Collective Action Clauses and the Restructuring of Sovereign Debt 9783110314526, 9783110314472

Table of contents :
Preface
The Authors
Introduction
The Euro Area’s Collective Action Clause – Some Questions and Answers
Private Sector Involvement in Sovereign Debt Restructuring: Model CACs for Europe
Identical Collective Action Clauses for different Legal Systems: A European Model
Towards a Level Playing Field
An Introduction to the Euro Area’s Model Collective Action Clause
Enfranchisement and Disenfranchisement in Collective Action Clauses
Collective Action Clauses and Litigation
Bondholder Resolutions in the Courtroom
Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States?
The Future of Sovereign Debt in Europe
CACs and the Restructuring of Sovereign Debt – How Would Markets React?
The Aggregation Clause in Euro Area Government Securities: Game Changer or Flavor of the Month? – Background and the Greek Experience
Sovereign Debt Restructuring: Lessons from History
A Resolvency Proceeding for Defaulting Sovereigns
Appendices
Appendix 1: Art. 12(3) ESM Treaty
Appendix 2: EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause
Appendix 3: EFC Sub-Committee on EU Sovereign Debt Markets Supplemental Provisions to Model Collective Action Clause
Appendix 4: EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause Supplemental Explanatory Note
Appendix 5: EFC Sub-Committee on EU Sovereign Debt Markets Draft Model Collective Action Clause (distributed for Comment in July 2012)
Appendix 6: EFC Sub-Committee on EU Sovereign Debt Markets Collective Action Clause Explanatory Note dated 26 July 2011

Citation preview

Collective Action Clauses and the Restructuring of Sovereign Debt ILFS

Institute for Law and Finance Series

Edited by Theodor Baums Andreas Cahn

Volume 12

Collective Action Clauses and the Restructuring of Sovereign Debt

Edited by Klaus-Albert Bauer Andreas Cahn Patrick S. Kenadjian

ISBN 978-3-11-031447-2 e-ISBN 978-3-11-031452-6 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. © 2013 Walter de Gruyter GmbH, Berlin/Boston Cover image: Medioimages/Photodisc Data Conversion: Werksatz Schmidt & Schulz GmbH, Gräfenhainichen Printing and Binding: Hubert & Co. GmbH & Co. KG, Göttingen ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Preface The volume contains articles based on presentations given at a conference hosted by the Institute for Law and Finance of Goethe University on October 27, 2011. Collective action clauses are an example of the typical dichotomy of financial regulation: While the problems are economic in nature, the solutions need to be implemented by law. The Institute for Law and Finance strives to bring together law and finance in order to foster a better mutual understanding of both disciplines and to improve the regulation of financial markets. Thus, the organizers are particularly pleased that eminent experts from the fields of law and finance agreed to participate in the event and to share their views on and experiences with collective action clauses. The presentations given at the conference have been updated in 2012 to reflect recent developments. Andreas Cahn

Table of Contents Preface   V The Authors 

 IX

Introduction Klaus-Albert Bauer The Euro Area’s Collective Action Clause – Some Questions and Answers   3 Private Sector Involvement in Sovereign Debt Restructuring: Model CACs for Europe Antonio Sáinz de Vicuña y Barroso Identical Collective Action Clauses for different Legal Systems: A European Model   15 Claus-Michael Happe Towards a Level Playing Field 

 25

David G. Sabel An Introduction to the Euro Area’s Model Collective Action Clause 

 29

Christian Hofmann Enfranchisement and Disenfranchisement in Collective Action Clauses   45 Collective Action Clauses and Litigation Lachlan Burn Bondholder Resolutions in the Courtroom 

 73

Boris Kasolowsky / Smaranda Miron Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States?   85

VIII 

 Table of Contents

The Future of Sovereign Debt in Europe Martin Wiesmann CACs and the Restructuring of Sovereign Debt – How Would Markets React?   103 Patrick S. Kenadjian The Aggregation Clause in Euro Area Government Securities: Game Changer or Flavor of the Month? – Background and the Greek Experience   113 Christian Kopf Sovereign Debt Restructuring: Lessons from History  Christoph G. Paulus A Resolvency Proceeding for Defaulting Sovereigns 

 149

 181

Appendices Appendix 1 Art. 12(3) ESM Treaty   209 Appendix 2 EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause   211 Appendix 3 EFC Sub-Committee on EU Sovereign Debt Markets Supplemental Provisions to Model Collective Action Clause  Appendix 4 EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause Supplemental Explanatory Note   227 Appendix 5 EFC Sub-Committee on EU Sovereign Debt Markets Draft Model Collective Action Clause (distributed for Comment in July 2012)   239 Appendix 6 EFC Sub-Committee on EU Sovereign Debt Markets Collective Action Clause Explanatory Note dated 26 July 2011   251

 225

The Authors

Dr. Klaus-Albert Bauer Klaus-Albert Bauer is a partner of Freshfields Bruckhaus Deringer specializing in banking and capital markets law. He is based in Frankfurt. Klaus-Albert Bauer completed his legal education at the universities of Würz­ burg, Geneva, Tübingen as well as at Columbia University in New York from where he holds a master of laws degree (LL.M.). In addition, he holds the degree of doctor of law (Dr.jur.) from the University of Tübingen. From 1979 to 1984, he worked as a research assistant at the Max-Planck-Institute for Foreign and International Patent, Copyright and Competition Law in Munich. In 1984, he joined Westrick & Eckholdt, a predecessor firm of Freshfields Bruckhaus Deringer. In the mid 1990s he practised in his firm’s Tokyo office and opened its Moscow office. From 2003 until 2007 he was Office Managing Partner of Freshfields Bruckhaus Deringer’s Frankfurt office. Klaus-Albert Bauer is the author of a book on international banking supervision as well as of numerous other publications. He has been teaching international business transactions at Bucerius Law School. Since its foundation, KlausAlbert Bauer is a member of the faculty (Lehrbeauftragter) at the Institute for Law and Finance at the Goethe-University Frankfurt (ILF). For purposes of disclosure it is stated here that Freshfields Bruckhaus Deringer advised the German Ministry of Finance in connection with the drafting of the model CAC. The views expressed by Klaus-Albert Bauer in his article are entirely his own. Lachlan Burn Lachlan Burn is an English solicitor and a partner of Linklaters, which he joined in 1974. He specialises in capital markets work, including debt, equity and derivative securities. He is a member of the Primary Markets Group of the London Stock Exchange and the Legal and Documentation Committee of the International Capital Markets Association. He was formerly a member of the Bank of England’s Legal Risk Review Committee from 1992 to 1994, the Financial Markets Law Committee from 2002 to 2006 and the United Kingdom Listing Authority’s Advisory Committee from 1999 to 2011. He is co-editor of the Capital Markets Law Review, published by Oxford University Press. Prof. Dr. Andreas Cahn Andreas Cahn studied law at the Johann Wolfgang Goethe-University Frankfurt/ Main and at the University of California at Berkeley, where he earned an LL.M. After his Second State Examination in Frankfurt he worked for 6 years as a research assistant at the University of Frankfurt. During this period of time he wrote his doctoral thesis on problems of managers’ liability (published in 1996)

XII 

 The Authors

as well as his post-doctoral thesis on legal aspects of intra-group financing (published in 1998). In 1996 he took up the Chair of Civil Law, Commerce Law and Corporate Law at the University of Mannheim. Since October 2002 he is Director of the Institute for Law and Finance at Goethe-University in Frankfurt. He has published extensively on corporate law, capital markets law, the law of products liability, general civil law as well as on civil procedure. He is co-publisher of “Der Konzern”, a law journal focusing on company law, taxation and accounting of corporate groups, of “Corporate Finance law”, a journal with a focus on current legal issues corporate finance, co-editor of the Institute for Law and Finance Series and member of the editorial board of the law journal “European Company Law”. Dr. Claus-Michael Happe Claus-Michael Happe is Head of Division in the German Ministry of Finance, responsible for the Paris Club and all international debt related questions. His portfolio also covers the International Development Banks and other debt related initiatives in the G 20 context. He has studied law in Berlin and Münster and practiced in several disciplines of law in German Federal Ministries, the European Commission and a private Law Firm. Prof. Dr. Christian Hofmann Christian Hofmann is a professor of law at the Private University in the Principality of Liechtenstein (UFL). Prior to this appointment (and at the time of the conference), he was a senior legal counsel for Deutsche Bundesbank. His former academic positions include visiting research positions at NYU and UC Berkeley and teaching positions at Humboldt University Berlin, Goethe University Frankfurt and University of Cologne. His teaching and research focus is in banking and company law in which he has published numerous papers and authored, edited and contributed to several books. Boris Kasolowsky Boris Kasolowsky is a partner in Freshfields Bruckhaus Deringer’s international arbitration group. He is based in Frankfurt, having previously practised in Freshfields’ London and Vienna offices. He regularly appears as counsel and sits as arbitrator in commercial arbitrations concerning complex financial instruments, joint venture and post-M&A disputes, and infrastructure and oil and gas projects. He also represents and advises governments and commercial entities on disputes under relevant bilateral and multilateral investment treaties, including under the Energy Charter Treaty. Boris Kasolowsky holds a law degree from Oxford University, a master’s degree from the School of Oriental and African Studies, London



The Authors 

 XIII

University, and a doctorate from Hamburg University. He is qualified as a solicitor (England and Wales) and German Rechtsanwalt. He speaks English, German, French and Arabic. Patrick S. Kenadjian Patrick S. Kenadjian is currently an Adjunct Professor at the Goethe University in Frankfurt am Main, Germany, where he teaches courses on the financial crisis and financial reform and comparative public mergers and acquisitions at the Institute for Law and Finance. He speaks frequently on topics related to financial reform, including too big to fail, the architecture of financial supervision and the new regulatory environment in the US and the EU. Mr. Kenadjian is also Senior Counsel at Davis Polk & Wardwell, LLP in their London office. He was a partner of the firm from 1994 to 2010, during which time he opened the firm’s Tokyo and Frankfurt offices in 1987 and 1991, respectively and spent over 25 years in their European and Asian offices. His practice includes cross-border securities offerings, especially for financial institutions, mergers and acquisitions, privatizations and international investments and joint ventures, as well as general corporate advice, with an emphasis on representing European clients. He has been active in securities transactions for issuers in Asia and Europe, particularly on initial public offerings and privatizations in Germany, Austria, Italy and Switzerland. He has represented bidders and targets in cross-border acquisitions throughout Europe, in particular in France, Germany, Italy, Switzerland, the United Kingdom and the United States. Mr. Kenadjian has also represented European and Asian issuers in U.S. debt private placements. He speaks French, German and Italian. Christian Kopf Christian Kopf is a partner in the London office of Spinnaker Capital, an Emerging Markets investment management firm. He is in charge of economic research and investment strategy, co-ordinates macro analysis carried out at Spinnaker’s offices in Dubai, Hong Kong, London, São Paulo and Singapore and chairs the firm’s macro meeting. Christian Kopf has spoken on various aspects of international finance at meetings of the Brookings Institution, the Committee on the Global Financial System, EBRD, EMTA, Euro50 Group, G-20, IFF, IIF, the Paris Club, and at several industry conferences. He authored papers on sovereign debt that were published by Deutsche Bank Research and by the Centre of European Policy Studies. From 1999 to 2006, Christian Kopf was a senior portfolio manager at DWS in Frankfurt, Germany, responsible for Global Emerging Markets Fixed Income. Prior to working in the financial industry, he conducted sustainability research at the Wuppertal Institute for Climate, Environment and Energy.

XIV 

 The Authors

Christian Kopf graduated from Witten/Herdecke University in Germany with a Masters degree in Economics and holds the Chartered Financial Analyst designation from the CFA Institute. He was a scholarship holder of Evangelisches Studienwerk, the Protestant institute for the promotion of academic talent, and of the Congress-Bundestag Youth Exchange Program. Christian Kopf grew up in Hessen, Colombia, Minnesota and France and he is fluent in German, Spanish, English and French. Smaranda Miron Smaranda Miron is a Romanian-qualified lawyer. She received an LLB from the Alexandru Ioan Cuza University of Iasi, Romania and holds an LL.M in European Economic Law from Europa Institut, Saarland University, Germany. Currently, she is pursuing an external LL.M in International Dispute Settlement (focusing on human rights and the law of treaties) with the University of London. At the time of writing this article, Smaranda was based in Frankfurt and was working in Freshfields Bruckhaus Deringer LLP’s International Arbitration Group. During her time at Freshfields, she has been dealing with both investment and commercial disputes, conducted under various arbitration rules, such as ICSID, UNCITRAL, LCIA and PCA. She speaks Romanian, English, German and basic French. Prof. Dr. Christoph Paulus Christoph Paulus is professor of law at the Law School of the Humboldt Universität zu Berlin, Germany, a position that he has held since 1994. He wrote his Dissertation (Dr. iur.) and his Habilitation at the University of Munich; and he earned an LL.M. from the University of California at Berkeley in 1984. As a Feodor Lynen Fellow of the Alexander v. Humboldt-Stiftung he had been at UC Berkeley in 1989 and 1990. Since 1998, he has served several times as a Consultant to the International Monetary Fund in Washington, D.C., thereby inter alia preparing the IMF’s brochure on “Orderly & Effective Insolvency Procedures”. Additionally, in 2006, he has been appointed as a Consultant of World Bank in Washington, D.C., regarding inter alia insolvency laws and legislation. From November 2006 through November 2011, he served as Adviser of the German delegation for the UNCITRAL deliberations on group insolvency law and further topics. He is member (and currently one of the directors) of the International Insolvency Institute, of the International Association of Procedural Law, of the American College of Bankruptcy, of the International Academy of Commercial and Consumer Law and – as an extraordinary member – of the Instituto Iberoamericano de Derecho Concursal.



The Authors 

 XV

He has held guest professorships i. a. in Paris (Panthéon-Assas) / France, Cape Town / South Africa, Fukuoka / Japan, Brooklyn School of Law / USA, University of Sydney and Tongji University in Shanghai / China. Since 2009, he is one of the directors of the Institut für Interdisziplinäre Restrukturierung (Institute for interdisciplinary restructuring) at his university. David G. Sabel David G. Sabel is a partner at Cleary, Gottlieb, Steen & Hamilton LLP, based in their London office. Mr. Sabel was previously resident in the firm’s New York and Brussels offices. Mr. Sabel has for many years been the principal outside legal adviser to the Government of the Russian Federation with respect to financial matters. Mr. Sabel was the partner in charge of the Cleary Gottlieb team that assisted the Russian Government in the restructuring of the external debt of the former Soviet Union for which the Russian Government agreed to be responsible (Paris Club, London Club and foreign trade debts), and later assisted the Russian Government in restructuring its domestic debt following the Russian financial crisis of 1998. More recently, Mr. Sabel has advised the Russian Government with respect to its Eurobond issuances and sovereign wealth funds. Cleary Gottlieb Steen & Hamilton LLP advised the EFC Sub-Committee on EU Sovereign Debt Markets on the drafting of the model CAC. The views expressed by Mr. Sabel in this article are solely his own. Mr. Sabel would like to thank David Billington and Matthew Hamilton-Foyn of Cleary Gottlieb for their assistance in the preparation of this article. Antonio Sáinz de Vicuña Antonio Sáinz de Vicuña has been General Counsel of the European Central Bank since 1998. He is the Chairman of the Legal Committee (LEGCO) of the ESCB. He is a graduate in Law and in Economic Sciences of the Universidad Complutense de Madrid, and has a Master’s in International Law from Cambridge University. He entered the Government Legal Service in 1974, and was subsequently posted as legal counsel in the ministries of Finance, Economy and Foreign Affairs, where he became Chief Legal Adviser in 1985. After a five-year period in a commercial bank, in 1994 he joined the European Monetary Institute as General Counsel. He is the author of one book and of various publications in the field of community, international and banking law. He is Chairman of the European Financial Market Lawyers Group, Vice-President of the Committee on International Monetary Law (MOCOMILA), and a member of the Advisory Board of the European Capital Markets Institute, of the High Level Group on Banknote counterfeiting, and of the Institute for Law and Finance of the Goethe University of Frankfurt. He was

XVI 

 The Authors

formerly a member and rapporteur of the Working Group of Legal Experts of the European Monetary Institute. Martin Wiesmann Martin Wiesmann (47) joined J.P. Morgan in July 2007 as Managing Director and Member of the Executive Board of JPMorgan Chase Bank, Frankfurt. Since August 2012, he is Head of Investment Banking Germany, Austria. Martin Wiesmann is responsible for a broad portfolio of corporate clients and J.P. Morgan’s business with the Public Sector. He has gained 20 years of experience in the banking industry. Prior to joining J.P. Morgan, he worked for 12 years in the investment banking business of Deutsche Bank in Frankfurt. Mr. Wiesmann has worked on a large number of M&A and capital market transactions in the transport, industrials and financial institutions sector, and he has gained broad expertise in the area of privatisation. Mr. Wiesmann holds a Master of Arts in Political Science from the University of Bonn, a Certificate of Political Studies of the Institut d’Etudes de Paris (“Science Po”) and an MBA from the Josef Katz School of Business in Pittsburgh.

Introduction

Klaus-Albert Bauer

The Euro Area’s Collective Action Clause – Some Questions and Answers During much of the time prior to the ILF conference on “Collective Action Clauses and the Restructuring of Sovereign Debt” held on October 27, 2011, the focus of public attention, was on the second part of the title – “Restructuring of Sovereign Debt”. Almost on a daily basis, the destiny of the Euro Area and of individual countries with high volumes of sovereign debt was discussed. Markets were volatile. In the morning hours of the day of the ILF Conference, a deal was made in Brussels providing for a 50 % “haircut” for private investors in Greek bonds on a voluntary basis.1 Many slides had to be updated virtually in the last minutes preceding the start of the Conference. At the time of writing these lines (mid December 2012), there is relative calm “out there”. Two major international agreements, the Treaty establishing the European Stability Mechanism (ESM Treaty) and the Treaty on Stability, Coordination and Governance in the Economic and Monetary Union (Fiscal Compact) are in place.2 Greece has gone through a successful restructuring of its bond debt including a recent buy-back programme. This is a good time to take a breath and focus more closely on the Collective Action Clauses which gave rise to our conference in the first place. To kick our book off, here are some questions and some answers suggested by this co-editor.

What are Collective Action Clauses? Generally, one understands collective action clauses (CACs) to refer to a clause in bond terms and conditions which provides inter alia for rules of majority voting to change the terms and conditions themselves. CACs could therefore, for the sake of simplicity and ignoring possible other features of a collective action clause, also be referred to as “majority voting provisions” or, keeping their most important purpose in mind, as “debt rescheduling clauses”. The effect of a bondholder vote passed with the required majority is to bind all holders of the respective bonds whether they have participated in the vote or not.

1 Cf. the details given in Christian Kopf’s contribution to this volume, p. 149 below. 2 The ESM Treaty has entered into force on September 27, 2012. As of December 14, 2012 eleven countries had ratified the Fiscal Compact.

4 

 Klaus-Albert Bauer

Are Collective Action Clauses new? No. They have been used in many corporate bond issues since the 19th century (with English law being the front runner). Many emerging market sovereign bond issues under English law and (since 2003) under New York law contain collective action clauses.

Have Euro Area sovereign bonds traditionally included CACs? No. The vast majority of presently outstanding sovereign debt issued by Euro Area countries contain no CACs. There are, however, exceptions. In 2003, following the G-10 recommendations, the President of Ecofin had supported the use of CACs for international bond issues. In fact, a number of foreign currency bonds or bonds under foreign law issued by many European countries during the past decade contain collective action clauses.

Are the collective action clauses used in the market today uniform? No. Many sovereign issues tend to follow one of the existing model provisions for collective action clauses, in particular the model clauses recommended by the Group of Ten in 2002 or the IPMA recommendations of 2004. The actual provisions of CACs used in the market differ to a great extent.

What happens after January 2013? After this date all newly issued sovereign bonds in the Euro Area with a maturity exceeding one year are to contain identical collective action clauses. It is important to note, however that the new rules only apply to central government issued securities and exclude regional and municipal bonds which may be issued without inclusion of CACs.

What is the legal basis for this? In Article 12 of the ESM Treaty all Euro Area member countries have undertaken to include collective action clauses with identical legal effect in their sovereign bond issues. The actual text of the clause is not attached to the Treaty itself. It was



The Euro Area’s Collective Action Clause – Some Questions and Answers 

 5

developed by the EFC Sub-Committee on Sovereign Debt Markets and approved by the EU’s Economic and Finance Committee (EFC) on November 18, 2011.3

How long will it take until all outstanding sovereign bonds in the Euro Area contain CACs? Sovereign bonds with and without CACs will co-exist for a very long time.4 First, all existing bonds (some of which have a maturity of 30 years) are not affected by the introduction of the Euro CACs. Second, even for newly issued bonds there is a phasein corridor for cases where the aggregate principal amount of a series of bonds existing prior to January 1, 2013 (with no CACs) is increased (so called “tapping”).5

What is the purpose of the Euro Area CAC? There seems to be a double purpose. First there is the general goal of any collective action clause, i.e. to facilitate restructuring of an issuer’s debt. The logic is that if a defined majority accepts the restructuring, then it should also be binding for everyone, thereby eliminating the so-called “hold-out problem”. Second, in a European context, the inclusion of the collective action clause was part of what has come to be known as Private Sector Involvement (PSI). In plain language: Bilateral help (rescue measures) between sovereign states within the Euro Area was conditional on a bundle of measures which should ensure that private investors recognize and bear the risk of their investment decisions. If one follows this logic, collective action clauses would also have a “disciplinary function”.6

What is the main content of the Euro Area CAC? The obligations of an issuer to pay principal and interest can be changed with binding effect for all holders by a bondholder resolution. This resolution can be taken in a bondholder meeting or in writing. Here are the details for the restructuring of a single series of bonds:

3 For the final text of the Euro Area CAC, for supplemental provisions thereto and for an earlier draft see Appendices 2, 3 and 5 below. For an overview of the drafting process see David Sabel’s contribution in this volume at p. 29 below. 4 See Wiesmann p. 103 below. 5 For details cf. Supplemental Explanatory Memorandum, Appendix 4, below. 6 Cf. the contributrion of Claus Happe in this volume, p. 25 below.

6 

 Klaus-Albert Bauer

Bondholder Meeting

Written Resolution

Majority Requirement

75 % of principal amount of outstanding bonds repre­ sented in the meeting

66 2/3 % of principal amount of all outstanding bonds of the series concerned

Quorum (minimum participation in meeting required, calculated as percentage of principal amount of all outstanding bonds of the series concerned)

66 2/3 %

n.a.

If more than one issue of bonds is to be restructured at the same time, a lower per issue majority as set out here Bondholder Meeting

Written Resolution

Majority Requirement

66 2/3 % of principal amount of outstanding bonds represented in the meeting

50 % of principal amount of all outstanding bonds of the series concerned

Quorum (minimum participation in meeting required, calculated as percentage of principal amount of all outstanding bonds of the series concerned)

66 2/3 %

n.a.

is sufficient for a particular series concerned if, on an aggregate basis over all different series included in the restructuring, the following majority is reached Bondholder Meeting

Written Resolution

Majority Requirement

75 % of the aggregate principal amount of outstanding bonds represented in all bondholder meetings of the series of bonds included in the restructuring

66 2/3 % of the aggregate principal amount of all outstanding bonds of all series of bonds included in the restructuring

Quorum (minimum participation in meeting required, calculated for each series separately as percentage of principal amount of all outstanding bonds of the respective series)

66 2/3 %

n.a.



The Euro Area’s Collective Action Clause – Some Questions and Answers 

 7

The legal technique to provide for a lower per-issue minimum majority in case of a bondholder vote encompassing several series of bonds has come to be known as aggregation or, in the context of the Euro Area CAC, as cross-series modification.7 The issuer may decide whether it proposes one or several single issue restructurings or one or several restructurings on an aggregate basis. The Euro CAC also deals with the issue of who is entitled to vote in order to avoid corrupting the outcome by having the issuer or certain holders close to the issuer participating in the vote.8

Does the Euro Area CAC solve the “hold-out problem”? To a certain extent but not entirely. An investor or a group of single investors holding 50 % of the outstanding principal amount of a particular bond issue would have a “secure” position to block a restructuring proposal for the respective series even if the restructuring is based on the cross-series modification/aggregation provisions and, over all series of bonds concerned, the required majorities under the aggregation provisions were reached. If a restructuring is attempted on a single issue basis, 33 1/3 % of the outstanding principal amount would give a “holdout” creditor a “secure” blocking position.

Does the Euro Area CAC give better protection against hold-out creditors than collective action clauses traditionally used in the market? Arguably yes. As set out above, many collective action clauses used by sovereign issuers in some of their foreign law issues follow the standards of the G-10 recommendations or the IPMA recommendations of 2004. All these model clauses provide for majority voting on a per issue basis, i.e. with no reduced majority requirements in case of a restructuring of more than one bond issue. For debt restructurings, both the G-10 recommendations and the IPMA recommendations require a majority of 75 % of outstanding principal amount of bonds of a particular series. In other words, a holder of 25 % of the outstanding principal amount of bonds of a series would always be able to block a restructuring resolution concerning the particular series of bonds. The Euro CACs therefore seem to afford

7 For a detailed discussion of these clauses see the contributions of David Sabel and Patrick Kenadjian in this volume on pp. 29 and 113, respectively. 8 For a detailed analysis of the “disenfranchisement” provisions included in the Euro Area Model CAC and their rationale see the contributions by Christian Hofmann and David Sabel in this volume, pp. 45 and 29 below.

8 

 Klaus-Albert Bauer

better restructuring possibilities by requiring lower per bond issue majorities in case of written resolutions and, most importantly, in case of cross-series modifications, thereby raising the threshold of a “secure” blocking position from 25 % to 50 % of the outstanding principal amount of bonds of a particular series.

Did the Greek restructuring of February/ March 2012 use CACs? Yes. Part of Greece’s outstanding debt under foreign law had a collective action clause included. Some of these issues were restructured following a bondholder resolution reaching the required majorities. In other issues, an amendment of terms was either not sought or not backed by a sufficient number of votes. As for Greek law bonds, they originally had no CACs included. The Greek legislator had “retrofitted” CACs into existing bonds on February 23, 2012 (see next question) which were then used to restructure the Greek law debt.9

Were the Greek retrofit CACs for Greek law bonds the same as the Euro Area Model CACs? No. The Greek retrofit CACs were very simple. The Greek Bondholder Act of February 23, 2012 provided for a full aggregation of votes across all series of bonds without any counting of votes on a per-issue basis. The overall quorum was 50 %, and the majority required was 66 % of the aggregate nominal amount.

What was the result of the vote of Bondholders for the restructuring proposal concerning Greek law bonds with “retrofit” CACs? In the actual vote, the retrofit CAC thresholds were easily achieved: of some 177 bn. Euro principal amount outstanding, 161 bn. Euro participated in the vote (91 %) of which 152 bn. Euro (94 %) voted in favour of the restructuring proposal. As a consequence, 100 % of the Greek law debt included in the proposal was restructured.

9 For a detailed description of the Greek restructuring see Jeromin Zettelmeyer, Christoph Trebesch and Mitu Gulati The Greek Debt Exchange: An Autopsy, Draft 11 September 2012, available at http://ssrn.com/abstract=2144932; cf. also the contribution by Patrick Kenadjian in this volume, p. 113 below.



The Euro Area’s Collective Action Clause – Some Questions and Answers 

 9

If Greece could legislate CACs “ad hoc” into all outstanding Greek law bonds and complete a successful restructuring on this basis, why is it necessary to equip all new European government securities issued after January 1st, 2013 with CACs? In fact, not everybody agrees on the necessity of having the Euro Area CAC. Christian Kopf argues in this volume that the Euro Area CAC is not necessary as foreign law bonds by Euro Area member states tend to already have CACs included in the terms and conditions and as the overwhelming majority of Euro Area member states’ debt is governed by local law which would allow “Greek-style” retrofit CACs should they ever be needed. Claus Happe and Matthias Wiesmann take a different view as they stress the disciplinary function of the Euro Area CAC. As to Christian Kopf’s premise, namely that a national legislator can introduce CACs with effect for outstanding bonds at any time, see next question.

Can a national legislator introduce (“retrofit”) CACs for its outstanding bond issues governed by its own law at any time? Arguably yes. There are, however, limits as to what a legislator can “legally” do. These limits could be found in the sovereign’s constitution, in bilateral investments treaties (BITs), in multi-lateral conventions (e.g. human rights) and in general principles of international law. In this volume, Lachlan Burn10 forcefully argues that “Greek-style” retrofit CACs violate the rule of law. Boris Kasolowsky and Smaranda Miron11 point out that many BITs may provide protection against retrofit CACs. A recent article in the Harvard Business Law Review examines under which circumstances retrofit CACs might qualify as an expropriation which would give rise to claims in a US court.12 Finally the constitutional issue would need to be carefully explored on a case by case basis. So, stopping short of trying to give a concise answer to the question, we surmise that the Euro Area’s approach of providing new sovereign

10 p. 73 below. 11 p. 85 below. 12 See Melissa Boudreau Restructuring Sovereign Debt Under Local Law: Are Retrofit Collective Action Clauses Exproprietary, Harvard Business Law Review (Online Edition) available at http:// www.hblr.org/2012/05/retrofit-collective-action clauses/See also Christian Kopf, p. 173 below. For an ex ante analysis of the Greek situation see Mitu Gulati/Lee Buchheit How to Restructure Greek Debt, Duke Law Working Papers, No. 47, Duke University (2010).

10 

 Klaus-Albert Bauer

debt securities with collective action clauses is certainly more “civilized”13 than ignoring the issue and legislating post factum.

Once the Euro Area CAC is in use, could a national legislator change it unilaterally in case of need? Following the Kopf / Gulati analysis set out above, the answer seems to be yes. So, in a number of years, a state wishing to restructure its debt might try to change the rules if it finds that a “Greek-style” retrofit full aggregation would increase the chances of success for a restructuring. As we have seen, under the Greek-style full aggregation approach no blocking position for any particular bond issue would work if a certain (rather low) overall majority of bondholders approved the proposal. There are, however, two important considerations: first, while a retrofitting introduction of collective action clauses may raise eyebrows it would be an even stronger breach of investor confidence to unilaterally change the rules after they have been explicitly set out in the terms of the bonds. Second, Article 12 of the ESM Treaty obliges the Euro Area member states to issue all bonds with identical collective action clauses. A natural understanding of this clause would seem to prohibit not only the issue of bonds with non-conforming collective action clauses but also the later unilateral amendment of Euro Area CACs to suit a particular issuer’s needs in times of crisis.

If an issuer issues otherwise identical bonds including the Euro Area Model CAC under its own law and under the law of another jurisdiction, will bondholders be treated in the same way in case of a restructuring? Basically, this is just another way to ask the preceding question once more. If a proposal for a debt restructuring is made to all bondholders, the Euro Area CAC requires counting of votes on a per issue basis. Therefore, depending on the outcome of bondholders’ votes, the destiny of each bond may be different. If the same percentage of holders votes in favour of the proposal for each of the bonds concerned, the result will be the same for each issue. However, if the state in question decides to change the rules for its domestic law governed bonds in its favour, the results may be different.

13 Antonio Sáinz de Vicuña’s article (see p. 15 below) emphatically describes insolvency law as “a triumph of civilisation over force”.



The Euro Area’s Collective Action Clause – Some Questions and Answers 

 11

Concluding Remark Just a few years ago, very few people would have foreseen the present crisis of sovereign debt in Europe. At the same time it would have been hard to imagine that, in responding to the crisis, all countries of the Euro Area would commit to using uniform collective action clauses in their debt issuances. Having gone through this experience, one is well advised to exercise modesty when it comes to predicting what lies ahead. Antonio Sainz de Vicuna, in his contribution below,14 sees collective action clauses as an interim step towards a new order which would eventually include a sovereign debt resolution mechanism.15 While it may take rather long to go down that road, this author also believes that the Euro CAC is an important first milestone on the way to a new governance for the Eurozone.

14 p. 15 below. 15 See also Paulus p. 181 below.

Private Sector Involvement in Sovereign Debt Restructuring: Model CACs for Europe

Antonio Sáinz de Vicuña y Barroso

Identical Collective Action Clauses for different Legal Systems: A European Model1 Introduction Insolvency law is – within the long history of mankind – a relatively modern development, a triumph of civilisation over force. Since the birth of Man until the 19th century laws would protect creditors and would provide him with powers to obtain the due repayment by his debtor. In ancient Greece, a creditor would have the right to enslave its defaulting debtor together with his wife, children and servants and use the proceeds of their work to achieve reimbursement. Similarly, Roman law provided for the possibility to throw into prison a debtor unless he would transfer his assets to settle with his creditor (through the contracts named “cessio pro soluto”-actual payment- or “cessio pro solvendo”-assignment of future income to the creditor-). Some more humane relief for debtors were foreseen in the Torah, whereby such enslavement of a debtor had a time limit of seven years where a Sabbatical relief was granted; or even more, in the Holy Quran Verse 280 stated “If someone is in hardship, let there be a postponement of his dues until a time of ease. But if you dispense him as a charity, then this is better”. The Roman Catholic Church provided immunity for those persons being chased from their home because of their debts when they sought refuge in religious places, and intermediated their debt with creditors. Overall, defaulting debtors were assimilated to thieves and robbers2 and already in medieval ages Courts could void operations realised by the debtor to protect his assets and economically survive when a default was forthcoming (i.e. a kind of modern “suspect period”). Only in the 19th century Europe did start to see bankruptcy statutes in several nations3, starting with the Napoleonic Code de Commerce of 1808, which

1 This article is based on a conference given in the House of Finance of the W. Goethe University of Frankfurt am Main in October 2011. It represents the personal views of the author and not necessarily those of the European Central Bank. The PowerPoint presentation is available at http:// www.ilf-frankfurt.de/uploads/media/Sanz_de_Vicuna_-_CAC_-_European_Model_03.pdf. 2 Castilian Act 1502: defaulting debtors were “públicos ladrones y verdaderos robadores.” Novísima Recopilación de las Leyes de España (1831). 3 In a few isolated cases, already in the 18th century some judicial procedure statutes gave jurisdiction to facilitate arrangements between debtor and creditors. E.g. the English Act 1705 of Queen Anne, giving powers to judges to adjudicate based on agreements of a super-majority of creditors.

16 

 Antonio Sáinz de Vicuña y Barroso

addressed the organisation of an insolvency situation under judicial control and with a ranking of credits. This is what has evolved until this day, perhaps with an overall shift from being initially targeting an orderly liquidation of the debtor’s estate towards the more contemporary idea that creditors are best served if the debtor is allowed to continue his business under creditors’ control. Common feature of early insolvency European acts in the 19th century was the transfer of the control of the debtor’s assets to the creditors, frequently accompanied with the judicial prohibition to the debtor to continue his trade, do any other trade, or borrow money (so-called “mort civile”), and with some penal consequences that hindered forever the reputation of the insolvent trader (“stigma”)4. Turning to the case of debtors who are sovereigns, one could see a certain parallelism in the evolution of the Law of debt default… but with a time gap of not less than two centuries: whilst modern bankruptcy procedures are in place for ordinary debtors in all European jurisdictions since the beginning of the 19th century, there is still today a legal lacuna with regard to sovereigns, both regarding their own ‘domestic’ public sector indebtedness and vis-à-vis foreign creditors. Up to this day, 21st century, there is not a general bankruptcy procedure for sovereign debtors. However, sovereign defaults have existed, both within Europe and outside. What was the treatment of such cases? For a long period of time that extends itself until the first half of the 20th century, sovereigns defaulting to foreign creditors were subject to measures that could compare with the case of private debtors before the 19th century: legitimate coercion by creditor countries against the debtor. Indeed, failing to pay meant possible use of force by the lending nation. Some cases suffice to prove this: –– In 1773 the Royal Navy blockaded the harbour of Boston, Colony of Massachusetts, to collect the taxes due by such port on the tea imported into it, giving rise to the famous Boston Tea Party conflict that inter alia eventually lead to the independence of the United States from Britain; –– In 1850 the Royal Navy blockaded the port of Piraeus in Greece, to collect the moneys owed to the British by the Greek State; –– In 1902 the Royal Navy together with the German Kaiserliche Marine blockaded the Venezuelan ports of La Guaira, Maracaibo and Puerto Cabello to

4 E.g. In England, until 1914 traders becoming insolvent were imprisoned as a guarantee of avoiding any influence of the debtor on the treatment of his assets, but also to force him to deliver all his assets for the insolvency estate, and until investigations on his conduct were finalised. Most European countries had abandoned imprisonment for insolvency already in the 1870s.



Identical Collective Action Clauses for different Legal Systems: A European Model 

 17

recover the amounts of Venezuelan defaulted debts to English and German lenders (some $166 million), and with Germany asking for the transfer to it of the island of Margarita as settlement, in order to establish there a naval base. Following US intervention (with a flotilla of US Navy under Admiral Dewey ready to intervene), an agreement was reached in Washington5 whereby against a nominal debt reduction and a maturity rescheduling Venezuela pledged 30% of Venezuelan Customs income as collateral for the debt repayment; such agreement entailed that such blockade was lifted by February 1903. In 1904 the US Congress adopted the so-called “Roosevelt Corollary to the Monroe Doctrine” according to which the US would have a right to intervene in cases of inability of American countries to pay their international debt, and in order to preclude actions in America of European powers. –– To ensure repayment of existing debts, the USA took military control of the Customs of the Dominican Republic in 1905, of Cuba in 1906, of Nicaragua in 1911, in Haiti in 1915; it took control of the Panamanian Treasury in 1918 and of el Salvador in 1921; in 1926 the USA imposed on Honduras a special ‘export tax’ pledged to the repayment of US loans to that country; also in 1926 Peru and Bolivia were subject to the intervention of an American Agent to monitor their customs income and the public finances to ensure repayment of outstanding loans; the US military occupations of Cuba (1917–1922), Haiti (1915–1934) and the Dominican Republic (1916–1924) had several aims but including the wish to ensure due repayment of debts owed to American lenders. –– Failure to pay in 1922 the War Reparations imposed on Germany in the Treaty of Versailles, caused France and Belgium in January 1923 to send their armies to Germany and occupy the Ruhr valley, so that the coal, iron, steel and timber produced there would be confiscated by French and Belgian commissioners and dedicated to the repayment of existing debt. Such occupation – which caused 130 German deaths, plus civil unrest in Germany, including the Munich Brauhaus putsch of Hitler – lasted until 1925 when a rescheduling of the debt was agreed thanks to the mediation of the USA (“Dawes Plan”). An improvement into a more civilised manner to obtain repayment from a sovereign debtor came in the 1920s thanks to the establishment within the League of Nations of a Financial Committee6, empowered to broker sovereign debt resched-

5 So-called “Washington Protocols 1903”. 6 The League of Nations considered the case of non-payment of the War Reparations by Germany and the Franco-Belgian invasion of the Ruhr, but abstained from action as it was consistent with the Treaty of Versailles.

18 

 Antonio Sáinz de Vicuña y Barroso

uling in Europe arising from the sovereign debt crisis that arose as a consequence of the end of the gold standard in the context of the Great War 1914–1918, whereby European powers exhausted their gold holdings as a consequence of the war. Cases such as Austria, Hungary, Bulgaria or Greece were settled through such Committee, whose powers were wide enough – e.g. – to order the establishment of a new independent central bank in Greece with an international statute (“Geneva Protocol of 15.9.1927”) that inter alia permitted the presence of non-Greek officials from creditor countries in crucial positions. Use of force against debtor countries was thus replaced by high-level diplomacy. In a further move towards an “institutionalised” method of peaceful renegotiation of sovereign debts, the second half of the 20th century saw the birth of the Paris Club7, an informal forum managed by the French Trésor where sovereign debts of countries under distress are re-negotiated by sovereign lenders, setting the parameters for corollary sovereign debt re-negotiations with private lenders in the so-called “London Club”, an ad hoc gathering of actual private lenders aimed at adapting their loans to the re-scheduling conditions agreed by the Paris Club. It is however in the 21st century that the international community took more decisive steps towards an articulation of legal mechanisms to renegotiate sovereign debts. In 2002, following the default of Argentina in 2001 and in view of the aggressive actions by holdout speculative creditors (so-called “vulture funds”) to exact from Argentina the full amount owed to them, impeding de facto further access to the market of the defaulting sovereign, the IMF decided to foster the worldwide use of “collective action clauses” (CACs), whereby sovereign debt issuance contracts would have to contain a mechanism whereby a supermajority of bondholders would be able to bind all of them. The same reasons that originated the bankruptcy provisions of the Napoleonic Code de Commerce two centuries before: a collective procedure to renegotiate debts or liquidate insolvent estates without allowing for disruptive individual actions by stand-alone creditors. The IMF also discussed in those years a possible bankruptcy framework for sovereigns: the Sovereign Debt Restructuring Mechanism8, whereby an international independent body would have by Treaty the jurisdictional power to organise the orderly settlement of sovereign debts under parameters similar to a

7 http://www.clubdeparis.org/ ; it is composed by 19 “first world” permanent members, plus 13 developing countries as ‘associated members’, plus a dozen IFIs and regional development banks, the European Commission and the UNCTAD, as observers. The Paris Club has reached 425 agreements with 89 different debtor countries since its establishment in 1956, for a total amount of $ 556 billion. 8 See http://www.imf.org/external/pubs/ft/exrp/sdrm/eng/index.htm



Identical Collective Action Clauses for different Legal Systems: A European Model 

 19

national bankruptcy court: listing and ranking of debts, control over the administration of debtor’s income, and decision-making capacity with binding effect worldwide. Such proposal failed within the IMF because it did not obtain sufficient support by its members9. But, in this writer’s view, it went into the same direction as general bankruptcy laws have gone in history.

The Collective Action Clauses in Europe Collective Action Clauses (CACs) were not an invention of the IMF: they existed already in the London bond market since 1879, as a civilised way to organise collective decision-making when there is a plurality of creditors, ensuring also that the collective interest dominates private – often insolidaire – interests. The IMF decided in 2002 to establish a “G-10 Working Group on Contractual Clauses” to recommend the use of CACs, for which the London practice served as a basis, to all international sovereign issuers10. Some demands to link the use of such CACs by sovereign issuers to eligibility conditions to IMF funding did not prosper. But their introduction into issuances placed in the New York and in European markets since 2003 was easily accepted by investors, who had seen the damage done to Argentina by the disruptive actions of holdout creditors (which have caused updates to the doctrine of sovereign immunity). The European Union explicitly supported in 2003 the use of CACs, but only for international issuances11. The IMF and the World Bank as well12. Many European countries13 have introduced CACs in their international issuances since 2002. The G-10 Group refrained from recommending globally a fully harmonised CAC model. A group of trade associations produced in February 2003 a set of model CACs for bonds issued in New York and London markets14; however, CACs

9 See http://www.cilae.org/publicaciones/SDRM.pdf 10 See http://www.imf.org/external/np/psi/2003/032503.htm 11 Statement of the President of the Ecofin April 2003: “The EU will use contractual provisions based on the framework developed by the G10, and where necessary in accordance with applicable law and adjusted to local practice, in their central government bonds issued under a foreign jurisdiction and/or governed by a foreign law by the end of this year.” 12 April 2003 Communiqué of the International Monetary and Financial Committee, and the IMF Guidelines for Public Debt. 13 Romania, Greece, Sweden, Austria, Cyprus, Estonia, Hungary, Italy, Lithuania, Poland, Slovakia, the United Kingdom. 14 See for instance the London model in: http://www.icmagroup.org/assets/documents/ 310103%20Final%20Model%20Clauses%20and%20Code%20of%20Conduct.doc.PDF

20 

 Antonio Sáinz de Vicuña y Barroso

are negotiated individually for every issuance, without a globally-accepted prescribed set of terms; as a result, majorities needed for several decisions, aggregation clauses, disenfranchisement provisions, etc, show a relatively wide diversity. Some authors15 opined that a certain diversity was tolerable, if the basic overall parameters were respected. In 2010, arising from the sovereign debt crisis that erupted in April 2010, a political decision was taken to make sovereign bondholders support part of the costs of facilitating the restructuring of the public finances of the sovereign; in May 2010 the Member States of the euro area agreed to lend money to the Hellenic Republic in a collective effort to facilitate the adoption of fiscal austerity measures, the increase in the tax collection, the orderly privatisation of part of the assets owned by the Hellenic Republic, plus a number of structural reforms aimed at putting the Greek public sector in a financially sustainable shape. In November 2010, the Eurogroup decided that “standardised and identical Collective Action Clauses (CACs) will be included in the Terms and Conditions of all new euro area government bonds”, as a way to ensure that bondholder creditors of the euro area sovereigns should – in the unlikely event of another sovereign crisis in Europe- share the burden of the re-structuring measures by accepting a re-scheduling of their credits negotiated under a harmonised generally-agreed CAC system. In doing that, the political leaders of the euro area followed the above-described growing universal pattern to have legal instruments in place to facilitate civilised and organised work-outs of difficult financial positions of sovereigns; they were also consistent with the above-mentioned Ecofin statement of April 2003 recommending the insertion of CACs by European sovereign issuers. On 2 February 2012 the Treaty Establishing the European Stability Mechanism was signed by 17 euro area Member States. The final version16 of Paragraph 3 of Article 12 of the Treaty states: “Collective action clauses shall be included, as of 1 January 2013, in all new euro area government securities, with maturity above one year, in a way which ensures that their legal impact is identical”.

On 26 March 2012 the standardised CACs to be applied as of 1 January 2013 were made public by the EU Economic and Financial Committee17. Its content builds upon international practice, and therefore is conceptually aligned and broadly

15 F. Elderson and M. Perassi, in Euredia July 2003: “Collective action clauses in sovereign foreign bonds: towards a more harmonised approach.” 16 Previous drafts had considered later dates for introducing harmonised CACs. 17 http://europa.eu/efc/sub_committee/cac/cac_2012/index_en.htm



Identical Collective Action Clauses for different Legal Systems: A European Model 

 21

consistent with existing CACs, although it differs in a number of particular respects18. Since the CACs apply only as of 2013, for a certain while pre-existing sovereign bonds without CACs will coexist with post-2013 bonds having the new CACs. So the scenario as of 2013 will be of co-existence of sovereign bonds with and without CACs, and of bonds with euro area standardised CAC terms and conditions, and of bonds with different CAC terms and conditions. To meet the political requirement that “the legal impact [of CACs] is identical” the European Economic and Financial Committee19 discussed several options: –– The use of a EU legal act; this road was not pursued because it was felt that no clear legal basis existed in the Treaty. –– The incorporation as an Annex to the ESM Treaty of the Common Terms of Reference where the ‘model CACs’ are embodied; this possibility had some support within euro area Member States, but not unanimous support; from a technical point of view, it had the inconvenience of lack of flexibility to allow for its adaptation if and when the particular terms of the financial instrument, the special profile of the investor or the specific market conditions so required, as it would indeed freeze once and forever the CACs, requiring an amendment to the ESM Treaty should the need arise for future adaptations; finally, it was deemed somehow exogenous to an international Treaty to have a mandatory provision belonging to the contractual domain. –– The idea of separating the CACs from the general terms and conditions of sovereign issuances, and submitting the CACs’clause – not the overall issuance – to the law and jurisdiction of a single pre-determined EU Member State. This private international law technique – known as dépeçage – would have provided a single legal system to interpret and enforce the CACs, irrespective of who the sovereign issuer was. The identical application of the CACs in Europe would have been achieved. However, it was not accepted as it was considered politically unacceptable for a sovereign to submit itself to the laws and jurisdiction of a foreign state in all its debt sovereign issuances, both the domestic and international.

18 See a legal assessment of variations vis-à-vis the CAC model hitherto used in the London market in http://www.linklaters.com/pdfs/mkt/london/A14950441_0_11_120508_CAC_client_ memo_MND.pdf 19 The real work on CACs was effected in the Sub-Committee of EU Sovereign Debt Markets, which reports to the Economic and Financial Committee: http://europa.eu/efc/sub_committee/ index_en.htm

22 

 Antonio Sáinz de Vicuña y Barroso

–– The euro area Member States finally settled on the idea of having the CACs in the format of “Common Terms of Reference” sufficiently elaborated and detailed so as to ensure consistent application by all, but documented in an informal manner, basically on the Commission’s internet site; euro area Member States would use such Terms as a rule, but because of its lack of legal format adaptations might be possible. Because under this model the CACs would be just a clause in the Terms and Conditions of each issuance, domestic or international, it would be subject to the Law and to the Courts under which the issuance is made; this entails a risk that the application of the CACs might not be “identical” everywhere. Such risk was, however, considered to be within an acceptable level, since (a) the prospect of another case20 in Europe of a rescheduling of the sovereign debt of a EU Member State was becoming more and more remote as the reforms that had been introduced or were in the process of being introduced21 in the euro area become operative making a second case implausible; and (b) any judgement by a national court having to interpret the standardised CACs should in all reasonable logic take into account judgements of courts of justice located in other jurisdictions on same issues.

The choice of an international jurisdiction A substantial majority of bonds issued by European sovereigns are issued under the national law of the issuing State or entity, and placed under the jurisdiction of the national courts. In the case of the Hellenic Republic some 93 % of Greek Government bonds were domestic, based on the established Greek legal provisions for debt issuance and subject to the jurisdiction of the Hellenic courts. Under the CAC model agreed by the European Economic and Financial Committee, the same approach applies to all euro area sovereigns, irrespective of whether issuances are subject to national law and domestic jurisdiction, or are international.

20 The only case has been the rescheduling of the debt of the Hellenic. 21 The forthcoming start of a financially powerful and operationally flexible European Stability Mechanism (ESM), together with the enhanced governance of the fiscal discipline and of the macroeconomic convergence of euro area Member States (the arrangements known as the “sixpack”, the “two-pack” and the Treaty on Stability, Coordination and Governance (nicknamed the “Compact”), plus the plans agreed in the Summit and in the Euro Area Summit of June 2012 for a Financial Union, a Fiscal Union and an Economic Union.



Identical Collective Action Clauses for different Legal Systems: A European Model 

 23

It is submitted that risks are higher in the case of domestic issuances that – in a major crisis – (a) the sovereign may use its capacity to enact legislation to unilaterally alter or amend the terms of its issuance, including the inserted CACs22; and (b) that a domestic court may be subject to and have to apply national laws that provide for procedural privileges to the sovereign, such as immunity from execution, or budgetary laws that delay or otherwise affect the enforcement of judgements – for instance by requiring budgetary consent –, or in any other manner favour or protect the sovereign. The credibility of CACs depends on the avoidance of the legislative arm of the sovereign using its power to change their terms or otherwise influence its application in any manner. The enforcement of the CACs should be outside the powers of the issuing sovereign. Would that be an ideal case for an international jurisdiction? Article 37 of the ESM Treaty provides for the jurisdiction of the Court of Justice of the European Union for disputes about the interpretation of the ESM Treaty; however, the locus standi for such jurisdiction is limited to the signatories to the ESM Treaty – i.e. the euro area Member States – and the ESM itself. That means that individual holders of sovereign bonds cannot sue the issuing sovereign in the European Court, but only in the national court designated in the terms and conditions of such bonds. Whether such designated national court may seek a preliminary ruling from the European Court is to be answered in the negative, since (a) there is no provision for this in the ESM Treaty, (b) the text of the CACs is located outside the ESM Treaty and thus is not part of it, and (c) the interpretation of the CAC is neither related to the interpretation of the ESM Treaty nor to EU Law. The proposal that was suggested – and rejected – when the ESM was being discussed in the Economic and Financial Committee was to submit the CACs, only the CACs and not the other clauses of a sovereign debt issuance (i.e. “dépeçage”), to the arbitration of the Permanent Court of Arbitration in The Hague23. Such court is a body established in 1899 under public international law, whose statute provides for fact finding, arbitration and conciliation in disputes involving various possible combinations of States, State entities, private parties, and international organisations; it has pre-established arbitration rules and a permanent Secretariat to administer them; it has a long track record of arbitrating disputes between

22 On 23 February 2012 the Parliament of the Hellenic Republic adopted Law 4050/2012 on Amendment of Securities Issued or Guaranteed by the Greek Government by Consent of the Bondholders, that affected the public debt issued under Greek law and subject to Greek courts. 23 http://www.pca-cpa.org

24 

 Antonio Sáinz de Vicuña y Barroso

bondholders and public issuers24; it has its seat in the Peace Palais of the International Court of Justice in The Hague; and it could provide a neutral forum for the interpretation of CACs where biases in favour of the issuing sovereign could not be even conceived. The rejection of that idea came out of the desire to keep public debt under existing practices, where a national court is designated or implicit25. However, in line with the evolution of debt recovery in the case of private debtor, summarised in the introduction to this article, it is submitted that the day will come where proper insolvency procedures for sovereigns will exist. Painful negotiations under CACs, litigation by dissatisfied creditors, will be replaced by a pre-established mechanism that will organise in a neutral manner the workouts of States under financial difficulties. The image of speculative holdout bondholders (so called “vulture funds”) escaping sensible rescheduling proposals and trying to seize assets of the sovereign all around the globe should never be repeated. As mentioned earlier in this article, the IMF tried this in 2002–2003 with its Sovereign Debt Restructuring Mechanism26. Regretfully, such proposal did not meet the favour of important financial centres and did not get finally enough support within the IMF. The current sovereign crisis in Europe has triggered renewed initiatives for such pre-established arrangements27. So far, it has resulted in expanding the use of CACs, which is a positive development. But – as happened in the domain of private debtors – the final destination of mankind in this similar domain of sovereign debt is without doubt that of a permanent procedural and institutional system to minimise damages and enhance the general interest.

24 See a list in http://www.pca-cpa.org/showpage.asp?pag_id=1029 25 The allegation was even made by a EU Member State that submission of the sovereign issuer to arbitration was unconstitutional. 26 See footnote 7. 27 See http://www.iir-hu.de/en/veranstaltungen/intl-conference-legal-procedure/ ; http://www.voxeu.org/article/european-debt-restructuring-mechanism-tool-crisis-prevention ; http://www.indiana.edu/~econdept/workshops/Fall_2010_Papers/EMSDR_191010.pdf ;

Claus-Michael Happe1

Towards a Level Playing Field Since the title chosen by the organizational board of this conference is without any punctuation mark, it gives raise to “a”, if not to say “the”, first and fundamental question: Will the mandatory implementation of a model set of CACs (as drafted) into all euro area sovereign bonds lead to a level playing field(!) or is this effect more questionable(?). Before coming to that, let us reflect a little on what is meant by the term “level playing field”. What is the playing field that needs to be in equilibrium? Finance Ministers made clear at a very early stage, i.e. March 2010, that CACs will serve to “facilitate agreement between the sovereign and its private-sector creditors” and will be introduced in a way “which preserves a level playing field among euro area Member States”. With this in mind, one may feel tempted to reflect about the level playing field “to be preserved”. (1) One angle to look at CACs could be: Are they neutral in respect of refinancing costs? Neutral in the sense that they should not allow one Member State to draw more advantage from the introduction of CACs than others. This could – inter alia – easily be the case if some of the model clauses were made optional. If – for example – the aggregation clause were not mandatory, it would be an advantage not to implement that clause in order to make a cross-series restructuring very difficult if not impossible. Investors as a consequence would be in a much more comfortable position and presumably honor that extra comfort by preferable conditions. Refinancing via bonds would become much cheaper. (2) The other angle to look at the level playing field is more fundamental in nature: if a level playing field is to be preserved, what is then meant by the level playing field that could potentially be endangered by the introduction of CACs? To shed a light on the latter question first, it is worth taking a brief historical detour into the construction of the euro zone: the report of the “Delors-Group”2 in 1989 stated that an Economic and Monetary Union would be the final result of the economic integration of Europe, but even after having achieved this result, the union would consist of individual nations with different economic, social, cul-

1 The opinions expressed here are those of the author and do not necessarily reflect the position of the Federal Ministry of Finance. 2 As described by Hans Tietmeyer Herausforderung Euro (2005).

26 

 Claus-Michael Happe

tural and political particularities. Because of this plurality, member states – says the report – need to enjoy a certain degree of autonomy in respect of their economic decisions. Nevertheless, a high degree of convergence was deemed to be absolutely necessary for any proper functioning. That statement on convergence was the substance. The rest remained more or less opaque: Mobility of workers and flexibility of wages, macro-economic coordination and mandatory rules on budgetary policies were just mentioned as proposals – as being ‘nice to have’. Further discussions very much focused on a common monetary policy guaranteed by an independent European Institution. This topic was a sensation. Would policy makers be able to compromise on an independent European Central Bank, modeled on the Bundesbank and accepted by France, Italy and others? Who would have bet on this? Discussions afterwards concentrated on issues related to that complex: the task and structure of the bank, criteria for a non-monetary economic policy and its surveillance including safeguarding the necessary fiscal discipline. But how and to what extent could such a safeguard mechanism be implemented? During the discussions it became clear to everybody who was willing to listen: by abandoning the exchange rate mechanism in a monetary union, one abandons a very important warning system and a smoothly-functioning safeguard tool. Exchange rates typically react at a very early stage of a latent crisis, perhaps much earlier that the interest rates do. As Hans Tietmeyer notes: “In a monetary union characterized by solidarity, erroneous fiscal trends of individual countries will – for the respective countries – hardly be sanctioned by the Financial Markets. … Whereas in the good old times creditors asked for an additional spread as compensation for national inflation and depreciation risks, that can not be the case in a monetary union. The related moral hazard problem must not be underestimated.”3 We now know the outcome: an imperfect Stability and Growth Pact that has been refurbished in recent months. This shows the dichotomy at stake – an instance of déjà vu from our current perspective – between economic and monetary integration. What does that mean in terms of a level playing field? Due to the stipulated pluralism the playing field was uneven from the very beginning: the economic, social, cultural and political particularities prevailed, but monetary integration was achieved – following a set of economic convergence parameters. Interest rates for borrowed money adapted in a very short time. Despite the economic differences, stakeholders in financial markets treated the playing field as flat. They did so despite the fact that the Maastricht Treaty explicitly provides for a nobail-out clause in Article 125. Every bond from every euro zone member state was

3 Hans Tietmeyer Herausforderung Euro, p. 229.



Towards a Level Playing Field 

 27

regarded as being risk free. There was only very little difference between interest rates – for example, the spread between German and Greek bonds. As we know today, that was a fiction. A fiction also echoed by the Basel II rules for non IRBA Banks: zero equity was required for keeping these bonds in the banking books. The result was that cheap money poured into countries with diverse economic, social, cultural and political conditions. In some cases, this led to extensive consumption and high debt ratios. The party, which was fueled by a fiction of equality, certainly could have kept going for longer – without the Lehman crash. Over night and around the globe, countries had to safeguard their banking systems by providing enormous guarantees to the banking institutions and taking on a mass of private risk in their public budgets. Rating agencies – heavily criticized during the Lehman crisis – chose to polish their reputation, became more wary and – perhaps consequently – identified unsustainable debt burdens, i.e. higher risks of default, notably for members of the euro zone. The fiction of the level playing field was thrown over board within hours. Welcome to reality. As to the initial question of the level playing field to be preserved: what kind of level playing field is meant? Did it ever exist? Does it need to be preserved? Nevertheless, vigilant investors put their money into this – highly tempting – fiction, despite the no-bail-out provision. While deciding what to invest in, they obviously believed more in an implicit guarantee of the other member states than in the letter of the treaty. Has economic realism defeated judicial positivism? The political debate has since identified many issues that require readjustment, as German Finance Minister Schäuble outlined already in an FTD article as early as March 12th 2010. Among institutional reforms (many of them were already agreed upon) one issue needs to be highlighted in particular: Since in a monetary union exchange rates can no longer serve as an early warning tool, the interest rate as a red flag must be preserved. This market instrument forces countries, parliaments and governments to take the necessary decisions, and reduces incentives for pursuing poor policies. This lesson needs to be learned and needs to be implemented. It is reasonable to expect that the introduction of model CACs in all Eurozone countries will help to keep investors vigilant and will lead to differentiated spreads that reflect the fiscal performance of each country. Let me come back to the first angle. Are the model CACs as drafted neutral, and do they not harm the refinancing conditions of Member States? Much has been written about the interaction of CACs and refinancing costs. Let me quote for many others “Bradley and Gulati: Collective Action Clauses for the Eurozone: an Empirical Analysis”. They concluded that CACs are value enhancing for weaker issuers, implying that the cost savings in terms of making restructurings easier are greater than the cost enhancing as a result of increased incentives for governments to engage in risky behavior. The model CACs under negotiation in the Euro-

28 

 Claus-Michael Happe

zone hopefully serve to enhance the validity of that statement. Model CACs are to be introduced by all Eurozone countries from 1 January 2013 into their securities with maturity longer than one year, be they issued under domestic or foreign law. CACs introduced by all members serve as a measure of last resort to restructure – if unavoidable – a country’s debt in an orderly, dialogue-based manner. They will assist in fostering an early dialogue between debtors and bondholders and prevent individual creditors from blocking negotiations on specific debt restructurings, rather than sending signals of an upcoming restructuring. Advocates of that view certainly ignore the fact that the Model CACs are directed to the future and do not alter any current situation or debt profile. Although these CACs are based on existing CACs under UK and New York Law, they have been remodeled in order to find a distinct balance between an effective restructuring and creditor interests. Concluding on the first angle: the ESDM Group did its best to draft a set of model CACs that balance the interests of investors and issuers. Since the CACs are to be implemented in such a way that they have identical legal effects in all member states, they are very likely to be neutral in terms of refinancing conditions.

David G. Sabel

An Introduction to the Euro Area’s Model Collective Action Clause I. Background On 28 November 2010, euro area finance ministers announced a number of policy measures intended to safeguard financial stability in the euro area, among them the mandatory inclusion of standardised collective action clauses (“CACs”) in all new euro area government securities issued after 30 June 2013. This date was subsequently brought forward to 1 January 2013. CACs, long familiar in English-law governed corporate bonds and, more recently, in emerging market sovereign bonds as well, provide an orderly contractual mechanism for amending a series of bonds without the consent of all the affected bondholders. This is achieved by having the bondholders agree in advance to be bound by amendments that are approved by a specified supermajority of the affected holders. CACs should, in principle, be attractive to issuers and to investors, because they facilitate the orderly restructuring of an issuer’s obligations on terms broadly acceptable to both constituencies. They do this, first, by minimizing the risk that a widely supported restructuring will be delayed or frustrated by a handful of dissenting bondholders, and later by preventing hold-out creditors from “freeriding” on the concessions made by their co-investors.1 In light of market developments in the euro area, it is not at all surprising that euro area authorities recently concluded that all euro area sovereign bonds should contain a CAC and, this being Europe, that the same model CAC should be adopted by all euro area sovereign issuers. The EFC Sub-Committee on EU Sovereign Debt Markets, chaired by Philippe Mills, the Chief Executive of Agence France Trésor, was tasked with drafting the model euro area CAC. The Sub-Committee includes representatives from all 27 EU Member States and from the European Commission, the European Central Bank, the European Investment Bank and the European Financial Stability Facility. In July 2011, the Sub-Committee distributed a draft of the model CAC to market

1 The ability of CACs to address the problem of hold-out creditors should not be overstated. An individual creditor, or a group of creditors acting in concert, may defeat the operation of a CAC by acquiring sufficiently large holdings to make it impossible in practice for an issuer to obtain the requisite approval necessary under the CAC to legally bind dissenting bondholders.

