: they are motivated to reduce the ex ante costs of a bond issue even where expected ex post costs are thereby increased by an even greater amount. (187) This hyperbolic discounting problem exists whenever contracting parties pay more attention to the ex ante or front-end costs of the contract than to the ex post or back-end costs of a later default. (188)

This, then, presents a problem that seems unique to boilerplate that has become a black hole. In the case of boilerplate that has been litigated or otherwise updated over time, for example by an industry group, and thus has a recognized contextual meaning, coordination is more tractable. The extant interpretation economizes on transaction costs by providing a focal point that aligns the parties' expectations and thus permits them to solve the coordination problem more efficiently. (189) Parties who participate in the sovereign debt market are involved in a mixed motive game. They coordinate on certain expectations but have conflicting interests on others. One way they align their expectations is through shared meaning. When the parties can communicate a shared meaning, experiments show that their "cheap talk" facilitates coordination. (190) As Schelling famously noted, when the problem is selecting one means of coordinating among many, focal point solutions stand out and attract the attention of both parties. (191) Once announced, the focal point default economizes on costly precontractual communications: this function is especially valuable when the parties have different possible ways to coordinate, and there is no consensus on how to do so.

This means of coordination is absent in the case of contractual black holes because any attempt at a collective revision to the clause first has to be "settled" among the key players, since those same private interests demand "standard" legal terms that reduce ex ante costs. (192) The absence of any focal point of meaning conflicts with the high demand for standard terms. If the term in question is one that is on a proverbial checklist, then it is the type of term that an investor expects to see in a sovereign bond. (193)

This collective action problem is exacerbated by the agency costs that seem to pervade the sovereign bond market (194) and that explain the apparent inconsistency between the expressions of distress over the pari passu litigation by the elite bar and their concomitant unwillingness or inability to effect any change in the standard boilerplate language. It also explains why there finally was substantial movement to revise the pure sovereign issuances in late fall 2014 and why this apparently coordinated decision to revise pari passu was not followed in the quasi-sovereign bonds that were issued during the same time frame. In what follows, we support the collective action/agency cost story by describing in more detail the respective individual and collective interests of each of the principal parties.

2. The Incentives of the Elite Lawyers. The private interest of each of the lawyers who dominate the sovereign bond market is to process bond issues at the least ex ante cost and as quickly as possible, notwithstanding expected default costs. (195) This single-minded focus on reducing front-end contracting costs is simply a reflection of the fact that the "legal terms," for which the lawyers are responsible and that form the standard boilerplate, are seen as immaterial to both sellers and buyers in the initial pricing of the bonds. Thus, any change in the risk of default that results from a change in the legal terms of the contract is ignored by both the debt managers, who act as agent for the sovereign, and the investment bank, which serves as agent for the investors. In short, the ex ante legal meaning of pari passu is irrelevant to both sides of the transaction as is the fact that this particular clause has no understood ex ante meaning: it is a black hole with random variations that are meaningless encrustations. This makes pari passu a much more difficult problem to repair once the inertia costs of an aberrant interpretation become salient. Nevertheless, the pari passu clause remains a part of the bonds' contractual boilerplate because it is part of the "standard form," and standardization is valued because it reduces ex ante contracting costs. (196)

When Elliott Associates succeeded in having a court adopt the ratable payments interpretation in Brussels in 2000, the elite bar was outraged but not because any lawyer's individual interests were imperiled. As we know, their clients did not care. The lawyers were outraged, in unison with the public sector and other collective groups, because they saw that the ratable payment interpretation imperiled the health of the industry itself. If bonds in default cannot be restructured, then over time the pressure from the vulture funds whenever a sovereign faces default will reduce the demand for issuing debt in this form and the robust market for sovereign bonds will be negatively impacted. This means a decline in a lucrative legal business. The collective interests of the lawyers who dominate this industry is to maintain a thriving sovereign bond market where bond issues are produced on an assembly line. (197) And this way of doing business was threatened.

At the same time, these lawyers had no incentive to revise the standard terms for their individual clients. The debt managers for the sovereigns do not care about the legal terms at the time of issuance: they do not regard the legal terms as relevant to the initial pricing of their bonds because they know that the investment banks charged with marketing the bonds only care about having the standard form. Moreover, the clients affirmatively discourage individualized deviations from the boilerplate formulation because nonstandard legal terms make the initial issuance more difficult and costly to get to market. Thus, the lawyers repeatedly demanded that the state solve the problem but did nothing themselves other than to offer empty platitudes about why they failed to act. (198)

This saga continued for nearly 15 years until one lawyer at the October 2014 Columbia conference committed a gaffe by telling the truth: "We don't know how to respond to this problem because the interests of our clients are not identical, and many clients do not ask for or want any change in the standard legal terms." (199) The evidence points to the fact that this session, and the subsequent meeting of an elite subset of the same basic group a few weeks later, was the impetus for coordinating a move to a revised (though equally ambiguous) pari passu clause. The lingering agency problem for the elite bar is that the de jure client is the "true sovereign"--the people or at least the duly constituted government--and not just the debt managers. And it is not at all clear that those interests are advanced by treating all sovereigns as having the same default risk and issuing standardized boilerplate for sovereign bonds despite the apparent variance in the probability of a future restructuring between developed countries and developing nations.

3. The Sovereigns' Incentives. The sovereign's interests are also skewed by an agency problem. Sovereigns have a long-term interest in having the capacity to restructure their debt. But because they have not incentivized their agents correctly, those interests are underrepresented in the state's bureaucracy. Thus, the debt managers care primarily about what the investors claim to care about--the business terms that they believe do influence bond prices--and they want the legal terms to remain unchanged and uncontroversial so as to secure the best initial price at the lowest issuance cost. (200) In short, the sovereign's agents engage in hyperbolic discounting because that is what they are paid to do. Nevertheless, theory predicts that some (many) sovereigns will have issued bonds prior to the revisions to pari passu that risk imposing substantial costs on their country's citizens in the future.

4. The Investors' Incentives. But why don't the investors who buy the bonds care about the ability to restructure in the event of default? This is a difficult question, but one hypothesis is that it is too costly to try and match a given sovereign with the optimal clause. Some sovereigns may present a measurable default risk while others may not, and the information to make particularized ex ante calculations is costly to acquire, especially in a world of encrusted boilerplate of uncertain meaning. The same holds for the information needed to quantify how changes in pari passu will alter the present value of future repayment if default occurs. A rationalist skeptic might ask: Why is the market so imperfect? Behavioral theory may explain part of the answer as being a function of excessive discounting. (201) Another consideration is the fact that this is a liquid market where bonds can easily be resold on the secondary market. In such an environment, the business terms and the bond's rating are a good enough proxy for future default risks especially where boilerplate terms have been stripped over time of comprehensible legal consequences.

But the preceding does not answer one remaining question: If Jay Newman, Elliott's legendary legal arbitrageur, prides himself on reading the bond contracts once default looms in order to capture rents as a holdout creditor, then doesn't it follow that he and other hedge fund hotshots will read the bond contracts at the time of issuance as well? (202) This implies that there should be arbitrage in the primary market where smart investors selectively buy bonds based on their reading of the legal terms: even if the initial purchasers do not plan to be there when default looms, they know that Jay Newman and others will pay a higher price for the bonds with better contract terms in that near-default scenario. (203) To be sure, investors act as though the risk of a default without the ability to restructure is too remote to affect initial price. But will the market adjust if hedge funds engage in arbitrage ex ante as well as ex post? Perhaps not. The tradeoff between the ex ante moral hazard risk of a future restructuring and the ex post returns of a successful restructuring to creditors is difficult to assess. So long as the initial investors only bear some of the price risk caused by legal arbitrage, it still may be rational for them to buy bonds without discriminating among legal terms that influence the costs of default.

5. The Role of the Public Sector and Industry Associations. The IMF, the Paris Club, the Institute of International Finance, ICMA, and so on only have collective interests. Why was it so hard to coordinate with the leading lawyers to solve the problem much earlier? The best inference from the data is that the pari passu clause was emptied of any context that could help determine its most plausible meaning, thus increasing the risks of error to actors who were motivated to revise their boilerplate. What should they do without any context that could point to a plausible meaning to use as a benchmark? (204) Under these conditions, the collective interests believed that the expected costs of litigation in the Second Circuit and the Supreme Court (discounted by the probability of prevailing) were lower than the costs of coordination given that the elite members of the bar were individually reluctant to do anything. Thus, so long as the courts could be expected to get it right in the end, coordination costs were too onerous.

This then leaves the last part of the story. How did they get it done? Recall that solving the collective action problem not only required a willingness to consider the collective interest over private interests but the parties had to coordinate around a common formulation in order to substitute a new standard term in place of the defective one. At the Columbia conference in October 2014, many of the elite lawyers were openly critical of the clarification proposed by the IMF-led process. Yet, once they were invited to a second meeting limited to the most elite among them and asked to participate in solving a common problem of global significance, the attendees quickly and easily reached an agreement and revisions began to appear from that point onward. Is it really true that status and flattery were all that was needed to do the trick? Was the Columbia conference an important first step for the elite bar to recognize that they all shared the same private and collective interests? (205) Our data cannot answer these last questions but we do know that the lawyers who were in the room at both the Columbia and New York Federal Reserve meetings in October 2014 represented the pure sovereign issuers doing offerings under New York and English law. (206) And that is precisely where the standard revision has now become the norm, whereas other lawyers who were not invited to the meeting with the IMF--often partners at the same law firms as the elite cohort--represent subsidiary sovereign interests that have yet to coordinate around a revision that rejects the ratable payments gloss.

B. What Should Courts Do with Contractual Black Holes?

The standard interpretive principle courts are instructed to use in ascertaining the meaning of a contract term to which both parties have manifested assent is to look for the shared intent of the contracting parties. (207) Intent, in turn, is determined both objectively and prospectively: A party is taken to mean what its contract partner could plausibly believe it meant when the parties contracted. (208) Textualist theories undergird the New York courts' doctrines of contractual interpretation that are designed to uncover the objectively reasonable, ex ante intent of the parties. Textualist jurisdictions, such as New York, use a "hard" parol evidence rule that gives presumptively conclusive effect to merger or integration clauses, (209) and, in the same spirit, this approach bars context evidence suggesting that parties intended to impart nonstandard meaning to language that, read alone, is unambiguous. (210)

There is a powerful justification for giving boilerplate terms in commercial contracts their plain or standard dictionary meaning: creating standard vocabularies for the conduct of commercial transactions is a valuable state function. (211) When a phrase has a set, easily discoverable meaning, parties who use it will know what the phrase requires of them and what courts will say the phrase requires. By insulating the standard meaning of terms from deviant interpretations, this strategy preserves a valuable collective good, namely a set of terms with clear, unambiguous meanings that are already understood by the vast majority of commercial parties. (212)

But the preceding exposes a dilemma that courts confront when applying a plain meaning analysis to standard boilerplate such as the pari passu clause: the interpreter must somehow distinguish between meaningful language and empty boilerplate. Moreover, this problem is not solved simply by arguing that courts should instead adopt a contextualist interpretive style. (213) Contractual black holes are acontextual: the variations in language among different version of the clause are essentially random and thus context evidence does not aid in determining plausible meanings. This suggests that both textualist and contextualist courts are well advised to interpose a further step when interpreting standardized terms in commercial contracts.