30 

 David G. Sabel

participants and other interested stakeholders for comment. A total of fifteen responses were received from institutional investors, securities and insurance trade associations, the European Central Bank, the International Monetary Fund, the London Stock Exchange and Euroclear. These comments were reflected in a subsequent draft of the model CAC, which was approved by the EU’s Economic and Finance Committee (EFC) on 18 November 2011. Copies of both the initial draft of the model CAC circulated for market comment and the final approved text of the model CAC and its supplemental provisions are attached as Appendices to this volume, together with two explanatory memoranda prepared by the Sub-Committee, summarizing its key terms and the reasons for the Sub-Committee’s drafting decisions. This article discusses two important aspects of the model CAC: its broad scope and its introduction of new substantive provisions that are either not found in current sovereign CACs, or are not dealt with nearly so comprehensively.2

2 The Sub-Committee’s two explanatory memoranda, and the submission of the initial draft of the model CAC for broad market comment, also distinguish it from other CACs now in the market. The Sub-Committee’s published views on the model CAC’s key provisions provide an authorized context for its interpretation, and should in practice be of assistance to market participants, courts and commentators in filling the model CAC’s inevitable gaps and in resolving its inevitable ambiguities, even if the Sub-Committee’s views are in principle subject to the same inherent interpretive limitations, once removed. By contrast, most other CACs are the result of either an issuer’s private negotiations with its underwriters or – move often than most market participants would like to admit – a not very carefully considered “mark-up” of an existing CAC. Even where the terms of a CAC are specifically negotiated, the issuer’s aim is often principally to obtain CACfree pricing for its bonds, while its underwriters seek to successfully offer those bonds at a price broadly satisfactory to both the issuer and the market. The comments received on the initial draft of the model CAC, and the changes made in the final version in response to those comments, should likewise facilitate its interpretation and reception in the market for three related reasons. First, those provisions that did not attract any comment – by far, the bulk of the initial draft – may be assumed to be well understood and agreed by market participants. Second, many of the provisions that were amended in the final version directly reflect changes requested by market participants, and the reasons for those changes are set out in both the comment letters received by the Sub-Committee and in the Sub-Committee’s own supplemental explanatory memorandum. Finally, with respect to those comments that were rejected by the Sub-Committee or implemented only in part, the Sub-Committee’s supplemental explanatory memorandum discusses the reasons for its decisions, and in effect establishes an outer boundary of the model CAC’s intended operation.



An Introduction to the Euro Area’s Model Collective Action Clause 

 31

II. Scope of Application The model CAC must be included in all euro area sovereign domestic and international debt securities with an original stated maturity of more than one year. This is a significant departure from existing emerging market practice, where CACs are generally found only in a sovereign’s international bonds, and are specifically tailored to the terms and conditions of those bonds, to then-prevailing market conditions, and (often) to the issuer’s desire to maximize investor participation in a proposed exchange of its existing bonds for new bonds having more favorable terms and conditions. These important differences are inevitably reflected in the provisions of the model CAC.

A. Domestic and International Obligations Sovereign CACs have historically been used almost exclusively in a State’s international offerings, which tend to be much more fully documented than a sovereign’s domestic obligations, which may not be meaningfully documented at all. A government’s domestic debt – governed by its own laws – is also subject to the principle of fiscal sovereignty, and thus may be unilaterally amended by the issuer, subject to any conflicting constitutional or treaty provisions. The United Kingdom’s abandonment of the gold standard in 1931, and Greece’s more recent adoption of a law amending its outstanding domestic bonds to include a new retrofitted CAC, are examples of a State exercising its fiscal sovereignty. Three other differences between a sovereign’s domestic and international obligations are worth noting. First, sovereign issuers very rarely, if ever, waive their immunity from suit or execution in their domestic obligations, but often, (through not invariably), do so in their international offerings. Second, sovereign domestic debt instruments are always (though often not expressly) governed by the sovereign’s own law, even if that sovereign’s international obligations are governed by New York or English law. And, finally, a government’s domestic obligations, unlike its international offerings, rarely, if ever, provide for acceleration upon the occurrence of an event of a default. These different arrangements are reflected in the model CAC in a number of ways. Consistent with the principle of fiscal sovereignty, a euro area sovereign’s proposal to change the law governing its debt obligations is treated in the model CAC as a “reserved matter” – that is, a matter requiring the highest level of bondholder approval – only if the bonds are governed by a law other than the issuer’s own law. This leaves the issuer free to amend its own law, and thus the terms of any of its bonds that are governed by its law, so long as the issuer has not previ-

32 

 David G. Sabel

ously limited its fiscal sovereignty either constitutionally or by treaty in a way that would prohibit the proposed modification. Recognizing that euro area governments are extremely unlikely in their domestic obligations to waive their sovereign immunity or to submit to the jurisdiction of another country’s courts, a proposed amendment affecting these matters is treated in the model CAC as a reserved matter only if the issuer has already waived its immunity or submitted to the courts of another sovereign in the bonds to be amended. A proposed change in the payment-related events that may result in a bond being declared due and payable prior to its stated maturity is likewise treated as a reserved matter only if the bond already contains an acceleration clause. These arrangements, taken together, preserve the freedom of action found in domestic sovereign debt securities, but are also very likely to be consistent, insofar as a sovereign’s international debt securities are concerned, with the corresponding provisions of the CACs now found in that sovereign’s international obligations.

B. All National Debt Obligations The model CAC is required to be included in all euro area national debt securities. For these purposes, an issuer’s debt securities means all of its bills, bonds, debentures, notes and other debt securities with an original stated maturity of more than one year. The Sub-Committee did consider, but ultimately rejected, proposals to extend the mandatory reach of the model CAC to regional and municipal governmental obligations or, even more broadly, to all obligations that are taken into account in determining a euro area Member State’s compliance with the Maastricht (euro convergence) criteria. The national debt obligations of euro area Member States include a broad range of financial instruments, including fixed and floating rate bonds, unsecured and collateralized bonds, short and long-term instruments, bonds denominated in different (or multiple) currencies, index-linked instruments (typically, linked to changes in a published inflation index) and zero-coupon obligations, both those issued without provision for the payment of interest and those resulting from the stripping of interest-bearing obligations. The model CAC will thus in all cases provide for the treatment of many types of financial instruments in addition to the specific debt obligation in which it is included. The Sub-Committee nonetheless decided to require that issuers include all of the provisions of the model CAC in all of their covered debt securities, because other provisions, though not immediately applicable, may be relevant



An Introduction to the Euro Area’s Model Collective Action Clause 

 33

to an issuer’s subsequent cross-series modification – that is, a restructuring exercise involving not only the bonds in question, but also one or more additional series of bonds. The model CAC’s substantive treatment of a cross-series modification is discussed below. The model CAC’s provisions dealing with index-linked and zero-coupon obligations benefitted from extensive consultation with Sub-Committee members active in these markets, and it is thought that the model CAC’s provisions in respect of these oblegations are more comprehensive than those found in one-off CACs drafted with the terms and conditions of a specific debt instrument in mind.3 If these provisions are well received by the market, it is likely that they will serve as a model for CACs that are included in index-linked and zero-coupon obligations issued by private and non-euro area sovereign issuers.

C. All Euro Area Member States The Treaty Establishing the European Stability Mechanism, signed by all 17 euro area Member States on 2 February 2012, provides that CACs will be included in all covered government debt obligations “in a way which ensures that their legal impact is identical” in all Member States. In order to ensure the identity of “legal impact” called for by the Treaty Establishing the European Stability Mechanism, each euro area Member State is required to deliver a legal opinion from its highest legal authority competent for such matters, confirming that the model CAC is legal, valid, binding and enforceable in accordance with its terms under the law of that Member State or, if the model CAC will be implemented by a law, regulation or decree, that the applicable legal act will come into force no later than 1. January 2013 and will not conflict with any other law, regulation or decree of that Member State. In advance of the delivery of the required legal opinions, each Member State has previously confirmed to the Sub-Committee that the model CAC, as approved by the EFC, will be enforceable under its own law. Where an enforceability issue was identified by a Member State in the course of drafting the model CAC, the provision in question was redrafted so as to be enforceable under the laws of all euro area Member States and the laws of England and New York, with the exception of the model provision dealing with technical

3 The provisions of the model CAC dealing with index-linked and zero-coupon obligations (and with proposed modifications involving bonds denominated in different currencies) are found in Section 2.6 of the model CAC.

34 

 David G. Sabel

amendments, which could not be redrafted without substantially departing from current market practice. The Sub-Committee, however, concluded that this provision was by its nature extremely unlikely to affect the legal impact of the model CAC, and accordingly agreed that euro area issuers may, at their option, include the model technical amendments clause in their debt obligations if the clause is enforceable under the applicable governing law.4 The Sub-Committee considered, but rejected, another more direct way of possibly ensuring the model CAC’s required identity of “legal impact” across the euro area. Under this alternate approach, the model CAC would in all cases have been governed by the same law – the laws of England, for example – regardless of the law governing the other provisions of the issuer’s debt securities. This approach was rejected in part because it was thought to be politically unacceptable to have a euro area issuer’s domestic obligations governed by the laws of another State; in part because a split-governing law, while not unheard of – it is sometimes referred to as dépeçage – would have added additional legal complexity and uncertainty to an already complicated clause; and in part because no assurance could be given that all euro area national courts would interpret in the same way a clause governed by the unfamiliar laws of another State. The euro-area scope of the model CAC also had significant consequences for the CAC’s bondholder meeting provisions.5 Bondholder meeting rules vary greatly from jurisdiction to jurisdiction, and while this rules are very important to the bond-modification process, they are not an essential (or even a required) element of most CACs. In light of the absence of bondholder meeting rules for domestic sovereign debt obligations, the Sub-Committee concluded that it would be helpful to prescribe a minimum set of rules for all euro area sovereign bondholder meetings, but to allow issuers to supplement those rules to take account of local laws and practice to the extent not inconsistent with the CAC’s mandatory meeting rules. It is expected, for example, that while the rules prescribing the manner in which notice of a meeting may be given to bondholders will vary from Member State to Member State, the supplemental rules adopted by individual issuers will in all cases respect the minimum notice period and mandatory notice information required by the model CAC. For reasons of transparency, any sup-

4 The model CAC includes two other supplemental provisions – a clause dealing with acceleration and one limiting the right of bondholders to bring legal action against an issuer – that may also be optionally included in the terms and conditions of a euro area sovereign issuer’s debt securities if they are enforceable under the applicable governing law. 5 The model CAC’s bondholder meeting (and related) rules are found in Sections 4.1–4.13 of the model CAC.



An Introduction to the Euro Area’s Model Collective Action Clause 

 35

plemental rules applicable to a bondholder meeting must be set out in the notice for that meeting.

III. New Substantive Provisions The model CAC also includes a number of substantive provisions that depart from existing practice. Those dealing with the disenfranchisement of issuer-held or issuer-controlled bonds, with issuer transparency and with cross-series modification are discussed below.

A. Disenfranchisement All CACs contain a clause disenfranchising bonds held by the issuer. This makes eminent sense, as the losses suffered by an issuer from the modification of its own bonds held by or for the account of the issuer are more than offset by the gains realized by the issuer from the resulting reduction in its debt service or debt stock or both. Issuer-held bonds are for this reason not counted in determining whether a proposed modification has been approved by the required supermajority of bondholders, and are also excluded from the definition of “outstanding” bonds” used in calculating whether a quorum is present and whether a bond is eligible to vote on a proposed modification. In an attempt to prevent issuers from achieving by indirection that which is denied them directly, many CACs go one step further and disenfranchise all bonds held by any legal entity that is directly or indirectly owned or controlled by the issuer, with “control” broadly defined to include not only actual control, but also the right to control, even if that right has never been exercised. The practical effect of these CACs is to disenfranchise bonds held by the issuer’s central bank or by any company the majority of whose shares are owned directly or indirectly by the State, regardless of whether the central bank or company is in fact acting on behalf of the issuer. The model CAC, while it expressly disenfranchises bonds held in the name of the issuer, departs from current practice in not automatically disenfranchising bonds held by State-owned or State-controlled entities. As explained by the SubCommittee in its explanatory memoranda, many State-owned or State-controlled entities are legally required to act for their own account or for the account of a thirdparty. In the model CAC, these entities are referred to as having “autonomy

36 

 David G. Sabel

of decision” in respect of their investments.6 For example, Article 130 of the Treaty on the Functioning of the European Union and Article 7 of the Statute of the European System of Central Banks and of the European Central Bank expressly prohibit euro area national central banks and members of their decision-making bodies from seeking or taking instruction from EU institutions or bodies, from any government of a Member State or from any other body. As a result, a euro area national central bank’s decision to vote for or against a proposed modification of bonds acquired in connection with its Eurosystem operations must, as a matter of law, be made by the bank acting in its own interest, even if the bank is owned by a Member State or a Member State has the right to appoint the bank’s governor and a majority of its directors. More generally, the model CAC enfranchises bonds held by State-owned or State-controlled entities if, as a matter of law: (i) The holder may not take instruction from the issuer on how to vote on the proposed modification. As explained above, bonds held by euro area national central banks will generally be enfranchised on this basis. (ii) The holder is required act in accordance with an objective prudential standard, in the interest of all of its stakeholders or in its own self-interest. This provision will enfranchise bonds held by a partially State-owned company that is required by law to act in the interest of all of its shareholders. (iii) The holder owes a fiduciary duty to vote on a proposed modification in the interest of a person other than a person whose holdings would themselves be disenfranchised under the model CAC. Subject to national legislation, this provision will enfranchise bonds held by State-owned pension funds. Consistent with this approach, the model CAC provides that the holder of a bond for these purposes is the entity having the ultimate right to determine how the bonds are voted, regardless of whether that entity is also their legal owner. This provision will generally enfranchise bonds held on repo by a State-owned company that does not have autonomy of decision, so long as the bonds would not otherwise be disenfranchised if legally held by the company’s repo counterparty. It is thought that this arrangement will be well received by the market, as it clarifies the treatment of a potentially large volume of bonds that is often not addressed in other CACs.

6 Sections 2.7 and 2.8 of the model CAC set out the rules governing whether a bond is “outstanding” for purposes of determining if a proposed modification has been approved by holders of the requisite majority of the affected “outstanding” bonds, and if a quorum is present at a bondholder meeting called to consider a proposed modification.



An Introduction to the Euro Area’s Model Collective Action Clause 

 37

In adopting the principle of autonomy of decision, the Sub-Committee rejected several objections raised by market participants, including a general concern expressed by some investors that holders of euro area sovereign bonds might in practice be influenced by the consequences for the euro, or for the euro area as a whole, of rejecting a proposed restructuring, as opposed to the restructuring’s direct effect on the value of the investor’s own holdings. In the SubCommittee’s view, bondholders often make investment decisions for reasons that go well beyond the value of their immediate holdings, and this has never been thought to be a reason for disenfranchising an investor’s bonds. For the same reason, the Sub-Committee rejected arguments that would have disenfranchised bonds held by State-owned or State-controlled entities merely because their interests are likely in many cases to be aligned with those of the issuer, or because they are otherwise likely to vote in favor of a proposed restructuring. As explained by the Sub-Committee, the disenfranchisement clause is intended to serve a very narrow purpose – to prevent issuers from acting through State-owned or Statecontrolled proxies – and has never previously been used to disenfranchise bonds that are held, for example, by an investor benefitting from a credit default swap that would be triggered by a proposed restructuring, even though the investor’s interests are in a narrow sense aligned with those of the issuer.

B. Transparency As recent events in Greece have made clear, any restructuring of a euro area sovereign’s bonds is likely to be politically and financially charged. It is thus vitally important that a restructuring implemented pursuant to a CAC – that is, without the consent of all affected bondholders – be seen to have been carried out in accordance with the CAC’s agreed terms and conditions.7 The Sub-Committee was

7 Under the model CAC, a reserved matter may be modified with the affirmative vote of not less than 75% of the aggregate principal amount of the affected outstanding bonds represented at a duly called meeting, or by a written resolution signed by holders of not less than 66 2/3% of the aggregate principal amount of the affected bonds then outstanding. The quorum for a meeting called to consider the modification of a reserved matter is not less than 66 2/3% of the aggregate principal amount of the affected series of bonds then outstanding. The lower threshold for an action in writing reflects the fact that this threshold is based on the total principal amount of bonds then outstanding, as opposed to the principal amount of then outstanding bonds represented at a duly called meeting.

38 

 David G. Sabel

especially mindful of this, and took unprecedented steps to ensure the transparency of the bondholder approval process under the model CAC.8 Under the model CAC, an issuer must publish promptly following its formal announcement of a proposed modification (but in all events not less than 10 days before the record date for that modification) a list of all the entities controlled by the issuer that, based on the issuer’s own enquiries, then hold bonds that would be affected by the proposed modification but do not have autonomy of decision. This measure is intended to give investors advance notice of the potential universe of bonds that will, and will not, be entitled to vote on the proposed modification. This information will not, however, be definitive, as there will inevitably be substantial trading in the bonds prior to the record date for the proposed modification. The model CAC goes on to provide that the issuer must appoint a calculation agent to calculate whether a proposed modification has been approved by the requisite majority of “outstanding” bonds – equal to the principal amount of bonds outstanding for corporate purposes more generally, less the principal amount of any bonds disenfranchised under the model CAC. The issuer is also required to provide to the calculation agent (and to publish) a certificate listing the total principal amount of bonds outstanding on the record date for the proposed modification, the principal amount of bonds deemed under the model CAC to be not “outstanding” on that date, and the identity of the holders of any disenfranchised bonds. Inasmuch as a bond may be disenfranchised only if it is held by an entity owned or controlled by the issuer, it is expected that the issuer will write to all of its owned or controlled entities requesting information as to the principal amount of affected bonds that they then hold. This will avoid the otherwise intractable problem of identifying the ultimate beneficial holders of potentially disenfranchised bonds. Finally, the issuer will determine as to each owned or controlled holder whether that holder has autonomy of decision under the model CAC. The information included in the issuer’s certificate will be conclusive and binding on the issuer and all affected bondholders unless a bondholder delivers a “substantiated written objection” to the issuer before the vote on the proposed modification, and that objection, if sustained, would affect the outcome of the vote taken. In the event a substantiated written objection is timely delivered, the information included in the issuer’s certificate will still be conclusive and binding if the objection is later withdrawn, if the objecting bondholder does not commence legal action on its objection within 15 days of the publication of the

8 Transparency issues are dealt with in Sections 2.9 and 3.1–3.4 of the model CAC.



An Introduction to the Euro Area’s Model Collective Action Clause 

 39

results of the vote, or if a court subsequently rules that the objection is not substantiated or would not have affected the outcome of the vote. These measures, believed to be substantially more detailed than those found in other CACs, are intended to balance the different interests of the issuer and the bondholders. Insofar as investors are concerned, the model CAC provides bondholders with the information they need to challenge the results of a vote taken in reliance on the issuer’s certificate, and affords them a reasonable opportunity to have their objections heard by a competent court. At the same time, the transparency provisions protect issuers against unsubstantiated investor action taken solely to delay or frustrate implementation of a modification approved under the model CAC.

C. Cross-Series Modification Most sovereign bond issuers have more than one series of outstanding bonds. In the event an issuer is unable to service all of its outstanding bonds, some – but by no means all – CACs provide that the issuer may seek to modify two or more series of its bonds through a concerted offer made to all of it affected bondholders, often referred to as an aggregated modification or, in the model CAC, as a cross-series modification. Among the advantages of a cross-series modification is an issuer’s ability to deal more or less simultaneously with all of its bondholders. Investors can thus be assured that their holdings will not be treated invidiously, and, if the restructuring is broadly successful, that the issuer will obtain sufficient debt relief to ensure that it can realistically service its restructured obligations. Cross-series modifications also minimize (but do not eliminate) the risk of free-riding by holdout creditors who might otherwise be able to prevent the restructuring of one or more series of an issuer’s bonds by accumulating a blocking position in those bonds. A cross-series modification thus addresses many of the first-mover and hold-out concerns that may otherwise complicate the restructuring of a sovereign’s debt securities. With a view to facilitating the broadest possible modification of an issuer’s bonds, while affording the greatest possible equality of treatment to the holders of those bonds, the model CAC’s cross-series modification provisions are more comprehensive than those found in most other CACs, and allow for both partial cross-series modifications (in which a proposed modification may be made legally effective as to some but not all of the affected series of bonds) and for cross-law modifications (in which bonds governed by different laws may be combined in a single cross-series modification).

40 

 David G. Sabel

Those CACs that do provide for cross-series modification typically require that the proposed modification be approved by the requisite super-majority of all the holders of all the affected series of bonds (an aggregate test) and also by a lesser majority of the holders of each individual series of bonds (an individualseries test). A cross-series modification can thus be understood as, in effect, a CAC that works at the series level, in that it legally binds all of the holders of all of the affected series (including dissenting holders) so long as the proposed modification is overwhelmingly approved in the aggregate by the affected bondholders, but with the important further protection that the holders of each individual series of affected bonds will not be bound by the decision of the group as a whole unless the holders of that series also vote in favour of the proposed modification, though not necessarily at the level that would be required on a stand-alone basis. Under the model CAC, more than one series9 of debt securities may be modified in relation to a reserved matter in a cross-series modification with the affirmative vote of (i) the holders of at least 75% of the outstanding principal amount of all the affected bonds, and (ii) the holders of at least 66 2/3% of the outstanding principal amount of each individual series of affected bonds, in each case as are represented at one or more duly called meetings.10 An issuer may include some – but not all – of its outstanding bonds in a single cross-series modification, and may also treat its remaining bonds in one or more additional cross-series modifications, in several single-series modifications, or in a combination of the two. The flexibility afforded the issuer in grouping its bonds of different series in different modification baskets is intended to minimize the likelihood of a hold-out series, an interest shared by both the issuer and those investors – the vast majority – that approve a cross-series modification on both an aggregate and individual-series basis. Allowing the holders of one series of bonds to have a say in whether the holders of another series of bonds will be legally bound by a proposed modification requires that all of the affected bondholders have the right to vote on the same proposed modification or, if a series of different modifications is proposed, on each of the proposed options. Under the model CAC, a cross-series modification may include one or more proposed alternative modifications, but only if all of

9 Under the model CAC, a “series” is defined as one or more tranches of debt securities that are identical in all respects save for their date of issuance or first payment date, and which are expressed to be consolidated and form a single series of debt securities. 10 A cross-series modification may also be implemented by means of a written resolution signed (i) by at least 66 2/3% of the aggregate outstanding principal amount of all the affected bonds, and (ii) by at least 50% of the outstanding principal amount of each individual affected series of bonds.



An Introduction to the Euro Area’s Model Collective Action Clause 

 41

the proposed modifications are addressed to, and may be accepted by, any holder of any bond that would be affected by the cross-series modification.11 Issuers are thus free under the model CAC to propose a menu of alternative options, and investors may then pick the option they prefer, effectively preventing the issuer from colluding with a group of favoured investors, and ensuring that all investors have the right to accept, from a common menu of options, the option that best suits their interests. If no option is acceptable to the holders of an individual series of bonds, the requisite minority of the holders of that series need only vote against its adoption to prevent it from coming into effect insofar as the holders of that series of bonds are concerned. As a result, under the model CAC – as under most CACs that provide for crossseries modification – a determined group of investors may prevent the adoption of a proposed cross-series modification by accumulating a blocking position in one of the affected series of bonds, causing the entire modification to fail because the individual-series test would not be satisfied as to all of the affected series of bonds. While an issuer is free under the model CAC to structure a cross-series modification in the manner it believes most likely to succeed, and may in principle propose a new cross-series modification that excludes a prior hold-out series, the outcome of a cross-series modification is likely to be difficult to predict in advance, and an issuer may find it difficult or impossible in practice to propose a second cross-series modification if the first one fails to satisfy both the aggregate and individual-series tests. In order to address these concerns, the model CAC provides, in the event a proposed cross-series modification is approved in respect of some but not all of the affected series, that an issuer may implement the proposed modification in relation to those series of bonds whose modification would have been approved if the cross-series modification had initially been proposed only in respect of the bonds of those series – a so-called partial cross-series modification. If, for example, a cross-series modification involving Series A, Series B and Series C bonds were to fail because the Series B bondholders vote overwhelmingly against the proposed modification, but would have succeeded, on both an aggregate and individual basis, if the modification had initially covered only the Series A and the Series C

11 Cross-series modification is dealt with in the model CAC in Sections 2.2–2.4. The requirement that all affected bondholders have the right to accept any of the restructuring options that are included in a cross-series modification is set out in Section 2.3.

42 

 David G. Sabel

bonds, then the proposed modification may be deemed to have been approved under the model CAC in respect of the Series A and Series C bonds only.12 Allowing an issuer unbridled discretion to implement a cross-series modification in respect of some – but not all – of the affected series of bonds would inevitably give rise to investor uncertainty as to whether the investor’s own bonds would be modified if a proposed cross-series modification were to be implemented only in part, thereby increasing the risk of investor hold-outs. Returning to the above example, the holders of the Series A bonds might well hesitate to support a proposed cross-series modification if they did not know in advance whether they would be legally bound by the modification if it were to be approved in the aggregate by the requisite majority of all the affected bondholders and by the requisite majority of the Series A and Series C bondholders individually, but not also by the requisite majority of the Series B bondholders. This concern is addressed in the model CAC by providing that an issuer must notify investors, in advance, of the conditions under which a partial cross-series modification will be deemed to have been approved, if it is approved in respect of some, but not all, of the affected series of bonds. Investors will, as a result, be able to make an informed investment decision as to whether they are in favour of a proposed cross-series modification even if it is only partially successful. This approach should be familiar to the market, as it is already standard practice for an issuer proposing to modify a series of bonds that does not include a CAC to specify in advance the “critical mass” (or principal amount) of bonds of that series which that must approve the proposed amendment before it will become legally effective. The model CAC does not distinguish between a cross-series modification involving bonds that are subject to the same governing law and one involving bonds that are governed by the laws of different jurisdictions. It is believed that none of the CACs currently in the market contemplate the possibility of a cross-law modification, presumably because the conflicting procedural rules for obtaining bondholder consents in different jurisdictions render impossible a single meeting of all affected bondholders. The model CAC avoids this problem by providing for separate meetings for each series of bonds, and then tallies the results of all the individual meetings to determine whether a proposed cross-series modification has been approved both in the aggregate and on an individual-series basis.

12 The rules governing partial cross-series modifications are found in Section 2.4 of the model CAC.



An Introduction to the Euro Area’s Model Collective Action Clause 

 43

IV. Conclusion CACs have been the subject of rapid and significant development in recent years, as new provisions have been introduced, and existing provisions modified and expanded, in response to changing market conditions. While the model CAC has the backing of euro area and EU authorities, it too will be subject to the market’s independent judgment, and it was in part for this reason that the Sub-Committee elected not to implement the model CAC by means of a treaty, as any amendment of the model CAC would then have required the treaty to be re-ratified by all 17 euro area Member States. It is certainly to be hoped that the model CAC will also contribute to the further development of both sovereign and private CACs, but here, too, the market will be the ultimate arbiter. This is only appropriate, because the recent revived interest in CACs on the part of the euro area (and other) authorities was largely a response to the IMF’s failed plan to introduce greater order and predictability to the sovereign debt restructuring process through a bankruptcy-like regulatory scheme for illiquid and insolvent sovereign debtors.13 CACs, by contrast, are correctly seen as a market-based mechanism that addresses many – but not all – of the issues that would have been dealt with in the IMF’s plan, and it is thus appropriate that the market should ultimately decide whether the model CAC meets the needs of both issuers and their creditors.

13 See “Sovereign Debt Restructuring and Dispute Resolution”, speech by Ms. Anne O. Kruger, First Deputy Managing Director of the IMF, at the Bretton Woods Committee Annual Meeting (June 6, 2002).

Christian Hofmann

Enfranchisement and Disenfranchisement in Collective Action Clauses I. The phenomenon of disenfranchisement 1. Majority provisions in Collective Action Clauses One of the core mechanisms of Collective Action Clauses is the principle of majority voting.1 Majority voting can be considered an effective tool to overcome the hold out problem.2 Without majority provisions it may pay off for bondholders to keep a low profile and let others compromise. The majority provisions in Collective Action Clauses eliminate any hold out strategy. Once the majority requirements are met, all affected bonds are modified in the same way. Amendments to the terms of payment apply equally to all bondholders, be it amendments to the interest rate, the maturity period or the principal owed. This contribution is if not literally, then in its main propositions, an excerpt from a forthcoming issue of the Texas International Law Journal.3

2. The right to vote When amendments are decided by majority vote, the individual right to vote becomes essential for the bondholders. The majority provisions in Collective Action Clauses enable a group of creditors to amend the terms of the bond contract with binding effect on any dissenting minority. By that, the majority provisions deviate from one of the most basic and internationally recognized principles of contract law, the principle of sanctity of contracts (or pacta sunt servanda) – the binding effect of the contractual terms initially agreed upon which can only be amended by party consent.4 Collective Action Clauses, however, by way of

1 In detail see the contribution by Bauer in this volume, p. 3–7. 2 On the hold out dilemma in detail see Buchheit/Gulati, 28 Int’l. Finan. L. Rev. 22, 23 (2009). 3 Christian Hofmann, Sovereign Dept Restructuring in Europe under the new model Collective Action Clauses, 49 Tex. Int’l. L.J. (forthcoming 2014). 4 See H. Beale (ed.), Chitty on Contracts, Vol. 1 (General Principles), 29th ed. 2004, para 1–5; E. McKendrick, Contract Law, Text, Cases and Materials, 3rd ed. 2008, p. 939 et seq.; G. Treitel,

46 

 Christian Hofmann

introducing majority provisions, usually combined with provisions on centralized enforcement powers, reduce the rights of the bondholders significantly.5 The individual right of being the sole master of the contractual rights agreed upon is being reduced to a right to cast one’s vote in a meeting of bondholders or a written resolution. As a consequence, without a right to vote, the bondholders lose any impact on the fate of their claims. This is what disenfranchisement clauses achieve: A disenfranchised bond gives no right to vote (as opposed to enfranchised bonds which do). What is more, it is considered not outstanding and therefore not counted when quorums and majority requirements are determined.6 And there is a further consequence: The bondholder is also cut off from any remedies against the amendment of the terms. Such remedies traditionally require that an enfranchised bondholder objects to the proposal and seeks relief in court. As a consequence, disenfranchisement clauses need justification, and given the severe consequences for the affected bondholders the requirements for justification must be fairly high.7

II. Justification of disenfranchisement To find this justification, it helps to look at the intention of disenfranchisement clauses.

1. Fairness of the vote Disenfranchisement is aimed at a different purpose than many other important provisions in CACs. Most clauses intend to reduce the holdout problem. In contrast, the disenfranchisement clause seeks to limit the influence of the issuer to guarantee the fairness of the voting procedure. Fairness of the vote can be assumed when one of the basic principles of contract law is guaranteed: As long as all parties pursue their own best interest, it can be assumed that a correct decision is being taken – correct in the sense that

The Law of Contract, 6th ed. 2004, p. 312 et seq. On the protection of contractual expectations see p. 5 et seq. 5 On centralized enforcement powers Buchheit/Gulati, 28 Int’l. Finan. L. Rev. 22, 23 (2009). 6 On typical majority requirements see the contribution by Bauer in this volume, fn. 2 above. 7 Compare Buchheit/Gulati, 48 UCLA L.R. 59, 77 (2000).



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 47

the result is the best possible compromise that parties with selfish and therefore conflicting interests were able to reach. In sovereign bond contracts the issuer pursues the interest of reducing its debt as much as possible whereas the bondholders aim to reduce their claims not further than to the level where the sovereign’s debt seems to regain sustainability. In the opposite case in which a bondholder tries to free the issuer from as much debt as possible, he pursues an interest which is identical to the issuer’s. In such a scenario, there is no basis for assuming that the majority decision reflects the best possible compromise. This observation leads to the core issue which is tackled by disenfranchisement clauses – the conflict of interest on the bondholder’s side.

2. Conflict of interest One conflict of interest is most obvious: If the same party participates in the voting on both sides of the vote he will pursue the interests which are related to the highest stakes he holds. Here, this would mean that the issuer participates in the voting of the bondholders because the sovereign holds some of its own bonds. It is evident that he would vote in favor of a restructuring. Therefore, if the issuer and the bondholder are identical in a legal sense, a sovereign that holds its own bonds must not be entitled to vote on an amendment of the bond terms and therefore its own debt. However, cases in which there is no legal identity of issuer and bondholder seem both more frequent and problematic. The bondholder may share economic interests with the issuer which distinguishes him from the rest of the bondholders and may therefore be much more inclined to agree to a proposed amendment than the rest of them. The reason is that for such a bondholder economic arguments prevail because he stands to gain more if the issuer’s obligation is eased than in case the original debt is being upheld. Whereas on the bondholders’ side the burden is shared, on the issuer’s the gains are not. If five million bonds are held by a number of investors and the proposed amendment intends to reduce the debt on each of them by the amount of one Euro, each bondholder loses one Euro on each bond. At the same time, the issuer gains five million Euros.

3. Irrelevance of mere motivations and indirect advantages However, the share of economic interests seems a rather vague criterion for legal consequences, too vague to justify the severe result of a loss of voting power. It

48 

 Christian Hofmann

would also hardly be a feasible criterion: Some bondholders may only consider a haircut to avoid further losses on their bond claims, whereas many others, in particular institutional investors and the financial industry have reason to fear further losses from a sovereign default, for instance from further debt owed to them by the sovereign, from a destabilization of the financial industry or merely from a further decline of the national or even world economy. These are the mere motivations of the bondholders which form no basis for a disenfranchisement of bonds – even if they are in line with the interests of the issuer. The same must be true for advantages or disadvantages which are only indirectly connected to an amendment of the bond terms. In spite of the principle of equal treatment that must apply to all amendments8 it seems possible that one bondholder stands to gain more from a significant reduction of the bond debt than another. If one bondholder is a creditor to further claims against the sovereign the relief resulting from the amendments of the bond obligations may result in a higher likelihood that those other claims can be enforced. These indirect advantages influence the bondholder’s vote, but since they are not directed by the issuer’s will they do not justify a disenfranchisement of the bondholder.9

4. Ownership or control Therefore, whereas a legal identity of issuer and bondholder should not be a mandatory criterion for disenfranchisement, a bond of some significance between them seems indispensable. Ownership or (legal or mere factual) control creates such a bond. When the issuer owns or controls the bondholder, it is not unlikely that the issuer will exercise its power and have the bondholder’s votes cast in its own interest, thereby forcing the bondholder on its side and creating a conflict of interest between the controlled and the independent bondholders. Therefore disenfranchisement clauses should require that the bonds are owned or controlled by the issuer. Indirect control should be included to avoid any circumvention of the disenfranchisement clause by simply placing another entity in between the issuer and the bondholder.

8 On the requirement of equal treatment cf. the contribution by Burn in this volume, p. 75–80. 9 On the issue of a potential abuse of the majority power see the short remarks at IX.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 49

III. Approaches to a common standard Since 2003, there has been a substantial number of Collective Action Clauses containing disenfranchisement clauses.10 For international issues, especially those following New York and English law, disenfranchisement provisions can be considered as a widely accepted standard. These provisions are based on the recommendations of the G10 working group on Contractual Clauses of 2002 which proposed that “(b)onds excluded from the outstanding amount used as a reference for voting provisions are those owned or controlled directly or indirectly by the issuer or its public instrumentalities”.11 Similar clauses are common in issuances of non-sovereign bonds. The US Trust Indenture Act states in sec. 316(a) that “in determining whether the holders of the required principal amount of indenture securities have concurred in any (…) consent indenture securities owned by any obligor upon the indenture securities, or by any person directly or indirectly controlling or controlled by or under direct or indirect common control with any such obligor, shall be disregarded”.12 However, regardless of this widely accepted starting point, disenfranchisement clauses differ significantly in their wording. Some provisions explicitly enfranchise or disenfranchise specific groups of bondholders, others are very general and abstain from qualifying any particular entity as enfranchised or disenfranchised. In some sovereign bond issues of Uruguay, for example, bonds owned by any public sector instrumentality of the issuer are qualified as non-outstanding. According to the terms, a public sector instrumentality is typically a department, ministry or agency of the government of the issuer or any corporation, trust,

10 On the more recent developments see Buchheit/Gulati, 51 Emory Law Journal 1317, 1358–1360 (2002); Buchheit/Gulati, 6 Capital Markets Law Journal (CMLJ) 317, 318 et seq. (2011). 11 See the Report of the G10 Working Group on Contractual Clauses of 26 September 2002, Annex no. 2. Based on that are the Collective Action provisions by the International Credit Market Association (ICMA), see the “Standard Collective Action Clauses for the Terms and Conditions of Sovereign Notes” of Nov 2004, available at http://www.icmagroup.org/ICMAGroup/files/c3/ c37b864f-2279-4cab-9f38-47077deb2cee.pdf (see annex 3). 12 The regulations of the TIA do not apply to sovereign bonds. However, the terms of sovereign bond issuances under New York law usually copy these rules and therefore apply them by choice. See Kenadjian, “Bond issues under New York and U.S. law”, in: Baums/Cahn (ed.), Die Reform des Schuldverschreibungsrechts, 2004, p. 245, 248.

50 

 Christian Hofmann

financial institution or other entity owned or controlled by the government of the issuer and the national central bank.13 In the case of Greece, only a small amount of less than 10 % of all debt securitized in sovereign bonds contains Collective Action Clauses – mainly bonds which hold an XS-ISIN and are subject to English law. For this group of bonds disenfranchisement clauses are in use. These clauses are detailed as they specify the term ‘control’ and exempt certain entities from the disenfranchisement regardless of their proximity to the issuer: Bonds which the issuer owns or controls directly or indirectly will be disregarded and deemed not to be outstanding. Control is defined as the power to, directly or indirectly, direct the management of an entity or to appoint a majority of the board of directors or other persons of an entity that holds bonds. As an exemption to this general approach, bonds owned, directly or indirectly, by the Bank of Greece or any of the Republic’s local authorities and other local authorities’ entities are not regarded as controlled by the Republic, and, consequently, deemed enfranchised.14

IV. Autonomy of decision In many cases the mere control that the issuer exercises over the bondholder is a sufficient justification for a disenfranchisement since the control results in the before described outcome that the bondholder’s own interests are non-existent, inferior or irrelevant. However, the justification for a disenfranchisement is not the control exercised by the issuer, but the likely result of such control – the inability of the bondholder to exercise his voting power independently. This is reflected in the explanatory notes of the G10 which emphasize that disenfran-

13 See the bond terms of Uruguay in the prospectus of April 10, 2003, available at http://www3. bcu.gub.uy/autoriza/sgoioi/prospsup1003ui.pdf and referred to by several authors, e.g. Buchheit/Gulati, 6 Capital Markets Law Journal (CMLJ) 317; Dey, Collective Action Clauses – Sovereign Bondholders Cornered?, Working Paper 2009. 14 The prospectus on the offering of 29 April 2004 (ISIN XS0191352847) provides at 10: “(…) any bonds that the issuer owns or controls directly or indirectly will be disregarded and deemed not to be outstanding. For this purpose, Bonds owned, directly or indirectly, by the Bank of Greece or any of the Republic’s local authorities and other local authorities’ entities will not be regarded as, or deemed to be, owned or controlled, directly or indirectly be the Republic. ‘Control’ means the power, directly or indirectly, through the ownership of voting securities or other ownership interests or otherwise, to direct the management of or elect or appoint a majority of the board of directors or other persons performing similar functions in lieu of, or in addition to, the board of directors of a corporation, trust, financial institution or other entity”.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 51

chisement clauses are intended to address creditor concerns about the manipulation of votes by a sovereign.15 The relevant criteria should therefore be whether it is at least potentially possible that the bondholder’s interests differ from those of the issuer and whether he is capable of pursuing them independently. Therefore, –– if a bondholder can at least potentially have interests which differ from those of the issuer and may pursue these interests regardless of the issuer’s will he should be enfranchised, and –– if the bondholder’s interests are necessarily identical with those of the issuer or if different interests exist but he is incapable of pursuing them, he should be disenfranchised. In other words, the ultimate criterion for voting rights of the bondholder should be whether he is able to cast his vote autonomously. However, the criterion of autonomy of decision is no common standard in Collective Action Clauses (yet). This explains why the voting rights of entities closely linked to the issuer have not been a topic for a wider discussion yet. It seems, however, almost mandatory to analyze the legal situation for constellations of high practical relevance.

V. Enfranchisement and disenfranchisement of public organizations and entities 1. The right to vote of the ECB and the Euro zone NCBs Some sovereigns disenfranchise their own national banks in Collective Action Clauses. This is a particularly common practice in the bond contracts of Latin American countries. It seems that in those countries the central banks are considered as an executive arm of the government without autonomy of decision. There is no uniform legal framework for the national central banks (NCBs) in the Euro zone. Some NCBs are public entities whose capital is fully held by the national government. Consequently, the governing board is appointed by the national government. The German Bundesbank is a representative of this group. According to sec. 2 sentence 2 Bundesbank Act, the capital of the Bundesbank is owned by the Federal Republic of Germany. In addition to that, the national gov-

15 See “Report of the G-10 Working Group on Contractual Clauses” of 26 September 2002, p. 5.

52 

 Christian Hofmann

ernment appoints the members of the executive board: Sec. 7 para. 3 sentence 1 Bundesbank Act regulates that “(t)he members of the Executive Board shall be appointed by the President of the Federal Republic of Germany. The President, the Vice-President and one other member shall be nominated by the Federal Government; the other three members shall be nominated by the Bundesrat (the upper house of Parliament representing the federal states) in agreement with the Federal Government”. Other NCBs are corporations established under private law. Their capital has been raised by shareholders, and these shareholders elect the governing board. The Bank of Greece belongs to this group of NCBs. The General Council of the Bank of Greece is elected by its shareholders. The number of shareholders amounts to roughly 19,000.16 A dominance of the public sector is prevented by Art. 8 para. 5 which determines that “(t)he State, as well as public enterprises, shall not, directly or indirectly, hold shares of the Bank amounting, in the aggregate of such holdings, to more than thirty five per cent (35%) of the nominal issued share capital”. The issue of disenfranchisement is highly relevant for the NCBs of the member states of the Euro currency union. These central banks have acquired substantial amounts of Euro zone sovereign bonds under the Securities Markets Programme.17 The overall holdings currently add up to the amount of 211 bill. €.18 For reasons of clarity: This amount does not include the (significantly larger amount of) bonds held by the central banks as collateral for loans granted to credit institutions in the open market business. These bonds do not grant any voting rights to the NCBs as long as the transferor of the collateral (the so-called counterparty to the NCB) has not defaulted on the duties resulting from the loan agreement.19

In the future, Euro zone bonds will contain disenfranchisement clauses,20 and therefore the issue of voting rights of the NCBs may become relevant. Unlike the IMF or in the future potentially the European Stability Mechanism (ESM), the

16 Information available at http://www.bankofgreece.gr/Pages/en/Bank/shareholders.aspx. 17 Decision of the ECB of 14.5.2010 establishing a Securities Market Programme, (ECB/2010/5), OJ EU 2010 L 124/8. 18 As of Dec 20, 2011. 19 On assets transferred to the NCBs as collateral in open market transactions see Hofmann, 6 CMLJ 456, 459–461 (2011). 20 See European Council document SN 35/1/11 REV1 “Statement by the Euro Area Heads of State or Government” of 9 December 2011 at 15: “(…) standardized and identical Collective Action Clauses will be included, in such a way as to preserve market liquidity, in the terms and conditions of all new euro government bonds”.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 53

Euro zone does not profit from a preferred creditor status and the central banks are therefore treated like any other bondholder. a) “control” of the national governments over the NCBs Where the government owns the capital of the NCB and appoints the governing board, the government and therefore the issuer “controls” the NCB according to common definitions in disenfranchisement clauses. However, the relevant aspect should neither be the aspect of ownership nor the procedure of the appointment but the future status of the appointed board members. As long as they are autonomous in their decisions and also cannot be dismissed from their position by the national governments (which is guaranteed by Art. 14.2 ECB statute), there is no basis for an assumption that the appointment procedure or ownership causes the governing board’s inability to pursue interests different from the government. b) autonomy of decision in Euro zone tasks Article 130 TFEU guarantees the independence of the NCBs of the European System of Central Banks from their governments: “When exercising the powers and carrying out the tasks and duties conferred upon them by the Treaties and the Statute of the ESCB and of the ECB, neither the European Central Bank, nor a national central bank, nor any member of their decision-making bodies shall seek or take instructions from Union institutions, bodies, offices or agencies, from any government of a Member State or from any other body. The Union institutions, bodies, offices or agencies and the governments of the Member States undertake to respect this principle and not to seek to influence the members of the decisionmaking bodies of the European Central Bank or of the national central banks in the performance of their tasks.” As a result, the NCBs may not take instructions from the issuer of the bonds on how to vote and can therefore be regarded autonomous in their decision. This will be relevant for all sovereign Euro zone bonds which the NCBs will buy after the introduction of CACs in sovereign bond issuing terms. The purchases under the Securities Markets Programme have been categorized by the ECB as monetary policy measures.21 These bonds can therefore be regarded as acquired in performance of a Euro zone task and as such covered by the guarantee in art. 130 TFEU.

21 See Decision of the ECB of 14.5.2010 establishing a Securities Market Programme, (ECB/2010/5), OJ EU 2010 L 124/8 at (3); press release of the ECB of 10 May 2010 – ECB decides on measures to

54 

 Christian Hofmann

c) Autonomy of decision in national tasks However, art. 130 TFEU only covers the mandate of NCBs in the ESCB. On top of their functions in the ESCB, national central banks also pursue tasks vested in them by the national legislature. These tasks are regulated by national laws. NCBs have the possibility to acquire sovereign bonds as part of their national asset management as long as the acquisition does not collide with monetary policy. The limit to nationally enticed transactions is set by Art. 14.4 ESCB-Statute: “National central banks may perform functions other than those specified in this Statute unless the Governing Council finds, by a majority of two thirds of the vote cast, that these interfere with the objectives and tasks of the ESCB. Such functions shall be performed on the responsibility and liability of national central banks and shall not be regarded as being part of the functions of the ESCB.”

As a result, the focus for evaluating whether NCBs are autonomous in their decision on the future of bonds acquired in pursuit of national tasks might shift to national laws. Without providing a final assessment of the wording of the statutes of all 17 Euro zone member countries, it can be stated that the vast majority of national laws grants autonomy of decision to the board of directors of the NCBs.22

address severe tensions in financial markets at 1.; speech by José Manuel González-Páramo, ECB executive board member, of 18 May 2010 on “The current phase of the crisis: The phasing-out in late 2009 and early 2010 and the renewed tensions”. 22 To mention a few of the provisions for the 17 Eurozone NCBs: –– Greece: Article 5a of the Statute of the Bank of Greece: “When carrying out the tasks conferred upon them, neither the Bank of Greece nor any member of its decision-making bodies shall seek or take instructions from the government or any organisation. Neither the government nor any other political authority shall seek to influence the decision-making organs of the Bank in the performance of their duties.” –– Malta: Art. 5 of the Central Bank of Malta Act: “(2) In accordance with the Treaty and the Statute, neither the Bank nor any member of the Board or any official of the Bank, when exercising any function, duty or power under this Act, shall seek or take instructions from the Government or any other body.” –– Slovakia: Article 56 of the Constitution of the Slovak Republic: “(1) The National Bank of Slovakia is an independent central bank of the Slovak Republic. The National Bank of Slovakia may, within its scope of power, issue generally binding legal regulations if it is so empowered by a law.” –– Spain: Article 1 of the Law of Autonomy of the Banco de Espana: “(1) The Bank is an institution under public law with its own legal personality and full public and private legal capacity. It shall pursue its activities and fulfil its objectives with autonomy from the administration, carrying out its functions as specified in this law and other legislation.”



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 55

However, even where the wording of the national laws may cast a shadow of doubt on the autonomy of the central bank’s decision-making bodies, the independence in matters of votes on bond terms is nevertheless guaranteed. The reason is that the guaranty provided by art. 130 TFEU covers the entire financial independence of the NCBs – therefore both in ESCB and in national matters.23 The underlying cause is that without thorough financial independence from the national governments the central banks might be limited in pursuing their tasks under the ESCB. The ECB’s Convergence Report of May 2010 states that “(r)ights of third parties to intervene or to issue instructions to an NCB in relation to the property held by an NCB are incompatible with the principle of financial independence.”24 The ECB has given an example for the application of this principle: “A Member State may not impose reductions of capital on an NCB without the ex ante agreement of the NCB’s decision-making bodies, which must aim to ensure that it retains sufficient financial means to fulfil its mandate under Article 127(2) of the Treaty and the Statute as a member of the ESCB.”25 A situation in which the national government ordered the NCB to cast its vote on a modification of the bond terms in a certain way would have a similar effect and lead to a similar level of threat to its financial independence.

d) Other factors creating a conflict of interest? aa) Pursuit of stability There remains one further potential indicator for a disenfranchisement of the central banks (ECB and NCBs). It can be assumed that the central banks’ motivation to vote in favor of a proposed restructuring differs from the motivation of private investors. Whereas private investors pursue commercial interests with their investment in sovereign bonds, the Euro system may intend to secure market, currency and financial stability with its purchases. It is most likely that this purpose also influences the decision on any amendments to the bond terms. However, as argued above (at II. 3.), the motivation of a bondholder cannot be the basis for disenfranchisement. The principle of contract law to leave the

23 ECB Convergence Report May 2010, p. 21–23. 24 ECB Convergence Report May 2010, S. 23. See also the assessment of Lietuvos Banka’s independence on p. 248 according to which the central bank’s institutional and financial independence is undermined since Lietuvos Banka is not the owner of its immovable property and thus not entirely free to decide about it. 25 ECB Convergence Report May 2010, p. 23.

56 

 Christian Hofmann

motivation for commitments entirely to the parties also applies to the process of voting on amendments to the bond terms.

bb)  Impact of art. 123 TFEU Art. 123 para. 1 TFEU prohibits the European central banks to grant overdraft facilities or any other type of credit facility to the member states (and its bodies and undertakings) as well as the direct purchase of debt instruments from them. The statute does not provide any provisions for the exercise of the right to vote by the central banks. However, the discussion on the issue whether the government purchase program by the Euro zone infringes the ban on monetary financing of the public sector could have an implication on the voting rights of the central banks. The scope of Art. 123 para. 1 TFEU has been discussed in the context of the Securities Markets Programme. It has been argued that purchases of Euro zone government bonds by the Euro system had the effect of monetary financing and should therefore be banned.26 Contrary to that, the ECB considers the purchase of sovereign bonds by the Euro zone NCBs on secondary markets compatible with the prohibitions of Art. 123 TFEU as long as they are necessary to “address the malfunctioning of securities markets and restore an appropriate monetary policy transmission process”27 – an opinion widely shared by commentators.28 Therefore, two approaches seem possible. If the scope of Art. 123 para. 1 TFEU is understood in a broader meaning, any debt burden relief granted by the Euro system seems prohibited by the provision. The Euro system would therefore be required to vote no on any restructuring proposal. If the provision was understood in a way that is limited to its wording, the Euro system could enjoy some

26 On the discussion about the Securities Markets Programme of the Euro zone central banks and its compatibility with the prohibitions in art. 136 TFEU see Seidel, Europäische Zeitschrift für Wirtschaftsrecht (EuZW) 2010, 521; Herrmann, Europäische Zeitschrift für Wirtschaftsrecht (EuZW) 2010, 645, 646; Häde, Europäische Zeitschrift für Wirtschaftsrecht (EuZW) 2009, 399, 400 et seq. On the disciplinary approaches underlying the concept of Art. 136 TFEU see Bini Smaghi, speech of 28 May 2010 on “The Group of Thirty”, 63rd Plenary Session, Session I: The Crisis of the Eurosystem. On circumventions of the prohibitions in art. 136 TFEU (formerly art. 104 EC) compare Council regulation (EC) No 3603/93 of 13 Dec 1993 specifying the application of the prohibitions referred to in Article 104 and 104(b) of the Treaty, OJ EC 1993 L 332/1, seventh recital. 27 Citation taken from Decision of the ECB of 14.5.2010 establishing a Securities Market Programme, (ECB/2010/5), OJ EU 2010 L 124/8 at (3). 28 This explains why numerous authors recommend that the ECB participate in restructuring programs, e.g. in a public offering by the EFSF to buy up sovereign debt at a discounted price, on that see Gros/Meyer, Debt reduction without default? CEPS Policy Brief No. 233, Feb. 2011.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 57

freedom in its decision on how to vote its bonds – as long as the vote is in conformity with its monetary policy tasks. Either way the Euro system would have to be regarded enfranchised. Potential limitations to the free exercise of its voting powers would not be connected to the issuer’s interests but result from its institutional responsibilities in the currency union. Any limitations would lie outside the scope of disenfranchisement clauses.

2. The right to vote of publicly owned banks Similar issues arise when government bonds are held by publicly owned banks.

a) Banks owned by the central government (issuer) If the bond issuer owns a public bank, it has the power to send civil servants into the governing board or to appoint external managers as directors of the bank. This indicates that the bank is controlled by the issuer. However, the directors of the bank may not only be entitled, but even forced to pursue interests which are not identical to the interests of the sovereign. The statutes of public banks usually set out general criteria which serve as binding guidelines for the directors’ decisions. Public banks often pursue two statutory purposes: They are business undertakings and as such required to generate profits. As a result, to give up claims for the benefit of the issuer of sovereign bonds would be contrary to the business interests of the bank and constitute a breach of duty by the directors. The second purpose pursued differs from the targets of commercial banks. Public banks regularly also pursue tasks in the best interest of the general public, e.g. to provide payment services and loans to all groups of the population, especially to those for whom access to financial resources provided by the commercial sector is not easily available.29 Neither purpose is compatible with a situation in which the issuer influences the vote on the bonds.

29 In most cases, publicly owned banks will seek to pursue both purposes equally. In some instances, however, the pursuit of profits may be (temporarily) incompatible with the public interest objectives. The US mortgage banks seem to pursue mainly tasks in the public interest and, if incompatible, neglect the pursuit of profit.

58 

 Christian Hofmann

Similar issues arise in private banks which are dominated by the issuer because he holds a voting majority of the stocks. The rescue efforts of European countries in the aftermath of the Lehman collapse have created several such scenarios. Depending on the legal structure of the corporation, the major shareholder can either appoint the executive board directly or choose the members of the supervisory board who for their part appoint the executive directors of the company. The second model is dominant in the continental European stock companies as well as in the Anglo-American style corporations, the first in the continental European “small” corporations (e.g. German Gesellschaft mit beschränkter Haftung [GmbH], French société à responsabilité limitée [s.à.r.l.], Spanish sociedad de responsabilidad limitada [SL], Italian societa a responsabilita limitata [SRL]).30 However, the mere power to appoint the directors of the corporation does not necessarily create control over those directors’ decisions. It can be considered an internationally accepted standard of corporate law that the directors are obliged to pursue the best interest of the company – regardless of the legal structure of the corporation.31 They owe a fiduciary duty to the company which is a combination of the multifold interests of the different groups of stakeholders, including not only the shareholders but other groups such as the creditors and employees of the corporation. To vote in the mere interest of the majority shareholder regardless of the corporate interest would breach this fiduciary duty for which the directors could be held liable by the corporation. This argues in favor of autonomy of decision on the directors’ side. However, there might be constellations in which these formal arguments seem entirely bogus. It goes without saying that the influence of majority shareholders on the directors rises with the amount of shares they own. This is true for any kind of corporation including banks. Therefore, a mere formal analysis of the legal formalities may be insufficient where the majority power of one shareholder seems well established.32 Such a scenario requires an analysis of the actual situa-

30 These two types of corporations are different from the Anglo-American style publicly and privately held corporations, on those see J. Ferrar/B. Hannigan, Ferrar’s Company Law, 4th ed. 1998, p. 42–44. 31 General principle of company law, see sec. 76 German Corporate Code (Aktiengesetz). On this principle generally P. Davies, Gower and Davies: The Principles of Modern Company Law, 8th ed. 2008, p. 506–525; E. Ferran, Company Law and Corporate Finance, 1999, p. 157 et seq. 32 On this P. Davies, Gower and Davies: The Principles of Modern Company Law, 8th ed. 2008, p. 366: “shareholder-centered nature of British company law”; B. Cheffins, Corporate Ownership and Control, 2008, p. 30 et seq.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 59

tion in order to decide whether the directors can indeed be considered independent from the majority shareholder and, as a result, autonomous in their decisions. Past situations in which the interests of the majority shareholder were affected may serve as indicators. Majority shareholders in the “small” corporation are even empowered to factual dominance by law. The applicable corporate codes usually entitle the shareholders to issue directions with binding effect on the directors.33 What is more, the shareholders can even declare the directors incompetent and decide in their place. By doing so, the dominant shareholder dominates the executive business of the corporation. He is, in theory, bound by the corporate interest and, arguably, owes a fiduciary duty to the company as well. However, from practical experience reflected in the numerous court cases on the matter we have learned that the majority shareholder does not differentiate between the interests of the corporation and his own. This leads to the conclusion: It is impossible to enfranchise or disenfranchise issuer dominated banks in general. The clauses need to be designed in a discretionary way to determine on a case-to-case basis whether the bank should be entitled to vote on the sovereign bonds it holds.

b) Banks owned by local authorities In a federal system or generally a system that grants a certain level of sovereignty to municipalities, some public banks are not owned by the central government and therefore the issuer of the bonds, but by local authorities. This can be illustrated by the German example: In the German federal system there are three levels of authority, the central government of the Federal Republic, the regional authorities of the 16 states and the local authorities of municipalities. If banks are owned by the states or municipalities (which regularly applies to savings banks) the owner is not identical with the issuer of the bonds. What is more, the board members of these banks are appointed by the regional or local authorities and therefore not under the issuer’s control. This leads to the conclusion that there are no grounds for the disenfranchisement of these bonds. However, one could argue that regardless of the formal exer-

33 As an example compare sec. 37 para. 1 of German GmbHG. See also P. Davies, Gower and Davies: The Principles of Modern Company Law, 8th ed. 2008, p. 366.

60 

 Christian Hofmann

cise of ownership and control these constellations should be treated equally to cases in which the bank is owned by the federal government. This would have to be based on the argument that the interests of the central and regional as well as local governments are identical and therefore any control on those banks would be exercised in the same way as if the issuer was the owner and controller of the banks. But such reasoning seems rather weak since state and municipal authorities can and often do pursue interests which (significantly) deviate from those of the federal government. The only interest that all those authorities really share is the common interest in avoiding a sovereign default since all levels of the public hand would stand to lose from such an incident. But once more this common interest is nothing more than the mere motivation of the bondholders enticing them to vote in favor of the issuer’s proposal. A city may wish to avoid a sovereign default because it fears that it will suffer from a drastic decline of its share in the public household and therefore be forced to close public facilities. However, this indirect financial dependence on the central government is no sign that it lacks the power to pursue its own interest. Its voting decision is merely influenced by the fact that the default of a sovereign sends out shockwaves which affect every entity within its territory (and in many cases numerous beyond its borders as well). As a result, these banks will usually be considered enfranchised. Yet, here as well the situation will have to be decided on a case-to-case basis. This means that in certain cases these bondholders must be deemed disenfranchised because the dominance of the issuer seems to affect the decision of the directors disenabling them to pursue their own interests.

3. The right to vote of other Euro zone member states Member States of the Euro zone may be the holders of Euro zone government bonds. The most likely scenarios involve once more public banks. If such banks hold government bonds and do not pass the autonomy test according to the above described criteria (at 2 a), the bonds are de facto held by a Euro zone sovereign. In this case, restrictions on the right to vote do not result from the disenfranchisement clauses in the bond terms. The issuer of the bonds and the sovereign that controls the bondholder are different Euro zone member states, and as a consequence there is no imminent danger that the bondholder might be forced to vote in the issuer’s best interest. The issue results from the “no bail out” clause in art. 125 TFEU. Sentence 2 provides that “(a) Member State shall not be liable for or assume the commitments of central governments, regional, local or other public authorities, other



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 61

bodies governed by public law, or public undertakings of another Member State, without prejudice to mutual financial guarantees for the joint execution of a specific project.” This ban on financial help from one member state to another would not result into a disenfranchisement of the bonds, but it might require a vote conformant with the requirements and therefore prevent a member state from approving any proposed relief for the issuer because it could be interpreted as a bail out. The answer to this issue is tied to the general approach to Art. 125 TFEU. If the imminent ban is limited to the immediate wording of sentence 2, it covers only the liability for foreign sovereign debt as well as the assumption of the commitments of other sovereigns. A debt restructuring would lead to a (partial) waiver for the obligations of one member state to another. One could argue that such a waiver shows similar effects and should therefore be covered by the ban. However, the existing rescue mechanisms in the Euro zone have been based upon a more restricted interpretation of Art. 125 sentence 2 TFEU pursuing the argument that a bail out is limited to an assumption of foreign debt as expressed by the wording of art. 125 sent. 2 TFEU. Loans granted by the Union or the Euro zone member states leave the existing debt unaltered and can therefore not be considered bail-outs.34 The opposite opinion emphasizes the reasons behind all provisions referring to financial help for member states. They consider household discipline and its enforcement by the power of the markets as the predominant founding principles of the currency union. The measures provided for in Art. 122 para. 2, 143 TFEU permit to grant assistance to member states in extraordinary situations. They should be regarded as exceptional and exclusive. Therefore Art. 125 TFEU should be interpreted as a general ban on any kind of financial assistance for member states in any other than the explicitly permitted scenarios.35 If this approach was taken, the ban of Art. 125 Sentence 2 TFEU would in any case prohibit every member state to vote yes on a restructuring proposal.