At a minimum, courts should be open to arguments that, as a matter of law, the clause in question has been emptied of meaning and functions as a black hole in the boilerplate. To be sure, the moral hazard risk of false claims by a party who has been disadvantaged by fate argues for a strong presumption against the existence of a black hole. But there should be an opportunity to admit evidence sufficient to overcome that initial presumption against encrustation. The evidence described above is illustrative of the proof that the parties alleging encrustation might proffer. Has the clause been repeated by rote over many years, without having been tested in litigation, where repetition has robbed the term of any obvious conventional meaning? Has the term been embedded in layers of legal jargon such that its intelligibility is substantially reduced and variations in the formulation of the term across contracts have no apparent significance? (214) Is a historic or original meaning of the term accessible in a fashion that makes sense in the contemporary context and are contemporary commercial actors aware of that meaning? Is there credible evidence that the particular provision was priced at the original issue stage? (215)

If a court finds strong evidence of rote usage and encrustation in standard boilerplate, the presumption of shared meaning is no longer apt: when encrusted boilerplate is repeated by rote for many years without legal challenge, and no party has reason to know a different meaning attached to the clause by another party, the clause has become a legal black hole. What we have learned about black holes from our study of the pari passu litigation and its aftermath is that third parties can play a constructive role in facilitating coordination by the market on a new standard term. A court intent on facilitating coordination by groups such as the IMF could adopt the presumption that the evidence of a black hole implies that the parties attached different meanings to the term in question. This presumption invokes the common law rule that if parties have attached different meanings to a term neither party is bound by the meaning of the other unless at the time of contracting one party did not know or have reason to know the meaning of the counter party, who in turn did know or have reason to know the meaning of the first party. (216) Applying this principle to the NML v. Argentina case, and assuming neither party knew or had reason to know of the other's different ex ante understanding, a court could find that neither party's interpretation of pari passu was legally relevant. Reading the clause out of the contract in this way permits textualist courts to maintain their commitment to plain language interpretation and also allows contextualist courts to continue to invite extrinsic evidence of the meaning of contested terms.

To be sure, the state could advance the parties' interest in solving their collective action problems in other ways, say, by engaging in just the sort of "public/private" coordination efforts that ultimately succeeded in resolving the pari passu saga. In general, regulatory solutions of that sort are preferable to a court's resolution of the black hole conundrum because they operate generally and not just in the particular case. For this reason, among others, the true lesson of our study may be that the IMF and other groups that constitute the "official sector" may be better able than courts to solve these problems over time as they gain experience and become more confident in their methods. (217) If this is so, then even though the pari passu case shows that the intervention of a public authority is sometimes required to solve contractual disputes that have third-party effects, relying on the courts rather than private ordering to craft the solution may not always be the best choice.


In this Article, we have sought to support three claims. First, contractual black holes can exist as a by-product of the standardization of boilerplate in commercial markets. The boilerplate production process can generate random variations in language that are not the product of rational contract design. Nonetheless, these variations persist and are cemented over time as part of the market standard. Second, while black holes often remain for many years as relatively harmless surplusage, they can generate substantial social costs once litigation results in an interpretation that introduces inefficiencies into the market. These encrustations then invite opportunistic litigation and require costly collective efforts to repair the now vulnerable terms. Third, those costs are a function of the inherently greater inertia costs that result from an aberrant interpretation of a black hole term and the greater difficulty market players face in overcoming the resulting collective action problem.

We have used the pari passu litigation as a prototypical exemplar of the substantial costs that result from inefficient interpretations of black holes, costs that are exacerbated when the interpretations are advanced by contractual arbitrageurs, such as Elliott Associates and NML Capital, Ltd. Indeed, the costs of pari passu in taxpayer payouts to holdout creditors are already enormous and likely to increase even further. (218) As we write this paper, Venezuela is on the verge of default on upwards of $75 billion of debt. (219) Among its bonds are those with different versions of the pari passu clause, including the low risk Rank version and the high risk Rank in Payment version. Comparisons of the price differentials among the bonds that isolate differences other than the contract provisions suggest that holdout activity is driving the widening spreads among the bonds, with an increasingly higher yield for the low risk Rank version that provides less opportunity for holdouts. (220) Here the resulting social cost is a function of the fact that, despite the revisions that are common in new bonds issued after October 2014, few if any sovereigns have sought to modify the terms of their older bonds.

A natural question to ask is whether our study is idiosyncratic. Perhaps the pari passu clause is a unique example of the costs of encrustation. We do not have a good answer to this question, other than to note the number of recent papers exploring similar problems in other standard markets (221) and that within the sovereign debt contract itself there are other terms that are potential black holes. An example described elsewhere is the standard negative pledge clause, a clause that is on the standard checklist, and appears in almost every sovereign bond. Yet, a negative pledge term seems to have little contemporary meaning since sovereigns stopped pledging assets as a backstop to their debt more than seventy-five years ago. (222) And even if they did pledge assets, what would it mean to have pledges that rank equally in a context where seizing a foreign sovereign's assets in its own country is impossible in the modern era?

Outside of the sovereign bond context we have seen suspected black holes similar to pari passu in local municipal bonds in the United States. A frequent practice with these municipal bonds, of which tens of billions of dollars are outstanding, is that they are backed by pledges of revenues of the local governments. (223) Sometimes the pledges of revenues are of the general tax revenues, and sometimes there are specific streams of revenues from a utility or similar entity. An issue that has become salient in the context of the Commonwealth of Puerto Rico, which has over $50 billion of these types of bonds at risk of default, is how these revenues can be enforced if the Commonwealth or some other such entity defaults. Would a federal or state judge be willing to issue an order asking the government in question to stop doing repairs on its roads or providing basic services to inhabitants in order to pay a contract arbitrageur? Alternatively, is there a risk that the court would order the local government to raise taxes to pay the creditors? And what if the inhabitants simply moved to avoid the taxes? These kinds of revenue pledges are ubiquitous in the U.S. municipal bond market--and there are many variations in how they are formulated, variations that are supposedly priced (224)--and yet, no one seems to know how they would work.

In short, black holes do exist in standard boilerplate contracts that are used all over the world to regulate important markets. And the principal lesson of the pari passu saga is that once a black hole is discovered and then exploited by a contractual arbitrageur, the social costs of coordinating a move to solve the problem collectively can be extremely high.


(1.) Alan Schwartz & Robert E. Scott, Contract Theory and the Limits of Contract Law, 113 YALE L.J. 541, 568-69 (2003). Intention is "determined objectively and prospectively: A party is taken to mean what" a "contract partner could plausibly believe it meant when the parties contracted." Id. at 569. As the arbiter of disputed interpretations, the court determines the meaning of whatever signals of intention the parties agreed to. While the court presumably knows what the default rules it implies in every contract mean, it does not know the intended meaning of those terms that were chosen by the parties.

(2.) The concept of the black hole derives from theoretical physics. Stephen Hawking's work on black holes suggested that no "information" can escape from a black hole--once it is pulled past the event horizon, it is lost. Horizon: The Hawking Paradox, BBC Two television broadcast (Sept. 15, 2005). In the sense we use the concept here, the parties' original understanding of what a clause meant can, in theory, be lost entirely by the process of repetition and the insertion of random variations: once drawn into a black hole, it is lost forever. In the last two years, however, Hawking has decided that he was wrong and while some information can escape, it is so degraded as to be virtually useless. See David Castelvecci, Physicists Split by Hawking Paper, 529 Nature 448, 448 (2016) (noting that Hawking's recent paper "suggest[s] a mechanism for transferring [some] information to the black hole"); Clara Moskowitz, Stephen Hawking Hasn't Solved the Black Hole Paradox Just Yet, SCI. AM. (Aug. 27, 2015), article/stephen-hawking-hasn-t-solved-the-black-hole-paradox-just-yel [] (quoting Hawking as staling "[t]he information about the ingoing particles is returned but in a chaotically useless form"). This, then, is the "grey hole" concept. See infra note 3 and text accompanying notes 18-42.

(3.) David Dyzenhaus has distinguished the different characteristics of legal black holes and grey holes. See generally David Dyzenhaus, The Constitution of Law: Legality in a Time of Emergency 41-42 (Cambridge Univ. Press 2006). In his terms, grey holes are situations where "there are some legal constraints ... but the constraints are so insubstanlial that they pretty well permit [an actor] to do as it pleases." Id. at 42. A black hole, in contrast, does not even pretend to constrain. It is a "lawless void." Id. at 41. Thus, in the sense we use here, the contractual black hole is the limiting case of a boilerplate term that lacks any meaning whatsoever. A grey hole, in contrast, cannot excuse all possible meanings, but the meaning that has survived is incapable of providing a basis for making legal distinctions among the variations in language that have appeared over lime. Id. at 42. Thus, even for grey holes, where some evidence on the meaning of the contractual term remains, this evidence may be so minimal or contradictory as to leave courts effectively with little guidance on how to apply this meaning in litigation. Since the interpretive issues are the same with both black and grey holes, we use the generic term "black hole" to refer to the range of problems caused by the loss of contextual meaning.

(4.) Douglas Baird, in analyzing this question, uses the term "skeumorph." See Douglas G. Baird, Pari Passu Clauses and the Skeumorph Problem in Contract Law, 67 Duke L.J. Online (forthcoming Oct. 2017).

(5.) Standardized terms provide a uniform, and therefore intelligible, system of communication. See, e.g., Lon Fuller, Consideration and Form. 41 colum. L. Rev. 799, 801-03 (1941) (describing the "channeling function" of legal formality). This standardizing "function is analogous to governmental provision of regulated standards of weights, measures and generic product names," and it offers contracting parties "similar communicative advantages." Charles J. Goetz & Robert E. Scott, The Limits of Expanded Choice: An Analysis of the Interactions Between Express and Implied Contract Terms, 73 Calif. L. Rev. 261, 286-88 (1985).

(6.) See Sharon Steel Corp. v. Chase Manhattan Bank, N.A., 691 F.2d 1039, 1048-51 (2d Cir. 1982), cert, denied, 460 U.S. 1012 (1983) (discussing the importance of boilerplate language in successor obligor clauses); Broad v. Rockwell Int'l Corp., 642 F.2d 929, 943 (5th Cir. 1980), cert, denied, 454 U.S. 965 (1981) ("A large degree of uniformity in the language of debenture indentures is essential to the effective functioning of the financial markets: uniformity of the indentures that govern competing debenture issues is what makes it possible meaningfully to compare one debenture issue with another...."); see also Howard S. Steel, Elnaz Zarrini & Arkady A. Goldinstein, NML Capital v. Argentina.' A Lesson in Indenture Interpretation, 8 Insolvency & Restructuring Int'l 31, 32 (2014) (discussing the uniformity in pari passu clauses).

(7.) For a description of black holes and the rote usage phenomenon, see infra text accompanying notes 19-32.

(8.) The eminent British lawyer, Philip R. Wood, has described the process of encrustation as akin to that of barnacles accumulating on a ship's hull. Philip Wood, Allen & Overy LLP, Life After Lehman: Changes in Market Practice 9 (2009).

(9.) Goetz & Scott, supra note 5, at 289; see infra text accompanying notes 19-32.

(10.) Mitu Gulati & Robert E. Scott, The Three and a Half Minute Transaction: Boilerplate and the Limits of Contractual Design 3, 51-52, 109-18 (Univ. of Chicago Press 2013) (describing the almost universal ignorance of market participants as to the purpose and meaning of (he pari passu clause in sovereign bond contracts).

(11.) For an English translation of the Brussels case, see Joint Appendix at A-1356, NML Capital, Ltd. V. Republic of Argentina, 699 F.3d 246 (2d Cir. 2012) (No. 12-105(L)).