34 In this respect Herrmann, Europäische Zeitschrift für Wirtschaftsrecht (EuZW) 2010, 413, 415; Louis, CMLR 2010, 971, 985. 35 Along these lines Häde, Europäische Zeitschrift für Wirtschaftsrecht (EuZW), 2009, 399, 403; Knopp, Neue Juristische Wochenschrift (NJW) 2010, 1777, 1779; Faßbender, Neue Zeitschrift für Verwaltungsrecht (NVwZ) 2010, 799, 800; i.E. auch Schröder, Die öffentliche Verwaltung (DÖV) 2011, 61, 62 ff.

62 

 Christian Hofmann

4. The right to vote of the EFSF Restrictions on the voting rights of the European Financial Stability Facility could be based on similar reasons as discussed for other Euro zone member states. The EFSF is a Luxembourg-registered company owned by the Euro zone member states.36 According to the latest decisions by the Euro group members the EFSF will engage in sovereign Euro zone bond purchases both in the primary and the secondary market. Formally, the EFSF is a private law entity and therefore not covered by the restrictions in Art. 125 TFEU. De facto, however, the EFSF is a substitute for the Euro zone member states. Therefore, the outlined issues about a potential conflict with the no bail out rule could apply to the EFSF as much as to individual member states. In addition to that, with its bond purchases the EFSF pursues interests which differ from those of private investors. Its future bond purchases will seek to provide market confidence and by that bring down the financing cost of the issuer. This motivation will also prevail in a vote on an amendment of the bond terms. Here, what applies to central banks applies to the EFSF as well. The motivation for the vote is outside the scope of disenfranchisement clauses.

VI. Transparency requirements Having established the justification requirements and applied them in various relevant situations, there remains one important aspect. The rules on disenfranchisement are of no practical value unless they are efficiently applied. Efficiency requires transparency of both disenfranchised and enfranchised bonds which means that the substantial underlying information must be disclosed to the bondholders.37 Since this information is (only) available to the issuer, the issuer must be the addressee of these disclosure duties.

36 See the EFSF Framework Agreement at B as well as the EFSF Articles of Incorporation at 1.1. These and other legal documents on the EFSF are available for download at http://www.efsf. europa.eu/about/legal-documents/index.htm. 37 On the significance of transparency by disclosure about ownership and control see Buchheit/ Gulati, 6 CMLJ 317, 323 (2011). Buchheit/Gulati, 28 Int’l. Finan. L. Rev. 22, 24 et seq. (2009), uses the example of Ecuador to illustrate how disenfranchisement clauses are pointless without appropriate disclosure.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 63

1. Disenfranchised bondholders The issuer should be required to provide a list of all entities holding its government bonds and required to pursue its interest. This list must be made public prior to any voting and will tell the representative of the bondholders who is in charge of organizing and conducting the vote which entities will be disenfranchised in the vote. Prior to this communication, the sovereign and the affected entities must have come to an agreement on whether the entity benefits from autonomy of decision and might therefore be enfranchised in spite of indications that it might be under the issuer’s control. For the other bondholders it is important to know which entities are disenfranchised. It helps to evaluate the chances for the proposal to pass. Another list, however, seems even more important. The destiny of the sovereign debt and the future of the bondholders’ claims may be decided by the enfranchised bondholders. It is therefore more important to know who holds bonds and will be entitled to vote.

2. Enfranchised bondholders A thorough list of all enfranchised bondholders would satisfy this interest for disclosure best. A duty to provide such a list would, however, seem an impossible task for the issuer. The bonds are freely tradable securities. The primary purchasers may pass them on, and they are not required to inform the issuer about these transactions. In addition to these practical obstacles, privacy protection is another issue. To provide a thorough list of all enfranchised bondholders would require bondholders who wish to remain anonymous to disclose their holdings. As a result, the disclosure should cover the group of entities which the bondholders will consider the most problematic: entities which hold close ties with the issuer’s government but are considered autonomous in their decision and therefore enfranchised. These entities are not included in a list of disenfranchised institutions because the issuer has dismissed the idea of its control over the entity. The creditors may disagree with the sovereign on this assessment, and including these entities in the list will enable the bondholders to intervene prior to a vote or to base legal action in the aftermath of the vote on the enfranchisement of those entities. Major interests of the entities which would argue against including them in the list seem far-fetched. Aspects of privacy protection seem of minor importance for entities closely affiliated to the sovereign. The list should therefore name all the bonds held by bondholders which are owned or controlled by the issuer. It should then specify which of those are con-

64 

 Christian Hofmann

sidered enfranchised and explain the underlying reasons, i.e. clarify why these entities are deemed autonomous in their decisions.

VII. Lack of disenfranchisement provisions One further question remains: What rules apply if disenfranchisement clauses do not exist? Does a lack of provisions in the bond terms grant the issuer the right to control the vote by using entities it owns and controls? The answer should be no. Precedents on this issue do not exist although a New York case based on similar facts seems to favor enfranchisement.38 In that case Brazil first provided one of its state-owned banks with a majority voting power and afterwards used this voting power to block an acceleration of the unmatured principal. The court clung to the wording of the contract which did not provide for any disenfranchisement of creditors affiliated with the debtor.39 Contrary to the opinion of the court, the course of action of the sovereign should be considered a manipulation of the vote. This result seems not only mandatory to guarantee the intended general acceptance of the majority vote on bond terms, it also seems the only feasible conclusion from a legal point of view: To use one’s own voting power to influence the result of a vote on one’s own debt seems incompatible with internationally recognized contractual standards such as the rule of good faith and fair dealing.40

VIII. The Euro zone plans The Euro zone has decided to include Collective Action Clauses into the contracts of any bonds issued by a Euro zone member from January 2013 onwards. The draft on Euro zone model Collective Action Clauses available at the time of writing this

38 CIBC Bank and Trust C. (Cayman) Ltd. v. Banco Central do Brasil, 886 F. Supp. 1105 (S.D.N.Y. 1995). 39 On the judgment in detail see Buchheit/Gulati, 51 Emory Law Journal 1317, 1340 (2002). 40 This is also the result of the evaluation by Buchheit/Gulati, 51 Emory Law Journal 1317, 1340 (2002). See also the contribution by Lechlan Burn in this volume, p. 73–83.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 65

contribution contains provisions on the disenfranchisement of bondholders. This draft requires a test of two steps to determine the disenfranchisement of the bondholder. First, the draft lists the scenarios in which the control by the issuer can be assumed. The bond must be “held by the issuer, a department, ministry or agency of the issuer, or a corporation, trust or other legal entity controlled by the Issuer or a department, ministry or agency of the Issuer”. In a second step, it must then be verified that the bondholder does not have autonomy of decision.

1. The wording of the draft The draft does not name specific institutions which are enfranchised or disenfranchised, but approaches the conflict by providing a general rule. The exact wording of the disenfranchisement clause (clause 2.7 of the current draft) is as follows: “In determining whether holders of the requisite principal amount of outstanding Bonds have voted in favour of a proposed modification or whether a quorum is present at any meeting of Bondholders called to vote on a proposed modification, a Bond will be deemed to be not outstanding, and may not be voted for or against a proposed modification or counted in determining whether a quorum is present, if on the record date for the proposed modification: (a) the Bond has previously been cancelled or delivered for cancellation or held for reissuance but not reissued; (b) the Bond has previously been called for redemption in accordance with its terms or previously become due and payable at maturity or otherwise and the Issuer has previously satisfied its obligation to make all payments due in respect of the Bond in accordance with its terms; or (c) the Bond is held by the Issuer, by a department, ministry or agency of the Issuer, or by a corporation, trust or other legal entity that is controlled by the Issuer or a department, ministry or agency of the Issuer and, in the case of a Bond held by any such above-mentioned corporation, trust or other legal entity, the holder of the Bond does not have autonomy of decision, where: (i) the holder of a Bond for these purposes is the entity legally entitled to vote the Bond for or against a proposed modification or, if different, the entity whose consent or instruction is by contract required, directly or indirectly, for the legally entitled holder to vote the Bond for or against a proposed modification; (ii) a corporation, trust or other legal entity is controlled by the Issuer or by a department, ministry or agency of the Issuer if the Issuer or any department, ministry or agency of the Issuer has the power, directly or indirectly, through the ownership of voting securities or other ownership interests, by contract or otherwise, to direct the management of or elect or appoint a majority of the board of directors or other persons performing similar functions in lieu of, or in addition to, the board of directors of that legal entity; and

66 

 Christian Hofmann

(iii) the holder of a Bond has autonomy of decision if, under applicable law, rules or regulations and independent of any direct or indirect obligation the holder may have in relation to the Issuer: (x) the holder may not, directly or indirectly, take instruction from the Issuer on how to vote on a proposed modification; or (y) the holder in determining how to vote on a proposed modification, is required to act in accordance with an objective prudential standard, in the interest of all of its stakeholders or in the holder’s own interest; or (z) the holder owes a fiduciary or similar duty to vote on a proposed modification in the interest of one or more persons other than a person whose holdings of Bonds (if that person then held any Bonds) would be deemed to be not outstanding under this Section 2.7.

The transparency requirements are listed at 3.2. They provide: “The Issuer will provide to the calculation agent and publish prior to the date of any meeting called to vote on a proposed modification or the date fixed by the Issuer for the signing of a written resolution in relation to a proposed modification, a certificate: (a) listing the total principal amount of Bonds and, in the case of a cross-series modification, debt securities of each other affected series outstanding on the record date for purposes of Section 2.7; (b) specifying the total principal amount of Bonds and, in the case of a cross-series modification, debt securities of each other affected series that are deemed under Section 2.7(c) to be not outstanding on the record date; and (c) identifying the holders of the Bonds and, in the case of a cross-series modification, debt securities of each other affected series, referred to in (b) above, determined, if applicable, in accordance with the provisions of Section 2.6.”

2. scenarios without autonomy of decision At (a) to (c) (ii) the draft lists several cases in which the bondholder lacks autonomy of decision. Included in this list are scenarios in which the bondholder needs the consent or instruction of another entity to cast a vote. The bondholder is disenfranchised if he is controlled by this other entity and not free to cast his vote autonomously. In a different scenario, the issuer controls the bondholder because it has the power to direct the management of that legal entity or to elect or appoint a majority of the board of directors or other persons performing similar functions as the board of directors in that legal entity. This control can be exercised, directly or indirectly, through the ownership of voting securities or other ownership interests and by contract or otherwise.41

41 The other constellations in which the draft orders a disenfranchisement are of minor impor-



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 67

3. Scenarios with autonomy of decision The draft presumes autonomy of decision where the bondholder under applicable law –– may not take instructions from the issuer on how to vote, or –– is required to act in accordance with an objective prudential standard, or –– is not directly or indirectly wholly owned by the issuer and required to act in its own interest, or –– owes a fiduciary or similar duty to vote on a proposed modification in the interest of one or more persons other than the Issuer, any department, ministry or agency of the Issuer or any corporation, trust or other legal entity controlled by the Issuer or any department, ministry or agency of the Issuer.

4. Assessment of the draft The approach of the draft to limit the wording of disenfranchisement clauses to general provisions and to stay silent on the situation of specific institutions seems well justified. As illustrated above (at IV. 2.), the situation of public banks, for example, depends entirely on the national regime. It seems impossible to generally enfranchise all savings banks in the Euro zone. Whereas it seems very likely that the vast majority will be enfranchised due to the reasons given above, there might be exceptional circumstances which support a disenfranchisement. The same applies to other entities closely linked to the issuer. This major advantage compensates for the obvious disadvantage of such a general approach. It provides less clarity than an explicit list of disenfranchised and enfranchised institutions in the bond terms. The competence to provide such a list shifts to the issuer as the one institution positioned best for such a task. The Euro zone draft does exactly that but falls short of the requirements supported here: according to the draft the list must name the disenfranchised bondholders whereas the enfranchised bondholders are not mentioned. What is more, the issuer provides a list (now called certificate) of all disenfranchised bondholders

tance and seem self-explanatory: the Bond was previously cancelled or delivered for cancellation or held for reissuance but not reissued, or the Bond has previously been called for redemption in accordance with its terms or previously become due and payable at maturity or otherwise and the issuer has previously satisfied its obligation to make all payments due in respect of the Bond in accordance with its terms.

68 

 Christian Hofmann

to the calculation agent. The agent relies on the correctness of the list, excludes all bondholders on the list and admits everyone else. The imminent danger is obvious. The power of decision shifts to the very party which holds the major interest in the enfranchisement of all entities which will vote in its favor. If the issuer is not required to also name bondholders that are deemed autonomous in their decision, the bondholders have reason to distrust the list and therefore the voting procedure. There is another aspect that requires close attention. Because of the issuer’s bias, the bondholders must be entitled to challenge the list. This requires efficient remedies against the list or at least against the outcome of the vote.

IX. Remaining issues Disenfranchisement clauses do not solve all issues related to the principle of majority voting. For all their benefits, majority provisions come at a price. The major concern related to them is their imminent risk for an abuse of the majority power. The abuse of majority power is a well-known phenomenon in corporate law that sometimes the legislators, more often the courts have been required to find answers for.42 By introducing majority provisions, Collective Action Clauses embrace this danger. It seems possible that enfranchised bondholders abuse their majority power to the detriment of the dissenting minority. Possible scenarios involve situations in which the majority stands to gain from a reduction of the issuer’s bond debt, therefore situations of indirect advantages which lead to an indirect inequality of treatment of the bondholders. Such situations are not covered by disenfranchisement clauses as long as the majority bondholders are not controlled by the issuer and therefore capable of pursuing their own interest. On this matter principles seem difficult to find. The abuse of majority power is a controversial matter in corporate law. What makes it even more difficult in the context of sovereign bonds is the fact that the bondholders are not bound by corporate ties which oblige to pursue the corporate purpose and can justify restrictions which otherwise do not apply to contractual partners. The abuse of majority power is a matter which has not been tackled by the Euro zone draft and therefore will have to be solved by national law. What is

42 The most prominent example being corporate law where the courts have built up a complex case law construction. See P. Davies, Gower and Davies: The Principles of Modern Company Law, 8th ed. 2008, p. 649–663.



Enfranchisement and Disenfranchisement in Collective Action Clauses  

 69

more, it has not been discussed widely so far. It seems that potential relief for disadvantaged minority bondholders might be granted on the basis of the abovementioned (at VI.) principles of good faith and fair dealing.

Collective Action Clauses and Litigation

Lachlan Burn

Bondholder Resolutions in the Courtroom I have been asked to discuss, in brief terms, how the operation of provisions such as the proposed eurozone collective action clause (or CAC) might be attacked in a court of law. I am qualified to advise on English law, and will not venture into other territory. I appreciate that this timidity on my part will appear to render much of what I have to say irrelevant, on at least two grounds. First, I suspect that hardly any eurozone sovereigns will be issuing bonds governed by English law. And, secondly, given that the United Kingdom is not part of the eurozone, it may well not incorporate the eurozone CAC in its bonds; so it will be of academic interest only to know how English law might apply to such a provision, as it is unlikely that it ever will1. However, some of the concepts that I will describe in the next twenty minutes or so may be of interest purely by way of comparison or indication of how lawyers might approach these matters. Much of the English case law in this area has, at its root, the desire to prevent bad faith and outright, deliberate unfairness without any other counterbalancing object; and I suspect that the laws of many other countries will have the same objective.

Contractual certainty Many civil lawyers will have come across the claim that common law (whether the English variety or the other varieties exported to parts of the USA and elsewhere) accords certainty to contracts – what you see on the page is what the courts will enforce. This contrasts with some other legal systems, that introduce concepts such as requirements for fairness and good faith, which can, in some cases, allow the court to readjust the terms of the contract to rebalance the bargain where it has been disturbed by events subsequent to its conception. Broadly speaking, I think that the claim for English law is true. Consumer contracts apart (which are subject to all sorts of statutory restrictions prohibiting unfair terms and the like), where commercial individuals or entities enter freely

1 It turns out that my misgivings were somewhat misplaced, in that the bonds offered by Greece in its recent exchange offer are governed by English law. However, it remains to be seen whether, and to what extent, this practice is adopted by other Eurozone Sovereigns; and I remain sceptical as to a wholesale adoption of English law in this market.

74 

 Lachlan Burn

into a bargain, the courts will not willingly interfere. The presumption is that they will have known what they were about when they entered into the contract and the courts should not second-guess their intentions or overturn their bargain. But notice the qualifications in what I have just said. “Broadly speaking”; “the courts will not willingly interfere”. The implication is that there are circumstances where the courts will intervene and it is on these that I want to concentrate in the next few minutes. I will try to illustrate the concepts with some relevant English cases (none, I fear, directly on the subject of CACs but some are closely analogous). I will then try to tease out some examples of how those concepts might apply in the context of a eurozone CAC.

Abuse of power I thought that a useful starting point would be what I take to be the legal root of the possible attack on a majority decision by the minority that is supposedly bound by it. I think that this is to be found in equitable principles that regulate the use of power by decision makers and prevent its abuse. Professor Sarah Worthington puts the concept well in her book on Equity2: “[The rules that regulate abuse of power by decision-makers] operate far more widely than Equity’s fiduciary rules. Their aim is not merely to address fiduciary self-interest or disloyalty; it is to regulate the exercise of all powers where the claimant’s interests may be compromised by the defendant’s decision. . . [T]hese powers do not prescribe decision-making strategies; they proscribe. This rather negative aim is designed to ensure that decisions are not grounded on irrelevant considerations or directed at achieving unacceptable ends.”

Professor Worthington then goes on to give an example involving the issue of shares by a company to alter voting majorities to enable one shareholder to be better placed to mount a takeover bid for the company than another shareholder – which was held to be an improper use of the power to issue shares, the proper use being, for example, the raising of capital for business expansion3. One way of limiting scope for abuse of power is to tightly circumscribe the power. So, in the context of a CAC, some of the power to make decisions is removed from the general power and becomes a “reserved matter”, which requires a larger majority. Or, in the case of the aggregation provision, as I understand it, the

2 Equity by Sarah Worthington, Clarendon Law Series. OUP 2nd ed. at page 142. 3 Howard Smith Ltd v Ampol Petroleum Ltd [1974] AC 821.



Bondholder Resolutions in the Courtroom 

 75

majority across a number of issues brought into the aggregation is subject to a confirming vote by individual issues, for it to be binding on them. However, it is not always easy, or even possible, to circumscribe in such a way as to deal with every possible circumstance that might arise in the future. Therefore many powers are drafted with considerable discretion given to those who are to exercise them. In this case, abuse is controlled by an overriding requirement that the power has to be exercised in good faith. For this purpose, bad faith is presumed (where it cannot be proved directly) if no reasonable person in similar circumstances could have made the same decision. Finally, there is the argument that powers must be exercised for proper purposes, after consideration of the relevant facts. So, the issue of shares by a company to someone who will vote in favour of a particular proposal purely so that he will be able to swing the vote will probably be found to be an improper use of the power of a company to issue shares. Contracts under English law may enjoy certainty, therefore; but that does not mean that a power (such as the power of the majority to bind the minority under a CAC) can be exercised in any way the majority (or the issuer) chooses. There may well be circumstances where an exercise of that power is so manifestly abusive that the English courts will interfere. In many ways, this is not an abrogation of the notion that contracts are enforced in accordance with their terms. It is, rather, merely a recognition that when the parties enter into contractual relations that include provisions such as, say, the CAC, they are not saying “if I turn out to be in the minority, I consent to your using your majority powers as unfairly as you like to abuse me as you see fit”. Indeed, if the bond said that in terms, investors would most likely object vociferously. So all the court is doing, in intervening and disallowing an abusive use of the majority power in enforcing the contract, is enforcing the terms of the contract (including those that are necessarily implied).

Some illustrations That is, I think, the main root on which an English lawyer would grow the argument against a majority decision under a CAC. As with most theory (particularly the legal sort) it sounds (and probably is) arid and difficult to digest. So some illustrations4, drawn from English case law, may be helpful at this point.

4 Between the delivery of this paper and its inclusion in the current volume a further interesting case has been decided in the High Court, namely Assénagon Asset Management S.A. v Irish Bank

76 

 Lachlan Burn

Perhaps the most helpful case is Redwood Master Fund, Ltd and Others v TD Bank Europe Limited and Others5. The facts are complicated, so the following summary is highly (I hope not over) simplified. In essence, the case involved a syndicated loan facility made to a company (the Borrower). The facility had three Tranches – Tranche A, which was a revolving credit facility (to be used for things like working capital requirements); Tranche B, which was a term loan facility (to be used for capital expenditure and acquisitions); and Tranche C (to be used to refinance certain other loan facilities). The original lenders under the facility agreed to participate in Tranches A and B in fixed proportions. However, as time ran on, trading in the secondary market resulted in these proportions being disturbed, so that some lenders in Tranche A had a much greater proportional commitment than they had in Tranche B – (in some cases, indeed, a Tranche A lender had no Tranche B involvement). The loan agreement contained a provision somewhat similar to the proposed eurozone CAC, in that it contemplated amendments being made to the agreement by majority vote, by value, across all three Tranches, with some matters (such as change to the amounts repayable and the maturity date) subject to unanimous vote. The Borrower got into financial difficulties. A steering committee of banks negotiated a restructuring arrangement with the Borrower, which was put to, and passed by, the requisite majority of the lending banks. However, the proposal involved (among other things) the reduction of the Tranche A and B commitments and loans. To achieve this in the proportions that the Borrower thought to be sustainable by it, the undrawn Tranche A loans were to be used to repay some of the outstanding Tranche B loans. Had the original proportions of Tranche A and Tranche B commitments and loans been maintained, this would not have been a problem – each bank would be lending in proportionately the same amount to

Resolution Corporation Limited [2012] EWHC 2090 (Ch). Briefly, the case involved an exchange offer of new notes for old. Acceptance of the exchange was required shortly before the date for a meeting of holders of the old notes. As part of the exchange terms, holders who accepted the offer were bound to vote in favour of a resolution put to that meeting, the main object of which was to reduce, very significantly, the payments due on the old notes to those (the minority) who had refused the offer. The case was argued on a number of grounds, including oppression of a minority by the majority. Mr Justice Briggs held that the vote by the majority was oppressive. In his words “the exit consent is, quite simply, a coercive threat which the issuer invites the majority to levy against the minority, nothing more or less. Its only function is the intimidation of a potential minority, based upon the fear of any individual member of the class that, by rejecting the exchange and voting against the resolution, he (or it) will be left out in the cold”. 5 [2002] EWHC 2703 (Ch)



Bondholder Resolutions in the Courtroom 

 77

repay its Tranche B loan; everybody would be treated the same. However, as some Tranche A banks had only very small, or even no Tranche B loans, they were in a different position from those who had maintained their proportions. Essentially, they would be lending new money so that other banks, more heavily committed under Tranche B, could be prepaid. And, to enable this to happen, the loan agreement was modified, pursuant to a majority vote of all the lenders. The main changes were a revision of the provision of the loan agreement that restricted use of Tranche A to working capital purposes. This was rewritten to permit use of Tranche A loans to repay Tranche B loans. In addition, certain defaults by the Borrower, that would have permitted the Tranche A lenders to refuse to make further loans, were waived, again by a substantial majority vote of all the lenders. Perhaps understandably, some of the Tranche A banks refused to sign the waiver agreement with the Borrower and refused to make the Tranche A loans when called upon to do so. Instead they took the case to court. In the words of the judge, Rimer J: “the claimants’ main point of complaint … is that the modified waiver letter discriminated against the A lenders as a class by imposing discriminatory disproportionate overall reductions in A and B, and, at a time when the future viability of [the Borrower] was in question, subjected the A lenders as a class … to an unfair exposure to risk solely for the purpose of removing an equivalent risk from the B lenders”.

The judge, on the evidence, rejected the notion that the majority banks (or the steering committee) had acted in bad faith in negotiating or voting for the restructuring. In other words, the second method for controlling abuse of power discussed above was not available. Rather than trying to “improve their position at the expense of the A lenders” or being motivated by “vindictiveness or malice towards the claimants and other A lenders”, the judge found that the deal that was eventually struck with the Borrower was the result of a proper two-sided negotiation that had, as its object, a restructuring of the Borrower’s debts that gave it the best chance of survival. This was not abusive, but a proper exercise of power. But it is interesting to note what might have constituted bad faith – vindictiveness or malice towards the claimants, or pursuit of one’s own interest at the deliberate expense of others. Or, as the judge says when describing the argument on the part of the Borrower: “it is no part of [their] case that [the relevant provision] can or should be regarded as conferring on the majority an unfettered power to act arbitrarily, capriciously or oppressively and [they] fully accept that the power is impliedly subject to certain restrictions in the manner of its exercise”.

78 

 Lachlan Burn

Again, the implication is that “arbitrary, capricious or oppressive” conduct might be problematic. These are strong terms that imply serious misconduct on the part of the person accused of abusive behaviour. If this is what amounts to “bad faith”, then it is not engaged in unknowingly or accidentally. Another interesting finding in this case is that the fact that the majority decision might result in different treatment for the different Tranches of the loan was not, in itself, abusive. If the majority voting provision was only to operate when all of the three Tranches were in equal positions, it would be deprived of much of its effect (because it was almost inevitable that some lenders in some Tranches would be worse off as a result of its operation). In the judge’s view, on the facts of the case, the majority voting provision was intended “to enable the majority of lenders to make decisions binding on all three classes, even though it might perhaps be capable of being said by one class that the decision could not be said to be for its benefit”. He then goes on to say “by signing up at the outset, each lender submits to the decision of the majority lenders at important forks in the road”. Provided there is no bad faith or improper purpose in the exercise of the majority powers, the court would not intervene. In this case, the motivation of those negotiating and voting for the restructuring was to allow the Borrower to continue to trade – not to victimise or punish the Tranche A lenders – and that was proper. So much for Redwood. Another case that is, I think, interesting is British America Nickel Corporation, Limited and Others v J. O’brien, Limited6. This did actually concern the majority voting provisions in a bond. In brief, a majority vote of bondholders was swung in favour of a resolution authorising an exchange of new bonds for existing bonds by a bondholder whose support for the resolution was obtained by the promise of a large block of ordinary stock. Viscount Haldane said: “To give a power to modify the terms on which debentures [bonds] in a company are secured is not uncommon in practice. The business interests of the company may render such a power expedient, even in the interests of the class of debenture holders as a whole. The provision is usually made in the form of a power … upon the majority of the class of holders … There is, however, a restriction of such powers, when conferred on a majority of a special class in order to enable that majority to bind a minority. They must be exercised subject to a general principle …; namely, that the power given must be exercised for the purpose of benefitting the class as a whole, and not merely individual members only.”

6 [1927] AC 369.



Bondholder Resolutions in the Courtroom 

 79

The judge held that the bondholder whose vote swung the decision was voting for his own personal benefit (to obtain the promised stock) and not purely as a bondholder, with the interests of other bondholders in mind. Therefore, the vote was void. The final illustration that I will offer to you is another case involving majority voting in the context of a bond – Goodfellow v Nelson Line (Liverpool), Limited7. Briefly, this case involved a bond issue that was jointly guaranteed by two entities. One of them subsequently became insolvent and its liquidator asked for the guarantee to be released in return for giving up certain rights. The issuer, believing that this was in its own best interests, offered bondholders another bond in exchange for their existing holdings. The new bond offered a higher interest rate, but was unguaranteed. The required majority vote in favour of the exchange could only be met if the surviving guarantor (who was also a bondholder) voted in favour. The proposals therefore included a provision that gave the surviving guarantor a generous payment for giving up its rights as guarantor (things like the annual guarantee fee); and the surviving guarantor duly voted in favour of the proposal. The validity of the vote was contested by some of the other bondholders. Interestingly, the case was argued on the grounds that the payment to the surviving guarantor was a bribe and therefore the surviving guarantor’s vote was invalid (bribery being a clear example of bad faith). On the facts, the judge, Parker J, disagreed with this and the vote was held to be valid. In doing so, he said: “The powers conferred by the trust deed on a majority of the debenture-holders must, of course, be exercised bona fide, and the Court can no doubt interfere to prevent unfairness or oppression, but, subject to this, each debenture-holder may vote with regard to his individual interests, though these interests may be peculiar to himself and not shared by other debenture-holders … Further, where, as in this case, there is, as between different holders, a diversity of interest, it may be necessary or advisable, as a matter of business fairness, to make special provision for special interests, and I do not think there is any equity precluding a debenture-holder voting for or against a scheme containing such special provision merely because he is interested thereunder.”

Again, the keystone of a successful argument against a majority vote is “bad faith”. Note also the references to the court’s power to intervene to “prevent unfairness or oppression”. But, equally important, note also that a bondholder can vote even

7 [1912] 2 Ch. 342.

80 

 Lachlan Burn

though his personal position is different from that of other bondholders – in that, for example, the proposal includes a benefit for him that others do not share. Thus, to take a modern example, a bondholder who is also on the right end of a credit default swap would not, simply by virtue of that fact, be prevented from voting. Nor would he have to avert his mind from the swap when voting.

The CAC context That is probably enough English case law for our purposes today (or, for many, for ever). To conclude, therefore, I will endeavour to put some of the concepts that I have outlined into the context of the proposed eurozone CAC. Before doing so, though, I want to make a couple of observations, some of which have already been touched on today. First, I think that it is very important to recognise that the eurozone CAC is somewhat different from the equivalent provision in a corporate bond or even a non-eurozone CAC. This is due to the fact that each eurozone state has the same currency – the euro. As a direct consequence of this, the potential divergence of, or competition between, the interests of different groups of bondholders is aggravated. Most bonds with majority voting provisions will disenfranchise the issuer and the issuer’s creatures (subsidiaries and the like) on the basis that they will (obviously) vote in favour of anything that benefits the issuer, however disastrous it might be for bondholders. But in the eurozone context the categories of investors who will be aligned with the issuer in this way are not limited to entities under the issuer’s control. They will include entirely separate bodies, such as other eurozone states and their agencies and supra-national organisations, such as some of the EU institutions. They may also include non-EU sovereigns for whom the euro is a reserve currency and the EU is a vital trading block and who will therefore disregard their interests as bondholders if their vote will enable the euro to survive. Secondly, the eurozone CAC will, I understand, contain aggregation provisions. I have not seen a draft of the proposed provision; but I would imagine that it will allow the issuer to put the same resolution to a number of issues, chosen by it, at the same time, with the resulting vote being based on a required majority by value across all the aggregated issues. It may also require that each issue pass a certain, lower, threshold in its voting for it to be bound. The significant point, in the context of this discussion, is that the decision of who is in, and who is out, of the aggregation is left entirely to the issuer. This would enable it, at least in theory, to pick issues that it believes will be likely to result in a vote going in



Bondholder Resolutions in the Courtroom 

 81

its favour, thus exposing it to the possible accusation of bad faith and resulting litigation. I must say, before giving some hypothetical examples, that they are exaggerated for the purposes of illustration; issuers and others involved in any restructuring debate will inevitably avoid unfairness and bad faith, if not for commercial reasons then because that is simply not how we do things in the EU. And I should reemphasise that these examples assume that the relevant legal principles arise from English law – which is unlikely to be the case given that the UK is not within the eurozone and there must be considerable doubt whether eurozone sovereigns will choose issue bonds governed by English law. With these caveats in place, though, let me offer the following examples of how the exercise of the eurozone CAC might conceivably be attacked by a disenchanted minority bondholder. The first example assumes that the issuer believes that it will not achieve the requisite majority. In order to swing the vote of one of the major holders, it privately promises that holder the position of adviser in the proposed privatisation of its publicly owned telecommunications company. I would think that that would, at the least, give minority bondholders scope to take the question to court on the grounds that the vote was obtained through bad faith (the promise amounting, perhaps, to bribery). Another example might be where the issuer, again uncertain of being able to obtain the requisite majority, issues new bonds to someone who will vote in favour (e.g. to save the euro). The issue is then brought into the aggregation arrangements. Again, the claim would be that this is bad faith. Finally, the issuer might, to extend maturities to 30 years across the board, use the aggregation provision to ensure that all its issues with a short time to run to maturity will be outvoted by holders of issues that have much longer to run. So, for example, a EUR 100 million issue with one year to run could be aggregated with 8 EUR 100 million issues with 10 years to run. The proposal – to extend maturities to 12 years across all issues – would be far more onerous to the first issue than to the other 8, who may well, as a result, vote in favour. Whether this would be open to the accusation of bad faith is perhaps more debatable than the other examples. But the risk is there, particularly if the only reason for structuring the aggregation in this way is to turn, by force, the short term issue into a longer term maturity.

82 

 Lachlan Burn

Some conclusions To conclude, I think that, despite the much vaunted certainty of English contract law, provisions such as the eurozone CAC could, depending on the circumstances, be attacked when exercised. Everything depends on the circumstances of each case (which makes it even more difficult than usual to predict the future); and the circumstances would have to be pretty extreme, or obviously “wrong”, for any attack to succeed in court. But I suspect that what will matter, in the context of a restructuring where the CAC is used in earnest, is not whether any claim would eventually be dismissed by a court. To find out what the court thinks, the parties to the restructuring will have to wait months or even years, and the restructuring will not wait for the puff of white smoke to appear above the court. A restructuring will not just be about a particular bond, or series of bonds. Issuers, even sovereigns, have other debts and liabilities that will no doubt be restructured at the same time, through the aegis of the Paris Club and the like. Each limb of the restructuring will, most likely, be interdependent or conditional on every other limb working. Even if it is not, the financial markets will be unlikely to start lending to the troubled sovereign again unless and until they are certain that the new footing has been firmly established. So, the objective should be not so much to ensure that the issuer has a viable defence, should proceedings ever be commenced, but to minimise the risk that proceedings would ever be brought in the first place. The way to do this is to ensure that the CAC is drafted in a way that is not overly one-sided – for example, so that the voting majorities are not too low and the arrangements for voting enable everyone to participate. I would also think that it would be very sensible to develop some kind of code that issuers would be expected to follow, so that the market knows in advance that the objective is to achieve a balanced and fair restructuring. So, for example, the code might indicate that all bondholders in all issues will be treated proportionately, wherever possible; and that aggregation will only be used to bring together issues that have similar commercial terms. It will also be important to ensure that, when the CAC is introduced into Eurozone sovereign bonds, there is no initial uncertainty as to the consequences of such introduction. The market is already discussing whether the risks associated with the CAC should be disclosed to investors and what those risks are – particularly in the light of recent experience in the context of the Greek restructuring. It will be important to manage the process, to avoid the possibility that such disclosure will get out of hand. Investors have been somewhat shaken by some of the recent events associated with the Greek restructuring – not least the retroactive insertion of a CAC



Bondholder Resolutions in the Courtroom 

 83

into existing bonds. Retroactive legislation of this sort runs counter to one of the most fundamental cornerstones of any civilized, free society – namely the rule of law. Investors acquired bonds in the belief that whatever contractual rights were embodied in them would be upheld by the courts and could only be altered with their agreement. So, when the debtor altered those rights, unilaterally and to its own advantage, by inserting a collective action clause, it is hardly surprising that the confidence of investors was badly shaken. And the fact that the perpetrator of this act was a European sovereign, and part of the European Union, who might be expected above all others to understand and uphold the principal of the rule of law, made the shock all the more palpable. It is easy to protest, but very much harder to make investors believe, that this was an isolated incident, and would not be repeated. Dogs that bite once tend to do so again. It will take time – probably a lot of it – to rebuild confidence. Again, part of the solution may lie in the adoption by Eurozone sovereigns of a code of conduct. The Institute of International Finance has, for some time now, been working on such a code – their “Principles for Stable Capital Flows and Fair Debt Restructuring” (September 2011). This is what they say about contractual rights. “Sanctity of contracts. Subject to their voluntary amendment, contractual rights must remain fully enforceable to ensure the integrity of the negotiating and restructuring process.” Precisely so. And precisely what did not happen in the Greek case. The voluntary adoption by Eurozone member states of such a code would, I suggest, help to resolve uncertainties created in the minds of many investors by recent events and thereby increase the chances of the market reaction to the introduction of the eurozone CAC being relatively neutral, because the scope for imaginations to run riot will have been restricted.

Boris Kasolowsky and Smaranda Miron*

Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? Sovereign debt restructuring exposes States to litigation risks. Such risks can be limited by providing for Collective Action Clauses (CACs)1 in sovereign bonds. This paper specifically addresses possible bondholder claims on the basis of bilateral investment treaties (BITs) and seeks to show how the inclusion of CACs in sovereign bonds can reduce the risks of such bondholder litigation directed against the States. Bearing in mind that the audience is made up mainly of capital markets and regulatory lawyers, the first part briefly summarises the protection available to investors, and particularly sovereign bondholders, under BITs. The second part provides an illustrative example using the evolution of the hypothetical State of Ruritania between 2013 and 2020. To describe Ruritania’s litigation risks, we will assume that Ruritania issues new bonds without CACs in 2013 and with CACs in 2014, and that the State is subsequently forced to restructure both editions of bonds in the year 2020.

I. What are BITs? When can bondholders rely on BITs? BITs are agreements between two States and thus instruments of public international law. Pursuant to these agreements, States undertake to promote and protect investments made by investors from one State in the territory of the other

* The authors are grateful to Margaret Artz, J.D. Candidate at Vanderbilt University Law School, for her thorough review of earlier versions of this article and helpful suggestions. All the views expressed in this article, and any remaining errors, are the authors’ own. 1 The European Stability Mechanism, ECB Monthly Bulletin, July 2011, pp 80–81 (“CACs are contractual provisions inserted into the terms and conditions governing bonds. […] [They] are designed to ensure orderly sovereign debt restructuring. They provide […] for a supermajority of bondholders (66⅔% or 75%) and the debtor to restructure outstanding bonds (e.g. modify key payment terms, or convert or exchange bonds). […] CACs imply that any restructuring modifications accepted by the specified majority of bondholders are conclusive and binding on all of the debt securities of a particular bond series—whether or not they have given their consent—which facilitates successful debt restructuring by overcoming the problem of ‘hold-out’ creditors.”).

86 

 Boris Kasolowsky and Smaranda Miron

State.2 While each BIT has to be interpreted individually,3 the approximately 2000 BITs4 that States have concluded over the last fifty years generally follow the same broad pattern and usually afford a similar level of protection for investors.5 Despite the inherent differences between various BITs, an arbitration tribunal’s interpretation of one BIT often influences later tribunals’ interpretation of the same concepts contained in different BITs, and is therefore helpful in understanding these treaties generally.6

2 See e.g. the Treaty between the Federal Republic of Germany and the Kingdom of Bahrain concerning the encouragement and reciprocal protection of investments of 5 February 2007, BGBl. 2008 II, 494 (The preamble reads: “The Federal Republic of Germany and the Kingdom of Bahrain, desiring to intensify economic co-operation between both States, intending to create favourable conditions for investments by nationals and companies of either State in the territory of the other State, […] have agreed as follows: […].”). 3 Vienna Convention on the Law of Treaties of 23 May 1969, entered into force on 27 January 1980. For a general view on the application of the VCLT as a source of customary international law, see Mark E. Villiger, Commentary on the 1969 Vienna Convention on the Law of Treaties, Martinus Nijhoff Publishers, Leiden, 2009. 4 The United Nations Conference on Trade and Development (UNCTAD) compiled a collection of around 1900 BITs, available at http://www.unctadxi.org/templates/DocSearch_779.aspx. However—to the authors’ knowledge—no exhaustive list of BITs has ever been successfully compiled. 5 See the United Kingdom Model Agreement for the Promotion and Protection of Investments of 2005, with 2006 amendments. 6 Saipem SpA v People’s Republic of Bangladesh, ICSID Case No. Arb/05/07, Decision on Jurisdiction and Recommendation on Provisional Measures, 21 March 2007, para 67 (“The Tribunal considers that it is not bound by previous decisions. At the same time, it is of the opinion that it must pay due consideration to earlier decisions of international tribunals. It believes that, subject to compelling contrary grounds, it has a duty to adopt solutions established in a series of consistent cases. It also believes that, subject to the specifics of a given treaty and of the circumstances of the actual case, it has a duty to contribute to the harmonious development of investment law and thereby to meet the legitimate expectations of the community of States and investors towards certainty of the rule of law.”). Cf., Principles of International Investment Law, Rudolf Dolzer and Christoph Schreuer, Oxford University Press, 2008, pp 35–36 (“Reliance on past decisions is a typical feature of any orderly decision process. […] In investment arbitration each tribunal is constituted ad hoc for the particular case. Therefore, it is more difficult to develop a consistent case law than in an international court such as the ICJ or ECHR. Yet, tribunals do rely on previous decisions of other tribunals whenever they can. Discussion of previous cases and of the interpretations adopted in them is a regular feature in almost every decision. At the same time, it is also well-established that tribunals in investment arbitration are not bound by previous decisions of other tribunals.”).

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 87

1. Which investors are protected? Which investments are protected? In order for investors—and therefore, as appropriate, bondholders—to take advantage of the BIT protections, two conditions must be fulfilled: both the investor and his investment in the host State need to qualify under the relevant BIT. Protected investors: Investors are typically protected under BITs if they are nationals of one of the contracting States or legal entities incorporated there. Arbitration tribunals interpreting the nationality requirement of legal entities have confirmed that “investor” is a reasonably broad category. To accept a company as an investor under a BIT, tribunals have readily disregarded the fact that the claimant investor is a special purpose vehicle with no meaningful activities at the place of incorporation.7 In other situations, tribunals permitted foreign claimant investors to take advantage of a BIT despite the fact that the investor entity was managed and beneficially owned by nationals of the host State.8 As a result, investors may structure their investments through legal entities incorpo-

7 Saluka Investments BV v The Czech Republic, Partial Award of 17 March 2006, paras 240–241 (“The Tribunal has some sympathy for the argument that a company which has not real connection with a State party to a BIT, and which is in reality a mere shell company controlled by another company which is not constituted under the laws of that State, should not be entitled to invoke the provisions of that treaty. […] However that may be, the predominant factor which must guide the Tribunal’s exercise of its functions is the terms in which the parties to the Treaty now in question have agreed to establish the Tribunal’s jurisdiction. In the present context, that means the terms in which they have agreed upon who is an investor who may become a claimant entitled to invoke the Treaty’s arbitration procedures. The parties had complete freedom of choice in this matter, and they chose to limit entitled “investors” to those satisfying the definition set out in Article 1 of the Treaty. The Tribunal cannot in effect impose upon the parties a definition of ‘investor’ other than that which they themselves agreed. That agreed definition required only that the claimant-investor should be constituted under the laws of (in the present case) The Netherlands, and it is not open to the Tribunal to add other requirements which the parties could themselves have added but which they omitted to add.”). Cf. OECD Model Tax Convention on Income and on Capital of July 2010, Article 7 (“In order for a legal entity to be allowed to take advantage of a double taxation treaty, it must not only be established in the relevant jurisdiction, but it must also be able to demonstrate substantive business activities in the respective jurisdiction.”). 8 Tokios Tokelés v Ukraine, ICSID Case No.ARB/02/18, Decision on Jurisdiction and Dissent of 29 April 2004 (Ukrainian businessmen had incorporated a Lithuanian company for the purpose of conducting their business. Despite the fact that the Lithuanian company’s business was entirely in the Ukraine and owned by Ukrainians, the majority of the Tribunal allowed the Lithuanian entity to sue the Ukraine on the basis of the Lithuanian-Ukrainian BIT.).

88 

 Boris Kasolowsky and Smaranda Miron

rated in jurisdictions that can take advantage of investment treaty protection or, as the case may be, better protection.9 Protected investments: Most BITs include a broad definition of protected investments. BITs commonly state that “every kind of asset” is protected and provide a non-exhaustive list of protected investments that typically includes shares, loans, other financial participations and claims to money.10 Arbitration tribunals have accordingly confirmed that promissory notes issued by Venezuela and acquired by the claimant investors on the secondary market11 and warranty bonds issued by Bangladesh12 qualify as investments for purposes of the applicable BITs. Even more relevant in the present context is the Abaclat and others v Argentina case.13 This ongoing arbitration concerns some 60,000 Italian bondholders suing Argentina in connection with the State’s default on its sovereign debt in December 2001. In its defence, Argentina argued that the bonds did not qualify as “investments” under the BIT, and also objected to the admissibility of the mass claim from the 60,000 bondholders under the BIT. The majority of the Tribunal, however, confirmed that Argentina’s sovereign bonds qualified as investments under the Argentina-Italy BIT.14 The majority also

9 OECD International Investment Perspective, 2006 (“Corporations are reported to begin structuring their transactions in such a way as to be able to benefit from the protection of different BITs.”). 10 The Germany-Bahrain BIT, for example, defines “investment” as “every kind of asset […] in particular: (a) movable and immovable property as well as any other rights in rem, such as mortgages, liens and pledges; (b) shares of companies and other kinds of interest in companies; (c) claims to money which has been used to create an economic value or claims to any performance having an economic value; (d) intellectual property rights, in particular copyrights, patents, utility-model patents, registered designs, trade-marks, trade-names, trade and business secrets, technical processes, know-how, and good will; (e) any right conferred by law or under public contract or any licenses, permits or concessions issued according to law […].” 11 Fedax NV v Venezuela, ICSID Case No. Arb/96/3, Award on Jurisdiction, 11 July 1997. 12 Saipem SpA v People’s Republic of Bangladesh, fn 6 above. 13 Abaclat and others v The Argentine Republic, ICSID Case No. Arb/07/5, Decision on Jurisdiction and Admissibility, 4 July 2011. 14 Ibid., para 387. (The Tribunal found that “[…] the present dispute arises out of an investment pursuant to Article 1 BIT and (if needed be) Article 25(1) ICSID Convention. In particular: (i) According to one view, the ‘double-barreled’ test developed with regard to the concept of ‘investment’ does not mean that the definition of investment provided by the two States in a BIT has to fit into the definition deriving from the spirit of the ICSID Convention. Rather, arguably, it is the investment at stake that has to fit into both of these concepts, knowing that each of them focuses on another aspect of the investment; (ii) […] the relevant bonds and Claimants’ security entitlements therein are both to be considered ‘investments’ pursuant to Article 1(1) lit. (c) BIT; (iii) […] Claimants’ purchase of security entitlements in Argentinean bonds constitutes a contribution

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 89

allowed the mass claim to be heard in a single proceeding (rather than in 60,000 separate proceedings).15 In practice, therefore, it is likely that foreign bondholders will qualify as investors, and that their investment in sovereign bonds will qualify as an investment, provided that an applicable BIT is in place.

2. What substantive protection do BITs provide? Depending on the actual measures taken by the State (or, as the case may be, on the State’s failure to act), some or all the rights of the investors under the BIT might be infringed. Generally, BITs provide for numerous investor protection standards, but the most relevant for bondholders are likely to be: –– the State’s obligation to pay prompt and adequate compensation in the event of a lawful expropriation,16 –– the fair and equitable treatment guarantee,17

which qualifies as ‘investment’ under Article 25 ICSID Convention; (iv) [t]he relevant bonds and Claimants’ security entitlements are both to be considered made ‘in compliance with the laws and regulations of [Argentina] pursuant to Article 1(1) BIT; (v) [t]he bonds and Claimants’ security entitlements therein are both to be considered ‘made in the territory of Argentina.’”). Cf., Dissenting Opinion dated 28 October 2011, paras 34–119 (In his dissent Professor Georges Abi-Saab opined that such a broad definition should not be adopted because “the alleged investment is totally free-standing and unhinged, without any anchorage, however remote, into an underlying economic project, enterprise or activity in the territory of the host State. … As both the ICSID Convention and the BIT require that the investment be made in the territory of the host State for the investment to be covered by the treaty and fall within the jurisdictional ambit of ICSID, and as this territorial link is lacking in the present case, I conclude that in the absence of a ‘protected investment,’ the case has to be dismissed, as falling outside the jurisdiction ratione materiae of the Tribunal.”). 15 Ibid, paras 534–547. Cf., Dissenting Opinion dated 28 October 2011, paras 176–193 (With regard to mass claims, Professor Abi-Saab stated that a mere acceptance to arbitrate does not cover collective claims actions and that special or secondary consent is needed for such actions.). 16 Germany-Bahrain BIT, article 4(2) (“Investments by nationals or companies by either Contracting State shall not be expropriated, nationalized or subjected to other measure the effects of which would be tantamount to expropriation or nationalization in the territory of the other Contracting State except for the public benefit and against compensation. […]”). 17 Ibid., article 2(1) (“Each Contracting State shall in its territory promote as far as possible investments by nationals or companies by the other Contracting State and admit such investments in accordance with its legislation. It shall in any case accord such investments fair and equitable treatment.”).

90 

 Boris Kasolowsky and Smaranda Miron

–– the State’s guarantee to treat investors from one State no less favourable than nationals and investors from other States (most favoured nation treatment),18 and –– the State’s obligation to satisfy its contractual obligations and undertakings (umbrella clause).19 For an audience of lawyers who work in the capital markets and regulatory arena it may be useful to say a few more words about each of these standards. No expropriation: Prompt and adequate compensation in the event of a lawful expropriation is the classic investor right. Expropriation may occur by a direct and deliberate formal taking, or indirectly, by measures resulting in a substantive deprivation of the use and value of the investment. Plain and brutal expropriation has fortunately become rarer, so arbitration tribunals have shifted their focus to a different type of expropriation with similar effects: indirect expropriation.20 Where an investor has been substantially deprived of the use and benefit of his investment by a measure (or a series of measures) of the State, he is entitled to claim compensation for expropriation even if he continues to retain nominal ownership over the investment.21 However, in order for an arbitration tribunal to hold that the State did indirectly expropriate the investment at issue, the level of interference with the investors’ rights and its

18 See e.g. the Treaty between the Federal Republic of Germany and the Sultanate of Oman concerning the encouragement and the reciprocal protection of investments of 25 June 1979, BGBl. 1985 II, 345, article 3, para (1) (“Neither Contracting State shall subject investments in its territory owned or controlled by investors of the other Contracting State to treatment less favourable than it accords to investments of its own investors or to investments of investors of any third State.”). 19 Germany-Bahrain BIT, article 8(2) (“Each Contracting State shall observe any other obligation it has assumed with regard to investments in its territory by nationals or companies of the other Contracting State.”). 20 Indirect expropriation is usually described in most BITs as any “measure the effects of which would be tantamount to expropriation or nationalization.” See fn 16 above. Cf., Rudolf Dolzer and Margrete Stevens, Bilateral Investment Treaties, The Hague, Nijhoff, 1995, p 99 (“Though there have been various attempts at clarifying and differentiating between different types of indirect expropriations, it appears that the term is frequently used interchangeably with expressions such as de facto, disguised, constructive, regulatory, consequential, or creeping expropriation.”). 21 Metalclad Corporation v The United Mexican States, ICSID Case No.ARB(AF)/97/1, Award of 30 August 2000, at para 103 (“[…] [E]xpropriation […] includes not only open, deliberate and acknowledged takings of property, such as outright seizure or formal or obligatory transfer of title in favour of the host State, but also covert or incidental interference with the use of property which has the effect of depriving the owner, in whole or in significant part, of the use or reasonably-to-be-expected economic benefit of property even if not necessarily to the obvious benefit of the host State.”).



Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 91

economic impact must be notable.22 An arbitration tribunal has found that even a very significant loss in the value of shares did not constitute a sufficient loss of enjoyment and use of the asset to constitute an expropriation.23 This suggests that in practice, it may be challenging for bondholders to argue successfully that sovereign debt restructuring amounts to indirect expropriation. Fair and equitable treatment: Most BITs include the fair and equitable treatment (FET) standard and investors rely on it in almost all investor-State arbitration proceedings. In fact, it is the basis on which investors are most likely to succeed in practice.24 Under the FET standard protection, investors are entitled to expect States to conduct their affairs transparently and to insist that States maintain a reasonably stable regulatory and legal environment.25 Investors have successfully argued a breach of the FET standard in circumstances where the conduct of the host State prompted legitimate expectations—typically by making a commitment to act or to refrain from acting in a particular manner—and subsequently disappointed them.26 Investors have also prevailed using the FET standard where an

22 LG&E Energy Corp. and others v Argentine Republic, ICSID Case No.ARB/02/1, Decision on Liability of 3 October 2006, para 190 (“In evaluating the degree of the measure’s interference with the investor’s right of ownership, one must analyze the measure’s economic impact—its interference with the investor’s reasonable expectations—and the measure’s duration.”). 23 CMS Gas Transmission Company v The Argentine Republic, ICSID Case No.ARB/01/8, Award of 12 May 2005, paras 263–264 (“[…] The Government of Argentina has convincingly argued that the list of issues to be taken into account for reaching substantial deprivation […] is not present in the instant dispute. In fact, the Respondent has explained, the investor is in control of the investment; the Government does not manage the day-to-day operations of the company; and the investor has full ownership and control of the investment. The Tribunal is persuaded that this is indeed the case in this dispute and holds therefore that the Government of Argentina has not breached the standard of protection laid down in Article IV(1) of the Treaty.”). 24 Thomas W Wälde and Borzu Sabahi, Compensation, Damages and Valuation in: Peter Muchlinski, Federico Ortino and Christoph Schreuer (eds.), The Oxford Handbook of International Investment Law, Oxford University Press, 2008, pp 1086–1089 (“Tribunals seem to prefer basing their awards on a breach of fair and equitable treatment rather than on ‘indirect expropriation.’”). 25 SPP v Egypt, Award and Dissenting Opinion of 20 May 1993, para 83 (“[…] Whether legal under Egyptian law or not, the acts in question were the acts of the Egyptian authorities, including the highest authority of the Government. These acts, which are now alleged to have been in violation of the Egyptian municipal system, created expectations protected by established principles of international law.”). 26 CME Czech Republic B.V. v The Czech Republic, Award of 13 September 2001, para 611 (“[…] The Media Council breached its obligations of fair and equitable treatment by evisceration of the arrangements in reliance upon which the foreign investor was induced to invest.”); Cf., Tecnicas Medioambientales Tecmed S.A. v United Mexican States, ICSID Case No.ARB (AF)/00/2, Award of 29 May 2003, para 167 (“[…] the Claimant was entitled to expect that the gov-

92 

 Boris Kasolowsky and Smaranda Miron

agency of the host State conducted government affairs without due process or the requisite judicial propriety.27 Thus, in the event that a government defaults on its bonds, or modifies the terms of the bonds without the consent of the bondholders, it would appear reasonably likely that this constitutes a breach of the FET standard. Most favoured nation treatment: The most favoured nation (MFN) treatment may be relevant in several situations. Most importantly, the MFN treatment guarantee allows investors to insist on being treated as investors from other States are treated. On the basis of an MFN provision, tribunals have allowed investors to import substantive treaty protection available in a BIT between the host State and a State unrelated to the investor into his BIT with the host State. Let us consider the hypothetical case of a German company that is denied due process in the courts of Malaysia. The BIT between Germany and Malaysia does not provide for FET, but includes an MFN clause. In any proceedings against Malaysia, the German investor could rely on the MFN clause in the German-Malaysian BIT to argue that an FET clause should be imported from the Dutch-Malaysian BIT into the BIT between Germany and Malaysia.28 MFN treatment is therefore always relevant for investors suing on the basis of a BIT that does not include all the requisite treaty standards. The MFN clause’s relevance in connection with the procedural protections will be addressed in more detail below. Umbrella clause: The last substantive protection addressed here is the socalled “umbrella clause.” Under the terms of most umbrella clauses, States commit to fulfil their contractual obligations and undertakings.29 The exact scope

ernment’s actions would be free from any ambiguity that might affect the early assessment made by the foreign investor of its real legal situation or the situation affecting its investment and the actions the investor should take accordingly.”). 27 Waste Management, Inc. v United Mexican States, ICSID Case No.ARB(AF)/00/3, Final Award of 30 April 2004, para 98 (“[…] the minimum standard of treatment of fair and equitable treatment is infringed by conduct attributable to the State and harmful to the claimant if the conduct […] involves a lack of due process leading to an outcome which offends judicial propriety—as might be the case with a manifest failure of natural justice in judicial proceedings or a complete lack of transparency and candour in an administrative process […].”). 28 Bayindir Insaat Turizm Ticaret Ve Sanayi AŞ v Pakistan, ICSID Case No.ARB/03/29, Decision on Jurisdiction of 14 November 2005, paras 208–235 (The Turkish claimant successfully imported the FET standard using an MFN clause provided in the BIT. The applicable BIT between Turkey and Pakistan did not expressly provide for an FET clause, but the claimant asserted that it was entitled to such treatment because it did include an MFN clause. The Tribunal accepted the claimant’s argument, since other BITs concluded by Pakistan with France, the Netherlands, China, Australia and Switzerland all contained an explicit FET clause.). 29 See fn 19 above.

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 93

of umbrella clauses is still subject to considerable debate.30 On one side, tribunals have relied on umbrella clauses to elevate each and every contractual obligation entered into by a State into a parallel treaty right.31 Yet, other arbitration tribunals have interpreted the umbrella clauses more narrowly. They have found that only contractual obligations and undertakings entered into by States in the exercise of their sovereign powers indicate the States’ intent to create an independent right under the relevant BIT.32 Whether or not the issuance of bonds is an exercise of sovereign power has—to the authors’ knowledge—not been settled thus far. In yet another case, the tribunal required States’ intention that contract claims be arbitrable as treaty claims under the BIT be evident from the wording of the BIT.33 Assuming the relevant BIT includes an investor-friendly umbrella clause, bondholders are likely to rely heavily on such provision. Compensation: If bondholders succeed in their claim that the host State has breached its duties under the relevant BIT, an arbitration tribunal may award them compensation. It is a basic principle of international law that the aggrieved

30 Nigel Blackaby and Constantine Partasides with Alan Redfern and Martin Hunter, Redfern and Hunter on International Arbitration, 5th ed., Oxford University Press, 2009, pp 506–508. 31 Eureko BV v Republic of Poland, Partial Award of 19 August 2005, paras 246 (“The plain meaning—the ‘ordinary meaning’—of a provision that a State ‘shall observe any obligation it may have entered into’ with regard to certain foreign investment is not obscure. The phrase ‘shall observe’ is imperative and categorical. ‘Any’ obligations is capacious; it means not only obligations of a certain type, but ‘any’—that is to say, all—obligations entered into with regard to investments of investors of the other Contracting Party.”) 32 Pan American Energy LLC, and BP Argentina Exploration Company v The Argentine Republic, ICSID Case No.ARB/04/8, Decision on Preliminary Objections of 27 July 2006, paras 91–117 (“[T]he Tribunal […] confirms that it has jurisdiction over treaty claims and cannot entertain purely contractual claims, which do not amount to a violation of the standards of protection of the BIT. […] The answer to the question […] whether the existence of a so-called ‘umbrella clause’ changes the Tribunal’s intermediary conclusion to the effect that it has no jurisdiction over purely contractual claims, and that it can only entertain treaty claims, is purely in the negative. Indeed, the Tribunal has jurisdiction only over treaty claims, the latter including, pursuant to the wording of Article VII(1), the claims based on the violation of an investment agreement entered into by the foreign investor with the State as a sovereign.”). 33 SGS Société Générale de Surveillance v Islamic Republic of Pakistan, ICSID Case No.ARB/01/13, Decision of the Tribunal on Objections to Jurisdiction of 6 August 2003, para 171 (“We believe, for the foregoing considerations, that Article 11 of the BIT would have to be considerably more specifically worded before it can reasonably be read in the extraordinarily expansive manner submitted by the Claimant. The appropriate interpretive approach is the prudential one summed up in the literature as in dubio pars mitior est sequenda, or more tersely, in dubio mitius.”).

94 

 Boris Kasolowsky and Smaranda Miron

investor should, to the extent possible, be restored to his ex ante position.34 Therefore, the tribunal may order the host State to pay the full nominal value of the bond plus interest, as well as all or part of the arbitration costs.

3. Procedural protection In addition to the substantive standards, investors benefit from certain procedural rights. The most important of these is the power of investors to refer their disputes with host States to international arbitration tribunals. The majority of BITs allow investors to initiate arbitration proceedings under the auspices of the World Bank’s International Center for the Settlement of Investment Disputes (ICSID)35 or ad hoc arbitration under the Arbitration Rules of the United Nations Committee on International Trade Law (UNCITRAL).36 ICSID arbitral awards benefit from a very favourable enforcement regime,37 while UNCITRAL and other ad hoc awards are generally enforceable under the New York Convention of 1958, whose provisions are almost as favourable.38

34 Case Concerning The Factory at Chrozów (Claim for Indemnity) (Merits), Germany v Poland, 1928 PCIJ (Series A), No.17, p 40 (“The essential principle contained in the actual notion of an illegal act—a principle which seems to be established by international practice and in particular by the decisions of arbitral tribunals—is that reparation must, as far as possible, wipe out all the consequences of the illegal act and re-establish the situation which would, in all probability, have existed if that act had not been committed. Restitution in kind, or, if this is not possible, payment of a sum corresponding to the value which a restitution in kind would bear; the award, if needed be, of damages for loss sustained which would not be covered by restitution in kind or payment in place of it—such are the principles which should serve to determine the amount of compensation due for an act contrary to international law.”). 35 ICSID Rules of Procedure for Arbitration Proceedings of April 2006. 36 UNCITRAL Arbitration Rules, as revised in 2010. 37 ICSID Convention, Art 53(1) (“The award shall be binding on the parties and shall not be subject to any appeal or to any other remedy except those provided for in this Convention. Each party shall abide by and comply with the terms of the award except to the extent that enforcement shall have been stayed pursuant to the relevant provisions of this Convention.”). 38 The United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards, Article III (“Each Contracting State shall recognize arbitral awards as binding and enforce them in accordance with the rules of procedure of the territory where the award is relied upon […].”).

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 95

However, certain older BITs (particularly those concluded before 1965)39 do not provide investors with the right to file a claim independently. Under such BITs, only States have standing to commence proceedings against one another.40 If an investor’s home State is party to an older BIT, the investor may attempt to improve his position by structuring his investment through an entity incorporated in a jurisdiction that is a party to a more advantageous BIT with the State in which he invested. This is, however, only possible when a dispute has not yet arisen.41 If a dispute has already emerged, the investor must rely on his home State to take up the claim against the host State of investment.42 However, since States are generally unwilling to disrupt their international relations for the purely economic interests of their nationals, such State versus State disputes are rarely, if ever, commenced. For this reason, it seems even less likely that home States would help their nationals against host States of investment with respect to such a politically sensitive issue as the host State’s default on sovereign bonds. Alternatively (and continuing under the assumption that the BIT does not provide for a direct investor right to initiate proceedings), an investor may attempt to invoke an MFN clause, if such a clause exists in the applicable BIT. He might argue that because investors from other countries are entitled to sue on the basis of their BIT, the MFN clause should bestow the same right on him. Thus investors have argued that if investors from other States are entitled to sue on the basis of their BIT, they should enjoy the same right under the MFN clause in their States’ BITs. Thus far, however, tribunals have responded unfavourably to investors’ efforts to establish a direct right to sue in reliance on the MFN clause.43 Arbitra-

39 The ICSID Convention was submitted for signature to the member states of the World Bank on 18 March 1965 and entered into force on 14 October 1966. 40 Agreement between the Federal Republic of Germany and the Kingdom of Greece concerning the promotion and reciprocal protection of investments of 27 March 1961, BGBl. 1963 II, 216. 41 Mobil Corporation and others v Bolivarian Republic of Venezuela, ICSID Case No.ARB/07/27, Decision on Jurisdiction of 10 June 2010, paras 167–207 (The Arbitration Tribunal held that the aim of restructuring of investments in order to gain access to ICSID arbitration through a BIT was a perfectly legitimate goal as far as it concerned future disputes. With respect to pre-existing disputes, the tribunal considered that to restructure investments only to gain jurisdiction under a BIT for such disputes constitutes an abusive manipulation of the system of international investment protection.). 42 This practice is known as diplomatic protection or espousal of rights. 43 Vladimir Berschader and Moïse Berschader v The Russian Federation, SCC Case No 080/2004, Award of 21 April 2006, paras 159–208 (The arbitration tribunal ruled that the MFN provision did not clearly and unambiguously provide for incorporation by reference of arbitration clauses in other BITs and declined jurisdiction under the BIT.)

96 

 Boris Kasolowsky and Smaranda Miron

tion tribunals’ decisions on this topic suggest that bondholders from States that do not benefit from a BIT which provides for a direct right to sue will not be able to enforce their treaty rights directly.

II. Ruritania between 2013 and 2020 Having briefly outlined the investment protection that might be available to bondholders, we turn our attention to the hypothetical State of Ruritania and the years of 2013 and 2020.

1. The Scenario In 2013, for the first time in the history of Ruritania, a President from the Ruritania Progress Party takes office. She has been elected on the basis of a programme promising significant investments in public infrastructure (publicly funded) and the concomitant creation of 100,000 new jobs. Shortly afterwards, the President’s Progress Party is also successful in parliamentary elections, and consequently holds the majority in the Ruritanian parliament. As promised during the election campaign, Ruritania proceeds to issue new bonds to fund the infrastructure projects and create employment. In the autumn of 2013, Ruritania issues bonds with a 4.5% yield that reach maturity in 2021. These bonds do not contain CACs. Early in 2014, Ruritania issues another series of bonds. The yield and the maturity date remain unchanged. However, the 2014 edition contains CACs providing that bondholders holding 80% of the 2014 bonds may collectively amend the contractual terms of these bonds. Ruritania’s political and economic situation changes dramatically in the following years. Like many other neighbouring States, Ruritania is hit by the “tripledip” global economic crisis starting in the year 2015. Due to massive delays and cost overruns, none of the major infrastructure projects proceeds at the previously anticipated pace. Capital flows into Ruritania decrease sharply from 2016 on wards. Starting in 2017, capital outflows increase substantially. By 2018, Ruritania’s credit rating is drastically downgraded, and the Progress Party loses its

Cf., Berschader separate opinion of Todd Weiler, paras 17–20 (“[An MFN clause] extends to procedural aspects of the dispute, including entitlement to pursue arbitration. […] [T]here is simply no reason to suppose that – absent some specific treaty language – any given MFN provision should be more or less narrowly defined.”).

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 97

majority in the parliament. In 2019, rumours of a possible restructuring, and even a default, start to circulate. The President is not re-elected and the Conservative Party takes office at the end of 2019. At the beginning of the year 2020, Ruritania decides to restructure its sovereign debt—specifically the bonds issued in the years 2013 and 2014. Ruritania invites bondholders to agree to a change in the terms of their bonds. Under the new terms, the bonds have a face value of 30% of the original nominal value, 2030 as the maturity date, and a yield tied to the development of Ruritania’s gross domestic product. Due to the effective lobbying of the major groups of bondholders and Ruritania itself, more than 80% of the holders of both the 2013 and 2014 bonds accept the offer voluntarily. The 80%+ majority holding the 2014 bonds votes to change the terms of the bonds in accordance with the CAC. The situation is more difficult as far as the 2013 bondholders are concerned. While 80% of the 2013 bondholders support a modification of the terms, a vociferous minority of powerful hedge funds and private investors is strongly opposed to any changes. In order to effect a change against the will of the minority, the Ruritanian parliament passes emergency legislation that retroactively introduces CACs into all bonds issued by Ruritania since 2005. Once the CACs have been introduced into the 2013 bonds by virtue of emergency legislation, the majority of the 2013 bondholders approve the changed terms in accordance with the CACs.

2. What remedies do the 2013 and 2014 bondholders have against Ruritania? The bondholders who voluntary took up the new bonds waived their rights under the old terms of the bonds. This is of little consequence, as they are unlikely to seek any remedies against Ruritania. However, the dissatisfied private investors and hedge funds that opposed the change in the terms of the bonds may well consider available remedies under the applicable BITs. As far as the 2014 generation of bondholders is concerned, it would appear difficult for any of them to succeed in proceedings against Ruritania. The minority of the 2014 bondholders always knew that the bonds provided for CACs. They never had any legitimate expectations that this could be otherwise. Therefore, the fact that the majority makes a decision to the detriment of the minority does not appear to be a per se breach of a BIT. However, the situation might be different with respect to the dissatisfied 2013 bondholders. Initially the 2013 generation bondholders could not be compelled to take up new bonds. Only the retroactive change in Ruritanian law made this

98 

 Boris Kasolowsky and Smaranda Miron

possible. While Ruritania might achieve its objective in compliance with domestic law,44 there are several grounds on which foreign investors might be able to assert investment treaty claims against Ruritania for having effected this retroactive change in law. Assuming the 2013 bondholders qualify as investors and the bonds as investments under the relevant BITs, aggrieved bondholders can argue that the retroactive introduction of CACs constitutes a breach of substantive BIT protections and claim compensation. While investors may generally rely on all available BIT standards, it is likely that the bondholders will focus on the FET standard and on the umbrella clause. As explained above, due to the characteristics of their investment, bondholders are likely to rely on the FET standard and on the umbrella clause. A breach of the investor’s legitimate expectations and a significant change in the legal stability, on which the investor relied in making his investment, constitute a breach of the FET standard. Aggrieved bondholders might argue that introduction of retroactive legislation shows a lack of good faith and transparency,45 as the legislative measures enacted by Ruritania are inconsistent with previous policies and unpredictable.46 Moreover, the dissatisfied 2013 bondholders may also contend that the enactment of retroactive legislation is a per se breach of the FET standard.47 In addition to potential FET claims, the legislative changes in the CACs legislation may amount to a breach of the BIT’s umbrella clause. The 2013 bonds

44 For present purposes we assume that there are no legal grounds to object to the legality of the Ruritanian law with retroactive effect. However, this fact alone cannot guarantee the legality of the measures under international law. Cf., the Vienna Convention on the Law of Treaties, fn 3 above, Article 27 (“A party may not invoke the provisions of its internal law as justification for its failure to perform a treaty. […]”). 45 Saluka v The Czech Republic, fn 7 above, para 307 (“A foreign investor protected by the Treaty [BIT] may in any case expect that the Czech Republic implements its policies bona fide by conduct that is, as far as it affects the investor’s investment, reasonably justifiable by public policies and that such conduct does not manifestly violate the requirements of consistency, transparency, even-handedness and non-discrimination.”). 46 LG&E v The Argentine Republic, fn 22 above, para 131 (“[…] the fair and equitable standard consists of the host State’s consistent and transparent behaviour, free of ambiguity that involves the obligation to grant and maintain a stable and predictable framework necessary to fulfil the justified expectations of the foreign investor.”). 47 Total S.A. v Argentine Republic, ICSID Case No.ARB/04/1, Decision on Liability of 27 December 2012, para 444 (“[…] the Tribunal concludes that the retroactive elimination of the Tierra del Fuego tax exemption effected by Argentina through the enactment of Resolution 776/06 is in breach of the fair and equitable treatment clause of the BIT.”).

 Can Collective Action Clauses in Sovereign Bonds Limit Litigation Risks for States? 

 99

provided for specific terms and did not include CACs. The only way the bond terms could be modified was with the consent of all bondholders. The retroactive introduction of CACs into the 2013 bonds, and the resulting change of the terms, arguably constitutes a breach by Ruritania of a commitment made to the 2013 bondholders in 2013. To overcome the restrictive interpretation of the umbrella clause, the bondholders might also assert that the issuance of a bond is not a mere commercial activity, but an act that Ruritania performed in its sovereign capacity. Apart from putting the claimant investors to proof of each and every requirement of their positive case, Ruritania might also argue that it had to adopt the emergency legislation out of necessity.48 To date, however, States have rarely succeeded with such an argument.49

48 Andrea K Bjorklund, Emergency Exceptions, in: Peter Muchlinski, Federico Ortino and Christoph Schreuer (eds.), The Oxford Handbook of International Investment Law, Oxford University Press, 2008, pp 520–522 (“States may raise as defences to their international obligations the doctrines of necessity and force majeure. Both have been recognised as part of customary international law by international tribunals, and both are included in the ILC’s Articles on State Responsibility as circumstances precluding wrongfulness.”). 49 See Bundesverfassungsgericht [2 BvM 1/03], Decision of 8 May 2007, para 27 (In the context of Argentina’s bond restructuring, the German Constitutional Court held that no rule of general international law can be ascertained entitling a State, vis-á-vis private individuals, to suspend the performance of due obligations for payment arising under private law by invoking necessity based on an inability to pay.). Cf., CMS v Argentina, fn 23 above, para 317 (“If strict and demanding conditions are not required or are loosely applied, any State could invoke necessity to elude its international obligations. This would certainly be contrary to the stability and predictability of law.”). Cf., LG&E and others v The Argentine Republic, fn 22 above, paras 215–266 (“[…]The Tribunal has determined that Argentina’s enactment of the Emergency Law was a necessary and legitimate measure on the part of the Argentine Government. Under the conditions the Government faced in December 2001, time was of the essence in crafting a response. Drafted in just six days, the Emergency Law took the swift, unilateral action against the economic crisis that was necessary at the time […] Based on the analysis of the state of necessity, the Tribunal concludes that, first, said state started on 1 December 2001 and ended on 26 April 2003; second, during that period Argentina is exempt of responsibility, and accordingly, the Claimants should bear the consequences of the measures taken by the host State; and finally, the Respondent should have restored the tariff regime on 27 April 2003, or should have compensated the Claimants, which did not occur. As a result, Argentina is liable as from that date to Claimants for damages.”).

100 

 Boris Kasolowsky and Smaranda Miron

III. Conclusion This paper has shown that based on Ruritania’s retroactive introduction of CACs into the 2013 bonds, the dissatisfied private investors and hedge funds holding 2013 bonds may have valid grounds on which to commence arbitration proceedings against Ruritania. By contrast, the dissatisfied 2014 bondholders cannot invoke the same legal arguments and remedies under the applicable BITs. By including CACs in the 2014 bonds, Ruritania avoided costly arbitration with investors holding bonds of this generation. Ruritania was thus well advised to include CACs in the 2014 bonds, and would have been well advised to do so already in 2013.

The Future of Sovereign Debt in Europe

Martin Wiesmann1

CACs and the Restructuring of Sovereign Debt – How Would Markets React? Introduction The world’s financial markets remain in their biggest crisis in decades. The crisis is not only deep and protracted but involves the two core geographies of the developed world: the US and Europe. It is profoundly shaking both the banking sector and the realm of sovereign credit and it is leading to major paradigm shifts in modern Finance. One of the causes of the first phase of the crisis in 2007/08 was that market participants and regulators treated any triple A paper, even highly leveraged paper such as CDO squared, as risk-free; relying too heavily on previous history. The second stage of the crisis erupted in part because market participants and regulators believed that Euro-denominated Government bonds were also risk-free be they issued by Greece, France or Germany. This view has been torn apart over the past few years as increasing government leverage and losses on government bonds have brought home to investors that credit risk exists in European government bonds. It will take a long time for our financial market infrastructure to fully draw and implement the right consequences from this experience, but it is important to understand the direction we are heading in. As much as the ongoing work around the need for resolution regimes and the debate about bail-in capital on the banking side, Collective Action Clauses (“CACs”) are a recognition that risk in the sovereign bond space needs to be reassessed. This goes first for recognising and dealing with the nature of risk; second, it relates to the question of who should bear these risks and how to ensure that other sovereigns and their taxpayers do not foot the bill for private investors’ mistakes; and finally, it relates to the policy response to the crisis in the Eurozone, which aims at reinstating the discipline of the market, and achieving governance changes that make Europe, the EU and the Eurozone, stronger both institutionally and structurally.

1 I would like to thank my colleagues Saul Doctor, Executive Director in J.P. Morgan’s credit research team in London, for his invaluable input based on his extensive research and analysis around restructurings in the realm of sovereign credit, and Julian Peise, Associate in J.P. Morgan’s Frankfurt advisory team, who supported this paper in particular with a deep dive into the history of academic research on the pricing impact of CACs.

104 

 Martin Wiesmann

These considerations explain why a concept that has reshaped Emerging Market credit has now entered the realm of sovereign risk in the developed world. In the Emerging Markets world, CACs have had a significant impact on the relationship between governments and investors in cases of a sovereign restructuring, much in the same way as international legal standards, under which most Emerging Market debt is now issued, have had. CACs are useful in the Emerging Market sovereign world to aid restructuring plans and recognise that Emerging Market sovereign debt bears credit risk. The introduction of CACs in the Eurozone may also help restructuring of sovereign debt, but more fundamentally it is a recognition that the world of “untouchable” Developed Market sovereign debt also bears credit risk. The Greek PSI restructuring shook governments, markets and other public institutions, including the ECB, into a new paradigm. From a capital market perspective, CACs, in the context of Eurozone credit, raise some fundamental questions: will the Eurozone become a hybrid credit, where investors have to recognise that a significant part of the region bears credit risk as a normality? Will the other part of that region start to take protective measures, driving forward the nationalisation of our European financial markets and thereby deepening the forces of political and financial disintegration? How will future Eurozone government debt, WITH credit risk, be financed if it is uncertain from today’s vantage point how to distinguish between risky and risk-free European sovereigns? Will the Eurozone manage to establish a new governance model on the basis of successful institutional and economic reform that achieves a new and healthy balance between the Eurozone’s core and its periphery, allowing the latter to re-establish sustained market access? Our analysis about the market reaction to CACs will not deal with their more philosophical role in that context. Much rather, it aims at providing an analytical basis to deal with the question, under what conditions CACs could have an impact on the pricing of government debt going forward. For that purpose, we look at: –– History and definitions of CACs –– Precedents of sovereign debt restructurings with CACs –– Impact of CACs on bond pricing –– Potential impact on the Eurozone sovereign debt market



CACs and the Restructuring of Sovereign Debt – How Would Markets React? 

 105

History and definitions of CACs

40

Total Issues

35

Source: IMF

Collective Action Clauses (“CACs”) refer to a wide range of bond clauses, establishing a framework under which the terms of the bond can be altered. They specify how bondholders are represented and outline majority-voting procedures, therefore allowing for a simplification and streamlining of debt restructurings. After being introduced in English corporate bonds, CACs were first used in sovereign bonds at the beginning of the 20th century. However capital markets did not adopt CACs on a large scale until the end of the 20th century. Only after prominent sovereign debt crises such as Mexico in 1995 or Argentina in 2001, were CACs increasingly used in sovereign bond documentation for bonds issued under international law by Emerging Market countries. This trend has been supported by the inclusion of CACs within New York law from 2003 onwards, a legal bond framework used by many Emerging Market countries. Figure 1 shows the increase in bond issues under New York law which include CACs. Today, almost all bonds issued by Emerging Market countries feature CACs (Figures 2). Issues Incl. CACs

30 25 20 15 10

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

1999

2000

1998

1997

1995

1996

0

1994

5

Figure 1: Use of CACs over time (Total number of NY law issues, number with CACs).

Excluding CACs 61 %

2001

Including CACs 39 %

Excluding CACs 3%

2006

Including CACs 97 %

Figure 2: Use of CACs over time (Use of CASs in bonds issued by Emerging Markets countries).

106 

 Martin Wiesmann

While there is no global standardized set of CACs, the following clauses are usually defined as being “core”2: –– Majority action clause: Changes of the bond terms which are agreed by a qualified majority are binding for all bondholders. –– Representation clause: Bondholders are allowed to appoint representatives who can act on their behalf. The clause effectively establishes a mechanism for coordinated interaction between the issuer and the holders of a bond. –– Enforcement clause: The clause prohibits the ability of a minority of bondholders to legally challenge other bondholders or the issuer in connection with a debt restructuring. –– (De-)Acceleration clause: Allows changes of due dates of payments and the bond maturity. While the general nature of standard CACs is usually consistent, specific provisions have evolved over time and can differ across bond jurisdictions. The most important provision for CACs, and a key in any debt restructuring, is the minimum

No. of Bonds

Minimum percentage votes to modify payment terms in New York bonds

75 %

85 %

100 %

a)

1995

2000

Minimum percentage votes to modify payment terms in English bonds No. of Bonds

19 %

2005 19 %

25 %

2010 38 %

50 %

b)

1995

2000

2005

Figure 3: Majority action clause within NY and UK bonds (Source: IMF) a) Majority action clause within NY bonds b) Majority action clause within UK bonds

2 Dixon/Wall 2000

2010

75 %



CACs and the Restructuring of Sovereign Debt – How Would Markets React? 

 107

percentage necessary to modify the payment terms in a bond. Figure 3 provides an overview of such minimum vote requirement for bonds governed by New York (“NY bonds”) and English (“UK bonds”) law: After historically requiring a 100% resolution, NY bonds today usually feature a 75% provision. In contrast, UK bonds used to require significantly lower thresholds to alter payment terms. Today most of UK bonds’ CACs ask for a 75% vote similar to NY bonds.

Precedents of sovereign debt restructurings with CACs While CACs have become a predominant feature of sovereign bonds of Emerging Market countries, the market experience of debt restructuring using such CACs is still limited. Along with the rather extreme recent debt restructuring of Greece in the Euro crisis, other relevant examples include Belize (2006) and Moldova (2002). One of the more well known restructurings with CACs is the Ukraine in 2002: The Russian Crisis in 1997 lead to the drying up of the lending market for Ukraine which itself had large refinancing needs in 1998. In February 2002 a comprehensive restructuring was offered for $2.5bn of Euro, USD and DM bonds. By accepting the restructuring, bondholders were also agreeing for an exchange agent to act as their proxy at a meeting which would use CACs to change the terms of the existing bonds. At the end of the process, the acceptance level for the restructuring was above 90%. Figure 4 provides an overview of historical sovereign defaults. While the use of CACs would have been an option in a large number of debt restructurings over the past years, they have been used in selected cases only. Usually they were used as a practical instrument to sweep-up holdout bondholders in an exchange or restructuring offer.

108 

 Martin Wiesmann

Country

Year

CACs included?

CACs used

Russia

1998

✓

×

Ecuador

1999

✓

×

Pakistan

1999

✓

×

Ukraine

2000

✓

✓

Moldova

2002

✓

✓

Uruguay

2003

✓

×

Dominican Republic

2005

✓

×

Belize

2006

✓

✓

Figure 4: Selected historical defaults and the use of CACs (Source: J.P. Morgan)

Impact of CACs on bond pricing CACs are an instrument used to establish a practical framework for debt restructurings which can help simplify and streamline the restructuring process. Should CACs therefore increase or decrease borrowing costs? The academic research on this question has examined the correlation between pricing and CACs and has found little or no evidence for any pricing effect.3 Figure 5 provides an overview of studies, the underlying data set and the respective findings. Potential explanations for the absence of a correlation include an assumed trade-off between: –– lower ex-post restructuring costs, as CACs allow for easier and faster debt restructurings, therefore effectively reducing pricing costs and –– higher ex-ante moral hazard costs as a debt restructuring gets more likely if CACs are an available instrument, effectively increasing pricing costs Such a trade-off could also be an explanation for empirical results from a few studies which found that CACs increase costs for riskier issuers, but reduce costs for financially sound issuers4: As an investment grade issuer is very unlikely to

3 Such studies include Dixon/Wall (2000), Becker/Richards/Thaicharoen (2001) or Gugiatti/Richards (2003). 4 Eichengreen, Moody (2004)



CACs and the Restructuring of Sovereign Debt – How Would Markets React? 

 109

Countries Paper

EM

DM

Sample time period

Findings

Tsataronis (1999)

✓

✓

1990–1999

no significant impact

Dixon / Wall (2000)

✓

1997–2000

no significant impact

Becker / Richards / Thaicharoen (2001)

✓

1991–2000

no significant impact

Gugiatti / Richards (2003)

✓

✓

1994–2003

no significant impact

Eichengreen / Mody (2004)

✓

✓

1991–2000

CACs reduce/increase pricing for low/high risk issuers

Bradley / Gulati (2011)

✓

✓

1990–2011

CACs reduce pricing for high risk issuers, no impact for low risk issuers

Figure 5: Academic research on pricing impact

find himself in a restructuring scenario, CACs are just an additional safety net for bondholders – lower ex-post restructuring costs more than outweigh any moral hazard costs. For a non-investment grade issuer however, ex-ante moral hazard costs outbalance any positive pricing effects. However, as most of these articles were performed using data from over a decade ago, when the use and variety of CACs was limited and inconsistent across jurisdictions, the practical significance of these results is limited. Recent research from Bradley and Gulati (2011) is based on a much more comprehensive data set and could therefore provide more meaningful results. Bradley and Gulati’s paper analyses the effect of certain CACs on borrowing costs of sovereigns. Their data set includes sovereign bonds issued by over 75 nations across a time period from 1990 to 2011. In total their sample includes 746 observations with NY and UK bonds issued primarily by Emerging Market countries. Bradley and Gulati’s empirical analysis examines the effect of certain independent variables, which represent specific CACs, on one dependent variable, which measures the costs of borrowing. Their analysis is performed on two sub-sets, one comparing NY and UK bonds, one comparing investment-grade and non-investment-grade issuers. The main empirical result of the study indicates that most CACs are associated with lower costs of borrowing for non-investment-grade issuers, while at the same time no significant impact is found for investment-grade issuers. Figure 6 provides an overview of the empirical results for different CACs.

110 

 Martin Wiesmann

Feature

Indication of empirical results¹

Majority action clause (measured as minimum vote required to alter terms)

Lower vote requirement reducing CoB

Mandatory meeting to alter terms

Requirement increasing CoB

Disenfranchisement clause

No significant impact

(Dis-)acceleration clause

Inclusion of clause reducing CoB

Representation clause

Inclusion of clause reducing CoB²

 For non-investment grade issuers only 2 Significant for UK bonds only

1

Figure 6: Empirical results from Bradley, Gulati (2011)

Potential impact on the Eurozone sovereign debt market In 2011, the European Council (“EC”) issued a position paper defining several CACs which could be used as a standard set in sovereign bonds issued in the Eurozone. The proposal intends to establish a common set of CACs but also provides specific wordings for each CAC. As they are based on New York and English law bonds, the included provisions are more or less consistent with other sovereign bonds already known to the market. If an EU state already issues his bonds under those international jurisdictions, the required changes in local laws would be very limited. If a state however issues sovereign bonds under his own local law, he would most likely have to implement much more significant changes to include the new clauses. In particular, the EC proposal includes the following clauses: –– Majority action clause –– Enforcement clause –– Aggregation clause –– Representation clause –– Disfranchisement clause (which bans the voting of bonds held by an issuer or related parties at bondholder meetings or in consent solicitations). A (de-)acceleration clause is explicitly not included. The proposal indicates a mandatory inclusion from 2013 onwards. However, any implementation would take some time to come into effect. If bonds which are currently outstanding do not get exchanged, an introduction of CACs could only take place for newly issued bonds. Assuming a scenario where debt levels largely



CACs and the Restructuring of Sovereign Debt – How Would Markets React? 

 111

remain constant, CACs get introduced if old bonds mature. Figure 7 provides an overview of maturity statistics of European countries. A simulation of the IMF (Figure 8) provides further evidence that a migration towards a full 100% CAC market would take two to three decades, with however a strong momentum in the first 10 years, where the share of CAC bonds could reach 70% to 80%. Country

Average maturity of outstanding bonds (years)

% maturing within the next 2 years

Austria

7.6

11%

Belgium

6.6

36%

Finland

5.4

31%

France

7.2

35%

Germany

6.2

38%

Greece

6.6

35%

Ireland

6.1

18%

Italy

6.9

34%

Netherlands

6.1

42%

Portugal

5.9

36%

Spain

6.6

41%

Figure 7: Average maturity data of European countries (Source: J.P. Morgan research) 100 % 80 %

Note: Assuming a mandatory inclusion of CACs in all new bonds issues from 2001 onwards

Bonds Excl. CACs

60 % 40 %

Bonds Incl. CACs

2035

2031

2033

2029

2025

2027

2021

2023

2019

2015

2017

2011

2013

2009

2005

2007

2001

0%

2003

20 %

Figure 8: Simulation of global CAC share in outstanding bonds (Source: IMF)

112 

 Martin Wiesmann

As the EC proposal was introduced in the context of the developing Euro crisis it may have contributed to the erosion of market confidence and the expectation of a restructuring in Greece. As a fairly new feature of sovereign bond issuance known so far in the Emerging Markets, it raised the fundamental question as to how the Eurozone will deal with sovereign risk in its currency area going forward. As a general concept however, the idea of a standardised set of CACs should genuinely be seen as a contribution to improving the way international capital markets work. To assess its potential impact on the pricing of Eurozone government debt in the future we can take that concept and benchmark it against two fundamentally different ways of how the Eurozone may evolve. Assuming the scenario of a fiscal union with strict surveillance of national budgets, transfer mechanisms and joint liabilities such as Eurobonds, the impact of CACs on borrowing cost would most likely be very limited to non existent: the Eurozone as a whole would be investment grade, and its members would fund themselves under a financing guarantee of the Union making a debt restructuring a very low probability event. By contrast a scenario in which fiscal authority remains at national state level and the Eurozone moves towards a “Maastricht II” governance model, whereby better controls help manage the region, but every member is fully responsible and liable for ensuring sustained market access, would in combination with the no bail out clause increase the probability of a restructuring. In such a scenario, CACs could have a much more significant impact on pricing. While investment grade issuers would be likely to still see no change in borrowing costs, noninvestment grade countries may experience decreasing costs as lower ex-post restructuring costs outweigh higher ex-ante moral hazard costs. In summary the market impact of an area wide introduction of CACs is very likely to be dependent on two factors: The level of political and fiscal integration the Eurozone migrates towards, and the stand alone credit quality of the issuer.

Bibliography Becker, T., Richards, A. and Thaicharoen, Y. Bond restructuring and moral hazard: Are collective action clauses costly? (2001) Bradley, M. and Gulati, M. Collective Action Clauses for the Eurozone: An Empirical Analysis (2011) Dixon, L. and Wall, D. Collective action problems and collective action clauses (2000) Eichengreen, B. and Moody, A. Do Collective Action Clauses raise borrowing costs? (2004) Gugiatti, M. and Richards, A. Do collective action clauses influence bondyields?: New evidence from emerging markets (2003) Gugiatti, M. and Richards, A. The use of collective action clauses in New York law bonds of sovereign borrowers (2004)

Patrick S. Kenadjian1

The Aggregation Clause in Euro Area Government Securities: Game Changer or Flavor of the Month? – Background and the Greek Experience 1. Introduction The European Council decided at its meetings on March 24/25 2011 that: “Collective Action Clauses (CACs) will be included in all new euro area government securities, with maturity above one year, from July 2013. The objective of such CACs will be to facilitate agreement between the sovereign and its private-sector creditors in the context of private sector involvement. The inclusion of CACs in a bond will not imply a higher probability of default or of debt restructuring relating to that bond. Accordingly, the creditor status for sovereign debt will not be affected by the inclusion of CACs.” The European Council communication dated 20 April 2011 which included the above language went on to state that the main features of the CACs will be consistent with those commonly used in the US and UK markets since the Group of Ten report on CACs of 2002, that these CACS “will include an aggregation clause enabling a super majority of bondholders across multiple bond issues subject to such a clause and subject to the law of a single jurisdiction to include a majority action clause where the needed majority of creditors for the restructuration would not be attained within a single bond issue.” The aggregation clause was considered but not adopted by the Group of Ten in 2002 and, in my view, constitutes the key advance in the Council’s new rules which might have the potential to make their version of CACs a game changer, as demonstrated by the use of what have been referred to as “retrofitted CACs” with aggregation clauses by the Hellenic Republic in its February 2012 exchange offers. In these exchange offers, as discussed more fully below, the use of the aggregation clause allowed Greece to move the acceptance rate for its Greek law governed bonds to 100%, thus enabling it to exchange 95.7% of the bonds subject to invitation, after taking into account the much lower acceptance rate of 61% on

1 The views expressed in this article are the author’s own.

114 

 Patrick S. Kenadjian

its foreign law bonds, as of March 9, 2012. After three extensions for the foreign law bonds, the overall result increased to 96.9% by April 25.2 Using “retrofitted CACs” with aggregation clauses, Greece achieved approval of its consent solicitations in all of its Greek law governed bonds, but achieved approval in only 17 of out 36 foreign law governed bonds which contained conventional CACs without aggregation clauses. Still, the tender rate for bonds with conventional CACs, at 78% was higher than the tender rate of 31% in a few series of foreign law bonds where no amendment was attempted, primarily due to local securities law issues in some foreign jurisdictions. While other exchange offers for sovereign debt have achieved comparable levels of acceptance, given the size of the haircut involved (comparable to that for Argentina which achieved only a 76% acceptance rate), the exchange was judged a great tactical success and the aggregation clause included in the “retrofit” clearly accentuated the disparity between the Greek law and foreign law bond acceptance rates. This paper will discuss the background to the adoption of the CAC provisions in the European Union, describe historical alternatives to their use and consider whether the Greek experience fundamentally changes the case for their use. The 24/25 March 2011 Council Conclusions also required a Member State to “engage in active negotiations in good faith with its creditors to secure their direct involvement in restoring debt sustainability” and made “adequate and proportionate private sector involvement” a precondition to receiving financial assistance from the official sector. Thus the initial Council decision on CACs tied them to facilitating private sector involvement (“PSI”) in sovereign debt restructurings and explicitly stated that aggregation clauses would be included in them. The thinking behind this decision was set forth in further detail in a July 23, 2011 communication from the Sub‑committee on EU Sovereign Debt Markets (the “Subcommittee”) of the Economic and Financial Committee (“EFC”) announcing a consultation of “market participants and other stakeholders” on CACs. The original link to PSI led many of us to think that CACs might well be dropped after the December 2011 European Council meeting which officially renounced the use of PSI in further Euro area sovereign debt restructuring after Greece.3 But the CAC project has survived the ostensible official demise of PSI. In fact, paragraph 3 of Article 12 of the modified version of the Treaty Establishing

2 The percentages in this and the next paragraph are derived from Zettelmeyer, Trebesch and Gulati (2012), pp. 5–6 and Appendix, Tables A3 and A4. 3 “We clearly reaffirm that the decisions taken on 21 July and 26/27 October concerning Greek debt are unique and exceptional….” European Council (2011A), paragraph 15.



The Aggregation Clause in Euro Area Government Securities 

 115

the European Stability Mechanism (“ESM”) even brought forward the effective date for inclusion of CACs in new Euro area government securities from the July date cited above to January 1, 2013. The Sub-committee posted the final version of the Common Terms of Reference dated February 2, 2012 (“Common Terms”) on its website as of March 26, 2012, together with a supplemental explanatory note. This final version included Section 2.2, an aggregation clause, referred to in the Common Terms as “cross-series modification.” As adopted, it allows bondholders of separate series voting on common changes of a “reserved matter” (changes in amounts payable, maturity, currency or place of payment, releases of guarantees or collateral, acceleration, seniority or ranking, governing law and court jurisdiction) who fail to meet the minimum vote required for such a change (75% of bonds represented at a meeting or 66 2/3% of bonds outstanding consenting to a written resolution) nonetheless to effect a change if their series has voted in favor of the change by a lower majority (66 2/3% of bonds represented or 50% of bonds outstanding consenting), so long as overall all the series voting, taken in the aggregate, have voted in favor by the higher majority. These percentages were increased from the initially proposed 66 2/3% and 50% respectively as a result of the comment process. That process also led to the quorum for action at a meeting to be set at 66 2/3% of bonds outstanding for both initial and adjourned meetings for reserved matters as opposed to 50% and 25% for such meetings on non-reserved matters. The Common Terms also allow in Section 2.3 partial crossseries modifications so that if not all series reach the lower majority those that do can be modified nonetheless. Finally, Section 2.3 also allows for a menu of alternatives to be offered to bondholders so long as all alternatives are open to each affected series. The intention of the aggregation clause is thus to reduce the ability of non-consenting creditors (“hold-outs”) to prevent a restructuring by blocking approval by a single series if the restructuring is otherwise approved by a large overall majority of the bondholders. The effect of the clause is a bit like that of the “cram-down” provision in Chapter 11 bankruptcy proceedings in the US which allows a plan of reorganization to be approved over the objections of a class of creditors if at least one impaired class of creditors has approved it, but the higher quorum and majorities required were meant to insure that no individual series could be singled out for “invidious treatment”.4

4 EFC Sub‑Committee (2012).

116 

 Patrick S. Kenadjian

2. Background: the IMF, the Group of Ten and the “Hold-out Problem” As the Council communication of April 20, 2011 made clear, the idea behind CACs as currently understood goes back to a Group of Ten Working Group on Contractual Clauses, formed in June 2002 at the behest of the United States and chaired by a former partner of mine, Randall Quarles, then Deputy Assistant Secretary of the US Treasury for International Affairs, to provide a market-based mechanism to facilitate expeditious and orderly restructurings of sovereign debt when debt crises occur. The problem the Working Group was meant to address was the perceived lack of co-ordination among creditors of foreign sovereigns. In the 1980s, when most sovereign debt was in the form of bank loans, the number of creditors was relatively small. An ad hoc group of large banks would negotiate with the sovereign and smaller banks would be pressed to accept their deal by their larger peers, their regulators and multi-lateral organizations such as the IMF and the World Bank.5 With the shift of sovereign borrowing to the bond markets and the rise of secondary markets for bank loans and sovereign bonds, the class of creditors expanded to include investors, such as hedge funds, who were less susceptible to pressure from bank regulators and governments and thus thought to have less incentive to agree to a restructuring. Thus was born the perception of the “hold-out” problem, delicately referred to officially as “the creditor coordination” problem, as a key issue to be resolved, although it was also sometimes referred to, in retrospect, as “the problem that wasn’t”.6 Christian Kopf’s accompanying article and the excellent overview of sovereign debt restructurings between 1950 and 2010 by the IMF Staff published in August 2012 show that the problem was vastly overestimated.7 Nonetheless, the Group of Ten Working Group was meant to tackle this problem, but it was also a counter-initiative to the proposal by the International Monetary Fund in 2002 to institute an international bankruptcy scheme for sovereigns, the “Sovereign Debt Reduction Mechanism” (“SDRM”).8 The objectives of the Working Group, as set out in its September 26, 2002 report, were (i) to foster early dialogue, coordination and communication among a sovereign and its creditors, (ii) to provide effective means for creditors and debtors “to re-contract, without a minority of debt-holders obstructing the process” and (iii) ensuring

5 Verdier (2004). 6 Bi, Chamon and Zettelmeyer (2010). 7 Das, Papaioannou and Trebesch (2012), at 28. 8 Krueger (2002).



The Aggregation Clause in Euro Area Government Securities 

 117

that disruptive legal action by individual creditors does not hamper a workout that is underway. It was the second objective, helping to ensure agreement, that the inclusion of what was then called a “majority amendment clause” permitting amendments of payment terms with the approval of a super majority of bondholders was meant to achieve. The Working Group stated they considered it “perhaps the most critical component” of the package. Bonds issued under English law had long provided for majority amendment, whereas bonds issued under New York law had largely followed the model of the U.S. Trust Indenture Act of 1939, adopted to counter abuses in corporate bond amendments during the Great Depression which required unanimity for changes in principal amount, maturity and interest rate. The Report recommended adoption of a majority amendment clause in sovereign bonds governed by New York law and provided the text of a model clause. The Working Group rejected as “not practicable within a contractually based mechanism” the adoption of an aggregation mechanism across a range of different types of creditors for voting purposes under the majority amendment clause. The Working Group thought aggregation merited further exploration, especially in the context of instruments issued under a single master agreement. The Working Group’s recommendations were widely adopted – without aggregation – by emerging market sovereigns issuing external debt under New York law, the IMF’s SDRM proposal was quietly shelved and, as Bradley and Gulati (2011) note, by 2010, over 90% of all New York law governed sovereign bonds were being issued with a majority amendment clause.9

3. The European Union and CACs a. “Leading by example” The European Union has long been a supporter of majority amendment clauses. As far back as April 2003, a little over six months after the Working Group published its recommendations, the Member States agreed “to lead by example in including collective action clauses (CACs) in their international debt issuance to promote international efforts for orderly restructurings in the event of sovereign debt crises.”10 The idea was to remove any stigma associated with adopting CACs

9 Bradley and Gulati (2012). 10 ECFIN (2004).

118 

 Patrick S. Kenadjian

by stating that the intention was to set an example to other sovereigns, not to suggest that any Member State, nor any accession state, might need to or intend to ever use the clause. In September 2003 the EFC agreed a set of core clauses which were expected to be included in Member States’ loan documentation. It is a small irony of history that Greece held the rotating presidency of the EU at the time. The core clauses included collective representation, majority restructuring, majority enforcement and disenfranchisement provisions. The majority restructuring provisions did not include an aggregation clause and the 2003 CACs initiative applied only to “international debt issuances,” meaning debt issued in foreign currencies or under foreign law, or both. The ECFIN (2004) report noted that, among others, Greece, Italy and Spain, had included majority restructuring provisions in their international debt issuances, Greece, for example, in its €1  billion 2004 issuance under English law, Italy in a series of dollar denominated issuances in 2003 and 2004 under New York law and Spain in a US$ 1.5 billion issuance in 2004 under English law. Issuances in Euros under a Member State’s domestic law were not covered by the 2003 initiative. That, and the inclusion of aggregation clauses, would have to wait until 2011, at which point the issue was no longer “leading by example,” but rather providing an additional tool to the Member States and the Eurogroup to reduce the possibility of a disorderly default by a Member State of the Eurozone.

b. Preventing a disorderly default in an inter-connected financial system In the meantime the world had changed. The financial crisis of 2008/2009 had demonstrated how tightly linked the financial systems of the OECD countries were and how quickly contagion could jump from one financial system to another. While the contagion mechanisms are still rather imperfectly understood, two conduits at least were identified in the aftermath of the crisis, the similarity of the assets held by financial institutions across borders and the unpredictable effects of derivative instruments, in particular of credit default swaps (“CDSs”). The similarity in assets was driven in part by bank regulation, in particular the capital adequacy rules developed by the Basel Committee on Bank Supervision in their second Accord (“Basel II”), which assigned risk weights to bank assets for purposes of evaluating the banks’ capital adequacy. A zero weight was allocated to bonds issued by a bank’s sovereign in the sovereign’s domestic currency. The European Union took this rule and, in the Capital Markets Directive, expanded it to provide a zero risk weight to any bonds issued in Euros by any



The Aggregation Clause in Euro Area Government Securities 

 119

Eurozone sovereign for any Eurozone bank. Thus, a French or a German bank would get the same zero risk weight for holding a relatively low yielding French or German bond as for holding higher yielding Greek, Irish, Italian, Portuguese or Spanish bonds, but if it held the latter it would earn a better return. So it was not surprising to find that at the beginning of the sovereign debt crisis in 2010 the portfolios of French and German banks contained significant amounts of such bonds. Eurozone core banks’ exposure to Greece just about quintupled during the decade between 1999 and 2009. This increase was rather typical of the general increase in exposure to GIIPS (Greece, Ireland, Italy, Portugal and Spain) debt, which ranged from a low of 217% for Italy to 554% for Spain. In absolute terms, the exposure to Italy was the greatest and almost as large as the exposure to the balance of the GIIPS together.11 To quote Baldwin and Gros, “[t]his interconnectedness is critical to understanding why Eurozone leaders could not afford to let Greece face its debt problems alone. … One can never know what history looks like in parallel universes, but the Greek crisis might well have remained a Greek problem if Eurozone core nation banking sectors had not been so exposed to the GIIPS.”12 This meant that a default by one of these sovereigns could have a direct negative effect on banks in other Member States. Yilmaz (2011) and colleagues calculate the volatility connectedness across the 14 leading banks in six Eurozone countries (including France, Germany, Italy and Spain) plus Switzerland as measured by the net connectedness in the movement of their stock prices and concludes that “[t]he banks in the EU and Switzerland are tightly connected in a state of high volatility.”13 There could also be a more indirect effect, depending on the number of CDSs written on the sovereign debt involved, who had written them and who held them. Thus all Member States, especially those in the Eurozone, found they had a rather direct stake in what happened to the debt of every other Member State. The more fragile their domestic banking system and the larger their banks’ holdings of such sovereign debt, the greater the stake. The concern was often phrased as being the preservation of the Euro, but the health of a banking system whose recovery from the financial crisis was at best uncertain was also clearly a factor. The banks which held this sovereign debt and their governments were in a quandary. To encourage the banks to dump their holdings of “peripheral” sovereign debt would risk two negative consequences. The sales would drive down

11 Baldwin and Gros (2010). 12 Baldwin and Gros (2010), at 13. 13 Yilmaz (2011).

120 

 Patrick S. Kenadjian

the price of such debt, increasing the costs of borrowing for those sovereigns and exacerbating their financial distress. The sales would also diminish the value of that sovereign’s debt held by all other banks, both as collateral and under applicable accounting rules, leading to two sets of problems, an immediate liquidity squeeze and an eventual income statement loss. The most immediate problem was that a decline in the secondary market value of the bonds would entail a corresponding decline in the collateral value of those bonds. Since many banks had financed the bonds in the repurchase (“repo”) market, this would require the banks to post additional cash margin to maintain the financing, which many banks did not have the liquidity to provide. This would in turn lead to further sales of bonds. Also, under applicable accounting rules, the banks would have to record losses immediately to the extent they held those bonds in their trading books, i.e. as trading assets or as assets available for sale, under the “mark to market” accounting rules. If they held the bonds in their banking book, i.e. as held to maturity, they might be able to avoid losses for the moment, so long as they could convince their auditors the decline in price did not reflect a permanent impairment of the value of the bonds, but they would be sitting on potentially large losses which could negatively affect their future stability should an impairment occur in the future, for example through restructuring or default by the sovereign. Some banks might even not be able to dispose of the sovereign bonds if they wanted to. This would likely be the case of banks which had issued “covered bonds” and segregated the sovereign bonds on their books as part of the “cover” for those covered bonds. Coincidentally or not, the German banks who ended up with the highest concentration of Greek sovereign debt, Commerzbank, Hypo Real Estate and some of the Landesbanken, were also, directly or though subsidiaries, among the largest issuers of Pfandbriefe, the German covered bonds. Those banks’ only recourse might be to hedge this exposure by purchasing CDS protection for the sovereign bonds, but that protection was becoming more expensive as the sovereign debt crisis progressed. The governments also understood, after the round of bank bail-outs in 2008/2009, how expensive and unpopular a further round of direct support for their banks might be and the experience of Ireland had shown that there was a point at which such support could imperil the state’s own creditworthiness. While the size of Ireland’s banking sector as a multiple of GDP (about seven times) was exceptionally high, multiples of two to four were not unusual. Hence the search for a way to control the potential fallout from a “peripheral” Eurozone sovereign default, which eventually led to CACs for all sovereign debt with a maturity in excess of a year, as one piece of the puzzle.



The Aggregation Clause in Euro Area Government Securities 

 121

The move from just international debt to all sovereign debt can in part be seen as pure pragmatism. Over 90% of Greece’s debt was in euros and governed by Greek law, so excluding bonds issued under domestic law would have excluded 90% of Greek sovereign debt. And the Greek situation was not exceptional. Recent research by Nomura Securities estimates that only 7% of sovereign bonds issued in the Eurozone are governed by foreign law.14 The move also put domestic debt which is largely issued without collective action clauses on a par with international debt, which typically does include such clauses.15 The move can also be seen as a recognition that, as Christian Kopf was one of the first to point out, in his brilliant piece for CEPS, the Euro is the equivalent of a foreign currency for Member States that are no longer in control of the printing press for the Euro.16 Thus, as a former Governor of the Central Bank of Argentina has pointed out, Greece (or Italy) borrowing in Euros is closer to Argentina borrowing in US dollars than to the UK borrowing in Pounds Sterling.

4. Emerging Markets Sovereigns and CACs a. To use or not to use But, to return to emerging market sovereigns, where majority amendment clauses got their start on the sovereign debt stage, one peculiarity must be noted: despite the presence of such clauses in their debt, most sovereigns continued to restructure their debt using exchange offers in which they offered existing bondholders a menu of alternative new instruments in exchange for their existing debt, rather than asking them to vote on amending the payment terms of that debt. To the extent they asked investors to vote on anything, it was on “exit consents” relating to non-payment terms, as discussed more fully below. The new instruments could have a number of features, but, starting with the “Brady Bonds” issued to resolve the Latin American debt crisis of the 1980s, in general there was at least one “par bond” whose principal amount would be substantially equivalent to that of the old bonds but paid a lower interest rate and had a much longer maturity, and one “discount bond” whose par value was lower than that of the old bonds, roughly reflecting the current market value of the old bonds, but paid a higher interest rate

14 Nordvig and Firoozye (2012). 15 Allen & Overy (2011). 16 Kopf (2011).

122 

 Patrick S. Kenadjian

than the par bonds. The par bonds were meant to appeal to the initial investors who had bought at par, not written down their value and did not want to record a loss on the exchange. The discount bonds were meant for more recent investors who had purchased the old bonds at a discount in the secondary market. Gros and Mayer (2011) note that the first EU-wide stress test in the summer of 2010 showed that close to 90% of all government debt held by banks was in the “banking book”, thus not subject to mark to market.17 The European Banking Authority (“EBA”) in announcing the results of its latest stress tests noted that as of September 30, 2011 only 43% of EEA sovereign debt was so held. Whatever the exact amount, these banks are the natural market for the “par bonds.” Sturzenegger and Zettelmeyer (2007) in their excellent survey of sovereign debt restructuring cite only two direct uses of majority amendment clauses in sovereign debt restructurings, once by Moldova (June 2002) and once by Uruguay (May 2003) as well as one indirect use (i.e. as a threat) by the Ukraine in 2000 in connection with something that looked very much like a traditional exit consent procedure in an exchange offer.18 In contrast, three times as many sovereign debt restructurings have been accomplished through the use of exchange offers. It is likely that the majority amendment clauses involved, with a few exceptions noted below, followed the G-10 model and did not provide for aggregation, so that CACs as proposed for the Euro zone would possess that one, perhaps crucial, difference. Would this make CACs a game changer for sovereign debt restructuring or merely another potential tool in the arsenal of a sovereign debtor? Prior to the Greek exchange offers of 2012, the question could only be answered analytically and theoretically since, although Argentina, the Dominican Republic and Uruguay have included aggregation clauses in their sovereign bonds, none of them had actively used these clauses in a restructuring.19 The IMF 2012 study includes an interesting table showing sovereign restructurings with and without the use of CACs from which it is hard to establish a correlation between their use and the percentage of hold-outs. Pakistan did not use them in 1999 and had 1% hold-outs. The Seychelles and Ukraine used them in 2009 and 2000 respectively and had 16% and 3% hold-outs respectively. The consensus view on the utility of

17 Gros and Mayer (2011). 18 In a traditional exit consent, bondholders who exchange their old bonds are asked to consent to changes in the non-payment terms and conditions of those bonds to make them less attractive to non-exchanging bondholders. Such changes could involve the removal of covenants, both negative (e.g. a negative pledge) and affirmative (e.g. to maintain a listing for the bonds) but do not affect principal, interest, currency of payment or maturity, the so-called “reserved matters” under CACs. Bi, Chamon and Zettelmeyer (2010). 19 Bradley and Gulati (2011), Dey (2009).



The Aggregation Clause in Euro Area Government Securities 

 123

CACs prior to the Greek experience was perhaps best summed up by Alessandro Leipold as follows: “collective action and aggregation clauses no doubt help at the margin but have not shown themselves to be decisive in debt restructurings.”20

b. How the game is played: tactics and fundamentals To understand the calculus involved for both creditors and sovereign debtors it is important to understand that exchange offers, while do-able, are complex and potentially expensive undertakings which expose both sides to risks and costs. Sovereign debt restructurings implemented through debt exchange offers take time. Three months is considered fast, ten to twenty months normal and, in particularly contentious cases, such as Argentina, forty months was necessary.21 Nouriel Roubini, who has written extensively on the area, considers six months the minimum normal time period. The reasons for these time lags lie on both sides. Governments are generally reluctant to restructure if they can avoid it, out of fear of being shut out of the international capital markets on which they are more and more reliant for their financing.22 Their creditors often profit from a long drawn out process which gives them time to trade out of their debt, write it down gradually or build up reserves to absorb the losses the restructuring will generate for them. This was famously the case for US banks holding Latin American debt in the 1980’s. With changed accounting rules, such a delay is less likely today, but even when the parties are convinced a restructuring is imminent and that a speedy agreement will benefit both parties, the issues involved are difficult ones. In effect, they revolve around the question of how large the pie which can be divided up among the creditors and the debtor (essentially the government’s share of its country’s GDP) is and how it can be divided up in a way which makes the new debt sustainable for the debtor (based on the portion of the primary surplus the debtor can expect to generate which can reasonably be expected to be dedicated to debt service, at

20 Leipold (2011). 21 Dhillon et al. (2006). 22 A 2009 survey found that it took a defaulting sovereign on average 5.7 years to regain partial market access and 8.4 years on average to regain full access. Richmond and Dias (2009). This used to be a problem primarily for emerging market sovereigns and the major examples have come from Latin America, Asia and the former Soviet Union’s successor states. But increased borrowing by all sovereigns has meant that with few exceptions most OECD countries now borrow as much, if not more, from foreign creditors as domestically and most of this comes via the capital markets.

124 

 Patrick S. Kenadjian

a given debt level).23 The primary surplus is the excess of government revenues over government expenses exclusive of debt service, expressed as a percentage of a country’s GDP. As Gianviti, Krueger, Pisani-Ferry and their co-authors (2010) point out, “the main asset of a sovereign debtor is its power and its capacity to tax, which is intangible in nature … [t]he economic value of this asset depends on the degree of hardship a country’s citizens are willing to bear to service its debt and on the government’s administrative capacity to raise revenues (i.e. on the primary surpluses the government is able to achieve) and is, therefore largely, an issue of ‘political judgment’.”24 As a consequence, “the maximum feasible primary balance is not known with certainty (until after it is reached).”25 Given that the creditors have no interest in making the first move in a process to reduce their possible recovery on the debt they hold, the first move is usually made by the sovereign. The inherent uncertainty of both the numbers involved and the political nature of the second calculation26 lead to a great deal of uncertainty as to the size of the pie. The difficulty of arriving at reliable numbers and the sensitivity of such numbers to even small changes in assumptions is demonstrated in the IMF’s Preliminary Debt Sustainability Analysis for Greece of February 2012. The sovereign thus usually opens the bidding by making a more or less aggressive first offer. Whether this move succeeds will depend on whether the creditors think they have more to gain out of delay and what their alternatives are to put pressure on the debtor, usually through litigation. Where the debtor has stopped making interest and principal payments on the existing debt and its economy is improving, the incentive for creditors to hold out in the face of an offer is small, unless they think they have a winning litigation strategy. Where, as has been the case so far within the European Union, the debtor has not instituted a payment moratorium or defaulted on payment, the incentive to hold out is greater, in particular with respect to bonds which have a relatively short remaining term. On the other hand, where the debtor’s economy has not yet stabilized and the prospects of successful litigation are uncertain, the incentive to hold out will be lower, especially for bonds governed by the debtor’s domestic law and containing few, if any, covenants protecting the creditors.

23 Gros and Alcidi (2011). 24 Gianviti, Krueger, Pisani-Ferry (2010). 25 Escolano (2010) at 11. 26 Keynes famously observed in “A Tract on Monetary Reform” (1923) that there was a limit to how much of its revenues a sovereign government could share with rentier investors to the detriment of its own citizens. Where citizens are being asked to make significant sacrifices to generate this primary surplus, the issue can become an acute one.



The Aggregation Clause in Euro Area Government Securities 

 125

However, beyond these tactical considerations, the bottom line given by an analysis of participation rates and speedy completion of recent debt restructurings is that offers which are seen as honestly reflecting the governments’ capacity to pay and which offered appropriate burden sharing between debtor and creditors have tended to be successful.27

5. The Greek Experience a. The February 2012 Exchange Offers The series of exchange offers and consent solicitations launched by the Hellenic Republic on February 24, 2012 can be considered a tactical masterpiece. They constituted the largest sovereign debt restructuring in history and resulted, by the initial deadline set for March 8, 2012, in an aggregate exchange rate of 95.7%, despite involving a haircut variously calculated at between 53.5% and 77.1%.28 For each €1,000 of initial principal amount, the exchanging bondholders received a package consisting of €315 of new Greek government bonds (“GGBs”) due between 2023 and 2042; €150 of European Financial Stability Fund (“EFSF”) notes due in 2013 and 2014, a variable amount of EFSF notes due in 2012 in discharge of interest accrued on old bonds, and detachable GDP linked notes, regardless of the initial due dates of the exchanged GGBs. The high exchange rate was achieved through the use of CACs or CAC like devices including, for Greek law governed bonds (“GLGBs”) an aggregation clause which allowed the Republic to increase the acceptance rate for those bonds to 100%. The acceptance rate among bonds governed by foreign law (“FLGBs”) was considerably lower, at 69% by the initial March 8 deadline. After three extensions of the offers for the FLGBs, the offers closed on April 25 with an overall exchange rate of 96.9%. While this overall rate is comparable to that achieved by Uruguay, Ecuador, Pakistan or the Ukraine, the haircuts (reduction in net present value of

27 IMF 2012, citing Bi, Chamon and Zettelmeyer. 28 A more sophisticated analysis by Zettelmeyer, Trebesch and Gulati (2012) put the actual aggregate haircut at 55–65%, depending on the remaining maturities of the original debt to be exchanged.

126 

 Patrick S. Kenadjian

the bonds involved) of those exchanges had been much lower. The closest haircut was that of Argentina, at 70% and its initial acceptance rate was about 76%.29 This result was achieved without recourse to many of the traditional instruments of debtor coercion, such as payment defaults, declaration of a payment moratorium or explicit threats of either. The Republic was under severe time pressure because it had GLGBs in the principal amount of €14.5 billion coming due on March 20, which it could not afford to pay, so it had to have implemented the exchange offers with respect to those bonds before that date so that they could be cancelled and replaced with new bonds before they became due. Although the broad economic outlines of the exchanges had been agreed on October 26, 2011, key elements of the exchange package had only been cleared with the “Troika” (the IMF, the ECB and the EU) on February 21, 2012, so that the offers had only been able to be launched on February 24. The Republic had initially applied rather minimal coercion on the offers. It had not defaulted on any payments. Its Parliament had adopted a CAC “retrofit” law in advance of the launch of the offers (effective February 23, 2012), the Greek Bondholder Act which included an aggregation clause with somewhat lower thresholds than those adopted in the final Common Terms and Conditions – a quorum of 50% versus 66 2/3% and a favorable vote of 66 2/3% versus 75% – which corresponded to the initially proposed percentages in prior drafts of the Common Terms. The Act allowed the government to make the clause applicable (“retrofit” it) to any outstanding debt for which it had received the required vote and quorum and required only an overall vote of 66 2/3% of all retrofitted bonds, with no separate vote required on a series by series basis to amend those bonds. The February 27, 2012 Invitation to Bondholders document noted that this power could be exercised at the government’s discretion but did not indicate whether the government would make use of that power and there was a good deal of uncertainty as to whether Greece’s public sector interlocutors, in particular the ECB, would acquiesce to its use. The closest the government came to threatening hold-outs with an outright default came at a meeting held in Frankfurt on March 5, 2012 at which, according to the Greek Public Debt Management Agency’s press release of March 6, “[t]he Republic’s representative noted that Greece’s economic programme does not contemplate the availability of funds to make payments to private sector creditors that decline to participate in PSI.” Finance Minister Venizelos was also quoted by Reuters on the same date as saying “[w]hoever thinks they will hold out and be paid in full, is mistaken.” Zettelmeyer, Trebesch and Gulati rate the

29 Allen & Overy (2011).



The Aggregation Clause in Euro Area Government Securities 

 127

Greek exchange offer as a 2 on a coercion scale of 1 to 10, with Uruguay’s 2003 exchange offer which merely extended maturities ranking as a 1 and Argentina’s 2005 external debt exchange ranking as a 9.30 The temptation to hold out must have been particularly great at the shorter end of the maturity scale of the outstanding bonds. This is always so, especially where there has been no payment default so that bonds due a few weeks before the exchange were paid in full, while the next bonds due took a haircut, but the “one size fits all” nature of the Greek exchange offer, where all bonds, regardless of maturity, got the same package of securities, exacerbated the situation. It meant that holders of the bonds due March 20, 2012 were being asked to take a haircut of close to 80%, while holders of the longer dated bonds were taking a haircut of 25 to 34%. The longest dated bond arguably got a premium.31 This disparity in treatment reflected the Greek intention to make their offers as simple as possible to understand, but there is nothing in the EU CACs Standard Terms which would prevent another Eurozone sovereign from doing the same thing, as neither the Standard Terms nor the accompanying commentary include any limitations or guidelines on how bonds can be aggregated. It has been suggested that it might be desirable to supplement at least the commentary or develop a code of conduct which would require a sovereign to offer more substantially equivalent consideration to any classes of debt which are aggregated for purposes of voting. The exchange offers involved 75 separate series of bonds, with maturities between 2012 and 2057.32 There was very little uniformity in the terms and conditions of the various bonds, especially those under domestic law, none of which had CACs. Each series, some of which were quite small, could be subject to separate hold-outs. In order to incentivize the existing creditors to tender Greece offered credit enhancements, which took several forms. Part of the principal (15% of the initial face amount) was in the form of one and two year EFSF bonds which Greece was to acquire as part of the €130 billion second bail-out package agreed on February 21, 2012. That credit was clearly superior to Greece’s. The new GGBs the Republic issued (equal to 31.5% of the initial face amount) were to be governed by English law and no longer Greek law, meaning Greece was renouncing

30 Zettelmeyer, Trebesch and Gualti (2012), Table 3. I note that my friend Christian Kopf in his contribution to this volume takes exception to characterizing the Greek exchange offers as noncoercive. I concede his points are well taken but think the truth lies somewhere between the two positions. 31 Zettelmeyer, Trebesch and Gulati (2012), Table 2. 32 The Hellenic Republic, Invitation Memorandum dated 24 February 2012, Annex I.

128 

 Patrick S. Kenadjian

any further right to amend them unilaterally, and to benefit from a Co-Financing Agreement with the EFSF which was meant to insure the EFSF could not be repaid prior to the holders of the new GGBs. Furthermore the new GGBs could not be further amended without the EFSF’s consent and its consent was also required for Greece to issue additional bonds beyond a permitted ceiling, set initially at €70 billion. Bondholders also received detachable GDP-linked notes meant to provide additional compensation if the Greek economy performed better than expected. In addition to these enhancements to the new bonds, the Greek government needed a temporary enhancement for some of its existing bonds which had been pledged by Greek and other EU banks as collateral for central bank loans. Because, during the time of the exchange offer, its existing debt was rated “selective default” by the credit rating agencies, that debt was no longer eligible as collateral for those banks to maintain their existing advances from the ECB or other national central banks. Thus the Greek government needed additional measures to ensure the existing lines of its domestic banks at the ECB (amounting to €73.4 billion as of the end of November 2011) were not called by the ECB. This was done through guarantees of up to €35 billion issued by the Eurozone governments in favor of the ECB. Exchange offers usually have the problem that any bonds tendered by creditors are locked in limbo until they are accepted by the sovereign and the offer closed. During that period, the creditors are not able to trade out of their positions even if it looks like the exchange may fail. Tendering late is the obvious remedy for this risk, but a sovereign is usually under no legal obligation to complete the exchange promptly after it expires and to the extent the enhancements it needs are subject to actions by third parties (approval by national parliaments to allow the issuance of EFSF bonds in Greece’s case for example), it may not be able to control the timing of the conditions involved, thus putting tendering holders at risk and diminishing their inclination to tender. Greece’s case was different in that it only agreed its second bail-out package on February 21 and so could only send out its invitation memoranda on February 24, but had €14.5 billion of bonds coming due on March 20, so could not afford to delay completion of the exchanges with respect to those bonds at least, so it completed the exchange promptly. Thus, the actual time period the exchange offers were open was quite short, but if one recalls that the initial PSI had been agreed on July 21, 2011, the restructuring had required seven and a half months.