(12.) NML Capital, Ltd. v. Republic of Argentina, No. 08 Civ. 6978, 2011 WL 9522565, at *2-3 (S.D.N.Y. Dec. 7, 2011).

(13.) NML Capital, Ltd. v. Republic of Argentina, 699 F.3d 246 (2d Cir. 2012).

(14.) NML Capital, Lid. v. Republic of Argentina, 727 F.3d 230 (2d Cir. 2013).

(15.) See generally Stephen J. Choi, Milu Gulati & Robert E. Scott, Variation in Boilerplate: Rational Design or Random Mutation, 19 Am. L. & Econ. Rev. (forthcoming 2017) (reporting on data concerning the use of pari passu clauses after June 2011).

(16.) In addition to the pari passu term, there are other terms in sovereign bonds themselves, such as the negative pledge clause, that may well have become black holes. One can also find such potential black holes elsewhere, such as the tax revenue pledges that underlie many municipal bonds in the United States. Insurance contracts appear to be another area with the potential for such terms. See Christopher C. French, Understanding Insurance Policies as Noncontracts: An Alternative Approach to Drafting and Construing These Unique Financial Instruments, 89 Temp. L. Rev. 535, 547-48 (2017) (describing conditions ripe for the generation of black holes). French explains:
   Many of the terms and conditions contained in standard form ISO
   [Insurance Service Office, Inc.] policies were drafted many years
   ago and are reused each time ISO issues new versions of the policy
   form. For example, the policy language in the 1943 New York
   Standard Fire Insurance Policy is still used today in some
   homeowners' insurance policies. Because much of the standard form
   policy language used today was drafted long ago, the original
   drafters are often dead or unknown. Documentation regarding the
   drafters' intent also rarely exists. Consequently, it is difficult,
   if not impossible, to discern the drafters' intent if the policy
   language is ambiguous.

Id. at 547-48.

(17.) A number of forthcoming papers, from an April 2017 conference on Contractual Black Holes, discuss the conditions under which black holes are more or less likely to occur. The papers are available at Contractual Black Holes, NYU Pollack Ctr. for Law & Bus., [https://].

(18.) The underlying canon of contract interpretation directs courts to give every term and clause in a contract a meaning, under the assumption that parties have drafted terms in a contract to convey their collective purposes. In NML Capital, Ltd. v. Republic of Argentina, the Second Circuit put it this way: "A contract should not be interpreted in such a way as would leave one of its provisions substantially without force or effect." NML Capital, Ltd., 699 F.3d at 258 (citing Singh v. Atakhanian, 818 N.Y.S.2d 524, 526 (N.Y. App. Div. 2006)).

(19.) The legal recognition of certain words and phrases preserves a reliable meaning to boilerplate terms. "Since the legal system retains ultimate power over interpretation and enforcement, parties cannot be certain what effect will be given to any formulation until it is tested." Goetz & Scott, supra note 5, at 278. "Definitional recognition does not change the optional character of these terms," but it "does confer upon them the status of 'invocations,' terms that once deliberately" included in a contract "have a legally circumscribed meaning that will be heavily--perhaps even irrebuttably--presumed." Id. at 282. What ultimately can result from rote usage is a term that is widely used because market participants assume it must have a meaning, but, in fact, it does not.

(20.) An example of a collective effort to update standard terms is the International Chamber of Commerce's publication of Incoterms, the international rules for the interpretation of trade terms. See Int'l Chamber of Commerce, Incoterms (2010). Each of the eleven Incoterms sets forth a number of substantive rules, including, most importantly, the point at which risk of loss passes from the seller to the buyer. Id. A contemporary example of updating comes from the International Swap Dealers Association (ISDA), whose rules, standard forms and definitions are continually revised by a standing committee. See Stephen J. Choi & Mitu Gulati, Contract as Statute, 104 Mich. L. Rev. 1129, 1139-42 (2006) (explaining the emergence and activities of ISDA); Anna Gelpern, Public Promises and Organizational Agendas, 51 Ariz. L. Rev. 57, 63-67 (2009) (describing ISDA's "contract-drafting and related activities").

(21.) For a discussion of the deterioration of the sealed instrument from a prima facie signal of legal enforcement to a meaningless rote incantation, see Robert E. Scott & Jody S. Kraus, Contract Law and Theory 141-43 (5th ed. 2013).

(22.) As Goetz and Scott explain:
   Ordinary social relations are similarly replete with phrases that
   have lost their literal meaning through rote usage. The person who
   asks 'How are you?' usually does not mean to inquire about your
   health. Similarly, 'Have a nice day,' an unusual expression of
   friendliness a few years ago has now lost most of its meaning and
   become a ritual salutation from personnel in many retail

Goetz & Scott, supra note 5, at 288 n.73.

(23.) Id. at 288.

(24.) Id.

(25.) Ross Bros. Constr. Co. v. State ex rel. Transp. Comm'n Highway Div., 650 P.2d 1080, 1082 (Or. Ct. App. 1982).

(26). See Gulati & Scott, supra note 10, at 73-118 (describing interviews with lawyers on both the issuer and investor sides regarding boilerplate drafting); cf. Anna Gelpern, Mitu Gulati & Jeromin Zettelmeyer, Why Are Sovereign Bond Contracts Sticky? A View From Government Debt Managers 8-27 (Duke Law Sch. Working Paper 2017), sol3/papers.cfm?abstract_id=2984293 [] (reporting on interviews with the debt managers and investors who hire the lawyers).

(27). Gulati & Scott, supra note 10, at 119-38 (describing this phenomenon in the context of the pari passu clause).

(28). Lisa R. Anderson & Charles A. Holt, Information Cascades in the Laboratory, 87 Am. Econ. Rev. 847, 847-48 (1997). We are grateful to Bert Huang for alerting us to this point.

(29). See Gulati & Scott, supra note 10, at 75-76 (quoting senior lawyers); Gelpern et al., supra note 26, at 11 (noting that debt managers ask lawyers to "paper the deal" by insisting that all standard legal provisions are present). Market participants report the importance of satisfying expectations of the "checklist" in other contexts relating to sovereign bond issuances as well; even if the item on the list is itself meaningless. See Elisabeth de Fontenay, Josefin Meyer & Mitu Gulati, The Sovereign Debt-Listing Puzzle 23-24 (Duke Law Sch. Working Paper 2017) [] (finding that "responses were virtually unanimous in describing the task of obtaining an exchange listing for an international bond issuance as a 'check-the-box' item").

(30). See Gulati & SCOTT, supra note 10, at 75-76.

(31). For the classic discussions of the Darwinian process that will eliminate harmful mutations through capital structure and loan covenant contexts, see Merton H. Miller, Debt and Taxes, in 1 Selected Works of Merton H. Miller: A Celebration of Markets 103 (Bruce D. Grundy ed., 2002) and Clifford W. Smith, Jr. & Jerold B. Warner, On Financial Contracting, 7 J. Fin. Econ. 117, 123 (1979), respectively.

(32). MILLER, supra note 31, at 103.

(33). Scholars in contract law have largely ignored the study of how contracts are produced--even though assumptions about contract production and revision underlie many doctrines of contract interpretation. But see generally Robert Anderson & Jeffrey Manns, The Inefficient Evolution of Merger Agreements, 85 Geo. Wash. L. Rev. 57 (2017) (exploring how certain contracts are created and evolve over time); Barak Richman, Contracts Meet Henry Ford, 40 Hofstra L. Rev. 77 (2011) (same); D. Gordon Smith & Brayden G. Smith, Contracts as Organizations, 51 Ariz. L. Rev. 1 (2009) (same); Kevin E. Davis, Interpreting Boilerplate (NYU Ctr. for Law, Econ., & Org., Working Paper No. 10-21, 2010), papers.cfm?abstract_id=1618925 [] (same).

(34). For a discussion of the inertia costs that apply generally to standardized contracting, see Marcel Kahan & Michael Klausner, Standardization and Innovation in Corporate Contracting (or "The Economics of Boilerplate"), 83 Va. L. Rev. 713, 728 (1997).

(35). We use the term "random variation" in a specific sense: the initial change in the language of the boilerplate term was likely an intentional emendation by the agent charged with the duly of adapting the contract to particular client needs. But the lack of the error-correction mechanisms that exist in the case of variation of well-understood terms means that the durability of the linguistic variation is essentially random.

(36). The metaphor here is the concept of punctuated evolution as popularized by paleontologist Stephen Jay Gould. The idea is that evolution does not select against characteristics that are harmless. The evolutionary process is triggered only by an event that makes the characteristic undesirable. See S.J. Gould & R.C. Lewontin, The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme,

205 Proc. Royal Soc'y London 581 passim (1979). See generally Stephen Jay Gould, The Structure of Evolutionary Theory (Harvard Univ. Press 2002). In the context of this paper, the pari passu clause is not ripe for revision until holdout creditors persuade the federal district court in New York to adopt an interpretation that blocks an agreement to restructure Argentina's debt.

(37). Learning costs are the costs parties must expend in learning the meaning of the clause. The prediction from the learning cost literature is that the older and more widely used a term becomes, the better is the common understanding of what it means. Kahan & Klausner, supra note 34, at 719-25, 730-33. See also Goetz & Scott, supra note 5, at 286-88 (explaining how standard terms provide benefits in terms of reducing errors of ambiguity, inconsistency, and incompleteness); Tina L. Stark, Introduction, in Negotiating and Drafting Contract Boilerplate 3, [section] 1.02 (Tina L. Stark, Lauren Reiter Brody & Frances Kulka Browne eds., 2003) (observing that provisions that have been used repeatedly develop a "hallowed status" and have been "blessed").

(38). See GULATI & Scott, supra note 10, at 40-41 (suggesting moral hazard incentives that motivate free riding by private parties on the assumption that the official sector will bear even greater costs if the terms are not revised). For further discussion of collective action problems, see infra Part IV.A.

(39). See infra Part IV (discussing the implications of boilerplate). This method of thinking about contract interpretation techniques is articulated in Richard A. Posner, The Law and Economics of Contract Interpretation, 83 Tex. L. Rev. 1581 (2005).

(40). This argument has been advanced in the case of the pari passu controversy by Marcel Kahan & Shmuel Leshem, Moral Hazard and Sovereign Debt: The Role of Contractual Ambiguity and Asymmetric Information (2017) (unpublished manuscript) (on file with the Duke Law Journal) and by Mark L. J. Wright, The Pari Passu Clause in Sovereign Bond Contracts: Evolution or Intelligent Design?, 40 Hofstra L. Rev. 103 (2011). In another article, we test the "Rational Design" hypothesis and find little or no support for the claim that the failure to respond to the courts' interpretation was a rational attempt to reduce moral hazard costs by precluding a subsequent restructuring. Choi et al., supra note 15, at 28-34.

(41). See, e.g., GULATI & SCOTT, supra note 10, at 33-44 (providing an overview of the theories explaining the prevalence of contract boilerplate).

(42). For a contrasting example of sovereign boilerplate that has been revised much more rapidly when impacted by the same external events, see generally Choi et al., supra note 15.