The Aggregation Clause in Euro Area Government Securities 

 129

b. CACs and Exchange Offers Compared The Hellenic Republic combined the two major forms of debt restructuring in its February 2012 exchange offers. As noted above, it offered to exchange existing debt for a package of new securities, but it also solicited consents from its existing debtholders to amend the terms of any existing debt which was not exchanged to allow it (a) with respect to GLGBs, to carry out the exchange with respect to any non-tendered bonds; and (b) with respect to FLGBs to amend their terms to match those of the exchange. The votes were pursuant to the terms of the CACs included in its foreign law bonds and for its domestic law bonds in accordance with the terms of the Greek Bondholder Act, which effected the “retrofit” of a CAC-like mechanism in those Greek Law bonds, which lacked CACs. Any holder of GGBs which tendered its bonds also had to appoint irrevocably an entity designated by the Republic as its proxy to vote in favor of the amendments. As noted above, emerging market sovereigns have combined exchange offers with exit consents, which amend the non-payment terms of the bond which are not exchanged, to make them less attractive to potential hold-outs. The Greek exchange offers were among the first to combine exchange offers with a fundamental change in the payment terms of the existing bonds. The basic difference between the two forms of debt restructuring is that at the end of the day, even though exchange offers have tended to have high success rates, in many cases well over 90%, they do leave some bonds outstanding with their original terms still intact. In contrast, the great advantage of a majority action clause is the ability to bind all holders by a majority vote of the holders of a series or, in the case of an aggregation clause, by the majority vote of a super majority of all aggregated series, as well as of a lower percentage in each series to be amended. This of course assumes the sovereign has reached the required percentages for amendment. If the vote falls short, say 69% where 75% is required, as was the case with some FLGBs, the advantage of the two procedures are reversed, since the sovereign could accept the tenders, retiring 69% of the bonds, but would have failed to amend the bonds if all it had sought was an amendment. Thus, assuming the required percentage is achieved, amendment has the advantage of unanimity, but with it come at least two important disadvantages, (i) a majority vote binding on all holders will trigger a “credit event” under the most commonly used formulations for terms and conditions of credit default swaps written on the sovereign’s debt obligations and (ii) binding all holders may entail having to bind two kinds of holders a sovereign would not generally target for concessions: the

130 

 Patrick S. Kenadjian

public or official sector and retail investors, especially domestic retail investors, i.e. its own citizens.33 The latter feature makes CACs a bit of a blunt instrument. Sovereigns have sometimes tried to avoid requiring retail investors, particularly its own citizens, or its domestic banks, to share in the sacrifices required of international financial investors. This used to be easy when there was a clear difference between external debt and internal debt of sovereigns. In the era of sovereign bond issuances in Euros within the Eurozone this distinction has become less clear. For example, over 90% of GGBs were governed by Greek law, but over 61% of them were estimated to be held at one point by foreign investors. Lascelles (2011). The Eurozone, where the distinction between the Euro as domestic currency and as international reserve currency has become blurred, presents the most acute form of a problem which has been evolving at least since the 1990s, as the distinction between the sovereign debt instruments which domestic investors and institutional investors purchase has become increasingly less clear. Gelpern and Setser give an excellent example of this in their analysis of the Russian debt default on ostensibly domestic debt (Russian law governed and Ruble denominated instruments known as GKOs) but 40% held by foreign investors in August 1998.34 The default was resolved in a series of exchange offers in which domestic and foreign investors were treated differently. Even in a single exchange offer, small domestic holders can effectively be spared by either not targeting them or by setting minimum tender conditions which do not require their participation. so that they can continue to hold the old bonds. This is clearly not bulletproof as international holders can also seek to take advantage of the lower threshold for acceptance, but

33 It is also not clear how easily a CAC can be designed to provide the flexible menu of options exchange offers usually do. As noted above, in exchange offers since the Brady Bonds the sovereign has traditionally offered a “par bond” meant to be attractive for creditors who have bought their bonds at par and not written their value down (for example because they hold them to maturity and thus have been exempt from the requirement to mark them to market) and a “discount” bond meant to appeal to creditors who have bought their bonds at a discount in the secondary market and/or have to mark them to market on a continuous basis. Section 2.3 of the Common Terms and Conditions is intended to accommodate this device, so long as all alternatives are open to all affected series. However, the Greek exchange offers of February 2012 provided a single alternative for all holders described above. This lack of alternatives was ostensibly due to a decision taken at the Eurogroup October 26, 2011 summit that PSI required a visible haircut for all creditors, as well as an overall desire for as much simplicity in the structure as possible, rather than to a lack of flexibility of the CAC instrument, but so far we have no example of its use to provide a menu of alternatives. 34 Gelpern and Setser (2004).



The Aggregation Clause in Euro Area Government Securities 

 131

it is less coercive. In contrast, in a majority amendment procedure, they must be bound just like all of the institutional investors. The same applies to public or official sector entities, such as the ECB and national central banks (“NCBs”) who hold bonds of the series to be amended. In the past this has not been a problem, since public sector entities usually extended credit separately and on terms which made clear their priority over other creditors. This is particularly true of IMF lending and it will be the case for lending by the European Stability Mechanism (“ESM”) which “will enjoy preferred creditor status, junior only to the IMF loan.” European Council (2011)35. But in a world where the ECB and NCBs intervene in the market with a public policy purpose by buying bonds of the same series as those held by the private sector, if the sovereign has recourse to a CAC to amend the terms of a series some of whose bonds have been purchased by public sector entities, as demonstrated in the case of Greece, the sovereign may need to conduct a separate exchange with these entities before it undertakes the amendment procedure to avoid sweeping them into the sacrifices required of the private sector. Putting aside the question of whether such differential treatment is justified by systemic considerations, the Greek experience in 2012 demonstrates that this problem can be handled by doing a series of quick exchanges with a small number of public sector entities. The problem of retail investors is less susceptible of easy resolution since the sovereign is dealing with a large number of small investors and thus cannot use the same procedure to take them out. If the domestic investors to be spared are domestic banks, then a series of similar exchanges can be used, as Russia did in 1999.36 One final issue the Greek experience highlighted was that, where domestic banks hold a significant proportion of the sovereign’s debt (here Greek banks held about 24% of it), a significant haircut is likely to lead to the need to re-capitalize its banking sector. In Greece’s case this need was estimated between €24 to 30 billion, which of course had to be included in the bail-out package, resulting in a corresponding increase in Greece’s sovereign debt and thus a corresponding decrease in the net deleveraging produced by the exchange offers.

35 This priority granted to ESM loans raises a question about possibly triggering a Restructuring Credit Event under Credit Default Swap contracts by resulting in subordination of existing debt to it. The IMF’s seniority has been determined not to trigger a Credit Event since its loans are meant to help the sovereign. Since the same argument can be made for the ESM, recourse to its loans may likewise be held to be exempt from the trigger, but as of this writing, that question is still open and has recently been raised by analysts in connection with the currently proposed bail-out of Spanish banks. Citi (2012). 36 Gelpern and Setser (2004).

132 

 Patrick S. Kenadjian

c. Dangers of Triggering Sovereign CDSs i. A new problem in sovereign debt restructurings Another relatively new problem which arose in the Greek context was the potential triggering of sovereign CDSs. This potential problem had been flagged by some observers for a number of years, but there was relatively little data to go on to quantify it37. In the Greek case, the uncertainty as to how this would work and what effect it would have in practice contributed to slowing down the Greek restructuring process, as the ECB and certain Eurozone governments pressed for the exchange to be “voluntary” for a number of reasons, including in order not to trigger CDSs on Greek government bonds. From a legal point of view, it is quite clear that a “voluntary” exchange offer, no matter how “coercive” some of the bondholders may feel it to be, will not, under current settled interpretations of the current (i.e. 2003) version of the standard Credit Event (i.e. default) definition used in ISDA documentation for Western European sovereign CDSs, trigger a Restructuring Credit Event and thus a payment obligation under the relevant contracts.38 The reason is straightforward: an exchange does not involve a change in the terms of the existing debt that is binding on all holders. Thus, unless it is combined with another action which is covered by the Restructuring or another Credit Event, an exchange does not constitute a Credit Event under the 2003 text. Its status under the prior (1999) version is more ambiguous, as discussed below. In this connection, a sovereign needs to be careful what it says and what it may ask the creditors to vote on in connection with the exchange. For example, several authors suggest the sovereign should threaten to discontinue payments on old bonds that are not exchanged as a tactic to encourage tenders. This can be a dangerous tactic if the idea of using the exchange offer route is to avoid triggering the sovereign’s CDSs, as that statement could be interpreted to constitute a Repudiation or Moratorium Credit Event, independently of the exchange. In this respect, the March 5 statement cited above was a masterpiece of a non-threatening factual statement that there was no money budgeted to pay hold outs. In

37 Gelpern (2008). Das et al. (2012, at 58) note that CDSs were triggered in December 2008 when Ecuador refused to make an interest payment on its 2012 global bond and that this was followed by a standard ISDA auction of “cheapest to deliver” bonds which yielded a recovery rate approximately in line with the trading price of those bonds, but in the run-up to the Greek exchange offers everyone seems to have forgotten this event. 38 Allen & Overy (2011).



The Aggregation Clause in Euro Area Government Securities 

 133

contrast, it is quite clear that a modification of payment terms to the detriment of existing bondholders through a majority vote will trigger a Restructuring Credit Event, and require payment. The use of the term Credit Event in ISDA documentation has led to some confusion given the difference in treatment of restructurings under ISDA documentation and what the credit rating agencies consider a credit event for purposes of a ratings downgrade. The rating agencies generally regard an exchange which results in more disadvantageous terms for the holders of debt of an issuer in financial difficulty and which is conducted under an express or implicit threat of default as constituting a “selective default,” resulting in an “SD” or “D” rating for the duration of the exchange.39 Their action is totally unrelated to the triggering of Credit Events under ISDA documentation and can, in any event, be expected to be reversed upon completion of the exchange offer which, if successful, is usually also accompanied by an upgrade of the sovereign’s credit rating, as was eventually (though not immediately) the case for the February 2012 Greek debt exchanges.

ii. The nature of the problem The triggering of a Restructuring Credit Event can have two distinct consequences. First, it can influence the voting behavior of bondholders who have purchased CDS protection. Since a default means that they will be made whole and are thus what is sometimes referred to as “empty creditors,” they may be inclined to vote in favor of a restructuring even if it results in otherwise cutting existing bondholder recovery too deeply. In a sense, the presence of CDS contracts may effectively expand the pie available to be shared among creditors and the debtor by the amount of the net protection purchased. It is also possible that such holders may be inclined to vote against a restructuring in the hopes of triggering an actual payment default, which would also result in their receiving a payment. In an exchange offer, their motivation would normally be not to exchange in the hope that the offer fails and is followed by a payment default in which they would be made whole. Either way, the protection provided by the CDS skews their attitude towards the outcome of the vote, since the risk is held by the protection writer and not by them. The protection provider does not get to vote or decide whether the protection buyer does either. But it is also important to note that a portion of these CDSs may be held by investors which do not hold bonds, but are using them

39 Roubini (2011).

134 

 Patrick S. Kenadjian

to hedge a correlated risk, and thus have no vote and no bonds to exchange and consequently they may have no impact on the outcome of the vote or exchange offer.40 It is, however, fair to conclude along with Das et al. that it is difficult to assess whether the empty creditor problem will be a major concern for sovereign debt restructurings. As Das et al. point out, there is very little academic evidence of an effect from the corporate debt sector and, as discussed below, that sector typically has a significantly larger proportion of CDSs to debt outstanding than the public sector.41 Second, a triggering can contribute to the contagion effect of the restructuring to the extent the protection has been written by a few financial institutions which may incur substantial losses upon the triggering of the defaults, especially in the case of cash settled CDSs, where their counterparty does not have to deliver the underlying bond and simply receives a cash payment upon default.42 The existence of CDSs thus can amplify the consequences of a default or a restructuring through the use of CACs in a way which is difficult to predict, given that the CDS market is still an over-the-counter market with a relatively low degree of transparency, so that it is hard to say with certainty what effect a default on a single name will have on the overall stability of the financial markets. We do know what events count as Credit Events which trigger a payment obligation under CDS contracts. This depends on the formulation in the two ISDA sets of definitions of the term Restructuring used in sovereign CDSs, the 1999 Definitions and the 2003 Definitions. There are three types of Credit Events in Western European sovereign CDSs: failure to pay, repudiation/moratorium and restructuring. Failure to pay occurs when an interest or principal payment is not made when due. Repudiation/moratorium occurs when a sovereign announces it will not honor its debt or make payments on it, as Ecuador did in 2008. Restructuring involves a reduction in principal or interest, a change in maturity or currency of payment (to a non-G-7 or non-AAA rated OECD currency) or a (legal) subor-

40 On February 21, 2012, the Council of the EU adopted a Regulation on short selling and credit default swaps to prohibit the writing of such contracts on sovereign debt where the protection buyer does not have an “insurable interest” in the sovereign issuer. Council of the European Union (2012). If the Regulation comes into force as scheduled on November 1, 2012, it will eliminate this aspect of the CDS problem. 41 Das et al. (2012) at 59. 42 In the case of physically settled contracts, the protection buyer must deliver an eligible bond in order to receive payment. Since under existing law the buyer can buy protection without owning any bonds, a default may also trigger a “short squeeze” which results in the price of the bonds going up as protection buyers scramble to cover themselves with bonds eligible for delivery under the CDS contract and ISDA rules.



The Aggregation Clause in Euro Area Government Securities 

 135

dination done in a way which is binding on all holders regardless of whether they consent to the change or not, the route Greece eventually chose in February 2012. Both the 1999 and 2003 Definitions of Restructuring require a change in payment terms which result from an agreement between the debtor and a majority of holders sufficient to bind all holders. The 2003 Definitions are clearer in their formulation, with respect to the effects of a debt exchange43, but both have been interpreted by ISDA Determination Committees (the committees made up of ISDA members whose job it is to determine whether a Credit Event has occurred) to exclude voluntary debt exchanges where only the exchanging creditors are bound by the terms of the new bonds.44 Thus, the use of a CAC to amend the payment terms of a sovereign bond would be caught as a Credit Event under both the 1999 and the 2003 ISDA Definitions and would trigger payments under CDSs written on the sovereign (i.e. where the sovereign is the “reference entity”), while the same result achieved through a traditional debt exchange would not. So the question becomes, how much should a sovereign and its unhedged creditors care about that difference? Unless the institutions writing the net protection are overwhelmingly domestic banks whose collapse would exacerbate the sovereign’s domestic problems and overall financial situation (on the not unrealistic assumption that it would need to recapitalize at least some of them), the sovereign may not care much. But even then, this may only concern the sovereign at the margin since its banks can be expected to be even more directly affected by the haircut required of them on their holdings of sovereign debt and need restructuring anyway, as was the case with Greece. The creditors who are in favor of the restructuring, unless they have written the protection themselves, which would amount to doubling down on their exposure to the sovereign, may also be relatively indifferent, unless they fear for the solvency of some of their counterparties who may have written such protection. On the contrary, both the sovereign and the accepting creditors may welcome the ability to ensure a higher level of participation and thus a more effective restructuring. But actors that have a stake in the stability of the financial system, including the institutions of the European Union to which a Member State would normally turn for assistance in case of financial difficulties, can be expected to discourage any course of action which would run the risk of destabilizing financial insti-

43 The 1999 version had included the term “mandatory transfer” which was the subject of litigation in New York in connection with a 2001 Argentine mini-debt exchange. The Second Circuit Court of Appeals found the term ambiguous, which led ISDA to change the definition in 2003. Gelpern (2008). 44 Verdier (2004).

136 

 Patrick S. Kenadjian

tutions in any Member State which may be net writers of protection (as well as otherwise exposed to the sovereign’s debt) if this could have knock on effects for the financial system as a whole. In the Greek exchange offer case this difference of views opposed the government of Greece and the ECB in a classical situation of imperfect information.

iii. Quantifying the problem The question was, what did we think we knew about CDSs on Member State sovereign debt? We thought we knew that the net notional of CDS contracts was rather small compared to the stock of government debt outstanding. Based on their analysis of Depository Trust and Clearing Corporation (“DTCC”) (April 2011) and Organization for Economic Co-operation and Development (“OECD”) (December 2010) data, a June 2011 joint briefing paper put out by the Association for Capital Markets in Europe, ISLA, ISDA and ASSOSIM, estimated that net CDSs on European single name sovereigns amounted to between a low of 0.6% of government debt in the case of the UK and a high of 3.9% for Hungary.45 For Greece, the percentage was 1.2%, and an examination of DTCC data led these institutions to believe there had been relatively little change in net outstanding positions over the course of 2010. The reader was clearly meant to conclude that CDSs did not represent a serious multiplier risk for a sovereign default or restructuring shock. But was that reassuring conclusion a solid one? The absence of an increase in net outstandings would make sense in view of the increasing costs of protection on Greece and in view of the structure of the market where many transactions are back-to-back hedges. IOSCO’s June 2012 Report on the CDS market confirms that despite the overall growth in the sovereign CDS market as a percentage of the total CDS market from 15 to 25% between 2008 and 2011, the size of the CDS market for Eurozone sovereigns has remained stable. Furthermore, while the ratio of CDS gross notional to public debt has increased for some peripheral Eurozone sovereigns, the net notional amount for Ireland, Portugal and Greece has actually declined over the period.46 While IOSCO’s analysis would not have been available to the Greek decision makers, they would have had access to a September 2009 analysis of CDSs and financial stability in the Financial Stability Review of the Banque de France.47

45 AFME, ISLA, ISDA, ASSOSIM, Briefing on Sovereign CDS, 3 June 2011. 46 IOSCO (2012). 47 Duquerroy, Bex, Gauthier (2009).



The Aggregation Clause in Euro Area Government Securities 

 137

The authors’ first observation was not entirely reassuring: there is disagreement about the overall size of the market, with ISDA estimating it at $38.6 trillion, while DTCC, which believed it covered 95% of all transactions (in number of contracts), estimated it at $29 trillion. According to the authors, the IMF believed DTCC actually only covered 75% of transactions because DTCC only started requiring reporting in 2008, so older transactions were not recorded, nor were one-off “bespoke” trades which are not confirmed electronically. The Bank for International Settlements (“BIS”) estimate was $41 trillion. An often asked but never definitively answered question is how many over the counter bilateral trades go unreported to these organizations? Given how much use of ISDA documentation simplifies documenting transactions and given that more than 2,000 market participants (including banks, hedge funds and institutional investors) are parties to ISDA master agreements, it seems reasonable to think that number should not be too high, but it has been a point of uncertainty. The next question is what to make of these numbers, which are gross notional amounts, many of which represent offsetting contracts whereby one party having just written a contract covers its exposure thereunder in whole or in part by buying another contract? DTCC can also provide net notional positions per reference entity and the BIS gross market value, but all these estimates suffer from the potential for underreporting. Still, the Banque de France authors thought they could estimate net overall exposure, i.e. the maximum amount payable by protection sellers, at $25 trillion, or 9.5% of the gross notional amount. How robust was the estimate? A comparison with the experience of the Lehman Brothers default appears instructive. Initial media estimates of gross insurance claims had been $400 billion (compared to $150 billion of Lehman bond debt48). ISDA’s preliminary estimate was $7 billion. Based on DTCC’s records, the authors calculated the funds transferred from net protection sellers to net protection buyers at $5.2 billion, or 7% of the notional amount. Was this a fluke? According to the authors, DTCC’s data shows that the ratio of gross notional CDS amounts to net funds transferred has rarely topped 10% in the case of corporate reference entities. A preliminary conclusion from this analysis would seem to be that fears of CDS as a multiplier of a sovereign default shock has been overdone. There are two caveats to this conclusion. The first is that the historical data on payments

48 There is an important distinction to note here between the markets for corporate debt, in which gross notional amounts of protection written on any single name can occasionally be a multiple of the reference entity’s debt outstanding and that for sovereign debt where protection is typically only a fraction of outstanding debt. Net notional amounts are a fraction of overall debt outstanding for all classes of issuers. IOSCO (2012).

138 

 Patrick S. Kenadjian

upon default until the Greek experience related only to corporate rather than sovereign defaults. The second is that the authors also noted a very great concentration in writers of CDS protection. According to their data, in 2009 the 10 largest dealers accounted for 90% of trading volume by gross notional amounts, compared to 74% in 2004. With sovereign risk denominated in Euros carrying a zero risk weight for Eurozone banks under the Capital Requirements Directive, there was a possibility that concentration of exposure could be great among Eurozone banks. Somewhat paradoxically, if that concentration had been greatest among Greek banks, that would have been relatively good news since, as noted above, those banks had to be recapitalized anyway following a restructuring of GGBs and the EU rescue plan provided funds for this purpose. The issue would be more complex if the protection writers were German or French banks. Still, statistics from the second quarter of 2011 seemed to indicate that net notional outstandings on Greek debt were a relatively modest $5.5 billion, plus a further $1 billion from the Greek component of sovereign CDS indices, amounting to just 1.9% of cash bonds, compared to $78 billion in gross notional outstandings, equivalent to 26% of bonds outstanding. More recent reports, in mid-February 2012, cited contracts covering a net $3.2 billion outstanding at February 10, 2012, according to DTCC. The Greek government could thus have felt justified in concluding that if ever there was a situation in which to test the proposition that triggering CDSs written on a sovereign reference entity should not lead to catastrophic contagion, then Greece presented the most benign set of circumstances. The net notional amount was low, the event had been widely foreseen so there was none of the surprise element that surrounded the Lehman default and the triggering would be voluntary, in that Greece could choose after seeing the outcome of the exchanges whether to use the CACs or not. The absence of a Lehman like “jump to default” was a particularly important feature, since protection sellers post-Lehman and AIG, with the exception of central banks and sovereigns, are required to post collateral, so they are unlikely to suffer a severe shock where a default is as widely priced into the CDS for an Issuer as it was for Greece. Thus if the €3.2 billion net notional number was correct, at a 53.5% haircut, net cash transfer could be expected to be on the order of €1.71 billion.

iv. How it turned out In the end Greece did decide to use the “retrofitted CACs” on its domestic law bonds. Their use triggered a credit event under its CDSs and the world did not end. Of the €64 billion notional outstanding, a net €3.2 billion was determined by ISDA to be eligible for compensation. An auction was held on March 19, 2012



The Aggregation Clause in Euro Area Government Securities 

 139

under the auspices of ISDA to determine what payments were due the holders of the 4,323 contracts involved. The final payment was determined to be 78.5% of the net amount which came close to the estimate of the haircut the bondholders had suffered. At the end of the day, net payments of $2.89  billion were made. Since 70% of these positions were collateralized, meaning the protection sellers had posted cash or liquid securities as collateral for them and the average level of collateralization was 90% according to ISDA (cited in IOSCO 2012), the impact in terms of liquidity risk to the sellers of protection was limited.49 So, from a systemic point of view for Greece, the CDS issue turned out to be a non-event. IOSCO concluded in its report that “the impact of the credit event was remarkably low” with, in particular, no discernable impact on the CDS spreads of banks which had written protection as compared to banks which had bought protection. In fact, IOSCO found that “the exposure to Greece has been smaller than that of the Lehman Brothers default in September 2008. Moreover, the exposure to Greece was lower (and probably more collateralized) than that on Lehman Brothers and the recovery rate was higher…”50 So, protection buyers and markets had learned from the Lehman default to protect themselves more effectively. In this they were of course helped by the lack of a “jump to default” in the Greek case as there had been in the Lehman case. For the protection holders, there were a number of surprises. Since Greece closed its Greek law exchange offers in early March to avoid the March 20 due date on the €14.5 billion of debt that would otherwise have come due if those bonds were still outstanding, the bonds deliverable in the auction were not the old bonds, but the new bonds issued in the exchange. And only the new GGBs, not the EFSF bonds, were deliverable, since the EFSF is a different issuer. If the market had been fully convinced by the restructuring that Greece’s new debt was sustainable and the new bonds had traded closer to their new par, the result could have been less favorable for the protection on holders. Still, by luck or skill, the payment corresponded more or less to the loss the investors held actually suffered. Thus, a lasting legacy of the Greek exchange offers of 2012 may well be to have dissipated the fear of the unknown in triggering CDSs which originated with the Lehman failure. This is not to say that the consequences of such a triggering will always be benign. It is important to note that Greece is not necessarily representative of other distressed sovereigns in these respects. While the facts as we thought we knew them turned out to be accurate in the case of Greece, and there were no

49 IOSCO (2012), at 17. 50 IOSCO (2012), 17 and 18.

140 

 Patrick S. Kenadjian

serious consequences, net notional CDSs outstanding on Spanish and Italian sovereign debt are believed to be three to five times or four to six times as high as the net notionals on Greek sovereign debt.51 Figures published by DTCC in early June 2012 indicate net CDS outstanding for Spain close to €15 billion and for Italy in excess of €20 billion. Thus in future cases we may still need to weigh carefully how great the advantages of using CACs over traditional exchange offers are.

d. Other Factors to be considered: Hold-outs and Litigation It is true that exchange offers, even highly successful ones, always leave some old bonds outstanding. Whether this is a problem depends on the goal of the restructuring. If the sovereign’s main aim is to reduce outstanding payment obligations, getting in excess of 90% of the old bonds exchanged, as in the case of Uruguay (93%), Ukraine (97%), Russia (98%) or Pakistan (99%), or even 75%, as in the case of Argentina, may be enough.52 There is a clear trade-off between the size of the haircut and the success of the exchange in which the credibility of the numbers the sovereign can put on the table as to sustainability of the debt remaining outstanding after the transaction and the litigation alternatives available to potential hold-outs play important roles. There may be some situations in which the sovereign also needs to rid itself of an inconvenient covenant, such as a negative pledge or a cross default, in some outstanding bonds. In that case, the use of CACs with aggregation clauses may be tempting, although to the extent the covenant is a non payment covenant whose amendment would not trigger a Restructuring Credit Event, it could also be amended in an exit consent53, assuming either the bonds allowed non-payment terms to be amended by majority consent or were governed by the sovereign’s own law and could thus be “retrofitted” with a majority amendment clause or even unilaterally altered by act of the sovereign’s parliament. Relatively little attention has been paid to litigation alternatives in the context of the Eurozone sovereign debt crisis. A few news stories surfaced in early 2012 that hedge fund investors were looking into their ability to make claims concerning infringement of their property rights under the European Convention on Human Rights (the “Convention”) to the European Court of Human

51 Roubini (2011), Allen & Overy (2011). 52 Percentages are from IMF 2012A, at 24–25. 53 Subject to the limitations on the use of exist consents for English law bonds arising out of the Anglo Irish Bank case discussed below.



The Aggregation Clause in Euro Area Government Securities 

 141

Rights (“ECHR”), should Greece consider the use of retrofitted CACs, but these disappeared quickly. This may have been because the ECHR is a court of appeal, whose jurisdiction depends on the claimants having exhausted their appeals in the courts of a Member State. So litigation would have to start there and appeals would have to be exhausted there before the investors found their way to the ECHR. The “there” would be the Member State whose law governed the contract or transaction involved. For GLGBs that would be Greece. Foreign creditors generally consider litigation before the domestic courts of an emerging markets sovereign as a losing proposition. Whether this would apply equally in the case of a “peripheral” Eurozone sovereign is a case of first impression, but to the extent those courts are not known for their speed or commercial sense, litigating before them is probably not the preferred strategy. Whatever the arguments which could be advanced under Protocol 1 Article 1 of the Convention that an action by the Greek state has deprived an investor of its property, having to commence litigation in a Greek court in the context of the severe austerity measures being imposed on Greek citizens does not sound like a swift or winning strategy, in comparison to the international litigation strategies investors have been able to deploy against Latin American sovereigns in the past, which have been conducted largely in the courts of New York, England or Belgium. But even there the number of restructurings which have been subject to extensive litigation has been very small, Argentina and Dominica accounting for most of the recent cases.54 Another alternative for litigation is explored in Boris Kasolowsky’s excellent article in this volume on bilateral investment treaties. As Boris points out, there is a basis for using these treaties, initially intended to protect direct investment, to protect bondholders. However, as he also points out, these treaties are not uniform in their contents, with the older treaties providing less protection to bondholders and their coverage is less than universal. Greece, for example, has not signed a treaty with any of the United States, the United Kingdom, France, Italy or Spain, which leaves investors from those jurisdictions with no recourse under such treaties. Greece does have a treaty with Germany and nine other Member States, but the major jurisdictions for aggressive distressed debt investors appear to be cut off from this avenue of litigation. Also, the German treaty is a very early one and many of the other treaties with Member States are with Central and Eastern European states where Greece expected to be the investor other than the object of investment. Finally, to be able to use the treaty, the investment would have to be made by a person in the treaty country, or at least transferred to such a

54 IMF (2012A), at 46.

142 

 Patrick S. Kenadjian

person before the action which gave rise to the treaty violation was taken by the sovereign, so that the holder qualifies as an investor under the treaty. Consequently, the litigation avenue does not seem to provide a strong back up strategy for a potential hold-out, especially for bonds governed by Greek law. In the February 2012 exchange offers, coincidently or not, the greatest number of hold-outs were among the holders of foreign law governed bonds under which they had recourse to English rather than Greek courts. But since the retrofitted CACs did not apply to the English law bonds, this access did not provide bondholders with an avenue for challenging the retrofit.

e. The Rule of Law A final question concerns the rule of law. I have been using the term “CAC like device” to describe what Greece actually used to amend the terms of its GLGBs. Those bonds lacked the CAC clauses its FLGBs included, so that they needed to be “retrofitted” with something similar, which some experts have referred to as a “collective action mechanism”. The Greek Bondholder Act which did this was passed long after the bonds had been issued and enabled the Greek government to amend their terms unilaterally if a certain percentage of bondholders had tendered their bonds and consented to the changes. The EU CACs initiative applies only prospectively, to sovereign bonds issued on or after January 1, 2013. The initiative applies only to sovereign bonds, not to agency or regional government bonds which, in many EU countries, such as Spain and Italy, constitute a substantial percentage of public sector debt. If one takes as a rule of thumb that, on average, EU sovereigns refinance about 10% of their outstanding debt annually, it will be a number of years before newly issued debt containing CACs will constitute a substantial portion of Eurozone member state sovereign debt. Should any other sovereign want or need to restructure in the meantime, they will have to weigh the pros and cons of having recourse to a change in law with respect to their domestic law governed bonds similar to the Greek Bondholder Act. In doing so they will have to consider whether allowing a retroactive change in contracts they, their agencies and their regions have entered into, is consistent with commonly accepted notions of the rule of law. This question has constitutional, moral and commercial facets. The constitutional aspect may be the easiest to resolve: does Parliament have the right to allow public sector entities to alter their contracts retroactively in circumstances of dire public need and where a large majority of the counterparties are willing to agree to the change? I suspect the answer will most often be yes.



The Aggregation Clause in Euro Area Government Securities 

 143

The commercial aspect is also quite clear: the ability of overindebted sovereigns to borrow under their domestic law may be curtailed or made more expensive. I am told that in the wake of the Greek “retrofit”, the markets are already starting to price domestic and foreign law bonds differently. Markets are forgetful and this may not last, but for a while at least there may be a rebirth in the use of the English law for sovereign bonds.55 Also, as Christian Kopf points out in his excellent article in this volume, the sovereign’s overall borrowing costs may increase as a result if creditors feel they have been mistreated by the sovereign. The moral or ethical aspect may be the most difficult to resolve: if a sovereign can use exchange offers which do not interfere with the principle that the sover-

55 Or at least this seemed to be the case until July 27, 2012 when the Chancery Division of the English High Court upheld a challenge to the legality under English law of an exit consent in a sovereign related restructuring on the basis that it was oppressive towards the minority bondholders and was caught by the prohibition against issuers voting bonds beneficially held by them or for their account. Assénagon Asset Management S.A. v. Irish Bank Resolution Corporation Limited (formerly Anglo Irish Bank Corporation Limited) [2012] EWHC 2090(Ch). In the Anglo Irish case, the exit consent allowed the issuer to redeem any unexchanged debt for €0.01 per $1,000 principal amount. The judge held that it was not lawful for the majority to aid in the coercion of a minority by voting for a resolution which expropriates the majority’s rights for nominal consideration. The decision thus casts doubt on the legality under English law of any form of exit consent that imposes less favorable conditions on those who refuse to participate in the associated exchange offer. This part of the decision would not affect use of a CAC whose objective is to bind all holders to the same terms as the majority have agreed to for themselves, but it could seriously constrain future use of exit consents under English law to strip away covenant and other protections from non-exchanging bondholders, thus calling into question whether sovereigns will want to choose English law to govern their bonds. The other part of the decision held that the issuer had a beneficial interest in the votes at the time of the bondholder meeting which was held after the results of the exchange offer had been announced but before settlement because the tendered notes were subject to a specifically enforceable contract for sale by exchange and thus fell afoul of the prohibition in the indenture of the issuer voting notes in which it had a beneficial interest. Thus in any exchange offer where the consents must be voted at a bondholder meeting to be held after the exchange offer has been concluded, as is usual in England, will result in an invalidation of the votes of the bondholders. A different result should be reached in the context of a US style consent solicitation, where receipt of the required percentage of consents is usually a condition to acceptance of the tenders so that at the time of the vote the issuer has not yet accepted the tenders and thus does not have an enforceable interest in the notes. In conclusion, the Greek exchange offers, had they taken place under English law, would not have been vulnerable on substantive grounds, since they sought merely to bind non-exchanging holders to the conditions the majority agreed to rather than to oppress them, but might well have been invalidated on procedural grounds, to the extent the Greek state could be held to have had an interest in the exchanged bonds at the time of the bondholder vote.

144 

 Patrick S. Kenadjian

eign is as bound by its agreements as private parties are (pacta sunt servanda), what justifies the sovereign having recourse to the additional lever of a collective action mechanism to allow it to breach its word? For an advanced Western democracy, the question is not a trivial one.

6. Conclusion The Greek exchange offers of February 2012 can be divided into three parts: the offers to FLGBs where no amendment was attempted, mainly due to local securities law considerations; the offers for FLGBs which contained conventional CACs where amendments were proposed; and the GLGBs where amendments were proposed using “retrofitted CACs” with aggregation clauses. The participations rates were 31%, 78% and 100%, respectively. At first glance the lesson seems clear: having CACs is better than not having CACs and having CACs with aggregation clauses is best of all. However, in the absence of “retrofitting” legislation, CACs with aggregation clauses will come into being only slowly, as sovereign issuers replace existing debt without CACs with new debt containing them. That is occurring at the rate of approximately 10% a year on average in the Eurozone. Thus, if a sovereign issuer is faced with the need to restructure its debt in the next few years, it may need to have recourse to a “retrofit” as Greece did. That is clearly possible, but its effectiveness will depend on a number of considerations. The first is the percentage of its government debt governed by domestic law. The recent Nomura Securities study referred to above shows that on average foreign law governed bonds represent 7% of Eurozone sovereign debt overall. At 13%, Spain has among the highest percentages of foreign law governed sovereign bonds. Italy has only 6%. Nordvig and Firoozye (2012). Christian Kopf’s article, citing a Moody’s Weekly Credit Outlook study, cites lower figures for foreign law, 2.5% for Italy and 1.1% for Spain. The second is the extent to which the uncertainty as to the effects of triggering a Credit Event under CDSs referencing the sovereign’s debt has been dissipated in the minds of the market and the public sector actors. I think it largely has been. In addition, the Regulation on short selling and credit default swaps adopted by the Council on February 21, 2012 contains provisions requiring that significant positions in CDSs relating to EU sovereign debt issuers be notified to regulators and this should provide for increased transparency in these instruments. The third will be a balancing of costs. Prime among the circumstances which motivated Greece to retrofit its existing GGBs with CAC-like clauses and



The Aggregation Clause in Euro Area Government Securities 

 145

then use the retrofitted clauses was the need to avoid a massive debt payment of €14.5 billion due less than a month after commencement of their exchange offers. Had this debt not been due, the calculus between the cost of triggering a Credit Event under its sovereign debt CDSs and the savings from increasing the degree of acceptance of the offers might have been different. The use of the CAC-like clauses occasioned not inconsequential costs, in the form of additional Eurozone guarantees to allow its banks to continue to borrow against GLGBs pledged as collateral to the ECB and NCBs, during the period of time those bonds were rated SD as a result of the use of the CAC-like clauses. For governments who have extensively guaranteed debt issued by their domestic banks to allow this debt to be pledged to the ECB as collateral for the LTRO programs and other ECB and NCB extensions of credit, the cost could be higher than it was for Greece. The fourth will be whether the next sovereign debtor to restructure will be under comparable pressure to accomplish it as a “voluntary” restructuring. The fear of triggering a CDS default should be much lower. However, depending on how suddenly the crisis comes and what the public sector’s response to it is, the exposure of core Eurozone banks to the next sovereign’s debt could be greater, with greater knock-on effects on their solvency and that of their governments if they are called to recapitalize them. In the case of Greece the private sector banks had had almost two years to reduce their GGB exposure at a time when the ECB and NCBs were buying that debt to help stabilize Greece’s funding costs. This policy was not uncontroversial and may or may not be continued or replaced. The fact that Greece’s restructuring had been “telegraphed” so far in advance meant there was not the sharp price discontinuity known as the “jump to default” which can cause large jumps in the exposure of protection sellers upon default by the sovereign, such as occurred in the Lehman case. The fifth will be how inextricably linked domestic and foreign investors are in the debt to be restructured and whether the sovereign is content to treat all its creditors equally by using a CAC to bind them all to the same deal or menu of alternatives, or whether it may need to treat some domestic or international constituencies differently. While Greece has shown this can be done for a small number of creditors, the next sovereign may conclude that a series of exchange offers coupled with exit consents (subject to the caveats noted above about the use of exit consents under English Law after the Anglo Irish Bank case) might be a more flexible tool than a CAC retrofit.56

56 As this article was going to press, developments in litigation before the Federal courts in New York relating to the latest Argentine debt rescheduling were also casting doubt on the continued effectiveness of the traditional exchange offer as a viable instrument for sovereign debt

146 

 Patrick S. Kenadjian

The final consideration will be whether the use of a parliamentary “retrofit” is viewed as an acceptable action by the sovereign in question, or as one which contravenes the commonly accepted notion of the rule of law or the sovereign’s self image as a leading member of the European community of states, and actor on the world stage. Thus while only time will tell, at the date of this writing, I think it is too early to see aggregation clauses in CACs as either a game changer or the flavor of the month. I would rather be inclined to answer, as Zhou En Lai is alleged to have said when asked what he thought the historical impact of the French Revolution was, “it’s too early to say” or, more prosaically, to agree with Alessandro Leipold, and see them as helping at the margin, but not, in themselves, decisive in debt restructurings, at least within the Eurozone.

Bibliography AFME, ISLA, ISDA, ASSOSIM, Briefing on sovereign CDS, 3 June 2011. Allen & Overy, Sovereign state restructurings and credit default swaps, October 2011. Baldwin, Richard, Daniel Gross and Luc Laeven, 2010, Completing the Eurozone Rescue: What More Needs to be Done? June 2010. Bi, Ran, Marcos Chamon and Jeromin Zettelmeyer, 2010, The Problem that Wasn’t: Co-ordination Failures in Sovereign Debt Restructurings, June 2010. Bradley, Michael and Mitu G. Gulati, Collective Action Clauses for the Eurozone: An Empirical Analysis (2012). Bucheit, Lee and Mitu Gulati, 2011, Drafting a Model Collective Action Clause for Eurozone Sovereign Bonds, 2011. Citigroup Global Markets (Citi), Stealth subordination, Why ESM seniority will be most effective when subtle, 25 June 2012. Council of the European Union, 2012, Regulation adopted on short selling and credit default swaps, Press Release 6625/12, February 21, 2012. Council of the European Union, Statement by the Heads of State or Governments of the Euro Area and EU Institutions, 21 July 2011.

rescheduling under New York law. In NML Capital, Ltd. et. al., v. The Republic of Argentina et al, Docket No. 12-105(L), United States Court of Appeals for the Second Circuit (2013), hold-out Argentine bondholders were claiming the right to receive payment in full on their unexchanged bonds should Argentina make any payments on the restructured bonds, and both the trial court and the appeals court initially sided with the hold-outs. Should these initial orders become final, they could seriously complicate successful completion of exchange offers under New York law, since issuers could no longer credibly threaten that hold-outs would receive no further payments on their unexchanged bonds.



The Aggregation Clause in Euro Area Government Securities 

 147

Das, Udaibir, Michael G. Papaioannou and Christopher Trebesch, Sovereign Debt Restructurings 1950–2010: Literature Survey, Data and Stylized Facts, IMF Working Paper WP12/203, August 2012. Dey, Joy, 2009, Collective Action Clauses, Sovereign Bondholders Cornered? June 2009. Dhillon, Amrita, Javier Garcia-Fronti, Sayanteen Ghosal, Marcus Miller, 2006, Debt Restructuring and Economic Recovery: Analyzing the Argentine Swap, January 2006 (Dhillon, 2006). Duquerroy, Anne, Mathieu Gex, Nicolas Gauthier, 2009, Credit default swaps and financial stability: risks and regulatory issues, Banque de France Financial Stability Review No. 13, September 2009. Economic and Financial Committee of the European Union, Sub-Committee on EU Sovereign Debt Markets (EFC Sub-Committee), 2011, Invitation to participate in the consultation of market participants and other stakeholders on Collective Action Clauses to be included in euro area sovereign securities, 23 July 2011. Economic Sub-Committee, 2012, Model Collective Action Clause, Common Terms of Reference (17/02/2012) and Supplemental Explanatory Note 26 March 2012). Economic and Financial Committee of the European Union (ECFIN), 2004. Implementation of the EU Commitment on Collective Action Clauses in documentation of International Debt Issuance, 12 November 2004. Escolano, Julio, A Practical Guide to Public Debt Dynamics, Fiscal Sustainability and Cyclical Adjustment of Budgetary Aggregates, IMF, January 2010. European Council, 2011, 24/25 March 2011 Conclusions. European Council, Statement by the Euro Area Heads of State or Government, 9 December 2011 (European Council 2011A) Gelpern, Anna, Domestic Bonds, Credit Derivatives, and the Next Transformation of Sovereign Debt, Rutgers School of Law Newark (2008). Gelpern, Anna and Brad Setser, Domestic and External Debt: the Doomed Quest for Equal Treatment, Georgetown Journal of International Law, Summer 2004. Gianviti, Francois, Anne K. Krueger, Jean Pisani-Ferry, André Sapir, Jürgen von Hagen, 2010, A European Mechanism for Sovereign Debt Crisis Resolution: A Proposal, Bruegel, 9 November 2010. Gros, Daniel and Thomas Mayer, 2011, Debt reduction without default? CEPS Policy Brief No. 233, February 2011. Gros, Daniel and Cinzia Alcidi, 2011, Adjustment Difficulties and Debt Overhangs in the Eurozone Periphery, CEPS Working Document No. 347, May 2011. Group of Ten, 2002, Report of the G-10 Working Group on contractual clauses, 26 September 2002. International Monetary Fund (IMF), A Survey of Experiences with Emerging Market Sovereign Debt Restructurings, June 5, 2012 (IMF 2012A) IMF Greece: Preliminary Debt Sustainability Analysis, February 15, 2012 (IMF 2012). International Swaps and Derivatives Association (ISDA) 1999 ISDA Credit Derivatives Definitions (1999). IOSCO Committee on Risk and Research Subgroup, The Credit Default Swap Market Report, June 2012. ISDA, 2003 ISDA Credit Derivatives Definitions 2003. Kopf, Christian, 2011, Restoring financial stability in the euro area, CEPS Policy Brief No. 237, March 2011.

148 

 Patrick S. Kenadjian

Krueger, Anne A. 2002, A New Approach to Sovereign Debt Restructuring, 2002. Lascelles, Eric, 2011, The Slippery Slope from Sovereign Debt to Default, RBC Global Asset Management, May 2011. Leipold Alessandro, 2011, Thinking the Unthinkable, Lessons from Past Sovereign Debt Restructurings, The Lisbon Council e-brief, Issue 11/2011. Nordvig, Jens and Dr. Nick Firoozye, Rethinking the European Monetary Union, Nomura Securities, 2012. Richmond, Christine and Daniel A Dias, 2009, Duration of Capital Market Exclusion: An Empirical Examination, Draft July 2009. Roubini, Nouriel and David Nowakowski, 2011, CDS and Debt Restructuring: Does the Existence of Credit Derivatives Make Restructuring Harder? Roubini Global Economics, April 21, 2011. Sturzenegger, Federico and Jeromin Zettelmeyer, 2007, Has the Legal Threat to Sovereign Debt Restructuring Become Real? The Hellenic Republic, Invitation Memorandum dated 24 February 2012. Trust Indenture Act of 1939, 12 USC Ch 2A. Verdier, Pierre-Hugues, 2004, Credit Derivatives and the Sovereign Debt Restructuring Process, Harvard Law School, April 27, 2004. Whitehall, Christopher, International Financing Review (IFR), 2012, Hopes for Greek trigger to validate CDS, 9 February 2012 and Market prepares for Greek CDS trigger, 25 February 2012. Yilmaz, Kamil, 2011, Major European bank stocks are connected in a state of high volatility, Vox, 25 September 2011. Zettelmeyer, Jeromin, Christoph Trebesch, and Mitu Gulati, The Greek Debt Exchange: An Analysis in Historical Perspective (draft: 21 June 2012), 2012.

Christian Kopf

Sovereign Debt Restructuring: Lessons from History1 Table of Contents

Introduction: A Night in Brussels 1. Sovereign Debt Restructurings in Historical Perspective 2. Some Characteristics of Recent Sovereign Bond Exchanges 3. Desiderata for Future Debt Renegotiations 4. The Role of Collective Action Clauses in Euro Area Sovereign Debt Restructurings Conclusion

Introduction: A Night in Brussels In the wee morning hours of October 27, 2011, German Chancellor Merkel and French President Sarkozy left the meeting of the Heads of State and Government of the European Union in Brussels and retired into a small conference room. They were joined by the president of the European Commission, the Managing Director of the International Monetary Fund (IMF) and by a former French civil servant who had been retained as representative of a leading bank lobbying association, the Institute of International Finance (IIF). No members of the Greek government were present when Nicolas Sarkozy started off the meeting with a tirade against bankers and speculators and a reminder of the role of policymakers, who would be judged by history for their contribution in overcoming the European sovereign debt crisis. Staff members were whispering a German translation to Chancellor Merkel and had trouble keeping up with the president’s speed. Sarkozy concluded by expressing his gratitude for the presence of “chère Angela”, who would now take care of the “technical details”. Then Chancellor Merkel switched into English and dictated the terms of the planned Greek sovereign debt restructuring to the

1 The author would like to thank, without implication, Patrick Kenadjian, Ugo Panizza, Christoph Trebesch and Jeromin Zettelmeyer, and participants in conferences at the Institute for Law and Finance in Frankfurt am Main and at the European Bank for Reconstruction and Development in London for helpful comments, corrections and suggestions. The author retains responsibility for any remaining errors. The views expressed are attributable only to the author in a personal capacity. This article reflects developments as of August 2012.

150 

 Christian Kopf

IIF representative: a 50 per cent nominal haircut, interest rates as low as 2 per cent and an extension of maturities of up to 30 years. She made it clear that these terms were not negotiable but offered € 10 billion in EFSF collateral in order to make the deal agreeable to bondholders. The IIF representative immediately understood that this would mean a loss in excess of 85 per cent in net present value (NPV) terms for bondholders, a far cry from the 20 per cent NPV loss which the previous debt restructuring plan had targeted. It was now clear to everyone in the room that the proposal of a Greek sovereign debt restructuring with no nominal haircut and full EFSF collateralisation, which had been drafted by the French Banking Association in consultation with the French government, and which had been approved by the European Council in July 2011, was dead. However, the IIF representative was not ready to give up the fight: politely but firmly, he replied that € 10 billion in collateral would be insufficient to allow for a “voluntary” debt exchange – the German Chancellor would have to put more on the table. After some back-and-forth, Chancellor Merkel made her final offer: for every € 100 in original face amount, bondholders would receive € 15 in EFSF notes and € 35 in new Greek government bonds. In order to facilitate the restructuring of the stock of € 206 billion in privately-held Greek government bonds, euro area member states would thus provide up to € 32 billion in EFSF collateral. The IIF accepted, and in less than two hours, the terms of the largest sovereign debt restructuring in history were sealed.2 Much was gained in these two hours. Namely, after a long period of denial about the country’s debt overhang, Greece was given a limited but realistic chance to bring its public finances in order. But a lot was lost as well. The European Union, which prides itself as a champion of democracy, property rights and the rule of law, had allowed an expropriation in excess of € 100 billion, without giving either party in the respective contracts, namely the Hellenic Republic as debtor or private investors as creditors, an appropriate say in the process. Furthermore, the governments of twelve EFSF guarantors were forced to sign on to an agreement on a € 32 billion cash sweetener that was negotiated without their participation by two other EFSF guarantors, namely Germany and France. Finally,

2 The final terms of the February 2012 exchange offer differed slightly from the October 2011 agreement. Holders were invited to tender € 1,000 face amount of eligible Greek government bonds in exchange for a consideration consisting of € 315 aggregate face amount of 20 individual series of new Greek government bonds, € 315 notional amount of GDP warrants, € 75 face amount of one-year EFSF bonds, and € 75 face amount of two-year EFSF bonds, and for a variable face amount of six-month EFSF bills as accrued interest payments (Hellenic Republic 2012).



Sovereign Debt Restructuring: Lessons from History 

 151

in abrogation of the principle of preference avoidance3, Greek government bonds held by the European Central Bank, by the National Central Banks of the Eurosystem and by the European Investment Bank were exempt from the sovereign debt restructuring negotiated by Chancellor Merkel and President Sarkozy. This raises a pressing question: What can be done to allow for more orderly sovereign debt restructurings in Europe, if they were to become necessary again, and what is the role of Collective Action Clauses (CACs) in this process? The remainder of this paper is organised as follows. Section 1 asks whether conditions are in place that may lead to sovereign debt restructurings in Europe. Section 2 discusses some characteristics of recent sovereign debt restructurings, and section 3 elaborates on desiderata for future restructuring exercises. Section 4 then asks to what extent collective action clauses could facilitate future sovereign debt restructurings in Europe. The main findings of the paper are summarised in a concluding section.

1. Sovereign Debt Restructurings in Historical Perspective The debate about the European debt crisis, which started in earnest with Greece’s loss of market access in the spring of 2010, has been marked by strongly divergent views. On one end of the spectrum, it has been argued that restructuring the

3 According to the principle of preference avoidance, similarly situated creditors must receive an equal distribution of the debtor’s assets in liquidation, or an equal share of new claims in a debt restructuring. This principle has been established by English courts since the late 16th century: “There ought to be an equal distribution secundum quantitatem debitorum suorum; but if, after the debtor becomes a bankrupt, he may prefer one (who peradventure hath least need), and defeat and defraud many other poor men of their true debts, it would be unequal and unconscionable, and a great defect in the law, if, after that he hath utterly discredited himself by becoming a bankrupt, the law should credit him to make distribution of his goods to whom he pleased.” (Coke 1826, 505). It became a core element of UK and US bankruptcy legislation and it has also been established in other jurisdictions, e.g. in § 226 of the German Insolvency Statute. Under modern legislation, the principle of preference avoidance does not preclude discrimination between creditor groups per se. However, it establishes that all parties of the same creditor class shall be offered equal treatment, and that no undue preference shall be accorded to specific creditor classes. This principle was clearly violated in the Greek sovereign debt restructuring, when ECB holdings of Greek government bonds were assigned new ISIN codes and then exempted from the debt exchange.

152 

 Christian Kopf

sovereign debt of euro area member states is “unnecessary and undesirable”4 – a view that was supported by the fact that there hadn’t been a sovereign debt restructuring in an advanced economy since Germany negotiated a write-off of about half of its external public debt in 1953.5 On the other end of the spectrum, calls have been abundant to deal with public debt crises in the European Monetary Union using the tried and tested procedures of Emerging Markets sovereign debt workouts, notably the Brady Plan loan restructuring agreements that were negotiated in the 1990s6. This paper contends that both propositions can be rejected: membership in the European Monetary Union increases rather than eliminates the potential need for sovereign debt restructuring, but the Brady deal negotiation process between the London Club and the respective sovereign debtors offers a poor template for debt workouts in Europe. The view that sovereign credit risk is exclusive to Emerging Market economies typically ascribes the cause of sovereign debt crises to a weak debt servicing history7, weak political institutions8 or weak legal institutions9. Based on this narrative, it could be argued that the advanced economies of the European Union should not be affected by sovereign credit risk – and indeed, this has been the consensus among investors in European government bond markets over several decades. However, the main determinant of sovereign default risk is not a country’s history or its institutions, but its ability to issue debt in a currency the government can control10. It is virtually impossible for market forces to drive a government into default if it operates under a floating exchange rate regime and has issued its debt in domestic currency. The government could still face a buyers’ strike in which investors suddenly decide to redeem government bond holdings and take their money out of the country. But for every unit of domestic currency that is sold another unit of domestic currency would be bought, just at a different exchange rate. The domestic money stock would remain constant, and the exodus from the government bond market would lead to an increase in bank deposits (as long as nominal interest rates are reasonably high and there are no doubts about bank solvency). In the absence of safer alternatives, commercial banks would then lend to the government again. Even in the extreme case of a system-wide bank

4 Cottarelli et al. 2010. 5 Guinnan 2004. 6 Boone and Johnson 2010. 7 Reinhart et al. 2003. 8 Van Rijckegham and Weder 2009. 9 Kohlscheen 2007. 10 Kopf 2011.



Sovereign Debt Restructuring: Lessons from History 

 153

run, the government could direct the central bank to purchase government bonds and thereby prevent a sovereign default. By abandoning their domestic currencies, on the other hand, all member states of the European Monetary Union (EMU) have subjected themselves to sovereign default risk, regardless of the quality of their political and legal institutions and their payment record. The government can simply run out of cash if it loses market access as investors allocate the proceeds of maturing bonds to investments or bank accounts in other countries of the euro area. In the absence of market funding, the government won’t be able to borrow from the central bank either, as the European Central Bank (ECB) is prohibited from providing monetary financing of government expenditures under Article 123 of the Treaty on the Functioning of the European Union. By introducing the euro, EMU member states have thus turned themselves from “full sovereigns” into “subsidiary sovereigns”11 that have to borrow in a currency they no longer control and that have to rely on financial support to avoid defaults in the face of buyers’ strikes in their government bond markets.12 We can conclude that the re-introduction of sovereign default risk is a necessary consequence of the current institutional setting of the European Monetary Union. This institutional setting can be altered in a way that would render future debt restructurings unnecessary, e.g. by turning the ECB or another multilateral institution into an explicit lender of last resort to governments or by mutualising euro area public debt through joint issuance. But as long as the current economic and constitutional conditions prevail, it would appear desirable to move to an orderly workout process which reduces the market tensions and negative spill-over effects that have preceded Greece’s recent sovereign debt restructuring exercise. This raises the question whether the Brady deals of the 1990s between Emerging Markets borrowers and their lenders can offer a model for future sovereign debt restructurings in Europe. Over the past 200 years, international lending has occurred in waves of liquidity expansion and contraction, often driven by political changes or legal and financial innovations that led to self-reinforcing credit booms.13 As a result of market corrections that occurred with equal regularity, sovereign defaults have been bunched in temporal and sometimes regional clusters.14 The most important period of credit extension to Emerging Markets in post-war history has been

11 Goodhart 2011. 12 See also de Grauwe 2011 and Sims 2012. 13 Pettis 2001. 14 Sturzenegger and Zettelmeyer 2006, Reinhart and Rogoff 2009.

154 

 Christian Kopf

the “petrodollar recycling” of the 1970s, in which the rapid rise of global oil prices led OPEC member states to deposit large portions of their current account receipts with Western banks, which would in turn lend to African, Asian, Eastern European and Latin American sovereigns, generally in the form of floating rate dollardenominated bank loans. The strong rise in US interest rates in the early 1980s rendered the servicing of these loans unaffordable and led to a wave of sovereign defaults and subsequent restructurings of bank debt (see Figure 1). For the majority of countries involved, what followed was a lost decade of serial restructurings of bank loans with insufficient debt relief. This period was brought to an end by the Brady Plan, which allowed for a co-ordinated restructuring of bank loans into large and liquid sovereign bonds that were often collateralised by US Treasury zero-coupon notes. Between 1990 and 1998, a total of 22 Emerging Markets sovereigns exchanged syndicated bank loans and bilateral debt into tradable Brady bonds15. Soon, an active market for these new bonds developed, which allowed the borrowers to attract new investor interest and eventually to issue regular government bonds and retire the Brady debt. $ Billions of Original Claims 300 250 200

Greece

Brady Deals Bond Exchanges Bank Debt Restructurings

150 100

Argentina

50 0 1960

Peru 1965

1970

1975

1980

1985

1990

1995

2000

2005

2010

Figure 1: Sovereign debt restructurings with private creditors (Sources: Borensztein et al. 2006, Trebesch 2011, author’s calculations)

15 Brady bonds were issued by Mexico, Costa Rica, Morocco, Venezuela (in 1990), Uruguay, Nigeria (in 1991), the Philippines (in 1992), Argentina, Russia, Bolivia, Jordan (in 1993), Brazil, Bulgaria, the Dominican Republic, Poland (in 1994), Ecuador, Slovenia (in 1995), Panama, Croatia (in 1996), Peru, Vietnam (in 1997) and Côte d’Ivoire (in 1998). My classification differs slightly from Das et al. 2012, (p. 18), who have omitted the Brady deals of Morocco, Russia, Slovenia and Croatia.



Sovereign Debt Restructuring: Lessons from History 

 155

For the analysis of the process of sovereign debt restructuring, it is key to note that the bank debt restructurings of the 1980s happened during a period in which international lending was dominated by a small number of Western and Japanese banks which typically held long-term relationships with sovereign borrowers. Debt restructurings were negotiated by “Bank Advisory Committees”, which came to be known as the “London Club”. A small group of representatives from international banks would form an ad hoc committee to negotiate the terms of a debt restructuring and then try to gain universal approval for the agreed terms from all holders of the respective syndicated bank loans.16 The rise of the London Club culminated in the Brady deals of the 1990s, which themselves led to the demise of this institution. As the majority of the private sector claims against Emerging Markets were transformed from loans (which are generally held to maturity) into tradable instruments (which are subject to frequent mark-to-market), the role of commercial banks changed from being long-term creditors to becoming brokerdealers. Banks passed the ultimate sovereign default risk on to asset managers and started to generate revenues from trading debt instead of holding it. A new investor class of bondholders emerged and became dominated by insurance companies, investment funds and family offices. This change in the form of financial intermediation between sovereign borrowers and their creditors has not eliminated default risk or the need for debt restructuring mechanisms. Most governments that completed Brady bond deals during the 1990s have remained current on their debt until today, but six of these 22 countries did resort to another re-negotiation of their sovereign bonds (namely Argentina, Côte d’Ivoire, the Dominican Republic, Ecuador, Russia and Uruguay). They were joined by a small group of weak sovereigns that had not taken part in the original Brady plan. It is key to note that almost all sovereign debt workouts that took place since the early 2000s followed a process that differed sharply from the Brady deals of the previous decade: they took the form of debt reductions that “were not negotiated or even really discussed with bondholders but to a large extent were unilaterally imposed”.17 With the exception of the debt restructurings of Russia, Belize and Grenada, sovereign debtors and their legal advisors failed to recognize bondholder groups as interlocutors in the debt restructuring process.18 This is a direct consequence of the change in the distribution of sovereign creditors, from a small group of internationally active banks to a large group of bondholders. The

16 Sturzenegger and Zettelmeyer 2006, Das et al. 2012. 17 de Larosière 2005. 18 Das et al. 2012, p. 24–25.

156 

 Christian Kopf

Russian sovereign debt restructuring of 2000 was the “last dance” for the London Club.19 After the painful losses borne by institutions such as Deutsche Bank or Credit Suisse, commercial banks sped up the transfer of claims against risky sovereigns to asset managers. When Ecuador restructured its sovereign bonds in 2000, the government already decided not to recognize the negotiation committee that had been organized by bondholders, and opted for a “take-it-or-leave-it” strategy20 which forced creditors to accept an exchange into new securities. This became the norm in the following years. Changes in the sovereign debt renegotiation process are summarized in Table  1, which contrasts the two ideal types of lending relationships, namely bank intermediation versus market intermediation. Table 1: Ideal types of sovereign lending relationships Bank intermediation

Market intermediation

Claim holder

Banks

Asset managers

Claim distribution

Centralised

Atomised

Claim type

Bilateral or syndicated loans

Tradable bonds

Accounting treatment

Accrual accounting

Mark to market

Creditor representation

London Club

none

We can conclude that the institutional setting of sovereign debt restructuring procedures has been unable to keep up with the process of disintermediation between borrowers and lenders. The “Bank Advisory Committees” of the Brady deal area cannot serve as a model for future debt restructurings in Europe or elsewhere, because most international banks today are primarily active as broker-dealers and sovereign advisors, and no longer act as major holders of risky sovereign debt. Attempts by banks to resuscitate the London Club procedure of the Brady deal area, as in the early stages of the Greek sovereign debt renegotiations, lack legitimacy and are prone to conflicts of interest. The negotiations with the Greek government and its multilateral supporters were led by the CEO of Deutsche Bank, Josef Ackermann, although his institution had very little exposure to the sovereign. The credibility of Ackermann’s role as creditor representative was further eroded when Deutsche Bank was retained by Greece as “closing agent” for the debt exchange.

19 Since then, only a few minor bank loan restructurings took place, e.g. in Dominica and the Dominican Republic, each involving less than $ 200 million face value. 20 IMF 2003, p. 24.



Sovereign Debt Restructuring: Lessons from History 

 157

Within Deutsche Bank, work on the Greek debt restructuring was led by a group of investment bankers who had a clear objective of “getting a deal done”, in order to be able to charge fees to the Hellenic Republic – a major motivation that was absent in the London Club negotiations of the Brady Plan area.

2. Some Characteristics of Recent Sovereign Bond Exchanges We have established in the previous section that the need for sovereign debt restructurings may arise again in Europe, and that the London Club process does not offer a template for the re-negotiation of payment terms on government bonds, because it was modelled for the bygone era of government funding via syndicated bank loans. Loan renegotiations via Bank Advisory Committees have been replaced by sovereign bond exchanges. In order to gain guidance on a suitable design of sovereign debt restructuring processes, we need to take a closer look at recent examples of these bond exchanges. Over the past 14 years, 13 sovereigns have conducted distressed bond exchanges, in which they offered creditors to swap their holdings of existing government bonds into new securities with a reduced present value.21 Table 2 provides an overview of these transactions. The academic literature, research at multilateral institutions and policy initiatives have almost exclusively focused on one perceived flaw of this form of debt restructuring, namely the problem of holdouts.22 In the absence of a formal bankruptcy procedure, the restructuring of Emerging Markets government bonds had to rely on the contractual provisions of bond prospectuses. Most of the external public debt of Emerging Markets sovereigns is issued under State of New York law, and until 2003, bond covenants required unanimous approval by holders on changes to payment terms. This situation allowed for free-riding behaviour: creditors that did not consent to the proposed debt restructuring could insist on being paid in full, and rush to foreign courts in order to sue the sovereign borrower. In the famous Elliot vs. Peru case, an investment fund sued for the repay-

21 The present value loss, or “haircut”, which the debt restructuring entails is typically calculated by comparing the market value of the new debt to the market value of the original claims, both discounted using an identical exit yield. This method has been suggested by Sturzenegger and Zettelmeyer 2005 and has been applied in most subsequent studies. 22 See Rogoff and Zettelmeyer 2002 for a survey of the literature.

158 

 Christian Kopf

Table 2: Characteristics of recent sovereign debt exchanges Issuer

Transaction Year

Original claims, $ billions

Present Value loss

Participation rate

Pakistan

1999

0.61

15%

99%

Ukraine

2000

1.60

18%

97%

Ecuador

2000

6.70

38%

98%

Moldova

2002

0.04

37%

100%

Uruguay

2003

3.13

10%

93%

Dominica

2004

0.14

54%

72%

Grenada

2005

0.21

34%

97%

Argentina

2005

60.57

77%

76%

Dominican Rep.

2005

1.10

 5%

94%

Belize

2007

0.52

24%

98%

Seychelles

2010

0.32

56%

84%

Côte d’Ivoire

2010

2.94

55%

99%

Greece

2012

271.34

65%

97%

Sources: Das et al. (2012), Trebesch (2011), Zettelmeyer (2012), Cleary Gottlieb website, author’s calculations.

ment of nonperforming sovereign debt and obtained an attachment order from a New York court, which enabled it to intercept interest payments on restructured debt in Belgium. In order to avoid a technical default on its Brady Bonds, the government decided to settle with its creditor.23 In the case of Ecuador’s restructuring in 2000, a small group of holders refused to tender their holdings in the bond exchange. The Central Bank of Ecuador then decided to continue to honour these original bonds. Apart from potentially violating the principle of equal treatment of all holders of a certain class of claims, the holdout problem has been responsible for a lack of finality in debt restructurings. This can impose a large cost to the sovereign, in the form of continuous risk of litigations and attachment orders, and it prevents the normalisation of its relationship with creditors. An example is the case of Argentina, which is still struggling with holdout creditors seven years after its 2005 debt restructuring. Two solutions have been proposed to deal with this creditor coordination failure. The IMF championed a statutory approach of establishing a Sovereign

23 Moody’s 2000.



Sovereign Debt Restructuring: Lessons from History 

 159

Debt Restructuring Mechanism, but eventually saw its efforts frustrated by strong resistance from private sector creditors and bank lobbying organisations such as the IIF. Instead, the contractual approach of introducing Collective Action Clauses was chosen. These covenants allow a qualified majority of investors to vote on changes to the payment terms of bonds that are binding on all holders, thereby eliminating the ability of small groups of non-consenting creditors to obstruct the restructuring process or to gain payment preference. CACs had already been an established feature of English law bond contracts, and they have been successively included in sovereign bonds issued under New York law since 2003. The strong focus of policymakers and academics on the holdout problem in Emerging Markets sovereign debt restructurings may explain the prominence of CACs in proposals of the European Commission and the Eurogroup to improve the efficiency of potential future sovereign debt restructurings in Europe. However, it is difficult to find empirical evidence for the assertion that holdout creditors presented a major problem in recent sovereign debt restructurings. Table 2 shows that participation rates of 93% (Uruguay) to 99% (Pakistan) have been achieved in most bond exchanges, in spite of the fact that in most cases, CACs were not contained in the old bond contracts or not invoked by the issuer.24 The main reason why issuers have been able to achieve high participation rates in recent debt workouts in spite of the lack of majority restructuring provisions in bond contracts is the design of the restructurings as bond exchanges with “exit consents”, a legal technique that has been championed by the law firm Cleary Gottlieb.25 Given that the required unanimity made it close to impossible to achieve changes in payment terms on New York law sovereign bonds, the issuers instead presented their creditors with offers to voluntarily exchange existing instruments, which may already have been non-performing or on which it threatened default, into new securities with lower face value, longer tenor and/or lower coupon rates. In accepting the tender offers, the creditors were forced to consent to changes in the non-payment terms of the existing bonds, which can typically be achieved by simple majority under the covenants of State of New York law bonds. The purpose of this “exit consent” is to strip away protective covenants such as cross-default and acceleration clauses, listing requirements, the waiver

24 CACs only played a prominent role in the 2002 restructuring of a $ 40 million Eurobond issued by Moldova and in the 2009 restructuring of the Seychelles. CACs were also invoked on some bonds in the Ukraine restructuring of 2000 and on one bond in the Uruguay restructuring of 2003. In the Pakistan and Argentina restructurings, the issuer decided not to use CACs, and in other recent Emerging Markets sovereign debt restructurings, outstanding bonds did not contain CACs (Sturzenegger and Zettelmeyer 2006, Das et al. 2012). 25 Buchheit and Gulati 2000.

160 

 Christian Kopf

of sovereign immunity or of the requirement for IMF membership, and thereby make the instruments less attractive to holdout creditors. This technique presents a credible threat to creditors, which explains the high participation rates typically achieved in recent sovereign bond exchanges.26 The notable exception to this series of successful sovereign bond exchanges is the case of Argentina, which presented its creditors with an excessively harsh debt restructuring proposal in 2005. But even in the case of Argentina, the “voluntary” participation rate in the bond swap reached 76%, and it rose to 92% after the country re-opened the exchange offer to holdout creditors who saw their efforts to secure a better deal frustrated in the decade following the country’s default27. Until today, Argentina’s holdout creditors have been unable to secure payments through their litigation efforts. We can thus conclude with Panizza et al. that proposals to address the holdout problem “can perhaps be criticised (with the benefit of hindsight) for having barked up the wrong tree – creditor coordination failures did not, in the end, turn out to be a significant impediment to the debt renegotiations of the 1998–2005 period”.28 The fact that policymakers have focused on the wrong problem, namely low participation rates in the restructuring of bonds issued under State of New York law due to obstructions by non-cooperative creditors, does not imply that the recent sovereign bond exchanges offer a suitable model for the restructuring of public debt. Following the transformation of Emerging Markets bank loans into tradable bonds through the Brady deals of the 1990s, the overwhelming majority of claims against these sovereigns are now held by pension funds, mutual funds, insurance companies and hedge funds. However, in what has become the norm since the Russia and Ecuador bond exchanges of 2000, asset managers as well as individual investors with holdings of distressed or non-performing sovereign bonds have faced government representatives unwilling to engage in dialogues with their creditors.29 Then, all of a sudden, these bondholders would receive calls from their sales coverage at investment banks with exchange offers that typically entailed steep losses in present value terms. Creditors that had not been invited to participate in the drafting of the economic terms of the restructuring and had not

26 Bi et al. 2011. 27 Hornbeck 2010. 28 Panizza et al. 2009. 29 Even in the case of Russia’s restructuring of Brady bonds (Prins and Ians), negotiations only took place with commercial banks, and asset managers were not invited to the Bank Advisory Committee.



Sovereign Debt Restructuring: Lessons from History 

 161

been given access to detailed information to evaluate the payment capacity of the issuers would then be asked to decide within a few days whether they wanted to accept the offer or risk being left with an illiquid instrument stripped of important covenants via the “exit consent” technique. Sovereign bond exchanges often involve dozens or even hundreds of different instruments, and considering all of these claims on their own merit would raise transaction costs to a prohibitive level. Therefore, borrowers tend to make a very simple offer. In the case of the Brady deals and the Argentina restructuring of 2005, bondholders were offered a small menu of new bonds to choose from. In the case of Greece’s restructuring in 2012, creditors were given only one choice of new instruments. By reducing the complexity of the transaction, the borrower thereby unilaterally reduces the transaction costs and enables the involved parties to reach an agreement. In doing so, however, borrowers typically exploit existing information asymmetries to their favour and under-estimate their solvency. Law firms and investment banks acting on behalf of the sovereign borrower generally do their best to preserve this information asymmetry to the detriment of bondholders, in order to secure a high participation rate at economic terms that are favourable to the issuer. Atomised creditors have generally been unable to work against this approach. The fact that the lack of creditor organisation has tilted the bargaining power in recent debt restructuring episodes towards the debtor may hinder the efficient functioning of the market for sovereign debt in three ways. First, some unilaterally imposed bond exchanges have been characterised by excessive haircuts, which can lead to an increase in overall risk premia. Issuers have sometimes been able to negotiate a debt reduction that was larger than warranted, thereby offering their creditors an unduly low recovery on their claims. We can define the “fair recovery value” in a sovereign debt restructuring as “the ratio of the maximum debt that a country can pay to its stock of debt” under reasonable assumptions, including on the “politically and economically feasible (and sustainable) primary balance”.30 Note that under this definition, the “fair restructuring value” is a ratio that stabilises government debt relative to gross domestic product (GDP). An empirical estimate of this “fair recovery value” requires detailed debt sustainability analyses, which rely on three main inputs: the interest rate on public debt, the trend growth of the economy, and the primary surplus the government can be expected to achieve over time. The long-term interest rates on most of the stock of debt are set in the restructuring exercise. Barring large changes in the exchange rate or the monetary regime, a country’s nominal

30 Ghezzi 2012.

162 

 Christian Kopf

long-term trend growth can be estimated reasonably well based on population and productivity trends. Therefore, the “fair recovery rate” depends mainly on a country’s willingness to pay, namely its willingness to run a primary surplus. It can be argued that countries that have been able to severely reduce their government debt relative to GDP following a sovereign debt restructuring without having to rely on abnormally high primary surpluses may not have provided their creditors with a “fair recovery value” and have thus imposed an excessive haircut. Figure 2 shows the evolution of government debt and primary balance, both relative to GDP, of three countries that have completed sovereign bond exchanges over the past twelve years. Following Russia’s restructuring of Soviet-era debt, the country’s gross government debt has declined to 8 per cent of GDP. Russia did not achieve this through a serious fiscal austerity effort; instead, it benefitted from an overly generous haircut in the 2000 debt exchange and from a rise in oil prices. It can be argued that the favourable change in the country’s terms of trade could not have been foreseen at the time of the restructuring and merely presents “good luck”. But this contingency is precisely the reason why state-contingent payments had been included in the Brady deals of the 1990s: Mexico issued Value Recovery Rights, Venezuela and Nigeria issued oil warrants, the restructured debt of Bulgaria and Bosnia contained GDP options, etc. In Russia’s bond exchange of the year 2000, on the other hand, the increased bargaining power of the debtor meant that creditors were not offered warrants which would have allowed them to participate in a better-than-expected recovery of the sovereign’s ability to pay. The same applies to Argentina, which saw its government debt decline to 43 per cent of GDP in the years following the sovereign debt exchange, which was largely due to the aggressive haircuts imposed on government bonds. Argentina also offered its creditors only very limited participation in its economic recovery via GDP warrants. A large share of Argentina’s bondholders believed at the time that the debt relief demanded by the government was too high in light of the country’s capacity to pay, but they still participated in the tender offer because of their poor negotiating position. This assessment of a lack of bargaining power turned out to be correct in light of the inability of non-consenting creditors to secure payments from Argentina in the following years. In the case of Ukraine, the government ran a primary deficit in eleven out of twelve years following its debt restructuring, and only achieved a minimal primary surplus of 0.1 per cent of GDP in 2003. In spite of this lack of any meaningful fiscal effort, public debt declined to 10 per cent of GDP due to the country’s strong real appreciation. This shows that the country’s capacity to pay did not warrant imposing a haircut on creditors during the 2000 restructuring. From a welfare perspective, imposing larger or even excessive haircuts on sovereign creditors may not appear problematic at first sight, since this helps



Sovereign Debt Restructuring: Lessons from History 

 163

Argentina

140

15

Government Debt

120

Primary Balance

100 80

0

60 40 20 0

t–1 t t+1

t+3

t+5

t+7

t+9

–15

t+11

t–1 t t+1

t+3

t+5

t+7

t+9

t+11

Russia

140

15

Non-Oil Primary Balance

Government Debt

120 100 80

0

60 40 20 0

t–1 t t+1

t+3

t+5

t+7

t+9

t+11

–15

t–1 t t+1

t+3

t+5

t+7

t+9

t+11

Ukraine

140

15

Government Debt

120

Primary Balance

100 80

0

60 40 20 0

t–1 t t+1

t+3

t+5

t+7

t+9

t+11

–15

t–1 t t+1

t+3

t+5

t+7

t+9

t+11

All numbers in per cent of GDP. Figure 2: Government debt and primary balances after recent debt exchanges (Source: IMF)

put the country’s debt onto a sustainable trajectory. However, Grossman and van Huyck have argued convincingly that investors punish strategic defaults more severely than defaults that are “justified” by large negative shocks.31 Imposing excessively high haircuts can damage a country’s reputation and lead to persistently high discount rates on private investments. Cruces and Trebesch have

31 Grossman and van Huyck 1988.

164 

 Christian Kopf

shown that raising the haircut by one standard deviation (22 per cent) leads to a significant increase of country spreads (70 to 149 basis points) up to seven years post restructuring.32 Anecdotal evidence suggests that there are further long-term costs to a country that imposes excessive haircuts on its creditors and thereby demonstrates disregard for their property rights. Russia has been battling with very high levels of capital flight ever since its 1998 default, and when Finnish companies had to decide where to build large pulp and paper plants, they chose Uruguay over Argentina, in part because Uruguay had made an effort to protect the property rights of its creditors in restructuring its sovereign debt. Imposing excessive haircuts in sovereign bond exchanges thus speaks of the short-sightedness of country representatives negotiating the debt, as it imposes a long-term cost to the country. Second, unilaterally imposed bond exchanges that rely on the “exit consent” technique have in the past led to discrimination against traditional investors. This is mainly related to differences in the fee structure between investor groups, which make it profitable for hedge funds to invest in the legal and economic advice required to overcome information asymmetries and to evaluate the distressed sovereign’s ability to pay, or to litigate against an unfair deal. “Real money” investors such as pension funds and mutual funds, on the other hand, are often forced to accept a bad deal for their clients – if they haven’t sold their paper to hedge funds before the restructuring. Table 3 offers a numerical example to illustrate this point. We can consider two funds with € 1 billion in net assets each – a “real money” pension fund which pays a flat management fee of 80 basis points p.a. to its investment advisors and a hedge fund which pays a management fee of 2 per cent and a performance fee of 20 per cent to its management company. Both funds hold € 100 million in face value of bonds issued by a distressed sovereign that are trading at 40 cents on the euro. The sovereign now presents both funds with a bond exchange offer that would result in a recovery value of 50 per cent of the original claim. Both fund managers are also approached by members of a bondholder committee who believe that this exchange offer underestimates the country’s ability to pay and that there is a high probability the sovereign would accept a settlement at a recovery value of 70 per cent instead, if it is faced with the threat of litigation. The funds are invited to join the bondholder committee against a pro-rata participation in legal expenses amounting to € 250,000 each. In this setting, the manager of the “real money” fund will decide to accept the bond exchange offer, because it will typically not be able to charge legal expenditure to the fund, in addition to

32 Cruces and Trebesch 2011.



Sovereign Debt Restructuring: Lessons from History 

 165

Table 3: Hypothetical comparison of negotiation strategies of different investor groups Investor type

“Real money”

Hedge Fund

Fund size (net asset value, NAV)

€ 1,000 million

€ 1,000 million

Holdings of distressed issuer

€ 100 million face value, marked at a price of 40%

€ 100 million face value, marked at a price of 40%

Fee structure

0.80% management fee no performance fee

2.00% management fee 20% performance fee

Option 1: Participate in bond exchange, recovery value of 50%

Fund assets rise to € 1,010 million Fee income of € 0.08 million

Fund assets rise to € 1,010 million Fee income of € 2.2 million

Option 2: Litigation, recovery value rises to 70%

Fund assets rise to € 1,030 million Gross fee income of 0.24 million Legal expenses of € 0.25 million

Fund assets rise to € 1,030 million Gross fee income of € 6.6 million Legal expense of € 0.25 million

Chosen Strategy

Option 1: Settle for a recovery value of 50% and don’t engage lawyers

Option 2: Engage lawyers and litigate

Payout for investors

NAV gain of € 9.92 million

NAV gain of € 23.40 million

Payout for fund manager

Additional fees of € 0.08 million

Additional fees of € 6.35 million

the flat management fee that has been negotiated with the pension plan sponsor. The hedge fund manager, on the other hand, will be happy to join the bondholder committee, even if it cannot charge the legal expenses to the fund, because of the prospect of tripling both its fee income and its clients’ profits. This result is robust even if somewhat higher legal expenditures or lower hedge fund fees are assumed: “real money” funds typically won’t actively participate in debt renegotiations if their management companies have to bear the associated expenses, even if this means accepting a worse deal for their customers. In order to overcome this discrimination, it is common practise in the US corporate bond market that the issuer pays the legal expenses of bondholder committees in a debt restructuring. However, this provision does not exist for sovereign bond restructurings, which probably explains why it has been impossible to form bondholder committees in order to improve creditor representation in the past. The IMF calls for “good faith negotiations” with creditors in its “lending

166 

 Christian Kopf

into arrears policy” and the European Council concluded on 25 March 2011 that a member state under a European Stability Mechanism (ESM) programme which is pursuing a sovereign debt restructuring “will be required to engage in active negotiations in good faith with its creditors to secure their direct involvement in restoring debt sustainability”.33 However, both the IMF and the European Council have so far remained silent with respect to what “good faith negotiations” implies for bondholder representation. Third, unilaterally imposed bond exchanges have generally led to undue delay in restructuring sovereign debt that has become unsustainable. One could argue that the involvement of creditor committees would lead to longer delays in sovereign debt restructuring, but Trebesch has demonstrated empirically that these delays are typically debtor related and not creditor related.34 In addition to the long duration of debt renegotiations, there have often been delays in starting these negotiations, as politicians remain in denial about the state of affairs. This may be due to political economy issues: governments that lead their countries into IMF programmes or debt restructurings are generally overthrown or voted out at the next occasion, even if these are necessary steps for their country. From a purely welfare-maximising point of view, it is difficult to justify the delay of the restructuring of an insolvent sovereign: the country will not be shielded from elevated market risk premia on its public and private debt, which will manifest themselves in a severe output contraction, independently of whether or not the government ultimately decides to validate the market’s anticipation of a default.35 The three main problems with sovereign bond exchanges in their current form, namely excessive haircuts, discrimination against traditional investors, and undue delay, do not imply that bank loan restructurings through the London Club offer a better model in dealing with sovereign debt overhangs. Trebesch has calculated that bank loan restructurings in the 1980s and 1990s took even longer than sovereign bond exchanges, and a lack of realism with respect to the capacity to pay of Emerging Markets sovereigns prevented the necessary debt relief for too long, which resulted in a lost decade for Latin America and for other developing countries.36 However, the bargaining power in debt renegotiations may have shifted too far from the times of the London Club, when a creditors’ cartel could insist on draconian payment terms because of its ability to coordinate in order to block future lending to undisciplined debtors, to a setting in which sovereigns

33 European Council 2011. 34 Trebesch 2010. 35 Yeyati and Panizza 2011. 36 Trebesch 2008.



Sovereign Debt Restructuring: Lessons from History 

 167

can exploit information asymmetries vis-à-vis atomised and largely uninformed bondholders that have no representation at all in restructuring negotiations.37

37 Several strands of the literature challenge this interpretation of the historical evidence. Cruces and Trebesch (2011) show that on average, haircuts in bank loan restructurings were not higher than haircuts in sovereign bond exchanges. This result, however, does not rule out the existence of excessive haircuts in recent sovereign bond exchanges, since Cruces and Trebesch do not compare actual haircuts to haircuts that would be required to restore debt sustainability under reasonable assumptions. The collapse of public debt ratios in Russia and Argentina following the debt restructurings of 2000 and 2005, which occurred while these countries were running limited primary surpluses, suggest that at least some debt swaps in recent history entailed excessive debt relief. Enderlein et al. (2012) present an index to evaluate the coerciveness of sovereign debt restructurings. Summary statistics on this index seem to suggest that sovereign bond exchanges in the past 14 years were no more coercive than sovereign loan restructurings in the 1980s and 1990s. It remains unclear whether this coerciveness index offers a good proxy for the effects of differing bargaining power on debt negotiation outcomes. The index is a composite of nine dummy indicators that characterise observed government actions with respect to the payment and renegotiation of their debt, such as payment suspensions or threats of debt repudiation, etc. However, the concentration of bargaining power with the sovereign debtor does not need to manifest itself in payment suspensions or the breakdown of negotiations in order for a sovereign debt restructuring to be correctly qualified as coercive. Greece’s creditors knew that the sovereign was in a position to force the proposed change in payment terms upon them, and they did not need proof of a payment suspension or the declaration of a moratorium to make this judgement. In order to evaluate the coerciveness index, it is also worth comparing the debt restructurings of Argentina, the Dominican Republic and Greece. In 2005, the Dominican Republic offered its creditors to exchange bonds that were coming due in 2006 into new securities with maturity in 2011 which carried the same coupon, had no principal reduction, and an accelerated redemption via amortisation payments. The present value reduction has been estimated at 5%, the Dominican Republic did not attempt to drag along non-consenting creditors and it continued to pay these holdouts in full. Greece, on the other hand, forced its private creditors to accept a 50% reduction in the principal value of their claims, lowered coupon rates substantially and increased the maturity of the debt by up to 30 years. CACs were retrofitted into the existing contracts and invoked in order to make the change in payment terms binding on all holders of domestic debt. Official sector holders of government bonds were exempt from the debt restructuring, which resulted in higher haircuts for private creditors. The haircut suffered by Greek bondholders has been estimated at 65%. Argentina took a similar stance towards its creditors in the 2005 restructuring. Enderlein et al. would probably assign a coerciveness index value of 2 (on a scale from 0 to 10) to Greece’s recent sovereign debt restructuring, because the country didn’t miss any payments, declared no moratorium, had no data disclosure problems etc. They would assign an index value of 3 to the restructuring of the Dominican Republic, and they would assign a value of 9 to Argentina, because of the government’s rough negotiation tactics. However, most observers would agree that Greece’s debt restructuring was far more coercive than the approach chosen by the Dominican Republic, and indeed similar to Argentina’s bond exchange. This is not apparent in the index values. The IMF has argued that “a combination of high participation rates and speedy completion of most recent debt restructuring episodes likely suggest that the underlying offers were seen

168 

 Christian Kopf

It should be noted that the legal technique of tilting the bargaining power in sovereign bond exchanges towards the debtor through the use of exit consents may recently have lost some of its appeal, since the English High Court ruled these provisions as illegal under UK law on 27 July 2012 because they are oppressive towards minority bondholders. The court’s judgment was related to the use of exit consents in the restructuring of subordinated notes issued by Anglo Irish, and, “unless overturned on appeal, casts doubt on the effectiveness of exit consents to restructure debt”.38

3. Desiderata for Future Debt Renegotiations In light of the identified shortcomings of recent sovereign bond exchanges, we can now ask what can be done to improve the sovereign debt restructuring process. The answer is rather simple: sovereign borrowers and their creditors should apply the “Principles for Stable Capital Flows and Fair Debt Restructuring in Emerging Markets” which have been developed by a group of representatives from major sovereign issuers, institutional investors and investment banks following Argentina’s and Brazil’s sovereign debt crises.39 This code of conduct calls for transparency, cooperation, good faith negotiations and fair treatment of the affected creditors in a restructuring.40 As a precautionary measure, the Principles call for an improvement in investor communication. Lack of timely access to macroeconomic data and lack of policy transparency make it difficult to evaluate a sovereign’s capacity to pay. The resulting increase of transaction costs41 unnecessarily raises lending rates. This has been all too apparent in the European sovereign debt crisis that started in 2010. In the absence of effective communication channels between European governments and institutional investors, market participants came to rely on broker-dealer research or even blogs such as ZeroHedge or Roubini’s EconoMonitor. False rumours galore

by participating creditors as reasonable, in that they reflected governments’ capacity to pay and offered adequate burden sharing” (IMF 2012, p. 10). In light of the arguments presented above, it appears unreasonable to draw this general conclusion from the empirical evidence. High participation rates and speedy completion could also have been related to a strategy of blackmailing creditors. 38 Davis Polk 2012. 39 IIF 2005. 40 For the purpose of full disclosure, it should be noted that the author participated in preparatory work on this code of conduct. 41 Information and control costs, cf. Coase 1961.



Sovereign Debt Restructuring: Lessons from History 

 169

then drove up secondary market yields and contributed to an exodus of international investors.42 This could and should have been avoided by setting up investor relations offices that provide timely information via dedicated websites and email distribution lists and conduct regular conference calls which are open to all creditors. Many European governments are now starting to adopt this recommendation. If a debt restructuring cannot be avoided, the Principles call for measures to close the gap in the institutional infrastructure that has arisen due to the move from bank intermediation to market intermediation between sovereign borrowers and their creditors, namely the lack of creditor representation in sovereign debt workouts due to the disappearance of the London Club. Creditor representation through the formation of duly appointed bondholder committees should be encouraged. An important aspect of bondholder representation has not been included in the Principles: the sovereign debtor should cover reasonable legal and actuarial expenses of bondholder committees, in order to help overcome the problems of information asymmetry and discrimination against long-term investors. As discussed in the previous section, this already forms part of the standard covenants in US corporate bond contracts, and this practise should be extended to the sovereign bond market. Multilateral lenders have ample leverage to influence the form of debt renegotiations. They can and should use the “power of the purse” to encourage institutionalised bondholder representation in negotiations on sovereign bond exchanges.43 Thereby, it would be possible to avoid a repeat of the situation recounted in the introduction to this paper, in which representatives of third

42 As an example, on June 24th, 2010, Goldman Sachs published a research report claiming that the Spanish government would likely face a cash shortfall of € 12.6 billion in July 2010, which would likely have to be covered by credit lines with commercial banks if government bond issuance couldn’t be increased (Pérez de Azpillaga 2010). This estimate turned out to be false, but it was almost impossible for institutional investors to verify the numbers due to a lack of effective investor communication channels, and it greatly contributed to market nervousness about Spain’s ability to service its government debt. 43 It has sometimes been argued that the formation of representative creditor committees presents a practical problem due to the multitude of bond series and creditor groups in sovereign debt restructurings. This argument, however, is unconvincing. Corporate bond restructurings are often also characterised by a large variety of creditor claims, which typically does not impede the formation of bondholder committees. And indeed, representative bondholder committees have been formed in several recent sovereign debt restructurings, e.g. in the cases of Ecuador, Argentina and the Dominican Republic. However, these bondholder committees have often not been recognised by the sovereign and its legal advisors (Das et al. 2012), and several of these committees fell apart because of a lack of funding.

170 

 Christian Kopf

parties decided on the terms of a sovereign bond exchange without the participation of the sovereign debtor or of duly appointed representatives of its creditors. The mandatory inclusion of bondholder committees in the negotiation process would go a long way towards more orderly sovereign debt restructurings. However, setting up bondholder committees won’t be sufficient in dealing with holdouts. As discussed in the previous section, the collective action problems posed by non-cooperative creditors can be addressed by providing sufficient economic incentives to participate in a restructuring,44 and by setting economic disincentives for holdouts through defensive exit consent provisions or by introducing ambiguity about future payments on original claims.45 All of this can be achieved without having to rely on contractual novation such as the introduction of CACs. However, it is possible that the inclusion of CACs and aggregation clauses in sovereign bond contracts issued under international law offer a more effective way to deal with the holdout problem than the exit consent technique.46 Another crucial element of a more orderly debt restructuring process would be the return to the principle of preference avoidance that has guided corporate bankruptcy proceedings in the United Kingdom and other jurisdictions for more than 400 years. The principle of preference avoidance implies that all parties of the same creditor class should be offered equal treatment. This does not preclude assigning a preferred creditor status to another creditor class. Multilateral lenders that provide a distressed sovereign with financial assistance at concessionary terms should be given seniority, in analogy to the provisions of debtor-in-possession financing in corporate bankruptcy proceedings.47 However, in the case of Greece’s recent sovereign debt restructuring, ECB holdings of government bonds that had been purchased in the secondary market, on equal footing with private creditors, have been given de facto seniority by assigning different ISIN codes to these bonds only days before the exchange offer was launched. From a legal perspective, this clearly violates the principle of preference avoidance, since the seniority of a claim should be defined by its legal status, and not by the person of the holder. From an economic perspective, it is difficult to justify granting these ECB holdings seniority: these purchases did not provide the sovereign with “new

44 Uncoerced creditors will be economically incentivised to accept a sovereign debt restructuring proposal if it is based on a realistic assessment of the issuer’s ability to pay and provides for adequate debt relief along with optionality to participate in a better-than-expected economic recovery. Under these circumstances, the restructured debt should be expected to increase in value over time as the uncertainty about future payments is removed. 45 Zettelmeyer 2011. 46 Bi et al. 2011. 47 Kopf 2011.



Sovereign Debt Restructuring: Lessons from History 

 171

money” that would have allowed it more time to undertake an economic adjustment program, nor did they come at concessionary terms that would increase recovery values for private creditors in a restructuring.48 It thus appears desirable that in future sovereign debt restructurings, all holders of claims with identical ex ante legal status should be treated equally.

4. The Role of Collective Action Clauses in Euro Area Sovereign Debt Restructurings We have shown in the previous sections that the introduction of CACs in Emerging Markets government bonds that have been issued under New York law may have helped to deal with the holdout problem that has been the primary concern of policymakers, even if CACs do not help to overcome the dominant problems related to Emerging Markets sovereign bond exchanges, namely the risk of excessive haircuts, discrimination against traditional investors, and undue delay. Over the past two years, European policymakers have put considerable efforts into developing a more orderly sovereign debt restructuring mechanism, as evidenced by the work of the EFC Sub-Committee on EU Sovereign Debt Markets and various discussions in the Eurogroup. This work culminated in the European Council’s conclusion on 25 March 2011 that CACs will be included in all new euro area government securities with maturity above one year, from July 2013 onwards.49 The stated objective of this decision is “to facilitate agreement between the sovereign and its private-sector creditors in the context of private sector involvement”.50 The fact that European policymakers have chosen CACs as the main instrument to achieve the laudable objective of improving the process of sovereign debt renegotiations unfortunately demonstrates an undue focus on “the problem that wasn’t”51, namely the problem of holdouts. It also means that policymakers have mistakenly applied an instrument that may well be appropriate for an Emerging Markets context, where a large share of sovereign debt is typically denominated in foreign currencies and issued under New York law, to the European context, which is marked by a starkly different design of sovereign bond contracts.

48 Ghezzi 2012. 49 The effective date for the inclusion of CACs has since been brought forward to January 2013 by Article 12, para. 3 of the ESM Treaty. 50 European Council 2011. 51 Bi et al. 2011.

172 

 Christian Kopf

As an ECB Task Force report rightly concludes, “Collective Action Clauses … are aimed at … avoiding situations where a minority of bondholders block an agreement on a change in the terms of the contract, and ensuring that disruptive legal action by individual creditors does not hamper the debt workout”.52 We have explained in section 2 of this paper that this problem could arise with sovereign bonds issued under New York law contracts that require unanimous approval by holders on changes to payment terms. However, the bulk of the sovereign debt stock of euro area member states has been issued under domestic law, as Table 4 shows. Table 4: Governing law of outstanding euro area sovereign debt (Source: Carlson (2012)) Issuer

Portion of outstanding debt governed by domestic law

Portion of outstanding debt governed by foreign law

Austria

88.1%

9.5% English law, 1.6% German law, 0.1% Italian law

Belgium

99.8%

0.1% English law

France

100%

Germany

100%

Greece (pre exchange)

92.0%

6.4% English law, 0.1% Italian law, 1.4% Other

Ireland

90.8%

0.1% English law, 9.1% Other

Italy

97.5%

1.8% US law, 0.2% English law, 0.5% Other

Netherlands

100%

Portugal

97.4%

2.6% English law

Spain

98.9%

1.1% English law

Finland

89.4%

9.9% English law, 0.4% US law

Prior to the 2012 debt restructuring, 92% of Greece’s bonded government debt was governed by Greek law. This has very important consequences for the possible design of a debt restructuring: it means that the payment terms of the bonds can be unilaterally changed by a simple act of parliament. Maturities can be extended, coupon rates can be cut, or the principal repayment can be reduced in order to

52 ECB 2005.



Sovereign Debt Restructuring: Lessons from History 

 173

lower the sovereign’s debt burden.53 Any of these changes, if put into effect via domestic legislation, becomes binding on all holders. Holders of domestic government bonds will not be able to litigate in English or New York courts or attempt to freeze Greek assets abroad. In another important difference to the Emerging Markets context, government bond contracts in the euro area are typically not based on offering circulars with detailed covenants and several hundred pages of information on the borrower’s economic and political conditions, but on a simple term sheet that contains no provisions on creditor protection. Table 5 provides a summary of the legal characteristics of government bonds issued by five euro area member states. Prior to the 2012 debt exchange, domestic Greek government bond contracts contained no “negative pledge” clause. This implied that Greece could pledge any state income to third parties, such as airport taxes or lottery income, and these income streams would not be available to holders of sovereign debt in the event of a restructuring. As there was no “cross default” provision, Greece could undertake a selective default without triggering a restructuring of its entire stock of public debt. In the absence of a “pari passu” clause, Greece was also in a position to accord payment preference to certain obligations in a restructuring. Table 5: Investor protection covenants in euro area domestic government bonds (Source: Gulati and Zettelmeyer (2012)) Acceleration Modification Cross Default Neg. Pledge

Pari Passu

Greece (pre exchange) Yes

No

No

No

No

Ireland

No

No

No

No

No

Italy

Yes

No

Yes

Yes

Yes

Portugal

No

No

No

No

Yes

Spain

No

No

No

No

Yes

Euro area governments are thus in a position that allows them to dictate the conditions of a restructuring of between 88% (in the case of Austria) and 100% (in the cases of France, Germany and the Netherlands) of their government bonds. They could also use this power in a more cooperative way, e.g. by negotiating the terms of a debt restructuring and then applying these terms to non-consenting

53 Alternatively, the Greek parliament could have introduced a withholding tax of, say, 50% on principal and interest payments of Greek government bonds, which would have achieved the same objective.

174 

 Christian Kopf

creditors via a “mop up law”, once agreement has been reached with a supermajority of bondholders.54 In either case, the fact that a very large portion of bonded euro area sovereign debt is governed under domestic law means that the problem of holdouts, which may arise in an Emerging Markets context, does not apply to the euro area. As the problem of holdouts does not need to arise, there is no need to deal with it via CACs. Several objections have been brought forward against this conclusion. First, some argue that CACs should still be introduced into foreign law bonds issued by euro area governments. However, almost all of this debt has been issued under English law and already contains CACs. Others have argued that bondholders could seek legal remedy against a unilateral restructuring of domestic law bonds issued by euro area sovereigns. One avenue would be to sue a member state in the European Court of Human Rights on the basis of a violation of the right to the peaceful enjoyment of one’s possessions. However, the Protocol to the Convention for the Protection of Human Rights explicitly allows sovereigns to enforce laws that deprive natural and legal persons of their property if this is in the public interest. Another avenue would be to invoke investor protection rules contained in Bilateral Investment Protection Treaties (BITs), as advocated by Boris Kasolowsky and Smaranda Miron in their contribution to this volume. In most jurisdictions, however, BITs are not powerful enough to prevent the sovereign from exercising its sovereign right to change domestic legislation. Where this is not the case, BITs can be amended or cancelled. Another argument in favour of CACs in the euro area context has been brought forward by Bradley and Gulati: they present econometric estimates which suggest that CACs are associated with lower borrowing rates of weaker sovereigns, presumably because they allow for a reduction in transaction costs during debt renegotiation.55 However, this empirical result does not allow conclusions for the euro area context because the authors only compare borrowing costs of New York law bonds with and without CACs. The relevant comparison for the European context would be between the risk premia on New York law sovereign bonds with CACs and domestic law sovereign bonds, and Bradley and Gulati provide no empirical data on this. Finally, some observers have put forward the proposition that retrofitting CACs into the stock of domestic government bonds has allowed Greece to restructure its debt in a non-coercive way that relied on voluntary approval by a super-

54 Gulati and Buchheit 2010. 55 Bradley and Gulati 2011.



Sovereign Debt Restructuring: Lessons from History 

 175

majority of creditors. This argument, however, is deeply misleading. In preparation for the recent debt restructuring exercise, the Greek parliament passed the Bondholders Act 4050/12, which allowed a cross-series modification of payment terms via a two-thirds majority vote of bondholders. The quorum for this vote was set at half of the amount outstanding of all eligible titles. Greece’s bond exchange proposal also employed the “exit consent” technique described in section 2 of this paper, which implied that any investor deciding to tender its old Greek government bonds in the exchange automatically gave consent to the changes in payment terms. As a result, it was sufficient for Greece to gain the approval of holders of one third of the stock of domestic government bonds outstanding in order to make changes of payment terms binding on all holders (i.e. Greece needed to reach a quorum of 50 per cent, and it needed the participation and/ or consent of two-thirds of this quorum). Reportedly, the Greek government also gave hints to domestic banks, insurance companies and social security funds who were collectively holding around half of the eligible securities,56 that their institutions would only be eligible for public support if they participated in the bond exchange. Furthermore, it was widely expected, and later on confirmed, that the ECB would not accept untendered Greek government bonds as eligible collateral for refinancing operations after completion of the debt swap. Investors obviously understood that under these conditions, the Greek government would easily meet the required votes that would allow it to invoke the provisions of the Bondholders Act, which explains the lack of resistance against the tender offer. The provisions of this Greek act have perverted the concept of collective action clauses. CACs have been designed as majority restructuring and enforcement provisions, which are meant to allow a qualified majority of bondholders to limit the ability of a non-cooperative minority of creditors to obstruct a debt restructuring that is beneficial to all parties, if full participation cannot be achieved. The Greek Bondholders Act, on the other hand, allowed a minority of bondholders, who happened to be under the indirect control of the Greek government, to validate a significant deterioration in payment terms that was made binding on all holders. There may be strong economic reasons to make a sovereign bond exchange involuntary in order to avoid the problem of holdouts, but this motivation should not be concealed by creating the illusion of a lack of coerciveness. This criticism of Greece’s use of CACs, which violated the principles for which majority voting provisions have been designed, does not apply to the final version of CACs that has been approved by European finance ministers. These uniform provisions, which will come into force in new sovereign bond contracts from 2013

56 Garcia Pascual and Ghezzi 2011, p. 10.

176 

 Christian Kopf

onwards, provide for a quorum of enfranchised holders of two-thirds of the principal amount outstanding for a change in payment terms to be considered at a bondholder meeting, and an approval threshold of three-quarters of the principal outstanding represented in the meeting to validate the proposed restructuring. As a result, a minority of bondholders will not be able to enforce their will.57 In sum, we have shown that the introduction of CACs may be a useful amendment to government bond contracts issued under New York law, but in the particular case of the euro area, CACs are not needed in order to overcome collective action problems in sovereign debt restructurings. 

Conclusion This paper argues that the current “rules of the game” of the European Monetary Union, namely a fixed exchange rate, open capital accounts and the prohibition of monetary financing of government debt, have re-introduced sovereign default risk to its member states. As long as this institutional setting remains in place, there is a need to allow for orderly sovereign debt restructurings. Negotiations between governments and Bank Advisory Committees of the London Club during in the 1980s and 1990s did allow for an institutionalised debt restructuring process. However, the London Club process was fraught with shortcomings: it led to long delays and the power of this creditors’ cartel often prevented adequate debt relief. Furthermore, the London Club process cannot be applied in current circumstances, as it was predicated on bank intermediation between sovereigns and their creditors via syndicated loans that were generally held by a small number of institutions. Since lending relationships have become dominated by market intermediation instead of bank intermediation, debt restructurings have typically taken the form of sovereign bond exchanges. These debt swaps have been efficient in the sense that they largely avoided the problems of holdouts via the legal technique of “exit consent”. But they have also been characterised by sometimes-excessive haircuts, discrimination against traditional investors, and undue delay. In part, this is due to the lack of creditor representation: since the year 2000, debt restructurings have typically been designed as a “take-it-or-leave-it” exchange offer. Sovereign borrowers have used

57 It should be noted that these final terms on quora and approval thresholds represent a significant improvement over the CAC terms that had initially been drafted by the law firm Clearly Gottlieb on behalf of the EFC Sub-Committee on EU Sovereign Debt Markets.



Sovereign Debt Restructuring: Lessons from History 

 177

their increased bargaining power by insisting on debt reductions that “were not negotiated or even really discussed with bondholders but to a large extent were unilaterally imposed”.58 This may at times suit the interests of elected officials, but there is a long-term cost associated with this strategy in the form of higher risk premia on future borrowings,59 and other, less tangible effects of damages to the country’s reputation. Finally, the recent Greek debt exchange has allowed for undue discrimination against private creditors by granting ECB holdings of Greek government bonds that had been purchased in the secondary market de facto seniority. A first requirement to overcome the identified shortcomings of debt exchanges would be a commitment by sovereign debtors, multilateral institutions and the ECB to respect the principle of preference avoidance in future debt restructurings. Furthermore, sovereigns and their creditors should commit to good faith negotiations based on transparency and cooperation. In the conclusions of its March 2011 meeting, the European Council decided that any EU member state under an ESM programme “will be required to engage in active negotiations in good faith with its creditors”60 in the context of a sovereign debt restructuring. In order to effectively implement this Council conclusion, EU member states that entertain a sovereign debt restructuring should be required to facilitate creditor representation by covering reasonable legal and actuarial expenses of duly appointed creditor committees. The formation of bondholder committees should also be encouraged by the IMF through conditionality in financial assistance programmes. However, an improvement in creditor representation will not be sufficient in dealing with the problem of holdouts. Contractual novations of New York or English law bonds such as the introduction of CACs and aggregation clauses can help to overcome the ability of small groups of non-consenting creditors to obstruct a debt restructuring process or to gain payment preference. However, CACs are not needed to overcome collective action problems in euro area sovereign debt restructurings, since government bonds have largely been issued under domestic law and can thus be restructured unilaterally, if needed. The focus on CACs risks diverting attention away from the key issues in redesigning the relationship between sovereign borrowers and lenders in the euro area, namely the need for fair treatment of private creditors, the need for better investor communication and the need for creditor representation in sovereign debt renegotiations.

58 de Larosière 2005. 59 Cruces and Trebesch 2011. 60 European Council 2011, p. 30.

178 

 Christian Kopf

Bibliography Bi, Ran, Marcos Chamon and Jeromin Zettelmeyer (2011), “The Problem that Wasn’t: Coordination Failures in Sovereign Debt Restructurings”, manuscript Boone, Peter, and Simon Johnson (2010): “Brady Bonds For the Eurozone”, Project Syndicate, 15 September 2010 Borensztein, Eduardo, Eduardo Levy Yeyati and Ugo Panizza (2005), Living with Debt: How to Limit the Risks of Sovereign Finance, Cambridge, Mass.: Harvard University Press Bradley, Michael, and Mitu Gulati (2011), “Collective Action Clauses for the Eurozone: An Empirical Analysis”, manuscript Buchheit, Lee, and Mitu Gulati (2000): “Exit Consents in Sovereign Bond Exchanges”, UCLA Law Review, Vol. 48, pp. 59–84 Coase, Ronald (1961), “The Problem of Social Cost”, Journal of Law and Economics, Vol. 3, pp. 1–44 Carlson, Sarah (2012) “Unilateral Action Threatened by Greece is Also Available to Other Sovereigns”, Moody’s Weekly Credit Outlook, 6 February 2012 Coke, Edward (1826): The Reports of Sir Edward Coke, Knt: In Thirteen Parts, Vol. 1, London: Joseph Butterworth and Son [The Case of Bankrupts, 1584] Cottarelli, Carlo, Lorenzo Forni, Jan Gottschalk, and Paolo Mauro (2010), “Default in Today’s Advanced Economies: Unnecessary, Undesirable, and Unlikely”, IMF Staff Position Note, No. 10/12, International Monetary Fund Cruces, Juan, and Christoph Trebesch (2011), “Sovereign Defaults: The Price of Haircuts”, CESifo Working Paper, No. 3604 Das, Udaibir, Michael Papaioannou and Christoph Trebesch (2012): “Sovereign Debt Restructurings 1950–2010: Literature Survey, Data, and Stylized Facts”, IMF Working Paper, No. 12/203, International Monetary Fund Davis Polk (2012): “Exit Consents Unlawful Under English Law”, Davis Polk Client Newsflash, 30 July 2012 De Larosière, Jacques (2005): “From Mexico to Argentina: What Have We Learned from Two Decades of Debt Crises?”, Princeton Institute for International and Regional Studies Monograph Series, No. 3, Princeton University De Grauwe, Paul (2011) “Governance of a Fragile Eurozone”, CEPS Working Document, No. 346, Centre for European Policy Studies ECB (2005), “Managing Financial Crisis in Emerging Market Economies: Experience with the Involvement of Private Sector Creditors”, prepared by an International Relations Committee Task Force, ECB Occasional Paper, No. 32, European Central Bank Enderlein, Henrik, Christoph Trebesch and Laura von Daniels (2012): “Sovereign Debt Disputes: A Database on Government Coerciveness during Debt Crises”, Journal of International Money and Finance, Vol. 31, No. 2, pp. 250–266 European Council (2011): “Conclusions of the European Council on 24/25 March 2011”, press release Garcia Pascual, Antonio, and Piero Ghezzi (2011): “The Greek Crisis: Causes and Consequences”, CESifo Working Paper, No. 3663, Center for Economic Studies and Ifo Institute for Economic Research Ghezzi, Piero (2012): “Official lending: Dispelling the lower recovery value myth”, VoxEU.org column, 21 June 2012, http://www.voxeu.org/article/official-lending-dispelling-lowerrecovery-value-myth



Sovereign Debt Restructuring: Lessons from History 

 179

Gulati, Mitu and Lee Buchheit (2010), “How to Restructure Greek Debt”, Duke Law Working Papers, No. 47, Duke University Gulati, Mitu, and Jeromin Zettelmeyer (2012), “Making a Voluntary Greek Debt Exchange Work”, Capital Markets Law Journal, forthcoming Guinnan, Timothy (2004), “Financial Vergangenheitsbewältigung. The 1953 London Debt Agreement”, Economic Growth Center Discussion Paper, No. 880, Yale University Goodhart, Charles (2011): “Global Macroeconomic and Financial Supervision – Where Next?”, NBER Working Paper, No. 17682, National Bureau of Economic Research Grossman, Herschel, and John Van Huyck (1988), “Sovereign debt as a contingent claim: Excusable default, repudiation, and reputation”, American Economic Review, Vol. 78, pp. 1088–1097 Hellenic Republic (2012): Invitation Memorandum dated 24 February 2012 Hornbeck, Jeff (2010): “Argentina’s Defaulted Sovereign Debt”, Report for Congress, Congressional Research Service IIF (2005), Principles for Stable Capital Flows and Fair Debt Restructuring in Emerging Markets, Washington, DC: Institute of International Finance IMF (2003): “Reviewing the Process for Sovereign Debt Restructuring within the Existing Legal Framework”, paper prepared by the Policy Development and Review, International Capital Markets, and Legal Departments, International Monetary Fund IMF (2012): “A Survey of Experiences with Emerging Markets Sovereign Debt Restructurings”, IMF Policy Paper, prepared by the Monetary and Capital Markets Department, International Monetary Fund Kohlscheen, Emanuel (2007): “Why are there Serial Defaulters? Evidence from Constitutions”, Journal of Law and Economics, Vol. 50, No. 4, pp. 713–730 Kopf, Christian (2011): “Restoring financial stability in the euro area”, CEPS Policy Brief, No. 237, Centre for European Policy Studies Moody’s (2000), How to Sue a Sovereign: The Case of Peru, New York: Moody’s Investor Service Panizza, Ugo, Federico Sturzenegger and Jeromin Zettelmeyer (2009), “The Economics and Law of Sovereign Debt and Default”, Journal of Economic Literature, Vol. 47, No. 3, pp. 653–700 Pérez de Azpillaga, Javier (2010): “The cash position of the Spanish state”, Goldman Sachs European Weekly Analyst, No. 10/23, 24 June 2010, p. 8 Pettis, Michael (2001): The Volatility Machine, Oxford: Oxford University Press Reinhart, Carmen, Kenneth Rogoff and Miguel Savastano (2003), “Debt Intolerance”, Brookings Papers on Economic Activity, No. 1, pp. 1–74 Reinhart, Carmen, and Kenneth Rogoff (2009): This Time is Different: Eight Centuries of Financial Folly, Princeton: Princeton University Press Rogoff, Kenneth, and Jeromin Zettelmeyer (2002): “Bankruptcy Procedures for Sovereigns: A History of Ideas, 1976–2001”, IMF Staff Papers, Vol. 49, No. 3, pp. 470–507 Sims, Christopher (2012): “Gaps in the institutional structure of the euro area”, Banque de France Financial Stability Review, No. 16, pp. 217–223 Sturzenegger, Federico, and Jeromin Zettelmeyer (2005): “Haircuts: Estimating Investor Losses in Sovereign Debt Restructurings, 1988–2005”, IMF Working Paper, No. 05/137, International Monetary Fund Sturzenegger, Federico, and Jeromin Zettelmeyer (2006): Debt Defaults and Lessons from a Decade of Crises, Cambridge, Mass.: MIT Press Trebesch, Christoph (2008): “Delays in Sovereign Debt Restructurings”, manuscript

180 

 Christian Kopf

Trebesch, Christoph (2010): “Sovereign Debt Restructurings 1950–2010: A New Database”, manuscript Van Rijckeghem, Caroline, and Beatrice Weder (2009): “Political Institutions and Debt Crises”, Public Choice, Vol. 138, No. 3, pp. 387–408 Yeyati, Eduardo Levy, and Ugo Panizza (2011), “The elusive costs of sovereign defaults”, Journal of Development Economics, Vol. 94, No. 1, pp. 95–105 Zettelmeyer, Jeromin (2011): “Resolving the European debt crisis: Lessons from past experience”, presentation at the Bruegel-PIIE conference in Chantilly, September 2011, manuscript Zettelmeyer, Jeromin (2012): “Greece’s Debt Restructuring: Taking Stock and Looking Forward”, manuscript

Christoph G. Paulus

A Resolvency Proceeding for Defaulting Sovereigns A. Historical Overview A sovereign can go bust as can be seen from a rich history of failing states. For centuries (if not millennia), however, lawyers ignored this phenomenon and proclaimed that a state is incapable of becoming bankrupt. If evidence were needed for this statement, just look at s. 12 par. 1 no. 1 of the German Insolvency Ordinance which states: “Insolvency proceedings may not be opened for the assets owned by the Federation or a Land”. One of the reasons for this attitude is that the former Bankruptcy Ordinance (Konkursordnung) – thereby following a long lasting tradition of several millennia – offered just one option for the solution of the insolvency common-pool-problem,1 namely liquidation. That is why still today there is the wide-spread belief that in case of an insolvency of, for instance, Greece her islands, the Apollon temple in Delphi, or the Erechteion’s caryatids were to be sold. As a matter of fact, such perception is nonsense (in Germany) ever since 1999 when the new Insolvency Ordinance entered into force and with it the new option of rescuing a debtor by means of an insolvency plan. What is needed, accordingly, is that this option becomes ingrained in the stakeholders’ brains. Another reason for the said old-fashioned attitude is that at least for centuries – if not millennia2 – it was taken for granted that a sovereign would be able to get rid of its debts by increasing the tax revenues or by printing new money and, thus, inflating away its debts. However, referring to both options implies a good deal of cynism: To require the increase in tax revenues means that, ultimately, a sovereign can go bust only when and if the last of its citizens is living on or below minimum subsistence level. And the hint to a solid inflation means recommending to play with the fire; especially Germany has had in the twenties of the last century its more than painful experiences with this type of game.

1 Cf. Paulus, Freiheit und Gleichheit als Grenzmarkierung zwischen Zivilrecht und Insolvenz­ recht, Festschrift Medicus zum 80. Geb., 2009, p. 281, 286 ff. 2 About the history of defaulting sovereigns see, for instance, Manes, Staatsbankrotte, 3rd ed., 1923, p. 24 ff.; Lingelbach (ed.), Staatsfinanzen – Staatsverschuldung – Staatsbankrotte in der europäischen Staaten- und Rechtsgeschichte, 2000.

182 

 Christoph G. Paulus

There is even one more reason which offers lawyers even some justification not to bother with developing ideas regarding failing states – a reason which comes into play when (or to the degree that) the debts are owed to that sovereign’s own citizens (present day main example: Japan) or if the law governing those debt instruments is the domestic one (important example: 90 % of the Greek bonds are issued under Greek law). These cases had been quite common in earlier days and they permitted the respective sovereign to cut back or even eliminate its debts by a legislative act. However, as a consequence of the increasing globalization of (and on) the capital markets,3 this remedy has become scarce; since it has become more and more common that bonds are issued on a global scale and that they are governed by foreign (most prominently English law or New York) law.

B. Recent developments I. In high speed For the last 25 years or so, a closer look into the evolution of events reveals an increasing interest in looking for and coming to a structure for the defaulting sovereign phenomenon.4 To be sure, the speed of what is called evolution here resembles more that of a tortoise than that of Achilles. But nevertheless: After centuries of ad hoc-solutions which used instruments ranging from canon boat policy5 to negotiations – the latter ones partially institutionalized in the Paris Club and the London Club6 – it had been originally the NGOs (non governmental organizations) which articulated the idea of a legally structured procedure. In the course of time, they received powerful and influential support by churches, especially by Pope John Paul II, who appealed to the global public in his 1994

3 Informative with respect to the current global financial crisis Brummer, Networks (In-)Action?, in: Cissé/Bradlow/Kingsbury, International Financial Institutions and Global Legal Governance (The World Bank Global Review, Vol. 3), 2012, p. 323 ff. 4 For what follows cf. Paulus, Rechtliche Handhaben zur Bewältigung der Überschuldung von Staaten, RIW 2009, p. 11 ff. 5 Cf. Manes, op. cit., p. 157 f., 181 ff., 210. Regarding the so called Drago/Porter-Doctrine see, i.a., v. Liszt, Völkerrecht, 11th ed., 1918, p. 152; Paulus, in: Wolfrum (ed.), Encyclopedia of Public International Law, s.v. “debts” par. 11 ff. 6 As to both Clubs see Reinisch, State Responsibility for Debts, 1995, p. 18 ff. Additionally, infra sub C II.



A Resolvency Proceeding for Defaulting Sovereigns 

 183

Apostolic Letter “Tertio Millennio Adveniente” (par. 51)7 to treat the first year of the new millennium as a “Jubilee year”8 – which implied an all-encompassing discharge of the debts of all poor countries of this world.

II. The new millennium It is in this context that, right at the beginning of the 21st century, Argentina got into difficulty and became insolvent. This was an incident grave enough that the International Monetary Fund (IMF), in November 2001, saw the need and took the opportunity to present its own proposal of what was labelled “Sovereign Debt Restructuring Mechanism” (SDRM). This move was surprising for all and it made the idea of a legal concept enter the official stage. This SDRM contained quite a precise structure which was modelled to a certain degree on commercial bankruptcy law9 – in particular Chapters 11 and/or 9 of the US Bankruptcy Code. As a matter of fact, the obvious differences between a sovereign and a commercial entity (or even individual) as the debtor of such a proceeding determined the special diligence as to the degree of adaptation. The single most important feature which ist not transferable from the commercial context into the new one is the unconditioned exclusion of any liquidation pattern. The only option of a newly introduced sovereign default procedure must and can be a restructuring procedure. The only goal can be to steer the sovereign through its troubled waters into the calm sea of debt sustainability.

7 See also the later Apostolic Letter by John Paul II from late 2000 “Novo Millennio Ineunte” (par. 14). 8 Cf. Leviticus (3 Mose), 25, 8–15. About the Jubilee Year and its christian reactivation by Pope Boniface VIII in the year 1300 (bull: Antiquorum habet fida relatio) cf., e.g., Thurston, Holy Year of the Jubilee, in: Herbermann (ed.), The Catholic Encyclopedia (1907–1912), Vol. 7, available at: http://oce.catholic.com/index.php?title=Holy_Year_of_Jubilee. The historical roots of this idea reach far back behind the mentioned development deep into the cultures of Mesopotamia, cf. Scheuermann (ed.), Das Jobeljahr im Wandel (2000); a detailed historical and theological interpretation is given by Bergsma, The Jubilee from Leviticus to Qumran (2007); see additionally A.Michel (ed.), Èthique du Jubilé – vers une réparation du monde?, 2005. 9 Cf. Krueger, A New Approach To Sovereign Debt Restructuring, available at: www.imf.org/ external/pubs/ft/exrp/sdrm/eng/sdrm.pdf; idem, International Financial Architecture for 2002: A New Approach to Sovereign Debt Restructuring, available at: www.imf.org/external/np/ speeches/2001/112601.HTM; Geithner/Gianviti, Proposed Features of a Sovereign Debt Restructuring Mechanism, available at: www.imf.org/external/np/pdr/sdrm/2003/021203.pdf; Hagan, Designing a Legal Framework to Restructure Sovereign Debt, 36 Georgetown Journal of International Law (2005), p. 299.

184 

 Christoph G. Paulus

The IMF’s efforts came to a sudden end in mid-2003 following the intervention of the US administration. The argument was that the attempt to restructure defaulting sovereigns should be left to market mechanisms. They were already being used in the form of so called Collective Action Clauses (CACs). Accordingly, instead of a statutory approach10 – as envisaged by the IMF – it was mandated that the contractual approach was to be followed. These clauses were to be included in indentures where they determine for the case of a specified event of crisis that the terms and conditions of the respective bond issuance are subject to change on a majority vote (rather than unanimity). Usually that majority is a qualified majority of 75 %. Since 2009, such CACs are admissible also for indentures under German law, cf. s. 5 SchVG. They have been refined in the course of time.11 Primarily, they serve the purpose of preventing creditors from torpedoing any restructuring attempt by insisting on full compliance with the previously agreed contractual terms – to be sure, they invoke thereby the fundamental legal principle of pacta sunt servanda. However, when and if such insistence crosses an imaginary level of shamelessness those creditors earn the label of ‘vulture fund’ – a term which provokes very unhappy precedences12. Their characteristic smell is that debts are bought on the secondary market at a high discount the full payment of which is thereafter sought from the debtor state with all legal power, pressure, and sophistication. However, in the meantime CACs have evolved and serve additional puposes beyond the repression of opposing creditors. Nevertheless, the present European reader will easily recognize that such CACs do not solve a problem which is highly visible in the cases of Greece, Ireland, Portugal and others: In most cases of a defaulting sovereign, a reduction of the debt burden alone does not cure the difficulties; what is regularly needed in addition, is a corresponding restructuring of the national economy, policy and social attitude.13 For this, CACs do not offer a solution.

10 Cf. Paulus, A Statutory Proceeding for Restructuring Debts of Sovereign States, RIW 2003, p. 401 ff. 11 On CACs, see just Bradley/Gulati, Collective Action Clauses for the Eurozone: An Empirical Analysis, available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1948534; Hofmann/ Keller, Collective Action Clauses, ZHR 175 (2011), p. 684. 12 Infamous examples are: Donegal International Limited v. Republic of Zambia and Anr., High Court of Justice in London, [2007] EWHC 197 (Comm); Kensington International Limited v. Republic of the Congo, [2007] EWHC 1632 (Comm), for this see Buchheit/Pam, The Pari Passu Clause in Sovereign Debt Instruments, 51 Emory L.J. 869 (2004); Elliot Associates L.P. v. Banco de la Nacion and the Republic of Peru, 194 F.3d 363 (2nd Cir., 1999). 13 In Germany, a striking example of the uselessness of mere financial support is the fiscal situation of the Länder Saarland, Bremen and Berlin. See also Bazinas/Sakkas, The Odyssey of a



A Resolvency Proceeding for Defaulting Sovereigns 

 185

III. European The development as described so far is not yet complete – as can be seen from the aforementioned countries. Whereas the discussion of the present topic has hitherto been (more or less) confined to a global perspective – more precisely: to a developing countries’ issue (the German coalition agreement from 2009 had exactly this context in mind when it included the concern to promote the institutionalisation of a sovereign debt restructuring tool) –, the Greek crisis has opened a new stage. This has its disadvantages, but it also comes with advantages. Regarding disadavantages, daily media information make them evident. For the European reader, it is important to recognize that sovereign default is not reserved for developing countries but is something that can happen in front of one’s own doorstep (or even within one’s own house). At the latest now it should have become clear that the search for a fundamental solution of this problem must not be left exclusively to the respective development aid ministry but should include various other resorts, such as finance, justice, economy, and maybe even foreign affairs. Regarding advantages, it sounds more cynical than it is meant to be – but here, too, the motto should be: never waste a solid crisis. After all, the Eurozone comprises a manageable number of states which can easier introduce and practice what appears, as of now, not yet tolerable for the globe as a whole. In a small group like the member states of the Eurozone, the disciplining effect of a legal procedure (for the details, see infra sub D) can more readily unfold than on a worldwide scale. It is on this assumption that the subsequently presented proposal is founded.

C. Some observations of the extra-legal environment Before that proposal’s presentation, however, some important information about the factual context of the problem shall be given which will hopefully make it easier to understand.

Greek Sovereign Debt Restructuring and the Role of Privatizations, INSOL World – Third Quarter 2011, p. 15 f.

186 

 Christoph G. Paulus

I. Terminology To begin this part with a terminological admonition, might come as a surprise; but long lasting experience makes clear that it is of utmost importance – at least outside the USA. The present topic is not one of dogmatic finesse in which each and every detail needs to be aligned to other dogmatic concepts; it is rather one that is supposed to serve a very urgent practical need – the sooner the better. Therefore, it appears to be indispensable to show respect and to take into consideration existing psychological sensitivities. Accordingly, in order to get around the negative connotations of the words “bankruptcy” or even “insolvency”14 a name must be found which is free from them but does, nevertheless, indicate what it is all about. It is proposed, therefore, to flag the proceeding as “resolvency proceeding”15 – thereby indicating that, in contrast to insolvency, the one and only goal is to get the debtor state back to solvency; an alternative would be to call it “restructuring proceeding”.16 As an aside, it is one of the consequences of those connotations that, irrespective of the usual disorientation in case of such crises and the consequential stumbling into ad hoc-solutions, the preferred way of dealing with a debt crisis is to enter into negotiations. After all, they can be conducted behind closed doors. This option, however, is usually not available for bond creditors to the degree that they are scattered all over the globe; this disadvantage might possibly lead one day in the nearer future to a renewed attractivity of borrowing within the official sector.17 From a specific German stance, however, it should be mentioned and recalled that the negotiation solution rescued Germany from its last insolvency:

14 Regarding the history of the stigma of bankruptcy, most recently Paulus, Ein Kaleidoskop aus der Geschichte des Insolvenzrechts, JZ 2009, p. 1148 ff. Even in the USA where the “flight into bankruptcy” had to be reduced through various statutory amendments, the stigma is felt and feared among many citizens, see Skeel, Bankruptcy Phobia, available at: http://papers.ssrn.com/ sol3/papers.cfm?abstract_id=1553366. 15 See already Paulus, Resolvenzrecht als Mittel zur Haushaltsdisziplin, Wirtschaftsdienst – Zeitschrift für Wirtschaftspolitik 2010, p. 793; idem, Die Eurozone als Probefahrt für ein Resolvenzrecht für Staaten, RIW-Editorial, 2010 Heft 9, p. 1. 16 The disadvantage of this terminology is that it was recently introduced as the name for a particular proceeding for rescuing certain credit institutions, cf. s. 7 ff. Kreditinstitute-Reorganisationsgesetz (KredReorgG). 17 Regarding Africa, for quite some time it is China that steps in as a bilateral lender, see Paulus, What Constitutes a Debt in the Sovereign Debt Restructuring Context?, in: Ligustro/Sacerdoti (ed.), Liber amicorum in onore di Paolo Picone, 2010, p. 231.