(43). If in fact inertia costs peculiar to black holes do contribute to the extraordinary delay, then an inefficient interpretation of such a clause will impose additional social costs on the market as parties continue to write contracts that fail to maximize the joint gains from contracting. In the case of the pari passu clause, the extraordinary inertia costs cited by market players as explanations for inaction can be grouped in four distinct categories: (1) Legacy Debt Costs: A change to the contract language that closes off holdouts in cases involving a newly issued bond might put the unrevised clauses in prior bonds of that sovereigns at greater risk of enabling holdouts, if the court draws a negative inference about the meaning of the old clauses from the fact that there was a substantial change in the new issues. (2) Market Reaction Uncertainty: There will be uncertainty as to how investors will react to a new formulation of a commonly used clause, especially where the clause lacks a well-understood meaning in the market. (3) Idiosyncrasy Costs: Investors have a preference for the "standard" package of terms. Changing a term poses the risk that the bond will be viewed as idiosyncratic, thus increasing learning costs. (4) Legal Uncertainty: How courts will interpret the new term remains an uncertainty until the term is tested in litigation: the uncertainty is greater if the lack of a sealed market understanding means that courts must interpret the clause in an an evidentiary vacuum. For a discussion of these concepts, see Gulati & Scott, supra note 10, at 73-108.

(44). We lake the standard sovereign bond contract that includes the ubiquitous pari passu clause to be prototypical, not representative, of boilerplate contracts of its type: Its elements are not those most frequently to be found in an empirical survey of standard form contracts with some encrusted terms. Rather, the standard contract contains the essential features that, in repeating boilerplate terms untested by litigation over long periods, define the category. In this sense, the sovereign bond contract is a prototype or central exemplar of a distinct class of contracts, in the way that robins and swallows are prototypes of birds, while chickens, ostriches and penguins, despite their many similarities to robins and swallows, are not.

(45). The following Part draws from Gulati & Scott, supra note 10.

(46). This part draws on Gulati & Scott, supra note 10, at 12-17.

(47). For a discussion of the pari passu variation used in the Peruvian debt contracts, see infra note 93 and accompanying text. The English translation of pari passu is "in equal step." The clause has been present in sovereign debt instruments since at least the early nineteenth century. See Benjamin Chabot & Mitu Gulati, Santa Anna and His Black Eagle: The Origins of Pari Passu?, 9 Cap. Mkts. L.J. 216, 216-17 (2014); see also Pablo Triana, The First Foreign Sovereign Pari Passu? It's All Scottish to Me (2015) (unpublished manuscript), [] (describing a similar clause drafted in the 1800s).

(48). See supra note 11.

(49). For discussions of the Brussels case and its aftermath, see, e.g., Robert A. Cohen, "Sometimes a Cigar is Just a Cigar": The Simple Story of Pari Passu, 40 Hofstra L. Rev. 11, 15-17 (2011) and Manuel Monteaguodo, Peru's Experience in Sovereign Debt Management and Litigation: Some Lessons for The Legal Approach to Sovereign Indebtedness, 73 Law & Contemp. Probs. 201, 208-10 (2010).

(50). E.g., Lee C. Buchheit & Jeremiah S. Pam, The Pari Passu Clause in Sovereign Debt Instruments, 53 Emory L.J. 869, 876 (2004); Rodrigo Olivares-Caminal, To Rank Pari Passu or Not to Rank Pari Passu: That is the Question in Sovereign Bonds after the Latest Episode of the Argentine Saga, 15 Law & Bus. Rev. Ams., 745, 769 (2009); Charles G. Berry, "Pari Passu" Means What Now? N.Y.L.J., Mar. 6, 2006, at 1.

(51). See, e.g., Patrick Wautelet, Vulture Funds, Sovereign Debtors and Creditors, in Insolvabilite des Etats et Dettes Souvarines 103, 142 (Matthias Audit ed., 2011) ("While it was unanimously decried by commentators, who expressed surprise at the interpretation accepted by the Court of Appeal, the decision of the Court of Appeal was seen as a blessing by the community of sovereign debt creditors which had until then not paid much attention to the pari passu clause." (footnote omitted) (emphasis omitted)).

(52). See Anne Krueger, First Deputy Managing Director, Int'l Monetary Fund, A New Approach to Sovereign Debt Restructuring (Nov. 26, 2011), ~kathrynd/IMFDebtRestructuring.Krueger.pdf [].

(53). See Gulati & Scott, supra note 10, at 53-119 (reporting both on data and interviews with lawyers).

(54). Id. at ch. 5.

(55). Other scholars were also puzzled by the bond lawyers' failure to correct in their contracts what they were asserting was a grievous error in court interpretation. See Stephen Nelson, Market Rules: Social Conventions, Legal Fictions and the Organization of Sovereign Debt Markets in the Long Twentieth Century, in Contractual Knowledge: One Hundred Years of Legal Experimentation in Global Markets 118, 141-43 (Gregoire Mallard & Jerome Sgard eds., 2016); Leland Goss, NML v. Argentina: the Borrower, the Banker and the Lawyer: Contract Reform at a Snail's Pace, 9 Cap. Mkts. L.J. 287, 288-90 (2014); Umakanth Varottil, Sovereign Debt Documentation: Unraveling the Pari Passu Mystery, 7 DePaul bus. & comm. L.J. 119, 122-25 (2008).

(56). Gulati & Scott, supra note 10, at chs. 6-8.

(57). The context was a request for injunctive relief. See id. at ch. 1.

(58). Peru's lawyers in the case were from the firm Cleary Gottlieb Steen & Hamilton, probably the leading sovereign side law firm in the world. See id. at ch. 1; see also Michael Bradley, Irving De Lira Salvatierra & Mitu Gulati, Lawyers: Gatekeepers of the Sovereign Debt Market?, 38 INT'L Rev. L. & ECON. 150, 162 (2014) (describing the market for lawyers in the sovereign debt context, and the key players there).

(59). Gulati & Scott, supra note 10, at 170-75. This version of the pari passu clause is reproduced infra note 87 and accompanying text.

(60). See Allied Bank Int'l v. Banco Credito Agricola de Cartago, 757 f. 2d 516, 522-23 (2d Cir. 1985). For a discussion of the case, see Federico Sturzenegger & Jeromin Zettelmeyer, Debt Defaults and Lessons from a Decade of Crisis 64-66 (2006).

(61). Judge Griesa's attempts to manage the aftermath of the Argentine default and the various litigations is discussed in Marcus Miller & Dania Thomas, Sovereign Debt Restructuring: Judge Griesa, the Vultures and Creditor Rights, 30 World Econ. 1491 (2007).

(62). Order, NML Capital, Ltd. v. Republic of Argentina, No. 08 Civ. 6978 (TPG), 09 Civ. 1707 (TPG), 09 Civ. 1708(TPG) (S.D.N.Y. Dec. 7, 2011), 2011 WL 9522565, at *2. As in Brussels, the judge granted an injunction. This order applied to the institutions that might have otherwise helped Argentina pay the non-holdout creditors-they were at risk of being in contempt of court if they aided or abetted the debtor. NML Capital, Ltd. v. Republic of Argentina, No. 1:08-cv-06978, l:08-cv-01707, 1:08-cv-01708 (S.D.N.Y. Feb. 23, 2012).

(63). Some even intimated that the judge's advancing age (he was an octogenarian) may have played a role in his decision. To get a flavor of some of the critiques, see, e.g., Floyd Norris, The Muddled Case of Argentine Bonds, N.Y. Times (July 24, 2014), 2014/07/25/business/rulings-add-to-the-mess-in-argentine-bonds.htmI [] and Linette Lopez, Check Out This Crazy Argentine Propaganda Poster With an American Judge's Head on a Vulture's Body, Bus. Insider (Aug. 13, 2014, 5:35 PM), http:// [ VT7L-X5M6]. See also Theresa Monteleone, A Vulture's Gumble: High-Stakes Interpretation of Sovereign Debt Contracts in NML Capital, Ltd v. Republic of Argentina, 8 Cap. Mkts. L.J. 149. 160-65 (2013); Tim Samples, Rogue Trends in Sovereign Debt: Argentina. Vulture Funds, and Pari Passu Under New York Law, 35 Nw. J. Int'l L. & Bus. 49, 81-85 (2014); W.M.C. Weidemaier, Sovereign Debt After NML v. Argentina, 8 Cap. Mkts. L.J. 123, 130-31 (2013).

(64). Brief for the United States of America as Amici Curiae Supporting Appellee, NML Capital, Ltd. v. Republic of Argentina, 727 F. 3d 230 (2d Cir. 2013). A leading New York law firm. Shearman & Sterling LLP, has usefully consolidated and organized all of the key filings in the NML v. Argentina case. See Argentine Sovereign Debt, Shearman & Sterling LLP, http:// [].

(65). Brief for the Clearing House Association as Amici Curiae Supporting Non-Party, NML Capital, Ltd. v. Republic of Argentina, 727 F. 3d 230 (2d Cir. 2013); Brief for the Clearing House Association as Amici Curiae Supporting Reversal, NML Capital, Ltd. v. Republic of Argentina, 727 F. 3d 230 (2d Cir. 2013).

(66). Brief for the American Bankers Association as Amici Curiae Supporting Non-Party, NML Capital, Ltd. v. Republic of Argentina, 727 F. 3d 230 (2d Cir. 2013).

(67). NML Capital, Ltd. v. Republic of Argentina, 699 F. 3d 246, 250-51 (2d Cir. 2012).

(68). The amici were Anne Krueger and Joseph Stiglitz, respectively. Brief for Joseph Stiglitz as Amicus Curiae Supporting Petitioner, Republic of Argentina v. NML Capital, Ltd., 134 S. Ct. 2819 (2014) (mem.) (No. 13-990). The countries filing briefs at the U.S. Supreme Court, in opposition to the lower court's pari passu ruling were Brazil, France, Mexico, and the United States. Brief for the Federative Republic of Brazil as Amicus Curiae in Support of Petitioner, Republic of Argentina v. NML Capital, Ltd., 134 S. Ct. 2819 (2014) (No. 13-990); Brief for the Republic of France as Amicus Curiae Supporting the Republic of Argentina's Petition for a Writ of Certiorari, Republic of Argentina v. NML Capital, Ltd., 134 S. Ct. 2819 (2014) (No. 13-990); Brief for the United Mexican States as Amicus Curiae Supporting Petitions for Writs of Certiorari, Republic of Argentina v. NML Capital, Ltd., 134 S. Ct. 2819 (2014) (No. 13-990); Brief for the United States as Amicus Curiae in Support of Petitioner, Republic of Argentina v. NML Capital, Ltd., 134 S. Ct. 2250 (2014) (No. 12-842).

(69). See NML Capital, Ltd. v. Republic of Argentina, 134 S.Ct. 2819, 2819 (2014) (mem.). Not only did the Supreme Court decline to reinterpret pari passu, but in a related dispute over the scope of discovery allowed NML, the Court ruled in a fashion that facilitated NML's efforts to pursue its litigation overseas. Karen Halverson Cross, U.S. Supreme Court Denies Certiorari and Affirms Discovery in Bondholder Litigation Against Argentina, Am. Soc'y Of Int'l Law: Insights (Oct. 15, 2014), -denies-certiorari-and-affirms-discovery-bondholder []

(70). For a description of the "back story" from a senior official at the U.S. Treasury, see Mark Sobel, Strengthening Collective Action Clauses: Catalyzing Change-The Back Story, 11 Cap. Mkts. L.J. 3, 6-9 (2016).

(71). Id.

(72). Id. For additional detail on the reform efforts, see Goss, supra note 55, at 287-97 and Anna Gelpern, Brad Setser & Ben Heller, Count the Limbs: Designing Robust Aggregation Clauses in Sovereign Bonds, in Too Little, Too Late: The Quest to Resolve Sovereign Debt Crises 109 (Martin Guzman, Jose Ocampo & Joseph Stiglitz eds. 2016).

(73). Goss, supra note 55, at 289. The final version of the proposed ICMA clauses was issued roughly six months later in June 2014.

(74). For a description of the IMF's actions, see Nikita Agarwal & Chanda De Long, Strengthening the Contractual Framework for Sovereign Debt Restructuring, 11 Cap. Mkts. L.J. 25, 26-27 (2016). See also Elaine Moore, IMF Recommends Overhaul of Sovereign Bonds, FIN. Times (Oct. 6, 2014), [] (detailing reforms proposed by the IMF).

(75). See generally G20 Finance Ministers and Central Bank Governors, Communique for the G20 Summit in Ankara, Turkey 6 (2015), http:// [] ("We call on the IMF, in consultation with other relevant parties, to continue to promote and monitor the progress on the implementation of the strengthened collective action and pari passu clauses....").

(76). According to the Report, over 60 percent of new issues under both English and New York law have amended their pari passu clauses in the post-October 2014 period. See Agarwal & De Long, supra note 74, at 26.

(77). See Gulati & Scott, supra note 10, at 26.

(78). See Int'l Monetary Fund, Progress Report on the Inclusion of Enhanced Contractual Protections in International Sovereign Bond Contracts 8 (2015), []. For an illustration of the reduction in sovereigns employing pari passu clauses that ranked payments equally, see infra fig. 5.

(79). The overwhelming sentiment (as publicly articulated in the press, and what we have heard in interviews for this project) is that the interpretation of pari passu by the courts in Brussels and New York reduced the efficiency of the sovereign bond market. Nevertheless, there is the counterargument that observed variations in pari passu are a rational precommitment by debtors with different risk profiles to either foreclose or invite the possibility of a future restructuring attempt. See, e.g., Kahan & Leshem, supra note 40, at 15-17; Wright, supra note 40, 112-14. In a companion empirical paper, we test the "Rational Design" hypothesis and find it wanting. See Choi et al., supra note 15, at 32-33. That evidence is also consistent with the view of this case in the financial press:
      So the people who thought the holdouts' pari passu theory was
   crazy seem to be right. The world seems to be unambiguously worse
   off than it would be if the U.S. courts had rejected the theory.
   Argentina is worse off (it's in a default crisis), the exchange
   bondholders are worse off (their interest payments have stopped),
   and the holdout bondholders are no better off, since it's not like
   they're getting paid anything either.

      More generally, it's crazy because it makes sovereign debt
   restructuring impossible. If you know your choice is to take 30
   cents on the dollar or to lake zero, you'll probably take 30. But
   if you know that if everyone else takes 30, you can still sue for
   full repayment, then you'll have no incentive to negotiate. The
   courts' reading of the pari passu clause means that those holdout
   bonds can never be compromised; every single holder of those bonds
   can demand full payment. This is a problem that Argentina currently
   faces: Even if the holdout bondholders who sued Argentina do want
   to reach a settlement for less than 100 cents on the dollar, they
   have no way of binding other holdout bondholders.

Matt Levine, How Should Future Argentinas Treat Future Vultures?, Bloomberg: View (Aug. 29, 2014, 5:00 PM), -argentinas-treat-future-vultures [].

(80). Also outside the dataset are bonds that might be issued in some non-AAA countries to purely domestic investors, usually, captive domestic banks, in which bankers, lawyers, and contracts are not involved.

(81). In a companion paper, with an econometric analysis, we code for additional variables such as the bond ratings and offering amounts. See Choi et al., supra note 15, at 25.

(82). Supranationals include entities such as the Asian and African Development Banks.

(83). This latter aspect is important because the supposedly damaging impact of the ratable payments interpretation is most relevant in the context where there is no bankruptcy court to oversee a restructuring and, if necessary, force a resolution. We only coded these bonds if they used pari passu clauses in the period prior to December 2011 (because our interest is whether and how the clauses changed in these bonds in response to the case).

(84). Gulati & Scott, supra note 10, at ch. 4.

(85). See Mitu Gulati & Kenneth Klee, Sovereign Piracy, 56 Bus. Law. 635, 639-40 & 639 n. 26 (2001).

(86). In prior work, we suggested a story for why the clause might have made sense in the era of gunboat diplomacy in the 1800s. With the gunboats, after all, the foreign creditors could take over the debtor's ports, where the tax revenues were collected, and, in effect, liquidate the assets. But almost no one among the parties to the current transactions seemed aware of that earlier interpretation nor, more importantly, seemed to care. Gulati & Scott, supra note 10, at chs. 8-9.

(87). See Lee C. Buchheit & Sofia D. Marios, What To Do About Pari Passu?, 8 Butterworths J. Int'l banking & fin. L. 491, 491 (2014) (emphasis added) (emphasis omitted). The evolution of the different formulations of the clause is described in Mark Weidemaier, Robert E. Scott & Mitu Gulati, Origin Myths, Contracts, and the Hunt for Pari Passu, 38 L. & Soc. Inquiry 72 (2013).

(88). See Buchheit & Pam, supra note 50, at 878 (describing Lowenfeld's argument).

(89). Id.

(90). In our empirical tests of pricing effects for a companion article, we code these on a scale of 10, 5, 1, and 0, which roughly corresponds to the relative risk levels.

(91). See Weidemaier et al., supra note 87, at 88 fig. 2 and accompanying text (illustrating the increase in similar "mandatory law" phrases in English-law bonds). This provision has come to light in the context of Ukraine's recently concluded restructuring exercise. See Joseph Cotterill, Ukraine's Bonds: A Little Local Leverage?, fin. Times: Alphaville (Mar. 26, 2015), [] (providing examples of different sovereigns including similar provisions in their bonds).

(92). For discussions of the evolution of these different clauses and their risk levels, see Buchheit & Martos, supra note 87, at 491-93 and Weidemaier et at., supra note 87, at 84-90 figs. 1, 2 & 3. We had also envisioned coding some bonds as zero risk, where the pari passu clause had been deleted altogether, but no such observations appear in the data.

(93). See, e.g., Republic of Hond., Offering Circular, 7.50% Notes Due 2024, at 8 (2013), [https://perma. cc/EER8-4356] ("To ensure clarity on the point, Honduras intends to take the position that the pari passu clause in the terms and conditions of the Notes does not obligate it to pay Public External Indebtedness on a ratable basis.").

(94). For a classic case where arguments bearing on how a bond contract provision should be interpreted were made on the basis of risk disclosures in the prospectus and other public statements by company officials, see Metro. Life Ins. Co. v. RJR Nabisco, Inc., 716 F. Supp. 1504, 1514-15 (S.D.N.Y. 1989) ("In interpreting these contracts, this Court must be concerned with what the parties intended, but only to the extent that what they intended is evidenced by what is written in the indentures." (emphasis added)).

(95). The operative language from Paraguay's Offering Circular from January 2013 says:
      In ongoing litigation in federal courts in New York captioned
   NML Capital, Ltd. v. Republic of Argentina, the U.S. Court of
   Appeals for the Second Circuit has ruled that the ranking clause in
   bonds issued by Argentina prevents Argentina from making payments
   in respect of the bonds unless it makes pro rata payments in
   respect of defaulted debt that ranks pari passu with the performing
   bonds. The judgment has been appealed.

      We cannot predict when or in what form a final appellate
   decision will be granted. Depending on the scope of the final
   decision, a final decision that requires ratable payments could
   potentially hinder or impede future sovereign debt restructurings
   and distressed debt management.... Paraguay cannot predict whether
   or in what manner the courts will resolve this dispute or how any
   such judgment will be applied or implemented.

Republic of Para., Offering Circular, 4.625% Bonds Due 2023, at 17 (2013), [ 3GQM-LZ2A].

(96). For a description of these variations, see Lee C. Buchheit & Mitu Gulati, Sovereign Debt Restructuring After NML v Argentina, 12 CAP. Mkts. L.J. 224, 238 app. (2017) and Rodrigo Olivares-Caminal, The Definition of Indebtedness and the Consequent Imperiling of the Pari Passu. Negative Pledge and Cross-Default Clauses in Sovereign Debt Instruments, 12 Cap. Mkts. L.J. 164, 169-78 (2017).

(97). See Olivares-Caminal, supra note 96 at 170-72.

(98). See Buchheit & Gulati, supra note 96, at 225-28.

(99). Id.

(100). Id. at 231-32, 238 app. (describing the versions).

(101). Leading lawyers in the field also shared this assumption. See, e.g., Buchheit & Martos, supra note 87, at 491-92; Goss, supra note 55, at 288-90.

(102). See supra notes 65-70 and accompanying text.

(103). Buchheil & Martos, supra note 87, at 492 (emphasis added).

(104). Berry, supra note 50, at 1.

(105). See Int'l Monetary Fund, Progress Report on the Inclusion of Enhanced Contractual Protections in International Sovereign Bond Contracts 2 (2015), [].

(106). For our analysis, we start with 1, 583 issuances between June 1, 2011 to May 30, 2016, and examine whether the pari passu clause for a sovereign or quasi-sovereign's issuances on a particular day is different from the issuer's last issuance prior to the start of the dataset on June 1, 2011. For our analysis, we collapsed multiple issuances on the same day into one observation--leaving us with 1281 unique issuer-issue date observations--and treated a change as occurring if a particular change occurred for any of the issuances on that day relative to the last issuance prior to June 1, 2011.

(107). As an aside, the Second Circuit reaffirmed its decision in August 2013, NML Capital, Ltd. v. Republic of Argentina, 727 F. 3d 230, 245-48 (2d Cir. 2013), but that action does not seem to have had much of an effect.

(108). The Supreme Court's refusal to review the interpretation issue was likely anticipated by the market in any event since the Supreme Court almost never takes contract interpretation cases in the current era.

(109). See Lucy McNulty, The Future for Pari Passu, Int'l Fin. L. Rev., Mar. 2013, at 19, 19 (reporting from a survey of market actors in the sovereign debt industry that there was widespread agreement as to the need for contract reform, yet tremendous inertia as a result of the difficulties in coordinating a move to a new standard).

(110). Although imprecise, a measure of the costs of an erroneous pari passu decision are provided by the costs Argentina had to bear as a result of the decade-long litigation in New York. Among the direct indications of these costs are the facts that Argentina was denied access to the foreign capital markets for over a decade and the amounts it was eventually willing to pay holdout creditors to be able to gain access in 2016. For a detailed discussion, see generally Martin Guzman, An Analysis of Argentina's 2001 Default Resolution (CIGI Working Paper No. 110, 2016). Another illustration of the costs embedded in the use of a "risky" clause is the looming problem in Venezuelan bonds that are on the verge of default. See Robin Wigglesworth, Small Print on Venezuelan Debt Will Pique Wall Street's Interest, Fin. Times (June 17, 2016), https:// [].

(111). Our hypothesis that black hole clauses are subject to extraordinary inertia costs finds support in the reaction of the market to the Brussels decision in 2000. Rather than revise the pari passu clause to eliminate the risk of a ratable payments interpretation, the market chose instead to coordinate on revising the no-modification, or unanimous action, clauses in New York bonds that required unanimous approval of all creditors to change contract terms prior to a restructuring. Coordinating an agreement on revisions that lowered the vote to modify to 75 percent--thus increasing the burden on holdouts to obtain a blocking position--was accomplished in a little over two years, from late 2001 to mid-2003. This revision ameliorated but did not solve the holdout problem: it was preferred, we surmise, because, unlike pari passu, the no-modification clauses were well understood and familiar to the market. For discussion of this shift in contract language, see generally Anna Gelpern & Mitu Gulati, Public Symbol in Private Contract: A Case Study, 84 Wash. U. L. Rev. 1627 (2006).