A Resolvency Proceeding for Defaulting Sovereigns 

 187

After all, the London Agreement on German External Debts18 led to a substantial haircut from the victorious powers without which the so called “Wirtschaftswunder” would not have taken place. Be it noted that despite the unquestioned merits of Hermann Josef Abs in these negotiations, it were those creditors that waived (better: stayed, at least initially) a big part of their claims against their former enemy – a fact which is not sufficiently remembered in Germany.

II. Three categories of lenders A problem observable with almost all modern defaulting states19 stems from the way they are raising capital (cf. already supra sub A). Speaking in general terms, sovereigns can borrow money from three different types of lenders: from the public sector such as other sovereigns or multilateral institutions like IMF or World Bank, from private banks, or/and from the so – called private sector. The latter covers the wide range from the man on the street who ties up his savings to big private investors such as pension funds, etc. For some 20 years, this sector is the predominant lender on the capital market. Whereas a sovereign’s default is taken care of in case of public sector lending through the Paris Club20 and in case of private banks’ lending through the London Club, there is nothing comparable for the private sector.21 This is one of the reasons for today’s big problems within the Eurozone – as it had been before, for instance, in Argentina 10 years ago. It is striking and can hardly be rationally explained why this kind of sectoral treatment of creditors does still prevail nowadays. What is needed is to keep in mind the entire picture and to look for a comprehensive solution. Since lawyers usually introduce novelties by adapting well-known existing legal instruments to the new circumstances and adjusting them to the new surroundings22 it appears to be legiti-

18 London Agreement on German External Debts from February 27, 1953, BGBl. II, p. 333 (see additionally Gesetz betr. das Abkommen vom 27. Februar 1953 über deutsche Auslandsschulden vom 24.8.1953, BGBl. II, p. 331); for this, e.g., Abs, Entscheidungen 1949–1953: Die Entstehung des “Londoner Schuldenabkommens“, 1991. 19 For an interesting list of possibilities for a state to hide debts, cf. Chr. Mayer, Greift die neue Schuldenbremse?, AöR, 136 (2011), p. 266, 309 ff. (“Innovative Formen der Finanzierung von Staatsaufgaben”). 20 Cf. the website of the Club de Paris: http://www.clubdeparis.org. 21 The Collective Action Clauses which, according to art. 12 par. 3 ESM Treaty, shall be introduced from July 2013 on can be seen, however, as a specific private sector mechanism. 22 Paradigm for that is the drafting of the XII-Table-legislation in early republican Rome, around 450 BC: Instead of sitting down and drafting the requested statute, the 10-men Board started its

188 

 Christoph G. Paulus

mate under the given facts to look for a possible model in existing insolvency law. After all, lawyers in this field have been dealing with the common pool problem for millennia and they have produced rules which might fit at least partially for the common pool problem of a defaulting sovereign. More on that infra sub D.

III. Biases A further, highly essential factor for success or failure of any efforts to establish a legal proceeding for defaulting sovereigns are political reservations. For various reasons, politicians as a whole seem to have a deep distrust against this idea. One of the reasons is probably the already mentioned flaw of insolvency, another could be the subcutaneous joy of bullying the debtor state. Still others might be: the fear of a chaotic reaction of all stakeholders and, consequently, the power of the rating agencies; and finally, the fear of the population’s loss of trust in the stability and reliability of sovereign bonds. As a matter of fact, all these concerns are understandable. But assume these concerns were given a positive twist! Thus, for instance, the said flaw could be interpreted as a welcome disciplining mechanism – a sovereign’s fear to be forced to apply for such procedure. The same could be said about the ratings downgrades by the agencies: today, the fear alone that such step will be taken has the potential – as can amply be observed in the European events in summer and fall 2011 – to paralyze hosts of politicians. In this context, it might be noted as an aside, that the repeatedly given hint to the US provenance of those agencies is misleading in that any European counterpart would rightly be distrusted when and if it came to any better result. The problem of the “big three” is that their judgments (or better: opinions) are the reduction of the capital market’s diversity of opinions to the monopoly-like three opinions. As a consequence, their opinions receive over-proportional attention and influence.23 For the facilitation of market transactions spirits were invoked which one cannot get rid of anymore. After all, the events in summer and fall 2011 do also demonstrate that those spirits are haunting politics – in a way which makes everyone clearly recognize about the advantage of taking the handling of such a crisis out of the hands of politicians and putting it into the hands of a neutral institution which complies with strictly transparent and predictable legal rules.

work with traveling to Greece for studying potential models, cf. Paulus, Verbindungslinien zwi­ schen altem und neuem Recht, Festschrift Mayer-Maly, 2002, p. 563 ff. 23 This statement does not imply necessarily that these consequences are actually intended by those agencies.



A Resolvency Proceeding for Defaulting Sovereigns 

 189

As to the “subcutaneous joy“, this is hardly worthy of support. But the fear of a chaotic reaction to the establishment of a resolvency system of which kind soever is to be taken seriously. A closer analysis, however, reveals that this argument intermingles future realities with the present-day situation. Of course, in the beginning there is likely to be a mess when and if the new set of rules were introduced here and now. However, it would be wrong to assume that this messy situation would last forever. And it is still another scenario when and if the envisaged new set of rules is announced to come into effect somewhen in the future; since then the stakeholders can adjust themselves to the new situation in advance and there will be a sort of floating transition period. The argument is also to be taken very seriously that the “man on the street” (another word for constituent) might be losing his or her confidence in the reliability and stability of state bonds. It implies, however, that a legal proceeding would blindly accomplish the equal treatment of all creditors and that, for this very reason, no further thoughts should be given to this idea. Even when leaving aside the implicit (upon closer inspection: strange) assumption of this argument that the investment into a state bond must in any case carry a return, even then this argument misconceives the flexibility of legal instruments. Take, for instance, the guarantee funds which protect small depositors. Additionally, there is even a specific insolvency protection (see infra sub D III 3) to put small creditors into one group and to give them full satisfaction of their claims.

IV. The role of law Finally and regrettably, there is one more issue to be addressed in the present extra-legal context. On first sight, it looks like a banality – namely the necessity to comply with legal rules! However, even leaving aside the more or less publicly announced statements of various leading politicians all over the world that they couldn’t care less about what the law tells them to do, the way how the European politicians dealt with and argued around the rules of artt. 119 ff. TFEU which were meant to provide for a bail-out prohibition, gives rise to a fear that rules are more and more meant to be a sort of “symbolic legislation” – i.e. legislation the compliance with which is derogated with a wink of the eye. To the degree that this sort of understanding gains ground, law and its steering function are undermined. A law which is enacted but not complied with, is not worth the paper on which it is written. In other words, the operating efficiency of any legal and juridic proposal is entirely dependent on the addressees’ acceptance; if they refrain from complying, the legal side of the game is over.

190 

 Christoph G. Paulus

D. A proposal for a proceeding To avoid the creation of a model which has no chance of becoming accepted in the political reality with all its rough edges, each one of the preceding extra-legal issues needs to be kept in mind. Following the example of the abovementioned (fn. 22) ancient Roman Board of Ten Men, the model to be elected is the modern commercial insolvency law.24 After all, this field of law has changed its appearance tremendously in the last decade or so – away from the solely liquidating mechanism towards a highly sophisticated and effective tool for restructuring the debtor. Whereas the former is absolutely unbearable in the sovereign context, the latter is all the more at least worthy of being seriously considered. I.e. not the “moneyfication” of the debtor’s estate is at stake, but just and exclusively the reconstitution of the debtor’s debt sustainability. For that purpose, the modern commercial insolvency law has been offering for roughly 40 years by now the option of the Chapter 11 proceedings (USA) or “Planverfahren” (D) respectively.25 Irrespective of numerous necessary modifications and adaptions, it is this option which can be used as a model for dealing with the situation of a defaulting sovereign.26 At the core of this “transplant“27 is the introduction of a court-like institution. To begin with this rather procedural step resembles a somewhat comparable innovation in the sixties of the 20th century when the debates about pros and cons of an international investment protection were heated. The problems appeared at that time unsolvable due to the antagonistic interests involved but yet a solution was found – namely the creation of the International Centre for the Settlement of Investment Disputes (ICSID) at the World Bank. This is simply an arbitration platform for disputes arising out of investment disputes which offers only procedural rules and support but no substantive law. As a consequence, the

24 The following proposal is a modification of a broader model which the present author is about to publish together with Steve Kargman (New York) and which is meant to address a global, i.e. not just European, audience: Reforming the Process of Sovereign Debt Restructuring: A Proposal for a Sovereign Debt Tribunal as a Means of Resolving Sovereign Debt Disputes. 25 It is of more than just academic interest that at this point of time when Europe is about to get acquainted with the idea of a resolvency proceeding for states, in the USA a heated debate has started about need or nonsense of introducing a bankruptcy procedure for their member states. See, e.g., Skeel, State Bankruptcy from the Ground Up, available at: http://ssrn.com/ abstract=1907359. 26 Moreover, the US Chapter 9 proceedings teach how to design a procedure which does not interfere with the continuing functioning of the debtor’s sovereignty. 27 For a historical description and interpretation of “legal transplants” see A.Watson, Legal Transplants: An Approach to Comparative Law, 1974.



A Resolvency Proceeding for Defaulting Sovereigns 

 191

arbitral awards rendered under the auspices of this Centre28 add nowadays up to a considerable body of public international investment protection law.29 This might be a precedent for the evolution of a sovereign debt restructuring law under the auspices of a respective institution.

I. Credible Threat Following the deliberations offered supra (sub C IV) about the absolute validity of law, it should be noted right at the outset that the envisaged resolvency proceeding displays its efficiency then and only then when everyone is ready to apply it. If this is the case, i.e. if this remedy is understood to be a credible threat, it is to be assumed that it will have a disciplining effect. After all, which head of government is willing to state in front of numerous cameras that his or her country has to file a petition at the resolvency court? Thus, if every actor on that stage is ready to comply with the legal proscription of not bailing out other member states – in any form30 – it follows that a resolvency proceeding is most effective if it is never applied. The strategy is, thus, comparable to the historical political motto “peace through deterrence”.

II. Plan proceeding as model Claiming that the envisaged proceeding must be solely directed towards the debtor’s turnaround (resolvency) is an indication for any insolvency lawyer that something like the German “Planverfahren” (plan proceeding) shall serve as the model. It is worth remembering that this plan proceeding was drafted after the US Chapter 11 proceedings which not only has a modified copy in the Chapter 9 pro-

28 As to the question of the role of precedents in this context cf. Kaufmann-Kohler, Arbitral Precedent: Dream, Necessity or Excuse? Arb. Int’l 23 vol. 3, 2007, p. 357. 29 See, e.g., Griebel, Internationales Investitionsrecht, 2008; Herdegen, Internationales Wirt­ schaftsrecht, 6th ed., 2007, p. 238 ff.; Krajewski, Wirtschaftsvölkerrecht, 2006, p. 167 ff. See additionally Tietje, Grundstrukturen, Rechtsstand und aktuelle Herausforderungen des internationalen Investitionsschutzrechts, in: Giegerich (ed.), Internationales Wirtschafts- und Finanzrecht in der Krise, 2011, p. 11 ff. 30 If it is deemed to be indispensable for political reasons, art. 122 TFEU could be instrumentalized here. However, it cannot be overstated that such support must be dependent on precise preconditions which should be as interpretation-proof as possible.

192 

 Christoph G. Paulus

ceedings (a quite successful procedure for rescuing defaulting municipalities)31 but which also was copied by many jurisdictions around the globe – particularly by Germany’s neighbouring states such as England, France, Italy, etc. Accordingly, because of this multiplicity of variants a specification is necessary before details of the envisaged proceeding can be displayed (III).

1. Small solution The IMF’s attempt – or rather its failure – to introduce fully-fledged insolvency proceedings seems to indicate that the time is not yet ripe for such an ambitious step. Obviously, one has to be more humble and should leave unresolved all substantive issues. Instead, a concentration on more or less mere procedural issues appears to be the right way to go. For example, the questions whether or not the commencement of a resolvency proceeding should be accompanied by an automatic stay, whether or not certain contracts should be rejectable, whether or not certain pre-commencement transactions should be subject to avoidability touch so many highly intricate issues that – at least for the time being – one better does without. After all, any increase of complexity reduces the likelihood of any idea’s realisation – particularly when it is located in a legal, political, and economic minefield like the present one.

2. Potential expansion to the entire globe The comparison made supra (sub D before I) with the evolution of ICSID to a public international investment protection law carries with it a further implication. As in that case, a newly created resolvency proceeding should have the potential to expand from its original European application field to a global one. Even if this sounds banal: it should be remembered that the sovereign default problem is not a European phenomenon. And it is also a seemingly banal con-

31 Cf. Naguschewski, Kommunale Insolvenz – Untersuchungen zu einem Insolvenzverfahren nach Vorbild des US-amerikanischen Chapter 9, 2011, p. 27 ff. For this context, see additionally v. Lewinski, Öffentlichrechtliche Insolvenz und Staatsbankrott – Rechtliche Bewältigung finanzieller Krisen der öffentlichen Hand, 2011; Hornfischer, Die Insolvenzfähigkeit von Kommunen, 2010; Cranshaw, Insolvenz- und finanzrechtliche Perspektiven der Insolvenz von juristischen Personen des öffentlichen Rechts, insbesondere Kommunen, 2007; Frielinghaus, Die kommunale Insolvenz, 2007; Paulus, Überlegungen zur Insolvenzfähigkeit von Gemeinden, ZInsO 2003, 869; idem, Insolvenzfähigkeit von Gemeinden – pro, NordÖR 2010, 338.



A Resolvency Proceeding for Defaulting Sovereigns 

 193

clusion that for all other extra-European sovereigns fighting with the problem of default the Bretton Woods Institutions may not offer sufficient help. Quite to the contrary, the need for a global resolvency proceeding is ubiquitous and urgent. Accordingly, the following deliberations are meant (and should so be read) to include the option for such an expansion – irrespective of the then necessary modifications.32 It is, thus, the ambition of the subsequent proposal that it could also be used as a blueprint for a global sovereign default restructuring mechanism.

III. Details of the proceeding The entirety of the preceding deliberations and precautionary admonitions make it advisable to establish a court-like institution – Sovereign Debt Tribunal (SDT)33 – which is in charge of managing, monitoring, and directing the proceeding. The specific peculiarity of this approach is that not only those assets remain entirely unaffected which are indispensable for the functioning of the sovereign’s tasks but also that the responsibility for the resolvency’s success or failure rests completely on the parties involved. After all, a plan proceeding is – like the Chapter 11 proceedings – nothing but an invitation to negotiation to those parties; what they do with this invitation and how they act after its acceptance or rejection is their business. But they have the chance to reach a common ground in such plan which then serves as the basis for the future relations between debtor and creditors. As a caveat it should be noted that the following deliberations are not meant to offer a definite solution. The present author is rather aware that the issues at stake are so complex, intricate, and multidimensional that these scribblings cannot present much more than just initial thoughts and ideas.

32 Cf. Kargman/Paulus (fn. 24). 33 There do exist prominent models for such a tribunal (outside of ICSID) – see, e.g., the IranUnited States Claims Tribunal, cf. Gibson/Drahozal, Iran-United States Claims Tribunal Precedent in Investor-State Arbitration, Suffolk Univ. L.S., Legal Studies Research Paper Series 07–15, April 3, 2007; but see also the less prominent equivalent for the restructuring of the Saddam-era debts of Iraq, cf. Deeb, Project 688: The Restructuring of Iraq’s Saddam-Era Debt, Cleary Gottlieb Restructuring Newsletter Winter 2007, p. 3 ff.

194 

 Christoph G. Paulus

1. Contractual approach The debate around the IMF’s SDRM was split into those who advocated for a statutory approach and those who preferred a contractual one.34 A bit simplifying, this conflict addresses the basis for the legally binding character of the respective approach. The IMF, for instance, had in mind to achieve such binding effect through an amendment of its Articles of Agreement; others preferred the implementation of a narrowly pre-defined model law into the national body of law.35 Both proposals fall into the category of a statutory approach which, for several reasons, offers all-in-all preferable solutions. However, its decisive drawback is the big effort and time needed for its realization. The contractual approach, most prominently represented by the abovementioned (sub B II) Collective Action Clauses leaves it to the respective sovereigns to determine the procedure in all its details. Irrespective of some considerable deficiencies, this approach seems to be – at least for the time being – the better way; after all, it allows for an expedited access to workable solutions for an orderly resolvency proceeding.36 There is no need to enter into an international agreement or to enact a statute; instead, it suffices to include certain contract clauses into loan agreements. In other words, all loan and bond agreements of a sovereign should be amended by one (or more) contract clause(s) which provide(s) for the following: In case of default and a subsequent petition of that sovereign, a certain court-like institution will be in charge of the proceeding. The details could either be included in that contractual clause or (probably more transparent) become part of that institution’s rules of procedure.37 In the latter case, the contractual clause would be brief and would just establish the institution’s competence to deal with the procedure in case of the sovereign’s default.

34 For what follows cf. Paulus, (as in fn. 10). 35 Paulus, Rechtlich geordnetes Insolvenzverfahren für Staaten, ZRP 2002, p. 383 ff. 36 Cf. Malaguti, Sovereign Insolvency and International Legal Order, Int’l Comm.L.R. 11, 2009, p. 307 ff. 37 That court-like institution should draft the rules on its own; as to the permissibility of such autonomous rule-setting, cf. just art. 30 par. 1 of the ICJ Statute; art. 16 of the Statute of the International Tribunal for the Law of the Sea; or art. 3 par. 2 of the Claims Settlement Declaration of the Iran-US-Tribunal.



A Resolvency Proceeding for Defaulting Sovereigns 

 195

2. Sovereign Debt Tribunal If the inclusion of such a clause in a sovereign’s loan agreements constitutes the first step on the way to come to a functioning resolvency regime, the second one is to establish the court-like institution which might be called Sovereign Debt Tribunal (SDT). The third – and ultimate – step will be the drafting and enactment of the procedural rules (either by means of elaborate contract clauses or by the SDT itself). Given the fact that those judges should rapidly develop expertise the pool of potential judges should be limited right from the outset.38 This does not exclude that these experts remain in their professional positions and act as judges only when and if the president of that SDT appoints them as part of the respective three-judges-panel. The advantage of this approach is that it reduces the tribunal’s current costs; they are to be paid only when acting as judges and they are then paid by the debtor sovereign. The only permanent staff would be the president and his office. The selection of the judges’ pool must be guided by various diversification criteria, i.e. different nationalities, professions, and backgrounds are to be found. The experience with the IMF proposal a decade ago show that the placement of the SDT is an important issue. It is, thus, not acceptable to connect the tribunal with any of the existing credit institutions such as the European Central Bank in Frankfurt a.M.; the same is true also for any of the Brussels Institutions as they are (or appear to be) guided by specific interests. Under these circumstances, it would be preferable to have a special and independent chamber established at the European Court of Justice in Luxembourg. However, it is to be feared that this amendment would take quite some time; accordingly, under the present circumstances it might be the best solution to have the SDT connected with the Permanent Court of Arbitration in The Hague. Not only was such an idea actually circulated at the time of its foundation – i.e. some 120 years ago –, a similar concept was also discussed only very recently by the Dutch government.

38 The IMF’s SDRM had provided for a similar selection mechanism, cf. Paulus, Die Rolle des Richters in einem künftigen SDRM, in: Gerhardt/Haarmeyer/Kreft (ed.), Insolvenzrecht im Wandel der Zeit, Festschrift für Hans-Peter Kirchhof, 2003, p. 421.

196 

 Christoph G. Paulus

3. The procedural steps For a better understanding of the proposal to be submitted here, it appears to be advisable to list beforehand the procedural steps of the envisaged resolvency proceeding; they resemble in various aspects the procedure of a “plan proceeding” or Chapter 11-type of proceeding as it is established and applied in many jurisdictions. The first step is the application by the debtor country (cf. infra sub 5); this must be accompanied by the presentation of a restructuring plan which describes meticulously how the debtor imagines the restructuring of its country, i.e. the plan must explain which concessions are requested by the creditors and which contributions the debtor itself is ready to undertake. As an aside, this is one further advantage of the present proposal over the CAC solution that it is built on a fully-fledged approach rather than just the creditors’ side. The plan proposal of the debtor is, as a matter of fact, subject to manifold changes, adaptations, and amendments in the subsequent discussions with the creditors; but at the beginning it serves as a kind of filter: since it is the SDT’s, i.e. the freshly appointed three judges’ first task to examine the feasibility, fairness and reasonableness of this draft plan. The creditors – and this means, generally speaking,39 all creditors40 of that sovereign – are packed together into classes (or groups, pursuant to German insolvency law terminology) which, of course, has to be done in compliance with rational and verifiable criteria. This is an essential feature of the proposed proceeding since discussion about and final voting on the plan will be through the groups. Even if the consent of each single group should be needed, such group formation implies that not every single creditor needs to concur; it rather suffices that the (however qualified) majority of, e.g., 66 %, 75 %, or 85 % of the creditors within one group do so. Thus, if only a simple majority within a group was required, the consent of 50 % + 1 would be sufficient. However, before it comes to the voting the debtor and its creditors must discuss the proposed plan. In order to reduce the mass of creditors and to enable meaningful discussions, one might consider allowing for the appointment of special representatives – as provided for in many modern Debenture Bond Acts (cf. § 7 SchVG). As a matter of fact, the parties involved in these discussions will

39 As to the problem of who might qualify as a creditor in a sovereign default, just see Paulus, (as in fn. 17). 40 This is in order to prevent a segmentation as it is done by the Paris Club or its London equivalent.

A Resolvency Proceeding for Defaulting Sovereigns 

 197

negotiate toughly; it is to be assumed that every side will strongly argue for its own benefit. A resolvency regulation is well advised, however, to abstain from any substantive prescriptions; the result should be left to the balance of powers. This is at least true as a general rule of thumb. This does not exclude – in order to come back to a previously mentioned concern of many critics of sovereign debt regulation – that different groups are treated differently. It would, thus, be possible to pack small creditors into a separate group which will receive 100 % satisfaction whereas the other groups of, e.g., institutional creditors accept a “haircut” of 50 %. The flexibility of potential solutions in this context is as large as the contractual freedom allows for adaptations to the individual case. It is not necessarily mandatory that these negotiations are connected with those which the debtor country is likely to conduct at the same time with potential new lenders; after all, in the situation envisaged here, the debtor will be in dire need of fresh money. The primary candidate for those negotiations will be the IMF or the ESM; but depending on the acceptability and the reaction of the capital market to the commencement of the resolvency proceeding, it cannot per se be excluded that the debtor is able to get money from the capital market (bonds) at tolerable interest rates. In either case the lender might be interested to participate in the plan negotiations. This would increase the disciplining effect on the debtor. At a certain point in time (cf. infra sub 7), the voting must be done. As a matter of fact, the debtor is dependent on the creditors’ majority’s consent and is, accordingly, again under a certain disciplining pressure since without an accepted plan, the debtor’s debt situation does not change. This dependency on the creditors is a kind of compensation for what will be described in more detail subsequently (sub 5): namely the exclusive right of the debtor to pull the trigger for commencing the procedure. However, this dependency could be mitigated by reducing the requirement of unanimity regarding the groups’ consent (attention! not the voting requirements within any one of the groups). If so wished one could think along the lines of the so called “cram down rule”, cf. s. 245 InsO, which allows under certain circumstances a plan to be accepted if the (simple or qualified) majority of groups do concur. If the plan is, thus, accepted the court must confirm it by examining the legal correctness of the proceeding so far. Given this requirement, it is advisable to have the court present all the time during the negotiations. These judges should probably function as moderators. If, however, the necessary majority for the plan’s acceptance is not achieved a second chance should be granted for improving repair; i.e. re-negotiations should be possible, although just for a limited period of time. If this second attempt also fails the European Monetary Union (for Eurozone debtors) could possibly provide for a whole range of sanctions up to the exclusion of that particular sov-

198 

 Christoph G. Paulus

ereign from the Union – which would likely be the most severe of such sanctions. However, the failure of the plan’s acceptance could likewise be caused by a creditor who tries to “hold out”. Due to the court’s moderation a situation like that could be sanctioned by withdrawing that creditor’s voting right or interpreting its vote as a “yes” after special investigation of the result’s fairness and reasonableness. Thereby, both sides could be disciplined.

4. Reason for commencement of procedure A fundamental issue of any proceeding like the one presented here is, as a matter of course, under which conditions the resolvency proceeding shall become commenced. The commercial insolvency law of, e.g., the German Insolvency Ordinance, s. 17 to 19, provides three possible reasons: insolvency, imminent insolvency, over-indebtedness. None of them is transferable to the realm of sovereign default and a resolvency proceeding. The IMF imposed in this context the requirement of “unsustainability of debts”, meaning that the debt burden has become too high to reduce the principal amount and that the debtor is cought in the debt trap. To be sure, this is just a very superficial description of what the IMF has in mind. But one might ask if there is any need at all for pinning down a precise reason for opening a resolvency proceeding. It might suffice to have instead a subsequent abuse control.41 As will be discussed in more detail below (sub 6 a), it is one of the court’s significant tasks to examine the commencement prerequisites. When and if they are not present the proceeding will not be commenced and the debtor country is obliged to look for alternative solutions – an intentionally unpleasant alternative to making proper use the procedure.

5. Right to petition The IMF’s proposal of the SDRM reserved the right to petition for the commencement of the proceeding exclusively to the debtor country. This was the correct approach as such restriction is indispensable with respect to the debtor’s sover-

41 Cf. Paulus (as in fn. 35), p. 385. Another alternative would be, following an English model, to leave the definition of the reason for commencement of procedure to the parties, cf. Steffek, Insolvenzgründe in Europa – Rechtsvergleich, Regelungsstrukturen und Perspektiven der Rechts­ angleichung, KTS 2009, 317, 348, 352.



A Resolvency Proceeding for Defaulting Sovereigns 

 199

eignty; otherwise a creditor’s petition would urge the debtor into a straitjacket of actions which, in the worst case, are not needed at all. Moreover – and probably most importantly – the likelihood of such option’s general political acceptance is minimal at best. Germany, France, or any other country would hardly ever agree to submit themselves to such a right of others.42 However, in light of what has earlier been said about the advantage of any insolvency procedure – namely its disciplining function for all stakeholders just by means of its mere existence – the present proposal might appear as unilaterally favouring the debtor’s position. Since if the debtor alone has the right to pull the trigger, it has therewith a bargain chip for its negotiations, for instance, at the Paris Club, London Club, or its negotiations with the private sector; the creditors have nothing to hold out against this. Nevertheless, this imbalance is for the reasons presented to be tolerated and must be compensated by other measures disciplining the debtor – for instance the abovementioned (sub 3) need for the creditors’ consent to the plan.

6. Competences of the SDT In this section, too, it should be noted right at the outset that what is intended at this initial point of the evolution of such a proposal is to give a non-exhaustive overview of potential competences. Further refinement will be needed in due course. a) Being mindful of this restraint, it is fair to conclude that the court’s single most important task is to examine the reason for commencement of proceedings and the potential abuse (see supra sub 4) by the petitioning sovereign. The court must accordingly review the debtor’s justification given in its plan for the commencement of the proceeding. Note that conferring this task upon the court underscores the care and attention needed for the selection of the judges’ pool; intimate political, economical and legal knowledge is indispensable for such an examination. Since what is to be verified is the debtor’s claim that all existing sources of income and other means have been exhausted. The options here at stake are innumerable: (further) privatizations43 might be as possible and rea-

42 The same is true for many other countries such as the USA – if the proposal were to become extended to a global model. However, for the Eurozone it appears to be imaginable to give the right to trigger the proceeding to a supranational institution such as the European Council. 43 Regarding the privatizations particularly of present Greece, see the very informative deliberations by Kargmann/Potamitis, A New Approach to Privatisation: An Unexplored Option for

200 

 Christoph G. Paulus

sonable as increasing certain taxes, saving options might be available by cutting salaries in the public sector (for instance, cutting back the 13th month’s salary), certain commodities might not yet have been (fully) exploited, or the gold reserves might be available – and be it partially – to beings old, etc. To get an idea of the extent of the possibilities it could be helpful to study the conditionality catalogs of the IMF or World Bank which these institutions have previously set up for borrowing states.44 b) The verification of claims might also be seen as one of the key tasks of the court. But the devil is here, too, in the detail: Which claims are to be included in the resolvency proceeding? Just foreign ones or also domestic ones? Claims just against the state or also those against the national bank, the subdivisions such as Länder, provinces, regions, etc., or state owned enterprises? Only contractual claims or also those based on other grounds?45 The preferable (but, of course, not only possible) answers to these selected questions are: Bearing in mind the need to overcome the selectivity of the Paris Club or London Club and to strive for an all-encompassing resolvency proceeding all claims should be included – be they foreign or domestic (possibly including the domestic tax- or wage claims); the debtor should include all those entities which have no separate legal existence (as for those which have such separate existence, the general insolvency law will be applicable); and there should be a restriction to contractually founded claims because for the time being the court’s competence can be founded only on a contract (see supra sub 1). c) It is by no means a marginal task of the SDT to moderate the negotiations (see supra sub 3) and to check the legitimacy of the group formation, i.e. whether or not objective, coherent criteria have been applied. The goal of this latter task is to prevent the debtor state from strategically binding all creditors together in one group which are likely to reject the plan. Accordingly, this task is one further piece of the puzzle to discourage the debtor state from abusing the proceeding. d) A final competence of the SDT could be the dispute resolution within the proceeding. The details should be elaborated by taking inspiration from existing commercial insolvency jurisdictions such as Austria or the US. They confer far reaching competences upon their insolvency courts regarding the resolution of

Greece in Privatising Troubled State-Owned Enterprises, International Corporate Rescue 2011, p. 389. 44 It is more than likely that cooperation of the judges is necessary at this point of the proceeding both with Eurostat and IMF; since these institutions have unique sources of information for the debtor’s economy. 45 For this, see just Hagan (as in fn. 9), p. 299 ff.



A Resolvency Proceeding for Defaulting Sovereigns 

 201

disputes between the parties. Such a concentration (in the commercial context labelled as “vis attractiva concursus”46) serves the purpose of the proceeding’s acceleration and streamlining. In addition to the dispute resolution, one might also think about permitting the judges to serve as mediators or conciliators – thereby making references to the respective powers of ICSID tribunals.

7 . Timing The time factor is of eminent importance in any restructuring and, accordingly, in any resolvency proceeding as well. Therefore, it is key to take this facet into consideration and to provide for rather strict time frames in order to prevent strategic abuse by either side in prolonging or abbreviating the procedure to one’s own benefit. However, it is to be assumed that, in most cases, it will be the debtor country which pushes for acceleration; since the earlier the plan is accepted the earlier the state can begin with the realization of the resolvency measures (see infra sub 8). Based on this assumption the focus of timing rules should be on disciplining the creditors. Those rules could be, for instance, the right to ad hocinterventions of the judges or fixed time frames after which the majority requirements could change.47

8.  Plan realization If the plan is accepted by the prescribed majority vote and if the court has ultimately certified the legality of the procedure, this court order will be the basis for all subsequent legal changes and obligations arising from the plan. The creditors’ “haircuts” are effective as of this time as are the debt deferral agreements, etc. But as of this time the debtor state, too, is obliged to begin with all those measures which, pursuant to the plan, are to be done by the debtor – for instance, cutting back of salaries, privatization operations, exploring new (or increasing existing) taxing sources, etc.48

46 See just Paulus, Insolvenzrecht, 2nd ed., 2012, par. 47. 47 If the debtor state should abuse such time frame the court could be allowed to prolong the deadlines. 48 Some of these operations have already been mentioned supra (sub 6 a). This is not so surprising in light of the possibility that the political chances for their realization might be greater when done in the context of the resolvency plan – i.e. after the creditors’ consent to do their part for the sovereign’s restructuring.

202 

 Christoph G. Paulus

However, it is possibly a not entirely unlikely scenario that the debtor, once the plan has been accepted and the debt burden is accordingly reduced, becomes somewhat hesitant or less enthusiastic to comply with the imposed obligations of the plan: all of a sudden, the promised tax increase shall be postponed until after the elections next year, or the sale of certain shares is “unfortunately” presently not appropriate, etc. In such a case, the creditors would have recourse to the courts (whichever these might be in the given case) which, even if successful in the end, would take months if not years. Therefore, effective resolvency rules should provide for ongoing supervision by the court in this plan-realization-period in order to discipline the debtor here, too. The sanction against such retardedness of the debtor could be modelled pursuant s. 255 InsO: the court could be given the power to revoke the plan – with the consequence that the status quo ante is re-established and all claims exist as they had before the plan acceptance.

IV. Political Facts It is tempting to set the present proposal in correlation to the European Stability Mechanism (ESM) as established following the adoption of the European Council Decision 2011/199/EU. A number of similarities exist in both concepts. For instance, the ESM shall be an institution with its own legal personality that has, pursuant to art. 31 par. 1 ESM Treaty, its seat in Luxembourg; the ESM shall cooperate closely with the IMF, deliberation 8; the ESM shall undertake a rigorous analysis of the public-debt sustainability, art. 13 par. 1 lit. b ESM Treaty; the private sector can possibly be included, deliberation 12; Collective Action Clauses shall be included in future debt instruments, art. 12 par. 3 ESM Treaty; and the institution shall be granted a preferred creditor status, deliberation 13. All these are elements which are, more or less, also to be found in the present proposal for a resolvency proceeding. The main difference, however, between the ESM and a resolvency proceeding is that the former concept leaves considerable room for political intervention whereas the resolvency proceeding is guided primarily by the “blindness” of a legal procedure. Once set in motion, the action is in the hands of neutral judges with little if any potential for influence from the political side.

V. Once again: problems of the transition period It has already been mentioned (supra sub C III) that there exist enormous reservations regarding the introduction of a resolvency proceeding (in whichever form).



A Resolvency Proceeding for Defaulting Sovereigns 

 203

However, the arguments brought forward in this context often intertwine the existing situation with the transition situation and the future situation. Regarding the existing situation, it should suffice to refer to what happened in spring 2010 when the Greek crisis became evident. There was obvious need for rapid action which, for several – primarily political – reasons, could not be performed. At this occasion – and ever since – it became obvious how advantageous it is when there is a transparent procedure at hand that prescribes which steps are to be undertaken. As a general observation, it is no over-statement that the outstanding advantage of any legal procedure is that it provides guidance and structure in a chaotic situation.

1. Chaos In the transition period, established rights might appear to be threatened; creditors are more than likely to call for constitutional, political or other help. A protection against this scenario is to introduce the proposal of the resolvency proceeding not before 2013, i.e. the time when anyway Collective Action Clauses shall be introduced, art. 12 par. 3 ESM Treaty. This preparation period is helpful also for all agreements on a credit-default swap (CDS); the “insurer” has time to weigh the risk that might be at stake.49 Apart from these scenarios, it is by no means a peculiarity of the present proposal that a new legal regime has to provide special rules for a transition period. They might address, as one example among many, the necessary adaptations of commercial balance sheets or tax laws: the question whether or not the commencement of a resolvency proceeding leads to a total write-off of credit institutions’ claims needs to be answered one way or the other. In this transition period it is fair to predict that the decision making of the court will be based on shaky ground. After all, precedents do exist if at all only in a very limited amount. However, this uncertainty is a consequence of the introduction of any new area of law; it can, nevertheless, be mitigated to a certain degree by the authority and integrity of the deciding judges.

49 The situation is probably more complicated when a country retroactively introduces such rules. As to the possibility to change retroactively Collective Action Clauses, see Buchheit, Restructuring a Nation’s Debt, IFLR June 2011, p. 46.

204 

 Christoph G. Paulus

2. Structure Once the resolvency proceeding is well established, it is to be assumed that the resolvency proceeding’s disciplining function will come into full effect. At least some sovereigns will have increased difficulties to get access to credit, and the creditors will be more diligent in selecting those to whom they lend. Presumably, creditors will demand more security which, in turn, might reduce the average total lending. This, however, would be a great success of the resolvency proceeding’s introduction – after all, the volume of credit granted is presently the main concern of many countries around the globe, including Germany.50 For the sake of completeness, it shall be added that the addressees of these disciplining functions will rather sooner than later look for loopholes to get around the consequences of the resolvency proceeding or to make it work for their individual benefit. It is, however, at this point in time too early to search for solutions for these future problems. Forces should be concentrated to get this type of proceeding established after thousands of years of experience with failing states which have, in most cases, lead to nothing but ad hoc-solutions.

E. Conclusion It should have become clear by now that the time is ripe to come to a transparent and predictable solution for overindebted states after innumerable (and highly cost-intensive) ad hoc-reactions – a solution which complies with the modern requirements of the rule of law.51 Even though this is a global mandate, it presents itself in these days most urgently within the Eurozone. The preconditions for the creation of a respective proceeding are here as favourable as the need for it is pressing. The fears usually connected with the introduction of such proceeding can be minimized; all that is presently needed is the humbleness not to strive for a big solution. Taking this for given, the legal side of the solution how to deal procedurally with the problem of defaulting sovereigns is rather uncomplicated. All that

50 Imagine the scenario that the presently exceptionally low interest rates become higher. Then easily the budget item “debt redemption” becomes the single biggest item in the entire budget. 51 Cf. Buckley, The Bankruptcy of Nations: An Idea Whose Time Has Come, 43 The International Lawyer, p. 1189 (2009). Regarding the requirement “rule of law” cf. Paulus, A Standing Arbitral Tribunal as a Procedural Solution for Sovereign Debt Restructurings, in: Braga/Vincelette (ed.), Sovereign Debt and the Financial Crisis – Will This Time be Different?, 2010, p. 317.



A Resolvency Proceeding for Defaulting Sovereigns 

 205

is needed is for a contractual clause to be included in all future loan and bond agreements which provides for the competence of a court-like institution – herein called Sovereign Debt Tribunal (SDT) – for dealing with that sovereign’s default and defines the circumstances in which such default exists. What is important is that the handling of a scenario like the Greek one since Spring 2010 is taken out of the hands of the politicians and instead is put into those of a neutral institution composed of experts. The solution presents itself that this clause should be made obligatory from 2013 on when, all Euro-denominated bond issuances shall include a Collective Action Clause. The creation of such a neutral institution, the SDT, is less complex and costly than it might appear on first sight. Moreover, the creation of a procedure52 could also be quite uncomplicated. In a nutshell, it must provide for the following steps: the right to initiate it must be left with the defaulting sovereign; the application must be accompanied by a detailed restructuring plan which is initially examined by the three judges who will be appointed by the SDT’s president at the time of the application. If the result of that initial examination is that the application is abusive, the application will be rejected; otherwise this plan will serve as the basis for subsequent discussions with the creditors and will, thus, be subject to amendments and changes. There must be a time limit for these discussions after which the vote of the creditors is cast over this – amended and changed – plan. If the result is negative, the sovereign should be given a second chance for adjustment of that plan within a restricted period of time. If, however, the plan is accepted by the required majority the fulfillment of the obligations should be supervised by the judges. If the sovereign does not comply thereby with the terms of the plan, upon determination of the judges the plan should be declared to have lapsed – with the consequence that the original obligations re-emerge.

52 The procedure can be delineated either in the contractual clause or – preferably – in the court’s self-imposed rules.

Appendices

Appendix 1 Art. 12(3) ESM Treaty Collective action clauses shall be included, as of 1 January 2013, in all new euro area government securities, with maturity above one year, in a way which ensures that their legal impact is identical.

Appendix 2 EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause Common Terms of Reference 1. General Definitions (a) ‘debt securities’ means the Bonds and any other bills, bonds, debentures, notes or other debt securities issued by the Issuer in one or more series with an original stated maturity of more than one year, and includes any such obligation, irrespective of its original stated maturity, that formerly constituted a component part of a debt security. (b) ‘zero-coupon obligation’ means a debt security that does not expressly provide for the accrual of interest, and includes the former component parts of a debt security that did expressly provide for the accrual of interest if that component part does not itself expressly provide for the accrual of interest. (c) ‘index-linked obligation’ means a debt security that provides for the payment of additional amounts linked to changes in a published index, but does not include a component part of an index-linked obligation that is no longer attached to that index-linked obligation. (d) ‘series’ means a tranche of debt securities, together with any further tranche or tranches of debt securities that in relation to each other and to the original tranche of debt securities are (i) identical in all respects except for their date of issuance or first payment date, and (ii) expressed to be consolidated and form a single series, and includes the Bonds and any further issuances of Bonds. (e) ‘outstanding’ in relation to any Bond means a Bond that is outstanding for purposes of Section 2.7, and in relation to the debt securities of any other series means a debt security that is outstanding for purposes of Section 2.8. (f) ‘modification’ in relation to the Bonds means any modification, amendment, supplement or waiver of the terms and conditions of the Bonds or any agreement governing the issuance or administration of the Bonds, and has the same meaning in relation to the debt securities of any other series save that any of the foregoing references to the Bonds or any agreement governing the issuance or administration of the Bonds shall be read as references to such

212 

 Appendices

other debt securities or any agreement governing the issuance or administration of such other debt securities. (g) ‘cross-series modification’ means a modification involving (i) the Bonds or any agreement governing the issuance or administration of the Bonds, and (ii) the debt securities of one or more other series or any agreement governing the issuance or administration of such other debt securities. (h) ‘reserved matter’ in relation to the Bonds means any modification of the terms and conditions of the Bonds or of any agreement governing the issuance or administration of the Bonds that would: (i) change the date on which any amount is payable on the Bonds; (ii) reduce any amount, including any overdue amount, payable on the Bonds; (iii) change the method used to calculate any amount payable on the Bonds; (iv) reduce the redemption price for the Bonds or change any date on which the Bonds may be redeemed;1 (v) change the currency or place of payment of any amount payable on the Bonds; (vi) impose any condition on or otherwise modify the Issuer’s obligation to make payments on the Bonds; (vii) except as permitted by any related guarantee, release any guarantee issued in relation to the Bonds or change the terms of that guarantee;2 (viii) except as permitted by any related security agreement, release any collateral that is pledged or charged as security for the payment of the Bonds or change the terms on which that collateral is pledged or charged;3 (ix) change any payment-related circumstance under which the Bonds may be declared due and payable prior to their stated maturity;4 (x) change the seniority or ranking of the Bonds; (xi) change the law governing the Bonds;5 (xii) change any court to whose jurisdiction the Issuer has submitted or any immunity waived by the Issuer in relation to legal proceedings arising out of or in connection with the Bonds;6

1 To be included if the Bonds are redeemable. 2 To be included if the Bonds are guaranteed. 3 To be included if the Bonds are collateralised. 4 To be included if the Bonds are subject to acceleration. 5 To be included if the Bonds are governed by a foreign law. 6 To be included, as appropriate, if the Issuer has submitted to the jurisdiction of a foreign court or expressly waived its immunity.



Appendix 2 

 213

(xiii) change the principal amount of outstanding Bonds or, in the case of a cross-series modification, the principal amount of debt securities of any other series required to approve a proposed modification in relation to the Bonds, the principal amount of outstanding Bonds required for a quorum to be present, or the rules for determining whether a Bond is outstanding for these purposes; or (xiv) change the definition of a reserved matter, and has the same meaning in relation to the debt securities of any other series save that any of the foregoing references to the Bonds or any agreement governing the issuance or administration of the Bonds shall be read as references to such other debt securities or any agreement governing the issuance or administration of such other debt securities. (i) ‘holder’ in relation to a Bond means [the person in whose name the Bond is registered on the books and records of the Issuer]7 / [the bearer of the Bond]8 / [the person the Issuer is entitled to treat as the legal holder of the Bond]9, and in relation to any other debt security means the person the Issuer is entitled to treat as the legal holder of the debt security under the law governing that debt security. (j) ‘record date’ in relation to any proposed modification means the date fixed by the Issuer for determining the holders of Bonds and, in the case of a crossseries modification, the holders of debt securities of each other series that are entitled to vote on or sign a written resolution in relation to the proposed modification.

2. Modification of Bonds 2.1 Reserved Matter Modification. The terms and conditions of the Bonds and any agreement governing the issuance or administration of the Bonds may be modified in relation to a reserved matter with the consent of the Issuer and:

7 Include (subject to footnote 9) if the Bonds are registered bonds, regardless of whether held in global form by a common depositary or custodian. 8 Include (subject to footnote 9) if the Bonds are bearer securities, regardless of whether held in global form by a common depositary or custodian. 9 Include if under applicable law the person entitled to vote the Bond in relation to the Issuer is not the bearer of the Bond or the person in whose name the Bond is registered on the books and record of the Issuer.

214 

 Appendices

(a) the affirmative vote of holders of not less than 75% of the aggregate principal amount of the outstanding Bonds represented at a duly called meeting of Bondholders; or (b) a written resolution signed by or on behalf of holders of not less than 66 2/3% of the aggregate principal amount of the Bonds then outstanding. 2.2 Cross-Series Modification. In the case of a cross-series modification, the terms and conditions of the Bonds and debt securities of any other series, and any agreement governing the issuance or administration of the Bonds or debt securities of such other series, may be modified in relation to a reserved matter with the consent of the Issuer and: (a) (i) the affirmative vote of not less than 75% of the aggregate principal amount of the outstanding debt securities represented at separate duly called meetings of the holders of the debt securities of all the series (taken in the aggregate) that would be affected by the proposed modification; or (a) (ii) a written resolution signed by or on behalf of the holders of not less than 66 2/3% of the aggregate principal amount of the outstanding debt securities of all the series (taken in the aggregate) that would be affected by the proposed modification; and (b) (i) the affirmative vote of more than 66 2/3% of the aggregate principal amount of the outstanding debt securities represented at separate duly called meetings of the holders of each series of debt securities (taken individually) that would be affected by the proposed modification; or (b) (ii) a written resolution signed by or on behalf of the holders of more than 50% of the aggregate principal amount of the then outstanding debt securities of each series (taken individually) that would be affected by the proposed modification. A separate meeting will be called and held, or a separate written resolution signed, in relation to the proposed modification of the Bonds and the proposed modification of each other affected series of debt securities. 2.3 Proposed Cross-Series Modification. A proposed cross-series modification may include one or more proposed alternative modifications of the terms and conditions of each affected series of debt securities or of any agreement governing the issuance or administration of any affected series of debt securities, provided that all such proposed alternative modifications are addressed to and may be accepted by any holder of any debt security of any affected series.



Appendix 2 

 215

2.4 Partial Cross-Series Modification. If a proposed cross-series modification is not approved in relation to a reserved matter in accordance with Section 2.2, but would have been so approved if the proposed modification had involved only the Bonds and one or more, but less than all, of the other series of debt securities affected by the proposed modification, that cross-series modification will be deemed to have been approved, notwithstanding Section 2.2, in relation to the Bonds and debt securities of each other series whose modification would have been approved in accordance with Section 2.2 if the proposed modification had involved only the Bonds and debt securities of such other series, provided that: (a) prior to the record date for the proposed cross-series modification, the Issuer has publicly notified holders of the Bonds and other affected debt securities of the conditions under which the proposed cross-series modification will be deemed to have been approved if it is approved in the manner described above in relation to the Bonds and some but not all of the other affected series of debt securities; and (b) those conditions are satisfied in connection with the proposed cross-series modification. 2.5 Non-Reserved Matter Modification. The terms and conditions of the Bonds and any agreement governing the issuance or administration of the Bonds may be modified in relation to any matter other than a reserved matter with the consent of the Issuer and: (a) the affirmative vote of holders of more than 50% of the aggregate principal amount of the outstanding Bonds represented at a duly called meeting of Bondholders; or (b) a written resolution signed by or on behalf of holders of more than 50% of the aggregate principal amount of the outstanding Bonds. 2.6 Multiple Currencies, Index-Linked Obligations and Zero-Coupon Obligations. In determining whether a proposed modification has been approved by the requisite principal amount of Bonds and debt securities of one or more other series: (a) if the modification involves debt securities denominated in more than one currency, the principal amount of each affected debt security will be equal to the amount of euro that could have been obtained on the record date for the proposed modification with the principal amount of that debt security, using the applicable euro foreign exchange reference rate for the record date published by the European Central Bank; (b) if the modification involves an index-linked obligation, the principal amount of each such index-linked obligation will be equal to its adjusted nominal amount;

216 

 Appendices

(c) if the modification involves a zero-coupon obligation that did not formerly constitute a component part of an index-linked obligation, the principal amount of each such zero-coupon obligation will be equal to its nominal amount or, if its stated maturity date has not yet occurred, to the present value of its nominal amount; (d) if the modification involves a zero-coupon obligation that formerly constituted a component part of an index-linked obligation, the principal amount of each such zero-coupon obligation that formerly constituted the right to receive: (i) a non-index-linked payment of principal or interest will be equal to its nominal amount or, if the stated maturity date of the non-index-linked payment has not yet occurred, to the present value of its nominal amount; and (ii) an index-linked payment of principal or interest will be equal to its adjusted nominal amount or, if the stated maturity date of the indexlinked payment has not yet occurred, to the present value of its adjusted nominal amount; and (e) For purposes of this Section 2.6: (i) the adjusted nominal amount of any index-linked obligation and any component part of an index-linked obligation is the amount of the payment that would be due on the stated maturity date of that indexlinked obligation or component part if its stated maturity date was the record date for the proposed modification, based on the value of the related index on the record date published by or on behalf of the Issuer or, if there is no such published value, on the interpolated value of the related index on the record date determined in accordance with the terms and conditions of the index-linked obligation, but in no event will the adjusted nominal amount of such index-linked obligation or component part be less than its nominal amount unless the terms and conditions of the index-linked obligation provide that the amount of the payment made on such index-linked obligation or component part may be less than its nominal amount; and (ii) the present value of a zero-coupon obligation is determined by discounting the nominal amount (or, if applicable, the adjusted nominal amount) of that zero-coupon obligation from its stated maturity date to the record date at the specified discount rate using the applicable market day-count convention, where the specified discount rate is: (x) if the zero-coupon obligation was not formerly a component part of a debt security that expressly provided for the accrual of interest, the yield to maturity of that zero-coupon obligation at issuance or,



Appendix 2 

 217

if more than one tranche of that zero-coupon obligation has been issued, the yield to maturity of that zerocoupon obligation at the arithmetic average of all the issue prices of all the zero-coupon obligations of that series of zero-coupon obligations weighted by their nominal amounts; and (y) if the zero-coupon obligation was formerly a component part of a debt security that expressly provided for the accrual of interest: (1) the coupon on that debt security if that debt security can be identified; or (2) if such debt security cannot be identified, the arithmetic average of all the coupons on all of the Issuer’s debt securities (weighted by their principal amounts) referred to below that have the same stated maturity date as the zerocoupon obligation to be discounted, or, if there is no such debt security, the coupon interpolated for these purposes on a linear basis using all of the Issuer’s debt securities (weighted by their principal amounts) referred to below that have the two closest maturity dates to the maturity date of the zero-coupon obligation to be discounted, where the debt securities to be used for this purpose are all of the Issuer’s index-linked obligations if the zero-coupon obligation to be discounted was formerly a component part of an index-linked obligation and all of the Issuer’s debt securities (index-linked obligations and zero-coupon obligations excepted) if the zerocoupon obligation to be discounted was not formerly a component part of an index-linked obligation, and in either case are denominated in the same currency as the zero-coupon obligation to be discounted. 2.7 Outstanding Bonds. In determining whether holders of the requisite principal amount of outstanding Bonds have voted in favour of a proposed modification or whether a quorum is present at any meeting of Bondholders called to vote on a proposed modification, a Bond will be deemed to be not outstanding, and may not be voted for or against a proposed modification or counted in determining whether a quorum is present, if on the record date for the proposed modification: (a) the Bond has previously been cancelled or delivered for cancellation or held for reissuance but not reissued; (b) the Bond has previously been called for redemption in accordance with its terms or previously become due and payable at maturity or otherwise and

218 

 Appendices

the Issuer has previously satisfied its obligation to make all payments due in respect of the Bond in accordance with its terms;10 or (c) the Bond is held by the Issuer, by a department, ministry or agency of the Issuer, or by a corporation, trust or other legal entity that is controlled by the Issuer or a department, ministry or agency of the Issuer and, in the case of a Bond held by any such above-mentioned corporation, trust or other legal entity, the holder of the Bond does not have autonomy of decision, where: (i) the holder of a Bond for these purposes is the entity legally entitled to vote the Bond for or against a proposed modification or, if different, the entity whose consent or instruction is by contract required, directly or indirectly, for the legally entitled holder to vote the Bond for or against a proposed modification; (ii) a corporation, trust or other legal entity is controlled by the Issuer or by a department, ministry or agency of the Issuer if the Issuer or any department, ministry or agency of the Issuer has the power, directly or indirectly, through the ownership of voting securities or other ownership interests, by contract or otherwise, to direct the management of or elect or appoint a majority of the board of directors or other persons performing similar functions in lieu of, or in addition to, the board of directors of that legal entity; and (iii) the holder of a Bond has autonomy of decision if, under applicable law, rules or regulations and independent of any direct or indirect obligation the holder may have in relation to the Issuer: (x) the holder may not, directly or indirectly, take instruction from the Issuer on how to vote on a proposed modification; or (y) the holder, in determining how to vote on a proposed modification, is required to act in accordance with an objective prudential standard, in the interest of all of its stakeholders or in the holder’s own interest; or (z) the holder owes a fiduciary or similar duty to vote on a proposed modification in the interest of one or more persons other than a person whose holdings of Bonds (if that person then held any Bonds) would be deemed to be not outstanding under this Section 2.7. 2.8 Outstanding Debt Securities. In determining whether holders of the requisite principal amount of outstanding debt securities of another series have voted in

10 The reference to the Bond having previously been called for redemption to be included if the Bond is redeemable.



Appendix 2 

 219

favor of a proposed cross-series modification or whether a quorum is present at any meeting of the holders of such debt securities called to vote on a proposed cross-series modification, an affected debt security will be deemed to be not outstanding, and may not be voted for or against a proposed cross-series modification or counted in determining whether a quorum is present, in accordance with the applicable terms and conditions of that debt security. 2.9 Entities Having Autonomy of Decision. For transparency purposes, the Issuer will publish promptly following the Issuer’s formal announcement of any proposed modification of the Bonds, but in no event less than 10 days prior to the record date for the proposed modification, a list identifying each corporation, trust or other legal entity that for purposes of Section 2.7(c): (a) is then controlled by the Issuer or by a department, ministry or agency of the Issuer; (b) has in response to an enquiry from the Issuer reported to the Issuer that it is then the holder of one or more Bonds; and (c) does not have autonomy of decision in respect of its Bondholdings. 2.10 Exchange and Conversion. Any duly approved modification of the terms and conditions of the Bonds may be implemented by means of a mandatory exchange or conversion of the Bonds for new debt securities containing the modified terms and conditions if the proposed exchange or conversion is notified to Bondholders prior to the record date for the proposed modification. Any conversion or exchange undertaken to implement a duly approved modification will be binding on all Bondholders.

3. Calculation Agent 3.1 Appointment and Responsibility. The Issuer will appoint a person (the ‘calculation agent’) to calculate whether a proposed modification has been approved by the requisite principal amount of outstanding Bonds and, in the case of a cross-series modification, by the requisite principal amount of outstanding debt securities of each affected series of debt securities. In the case of a cross-series modification, the same person will be appointed as the calculation agent for the proposed modification of the Bonds and each other affected series of debt securities. 3.2 Certificate. The Issuer will provide to the calculation agent and publish prior to the date of any meeting called to vote on a proposed modification or the date

220 

 Appendices

fixed by the Issuer for the signing of a written resolution in relation to a proposed modification, a certificate: (a) listing the total principal amount of Bonds and, in the case of a cross-series modification, debt securities of each other affected series outstanding on the record date for purposes of Section 2.7; (b) specifying the total principal amount of Bonds and, in the case of a crossseries modification, debt securities of each other affected series that are deemed under Section 2.7(c) to be not outstanding on the record date; and (c) identifying the holders of the Bonds and, in the case of a cross-series modification, debt securities of each other affected series, referred to in (b) above, determined, if applicable, in accordance with the provisions of Section 2.6. 3.3 Reliance. The calculation agent may rely on any information contained in the certificate provided by the Issuer, and that information will be conclusive and binding on the Issuer and the Bondholders unless: (a) an affected Bondholder delivers a substantiated written objection to the Issuer in relation to the certificate before the vote on a proposed modification or the signing of a written resolution in relation to a proposed modification; and (b) that written objection, if sustained, would affect the outcome of the vote taken or the written resolution signed in relation to the proposed modification. In the event a substantiated written objection is timely delivered, any information relied on by the calculation agent will nonetheless be conclusive and binding on the Issuer and affected Bondholders if: (x) the objection is subsequently withdrawn; (y) the Bondholder that delivered the objection does not commence legal action in respect of the objection before a court of competent jurisdiction within 15 days of the publication of the results of the vote taken or the written resolution signed in relation to the proposed modification; or (z) a court of competent jurisdiction subsequently rules either that the objection is not substantiated or would not in any event have affected the outcome of the vote taken or the written resolution signed in relation to the proposed modification. 3.4 Publication. The Issuer will arrange for the publication of the results of the calculations made by the calculation agent in relation to a proposed modification promptly following the meeting called to consider that modification or, if



Appendix 2 

 221

applicable, the date fixed by the Issuer for signing a written resolution in respect of that modification.

4. Bondholder Meetings; Written Resolutions 4.1 General. The provisions set out below, and any additional rules adopted and published by the Issuer will, to the extent consistent with the provisions set out below, apply to any meeting of Bondholders called to vote on a proposed modification and to any written resolution adopted in connection with a proposed modification. Any action contemplated in this Section 4 to be taken by the Issuer may instead be taken by an agent acting on behalf of the Issuer. 4.2 Convening Meetings. A meeting of Bondholders: (a) may be convened by the Issuer at any time; and (b) will be convened by the Issuer if an event of default in relation to the Bonds has occurred and is continuing and a meeting is requested in writing by the holders of not less than 10% of the aggregate principal amount of the Bonds then outstanding.11 4.3 Notice of Meetings. The notice convening a meeting of Bondholders will be published by the Issuer at least 21 days prior to the date of the meeting or, in the case of an adjourned meeting, at least 14 days prior to the date of the adjourned meeting. The notice will: (a) state the time, date and venue of the meeting; (b) set out the agenda and quorum for, and the text of any resolutions proposed to be adopted at, the meeting; (c) specify the record date for the meeting, being not more than five business days12 before the date of the meeting, and the documents required to be produced by a Bondholder in order to be entitled to participate in the meeting; (d) include the form of instrument to be used to appoint a proxy to act on a Bondholder’s behalf; (e) set out any additional rules adopted by the Issuer for the convening and holding of the meeting and, if applicable, the conditions under which a crossseries modification will be deemed to have been satisfied if it is approved as to some but not all of the affected series of debt securities; and

11 To be included if the Bonds contain events of default. 12 The term ‘business day’ will be defined elsewhere in the Bond documentation.

222 

 Appendices

(f) identify the person appointed as the calculation agent for any proposed modification to be voted on at the meeting. 4.4 Chair. The chair of any meeting of Bondholders will be appointed: (a) by the Issuer; or (b) if the Issuer fails to appoint a chair or the person nominated by the Issuer is not present at the meeting, by holders of more than 50% of the aggregate principal amount of the Bonds then outstanding represented at the meeting. 4.5 Quorum. No business will be transacted at any meeting in the absence of a quorum other than the choosing of a chair if one has not been appointed by the Issuer. The quorum at any meeting at which Bondholders will vote on a proposed modification of: (a) a reserved matter will be one or more persons present and holding not less than 66 2/3% of the aggregate principal amount of the Bonds then outstanding; and (b) a matter other than a reserved matter will be one or more persons present and holding not less than 50% of the aggregate principal amount of the Bonds then outstanding. 4.6 Adjourned Meetings. If a quorum is not present within thirty minutes of the time appointed for a meeting, the meeting may be adjourned for a period of not more than 42 days and not less than 14 days as determined by the chair of the meeting. The quorum for any adjourned meeting will be one or more persons present and holding: (a) not less than 66 2/3% of the aggregate principal amount of the Bonds then outstanding in the case of a proposed reserved-matter modification; and (b) not less than 25% of the aggregate principal amount of the Bonds then outstanding in the case of a non-reserved matter modification. 4.7 Written Resolutions. A written resolution signed by or on behalf of holders of the requisite majority of the Bonds will be valid for all purposes as if it was a resolution passed at a meeting of Bondholders duly convened and held in accordance with these provisions. A written resolution may be set out in one or more document in like form each signed by or on behalf of one or more Bondholders. 4.8 Entitlement to Vote. Any person who is a holder of an outstanding Bond on the record date for a proposed modification, and any person duly appointed as a proxy by a holder of an outstanding Bond on the record date for a proposed modification, will be entitled to vote on the proposed modification at a meeting of



Appendix 2 

 223

Bondholders and to sign a written resolution with respect to the proposed modification. 4.9 Voting. Every proposed modification will be submitted to a vote of the holders of outstanding Bonds represented at a duly called meeting or to a vote of the holders of all outstanding Bonds by means of a written resolution without need for a meeting. A holder may cast votes on each proposed modification equal in number to the principal amount of the holder’s outstanding Bonds. For these purposes: (a) in the case of a cross-series modification involving debt securities denominated in more than one currency, the principal amount of each debt security will be determined in accordance with Section 2.6(a); (b) in the case of a cross-series modification involving an index-linked obligation, the principal amount of each such index-linked obligation will be determined in accordance with Section 2.6(b); (c) in the case of a cross-series modification involving a zero-coupon obligation that did not formerly constitute a component part of an index-linked obligation, the principal amount of each such zero-coupon obligation will be determined in accordance with Section 2.6(c); and (d) in the case of a cross-series modification involving a zero-coupon obligation that did formerly constitute a component part of an index-linked obligation, the principal amount of each such zero-coupon obligation will be determined in accordance with Section 2.6(d). 4.10 Proxies. Each holder of an outstanding Bond may, by an instrument in writing executed on behalf of the holder and delivered to the Issuer not less than 48 hours before the time fixed for a meeting of Bondholders or the signing of a written resolution, appoint any person (a “proxy”) to act on the holder’s behalf in connection with any meeting of Bondholders at which the holder is entitled to vote or the signing of any written resolution that the holder is entitled to sign. Appointment of a proxy pursuant to any form other than the form enclosed with the notice of the meeting will not be valid for these purposes. 4.11 Legal Effect and Revocation of a Proxy. A proxy duly appointed in accordance with the above provisions will, subject to Section 2.7 and for so long as that appointment remains in force, be deemed to be (and the person who appointed that proxy will be deemed not to be) the holder of the Bonds to which that appointment relates, and any vote cast by a proxy will be valid notwithstanding the prior revocation or amendment of the appointment of that proxy unless the Issuer has received notice or has otherwise been informed of the revocation or amendment

224 

 Appendices

at least 48 hours before the time fixed for the commencement of the meeting at which the proxy intends to cast its vote or, if applicable, the signing of a written resolution. 4.12 Binding Effect. A resolution duly passed at a meeting of holders convened and held in accordance with these provisions, and a written resolution duly signed by the requisite majority of Bondholders, will be binding on all Bondholders, whether or not the holder was present at the meeting, voted for or against the resolution or signed the written resolution. 4.13 Publication. The Issuer will without undue delay publish all duly adopted resolutions and written resolutions.

5. Publication 5.1 Notices and Other Matters. The Issuer will publish all notices and other matters required to be published pursuant to the above provisions: (a) on [insert the Issuer’s website for financial notices]; (b) through [insert clearing system];13 and (c) in such other places, including in [insert the Issuer’s official gazette], and in such other manner as may be required by applicable law or regulation.

13 To be included if the Bonds are cleared through a central depositary system.

Appendix 3 EFC Sub-Committee on EU Sovereign Debt Markets Supplemental Provisions to Model Collective Action Clause Supplemental Provisions 1. Technical Amendments 1.1 Manifest Error, Technical Amendments. Notwithstanding anything to the contrary herein, the terms and conditions of the Bonds and any agreement governing the issuance or administration of the Bonds may be modified by the Issuer without the consent of Bondholders: (i) to correct a manifest error or cure an ambiguity; or (ii) if the modification is of a formal or technical nature or for the benefit of Bondholders. The Issuer will publish the details of any modification of the Bonds made pursuant to this Section [•] within ten days of the modification becoming legally effective.

2. Acceleration and Rescission of Acceleration1 2.1 Acceleration. If any event of default occurs and is continuing, the holders of not less than 25% of the aggregate principal amount of the outstanding Bonds may, by written notice given of the Issuer, declare the Bonds to be immediately due and payable. Upon any declaration of acceleration properly given in accordance with this Section, all amounts payable on the Bonds will become immediately due and payable on the date that written notice of acceleration is received by the Issuer, unless the event of default has been remedied or waived prior of the receipt of the notice by the Issuer.

1 To be included only if the Bonds provide for acceleration.

226 

 Appendices

2.2 Rescission of Acceleration. The holders of more than 50% of the aggregate principal amount of the outstanding Bonds may, on behalf of all Bondholders, rescind or annul any notice of acceleration given pursuant to Section 2.1 above.

3. Limitation on Sole Holder Action2 3.1 No Bondholder will be entitled to institute proceedings against the Issuer or take steps to enforce the rights of the Bondholders under the terms and conditions of the Bonds unless the [trustee/fiscal agent], having become bound to proceed in accordance with these terms and conditions, has failed to do so within a reasonable time and such failure is continuing.

2 To be included only if the Bonds provide for a fiscal agent or trustee.

Appendix 4 EFC Sub-Committee on EU Sovereign Debt Markets Model Collective Action Clause Supplemental Explanatory Note 1. Introduction On 26 July 2011, the EFC Sub-Committee on EU Sovereign Debt Markets (the “Committee”) distributed for comment by market participants and other interested stakeholders a draft of the model collective action clause (“CAC”) to be included in all euro area government securities. The draft CAC was accompanied by an Explanatory Note dated 26 July 2011 that, among other things, summarized the key provisions of the model CAC. A total of fifteen responses were received from institutional investors, securities and insurance trade associations, the European Central Bank, the IMF, the London Stock Exchange and Euroclear. Following a careful review of the comments received and further informal consultations with market participants and other interested stakeholders, the Committee has approved the enclosed final version of the model CAC. This Supplemental Explanatory Note discusses the principal changes made in the model CAC in response to market comments, and should be read together with the Committee’s earlier Explanatory Note. It also describes the process that will be followed by euro area Member States in implementing the model CAC.

2. Mandatory Introduction Date On 2 February 2012, a modified version of the Treaty Establishing the European Stability Mechanism was signed by all 17 euro area Member States. Paragraph 3 of Article 12 of the Treaty states: “Collective action clauses shall be included, as of 1 January 2013, in all new euro area government securities, with maturity above one year, in a way which ensures that their legal impact is identical.”

As a result, the model CAC will become mandatory in all new euro area government securities issued on or after 1 January 2013, and not 1 July 2013, as previously contemplated.

228 

 Appendices

The introduction of the model CAC will not affect any euro area government securities issued prior to 1 January 2013 unless those securities include a collective action clause that allows for their modification on a cross-series basis on the terms contemplated in the model CAC. There are no such bonds now outstanding, apart from the English-law governed bonds recently issued by the Hellenic Republic as part of its Private Sector Initiative, which include a forward-looking collective action clause based on the model CAC. The Committee does not expect that any other euro area government securities issued prior to 1 January 2013 will contain a collective action clause based on the model CAC. As a result and except as just noted, euro area government securities issued prior to 1 January 2013 will not be subject to modification as part of a cross-series modification pursuant to the model CAC.

3. Changes Made to the Model CAC Set out below is a summary of the most important changes made to the model CAC in response to comments received as part of the public consultation process.

A. Bondholder-Meeting Approval Thresholds and Quorums The modification of a reserved matter, if proposed at a duly called meeting of bondholders, now requires the affirmative vote of not less than 75% – up from 66 2/3% in the prior draft – of the aggregate principal amount of the outstanding bonds represented at that meeting. The quorum for any adjourned meeting of bondholders at which a reservedmatter modification will be considered has now been fixed at the same level – 66  2/3% of the outstanding principal amount of the affected bonds – as the quorum required at the initial meeting of bondholders called to consider that matter. No change has been made in the initial- or adjourned-meeting quorums for a non-reserved matter, which remain at 50% (initial meeting) and 25% (adjourned meeting) of the outstanding principal amount of the affected bonds. The Committee believes that it should not be easier, or more difficult, for a proposed modification to be approved at a meeting of bondholders than by an action in writing. The Committee also continues to be of the view that investors opposed to a proposed modification are more likely in practice to be represented at a meeting called to consider the modification, and that the bondholder-meeting and action-in-writing approval thresholds should accordingly be fixed at the same level, though with the former measured against the outstanding principal



Appendix 4 

 229

amount of affected bonds represented at a quorate meeting and the latter against the outstanding principal amount of all affected bonds. The Committee has nonetheless increased both the bondholder-meeting approval threshold and the adjourned-meeting quorum for reserved-matter modifications in light of the concern expressed by some market participants that the prior draft unfairly advantaged the approval of a reserved matter at a bondholder meeting. By contrast, the action-in-writing approval threshold, which had already been informally sanctioned by a number of market participants, attracted relatively little comment, and remains unchanged at not less than 66 2/3% of the aggregate principal amount of the bonds then outstanding. One consequence of the two changes, taken together, is that a single-series modification of a reserved matter will in all events require the approval of at least a majority in principal amount of the affected bonds (>75% × 66 2/3% > 50%). While the Committee continues to believe it extremely unlikely that an important single-series reserved-matter modification would ever have been approved under the prior version of the CAC without the consent of a majority of the principal amount of the affected bonds, the Committee thought it appropriate under current market conditions to rule out this possibility even as a theoretical matter. The Committee also considered two related suggestions made by some commentators first, that the model CAC require that a reserved-matter modification be approved only by an action writing, and second, that the bondholder-meeting reserved-matter approval threshold, if retained, be fixed at 66 2/3% of the outstanding principal amount of all affected bonds. The Committee ultimately rejected both suggestions. While it is true that most recent emerging market sovereign debt restructurings have been implemented by actions in writing, bondholder meetings are the norm in the euro area for corporate restructurings, and it is expected that local corporate rules will in many cases serve as the starting point for the procedural rules to be adopted by euro area governments for modifying their own debt securities. If the approval threshold for a reserved-matter modification was in all cases measured against the outstanding principal amount of the affected bonds, every bond not represented at a meeting of bondholders would automatically be treated as a vote in opposition to the proposed modification. This struck the Committee as unreasonable, and would in practice make it increasingly difficult, and ultimately impossible, for a proposed modification to be approved with declining levels of bondholder-meeting participation. As a result, the Committee concluded that this approach, while it would preserve formal equality of treatment, would in practice produce very different results depending on the level of bondholder representation.

230 

 Appendices

The Committee also notes that the model CAC, in fixing the reserved-matter quorum for all meetings at 66 2/3% of the outstanding principal amount of the affected bonds, achieves, by a different route, a comparable balancing of the interests of investors and issuers to that which would have been provided by the proposed alternative approach. In sum, the Committee believes that the model CAC as now drafted avoids the unfair consequences of the unrealistic assumption that all non-participating bondholders will vote against a proposed modification, while affording bondholders appropriate protection against a reserved-matter modification being approved over the voiced opposition of a substantial minority of investors.

B. Cross-Series Modification Two changes have been made to provisions of the model CAC dealing with crossseries modification. First, the bondholder-meeting approval thresholds have been increased to 75% of the outstanding principal amount of all affected bonds in the aggregate and to 66 2/3% of the outstanding principal amount of each individual series of affected bonds. In addition, every modification proposed in relation to any series of bonds covered by a cross-series modification must now be available to be accepted by every holder of an affected bond. By way of background, the Committee notes that it will be for the issuer to decide whether to pursue – or not to pursue – a cross-series modification. An issuer that elects to pursue a cross-series modification may include some – but not all-of its outstanding bonds in a single cross-series modification, and may also treat its remaining bonds in one or more additional crossseries modifications, in several separate single-series modifications or in a combination of the two. The flexibility afforded to the issuer in grouping bonds of different series in different modification baskets is intended to minimize the likelihood of a hold-out series, an interest shared by both the issuer and those investors – the vast majority – that approve a cross-series modification on both an aggregate and individual basis. A cross-series modification can be understood as, in effect, a collective action clause that works at the series level, in that it binds all of the holders of all of the affected series so long as the proposed modification is overwhelmingly approved in the aggregate by the affected group of bondholders, with the important further protection that holders of any individual series of affected bonds will not be bound by the decision of the group as a whole unless they also vote in favour of the proposed modification, though not necessarily at the level that would be required on a stand-alone basis. The rationale underlying a cross-series modification requires that all bondholders affected by a cross-series modification should have the right to vote on



Appendix 4 

 231

the same proposed modification or, if a menu of different options is proposed, on each of the proposed options. New Section 2.3 of the model CAC makes this point expressly by providing that a cross-series modification may include one or more proposed alternative modifications, but only if all of the proposed alternatives are addressed to and may be accepted by any holder of any bond that would be affected if the cross-series modification were to be approved. The issuer is thus given the right to propose a menu of alternative options, and investors the right to pick the option they prefer, ensuring that all investors have the right to accept from a common menu of options the option that best suits their interests. This arrangement, together with the veto right effectively granted to each individual series of affected bonds, should ensure that no series of bonds is singled out for invidious treatment. Together with the issuer’s self-interest in having a cross-series modification approved by all of the affected series (no holdouts), this provision should also lead to different series of bonds being grouped in a commercially sensible way in a cross-series modification. The Committee confirms that, under the model CAC, a separate meeting will be held for each series of bonds covered by a cross-series modification, though these meetings are likely to take place on the same day and, perhaps, even in the same place.

C. Partial Cross-Series Modification An issuer is now required to notify investors in advance of the conditions under which a partial cross-series modification will come into effect. Investors will, as a result, be able to make an informed investment decision as to whether they are in favor of a proposed cross-series modification even if it is only partially successful. This approach should be familiar to the market, as it is already standard practice for an issuer proposing to modify a series of bonds that does not include a collective action clause to specify the “critical mass” of bonds that must approve the proposed changes before they will become legally effective.

D. Zero-coupon and Index-Linked Obligations The provisions of the model CAC dealing with zero-coupon obligations have been substantially revised and new provisions added to deal with index-linked obligations and stripped index-linked obligations. The model CAC now includes comprehensive rules for calculating the principal amount of a zero-coupon and index-linked obligation. These amounts will in turn be used to determine whether a proposed modification has attracted the support of the requisite principal amount of the affected bonds.

232 

 Appendices

(1) Index-linked obligations The model CAC now provides special treatment for the adjustable principal and interest components of an index-linked obligation. The principal issue addressed by the Committee was whether, in allocating votes to the adjustable portion of an index-linked obligation, the value of that adjustable portion should be ‘crystallized’ on the record date for a proposed modification even if, in accordance with its terms, its value would not then be adjusted and could in theory be reversed over time. The Committee ultimately concluded that interim ‘crystallized’ values should be used, principally because the adjustable portion of an index-linked obligation is keyed to changes in a published price index, and in the euro area these changes have over many years been almost entirely a one-way street of increasing prices. The component parts of most index-linked obligations may be ‘stripped’ by the holder. The resulting zero-coupon obligations are afforded the same ‘crystallized’ treatment under the model CAC, though the votes allocated to a zerocoupon index linked obligation are subject to further adjustment in accordance with the rules applicable to zero-coupon obligations more generally, discussed below.

(2) Zero-Coupon Obligations Under the prior draft of the model CAC, the principal amount of a zero-coupon obligation was equal to its present value on the record date for a proposed modification. This remains the case in the final version of the CAC, though it is now expressly stated that the principal amount of a zero-coupon obligation will be equal to its nominal amount if the stated maturity date of the zero-coupon obligation has already occurred. The present value of a zero-coupon obligation is now calculated using one of three different discount rates: –– if the bond was originally issued as a zero-coupon obligation, the applicable discount rate is the zero-coupon obligation’s yield to maturity at issuance; –– if the zero-coupon obligation was ‘stripped’ from a couponed debt security, the applicable discount rate is the coupon on the underlying debt security; and –– if the zero-coupon obligation was ‘stripped’ from a couponed debt security that cannot be identified due to the fungibility of all of the issuer’s zero-coupon obligations maturing on the same date, the applicable default discount rate is the weighted average of the coupons on the issuer’s debt securities that have the same stated maturity date as the zero-coupon obligation or, if there are no such securities, the coupon interpolated on a linear basis using



Appendix 4 

 233

the weighted average of the coupons on the issuer’s debt securities having the two closest maturity dates to that of the zero-coupon obligation. The methods to be used in calculating the applicable discount rate, described above, are subject to further detailed rules set out in the model CAC. The Committee’s principal aim was to assign the same number of total votes to the zero-coupon component parts of a ‘stripped’ security as were allocated to that same security before it was ‘stripped’. Using the coupon on the underlying ‘stripped’ security as the discount rate achieves this goal with mathematical precision. Using the yield to maturity at issuance of a bond initially issued as a zerocoupon obligation produces the same result (assuming the ‘stripped’ couponed security was issued at par). By contrast, use of the default discount rate could in practice result in more or fewer votes being allocated to the former component parts of a ‘stripped’ security. To address this concern, the Committee tested the default methodology using actual debt securities now in the market, and concluded that that variance in the votes allocated fell within acceptable limits. Several related drafting changes have also been introduced. A ‘zero-coupon obligation’ and an ‘index-linked obligation’ are now defined in the model CAC, and these terms are used extensively in Section 2.6, which deals in a comprehensive way with the calculation of their principal amounts. The definition of ‘debt securities’ also now expressly confirms that all of the stripped component parts of a debt security are themselves debt securities irrespective of their stated maturity date. As result, even short-dated ‘stripped’ coupons will be treated as debt securities so long as the underlying couponed security was itself a debt security.

E. Disenfranchisement Several commentators suggested that consideration should be given to disenfranchising investors who have the same interests as an issuer, who are predictably likely to vote in favour of a proposed modification or who are likely to be motivated by circumstances other than the direct effect of a proposed modification on the value of their holdings. The Committee strongly disagrees. In the Committee’s view, disenfranchising an investor is a serious step with important legal consequences, and should be taken only for good cause. The Committee believes that neither an investor’s interests or motives, nor the predictability of an investor’s vote for or against a proposed modification, constitutes adequate grounds for disenfranchising an investor. For the Committee, the litmus test remains: is a bondholder acting in its own interest? If so, the bondholder should be enfranchised irrespective of whether the investor may also be directly or indirectly owned or controlled by the issuer for purposes of the model CAC.

234 

 Appendices

The legal status of euro area national central banks illustrates the Committee’s thinking on this issue. Article 130 of the Treaty on the Functioning of the European Union and Article 7 of the Statute of the European System of Central Banks and of the European Central Bank expressly prohibit euro area national central banks and members of their decision-making bodies from seeking or taking instruction from European Union institutions or bodies, from any government of a Member State of the European Union or from any other body. As a result, a euro area national central bank’s decision to vote for or against the proposed modification of securities acquired in connection with, for instance, its Eurosystem operations must, as a matter of law, be made by the bank acting in its own interest, even though the bank may well be owned or, for other purposes, controlled by its government (for example, the government may have the right to appoint the bank’s governor or a majority of its directors). In the Committee’s view, euro area national central banks accordingly have autonomy of decision in deciding how to vote on the proposed modification of any euro area government securities so acquired, and their holdings of these securities will be enfranchised under the model CAC.

F. Bondholder Meetings The Committee would like to emphasize that the model CAC does not prescribe the complete set of rules applicable to bondholder meetings. All euro area jurisdictions already have their own bondholder-meetings rules and practices, and the model CAC expressly contemplates that its bondholder-meeting rules may be supplemented by additional rules adopted and published by the issuer in advance of any meeting called to consider a proposed modification.

G. Transparency The transparency provisions of the model CAC have been strengthened in response to market comment. An issuer is now required to publish, promptly following the announcement of a proposed modification, but in no event less than 10 days prior to the record date for a proposed modification, a list (prepared on the basis of enquiries made by the issuer) of the disenfranchised companies and entities then holding affected bonds. The issuer must now also publish a list of the record-date holders of disenfranchised bonds and the total amount of their holdings (but not the amount of any individual investor’s holdings). In discussing this matter informally with market participants, many investors were more interested in having a list of the companies whose bondholdings will be disenfranchised, rather than (as was contemplated in the prior draft) a



Appendix 4 

 235

non-exhaustive list of the companies having autonomy of decision. The list now required by the model CAC will presumably enable investors to assess whether there are additional bondholdings that they believe should also be disenfranchised.

H. Calculation Agent A number of commentators suggested that the model CAC should require the appointment of a calculation agent at the time of issuance. The Committee expects, in the case of a euro area government’s international debt securities, that the calculation agent will in fact be appointed at the time of issuance and will most likely be an affiliate of the fiscal agent or trustee also appointed by the issuer. By contrast, the Committee believes that appointment of a calculation agent at the time of issuance of a euro area government’s domestic debt obligations would be inconsistent with current market practice and unduly burdensome. The Committee does not anticipate any problem in the appointment of a calculation agent closer to the time of a proposed modification, as the calculation agent’s responsibilities are limited and strictly administrative in nature.

I. Other Provisions The model CAC does not include a technical amendments clause, does not limit the right of a bondholder to accelerate its bonds or to bring legal proceedings against an issuer, and does not grant bondholders the right to rescind a prior acceleration notice. While these provisions are found in many (but not all) collective action clauses now trading in the international debt market, they were omitted from the model CAC because the CAC will be included in both an issuer’s international and domestic obligations, and the above-mentioned provisions are either inconsistent with current domestic government debt issuance practice or would raise serious legal concerns in some euro area jurisdictions. The Committee nonetheless believes that these provisions serve the same purpose as are intended to be served by the model CAC, and the Committee accordingly recommends that they be included in all euro area government debt securities beginning in 2013 to the extent they are consistent with an issuer’s existing practice and applicable law. The text of the recommended provisions is enclosed. The model CAC also does not provide for the mandatory appointment of a fiscal agent or trustee. While found in most international debt securities, fiscal agents and trustees are not at all common in domestic government debt issuances, and were for that reason not included in the model CAC. The Committee

236 

 Appendices

does not expect that the introduction of the model CAC will affect a euro area government issuer’s existing practice of appointing fiscal agents and trustees for international debt issuances. If a fiscal agent or trustee is appointed by a euro area government, the Committee anticipates that they will have their customary responsibilities, including, among other things, liaising with the issuer and assisting in the convening and holding of bondholder meetings.

J. Retail Savings Bonds A number of euro area governments sell through retail channels, without any fees, charges or commissions, so-called savings bonds or certificates that may not be transferred by their holder. These savings products are manifestly not capital market instruments – indeed, they are almost certainly not securities at all – and the Committee accordingly concluded that they should not be treated as ‘debt securities’ for purposes of the model CAC, even if in some cases they are styled as ‘savings bonds’. Euro area governments are, as result, neither required nor expected to include the model CAC in retail savings instruments that have the above-mentioned features.

K. Miscellaneous Drafting Matters A number of more technical drafting changes have been made to the revised model CAC: –– The definition of a ‘series’ now provides that all of the bonds comprising a ‘series’ of bonds must be identical in all respects except for their date of issuance or first payment date. –– If a bond provides for acceleration upon the occurrence of an event of default, the list of reserved matters now includes a payment-related change in the circumstances under which that bond may be accelerated prior to its stated maturity. –– Except as permitted by any related security agreement, the release of any collateral that is pledged or charged as security for the payment of a bond now constitutes a reserved matter. –– If a bondholder wishes to legally challenge the calculation agent’s reliance on information provided by the issuer in connection with a proposed modification, the investor must now do so within 15 days of the publication of the result of the vote on the proposed modification (rather than within 15 days of the vote itself, as provided for in the prior draft). –– The record date for a bondholder meeting may now be set at not more than five business days before the date of the meeting.



Appendix 4 

 237

–– If an issuer’s debt securities are cleared through a central depositary system, notices required under the model CAC must now also be published through that clearing system. –– The ordering of several sections of the model CAC has changed, in part as a result of the introduction of new Section 2.3, dealing with the availability of the complete menu of options in a cross-series modification.

3. Tapping At its summit held on 24–25 March 2011, the European Council concluded that euro area Member States should be allowed under agreed conditions to reopen (“tap”) debt issuances outstanding on the date of the CAC’s mandatory introduction, in order to preserve market liquidity. Neither the tapped issuance nor (to preserve fungibility) the debt securities issued as part of the reopening of that issuance will contain the model CAC. Year

Percentage

2013

45

2014

40

2015

35

2016

30

2017

30

2018

25

2019

25

2020

25

2021

25

2022

10

2023 onwards

 5

All of the euro area Member States have agreed that the aggregate face amount of euro area central government debt securities that may be issued by any Member State in any year as part of the reopening of an issuance outstanding on 31 December 2012 will limited to a stated percentage of the total face amount of all central government debt securities issued by that Member State in that year. The table below lists for each year the total face amount of a Member State’s central government debt securities that may be issued without the model CAC in connection with the reopening of a debt issuance which was outstanding on 31 December

238 

 Appendices

2012, expressed as a percentage of the total face amount of all of that Member State’s central government debt issuance for that year. In light of the uncertain liquidity of existing euro area government debt issuances, the Committee has agreed that the limits on tapping set out above will be revisited in 2015 and that Member States may in urgent cases request permission from the Committee to exceed the agreed limits. In the interim, the Committee will monitor Member States’ tapping of pre-model CAC issuances, based on information periodically supplied by each Member State.

4. Approval and Implementation The model CAC was approved by the Economic and Finance Committee (EFC) on 18 November 2011. The text of the model CAC, together with this Supplemental Explanatory Note and the Committee’s prior Explanatory Note of 26 July 2011, are now posted on the Committee’s website (http://europa.eu/efc/sub_committee/) and will shortly be posted on the website maintained by each Member State’s debt management office. The Committee has agreed a schedule for monitoring each euro area Member State’s progress in implementing the model CAC, with a view to its timely introduction in all euro area government securities issued after 31 December 2012. As part of this process and consistent with the “identical legal impact” mandate contained in the Treaty Establishing the European Stability Mechanism, each Member State will be required to deliver a legal opinion to the Committee from the highest State authority competent for such matters, confirming that the model CAC will be legal, valid, binding and enforceable in accordance with its terms under the laws of that Member State. The Committee expects to publish a report on the implementation of the model CAC prior to January 2013, when the CAC will become mandatory in all new bonds, notes and other debt securities with an original stated maturity of more than one year issued by euro area national governments. 26 March 2012

Appendix 5 EFC Sub-Committee on EU Sovereign Debt Markets Draft Model Collective Action Clause (distributed for Comment in July 2012) Draft Common Terms of Reference 1. General Definitions (a) ‘debt securities’ means the Bonds and any other bills, bonds, debentures, notes or other debt securities issued by the Issuer in one or more series with an original stated maturity of more than one year. (b) ‘series’ means a tranche of debt securities, together with any further tranche or tranches of debt securities that in relation to each other and to the original tranche of debt securities are (i) identical in all respects except for their issue date, and (ii) expressed to be consolidated and form a single series, and includes the Bonds and any further issuances of Bonds. (c) ‘modification’ in relation to the Bonds means any modification, amendment, supplement or waiver of the terms and conditions of the Bonds or any agreement governing the issuance or administration of the Bonds, and has the same meaning in relation to the debt securities of any other series save that any of the foregoing references to the Bonds or any agreement governing the issuance or administration of the Bonds are to be read as references to such other debt securities or any agreement governing the issuance or administration of such other debt securities. (d) ‘cross-series modification’ means a modification involving (i) the Bonds or any agreement governing the issuance or administration of the Bonds, and (ii) the debt securities of one or more other series or any agreement governing the issuance or administration of such other debt securities. (e) ‘reserved matter’ in relation to the Bonds means any modification of the terms and conditions of the Bonds or of any agreement governing the issuance or administration of the Bonds that would: (i) change the date on which any amount is payable on the Bonds; (ii) reduce any amount, including any overdue amount, payable on the Bonds; (iii) change the method used to calculate any amount payable on the Bonds;

240 

 Appendices

(iv) reduce the redemption price for the Bonds or change any date on which the Bonds may be redeemed;1 (v) change the currency or place of payment of any amount payable on the Bonds; (vi) impose any condition on or otherwise modify the Issuer’s obligation to make payments on the Bonds; (vii) except as permitted by any related guarantee, release any guarantee issued in relation to the Bonds or change the terms of that guarantee;2 (viii) except as permitted by any related security agreement, release any collateral that is pledged as security for the payment of the Bonds or change the terms on which that collateral is pledged;3 (ix) change the seniority or ranking of the Bonds; (x) change the law governing the Bonds;4 (xi) change any court to whose jurisdiction the Issuer has submitted or any immunity waived by the Issuer in relation to legal proceedings arising out of or in connection with the Bonds;5 (xii) change the principal amount of outstanding Bonds or, in the case of a cross-series modification, the principal amount of debt securities of any other series required to approve a proposed modification in relation to the Bonds, the principal amount of outstanding Bonds required for a quorum to be present, or the rules for determining whether a Bond is outstanding for these purposes; or (xiii) change the definition of a reserved matter, and has the same meaning in relation to the debt securities of any other series save that any of the foregoing references to the Bonds or any agreement governing the issuance or administration of the Bonds are to be read as references to such other debt securities or any agreement governing the issuance or administration of such other debt securities. (f) ‘holder’ in relation to a Bond means the person in whose name the Bond is registered on the books and records of the Issuer [or] the bearer of the Bond [or] the person the Issuer is entitled to treat as the legal holder of the Bond,6 and in relation to any other debt security means the person the Issuer is enti-

1 To be included if the Bonds are redeemable. 2 To be included if the Bonds are guaranteed. 3 To be included ifthe Bonds are collateralised. 4 To be included ifthe Bonds are governed by a foreign law. 5 To be included, as appropriate, if the Issuer has submitted to the jurisdiction of a foreign court or expressly waived its immunity. 6 To be included as appropriate, based on the legal form of the Bond.



Appendix 5 

 241

tled to treat as the legal holder of the debt security under the law governing that debt security. (g) ‘record date’ in relation to any proposed modification means the date fixed by the Issuer for determining the holders of Bonds and, in the case ofa crossseries modification, the holders of debt securities of each other series that are entitled to vote on the modification.

2. Modification of Bonds 2.1 Reserved Matter Modification. The terms and conditions of the Bonds and any agreement governing the issuance or administration of the Bonds may be modified in relation to a reserved matter with the consent of the Issuer and: (a) the affirmative vote of holders of not less than 66 2/3% of the aggregate principal amount of the outstanding Bonds represented at a duly called meeting of Bondholders; or (b) a written resolution signed by or on behalf of holders of not less than 66 2/3% of the aggregate principal amount of the Bonds then outstanding. 2.2 Cross-Series Modification. In the case of a cross-series modification, the terms and conditions of the Bonds and debt securities of any other series, and any agreement governing the issuance or administration of the Bonds or debt securities of such other series, may be modified in relation to a reserved matter with the consent of the Issuer and: (a) (i) the affirmative vote of not less than 66 2/3% of the aggregate principal amount of the outstanding debt securities represented at separate duly called meetings of the holders of the debt securities of all the series (taken in the aggregate) that would be affected by the proposed modification; or (a) (ii) a written resolution signed by or on behalf of the holders of not less than 66 2/3% of the aggregate principal amount of the outstanding debt securities of all the series (taken in the aggregate) that would be affected by the proposed modification; and (b) (i) the affirmative vote of more than 50% of the aggregate principal amount of the outstanding debt securities represented at separate duly called meetings of the holders of each series of debt securities (taken individually) that would be affected by the proposed modification; or (b) (ii) a written resolution signed by or on behalf of the holders of more than 50% of the aggregate principal amount of the then outstanding debt securi-

242 

 Appendices

ties of each series (taken individually) that would be affected by the proposed modification. A separate meeting will be called and held, or a separate written resolution signed, in relation to the proposed modification of the Bonds and the proposed modification of each other affected series of debt securities. 2.3 Partial Cross-Series Modification. If a proposed cross-series modification is not approved in relation to a reserved matter in accordance with Section 2.2, but would have been approved if the proposed modification had involved only the Bonds and one or more, but less than all, of the other series of debt securities that would have been affected by the proposed modification, that cross-series modification will be deemed to have been approved notwithstanding Section 2.2 in relation to the Bonds and debt securities of each other series whose modification would have been approved in accordance with Section 2.2 if the proposed modification had involved only the Bonds and debt securities of such other series, upon the Issuer’s notice to this effect published within 10 days of the last approval required for the above-mentioned modification to be approved in accordance with this Section 2.3. 2.4 Multiple Currencies and Zero-Coupon Obligations. In determining whether a proposed modification has been approved by the requisite principal amount of Bonds and debt securities of one or more other series: (a) if the modification involves debt securities denominated in more than one currency, the principal amount of each affected debt security will be equal to the amount of euro that could have been obtained on the record date for the proposed modification with the principal amount of that debt security, using the applicable euro foreign exchange reference rate for the record date published by the European Central Bank; and (b) [if the modification involves debt securities that do not expressly provide for the accrual of interest (a ‘zero-coupon obligation’), the principal amount of each zero-coupon obligation will be equal to its present value on the record date for the proposed modification, determined by discounting the face amount of the zero-coupon obligation to the record date on the basis of the actual days elapsed in a 360 day year at: (i) if the zero-coupon obligation was formerly a component palt of a debt security that expressly provided for the accrual of interest, the coupon on that debt security; or (ii) if the zero-coupon obligation was never a component part of a debt security that expressly provided for the accrual of interest, the yield to maturity of the zero-coupon obligation at its issue price.





Appendix 5 

 243

If a zero-coupon obligation was issued in more than one tranche, its issue price for these purposes will be the weighted arithmetic average of the issue prices of all the previously issued tranches of that zero-coupon obligation.]7

2.5 Non-Reserved Matter Modification. The terms and conditions of the Bonds and any agreement governing the issuance or administration of the Bonds may be modified in relation to any matter other than a reserved matter with the consent of the Issuer and: (a) the affirmative vote of holders of more than 50% of the aggregate principal amount of the outstanding Bonds represented at a duly called meeting of Bondholders; or (b) a written resolution signed by or on behalf of holders of more than 50% of the aggregate principal amount of the outstanding Bonds. 2.6 Outstanding Bonds. In determining whether holders of the requisite principal amount of outstanding Bonds have voted in favour of a proposed modification or whether a quorum is present at any meeting of Bondholders called to vote on a proposed modification, a Bond will be deemed to be not outstanding, and may not be voted for or against a proposed modification or counted in determining whether a quorum is present, if on the record date for the proposed modification: (a) the Bond has previously been cancelled or delivered for cancellation or held for re issuance but not reissued; (b) the Bond has previously been called for redemption in accordance with its terms or prev iously become due and payable at maturity or otherwise and the Issuer has previously satisfied its obligation to make all payments due in respect of the Bond in accordance with its terms;8 or (c) the Bond is held by the Issuer, a department, ministry or agency of the Issuer, or a corporation, trust or other legal entity controlled by the Issuer or a department, ministry or agency of the Issuer, and in the case of a Bond held by any such corporation, trust or other legal entity the holder of the Bond does not have autonomy of decision, where:

7 Several aspects of Section 2.4(b) are subject to ongoing technical review, including the treatment of stripped index-linked bonds. A more detailed discussion of some of the key issues relating to zero-coupon obligations may be found in the Explanatory Note dated 25 July 2011, distributed together with the draft CAC. The applicable day-count convention will depend on local practice. 8 The reference to the Bond having previously been called for redemption to be included if the Bond is redeemable.

244 

 Appendices

(i) the holder of a Bond for these purposes is the entity legally entitled to vote the Bond for or against a proposed modification or, if different, the entity whose consent or instruction is by contract required, directly or indirectly, for the legally entitled holder to vote the Bond for or against a proposed modification; (ii) a corporation, trust or other legal entity is controlled by the Issuer or by a department, ministry or agency of the Issuer if the Issuer or any department, ministry or agency of the Issuer has the power, directly or indirectly, through the ownership of voting securities or other ownership interests, by contract or otherwise, to direct the management of or elect or appoint a majority of the board of directors or other persons performing similar functions in lieu of, or in addition to, the board of directors of that legal entity; and (iii) the holder of a Bond has autonomy of decision if, under applicable law, rules or regulations and independent of any direct or indirect obligation the holder may have in relation to the Issuer: (x) the holder may not, directly or indirectly, take instruction from the Issuer on how to vote on a proposed modification; or (y) the holder, in determining how to vote on a proposed modification, is required to act in accordance with an objective prudential standard, in the interest of all of its stakeholders or in the holder’s own interest; or (z) the holder owes a fiduciary or similar duty to vote on a proposed modification in the interest of one or more persons other than a person whose holdings of Bonds (if that person then held any Bonds) would be deemed to be not outstanding under this Section 2.6. 2.7 Outstanding Debt Securities. In determining whether holders of the requisite principal amount of outstanding debt securities of another series have voted in favor of a proposed cross-series modification or whether a quorum is present at any meeting of the holders of such debt securities called to vote on a proposed cross-series modification, an affected debt security will be deemed to be not outstanding, and may not be voted for or against a proposed cross-series modification or counted in determining whether a quorum is present, in accordance with the applicable terms and conditions of that debt security. 2.8 Entities Having Autonomy of Decision. For transparency purposes, the Issuer will publish promptly following the Issuer’s formal announcement of any proposed modification of the Bonds, but in no event less than five days prior to the record date for a proposed modification, a non-exhaustive list of corporations,



Appendix 5 

 245

trusts or other legal entities that are then controlled by the Issuer or by a department, ministry or agency of the Issuer and have autonomy of decision. 2.9 Exchange and Conversion. Any duly approved modification of the terms and conditions of the Bonds may be implemented at the option of the Issuer by means of a mandatory exchange or conversion of the Bonds for new debt securities containing the modified terms and conditions. Any conversion or exchange undertaken to implement a duly approved modification will be binding on all Bondholders.

3. Calculation Agent 3.1 Appointment and Responsibility. The Issuer will appoint a person (the ‘calculation agent’) to calculate whether a proposed modification has been approved by the requisite principal amount of outstanding Bonds and, in the case of a cross-series modification, by the requisite principal amount of outstanding debt securities of each affected series of debt securities. In the case of a cross-series modification, the same person will be appointed as the calculation agent for the proposed modification of the Bonds and each other affected series of debt securities. 3.2 Certificate. The Issuer will provide a certificate to the calculation agent prior to the date of any meeting called to vote on a proposed modification or the date fixed by the Issuer for the signing of a written resolution in relation to a proposed modification: (a) listing the total principal amount of Bonds and, in the case of a cross-series modification, debt securities of each other affected series outstanding on the record date; and (b) specifying the total principal amount of Bonds and, in the case ofa crossseries modification, debt securities of each other affected series that are deemed under Section 2.6(c) to be not outstanding on the record date, determined, if applicable, in accordance with the provisions of Section 2.4. The Issuer will arrange for this celtificate to be available for inspection and copying, during normal business hours at the office specified by the calculation agent, by Bondholders and, in the case of a cross-series modification, holders of debt securities of each other affected series.

246 

 Appendices

3.3 Reliance. The calculation agent may rely on any information contained in the certificate provided by the Issuer, and that information will be conclusive and binding on the Issuer and the Bondholders unless: (a) an affected Bondholder delivers a substantiated written objection to the Issuer in relation to the certificate before the vote on a proposed modification or the signing of a written resolution in relation to a proposed modification; and (b) that written objection, if sustained, would affect the outcome of the vote taken or the written resolution signed in relation to the proposed modification. In the event a substantiated written objection is timely delivered, any information relied on by the calculation agent will nonetheless be conclusive and binding on the Issuer and affected Bondholders if: (x) the objection is subsequently withdrawn; (y) the Bondholder that delivered the objection does not commence legal action in respect of the objection before a court of competent jurisdiction within 15 days of the vote taken or the written resolution signed in relation to the proposed modification; or (z) a court of competent jurisdiction subsequently rules either that the objection is not substantiated or would not in any event have affected the outcome of the vote taken or the written resolution signed in relation to the proposed modification. 3.4 Publication. The Issuer will arrange for the publication of the results of the calculations made by the calculation agent in relation to a proposed modification promptly following the meeting called to consider that modification or, if applicable, the date fixed by the Issuer for signing a written resolution in respect of that modification.

4. Bondholder Meetings; Written Resolutions 4.1 General. The provisions set out below, and any additional rules adopted by the Issuer that are consistent with the provisions set out below, will apply to any meeting of Bondholders called to vote on a proposed modification and to any written resolution adopted in connection with a proposed modification. Any action contemplated in this Section 4 to be taken by the Issuer may instead be taken by an agent acting on behalf of the Issuer.



Appendix 5 

 247

4.2 Convening Meetings. A meeting of Bondholders: (a) may be convened by the Issuer at any time; and (b) will be convened by the Issuer if an event of default in relation to the Bonds has occurred and is continuing and a meeting is requested in writing by the holders of not less than 10% of the aggregate principal amount of the Bonds then outstanding. 4.3 Notice of Meetings. The notice convening a meeting of Bondholders will be published by the Issuer at least 21 days prior to the date of the meeting or, in the case of an adjourned meeting, at least 14 days prior to the date of the adjourned meeting. The notice will: (a) state the time, date and venue of the meeting; (b) set out the agenda and quorum for, and the text of any resolutions proposed to be adopted at, the meeting; (c) specify the record date for the meeting, being not more than five days before the date of the meeting, and the documents required to be produced by a Bondholder in order to be entitled to participate in the meeting; (d) enclose the form of instrument to be used to appoint a proxy to act on a Bondholder’s behalf; (e) set out any additional rules adopted by the Issuer for the convening and holding of the meeting; and (f) identify the person appointed as the calculation agent for any proposed modification to be voted on at the meeting. 4.4 Chair. The chair of any meeting of Bondholders will be appointed: (a) by the Issuer; or (b) if the Issuer fails to appoint a chair or the person nominated by the Issuer is not present at the meeting, by holders of more than 50% of the aggregate principal amount of the Bonds then outstanding represented at the meeting. 4.5 Quorum. No business will be transacted at any meeting in the absence ofa quorum other than the choosing of a chair if one has not been appointed by the Issuer. The quorum at any meeting at which Bondholders will vote on a proposed modification of: (a) a reserved matter will be one or more persons present and holding not less than 66 2/3% of the aggregate principal amount of the Bonds then outstanding; and (b) a matter other than a reserved matter will be one or more persons present and holding not less than 50% of the aggregate principal amount of the Bonds then outstanding.

248 

 Appendices

4.6 Adjourned Meetings. If a quorum is not present within thirty minutes of the time appointed for a meeting, the meeting may be adjourned for a period of not more than 42 days and not less than 14 days as determined by the chair of the meeting. The quorum for any adjourned meeting will be one or more persons present and holding not less than 25% of the aggregate principal amount of the Bonds then outstanding. 4.7 Written Resolutions. A written resolution signed by or on behalf of holders of the requisite majority of the Bonds will be valid for all purposes as if it was a resolution passed at a meeting of Bondholders duly convened and held in accordance with these provisions. A written resolution may be set out in one or more document in like form each signed by or on behalf of one or more Bondholders. 4.8 Entitlement to Vote. Any person who is a holder of an outstanding Bond on the record date for a proposed modification, and any person duly appointed as a proxy by a holder of an outstanding Bond on the record date for a proposed modification, will be entitled to vote on the proposed modification at a meeting of Bondholders and to sign a written resolution with respect to the proposed modification. 4.9 Voting. Every proposed modification will be submitted to a vote of the holders of outstanding Bonds represented at a duly called meeting or to a vote of the holders of all outstanding Bonds by means of a written resolution without need for a meeting. A holder may cast votes on each proposed modification equal in number to the principal amount of the holder’s outstanding Bonds. For these purposes: (a) in the case of a cross-series modification involving debt securities denominated in more than one currency, the principal amount of each debt security will be determined in accordance with Section 2.4(a); and (b) in the case of a cross-series modification involving zero-coupon obligations, the principal amount of each zero-coupon obligation will be equal to the principal amount of the zero-coupon obligation determined in accordance with Section 2.4(b). 4.10 Proxies. Each holder of an outstanding Bond may, by an instrument in writing executed on behalf of the holder and delivered to the Issuer not less than 48 hours before the time fixed for a meeting of Bondholders or the signing of a written resolution, appoint any person (a “proxy”) to act on the holder’s behalf in connection with any meeting of Bondholders at which the holder is entitled to vote or the signing of any written resolution that the holder is entitled to sign.



Appendix 5 

 249

Appointment of a proxy pursuant to any form other than the form enclosed with the notice of the meeting will not be valid for these purposes. 4.11 Legal Effect and Revocation of a Proxy. A proxy duly appointed in accordance with the above provisions will, subject to Section 2.6 and for so long as that appointment remains in force, be deemed to be (and the person who appointed that proxy will be deemed not to be) the holder of the Bonds to which that appointment relates, and any vote cast by a proxy will be valid notwithstanding the prior revocation or amendment of the appointment of that proxy unless the Issuer has received notice or has otherwise been informed of the revocation or amendment at least 48 hours before the time fixed for the commencement of the meeting at which the proxy intends to cast its vote or, if applicable, the signing of a written resolution. 4.12 Binding Effect. A resolution duly passed at a meeting of holders convened and held in accordance with these provisions, and a written resolution duly signed by the requisite majority of Bondholders, will be binding on all Bondholders, whether or not the holder was present at the meeting, voted for or against the resolution or signed the written resolution. 4.13 Publication. The Issuer will without undue delay publish all duly adopted resolutions and written resolutions.

5. Publication 5.1 Notices and Other Matters. The Issuer will publish all notices and other matters required to be published pursuant to the above provisions: (a) on [insert the Issuer’s website for financial notices]; and (b) in such other places, including in [insert the Issuer’s official gazette], and in such other manner as may be required by applicable law or regulation.

Appendix 6 EFC Sub-Committee on EU Sovereign Debt Markets Collective Action Clause Explanatory Note dated 26 July 2011 1. Introduction On 28 November 2010, euro area finance ministers announced a number of policy measures intended to safeguard financial stability in the euro area, among them the mandatory inclusion of standardised collective action clauses (“CACs”) in all new euro area government securities beginning in July 2013. This commitment was later confirmed by the Heads of Government and States of euro area Member States at their meeting on 11 March 2011 and by the European Council at the summit held on 24–25 March 2011. All of the euro area Member States have also agreed to take all action necessary to implement the standardised CAC. The decisions taken by the euro area authorities contemplate that the standardised CAC to be introduced in July 2013 will: –– be based on those commonly used in the UK and US after the G10 report on CACs of 26 September 2002; –– be included in all new euro area government securities with a maturity of more than one year; –– have uniformity of application and provide a level playing field for all euro area Member States; –– allow a proposed modification of a euro area government’s securities to be made binding on all holders of the affected securities if approved by holders of the requisite principal amount of the affected securities; –– facilitate the agreement of private-sector creditors to the possible modification of euro area government debt securities that contain a standardised CAC; and –– not increase the probability of a euro area issuer defaulting on or modifying its debt securities containing a standardised CAC. The EFC Sub-Committee on EU Sovereign Debt Markets (the “Committee”) was tasked with producing the standardised CAC called for by the euro area authorities. Following an extensive review process, the Committee has approved the enclosed draft CAC. The draft CAC will be the subject of the public consultation process described in the accompanying note from the Chair of the Committee. Subject to the outcome

252 

 Appendices

of that process, the Committee contemplates that the standardised CAC will form a part of the definitive documentation for all covered euro area government debt securities beginning in July 2013, including all domestic and international bonds, irrespective of their governing law, issued by euro area governments. While euro area governments will be allowed to reopen (‘tap’) existing securities issuances after July 2013 to the extent necessary to ensure adequate liquidity, the Committee envisions that issuers will avail themselves of this opportunity only in limited cases.

2. Scope of Application The standardised CAC will be mandatory for all new bonds, notes and other debt securities with an original stated maturity of more than one year issued by national euro area governments from July 2013, regardless of whether the debt security is listed on a securities exchange, is actively traded or was privately placed. The Committee considered making the standardised CAC also mandatory for debt securities issued by regional and local euro area governments, for actively traded syndicated loans contracted by covered borrowers and for debt securities guaranteed by covered guarantors. The Committee ultimately concluded not to make the standardised CAC mandatory for any of these governmental obligations because, among other things, they at present represent only a very small portion of total euro area governmental indebtedness. The Committee’s decision does not, however, preclude the voluntary introduction of a CAC in any euro area governmental obligation. For ease of reading, any euro area government debt security that must contain a standardised CAC is generally referred to as a bond in this Explanatory Note.

3. Summary of Key Provisions The draft CAC distinguishes between three types of modification: –– a reserved matter modification, involving the modification of a bond’s most important terms and conditions and requiring the highest level of bondholder consent; and –– a non-reserved matter modification, involving the modification of a bond’s less important terms and conditions and requiring a lower level of bondholder consent.



Appendix 6 

 253

Under the draft CAC, a modification may be proposed in relation to either a single series of bonds or more than one series of bonds – a so-called cross-series modification. Different approval thresholds apply to single-series and cross-series modifications. In all events, a proposed modification will require the issuer’s own consent, and may also require the consent of other interested parties (for example, a fiscal or paying agent) under other provisions of an issuer’s definitive documentation. Interested-party approvals are not addressed in the standardised CAC.

A. Single-Series Approval Thresholds The affirmative vote of holders of not less than 66 2/3% of the outstanding principal amount of the bonds represented at a meeting of bondholders at which a quorum is present is required to modify a reserved matter in relation to a single series of bonds. A reserved-matter modification may also be approved by a written resolution signed by holders of 66 2/3% of the principal amount of all outstanding bonds. The required approval threshold in the case of a non-reserved matter modification is a simple majority of the outstanding principal amount of the affected series of bonds represented at a duly called meeting of bondholders or a simple majority of the principal amount of all outstanding bonds if the modification is approved by a written resolution. A leading international securities trade association has previously expressed its support for a model CAC that allowed a reserved-matter modification to be approved by 66 2/3% of the outstanding principal amount of all affected bonds, and a non-reserved-matter modification to be approved by more than 50% of the outstanding principal amount of bonds represented at a quorate meeting. The Committee believes it appropriate that a reserved-matter modification may also be approved by at least 66 2/3% of the outstanding principal amount of bonds represented at a duly called meeting of bondholders. Even though bondholder meetings are rarely, if ever, attended by all holders, the likely disproportionate representation at a meeting of investors opposed to a proposed modification will in practice, in the Committee’s view, often make it as difficult to obtain the same nominal level of approval at a meeting of bondholders as of all outstanding bonds.

B. Cross-Series Approval Thresholds The draft CAC provides that more than one series of bonds may be modified in relation to a reserved matter with the affirmative vote of holders of at least:

254 

 Appendices

–– 66 2/3% of the outstanding principal amount of all the affected series of bonds (taken in the aggregate) represented at duly called meetings of the affected bondholders; and –– 50% of the outstanding principal amount of each affected series of bonds (taken individually) represented at a duly called meeting of the bondholders of each such series. The same approval thresholds are also required in the case of a cross-series modification approved by a written resolution, though, as with the modification of a single series of bonds, the approval thresholds are in that event expressed in terms of all affected bonds then outstanding. The Committee gave careful consideration to allowing a cross-series modification to be approved on an aggregate all-series basis only. The Committee ultimately rejected this approach notwithstanding its evident administrative advantages (and its use in many statutory insolvency regimes) because of the perceived unfairness of allowing a series of bonds to be modified over the objection of a majority of the affected holders in the absence of any supervising judicial authority, as well as related legal concerns that a single aggregate approval threshold might not be enforceable throughout the euro area. In the event a proposed cross-series modification is approved as to some but not all of the affected series of bonds, an issuer may implement the proposed modification in relation to those series of bonds whose modification would have been approved if the cross-series modification had initially been proposed only in respect of the bonds of those series. The draft CAC does not distinguish between a cross-series modification involving bonds governed by the same law and one involving bonds governed by different laws. The Committee is not aware of any legal obstacle to allowing a cross-law modification so long as the modification of each affected series of bonds is approved at a separate meeting of bondholders or a separate written resolution. In providing for a separate approval process for each affected series, the draft CAC avoids a possible conflict between the laws governing the bonds of different affected series. Cross-series modifications and their partial and cross-law variants are intended to facilitate the broadest possible modification of an issuer’s bonds, with a view to affording the greatest possible equality of treatment to the holders of an issuer’s bonds.

C. Quorums A proposed modification may not be approved at a meeting of bondholders unless a quorum is present. The quorum of any initial meeting called to consider a proposed modification is 66 2/3% of all then outstanding bonds in the case



Appendix 6 

 255

of a reserved matter modification and 50% of all then outstanding bonds in the case of the proposed modification of a non-reserved matter. The quorum for all adjourned meetings is 25% of the outstanding bonds irrespective of the type of modification to be voted on. The Committee is aware of the concerns expressed by some market participants that even the high quorums included in the draft CAC could theoretically result in a reserved matter being approved by a relatively small percentage of the total affected bonds then outstanding, especially in the case of a vote held at an adjourned meeting. The Committee believes these concerns are misplaced. Experience suggests that a significant percentage of the principal amount of outstanding bonds will in fact be represented at a meeting called to vote on the proposed modification of a euro area government’s debt securities. In the unlikely event a quorum is not present at the initial meeting called to vote on a reserved-matter modification, the Committee is of the view that the apparent indifference of the absent bondholders can fairly be understood as constituting their silent acquiescence to the proposed modification.

D. Disenfranchisement Consistent with market practice, the draft CAC disenfranchises government bonds held by the issuer or by any of its ministries, departments or agencies. An issuer’s holdings of its own bonds, and the bonds held by other governmental bodies, are therefore treated by the draft CAC as not outstanding for purposes of determining whether a proposed modification has been approved, and may not be voted for or against a proposed modification or counted towards the quorum required for a duly called meeting of bondholders. The Committee believes that disenfranchising an issuer’s holdings of its own bonds is appropriate because the losses suffered by the issuer from the modification of the bonds it holds, unlike the losses suffered by an ordinary market participant, are more than offset by the gains realised by the issuer from the resulting reduction in its debt service or debt stock or both. The Committee also agrees that an issuer should not be allowed to accomplish through indirection – by instructing government-controlled companies to vote in favour of a proposed modification – that which is denied the issuer directly. Where, however, a government-controlled legal entity (referred to below as a government-controlled ‘company’ regardless of its legal form) has autonomy of decision and is required by applicable law, rule or regulation to act independently of any instruction given by an issuer, the Committee is strongly of the view that it would be inconsistent with the issuer’s own laws to disenfranchise a company’s holdings of government securities on the basis of its assumed disregard

256 

 Appendices

of binding legal obligations. To do otherwise would not only undermine the presumption of lawful action enjoyed by all euro area companies, but would also call into question the issuer’s commitment to its own laws intended to ensure a company’s autonomy of decision. The draft CAC accordingly casts the net wide in treating a company as directly or indirectly controlled by an issuer even if the issuer has never sought to instruct the company on the day-to-day management of its affairs or investments, but does not disenfranchise a government-controlled company that has autonomy of decision under applicable law. In furtherance of this arrangement, the draft CAC includes three safe-harbours that expressly exempt companies having autonomy of decision from the otherwise applicable disenfranchisement of their holdings of government securities. The first safe-harbour is available to government-controlled companies that under applicable law may not take instruction from a euro area government issuer. The second safe­harbour is available to a government-controlled company that is required under applicable law to act in accordance with an objective prudential standard, in the interest of all of its stakeholders (and not just its shareholders) or in its own self interest. The final safe-harbour is available to an issuer-controlled company that owes a fiduciary (or similar) duty to one or more third parties independently of any obligation it may owe to its controlling parent. The Committee believes that all of the safe harbours are consistent with the underlying rationale of the disenfranchisement clause – to prevent an issuer from voting in favour of the modification of its own securities either directly or by instructing companies it controls to so vote their holdings of the issuer’s debt securities. The Committee also considered but rejected two other related grounds sometimes offered as a reason for disenfranchising issuer-controlled companies – their supposed predictability of action and their assumed motivation. The Committee does not believe that predictability of action constitutes a sound basis for disenfranchisement. The Committee is mindful in this regard that the actions of many private creditors are at least as predictable as those of government-controlled companies – for example, the holder of a credit default swap is almost certain to vote in favour of a proposed modification that will result in a ‘credit event’ – and yet it has never been suggested that private investors should be disenfranchised because their actions are often predictable in practice. In the Committee’s view, a creditor’s motivation provides an even less appropriate basis for disenfranchisement. Experience suggests that it is often difficult, if not impossible, to discern a creditor’s real motivation, or to distinguish among the arguably acceptable and arguably unacceptable motivations that may together inform an investor’s decision. In any event, it is not clear to the Commit-



Appendix 6 

 257

tee what motives would even arguably constitute grounds for an investor’s disenfranchisement. It will be recalled that issuers often pay early-participation fees to private creditors to encourage their public support for a proposed modification of the issuer’s bonds, without calling into question the creditors’ right to have their votes counted. To enhance the transparency of the voting process, an issuer is required under the standardised CAC to publish, promptly following the formal announcement of a proposed modification (but in any event not less than five days before the record date for that modification), a non-exhaustive list of issuer-controlled companies that have autonomy of decision.

E. Reserved Matters The list of reserved matters included in the standardised CAC is broadly consistent with the list of reserved matters found in other CACs currently in the market. In light of the wide range of debt securities covered by the draft CAC – bearer, registered and dematerialised bonds, zero-coupon and interest bearing bonds, redeemable and non-redeemable bonds and domestic and international bonds, among others – the list of reserved matters is necessarily expressed in more general terms than might otherwise have been the case, and often combines items that sometimes receive separate treatment in other CACs. For example, the draft CAC treats as a reserved matter a reduction in ‘any amount’ payable on a bond. This formulation covers, among other things, any reduction in the principal, interest or ‘additional amounts’ that may be payable on that bond. The standardised CAC also treats as a reserved matter a modification that would impose a condition on or otherwise modify an issuer’s obligation to make payments on a bond. This formulation covers, among other things, a change in the nature of the issuer’s obligation – for example, a modification of the issuer’s unconditional obligation to make bond payments – or the substitution of a new obligor in place of the original issuer through an exchange of existing bonds for a new series of bonds. Still other reserved matters listed in the standardised CAC are relevant for some, but not all, government securities. The draft CAC accordingly treats these items as reserved matters only where they are relevant to an investor. For example, the standardised CAC treats a change in the law governing a bond as a reserved matter only if a bond is governed by a law other than the law of the issuer. The Committee believes there is no need to treat a change from the issuer’s own domestic law as a reserved matter because the issuer already has the power, at least in theory, to adopt any desired modification by means of domestic legislation without changing the law governing its bonds. In so treating a change in gov-

258 

 Appendices

erning law, the Committee does not mean to suggest that euro area government issuers will in fact exercise their sovereignty in relation to any series of bonds, or that modifying the terms of a bond by means of domestic legislation does not itself raise difficult legal issues. Other items on the list of reserved matters are manifestly not relevant to some euro area government securities. For example, a change in the terms on which collateral is pledged to secure payment of a series of bonds is relevant only in the case of bonds that are collateralised in the first place. These items are footnoted in the standardised CAC to indicate when they are to be included in the terms and conditions of a bond.

F. Technical Amendments The standardised CAC does not include a technical amendments clause because, in the Committee’s view, it is not an essential element of a CAC. Under the technical amendments clause found in many international debt obligations, an issuer may amend the terms and conditions of a bond, or an agreement governing the issuance or administration of a series of bonds, without the consent of bondholders (a) to correct a manifest error or cure an ambiguity, (b) if the modification is of a formal, minor or technical nature, or (c) if the modification does not materially prejudice the interests of bondholders. All euro area Member States will be free to include an international-style technical amendments clause in the definitive documentation for their debt securities or a technical amendments clause of the type customary in the issuer’s own market.

G. Calculation Agent All euro area issuers will be required to appoint a calculation agent to determine whether a proposed modification has been approved by the requisite principal amount of outstanding bonds. For ease of administration, the same person will be appointed as the calculation agent for each series of affected bonds in the case of a cross-series modification. The Committee anticipates that the calculation agent for any bonds will often be the same person as the fiscal or paying agent appointed by the issuer for those bonds. The calculation agent may rely on a certificate prepared by the issuer, listing the principal amount of outstanding bonds and specifying the principal amount of bonds then deemed to be not outstanding because they are held by the issuer or by a company controlled by the issuer that does not have autonomy of decision. Any information relied on by the issuer will be conclusive and binding on the issuer and bondholders, unless an affected bondholder timely delivers a sub-



Appendix 6 

 259

stantiated written objection that, if sustained, would affect the outcome of the vote taken on a proposed modification. The issuer and affected bondholders will also be bound by any information relied on by the calculation agent if the objection is later withdrawn or a court subsequently rules that the objection is either not substantiated or would not have affected the outcome of the vote taken.

H. Bondholder Meetings The standardised CAC includes mandatory rules for holding bondholder meetings, intended to ensure that the CAC operates in the same manner and with the same legal effect in all euro area Member States. In light of the wide diversity of bondholder meeting rules and practices in the euro area, issuers may adopt supplementary rules for the holding of bondholder meetings that are consistent with the mandatory rules included in the standardised CAC. For example, the documentary evidence required to be presented by a bondholder in order to vote on a proposed modification will vary depending on the nature – registered, bearer or dematerialised – of the debt securities in question. The standardised CAC accordingly contemplates that the issuer will specify the required evidence in the notice convening a bondholder meeting.

I. Zero-Coupon and Index-Linked Obligations The draft CAC provides special treatment for zero-coupon obligations, by which is meant both bonds originally issued without express provision for the accrual of interest, as well as so-called stripped bonds consisting of the component parts of a bond that originally expressly provided for the accrual of interest. In the absence of special treatment, the holder of a zero-coupon obligation would enjoy preferential voting rights as compared to the holder of an interest-bearing bond. An example may be helpful in illustrating the problem. If an interest-bearing bond were to be stripped into its component parts, and the holders of the stripped components were to be afforded voting rights based on the face amount of their holdings, the total voting rights enjoyed by the holders of all the stripped components would exceed the total votes that would otherwise have been cast by the holder of the original interest-bearing bond. In order to remedy this situation, the draft CAC adjusts the face amount of each stripped component so that the holders of the stripped components have, in the aggregate, the same voting rights as the original investor would individually have had in relation to the underlying interest-bearing bond. The draft CAC includes a similar mechanism for obligations issued without any express provision for the accrual of interest. Consistent with the approach adopted more gen-

260 

 Appendices

erally for interest-bearing obligations, the draft CAC assumes, in the case of a stripped bond, that the original interest-bearing bond was issued at par. The draft CAC does not include a special rule for amortising bonds, indexlinked obligations or other obligations whose principal amount may change over time, as any increase or reduction in their principal amount is already provided for in their terms and conditions. In determining whether a proposed modification has received the requisite approval, the principal amount of an obligation whose principal amount may change over time will be its outstanding principal amount on the record date for the proposed modification. The treatment of zero-coupon obligations inevitably raises complicated technical issues (for example, the treatment of stripped index-linked bonds), and the provisions of the draft CAC dealing with zero-coupon obligations are subject to ongoing technical review.

J. Governing Law and Enforcement The Committee gave careful consideration to having the standardised CAC governed by the law of one euro area Member State, to including the standardised CAC in a treaty binding on all euro area Member States, and to having all disputes arising out of or relating to the standardised CAC be heard and resolved in an agreed international forum or before the courts of one euro area Member State. The Committee ultimately concluded that the proposed arrangements either would not significantly enhance the uniformity of application of the CAC throughout the euro area, or that the desired uniformity could be achieved through other means more consistent with well-established market practice. The CAC included in any bond issued by a euro area Member State will instead be governed by the law that governs that bond more generally, and any dispute arising out of or relating to the CAC included in a bond issued by a euro area Member State will be resolved on the same basis and before the same courts as are all other disputes arising out of or in connection with any other provision of that bond.

4. Drafting Notes For ease of reading, the draft CAC refers to the covered debt securities as Bonds. The actual CAC included in each issuer’s definitive documentation will of course identify the offered securities by their proper name. The standardised CAC also refers to bonds or other debt securities that will be ‘affected’ by a proposed modification. These references are in all cases to be understood as references to the bonds and other debt securities whose terms and conditions are the subject of the



Appendix 6 

 261

proposed modification, and not to any other bonds or debt securities that may be economically affected by the proposed modification. 26 July 2011