(112). As contrasted to "quasi-sovereign" issues, which are issuances by cities, stales, regions within the sovereign state or issuances by some public or private corporate entity backed by a sovereign guarantee.

(113). The restructuring of Greek local-law governed bonds engineered via domestic legislation in March 2012 is an example of the vulnerability of local-law bonds, as compared to foreign ones. See generally Jeromin Zettelmeyer, Christoph Trebesch & Milu Gulali, The Greek Debt Restructuring: An Autopsy, 28 econ. Pol'y 513 (2013) (contrasting the treatment of the bonds governed under different laws).

(114). We examine whether the pari passu clause for a sovereign's issuances on a particular day differs from the sovereign's last issuance prior to the start of the dataset on June 1, 2011. For our analysis, we collapsed multiple issuances on the same day into one observation--leaving us with 545 unique issuer-issue date observations for pure sovereigns under foreign law--and treated an amendment as occurring if a particular modification occurred for any of the issuances on that day relative to the last issuance prior to June 1, 2011.

(115). The first bond to disavow explicitly the ratable payments interpretation was an issuance by Belize in June 2013, roughly eighteen months after the NML v. Argentina decision in the trial court. See Buchheit & Martos, supra note 87, at 493. Belize was a first mover partly because it had just gone through a restructuring, the first restructuring after the NML v. Argentina decision. See Robin Wigglesworth, Belize Does "Superbond" Deal with Lenders, Fin. Times (Feb. 13, 2013), [],

(116). Federated Republic of Braz., Prospectus Supplement, 5% Global Bonds Due 2045, at S-8 (2014), d761027d424b5.htm [].

(117). This sense of disapproval would be amplified further if one were to look at the amicus brief filed by the government of Brazil asking the U.S. Supreme Court to hear the pari passu dispute and reverse the lower court. Brief for the Federal Republic of Brazil as Amici Curiae Supporting Petitioner, supra note 68, at 3-7.

118. For details on how the group was constructed, see infra note 142.

119. A key participant explained:
      The Columbia meeting set the cat among the pigeons. [What we saw
   at the meeting] was a good deal of unhappiness among the NY lawyers
   (and some of the large Latin American sovereign issuers) with the
   drafting of the [] clauses. [That drafting had been done primarily
   by the English lawyers in collaboration with ICMA and had an
   English law style], [In response to the problem] Cleary Gottlieb
   and Sullivan & Cromwell quickly drafted "New York versions" of the
   clauses, intended to be substantively the same but written in a
   more plain-speaking, Yankee homespun manner. FRBNY was chosen
   because of its gravitas-the participants needed to understand this
   was important [and that meeting at Columbia might have had the
   opposite effect].

Email from a meeting participant to Authors (Aug. 13, 2016) (on file with authors).

(120). According to one source:
      [It was made clear to the participants] that no one wanted to
   open the door on substantive points being renegotiated. [The goal
   was to] spend a sufficient amount of time describing the
   collaboration/engagement/orchestration between the parties and the
   significant period of discussion on the substance that preceded it
   in the hope that we could obtain buy in.

      [Lengthy calls were had, I believe] with [many] of the
   participants on the contract reforms and it was only after they
   understood the level of discussion and collaboration and the
   "settled" nature of the end product (and its advantages fully) that
   they came on board with not unpicking key elements.

Email from a meeting participant to Authors (Aug. 13, 2016) (on file with authors).

(121). According to a participant:
      Sitting at that table at the Fed were lawyers representing a
   substantial pari of the Latin American sovereign issuer and
   underwriter community. So when the U.S. Treasury [representative]
   pronounced at the meeting that a consensus seemed to have been
   reached, everyone left with the sense that they were morally
   committed to encourage their clients to use the new NY clauses. Or
   at least not to discourage their clients from using the clauses.

Email from a meeting participant to Authors (Aug. 14, 2016) (on file with authors).

(122). There were more conversations about this topic at the annual World Bank/IMF meeting during the roughly three-week period between the two meetings. In particular, there were multiple sessions during the IMF/World Bank meetings for debt managers around the globe at which experts on the pari passu litigation were asked to conduct seminars explaining the implications of the ratable payments interpretation and how the contracts could be reformed.

(123). See generally Choi et al., supra note 15 (using statistical survival models to gauge which event spurred the adoption of new pari passu clauses).

(124). See Weidemaier et al., supra note 87, at 84 fig. 1 (describing the data from 1960 to 2011 for pure sovereign issuers).

(125). See generally Choi et al., supra note 15 (detailing the evidentiary shifts, particularly during the period between December 2011 and October 2014).

(126). Id. at 22-32.

(127.) Id. at 3-4; see also Anderson & Manns, supra note 33, at 80 (calling this the "artisanal model").

(128). The value of having a standard form that everyone understands and can price accurately produces positive network effects. Kahan & Klausner, supra note 34, at 731-33.

(129). Scholars use the assumption of Rational Design to attempt to explain variation in the pari passu context. See generally Benjamin Chabot & Veronica Santarosa, Don't Cry for Argentina (or Other Sovereign Borrowers): Lessons from a Previous Era of Sovereign Debt Contract Enforcement, 12 Cap. Mkts. L.J. 9 (2017) (explaining the rationality behind lending to sovereigns); Kahan & Leshem, supra note 40 (discussing the strategic consequences of pari passu clauses); Wright, supra note 40 (discussing rationality of actions).

(130). See supra note 129.

(131). Choi et al., .supra note 15, at 4.

(132). E.g., Anderson & Manns, supra note 33, at 80-84; Richman, supra note 33, at 81-82.

(133). Choi et al., supra note 15, at 31-34.

(134). We performed a chi-squared test of the proportions of the different types of pari passu clauses for each year in our study time period as depicted in Figure 5. The chi-squared test indicated that the difference in the proportions for the different types of pari passu clauses across the years is significant (prob. = 0.000).

(135). See Weidemaier el al., supra note 88, at 84-89.

(136). See Choi el al., supra note 15, at 11.

(137). The ICMA recommended clause from August 2014 reads as follows:
      The Notes are the direct, unconditional and unsecured
   obligations of the Issuer and rank and will rank pari passu,
   without preference among themselves, with all other unsecured
   External Indebtedness of the Issuer, from time to time outstanding,
   provided, however, that the Issuer shall have no obligation to
   effect equal or ratable payment(s) at any time with respect to any
   such other External Indebtedness and, in particular, shall have no
   obligation to pay other External Indebtedness at the same time or
   as a condition of paying sums due on the Notes and vice versa.

Int'l Capital Mkt. Ass'n, Standard Pari Passu Provision for the Terms and Conditions of Sovereign Notes (2014), Resources/ICMA-Standard-Pari-Passu-Provision-August-2014.pdf [ JY].

(138). Choi el al., supra note 15, at 25, 26.

(139). Id. at 27.

(140). Id. at 26.

(141). See infra Pari IV.A.1 and IV.A.2 respectively.

(142). We explained that we would present our initial findings and hoped to gather their reactions. On hearing about the meeting, a number of other senior lawyers, bankers, and policymakers also asked to be included in the discussion. This resulted in a session with roughly thirty senior sovereign debt lawyers from both the private and public sectors.

Our process of selection was simple. Using prior research on the role of lawyers in the sovereign debt market, we invited the senior partners from the ten leading firms in New York and London, the two leading jurisdictions for sovereign debt issuances. This gave an initial set of roughly twenty invitees. Three of the London lawyers who were not planning to be in the United States for the World Bank/IMF meetings later that week declined. However, a number of those invited asked whether they could bring colleagues from their firms. This resulted in a total of twenty-four attendees from major law firms. In addition, we invited senior counsel from the major official sector institutions such as the U.S. Treasury, the IMF and the Federal Reserve, and the major industry groups such as the Emerging Markets Traders Association and ICMA, which gave a total of thirty participants.

(143). We also asked these questions at a meeting in Paris, in October 2014, with a number of European sovereign debt lawyers and policymakers. This meeting was organized by Rodrigo Olivares-Caminal of Queen Mary Law School. The format of this second conference followed the presentation plus audience questions formal, and there were about a dozen lawyers and policymakers. We presented the same data and took notes on the responses, which were substantially similar to the ones we received in New York.

(144). The majority of this latter set of interviews were done as part of a project that one us is doing with Anna Gelpern and Jeromin Zettelmeyer that is focused on the question of how market participants think (and talk) about the price impact of making contract modifications. See Gelpern et al., supra note 26, at 4-13.

(145). See supra notes 59-69 and accompanying text.

(146). Audio tape: Columbia Law School Meeting (Oct. 2014) (on file with authors).

(147). Accounts of how Argentina had angered and frustrated Judge Griesa are numerous. See, e.g., Tomas M. Araya, A Decade of Sovereign Debt Litigation: Lessons From the NML v. Argentina Case and the Road Ahead, 17 Bus. L. Int'l 83, 99 (2016) ("The 2012/2014 decisions from Judge Griesa and the Court of Appeals were undoubtedly, connected with the Republic's attitude of disrespect of the US judicial system...."); see also Sheelah Kholhatkar, A Judge's Rage Goes Ignored While Argentina Sidesteps Holdout Hedge Funds. Bloomberg: Businessweek (Sept. 30, 2014), argentina-vs-dot-holdout-hedge-funds-judge-griesas-rage-is-ignored [] (discussing Judge Greisa's anger toward Argentina); Alexandra Stevenson, Frustrated Judge Scolds Argentina but Does Not Hold It in Contempt, N.Y. Times: Dealbook (Aug. 21, 2014), [] (discussing Judge Griesa's frustration with Argentina).

(148). See generally Weidemaier, supra note 63 (discussing the passage of the "Lock Law," among other factors, that resulted in the court's decision).

(149). See NML Capital, Ltd. v. Republic of Argentina, 699 F. 3d 246, 258-60, 264 n. 16; see also Buchheit & Gulati, supra note 96, at 2-3 (discussing this aspect of the cases).

(150). The appellate court opinion focused on the use of the word "payment" in Argentina's pari passu clause but made little attempt to interpret the meaning of the text in harmony with the other words in the bond contract. NML Capital, Ltd., 699 F. 3d at 258-60. In reports issued under the auspices of the Bank of England, a group of the most eminent U.K. sovereign debt lawyers has on multiple occasions expressed its view in writing that U.K. courts would not make that mistake as a matter of English law. See Fin. Mkts. Law Comm., Issue 79-Pari Passu Clauses 22 (2005); Fin. Mkts. Law Comm., Role Use and Meaning of Pari Passu Clauses in Sovereign Debt as a Matter of English Law 1 (2014); Fin. Mkts. Law Comm., Analysis of the Role, Use and Meaning of Pari Passu Clauses in Sovereign Debt as a Matter of English Law 10-11 (2015); see also Lachlan Burn, Pari Passu Clauses: English Law After NML v. Argentina, 9 Cap. Mkts. L.J. 2, 2-9 (2013) ("There is almost no risk that English courts, faced with similar facts to NML v Argentina, would adopt the 'payment' interpretation."); Tolek N. Petch, NML v Argentina in an English Legal Setting, 9 Cap. Mkts. L.J. 266, 271 (2013) ("It follows that the approach endorsed by the Court of Appeals would be found wanting in an English court. There is therefore no reason to revisit standard pari passu wording in sovereign bond documentation issued under English law.").

(151). Kensington Int'l Ltd. v. Congo [2003] EWHC (Comm) 2331 [91]; Kensington Int'l Ltd. v. Congo [2003] EWCA (Civ) 709 [13] (approving the decision); see also Global Law Intelligence Unit, Allen & Overy, The Pari Passu Clause and the Argentine Case 10-11 (2012), clause%20and%20the%20Argentine%20case.pdf [] (discussing the case).

(152). Former Lord Chief Justice Phillips, in his filing in support of the plaintiffs, wrote:
      An English court would approach the problem of interpreting the
   [pari passu] clause with a wish, if possible, to give it the same
   meaning as that which it now bears under the law of New York.... I
   consider that the application of the principles of construction
   that apply under English law would result in it doing just that.

Declaration of Lord Nicholas Phillips K.G., P.C. at 24. Red Pines LLC v. Republic of Argentina, No. 1:14-CV-09427 (S.D.N.Y. Mar. 2, 2016). As his declaration tells us, Judge Phillips was quite familiar with the NML Capital, Ltd. v. Argentina litigation, since the appeal in that case in 2011 had come before him; although the pari passu interpretation issue was not raised. Declaration of Lord Nicholas Phillips K.G., P.C., supra, at 2.

(153). The data we collected for this paper shows that, as of May 31, 2016, not a single issuer switched from New York to English law in the wake of Judge Griesa's decision in December 2011.

(154). For a discussion of the lawyer's views, sec infra Part III.E.1.d.

(155). This lack of a negative market reaction perceived by the respondents is consistent with what the most recent research shows as well. See Faisal Z. Ahmed & Laura Alfaro, Market Reactions to Sovereign Litigation, 12 cap. Mkts. L.J. 141, 154 tbl.1, 156 fig. 3 (2017) (finding no abnormal negative returns from the NML case for any Latin American sovereign issuers' bonds, other than Argentina itself).

(156). Buchheit & Martos, supra note 87, at 492 (emphasis added).

(157). See Anna Gelpern, Courts and Sovereigns in the Pari Passu Goldmines, 11 CAP. MKTS. L.J. 251, 254-56 (2016); Sung Hui Kim, Pari Passu. The Nazi Gambit, 9 CAP. MKTS. L.J. 1, 243 (2014).

(158). Kim looked at diplomatic correspondence in the archives in Geneva. Kim, supra note 157, at 242-50. Gelpern looked at transcripts of court decisions and related memos in archives in Geneva. Gelpern, supra note 157, at 253-54.

(159). Gelpern, supra note 157, at 272-73; Kim, supra note 157, at 249-50.

(160). Chabot & Gulati, supra note 47, at 235-36. Pablo Triana provides another historical examination of the origins of the pari passu clause. See generally Triana, supra note 47 (finding a pari passu clause in the bonds of a sub sovereign, the City of Edinburgh, in 1838).

(161). Email from a meeting participant to Authors (Aug. 14, 2016) (on file with authors).

(162). As an example, the Financial Times quotes the Nobel laureate Joe Stiglitz as saying of the case: "It's a disaster for the world.... It sets an enormously bad precedent and will cause a lot of anxiety in the global financial system." Robin Wigglesworth & Elaine Moore, Sovereign Debt: Curing Defaults, Fin. Times (June 7, 2016), [].

(163). Bradley el al., supra note 58, at 153.

(164). Id.

(165). Id.

(166). See generally Gelpern et al., supra note 26 (describing interviews conducted with government debt managers between 2013 and 2017).

(167). Our colleague, who ran three separate training sessions for the debt managers in Washington, D.C. during the period between the first Columbia Law meeting and the New York Federal Reserve meeting, said that the debt managers struck her as annoyed with the New York decisions (at least, more so than they were when we met with them some months later). In an email, she wrote:
   [T]he debt managers were angrier about the [NY] decisions than the
   lawyers-reaction was anger and disbelief, that is why I was invited
   to present to the debt managers as a group [three] limes in
   different places (not to or with their lawyers), and that is why
   all those [debt managers] agreed to see us-they felt like they
   needed an independent understanding of the matter. But they also
   thought the problem was with the court and had to be fixed
   somewhere in a more centralized way, perhaps in the courts or in
   the legislature (this was [especially] true of non-lawyers and
   people from continental legal [systems]). The contracts were a
   fallback, risk management on the margins pending global solution.
   The [question] for [the debt managers] was whether their existing
   [contracts] expose them to such a degree that they cannot afford to
   wait. And for most, the answer was no.

Email from a meeting participant to Authors (Jan. 10, 2016) (on file with authors).

(168). See generally Anna Gelpern & Mitu Gulati, How Much Is This Clause'? Debt Managers on Pricing Bond Contract Terms 11-12 (Aug. 25, 2016) (unpublished manuscript) (on file with the Duke Law Journal) ("Once the sovereign ... settles on ... price, the lawyers may be asked to 'paper the deal' in line with this decision.... In other words, price-setting occurs apart from any variation in legal terms.").

(169). See id. at 3, 11.

(170). See generally id. ("The widely-held view was that, in the ideal world, non-financial ('legal') terms in sovereign bonds should be irrelevant.").

(171). Id. at 42.

(172.) See id. at 21 (explaining, in the context of drafting contracts for sovereign guarantees and issuances by sub-sovereigns, like regions, that there is often a strict division of responsibility with the debt managers for the sovereigns generally not coordinating with the managers for the quasi sovereigns about what terms to use).

(173.) On the issue of the large number of sovereign debt guarantees outstanding and the prospect of them all migrating to the sovereign's balance sheet in the event of a crisis, see Elaine Moore & Jonathan Wheatley, Fears Mount Over Rise of Sovereign-Backed Corporate Debt, Fin. Times (Jan. 5, 2016), [https://].

(174.) One senior debt manager for a country that had not, as yet in early 2016, changed its clauses explained:
   We will change the clauses in our foreign law bonds, the next lime
   we do our issuance. Our outside lawyers know the IMF's
   recommendations. The local law bonds are different. We have pari
   passu clauses in them. But we have only one lawyer in our
   department. We put pari passu clauses in because we wanted outside
   investors to buy local bonds-and investors are used to seeing these
   clauses. They have a check list. Pari passu, negative pledge,
   exchange listing ... there are a set of things they look to see
   whether they are present. If so, they are okay with the bond. So,
   we put pari passu in. Seemed okay, since everyone else has it. [X
   country] also has it in their local law bonds. We are not a strong
   issuer; we need to make investors comfortable with our bonds. We
   only now noticed [after reading your article] that there are
   differences in the wording of our local law and foreign law pari
   passu clauses. We will change that.

Email from debt manager to Authors (Aug. 14, 2016) (on file with authors).

(175.) One debt manager explained why avoiding the IMF's black list was important:
   The investors have not demanded it [change] yet. But they want the
   standard forms. And if the standard has changed, they will want the
   new standard. Also, we don't want to be on the list of countries
   that have not fixed their contracts. We were on the list that the
   IMF put out. That was not good. But the bond we did was a small
   one, private issuance, and the investor did not ask. We tried to
   explain to the IMF, but they put us on the list [of those who had
   not changed) anyway. Not good. Next lime, we will have changed.

Email from debt manager to Authors (Aug. 14, 2016) (on file with authors).

(176.) The one debt manager who had considered doing an exchange of his old bonds appeared to wish to be seen as a leader and innovator in the world of debt managers. But, as of this writing, no real steps seem to have been taken in his office to engineer an exchange of the type we describe.

(177.) This exhortation to change the terms of earlier-issued bonds came publicly from the director of the IMF, Christine Lagarde, at the June 2016 meetings of the Paris Club. Email from Isabelle Couet to Authors (2016) (on file with authors).

(178.) See Gelpern & Gulati, supra note 168, at 4, 50-51.

(179.) Id. at 1-2.

(180.) Id. at 50.

(181.) Id. at 50-51.

(182.) Id.

(183.) Id. at 51.

(184.) Id. at 50.

(185.) The standard view in contract theory assumes that sophisticated parties have a better understanding of what the terms in their contracts mean than do courts, and thus have a corresponding incentive and capacity to revise those terms when necessary. For articulations of this view, see generally Meredith R. Miller, Contract Law, Party Sophistication and the New Formalism, 75 Mo. L. Rev. 493 (2010) and Schwartz & Scott, supra note 1. The economics literature on standard form contracts suggests that boilerplate contract terms might sometimes be slow to change because of network or learning externalities--such as the value of having identical terms that everyone understands and can easily and accurately price and litigate because understanding their legal meaning has become widespread. That rationale may apply in the case of some types of boilerplate contract terms, but the black holes that we examine are clauses whose meaning has dissipated over time and whose pricing is nonexistent. The classic statement of the externality argument is in Kahan & Klausner, supra note 34, at 715-17. For an argument that the Kahan and Klausner argument does not apply to the pari passu context, see Gulati & Scott, supra note 10, at ch. 6.

(186.) This is not the case with all the sovereigns in the data because some of the highest-rated issuers have no legal clauses at all in their pure sovereign bonds, but have legal clauses, including pari passu, in their sub-sovereign issuances. When we spoke to the debt managers of these issuers, however, they indicated that they would of course revise the pari passu clauses in their pure sovereign bonds if that were needed. But since they issue sovereign bonds without any legal terms, there is no problem with those bonds. When asked about their quasi-sovereign bonds, they generally expressed surprise.

(187.) Excessive discounting by agents thus leads to bond issuances that are less efficient than they could be. An efficient sovereign bond contract optimizes total contracting costs by trading off the ex ante or front-end costs of the contract and the ex post or back-end costs of default. For discussion, see Robert E. Scott & George G. Triantis, Anticipating Litigation in Contract Design, 115 Yale L.J. 814, 822-25 (2006).

(188.) See id. at 822-39 (describing the costs when contracting parties prioritize front end or back end costs at the expense of the other).

(189.) The presence of a focal point seems to explain the market's ability to coordinate within several years to reach the unanimity required among investors to revise collective action clauses. See generally Choi el al., supra note 15 (discussing focal points as a possible means of encouraging parties to coordinate).

(190.) See, e.g., Vincent Crawford, A Survey of Experiments on Communication via Cheap Talk, 78 J. Econ. Theory 286, 287 (1998) ("[W]hen players' preferences are sufficiently close, communication via cheap talk can be informative.").

(191.) Thomas C. Schelling, The Strategy of Conflict 54-55 (1980).

(192.) In short, the issuers faced a tradeoff: They could either insert language shutting down the hedge funds who might buy up the bonds in times of future distress, but in doing so their action could be construed as watering down a pledge that purchasers had come to expect as a promise of equal treatment. Ex ante, the issuers did not know which move would affect the price of their bonds more. So they adopted a wail and see attitude. Eventually, once they had some assurance that most other sovereign issuers would change the language, meaning they would be at no competitive disadvantage, the big issuers at least agreed to a change.

(193.) One way to understand the stickiness of terms that make it on to the checklist is to see them as category defining terms. Scholars in sociology have long observed the importance attached, even by sophisticated market actors, to whether products fit certain defined categories--for example, a vehicle is not a car unless it has four wheels, and only if it is a car will it get rated by car magazines and evaluated by car experts and sold by car dealers and so on. See, e.g., Ezra W. Zuckerman, The Categorical Imperative: Securities Analysts and the Illegitimacy Discount, 104 AM. J. Soc. 1398, 1398-1406 (1999).

(194.) See Gulati & Scott, supra note 10, at ch. 10.

(195.) Id.

(196.) For discussions of the value of standardized provisions, see Goetz & Scott, supra note 5, at 286-88 and Kahan & Klausner, supra note 34, at 719-29.

(197.) For discussion of the three-and-a-half-minute transaction and the mass production of boilerplate contracts, see Gulati & SCOTT, supra note 10, at ch. 10. See also Richman, supra note 33, at 79-82 (drawing an analogy to Henry Ford's production line for cars).

(198.) The exception in this story should be the restructuring lawyers and perhaps their clients. These lawyers are going to face the consequences of having suboptimal contract terms. So, one should predict that when they have the opportunity to remedy terms after a restructuring, they will. And, as reported, we see precisely that--the bonds that have the first changes lend to be those in restructurings. See Choi et al., supra note 15, at 26.

(199.) Audio tape: Columbia Law School Meeting, supra note 146.

(200.) Gelpern & Gulali, supra note 168, at 11-12.

(201.) See Steffen Andersen, Glenn W. Harrison, Morten 1. Lau & E. Elisahet Rulstrom, Discounting Behavior: A Reconsideration, 71 Eur. Econ. Rev. 15, 15-16 (2014); Shane Frederick, George Loewenstein & Ted O'Donoghue, Time Discounting and Time Preference: A Critical Review. 40 J. Econ. Lit. 351, 393-94 (2002).

(202.) See Gulati & Scott, supra note 10, at 158-59; see also Michelle Celarier, Mitt Romney's Hedge Fund Kingmaker, Fortune (Mar. 26, 2012), [] (describing Elliott's strategy of figuring out what the documents actually say).

(203.) There are a number of recent research papers showing that key contract terms such as governing law come into play in near-default scenarios. See, e.g., Andrew Clare & Nicolas Schmidlin, The Impact of Foreign Governing Law on European Government Bond Yields 2-5 (Mar. 8, 2014) (unpublished manuscript), available at abstract_id=2406477 []; Julian Schumacher, Marcos Chamon & Christoph Trebesch, Foreign Law Bonds: Can They Reduce Sovereign Borrowing Costs? 1-5 (Mar. 1, 2015) (unpublished manuscript), [https://].

(204.) It is useful to compare a clause such as pari passu where the range of meanings is unaided by any context evidence with ordinary vagueness. For example, a contract that calls for the delivery of 100 "dark red" Macintosh apples is vague because the term does not precisely determine what constitutes a conforming apple; how dark red must the apples be? Nevertheless, in such a ease courts can revert to context to resolve the meaning of the contract term. Pari passu lacked any such context to aid in fixing its meaning.

(205.) One of the key policymakers from the U.S. Treasury told us in January 2016 that, looking back, persuading ICMA to be involved was especially important. ICMA was, at the time, trying to demonstrate to the market that it was a key actor and worth joining. ICMA, in other words, had its own incentives to show the market that it was an important player and could engineer big changes. But our data show that the ICMA publication of draft clauses, which were released officially in August 2014 but drafts of which were circulated as early as January 2014, was not enough to induce change on the market. The catalysis appears to have occurred in October 2014.

(206.) Within this group, the lawyers representing Latin American sovereigns were disproportionately represented as they tend to do issuances primarily under New York law.

(207.) Jody S. Kraus & Robert E. Scott, Contract Design and the Structure of Contractual Intent, 84 N.Y.U. L. REV. 1023, 1025 (2009).

(208.) Schwartz & Scott, supra note 1, at 568-69.

(209.) Merger clauses are given virtually conclusive effect in New York. See Tempo Shain Corp. v. Bertek, Inc., 120 F.3d 16, 21 (2d Cir. 1997) ("Ordinarily, a merger clause provision indicates that the subject agreement is completely integrated, and parol evidence is precluded from altering or interpreting the agreement."); Norman Bobrow & Co. v. Loft Realty Co., 577 N.Y.S.2d 36, 36 (N.Y. App. Div. 1991) ("Parol evidence is not admissible to vary the terms of a written contract containing a merger clause.").

(210.) The New York courts' plain meaning rule addresses the question of what legal meaning should be attributed to the contract terms that the parol evidence rule has identified: when words or phrases appear to be unambiguous, extrinsic evidence of a possible contrary meaning is inadmissible. For a discussion of the different parol evidence rules and their effect on contract drafting, see generally Alan Schwartz & Robert E. Scott, Contract Interpretation Redux, 119 Yale L.J. 926 (2010).

(211.) See Robert E. Scott, The Case for Formalism in Relational Contract, 94 Nw. U. L. Rev. 847, 853-56 (20(H)).

(212.) Goetz & Scott, supra note 5, at 286-88.

(213.) Contextualists argue that formal interpretive rules that exclude certain categories of extrinsic evidence deprive the fact finder of indispensable information relevant to deciding the case and thus can distort the court's assessment of what the parties meant by their agreement. Ronald J. Gilson, Charles F. Sabel & Robert E. Scott, Text and Context: Contract Interpretation as Contract Design, 100 Cornell L. Rev. 23, 36 (2014). Contextualist jurisdictions, such as California, carry this view to its logical limit and reject the notion that words in a contract can have a plain or unambiguous--context free--meaning at all. Id. By the same logic they favor a soft parol evidence rule. Id. at 37. Here the test for integration admits extrinsic evidence notwithstanding an unambiguous merger clause or, absent such a clause, notwithstanding the fact that the writing appears final and complete on its face. See Pac. Gas & Elec. Co. v. G.W. Thomas Drayage & Rigging Co., 442 P.2d 641, 645 (Cal. 1968) ("[R]ational interpretation requires at least a preliminary consideration of all credible evidence offered to prove the intention of the parties."); Maslerson v. Sine, 436 P.2d 561, 564 (Cal. 1968) (admitting parol evidence to vary terms of deed on ground that "[e]vidence of oral collateral agreements should be excluded only when the fact finder is likely to be misled").

(214.) A critic might ask whether what we call "layers of legal jargon" or "encrustation" is being dismissed unfairly as meaningless. After all, these encrustations, like the insertion of the word "payment" into the pari passu clause sometime in the 1970s, were presumably chosen by a lawyer to make the clause clearer or more advantageous for the client. If the variation was consciously chosen, is it not the very opposite of meaningless jargon? Our response is that the key is whether the underlying core clause has any shared meaning. If not, then adding language intended to clarify increases rather than reduces the linguistic uncertainty infecting the core clause.

(215.) Based on what we have learned, it probably does not mean much if we do not find a pricing effect. See Gelpern & Gulati, supra note 168, at 11-12. But if we do find an effect, it probably means that the clause was intended to have a specific discernible meaning.

(216.) See Restatement (Second) Contracts [section] 201(3) (Am. Law Inst. 1981). In the case where one party does not know the meaning of the other and the other does know or have reason to know the meaning of the first party, the term is interpreted in accordance with the meaning asserted by the first party. See id. [section] 201(2).

(217.) We are grateful to Lisa Bernstein for this observation.

(218.) Among the largest of these payouts are the recoveries of holdouts against Argentina in March 2016, with recoveries estimated in the 1000 percent range for some hedge funds. See Martin Guzman & Joseph E. Stiglitz, Opinion, How Hedge Funds Held Argentina for Ransom, N.Y. Times (Apr. 1, 2016), [].

(219.) E.g., Dimitra DeFotis, 2 Venezuela Bond Strategies as Debt Default Looms, Barrons: Emerging Mkts. Daily (Mar. 22, 2017), -strategies-as-debt-default-looms-1490202932 [

(220.) See Elena Carletti, Paolo Colla, Mitu Gulati & Steven Ongena, Pricing Contract Terms in a Crisis: Venezuelan Bonds in 2016, Cap. Mkts. L.J. 540, 546 (2016) (comparing the pricing on Venezuelan sovereign bonds with similar maturities, but differing pari passu and collective action clauses); see also John Dizard, 'Complete Chaos in Caracas' Invites Vulture Fest, Fin. Times (Sept. 2, 2016), [ 4XHZ-WGMG] (discussing how vulture funds have begun to find investing in Venezuela attractive).

(221.) See supra note 17.

(222.) See Stephen J. Choi, Mitu Gulati & Robert E. Scott, Contractual Arbitrage, in oxford Handbook on Global Governance (Eric Brousseau, Jean-Michel Galmant & Jerome Sgaard eds., forthcoming 2017) (manuscript at 21-22), viewcontent.cgi?article=6304&context=facully_scholarship [].

(223.) The Commonwealth of Puerto Rico's bonds, that are in distress as of this writing, provide an illustration of the range of these pledges. See Tim Worstall, The Truly Horrible and Ghastly Mess of Puerto Rico's Bond Issuance, Forbes (May 6, 2017), timworstall/2017/05/06/the-lruly-horriblc-and-ghastly-mess-of-puerto-ricos-bond -issuance/#7405bea54224 []. For a general overview of municipal bonds, see Craig L. Johnson, Martin J. Luby & Tima T. Moldogaziev, State and Local Financial Instruments: Policy Changes and Management 12-14 (2014) and Neil O'Hara, The Fundamentals of Municipal Bonds 36 (6th ed. 2012).

(224.) See Mark D. Robbins & Bill Simonsen, Municipal Securities, in 2 Encyclopedia of Public Administration and Public Policy 795, 797-98 (Jack Rabin ed., 2003).

STEPHEN J. CHOI, Murray and Kathleen Bring Professor of Law, New York University,

MITU GULATI, Professor of Law, Duke University,

ROBERT E. SCOTT, Alfred McCormack Professor of Law and Director, Center for Contract and Economic Organization, Columbia University.

For comments, thanks to Karen Alter, Aditi Bagchi, Joseph Blocher, Patrick Bolton, Lisa Bernstein, Lee Buchheit, Bruce Carruthers, John Coyle, Elisabeth de Fontenay, Merritt Fox, Anna Gelpern, Tracey George, Ron Gilson, Clay Gillette, Victor Goldberg, Bert Huang, Kieran Healy, Marcel Kahan, Kim Krawiec, Tom Lee, Tom Merrill, Stephen Nelson, Un Kyung Park, Eric Posner, Dave Pozen, Mark Ramseyer, Barak Richman, Charles Sabel, Chris Serkin, Alan Schwartz, Mark Weidemaier and participants at workshops at the European University Institute, Northwestern University (Political Science), and at Columbia, Duke, Fordham, Vanderbilt, and Washington & Lee Law Schools.

Caption: Figure 1: Percent of Issuances by All Issuers with Any Change

Caption: Figure 2: Dollar Amount of Issuances for All Issuers Excluding those Issuers with a Rank Initial Pari Passu Clause

Caption: Figure 3: Percent of Pure Sovereigns under Foreign Law with Major Change

Caption: Figure 4: Percent of Pure Sovereigns under Foreign Law with Minor Change

Caption: Figure 5: Percentage of Sovereign Issuances by Year with Specific Type of Pari Passu Clause
Table I: Types of Pari Passu Clause

Pari Passu Clause           Risk Level
Pay Equally and Ratably     High
Rank Equally in Payment     Medium
Rank                        Low
Mandatory Law               Zero
COPYRIGHT 2017 Duke University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Choi, Stephen J.; Gulati, Mitu; Scott, Robert E.
Publication:Duke Law Journal
Date:Oct 1, 2017
Previous Article:An administrative right to be free from sexual violence? Title IX enforcement in historical and institutional perspective.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters