An Economic Analysis of Bilateral Investment Treaties by tuanthien123

VIEWS: 45 PAGES: 264

									Jan Peter Sasse

An Economic Analysis of Bilateral Investment Treaties
GABLER RESEARCH

Ökonomische Analyse des Rechts
Herausgegeben von
Professor Dr. Peter Behrens
Professor Dr. Thomas Eger
Professor Dr. Manfred Holler
Professor Dr. Claus Ott
Professor Dr. Hans-Bernd Schäfer
Professor Dr. Stefan Voigt (schriftführend)
Universität Hamburg, Fakultät für Rechtswissenschaft
und Fakultät für Wirtschafts- und Sozialwissenschaft




Die ökonomische Analyse des Rechts untersucht Rechtsnormen auf ihre gesell-
schaftlichen Folgewirkungen und bedient sich dabei des methodischen Instru-
mentariums der Wirtschaftswissenschaften, insbesondere der Mikroökonomie,
der Neuen Institutionen- und Konstitutionenökonomie. Sie ist ein interdisziplinäres
Forschungsgebiet, in dem sowohl Rechtswissenschaftler als auch Wirtschafts-
wissenschaftler tätig sind und das zu wesentlichen neuen Erkenntnissen über
Funktion und Wirkungen von Rechtsnormen geführt hat.
Die Schriftenreihe enthält Monographien zu verschiedenen Rechtsgebieten und
Rechtsentwicklungen. Sie behandelt Fragestellungen aus den Bereichen Wirt-
schaftsrecht, Vertragsrecht, Haftungsrecht, Sachenrecht und verwaltungsrecht-
liche Regulierung.
Jan Peter Sasse
An Economic Analysis
of Bilateral Investment Treaties
With a foreword by Prof. Dr. Thomas Eger




                                           RESEARCH
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.




Dissertation Universität Hamburg, 2010




1st Edition 2011

All rights reserved
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
Editorial Office: Stefanie Brich | Sabine Schöller
Gabler Verlag is a brand of Springer Fachmedien.
Springer Fachmedien is part of Springer Science+Business Media.
www.gabler.de

                   No part of this publication may be reproduced, stored in a retrieval system
                   or transmitted, in any form or by any means, electronic, mechanical, photo-
                   copying, recording, or otherwise, without the prior written permission of the
                   copyright holder.

Registered and/or industrial names, trade names, trade descriptions etc. cited in this publica-
tion are part of the law for trade-mark protection and may not be used free in any form or by
any means even if this is not specifically marked.

Cover design: KünkelLopka Medienentwicklung, Heidelberg
Printed on acid-free paper
Printed in the Netherlands

ISBN 978-3-8349-2756-9
Foreword

The global economy is characterized not only by a steady increase of international
trade, but also by growing flows and stocks of foreign direct investment (FDI). This
development has manifested itself especially since the mid-1980s. The importance of
FDI in providing foreign markets with goods and services has become comparable to
trade. FDI also constitutes a crucial source of external finance for developing
countries.
An important prerequisite for a high level of FDI is that investors view the political
risk of the host country as manageable. Political risk may, for example, comprise the
danger of expropriation of the investment without adequate compensation, or more
subtle regulatory measures with comparable effects, which are often referred to as
indirect expropriation. The most important legal instruments in international
investment law that may mitigate this kind of risk are bilateral investment treaties
(BITs). Since the first BIT, concluded between Germany and Pakistan in 1959, the
number of such treaties has risen to the impressive number of more than 2700 in 2010.
Despite the undeniable importance of FDI for the global economy and the growing
prominence of BITs, the economic analysis has mainly focussed on international trade
law and has thus far neglected the analysis of international investment law. This
contribution by Jan Peter Sasse seeks, and succeeds, to fill that gap in the existing
literature.
Jan Peter Sasse begins his research by providing a comprehensive analysis of the
economic and legal tools available to international investors who wish to safeguard
their assets abroad. In contrast to the domestic context, international law is generally
characterized by the absence of a supranational authority that may enforce legal
obligations through coercion. Consequently, investors have to rely on economic
devices (the exchange of hostages, for example) or on legal protection through
international law and the possibly unfavourable domestic legislation and judiciary of
host-states. Especially in the light of the lack of a truly multinational treaty on the
protection of FDI, bilateral investment treaties can be considered a cornerstone of
international investment law. Consequently, a number of recent empirical studies find
a positive relationship between the conclusion of BITs and the amount of FDI flows.
The economic analysis of the functioning of BITs must first of all illuminate the
relationship between the investor and the host-state. This relationship may be
characterized by problems of hold-up and asymmetric information. Based on the
rational choice principle as advanced recently by scholars of international law and
economics, Jan Peter Sasse shows to what extent the host-state can overcome the hold-
VI                                                                               Foreword

up problem through self-commitment. Furthermore, the author convincingly argues
that signaling is only of limited value in counteracting the asymmetry of information
between investors and host-states.
Apart from the relationship between investors and host-states, a thorough analysis of
bilateral investment treaties has to take into account the strategic interaction among
host-states that may consider BITs as an instrument in the competition for FDI. Jan
Peter Sasse analyses this competition in great institutional detail with special attention
devoted to the environmental conflicts that have emerged in the context of
international arbitration. His results show that BITs, if drafted and interpreted with
care, may help capture the beneficial effects of institutional competition and mitigate
the potentially detrimental effects. The author then contributes to two ongoing
discussions in the area of international investment law. Firstly, Jan Peter Sasse
provides an empirical analysis of the impact of BITs on the institutional quality of
developing countries. He finds that external effects of BITs on institutional quality
cannot be verified. The study thus lends empirical support neither to the optimistic
calls for BITs as serving as a positive example for developing countries, nor to the
pessimistic accounts that regard BITs as detrimental to domestic institutional quality.
Secondly, the author provides a systematic analysis of the issue of transparency vs.
confidentiality in international arbitration. Jan Peter Sasse explains why more
transparency may be harmful for the parties involved and that, as a consequence, a
movement towards more transparency in international investment arbitration will be
hard to achieve.
This publication in the field of international law and economics makes a valuable
contribution to our understanding of the functioning of bilateral investment treaties.
Jan Peter Sasse also provides an insightful and well-researched analysis of different
aspects of the protection of FDI through BITs, including institutional competition,
institutional quality, and transparency.




Prof. Dr. Thomas Eger
Acknowledgements

This thesis was written while I was a student at the Doctoral College for Law and
Economics (Graduiertenkolleg für Recht und Ökonomik) and, subsequently, a research
assistant at the Institute of Law and Economics, both at the University of Hamburg. It
would not have been successfully finished without the help and support of a number of
people. I am especially grateful to Thomas Eger, who supervised the thesis and
allowed me to benefit from his helpful comments and advice. Hans-Bernd Schäfer
agreed to act as the second reviewer and was of great support to me (and my fellow
colleagues) as the Speaker of the Graduiertenkolleg. I am grateful to him, other
professors and fellow colleagues, in and around the Graduiertenkolleg, for many
inspiring interactions und discussions during my time in Hamburg. Also, I want to
thank the chairman of the disputation committee, Manfred Holler, who not only
handled the formal dissertation process very efficiently, but also provided helpful
comments during the research process. My thanks go to the editors of the scientific
series “Ökonomische Analyse des Rechts” for giving me the opportunity to publish my
thesis in the series: Peter Behrens, Thomas Eger, Manfred Holler, Claus Ott, Hans-
Bernd Schäfer and Stefan Voigt. I had the pleasure of conducting research as a visiting
scholar at Columbia University (New York) and my gratitude goes to Avery W. Katz
for making this possible. The German Research Foundation (DFG) financially
supported my research both in Germany and the United States with generous research
grants and I am thankful for that.
Additionally, I am indebted to Anne van Aaken, Peter Behrens, Stephania Bonilla,
Eberhard Feess, Henning Fräßdorf, Andrew Guzman, Jonathan Klick, Frank Müller-
Langer, Jan Matauschek, Susan Russell, Stephan Wittig, Katherine Walker and
Tammy de Wright for helpful advice, comments and inspiration during my research
process. Focusing on interdisciplinary research may not be the obvious choice for a
young student of economics. I owe my passion for the economic analysis of law to two
professors who exposed me to this field of research while I was still a graduate student
at the Humboldt University in Berlin: Charles B. Blankart and Christian Kirchner. I
want to thank them for putting me on the right track.
Most importantly, I would like to take this opportunity to thank my father, my sister
and Katrin Stägert for their loving support, kind words, patience and encouragement.
This book is dedicated to the memory of my mother, Hanne Sasse.


Jan Peter Sasse
Summary Contents

FOREWORD...................................................................................................................V
ACKNOWLEDGEMENTS ........................................................................................... VII
SUMMARY CONTENTS ............................................................................................... IX
DETAILED CONTENTS............................................................................................... XI
LIST OF FIGURES ..................................................................................................... XIX
LIST OF ABBREVIATIONS........................................................................................ XXI
LIST OF VARIABLES ..................................................................................................... 1


1      INTRODUCTION ................................................................................................. 1
2       FOREIGN DIRECT INVESTMENT .................................................................... 6
3      ECONOMIC AND LEGAL PROTECTION OF FDI......................................... 17
4      THE ECONOMICS OF BITS ............................................................................. 67
5      BITS AND INSTITUTIONAL COMPETITION ............................................. 124
6      BITS AND INSTITUTIONAL QUALITY....................................................... 155
7      BITS AND TRANSPARENCY ........................................................................ 177
8       SUMMARY AND OUTLOOK......................................................................... 199


APPENDICES............................................................................................................. 205
BIBLIOGRAPHY......................................................................................................... 221
Detailed Contents

FOREWORD...................................................................................................................V
ACKNOWLEDGEMENTS ........................................................................................... VII
SUMMARY CONTENTS ............................................................................................... IX
DETAILED CONTENTS............................................................................................... XI
LIST OF FIGURES ..................................................................................................... XIX
LIST OF ABBREVIATIONS........................................................................................ XXI
LIST OF VARIABLES ..................................................................................................... 1


1       INTRODUCTION ............................................................................................... 1
    1.1 INTERNATIONAL LAW AND ECONOMICS ............................................................ 2
    1.2 STRUCTURE ........................................................................................................ 4


2       FOREIGN DIRECT INVESTMENT ................................................................ 6
    2.1 TRENDS AND FIGURES ........................................................................................ 6
    2.2 MULTINATIONAL ENTERPRISES AND FDI........................................................... 8
    2.3 FDI AND DEVELOPMENT .................................................................................. 12


3       ECONOMIC AND LEGAL PROTECTION OF FDI ................................... 17
    3.1 TIME INCONSISTENCY AND EXPROPRIATION RISK ........................................... 17
       3.1.1       Time Inconsistency................................................................................... 17
       3.1.2       On the Relevance of Expropriation Risk ................................................. 22
    3.2 THE ECONOMICS OF FDI PROTECTION ............................................................. 22
       3.2.1       Static Devices........................................................................................... 23
           3.2.1.1 Hostages................................................................................................ 24
           3.2.1.2 Collateral............................................................................................... 27
           3.2.1.3 Hands-Tying ......................................................................................... 29
           3.2.1.4 Union..................................................................................................... 30
XII                                                                                                      Detailed Contents

        3.2.1.5 Insurance ............................................................................................... 31
        3.2.1.6 Devaluation of Assets ........................................................................... 32
        3.2.1.7 Lobbying ............................................................................................... 32
      3.2.2       Dynamic Devices ..................................................................................... 33
        3.2.2.1 Expertise and Time ............................................................................... 33
        3.2.2.2 Repetition and Reputation .................................................................... 33
              3.2.2.2.1 Repetition...................................................................................... 33
              3.2.2.2.2 Reputation..................................................................................... 35
      3.2.3       Discussion ................................................................................................ 40
  3.3 LEGAL FDI PROTECTION .................................................................................. 40
      3.3.1       Domestic Regulation................................................................................ 41
      3.3.2       Customary International Law.................................................................. 41
      3.3.3       Multilateral Treaties ................................................................................ 43
        3.3.3.1 WTO ..................................................................................................... 44
        3.3.3.2 Energy Charter Treaty .......................................................................... 44
        3.3.3.3 NAFTA ................................................................................................. 45
      3.3.4       Investor-State Contracts .......................................................................... 45
      3.3.5       Bilateral Investment Treaties................................................................... 45
        3.3.5.1 Overview and History........................................................................... 46
        3.3.5.2 Treaty Practice and Treaty Interpretation ............................................. 47
        3.3.5.3 Preamble and Definitions...................................................................... 48
              3.3.5.3.1 Investor ......................................................................................... 48
              3.3.5.3.2 Investment..................................................................................... 49
        3.3.5.4 Admission ............................................................................................. 50
        3.3.5.5 Standards of Treatment......................................................................... 50
              3.3.5.5.1 Fair and Equitable Treatment ....................................................... 50
              3.3.5.5.2 Most-Favoured-Nation Treatment ................................................ 52
              3.3.5.5.3 Additional Standards..................................................................... 53
        3.3.5.6 Expropriation and Compensation ......................................................... 54
Detailed Contents                                                                                                          XIII

               3.3.5.6.1 Indirect Expropriation................................................................... 55
               3.3.5.6.2 Compensation ............................................................................... 56
          3.3.5.7 Public Concerns .................................................................................... 57
          3.3.5.8 The Settlement of Disputes................................................................... 58
               3.3.5.8.1 ICSID ............................................................................................ 59
               3.3.5.8.2 Remedies....................................................................................... 61
               3.3.5.8.3 Enforcement and Execution of Arbitral Awards .......................... 62
               3.3.5.8.4 Costs.............................................................................................. 63
               3.3.5.8.5 Empirical Aspects of Arbitration.................................................. 63
          3.3.5.9 BITs and Customary International Law ............................................... 64
          3.3.5.10 Summary ........................................................................................... 65


4      THE ECONOMICS OF BITS .......................................................................... 67
    4.1 THE (PERCEIVED) WEAKNESS OF INTERNATIONAL LAW ................................. 67
    4.2 THE EFFECT OF BITS ON FDI ........................................................................... 69
       4.2.1       Empirical Studies on the Effect of BITs on FDI ...................................... 69
       4.2.2       Discussion ................................................................................................ 72
    4.3 LAW AND ECONOMICS OF INTERNATIONAL LAW ............................................. 73
       4.3.1       Theories of International Law ................................................................. 74
       4.3.2       Rational Choice Approach....................................................................... 74
       4.3.3       Methodological Individualism ................................................................. 75
       4.3.4       The Three R's: Reputation, Reciprocity and Retaliation ......................... 76
    4.4 THE FUNCTIONING OF BITS.............................................................................. 78
       4.4.1       The Costs of BITs..................................................................................... 78
          4.4.1.1 Concluding BITs................................................................................... 79
          4.4.1.2 Breaching BITs: The Three R's Revisited ............................................ 79
               4.4.1.2.1 Reciprocity and Retaliation .......................................................... 79
               4.4.1.2.2 Reputation vis-à-vis Other States ................................................. 80
               4.1.1.2.3 Reputation vis-à-vis Investors ...................................................... 81
XIV                                                                                                      Detailed Contents

        4.4.1.3 Breaching BITs: Non-Reputational Costs ............................................ 83
        4.4.1.4 Summary ............................................................................................... 84
      4.4.2       Commitment and Signalling..................................................................... 84
      4.4.3       Commitment ............................................................................................. 85
        4.4.3.1 Repetition Revisited.............................................................................. 86
        4.4.3.2 The Commitment Game........................................................................ 86
              4.4.3.2.1 The Tribunal's Decision................................................................ 89
              4.4.3.2.2 The Arbitration Decision .............................................................. 89
              4.4.3.2.3 The Expropriation Decision.......................................................... 90
              4.4.3.2.4 The Investment Decision .............................................................. 91
        4.4.3.3 Equilibria............................................................................................... 92
        4.4.3.4 Hostages and Collateral ........................................................................ 93
        4.4.3.5 Extensions............................................................................................. 94
              4.4.3.5.1 Compliance and Enforcement in Third Countries ........................ 94
              4.4.3.5.2 Settlement vs. Trial ....................................................................... 95
              4.4.3.5.3 The Perils of Success .................................................................. 100
        4.4.3.6 Summary ............................................................................................. 101
      4.4.4       Signalling ............................................................................................... 102
        4.4.4.1 Reputation Revisited........................................................................... 102
        4.4.4.2 Signalling Theory ............................................................................... 104
        4.4.4.3 Rights and Treaties as Signals ............................................................ 105
        4.4.4.4 The Signalling Game .......................................................................... 107
        4.4.4.5 Equilibria............................................................................................. 110
        4.4.4.6 Hidden Characteristics and Hidden Intentions ................................... 114
        4.4.4.7 Summary ............................................................................................. 116
      4.4.5       Beyond Commitment and Signalling ..................................................... 117
        4.4.5.1 Flexibility vs. Commitment ................................................................ 117
        4.4.5.2 BITs as Development Aid................................................................... 118
        4.4.5.3 Risk Aversion ..................................................................................... 119
Detailed Contents                                                                                                         XV

    4.5 EMPIRICAL STUDIES AND THE FUNCTIONING OF BITS ................................... 120
    4.6 DISCUSSION AND CONCLUSIONS .................................................................... 121


5       BITS AND INSTITUTIONAL COMPETITION ......................................... 124
    5.1 EFFICIENCY AND BITS .................................................................................... 124
    5.2 INSTITUTIONAL COMPETITION........................................................................ 127
       5.2.1      BITs and the Prisoner’s Dilemma ......................................................... 127
       5.2.2      The Economics of Institutional Competition ......................................... 129
          5.2.2.1 Tax Competition ................................................................................. 131
          5.2.2.2 Environmental Competition................................................................ 133
       5.2.3      Discussion .............................................................................................. 135
    5.3 BITS IN THE CONTEXT OF INSTITUTIONAL COMPETITION .............................. 136
       5.3.1      BITs and Self-Interested Governments .................................................. 137
          5.3.1.1 Political Economy and International (Trade) Law ............................. 137
          5.3.1.2 Political Economy and International Investment Law........................ 139
       5.3.2      BITs and the Prisoner's Dilemma Revisited .......................................... 142
          5.3.2.1 BITs and the Underprovision Hypothesis........................................... 142
          5.3.2.2 Distribution of the Benefits of FDI..................................................... 144
    5.4 PROVISIONS AND EVIDENCE ........................................................................... 145
       5.4.1      Non-Discrimination ............................................................................... 145
       5.4.2      Fair and Equitable Treatment (FET)..................................................... 146
       5.4.3      Indirect Expropriation ........................................................................... 147
       5.4.4      Evidence................................................................................................. 148
          5.4.4.1 Case Law............................................................................................. 148
          5.4.4.2 Analysis............................................................................................... 151
    5.5 DISCUSSION AND CONCLUSIONS .................................................................... 152


6       BITS AND INSTITUTIONAL QUALITY.................................................... 155
    6.1 THE EFFECT OF BITS ON INSTITUTIONAL QUALITY ....................................... 155
XVI                                                                                                       Detailed Contents

       6.1.1      Some Theory........................................................................................... 155
       6.1.2 Empirical Literature ............................................................................... 158
    6.2 MODEL AND DATA ......................................................................................... 160
       6.2.1      Fixed-Effects Model............................................................................... 160
       6.2.2       Variables ................................................................................................ 162
           6.2.2.1 Dependent Variables........................................................................... 162
           6.2.2.2 Independent Variables ........................................................................ 163
           6.2.2.3 Overview............................................................................................. 165
       6.2.3       Summary Statistics................................................................................. 166
       6.2.4       Regulatory Quality and the Rule of Law ............................................... 166
       6.2.5       Corruption.............................................................................................. 172
       6.2.6       Outliers, Economic Freedom and OECD BITs ..................................... 173
    6.3 CAVEATS ........................................................................................................ 173
       6.3.1       Selection and Composition of Variables................................................ 174
       6.3.2       Perception Based Indicators.................................................................. 174
    6.4 DISCUSSION AND CONCLUSIONS .................................................................... 175


7       BITS AND TRANSPARENCY....................................................................... 177
    7.1 LEGAL BACKGROUND .................................................................................... 177
    7.2 THE CASE FOR TRANSPARENCY ..................................................................... 181
       7.2.1      International Law .................................................................................. 181
       7.2.2       International Investment Law ................................................................ 183
    7.3 THE CASE FOR CONFIDENTIALITY .................................................................. 184
       7.3.1      Conventional Arguments for Confidentiality......................................... 185
       7.3.2       Structural Arguments for Confidentiality .............................................. 186
           7.3.2.1 Informational Ambiguities and the Incentive to Comply ................... 186
           7.3.2.2 Efficient Breach and Settlement Problems ......................................... 190
           7.3.2.3 The Incentive to Conclude an Agreement .......................................... 194
       7.3.3       More Transparency?.............................................................................. 196
Detailed Contents                                                                                                        XVII

    7.4 DISCUSSION AND CONCLUSIONS .................................................................... 196


8       SUMMARY AND OUTLOOK....................................................................... 199
    8.1 SUMMARY....................................................................................................... 199
    8.2 OUTLOOK........................................................................................................ 203


APPENDICES............................................................................................................. 205
    APPENDIX A ............................................................................................................. 205
    APPENDIX B ............................................................................................................. 207
    APPENDIX C ............................................................................................................. 211
    APPENDIX D ............................................................................................................. 216
BIBLIOGRAPHY......................................................................................................... 221
List of Figures

Figure 2.1: FDI Flows (Inward) 1970-2009 ................................................................... 7
Figure 2.2: FDI Flows (Inward) – Selected Countries and Regions............................... 8
Figure 3.1: Time Inconsistency .................................................................................... 19
Figure 3.2: Hostage-Mechanism................................................................................... 26
Figure 3.3: Collateral .................................................................................................... 28
Figure 3.4: Hands-Tying............................................................................................... 29
Figure 3.5: Time Inconsistency and the Good Host Country (HCgood)......................... 37
Figure 4.1: The Commitment Game ............................................................................. 88
Figure 4.2: The Signalling Game................................................................................ 109
List of Abbreviations

BIT             Bilateral Investment Treaty
CIEL            The Center for International Environmental Law
CIL             Customary International Law
DCF             Discounted Cash Flows
DSB             Dispute Settlement Body (World Trade Organization)
EBIT            Earnings Before Interest and Taxes
ECT             Energy Charter Treaty
EU              European Union
FCN             Treaty of Friendship, Commerce and Navigation
FDI             Foreign Direct Investment
FET             Fair and Equitable Treatment
FMV             Fair Market Value
FPI             Foreign Portfolio Investment
FTC             Free Trade Commission
GATS            General Agreement on Trade in Services
GATT            General Agreement on Tariffs and Trade
GDP             Gross Domestic Product
GNI             Gross National Income
HC              Host Country
ICC             International Chamber of Commerce
ICJ             International Court of Justice
ICSID           International Centre for the Settlement of Investment Disputes
IISD            International Institute for Sustainable Development
IMF             International Monetary Fund
LDC             Least Developed Countries
M&A             Mergers and Acquisitions
MAI             Multilateral Agreement on Investment
XXII                                                    List of Abbrevations

MIGA       Multilateral Investment Guarantee Agency
MFN        Most Favoured Nation
MNE        Multinational Enterprise
NAFTA      North American Free Trade Agreement
NE         Nash Equilibrium
NGO        Non-Governmental Organisation
OAS        Organization of American States
OECD       Organisation for Economic Cooperation and Development
OPIC       Overseas Private Investment Corporation
PBE        Perfect Bayesian Equilibrium
PCA        Permanent Court of Arbitration
PD         Prisoner's Dilemma
SCC        Stockholm Chamber of Commerce
TRIMS      Trade-Related Investment Measures
TRIPS      Agreement on Trade-Related Aspects of Intellectual Property
           Rights
UN         United Nations
UNCITRAL   United Nations Commission on International Trade Law
UNCTAD     United Nations Conference on Trade and Development
USD        US-Dollar
VCLT       Vienna Convention on the Law of Treaties
WGI        World Governance Indicators
WTO        World Trade Organization
List of Variables

CM   =     Cooperation gain for the MNE (profit)
CH   =     Cooperation gain for the host country (taxes, spillover etc.)
LM   =     Loss for the MNE in case of expropriation (Loser's payoff)
WH   =     Value of the assets to the host country if expropriated (Winner's payoff)
     =     Discount rate of the host country
X    =     Collateral/Hostage
D    =     Damages/Compensation
     =     Probability that the host country is reliable
q2   =     Probability that the host country is reliable as assessed by the MNE in
           period two
z    =     Probability that the bad type will play accommodation in the first period
v    =     Probability that MNE plays invest
y    =     Probability of error
T    =     Litigation costs
R    =     Reputational loss
S    =     Settlement costs
A    =     Costs of concluding the BIT
BR   =     Compliance costs of the BIT for the reliable type (arbitration risk)
BU   =     Compliance costs of the BIT for the unreliable type (arbitration risk)
1               Introduction

Foreign direct investment (FDI) flows amounted to 1,697 billion USD in 2008, while
global FDI stocks reached a level of more than 16,205 billion USD.1 These figures
underline the fact that FDI has gained an importance that is comparable to trade in
providing foreign markets with goods and services.2 In addition, FDI constitutes the
largest source of external finance for developing countries.3 Nevertheless, the global
financial crisis had a significant impact on FDI at the end of 2008, reducing flows by
approximately 14.2% compared to the all-time high of 1,978 billion USD in 2007. The
downward trend continued in 2009. Flows declined to 1,114 billion USD, but are
expected to rise in the next 3 years to about 1,600 - 2,000 billion USD in 2012.4 In any
case, the relative importance of FDI flows in the global economy remains unchanged.
However, FDI is prone to risk. In addition to the operational risk inherent in any
business activity, FDI suffers from political risk. Investors need to be concerned with
the protection of their investment from expropriation, be it direct or indirect, and any
other derogation of their assets by governmental actions. Several institutions on both
the domestic and international level have developed to mitigate this kind of risk. The
most important development in international law in this regard is the emergence of
Bilateral Investment Treaties (BITs). While the number of these treaties has been
steadily growing since the first treaty in 1959, the number of arbitrations based on
BITs and the controversies surrounding the content and impact of BITs has been
notably increasing in the last decade.
Unlike international trade law, the functioning and consequences of international
investment law in general and BITs in particular have not yet been analysed
thoroughly from an economic point of view. Indeed, there is a gap in the literature
regarding the economic analysis of Bilateral Investment Treaties that is worth
addressing. Developing a better understanding of the legal structure surrounding the
protection of FDI is an important enterprise, especially because FDI can substantially
contribute to the economic development of poor countries. The legal protection of FDI
does not only have important implications regarding the integration of the global
economy, but also with regard to issues of domestic public policy as BITs limit the
sovereignty of governments. The approach of law and economics is well-suited to


1
    UNCTAD (2009a), p.xix and p. 251.
2
    Comparing FDI stocks and trade flows for the year 2006, Sauvant (2008) argues that FDI "has
    become the most important vehicle to bring goods and services to foreign markets." See Sauvant
    (2008), p.1.
3
    World Bank (2007), p.314.
4
    See UNCTAD (2010), p. xvii.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_1,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
2                                                                                   Introduction

unite the legal and economic aspects of FDI protection and to advance our
understanding of these important issues.

1.1            International Law and Economics
The economic analysis of law is by now an integral part of many if not most areas of
law. It is virtually undisputed that – in addition to classical fields like antitrust law -
economic theories and methods have also improved the understanding of contract law,
property law and tort law. The application of economic models has also been expanded
to criminal law and family law. Yet, while the field of international law has always
received considerable attention by legal scholars, it has traditionally been ignored by
scholars with a law and economics background.5 Leading law and economics
textbooks do not or only very shortly deal with issues of international law.6 Further,
international legal scholars have mainly abstained from the use of economic methods.7
Nevertheless, for about the last decade, there has been a small but growing literature
on international legal issues. In 1997, Ronald A. Cass denoted international law and
economics scholarship as the "new kid on the block."8 Dunoff and Trachtman (1999)
provide one of the first systematic overviews on the economic analyses of international
law. The authors identify a number of issue areas where the insights of law and
economics can be fruitfully applied. The recent overview article by Sykes (2007)
illustrates that, while the bulk of work still analyses international trade law, a number
of contributions have now been made in various areas of international law, such as
security issues and international investment law. Nevertheless, most fields of
international law remain underdeveloped in this respect.
But what are the defining characteristics of the law and economics of international
law? Three aspects appear especially noteworthy when answering this question. First,
the core principle of economics is the rational choice principle. Consequently, rational
choice theory is a cornerstone of law and economics. Its application to international
law has been the subject of controversial discussion.9 The differences of the rational
choice approach as compared to other common approaches to the analysis of
international law and international relations can be found in Guzman (2002) and
Guzman (2008a). Similar to Guzman (2008a), this book takes a rationalist-
institutionalist approach. In particular, it is not assumed - unlike the "managerial
model" laid out by Chayes and Chayes (1995) - that states have a propensity to comply

5
    A notable exception is international trade law.
6
    See e.g., Cooter and Ulen (2004), Shavell (2004) or, for a short comparison of contracts and
    treaties, Posner (2003), p.136ff.
7
    Dunoff and Trachtman (1999), p.2.
8
    Cass (1997), p.2.
9
    See e.g., Snidal (2002) or Keohane (2002) for the main points of the discussion.
Introduction                                                                                       3

with international obligations.10 The standard rationalist assumptions encompass that
states "are assumed to be rational, self-interested, and able to identify and pursue their
interests. Those interests are a function of state preferences, which are assumed to be
exogenous and fixed. States do not concern themselves with the welfare of other
states, but instead seek to maximise their own gains or payoffs."11 An implicit part of
these assumptions about the state is that it acts as a unitary actor. This assumption of
the unitary state is very useful for the analysis of many aspects of international law,
such as questions of compliance. This approach, although a deviation from the
methodological individualism paradigm of economic theory, has also been applied in
other areas of economics. With regard to corporate behaviour, Posner (2007) argues:
"Economics made much progress in modelling the interactions of business forms
without peeking inside them but instead treating them as if they were individuals."12
Therefore and in line with the relevant literature, the arguments in this book will in
large part be based on this assumption. Nevertheless, at some points, it will be
necessary to deviate from this assumption as some aspects of bilateral investment
treaties may be better understood when assuming - in the public choice tradition - that
the interests of political leaders and citizens do not align.13 Please note this is not a
repudiation of the rational choice principle. Relevant actors are still assumed to be
rational. The public choice approach in the context of international relations simply
opens the "black box" of the state and allows different actors within the state to have
diverging, yet rational interests that influence international lawmaking. However, in
analyzing many aspects of international law, it is simply not necessary to open this
black box.
A second feature of law and economics as compared to other branches of legal theory
is its emphasis on the use of certain, often rather formal methodologies, most
prominently econometrics and game theory.14 Nevertheless, it should be noted that,
while these methods are characteristic to international law and economics, they are not
essential to the rational choice approach. For example, quantitative and statistical
methods can obviously also be employed to test hypotheses derived from theories not
applying the rational choice approach, while a verbal analysis may very well be based
on the rational choice approach.
A third important aspect is the distinction between the positive analysis and the
normative analysis usually employed in economics. While the positive analysis

10
     Chayes and Chayes (1995), p.3.
11
     Guzman (2008a), p.17.
12
     Posner (2007), p.136.
13
     Namely, in chapter 5 of this book.
14
     See e.g. Elkins, Guzman et al. (2006) and Goldsmith and Posner (2005). For a literature overview
     on economics applied to international law, see Sykes (2007).
4                                                                                        Introduction

attempts to explain legal rules and outcomes as they are (given the rationality
assumption), the normative analysis rather asks how the law should be. The normative
criterion employed in law and economics is the efficiency criterion. Economists
differentiate between Pareto efficiency and Kaldor-Hicks efficiency.15 A situation
where it is impossible to make changes that make at least one person better off without
making another person worse off is called Pareto efficient. Put differently, a change
that makes one person better off without making anybody worse off is called a Pareto
improvement. A less restricted notion of efficiency is the Kaldor-Hicks efficiency.
Here, a change is already considered efficient when the winners gain more than the
losers lose. In theory, the winners could compensate the losers for the loss and still be
better off. This compensation, though, does not actually have to take place. Therefore,
Kaldor-Hicks efficient changes are often also referred to as potential Pareto
improvements. While the positive analysis of international law based on the rational
choice principle is now widely used16, the application of the efficiency principle to
international (public) law is rare and not unproblematic. As Van Aaken (2008b) points
out, the problems with the use of the efficiency criterion in international law bears a
resemblance to problems with its use in constitutional law.17 Here, one of the
fundamental problems is that certain constitutional principles may be equally or even
more important than efficiency. Another problem related to efficiency and social
welfare concerns the level of analysis. From a normative perspective, it makes a huge
difference if only the contracting parties are taken into account or if global welfare is
taken as the benchmark. Given these considerations, the book at hand is mainly of a
positive nature with the notable exception of chapter 5. Where normative conclusions
are put forward, the difficulties with the efficiency criterion in international law will be
made explicit.

1.2               Structure
The book will proceed as follows: chapter 2 will provide an overview on the
underlying causes, effects and trends concerning foreign direct investment. By using a
simple game theoretic example, chapter 3 explains why FDI suffers from expropriation
risk and discusses the economic strategies and mechanisms that may help to diminish
this risk. In addition, this chapter outlines the legal landscape regarding the protection
of international investment, with a special focus on the nature and contents of BITs.
Chapters 4 and 5 are both devoted to the understanding of the functioning of Bilateral
Investment Treaties from different perspectives. Chapter 4 focuses on the relationship


15
     See e.g. Cooter and Ulen (2004), p.16f and p.48.
16
     Although the rational choice principle is - as mentioned before - not undisputed.
17
     Van Aaken (2008b), p.659.
Introduction                                                                            5

between the countries trying to attract FDI and the multinational enterprises (MNEs)
providing FDI. At the core of the chapter lies the question of why BITs function
despite the (perceived) weakness of international law. The potential commitment and
signalling properties of BITs will be analysed and discussed. It will be argued that the
scope for a signalling function is rather limited. In addition, the commitment function
is structurally similar to the exchange of hostages or collateral and therefore suffers
from the same incentive problems. In turn, chapter 5 examines the competition
dynamic between developing countries trying to attract FDI. While BITs are certainly
part of the institutional competition for FDI, a controversy exists regarding the
(welfare) implications of this institutional competition. BITs will be discussed in the
context of this controversy. The relevant question is whether BITs are either a
manifestation of the detrimental or of the beneficial aspects of this competition. The
analysis shows that BITs contain elements of both aspects. Furthermore, the findings
of this chapter underline the importance of the consideration of public policy concerns
in investment arbitrations.
The following two chapters focus on specific issues relating to bilateral investment
treaties. More precisely, chapter 6 analyses the impact that BITs may have on
domestic institutional quality using panel data in a fixed-effects model. A positive
effect can be excluded while rather weak evidence points in the direction of a negative
effect. Moreover, the results emphasize the importance of domestic reform for
institutional quality. Chapter 7 is concerned with the issue of transparency in
international investment arbitration. A lack of transparency has been criticised by a
number of observers. Without contradicting the arguments for transparency, this
chapter will discuss some structural implications of increased transparency compared
to confidentiality and aims to add new arguments to the discussion. Special attention
will be paid to the question of who would actually profit from increased transparency,
arguing that the benefits of transparency are widespread, while the benefits of
confidentiality directly accrue to the parties of a dispute. This explains the persistence
of confidentiality in international investment law. Chapter 8 summarises the main
results and identifies areas for further research.
2              Foreign Direct Investment


2.1            Trends and Figures
According to the OECD, foreign direct investment "reflects the objective of obtaining
a lasting interest by a resident entity in one economy (‘‘direct investor’’) in an entity
resident in an economy other than that of the investor (‘‘direct investment
enterprise’’). The lasting interest implies the existence of a long-term relationship
between the direct investor and the enterprise and a significant degree of influence on
the management of the enterprise. Direct investment involves both the initial
transaction between the two entities and all subsequent capital transactions between
them and among affiliated enterprises, both incorporated and unincorporated."18 As a
practical matter, an equity share of more than 10% is usually considered the threshold
for the control of an asset.19 In contrast to FDI, investment that does not aim at the
exercise of control is usually referred to as Foreign Portfolio Investment (FPI). To be
classified as FDI, it does not matter if the direct investment enterprise is incorporated
in the foreign country, and is thus a subsidiary or associate company, or not
incorporated and is thus a branch.20 A frequent distinction with regard to FDI is
between flows and stocks. Obviously, FDI stocks denote the value of FDI in a given
country at a given point of time, while FDI flows denote the amount of FDI flowing to
a given country in given period of time. Another important differentiation regarding
the nature of FDI is between Mergers and Acquisitions (M&As) and Greenfield
investment. While the notion of M&A is self-explanatory, Greenfield investment
refers to investment that includes the establishment of new production facilities such
as offices, buildings and factories.21
The last few years have seen a considerable growth of FDI. In 2006, FDI flows
reached $1,306 billion USD while total FDI stocks amounted to roughly $12,000
billion USD.22 Compared to 2005, this constitutes a growth in FDI flows of about
38%. As Figure 2.1 indicates, FDI has been on a growth path since the mid-1980s,
reaching a peak in the year 2000 and an all-time high in 2007. However, recent figures
for 2008 and 2009 evidence the negative consequences of the financial crisis and have
caused FDI to fall to 1,114 billion USD in 2009.23


18
     OECD (1996), 7f.
19
     UNCTAD (2007a), p.245.
20
     OECD (1996), p.9.
21
     UNCTAD (2006), p.15.
22
     UNCTAD (2007a), p.9.
23
     UNCTAD (2010), p.xix.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_2,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
Foreign Direct Investment                                                         7




         Figure 2.1: FDI Flows (Inward) 1970-2009

Source: Own diagram based on UNCTAD Data (http://stats.unctad.org/FDI/)


Foreign Direct Investment now constitutes the largest source of external finance for
developing countries.24 Nevertheless, the bulk of FDI in 2006, namely more than 60%,
went to developed countries.25 As figure 2.2 indicates, the countries of Western
Europe experienced the highest FDI inflows. Among the emerging and developing
regions, Asia appears to be the most attractive FDI location.




24
     World Bank (2007), p.314.
25
     Kekic and Sauvant (2007), p.6.
8                                                                     Foreign Direct Investment




         Figure 2.2: FDI Flows (Inward) – Selected Countries and Regions

         Source: Own diagram based on UNCTAD Data (http://stats.unctad.org/FDI/)


It is remarkable that, as Sauvant (2008) notes, FDI "has become the most important
vehicle to bring goods and services to foreign markets and to integrate national
production systems".26 More specifically, in 2006 there were more than 78,000 parent
companies with more than 780,000 foreign affiliates.27 Another noteworthy trend is
the growing importance of south-south flows (meaning flows between developing
countries) since the mid-1980s. Starting from about 4 billion USD in 1985, south-
south flows rose to almost 60 billion USD in 2004.28 Regarding the sectoral
distribution of FDI, a relative shift from manufacturing to services can be observed. Of
the global FDI stocks in 2005, 61% were in the service sector, compared to 49% in
1990.29 Manufacturing accounted for 30%, leaving the remaining 9% to the primary
sector.

2.2             Multinational Enterprises and FDI
Why do enterprises find it worthwhile to invest in foreign countries at all? Goods and,
to a growing degree, services could be produced in the home country of the investor
and simply be exported to the foreign markets. A number of theories of foreign direct

26
     Sauvant (2008), p.3.
27
     UNCTAD (2007a), p.12.
28
     UNCTAD (2006), p.117.
29
     UNCTAD (2007a), p.22.
Foreign Direct Investment                                                                         9

investment have been developed. This section will focus on transaction
cost/internalisation theories of FDI as developed by scholars like Buckley, Casson,
Hennart and Rugman.30 The following summary is based on Hennart (2001). The
transaction cost theory of FDI has its foundations in the writings of Coase (1937),
Alchian and Demsetz (1972) and Williamson (1985).31 As Hennart (2001) points out,
classical trade theorists explained FDI mainly as a result of differences in real interest
rates. However, this theory proved to be unable to explain the major flow and growth
patterns of FDI. Theories based on transaction cost economics revealed more
explanatory power in the context of FDI. In general, transaction cost economics
focuses on the costs of cooperation between economic agents as a result of cognitive
limitations and opportunism, namely information, enforcement, and bargaining costs.32
The specific question transaction cost theories of FDI seek to answer is why
interdependencies across national boundaries are organised through multinational
enterprises (and thus through FDI) that could also be organised through markets.
Interdependencies, in this context, arise when a "firm located in country X has some
assets which have potential value in country Y if successfully combined with some
country Y factors."33 The obvious answer is that multinational enterprises (MNEs) will
arise when they can organise these interdependencies more efficiently than markets.
As Hennart (2001) notes, this implies three conditions for the existence of MNEs: "(1)
interdependent agents must be located in different countries (otherwise we would have
a domestic firm), (2) the MNE must be the most efficient way to organize these
interdependencies (otherwise we would have international market transactions), and
(3) given condition (2) the costs incurred by MNEs to organize these
interdependencies are lower than the benefits of doing so."34
These conditions are likely to be met where know-how, reputation, raw materials and
components, distributions and marketing and, finally, financial capital is involved.
Why should MNEs in these areas work more efficiently than markets? It is, for
example, well-known, that the market for know-how suffers from information
asymmetries. The dynamic is similar to the market for lemons as described by Akerlof
(1970). Potential buyers of know-how cannot assess the quality of the product.
Obviously, the seller cannot reveal the know-how (if he did, he would give away his



30
     Hennart (2001), p.131. A competing theory of FDI that nevertheless shares many features with
     transaction cost theories is the eclectic paradigm. See Dunning and Lundan (2008).
31
     Dunning and Lundan (2008), chapter 4. Hennart (2001), on the other hand, emphasises that the
     transaction cost theories of FDI was developed independently of the writings of Williamson. See
     Hennart (2001), p.132.
32
     Hennart (2001), p.133.
33
     Hennart (2001), p.145.
34
     Hennart (2001), p.136.
10                                                                 Foreign Direct Investment

product for free). If patent systems cannot guarantee the property rights of the know-
how, the optimal way to organise the interdependency might be through an MNE.
In the case of reputation, an alternative way for a company from one country to exploit
its reputation in another country would be through franchise contracts. However, as
Hennart (2001) emphasises, free-riding poses a huge problem to franchising. A single
franchisee has the incentive to lower his quality. The reputational costs will be borne
by the whole franchise chain and only to a very small degree by the single franchisee –
especially when customers are mainly non-repeat customers. When it is very costly or
impossible to contractually define quality, ownership (via an MNE) may be a more
efficient way to exploit reputation across borders than franchising contracts.
The motivation for vertical integration across borders in the case of raw materials and
components is different. Here, the problem with market transactions is often the
absence of a large number of buyers and sellers. According to Hennart (2001), a small
number of market participants in raw materials markets is often the result of
economies of scale, high transportation costs, government barriers, and asset
specificity.35 Asset specificity means that the value of an asset is (mainly) specific to a
certain transaction and thus loses value if used otherwise. As described prominently by
Williamson (1985), the party undertaking investment in specific assets may be subject
to the hold-up problem, which is opportunistic renegotiation by the other party to the
point where the investment would not have been profitable in the first place.
Consequently, a value-creating cooperation might not be pursued when the hold-up
risk is high. Long-term contracts are a potential solution for this dilemma. However,
when long-term contracts are not feasible and transaction-specific investments are
considerable, ownership might be a sensible strategy to overcome the hold-up
problem.
Comparable to the example of raw materials, markets for distribution can be
inefficient and subject to hold-up problems as well. Distributors may be reluctant to
invest in manufacturer-specific distribution systems. As Hennart (2001) states: "When
the physical or intellectual investments necessary to effectively sell a product are large
and manufacturer-specific, and the environment is hard to predict, the integration of
manufacturing and distribution within an MNE will often be the best solution."36
The last example illustrating the purpose of MNEs concerns financial capital. As
mentioned before, classical trade theorists considered FDI as a result of differences in
the real interest rate. The preceding paragraphs demonstrated that a number of reasons
for the existence of MNEs do not require the existence of interest rate differences.


35
     Hennart (2001), p.139.
36
     Hennart (2001), p.140f.
Foreign Direct Investment                                                              11

Nevertheless, financial capital from country A might often be invested more profitably
in country B. The question here is why this should not always simply be done through
lending. The problem is that lending money creates problems of moral hazard on the
side of the borrower who may take too much risk or not exert enough effort. The
cross-national nature of the transaction may make the lender’s control of the funds
even more problematic. One solution to this problem could be the use of collaterals.
However, when collaterals are not available and the project is nevertheless profitable,
a sensible strategy could be the integration of lenders and borrowers into an MNE.
As these examples illustrate, the key question for the existence of MNEs and thus FDI
is why the organisational form of an MNE for the cross-national transaction is more
efficient as compared to the use of the market or contracts. The prerequisite is, of
course, that the cross-national transaction is profitable. This prerequisite is evidently
not sufficient for the occurrence of FDI. Indeed, in many cases, the cross-national
transaction will simply be executed through markets or certain types of contracts.
However, if the existence of hold-up risks or information asymmetries means that the
transaction is more efficiently structured along the hierarchical lines of a multinational
enterprise, FDI will occur. Therefore, the motives for FDI naturally bear a lot of
resemblance to the classical reasons for vertical integration as identified by the
transaction cost economics literature, adding the cross-national nature of the economic
transaction at hand.
A different theory of FDI that is often considered the predominant paradigm for direct
investment is called the OLI or eclectic paradigm.37 The OLI theory is for the most
part not incompatible with the transaction cost/internalisation theory of FDI. As a
matter of fact, internalisation is an integral part of the OLI framework. The main
difference is that the OLI framework emphasises, apart from internalisation (I), the
importance of ownership-specific advantages (O) and location-specific advantages (L).
Ownership-specific advantages are the advantages the potential MNE has that it
wishes to exploit across borders, while location-specific advantages are the advantages
of a potential host country of FDI compared to its competitors.38
As a practical matter, FDI is often classified into three categories: market seeking FDI,
resource/asset-seeking FDI and efficiency-seeking FDI. This classification is used by
the United Nations Conference on Trade and Development (UNCTAD).39 As the
transaction cost theory of FDI has illustrated, the existence of attractive markets or the
demand for natural resources is not sufficient for FDI to emerge. In addition,



37
     Ethier and Markusen (1996), p.2.
38
     See Dunning and Lundan (2008) for the description of the OLI paradigm.
39
     See e.g. UNCTAD (2003), p.85.
12                                                                Foreign Direct Investment

ownership must be a more efficient organisational structure than, for example, market
transactions. Nevertheless, the classification used by UNCTAD (2003) can be useful
in some respects.

2.3              FDI and Development
Before proceeding, it appears worthwhile to take a closer look at the impact of FDI on
the host country's economy. The term host country (HC) is used to denote the country
that receives the investment (as opposed to the home country of the investor). With
special regard to developing countries, Jain and Vachani (2006) provide an overview
on the potential effects of multinational enterprises (MNEs).40 These effects include
technology spillover, export spillover and productivity spillover, as well as effects on
wages, competition and consumer preferences. The empirical evidence on some of
these effects appears to be mixed. One reason is that spillover is hard to measure. As
Saggi (2002) notes, spillovers "do not leave a paper trail."41 The question of the effect
of FDI on the host country can be approached on two different levels. First, one could
look at the effect of FDI on micro-level data like wages, productivity, exports etc.
Alternatively, one could look at the growth rate of countries and directly analyse its
connection to direct investment.
Lipsey (2002) takes the first approach and summarises the main empirical findings of
the impact of FDI on wages, productivity, exports, and the introduction of new
industries. He observes that the impact of FDI on local wages is positive. Virtually all
studies on this matter show that foreign-owned companies pay higher wages than local
firms.42 The question is whether these findings reflect that MNEs might mainly
operate in high-wage sectors and/or that MNEs might have a higher propensity to hire
more highly qualified personnel than domestic firms for reasons unrelated to their
foreignness. The numerous studies on this question are very diverse in the sense that
they focus on many different countries and regions (including developing as well as
developed countries) and employ a diverse set of control variables (firm size,
education level, industry sector, etc.). As the overview in Lipsey (2002) illustrates,
foreign firms clearly tend to pay higher wages, independent of the region in which
they operate (and independent of a wide set of other controlled variables). Another
important question concerns the wage effect of MNEs on the labour market (that is, on
the wages paid by local firms) in general. Here, the evidence is more ambiguous.
Some studies find positive, while others find no or negative spillover of MNEs on



40
     Jain and Vachani (2006), p.8ff.
41
     Saggi (2002), p.208.
42
     Lipsey (2002), p.20.
Foreign Direct Investment                                                             13

wages in domestic owned firms.43 In sum, the effect of the presence of MNEs on the
average wage appears to be mainly positive, which indicates that even if there are
negative spillovers, these are offset by the higher wages paid by the MNEs
themselves.44
Concerning productivity spillover, a logical prerequisite is that foreign-owned firms
are on average more productive than domestic firms. Virtually all empirical studies on
this matter confirm that foreign firms indeed are more productive, in developing as
well as in developed countries.45 The spillover of productivity (or knowledge and
technology) to domestic firms is obviously much harder to measure. The main
transmission channel for the spillover of productivity to domestic firms is likely
backward linkages. These backward linkages describe the connection of MNEs to
suppliers. MNEs usually have an interest in establishing or sharing production
technology and quality-control techniques in a vertical direction in order to build
reliable production networks.46 On the other hand, MNEs have an interest to avoid
knowledge spillover in a horizontal direction. These spillovers may nevertheless take
place due to imitation, competitive pressures or the movement of employees.47 The
empirical evidence on the effect of FDI on productivity is mixed. Some studies even
seem to indicate negative spillover effects.48 As Lipsey (2002) points out, the
measuring of productivity spillover poses a number of econometric problems,
including the validity and availability of productivity indicators as well as problems
with model specifications.49 As a logical matter, it is hard to hypothesise that MNE
activity should not have any (positive) effect on domestic firms at all. As Moran
(2006) provocatively phrases, it "is possible to imagine in the abstract that foreign
investors enter a host economy and train local managers and workers who never leave
the foreign-owned firms, set up operations without any local firms copying their use of
machinery or their management techniques, and create supply chains with indigenous
companies that learn nothing new from the relationship, enjoy no scale effects, or, if
they do, use the novel skills to sell exclusively to the foreign subsidiaries who capture
all the benefits that result."50 Obviously, this scenario appears to be rather unlikely.
Nevertheless, given the mixed empirical findings on this issue, it is hard to make any
firm assertions. Lipsey (2002) concludes that "the mixed story for spillover, combined



43
     Lipsey (2002), p.32.
44
     Lipsey (2002), p.34.
45
     Lipsey (2002), p.40.
46
     Moran (2006), p.21.
47
     Moran (2006), p.24.
48
     Lipsey (2002), p.46.
49
     Lipsey (2002), p.49ff.
50
     Moran (2006), p.23.
14                                                               Foreign Direct Investment

with the strong evidence for superior productivity of foreign-owned firms, suggests
that overall productivity is improved by the presence of foreign-owned operations,
although that question is rarely, if ever, examined."51
The evidence of the effect of FDI on exports and the introduction of new industries is,
however, well-established and positive.52 The affiliates of MNEs are usually not only
more export-oriented than their domestic counterparts; there is also evidence that the
presence of an MNE also raises the export propensity of the domestic firms in that
industry.53 Consequently, the host country establishes a tighter linkage in the world
trading system through FDI. This assertion also partly holds true for imports. The
question here is whether trade and FDI are substitutes or complements. At first glance,
the answer given by the internalisation theory appears simple: FDI is described as an
alternative to trade and consequently, FDI and trade should be viewed as substitutes.
However, this simplistic view ignores a number of more complicated processes related
to MNE activity. Most importantly, the MNE may import intermediate goods
necessary for the production of the final good. Also, the presence of a firm's
production facilities for one product in a country may increase the demand of other
products of the same company in that country, even when the other products are
produced somewhere else. Potential channels might be, among others, increases in
demand due to the publicity related to the presence of the firm54 or more efficient
deliveries and distribution.55 While there are theoretical arguments both for
substitutability and complementarity, the majority of empirical studies finds strong
evidence for net complementarity.56 However, depending on the data aggregation level
(product specific data as compared to firm or industry specific data), substitution
effects can also be shown.57 In sum, while FDI substitutes for imports to some extent,
on the aggregate level, FDI leads to even more imports.
Examining the possible spillover effects of FDI on certain economic variables leads to
the more general question regarding the impact of FDI on economic growth. The
importance of the dissemination of ideas for economic development, also through FDI,
has been noted by Romer (1993)58, one of the co-founders of the endogenous growth
theory. However, the empirical studies on this issue show mixed results. While some
studies find no systematic effect of FDI on growth, other studies find that, for certain


51
     Lipsey (2002), p.59.
52
     Lipsey (2002), p.54.
53
     Lipsey (2002), p.53.
54
     Lipsey and Weiss (1984), p.305.
55
     Blonigen (2001), p.84.
56
     Blonigen (2001).
57
     See Swenson (2004) and Blonigen (2001).
58
     Romer (1993).
Foreign Direct Investment                                                              15

subsets of countries, there is a robust positive relationship.59 Two recent studies should
be mentioned here as illustrative examples. Li and Liu (2005) investigate the
relationship between FDI and growth in a panel data study for 84 countries for the
period between 1970 and 1999. From the mid-1980s onwards, the authors identify a
significant relationship between FDI and economic growth. On the other hand,
Carkovic and Levine (2005), using a panel data study for 72 countries from 1960 to
1995, do not find a robust, positive influence of FDI and economic growth.
A number of authors have tried to detect the reasons for the inconclusive results of the
impact of FDI on growth. Borensztein, De Gregorio et al. (1998) show that FDI can
have an important positive impact on growth, but only when the host country already
has a certain stock of human capital. In other words, a lack of human capital limits the
absorptive capability of the host country. Busse and Groizard (2008) explore the
importance of regulatory quality for the impact of FDI on growth. The authors find
that excessive business and labour regulations frustrate a positive growth effect of
FDI. Alfaro, Areendam et al. (2003) stress the importance of the stage of development
of the local financial markets for a positive impact of FDI. Balasubramanyam, Salisu
et al. (1996) emphasise the importance of trade policy, or, more specifically, trade
openness, for FDI to exert a positive effect on growth. A similar point is made by
Moran (2006), who uses a number of examples to illustrate that FDI has detrimental
effects where it is understood as part of an import substitution policy, while positive
effects accrue in the absence of trade protection.
This discussion illustrates that it would be inaccurate to simply imply that more FDI
will lead to higher economic growth rates. However, FDI can indeed work as a vehicle
for higher growth when certain host country characteristics and policies are met.
Nevertheless, as a number of authors emphasise, FDI and MNE activity can also have
a negative effect on certain host country variables. Some of these effects have already
been mentioned, like the possible negative effect on host countries’ productivity. Also,
as Moran (2006) has pointed out, where MNEs operate in protected markets as part of
an import substitution strategy, they may use inefficient technology and may also not
be able to exploit economies of scale, thus doing more harm than good to the host
country's economy.60 Other authors have raised concerns of the impact of MNE
activity on environmental and social standards61, on the competition in domestic
markets and the loss of market share of domestic firms62, the crowding out of domestic



59
     Lipsey (2002), p.55.
60
     Moran (2006), p.7ff.
61
     For a short overview, see Meyer (2004), p.269ff.
62
     See, e.g., Aitken and Harrison (1999).
16                                                               Foreign Direct Investment

investment63 and the possible negative balance of payments effects.64 Evidently, the
economic literature on these issues can be characterised as controversial. Nevertheless,
the majority of economists seem to emphasise the positive effects of FDI on the
respective host country. It also seems clear that some of the potential negative effects
of FDI can be addressed through adequate domestic policies and thus might, if at all,
only pose minor or temporary problems.
While the academic debate on the effects of FDI on host countries is far from being
over, as a practical matter, the issue seems to have been decided. Governments all over
the world seemingly consider the contribution of FDI to domestic welfare as positive
and adapt their regulatory systems to attract FDI. From the 184 regulatory changes in
2006 with regard to FDI as monitored by UNCTAD, 147, representing a share of
almost 80%, were favourable to FDI.65 For the time span of 1992 to 2001, Kobrin
(2005) finds that even 95% of the changes were liberalizing rather than restrictive.
Using cross-sectional regressions, Kobrin (2005) shows that external factors do not
play a significant role in the liberalisation decision – indicating that favourable FDI
policies were not imposed through external pressures. In addition, international
development organisations such as UNCTAD emphasise the importance of FDI for
development.66 Not surprisingly, developed countries represented by the OECD also
see MNE activity as mainly beneficial.67 Lastly, the predominant perception of the
beneficial nature of FDI with policy-makers is also reflected by the lasting trend to
protect international investment by means of international law, namely through
bilateral investment treaties (BITs).




63
     See, e.g., De Backer and Sleuwaegen (2003).
64
     UNCTAD (1997), p.88.
65
     UNCTAD (2007a), p.14.
66
     See, for example, the foreword by Ban Ki-moon to the World Investment Report 2007,
     UNCTAD (2007c), p.iii.
67
     See e.g. OECD (2002).
3               Economic and Legal Protection of FDI

The preceding section has shown that most countries perceive FDI as predominantly
beneficial. Obviously, the companies that engage in FDI prefer this form of transaction
to other forms like trade or licensing. If FDI is beneficial to both parties of the
transaction, why would FDI need any protection at all? The following section will
explain the inherent risk to FDI as the consequence of a time inconsistency problem
(section 3.1). MNEs, international organisations and policymakers have implemented
economic and legal measures to protect FDI. The former will be introduced in section
3.2. and the latter in section 3.3.

3.1             Time Inconsistency and Expropriation Risk
Investors operating in a risky international environment are naturally concerned with
the protection of their assets. Apart from the regular economic risks associated with
the operation of a business, a major concern for MNEs is political risk.68 A useful
definition of political risk goes back to Weston and Sorge (1972), taken up by Kobrin
(1979): political risks "arise from the actions of national governments which interfere
with or prevent business transactions, or change the terms of agreements, or cause the
confiscation of wholly or partially foreign owned business property."69 The main
problem here is the problem of the taking of property owned by foreigners through
expropriation or through regulatory measures (indirect expropriation). At the core of
this issue is usually the redistribution of assets from foreign businesses to the
government or third parties.

3.1.1           Time Inconsistency
The economic problem related to expropriation, be it direct or indirect, is a problem of
time inconsistency70 or hold-up.71 Once the investment is sunk, the host state has an
incentive to expropriate (or exploit, redistribute etc.) the assets in one way or another.
This could be considered a form of opportunism that is present in many if not most
sequential transactions. A useful definition of time inconsistency can be found in




68
     See, e.g., Busse and Hefeker (2007).
69
     Kobrin (1979), p.67.
70
     See Guzman (1998). The idea of time inconsistency in government action goes back to Kydland
     and Prescott (1977). Guzman (1998) uses the expression dynamic inconsistency instead of time
     inconsistency.
71
     See, e.g., Williamson (1983).


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_3,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
18                                                              Economic and Legal Protection of FDI

Drazen (2002).72 In his example, a government has to choose a tax rate for a certain
time period in the future. It can either decide on this tax rate immediately or it can wait
until the relevant period arrives. "Time inconsistency is said to arise if, though nothing
has changed (at least ostensibly), these choices are not equal (…)".73 From a game-
theoretic point of view, the interaction with respect to international investment
resembles a trust game.74 The problem of time inconsistency can in this way be
presented as a game-theoretic problem of subgame-(im)perfection.75 The underlying
logic can be illustrated in a simple sequential two-player game of complete
information involving a representative multi-national enterprise (MNE) and a host
country (HC). First, the MNE decides whether to invest or not. If no investment takes
place, the payoff for both players is zero. If the investment takes place, the host
country has two options: it can either accommodate or expropriate the investment.76 If
the host country accommodates, both player receive their share of the cooperation
gain: the MNE receives CM and the host state gains CH, where CH is a measure
indicating how much a country can gain through FDI, be it through taxes or spillover
effects, with CH ! 0 . Further, it is assumed that WH (with WH t 0 ) describes how much
the investment is worth to the host state if expropriated.77 In that case, the remaining
value of the investment to the investor (which is often negative and therefore rather a
loss or damage) is defined as LM. Consequently, the extensive form of the game looks
as follows:




72
     See Drazen (2002), p.101f. The focus of Drazen (2002) is, however, on time inconsistency
     problems where governments use that phenomenon to increase social welfare by misleading their
     constituents. The focus of this book is, on the other hand, how time inconsistency can be avoided
     when it is anticipated by the economic agents.
73
     Drazen (2002), p.102.
74
     For an illustration of the trust game, see, e.g., Kirstein (1999), p.20 ff or Schmidtchen (1994).
75
     Time Consistency and Subgame Perfection are not identical concepts. First of all, the notion of
     time consistency is usually related to macro-economic issues, while subgame perfectness (or
     perfection) is a game theoretic expression. More importantly, subgame perfectness is a useful
     solution concept in numerous dynamic games that have no reference to time consistency.
     Nevertheless, for the exposition of time consistency in a game theoretic context, subgame-
     perfectness seems to be the most practical way. The similarity between time consistency and
     subgame perfectness is also illuminated in McTaggart and Salant (1989) and explained in
     Rasmusen (2001), p.105. For a more game-theoretic reflection on Kydland and Prescott (1977), see
     Fudenberg and Tirole (1991), p.74ff.
76
     Please note that by expropriate, the judicial term implying the payment of compensation is not
     meant, but simply the appropriation of the assets of the investor/MNE. Yet, the use of the word
     appropriate or, alternatively, confiscate, could be equally misleading.
77
     Or how much the host country can win (thus WH) by taking the opportunistic action. The notation
     used here is in line with Dixit (2004), p.15.
Economic and Legal Protection of FDI                                                  19

                                       (0,0)
                       Don't Invest


                       M                              (CM,CH)
                                       Accommodate
                           Invest

                                        H


                                        Expropriate
                                                      (LM,WH)


Figure 3.1: Time Inconsistency



Notation
CM       =    Cooperation gain for the MNE (profit)
CH       =    Cooperation gain for the host country (taxes, spillover, etc.)
LM       =    Loss for the MNE in case of expropriation (Loser's payoff)
WH       =    Value of the assets to the host country if expropriated (Winner's payoff)
Player
There are two players: the host country (HC) and a representative investor or multi-
national enterprise (MNE)
Strategies
The game has two stages. In the first stage, the MNE decides whether to invest. When
there is no investment, the game is over. In the case of investment, there is a second
stage. The host country now must decide whether to accommodate or expropriate the
investment.
Payoffs
In the case where no investment takes place, the payoff to both players is normalised
to zero (second best opportunity). When the MNE invests and the host country
accommodates, the MNE generates a profit of CM and the HC gains spillover and taxes
of CH. In the case where the MNE invests and the HC expropriates, the MNE will
incur damages LM and the HC an expropriation gain WH. In the baseline (no
cooperation) case, we assume CM >0, LM <0 and WH > CH.
20                                                             Economic and Legal Protection of FDI

This kind of game is also known as the trust game, which is a one-sided version of a
prisoner’s dilemma.78 Assuming rationality, the solution of this game can be derived
straightforwardly using backwards induction. This solution concept requires starting
the analysis at the final decision node or final stage of the game. At the second (and
final) stage of the game, the host country must decide whether to accommodate or to
expropriate. If WH < CH, that is, if the investment yields more payoff to the host state
when it is in the hands of the MNE, the outcome is the cooperative outcome
(Invest/Accommodate). However, in many cases, the assumption WH > CH appears to
be more realistic. If that is the case, the host state will expropriate in the second stage.
As rationality and complete information are assumed, the MNE will anticipate this
and, if LM <0, decide not to invest. The subgame-perfect Nash-Equilibrium is thus
(Don't Invest/Expropriate) with payoffs of zero for both players. Clearly, the
equilibrium is Pareto-suboptimal. Both players could achieve a higher payoff through
the combination (Invest/Accommodate). However, this does not constitute a subgame-
perfect equilibrium.
Obviously, the subgame-perfect equilibrium depends on the respective payoffs. The
following summary illustrates this (with LM < CM, the gray-shaded area is the case of
time inconsistency):



Payoffs                                             Equilibrium

CH > WH and CM >0                                   Invest, Accommodate

CH > WH and CM <0                                   Don't Invest, Accommodate

CH < WH and LM <0                                   Don't Invest, Expropriate

CH < WH and LM >0                                   Invest, Expropriate



The definition of time inconsistency included the notion that choices differ even
though "nothing has changed". Figure 3.1 shows that this statement is not entirely
true.79 Between stage 1 and stage 2 of the game, the investor has made his investment
decision, albeit in a predictable way. The statement that "nothing has changed" must



78
     See Kreps (1990) for a discussion of the trust game. An application to legal aspects can also be
     found in Kirstein (1999).
79
     See also Drazen (2002), p.108.
Economic and Legal Protection of FDI                                                  21

consequently be understood in a way that there are no changes that could not have
been predicted in the first place.
Two clarifications with regard to the game structure as presented above must be made.
The first concerns the notion of expropriation. In this model, the term expropriation is
not used as a legal term. Here, any government measure that deprives the investor of
the benefit of his investment is subsumed under expropriation. This includes not only
outright seizure of the asset, but also partial expropriation, revenue expropriation or
other governmental provisions and measures that reduce the value of the investment,
but do not fit into the legal definition of expropriation. In the case of revenue
expropriation, governments do not expropriate the asset as such, but leave the asset in
the hands of the investor and expropriate only the surplus or revenue of the investor.
Consequently, expropriation as defined for the game described above includes
situations where states set operating conditions in a way to leave just enough revenue
to cover the operating costs, thereby providing sufficient ex-post incentive to operate,
but certainly not enough to invest in the first place. Therefore, the simple model above
captures, depending on the values and interpretation of the variables, all forms of
expropriation and asset deprivation. As long as LM <0 (the value of the investment to
the investor is negative after state intervention) and CH < WH (the value of the asset to
the state is higher if it is actually possessed by the state), we are dealing with the
underlying economic problem of time inconsistency.
A second preliminary note should be made regarding social welfare and the
differentiation between nationals and non-nationals. The model as specified above
does not make any statements about the objective function of the government. As a
matter of fact, time inconsistency can arise even when the government is considered to
be a benevolent social planner.80 In other words, if the government maximises social
welfare and redistributes the expropriated assets to its citizens (while ignoring the
welfare of the investor), the time inconsistency problem also applies to welfare-
maximizing governments. Drazen (2002) emphasises that, contrary to the standard
view, a conflict of interests is a necessary condition for time inconsistency to arise.81
In our model, the conflict of interests between the government /state and the investor,
once the investment is sunk, is obvious (as long as CH < WH). But this conflict of
interests also exists when dealing with a benevolent (with regards to its own citizens)
government. Nevertheless, the legal characteristics of the Bilateral Investment Treaties
show that many clauses foremost address governments that maximise their own
welfare and that this behaviour aggravates the time inconsistency problem. This point



80
     Drazen (2002), p.101. See also Fischer (1980).
81
     Drazen (2002), p.110 ff.
22                                                        Economic and Legal Protection of FDI

will be discussed in chapter 5. However, the disposition of the appropriated assets is
not relevant for this chapter. Here, the focus is on the time inconsistency problem as
such, how governments and investors structure their transactions to alleviate this
problem and how international law, and, more specifically BITs, can contribute to
reducing this problem.

3.1.2           On the Relevance of Expropriation Risk
Before proceeding, it appears be conducive to highlight that expropriation or the risk
of expropriation (and other forms of appropriation of international investment) is not
only a theoretical problem. Three points illustrate this assertion. First, a historical
examination of the issue reveals that expropriation has occurred frequently in the 20th
century, especially since 1917. The nationalisation in the context of the Russian and
the Mexican Revolutions in that year placed the issue of the protection of foreign
property on the international law agenda.82 Following World War II, almost all
countries in Eastern Europe and a number of countries in Latin America as well as
China and later Cuba expropriated foreign (as well as national) investments due to a
shift towards communism.83 Another wave of expropriations occurred in the 1970s,
when most Arab countries nationalised their oil industries.84 Recent events in South
America, particularly in Bolivia and Venezuela, highlight the importance of the issue
even today.85 Also, in highly industrialised countries like Germany, expropriation had
been part of the political discussion in the wake of the financial crisis.86 Second, there
is empirical evidence that political risk (which subsumes the risk of expropriation) can
influence FDI decisions.87 Third, the existing legal remedies in international law are
widely used. The rise of Bilateral Investment Treaties and the increasing number of
arbitrations, as illustrated in section 3.3, show that there is a certain "demand" for
these instruments – which would not be the case if investors and governments
considered the risk of a loss of FDI through government measures to be negligible.

3.2             The Economics of FDI Protection
If the logic of the one-sided prisoner's dilemma holds true, why does foreign direct
investment nevertheless flow to countries that do not offer any legal protection? Put
more generally, why would a party fulfil its contractual obligations in the absence of


82
     Lowenfeld (2008), p.470ff and Dolzer (1985).
83
     Lowenfeld (2008), p.483f.
84
     Lowenfeld (2008), p.484.
85
     See for example, The Economist, August 7th, 2008, "The Autocrat of Caracas".
86
     See for example, Frankfurter Allgemeine Zeitung, January 31st, 2009, "Steinbrück erwägt
     Verstaatlichung der HRE".
87
     Busse and Hefeker (2007).
Economic and Legal Protection of FDI                                                  23

coercive power? There are a number of devices that can foster cooperation. For one, if
we add the factor of time to the game presented in section 3.1.1, cooperation might
emerge automatically without the need for the parties to take specific actions that
make cooperation more likely. Due to the time dimension, we refer to these devices as
dynamic devices. Before turning to these dynamic devices, the static prisoner's
dilemma situation as presented in section 3.1.1 will be discussed first.

3.2.1            Static Devices
Obviously, expropriation will not occur when CH > WH, that is when the state can
receive a higher payoff through taxes, spillover, etc. compared to expropriation. This is
typically the case when the value of the asset to the state is low. The investor might
bring skills, like technological knowledge, organisational capabilities, or access to
markets that would be lost in the case of expropriation and thus make expropriation
pointless.88 In other words, the government cannot manage the asset as efficiently as
the investor and is thus better off leaving the investment in the hands of the investor
(thereby profiting from taxes and spillover). This is especially true when the taxes and
spillover effects are large. However, in these cases, governments do not face a time
inconsistency problem in the first place. Put differently, there is no prisoner's dilemma
(“PD”) situation that is hostile to a cooperative outcome. If, however, governments
and investors do find themselves in a PD situation, the parties might be able to take
actions that transform the trust game into a game with an expected cooperative
outcome. As mentioned before, the central problem is the lack of coercive power in
international relations and law. Otherwise, the parties could simply contract on playing
(Invest/Accommodate).
In a seminal paper, Kronman (1985) identifies four mechanisms that enhance
transactional security and the exchange of promises in a state of nature, a state that in
some respects also characterises international relations. With reference to the classic
writings of Thomas Hobbes, yet with some modifications, Kronman (1985) defines the
state of nature as a situation where "those who make covenants are not subject to any
'common power' with the 'right and force sufficient to compel performance'."89 The
author emphasises that the state of nature is relational, meaning it describes a
relationship between two parties. This view of the state of nature is certainly
comparable to the relationship of a foreign investor vis-à-vis the host state. The four
mechanisms Kronman (1985) identifies are hostages, collaterals, hands-tying and
union. Acknowledging that the lines between these measures may be blurred,



88
     Eaton and Gersovitz (1984).
89
     Kronman (1985), p.7. Quotation marks in original.
24                                                                     Economic and Legal Protection of FDI

Kronman (1985) treats these four devices distinctively.90 This approach will be
followed here. In addition to the mechanisms identified by Kronman (1985), this paper
adds three additional ones: insurance as a mechanism to enhance transactional
security, the devaluation of assets as a strategy to avoid expropriation and lobbying. At
the core of each of these mechanisms is the modification of the payoffs of the original
PD game, namely CH, CM, WH and LM, to achieve a cooperative and thus a mutually
more beneficial outcome.

3.2.1.1           Hostages
Parties to a transaction that is not simultaneous might exchange hostages.91 More
specifically, the party that performs first (party A) might ask the party that performs
second (party B) for a hostage, that is, anything that is of value for party B. The
hostage is ideally of no or little value to party A or anybody else, but at any rate of less
value to party A than to party B.92 This feature makes the hostage attractive to party B:
it can be sure that the hostage will be returned after performance. The ultimate aim of
hostages-giving is to achieve "artificial simultaneity".93 However, the exchange of
hostages has certain problems, mainly the problem for party A to assess whether the
hostage is indeed (or to what extent) of value to party B. Also, and probably more
important, for party A it is virtually never optimal to carry out the inherent threat in
hostage-giving to destroy the hostage. This may allow party B to bargain for the added
value of the transaction.
How could hostages be used to increase transactional security in international
investment? Monaldi (2002) describes hostage-giving in the context of the Venezuelan
oil industry. In Venezuela, investment in the oil industry was significant in the first
half of the 20th century. At that time, external enforcement through the western States,
especially the U.S., and through powerful international oil cartels was strong and the
sovereignty of the Venezuelan government thus low. However, when external
enforcement mechanisms ceased to be effective and the government gained more
sovereign control in the late 1950s and the 1960s, revenue expropriation increased,
leading to less investment by multi-national companies, capacity shortages in the early
1970s and eventually to the nationalisation of the oil industry in 1976.94 Yet, when the
Venezuelan government decided to open the industry again to FDI in the early 1990s,
the country received substantial investment. As Monaldi (2002) asks, how "could this


90
     Kronman (1985), p.11.
91
     On hostages, see also Eger (1995), p.176 ff.
92
     This is the main difference to collateral, see section 3.2.1.2.
93
     Kronman (1985), p.13.
94
     Monaldi (2002), p.3.
Economic and Legal Protection of FDI                                                              25

happen in a country with a clear pattern of expropriation of oil revenues in times of
fiscal need, with a weak judicial branch, and more generally with an apparent lack of
domestic restraints to governmental opportunism?"95 Monaldi’s answer is that hostage
mechanisms played a crucial part in this development.96 He identifies at least two
distinct hostage mechanisms: first, the use of offshore assets and revenues as a
guarantee against expropriation and second, the use of future offshore receivables as a
hostage. The former will be described in a little more detail. As Monaldi (2002) points
out, the Venezuelan government used offshore assets and revenues by its state-owned
oil company PDVSA as a hostage. As a first step, the government made agreements
with the governments with regards to the conditions of the investment. In a second
step, PDSVA contractually agreed to compensate the foreign investors if these
agreements (between the investor and the government) were not respected by the
government. PDSVA held substantial assets in the U.S and Europe. These assets and
the revenues generated by the assets were the actual hostage. To avoid the potentially
biased Venezuelan court system, it was agreed that the contracts between the investors
and PDSVA should be covered by international arbitration.
The logic of hostages can be easily fitted in the time inconsistency model from above.
It is assumed that the HC offers a hostage that will be handed to the investor upon
investment. Now, imagine the extreme case where the hostage has no value to the
hostage-taker (the investor (MNE)), only to the hostage giver (the host country (HC)).
Loss of the hostage (in the following denoted by X where XH denotes the value of the
hostage to the host country and XM the value of the hostage to the MNE) reduces the
gain from expropriation from WH to WH - XH. If, as a consequence, CH > WH - XH, the
hostage will induce cooperation between the players. Nevertheless, the main problem
with hostages can also be demonstrated with a small modification of the game above.
Remember that often, the hostage taker has no incentive to destroy the hostage because
that may incur an additional cost or there may even be costs associated with keeping
the hostage (these costs will be denoted as T as they can be understood as a form of
transaction costs). In addition, the hostage has no or little value to the hostage taker.
We can easily extend the game tree to include a third stage with the choice of the
MNE between destroying or keeping the hostage and returning it:




95
     Monaldi (2002), p.4. See also Monaldi (2001).
96
     The importance of hostages for economic transactions in general in the absence of legal remedies
     has especially been emphasised by Williamson (1983).
26                                                                Economic and Legal Protection of FDI


                                                                 (LM+ XM - T, WH-XH)
           Don't Invest     (0,0)                     Keep


                                                M
            M
                           Accommodate              Return

                Invest                                           (CM, CH)
                              H
                                                                 (LM+ XM - T, WH-XH)
                                                      Keep
                             Expropriate

                                                M


                                                    Return
                                                                 (LM, WH)



Figure 3.2: Hostage-Mechanism



The game presented in figure 3.2 has an implicit assumption, namely that the hostage
will be transferred at the time of investment. Further, we assume that the host country
will expropriate retroactively if the MNE keeps or destroys the hostage despite prior
accommodation of the HC.97 If keeping or destroying the hostage is indeed costly to
the MNE while the value of the hostage itself is low, that is if XM - T <0, returning the
hostage is a dominant strategy for the MNE even if the HC expropriates. The HC will
know this and therefore identify the threat of the MNE to destroy the hostage as non-
credible. Therefore, the HC will expropriate if CH<WH and accommodate otherwise –
the bottom line is that we are dealing with the same game structure as if there had not
been an exchange of the hostage in the first place.
This insight can also be discussed in the context of the above described example of the
Venezuelan oil industry as presented by Monaldi (2002). In contrast to other
commitment mechanisms such as hands-tying, hostages require the action of the
hostage taker in the case of defection by the hostage giver.98 If enforcing the contracts
using international arbitration to actually collect the foreign assets of PDVSA would

97
     Which is not relevant for the point being, but will be relevant in the following section with regard
     to collaterals.
98
     Hands-tying will be discussed in section 3.2.1.3.
Economic and Legal Protection of FDI                                                               27

have been extremely costly or if the assets would have been more or less useless to the
investor (because, for example, a lack of a liquid market for these assets), the hostage
mechanism could not have been used to overcome the time inconsistency problem. In
other words, the fact that the foreign assets are valuable to the respective host country,
in this case Venezuela, cannot guarantee cooperation. If the assets are also very
valuable to the investor/MNE, they are rather referred to as collaterals. As the
following section will illustrate, collaterals solve the specific problem of non-credible
threats inherent with the hostage mechanism, but pose a different set of problems.

3.2.1.2           Collateral
The major difference between a hostage and collateral is that the collateral is equally
valuable to both parties to the transaction. Often, collaterals are assets with well-
defined market value. In this case, the asset (that is, the collateral) is a "direct
substitute"99 for the promised performance. The distinction between a hostage and
collateral can be blurry. Consider again the example of the Venezuelan oil industry as
described above. Whether the offshore assets are considered as hostages or as
collaterals depends on the valuation of the assets by the MNE (including the costs to
obtain them). If the MNE values possession of the offshore assets as much as the
original assets, the assets would be seen as collaterals rather than hostages.
While hostages may create ex-post bargaining power on the side of the hostage-giver
(as the hostage-giver knows that the hostage-taker values performance more than the
hostage), the reverse can be true with respect to collaterals. As Kronman (1985) puts
it: "By giving me collateral that is equal in value to the performance I have been
promised, you create an opportunity for bargaining that I can exploit, if I am skilful
enough, to appropriate the gain you expected to realize from our transaction."100 In
addition and equal to hostages, the use of collateral presupposes under normal
circumstances the existence of substantial offshore assets, which might be true for
some but certainly not all countries that are trying to attract foreign investment.
Consider the use of collateral in our baseline example. As in the hostages example, we
assume that the HC offers a collateral that is handed over if the MNE invests. It is also
assumed that, in contrast to the hostage, the collateral has equal value to the MNE and
to the HC, again denoted by X with X=XH=XM. Loss of the collateral makes
expropriation more costly to the HC relative to accommodation. Consequently,
collaterals can support a cooperative outcome of the game. The problem with



99
      Kronman (1985), p.16. For a description on the functioning of collaterals, see also Eger (1995),
      p.182ff.
100
      Kronman (1985), p.17.
28                                                    Economic and Legal Protection of FDI

collaterals is, as mentioned before, that the collateral taker might prefer keeping the
collateral over completing the original bargain.

                                                     (LM+ X - T, CH-X)
         Don't Invest    (0,0)                Keep


                                        M
         M
                         Accommodate        Return

             Invest                                  (CM, CH)
                           H
                                                     (LM+ X - T, WH-X)
                                              Keep
                          Expropriate

                                        M


                                            Return
                                                     (LM, WH)



Figure 3.3: Collateral



The game structure is similar to the game presented in figure 3.2. Equivalent to the
hostage game, the collateral game models the decision of the MNE to keep or return
the collateral as an additional, third stage. We assume as before that, if the MNE
chooses to keep the collateral although the HC played accommodate, the HC will
expropriate immediately. Therefore, the game sequence (Invest, Accommodate, Keep)
generates the same payoffs as (Invest, Expropriate, Keep) to the parties. Important to
understanding the problem with collateral is the payoff to the MNE. This payoff is
relevant for the decision whether the MNE in the third stage decides to keep the
collateral although the HC accommodated or if the MNE returns the collateral. The
MNE will do the former if LM+X-T>CM. As X, the value of the collateral, represents a
direct substitute for the promised performance, the inequality will often be true. The
HC anticipates this and might not offer the collateral in the first place. Cooperation
cannot be achieved, although the high value of the collateral will prevent the HC from
expropriating the assets. In the example of the Venezuelan oil industry, this would be
the case if the investor could seize the assets of PDSVA too easily, for example if
international arbitration had a strong bias in favour of the investor.
Economic and Legal Protection of FDI                                                29

3.2.1.3          Hands-Tying
The expression "hands-tying" is attributed, as Kronman (1985) notes, to Williamson
and Schelling and describes "actions that make a promise more credible by putting it
out of the promisor's power to breach without incurring costs he could otherwise have
avoided."101 The major difference between hands-tying and hostages is that hands-
tying is self-executive – it requires no additional actions by the parties to the
transaction. For FDI, probably the most widely used form of hands-tying is domestic
investment law enforced by an independent judiciary. In the absence of an independent
judiciary, hands-tying can and will take more complicated forms. To give an example,
imagine that the state or a state-owned bank lends some or all of the money required
for the investment project to the foreign investor. If the government expropriates, the
investor will obviously have to stop the repayments and also any interest payments.
Put differently, as the investor is no longer in possession of the assets that are
supposed to create the income stream flowing to the bank, this income stream will run
dry more or less automatically. If the government cares about the revenues, these are
extra costs triggered automatically by an event of expropriation, making this option
less attractive to governments.


                                       (0,0)
                        Don't Invest


                        M                             (CM,CH)
                                       Accommodate
                              Invest

                                        H


                                        Expropriate
                                                      (LM,WH-D)




Figure 3.4: Hands-Tying



The loss resulting from hands-tying is here denoted by D. As in the examples before,
the HC offers to "tie its hands" and will do so when the MNE invests. The main
advantage of hands-tying, as compared to hostages and collateral, is that the costs


101
      Kronman (1985), p.18.
30                                                     Economic and Legal Protection of FDI

result automatically in the case of non-cooperative behaviour – there is no need to
include a third stage into the game. HC will accommodate if D>WH-CH. One of the
problems with hands-tying is that it is often not an available option. Besides, while the
automaticity of the sanction is desirable as it enhances the credibility of the
commitment, it may be problematic in certain cases, especially when circumstances
have changed.

3.2.1.4          Union
The method of union is another method to align divergent interests. This method is
more comprehensive than the mechanisms described before in the sense that it aims to
eliminate rather than mitigate the divergence of interests.102 Obviously, the term union
describes the situation where two parties merge or integrate to become a single party.
In general, the method of union has its own limitations.103 First, union can often not be
completely achieved, meaning that it can usually not eliminate all conflicts of interest
between the parties. Second, the method of union might, under some circumstances,
increase the risk of exploitation in the short run, for instance through, as Kronman
(1985) puts it, "the relaxation of defenses"104 vis-à-vis the partner in the union, which
might entail sharing of information, etc.
In the context of FDI, transnational integration plays an important role. A huge
proportion of FDI is based on cross-border mergers and acquisitions (“M&A
activity”). In 2007, the amount of FDI resulting from M&A activity was 1,637 billion
USD.105 However, for the purposes of this paper, mergers and acquisitions must be
differentiated: while mergers might be able to mitigate political risks, this is not the
case for acquisitions where the acquired asset is wholly owned by the foreign
investors. In the case of a merger, at least part of the asset is owned by a domestic
party, which might, as described before, reduce the political risk. Obviously, the two
relevant parties to the original transaction, the government and the investor, cannot
merge. Nevertheless, a merger might take place, for example, between a state-owned
enterprise and a foreign company. In that case, the state has a direct incentive to
protect the interests of the new company (the same is obviously true for other
governance structures, like joint ventures). If the domestic company is, however, not
state-owned, the method of union only works if the government places value on the
welfare of the domestic partner to the union. It is important to note that the underlying
reasons for companies to merge are manifold and not primarily rooted in the protection


102
      Kronman (1985), p.22.
103
      See Kronman (1985), p.23.
104
      Kronman (1985), p.23.
105
      UNCTAD (2008a), p.xv.
Economic and Legal Protection of FDI                                                         31

of FDI. Nevertheless, the effect of the ownership structure on political risk is
potentially a strong argument in favour of a cross-country merger.

3.2.1.5          Insurance
Another way to countervail potential under-investment as a result of time
inconsistency is insurance. Many countries have agencies that provide insurance
against political risks like expropriation or similar government measures. The most
prominent might be the U.S. government agency OPIC, the Overseas Private
Investment Corporation. Like many state operated political investment agencies, the
OPIC has, inter alia, the purpose of promoting economic development. According to
the OPIC statute, the purpose of the agency is to "mobilize and facilitate the
participation of United States private capital and skills in the economic and social
development of less developed countries and areas, and countries in transition from
nonmarket to market economies, thereby complementing the development objectives
of the United States […]".106 In addition, there is also a private market for political risk
insurance. According to the World Bank's Political Risk Insurance Centre, the biggest
market place for political risk insurance is organised through Lloyd's, where
individuals and corporations can syndicate to provide political risk insurance.107
The Multilateral Investment Guarantee Agency (MIGA) plays a very important role in
the insurance of political risk. MIGA is part of the World Bank Group and had, as of
September 2010, 175 member countries, 150 of which are developing countries.108 The
preamble of the MIGA Convention emphasises, comparable to OPIC, the development
purpose of MIGA as a measure to promote FDI. As stated in the MIGA Convention,
Article 11, MIGA covers the following risks:
      x Currency Transfer
      x Expropriation and Similar Measures
      x Breach of Contract (by a host government)
      x War and Civil Disturbance
Please note that while insurance can induce investment in reality, it cannot overcome
the time inconsistency problem as such in a framework of complete information and
risk-neutral agents. Obviously, insurance does not really alter the incentives of
governments. If expropriation yields a higher payoff than accommodation, the HC will


106
      See OPIC Statute, sec. 231.
107
      See http://www.pri-center.com/directories/partner_specific.cfm?pgid=5&orgnum=34313   (Sep-
      tember 1st, 2010).
108
      See http://www.miga.org/quickref/index_sv.cfm?stid=1577 (September 1st, 2010).
32                                                             Economic and Legal Protection of FDI

expropriate to the detriment of the insurance company. Under complete information,
no insurance company would offer such an insurance contract. Only if there is
asymmetric information and the law of large numbers applies will a company offer
insurance under such conditions (unless its operation is motivated by development
goals). The demand for insurance naturally requires risk-aversion on the side of
MNEs.

3.2.1.6           Devaluation of Assets
Additionally, investors might adopt strategies to avoid expropriation. One of these
strategies is to make the assets deliberately less profitable to other owners.109 These
strategies often imply additional costs on the investor itself or even entail the
devaluation of the assets. One approach might consist of upgrading the investment
before it is actually worthwhile to do so in order to stay ahead in expertise. Put
differently, the investor will use (costly) new technology that is impossible for the host
country to operate even if using this technology would actually not be efficient. An
alternative approach is to locate different parts of the production chain in different
countries.110 Owning only one part of the production chain may make the asset useless
to the host country. Translated to our stage game, the MNE will try to take measures to
reduce the value of the asset to the host country, namely WH. The problem with this
strategy is that this can be costly, therefore also reducing CM and thus diminishing the
incentive to invest in the first place.

3.2.1.7           Lobbying
It has been argued that the time inconsistency problem applies to welfare and non-
welfare maximizing governments. Consequently, the measures that MNEs and
governments may take to mitigate time inconsistency are mostly independent of the
motivations of the government. A notable exception in this respect is lobbying (and
possibly corruption): the lobbying of MNEs for protection of their investment directly
uses the fact that there are principal-agent conflicts between the government (the
agent) and the citizens (the principal). As in the case of union (see section 3.2.1.4),
lobbying will change the welfare function of the government (and thus reduce the
value of WH in our example) – however, the difference is that lobbying should not be
possible if the state is welfare maximizing. The problem with lobbying is that it is
costly and does not necessarily lead to the desired outcome.111


109
      See for example Eaton and Gersovitz (1983).
110
      Eaton and Gersovitz (1983), p.90ff. Of course, this type of action equally distorts efficiency.
111
      Notwithstanding the problem that lobbying is a form of rent-seeking and can be detrimental to
      social welfare.
Economic and Legal Protection of FDI                                                   33

3.2.2             Dynamic Devices
Dynamic devices refer to mechanisms where time plays a crucial role. For one
example, the payoffs of the games may change over time. This is exemplified in
section 3.2.2.1. Also, reputation, which plays a crucial role in the area of investment,
can only develop over time. The role of repetition and reputation is discussed in
section 3.2.2.2.

3.2.2.1           Expertise and Time
A rational investor might be aware of the time inconsistency problem of the host state
and still find it worthwhile to invest, knowing that he will lose control over the assets.
Eaton and Gersovitz (1983) provide the following example: an investor may realise
that the only protection against expropriation is his monopoly over special knowledge.
Further, he knows that this knowledge will subsequently be passed over to the host
state over time so that finally, expropriation will occur. However, the profits up to the
point of expropriation might be high enough to make the investment profitable
anyway.

3.2.2.2           Repetition and Reputation
A crucial question is whether the time inconsistency problem as described above can
be reduced or even completely overcome through repeated interaction. It is
conventional wisdom in game theory that repeated "interactions give rise to incentives
that differ fundamentally from those of isolated interactions".112 The simplest way to
understand how repetition can induce MNEs to invest would be to consider an infinite
repetition of the stage game. Reputation, however, is more complex than mere
repetition and requires the existence of asymmetric information. Both approaches will
be discussed in the following sections.

3.2.2.2.1         Repetition
One explanation of the occurrence of FDI in the absence of legal protection may be
rooted in the repetitive nature of the game. In such a model, potential gains from
continuous cooperation and the threat of termination of that cooperation may be able
to exert a strong influence on the host country. Consider again the stage game as
presented in section 3.3.1. As in the baseline case, it is assumed WH>CH which implies
that cooperation would fail in a single-shot setting. According to the Folk Theorem,
the Nash equilibrium of the stage game, namely (Don't Invest/ Expropriate), need not



112
      Mailath and Samuelson (2006), p.2.
34                                                               Economic and Legal Protection of FDI

be the outcome of the infinitely repeated game.113 If the stage game is infinitely
repeated, any feasible average payoff combination that is Pareto-superior to the stage
game Nash equilibrium can emerge as a subgame perfect equilibrium when the
discounting of future payoffs is sufficiently small.114 We need to find a pair of
strategies that yield a cooperative result from which none of the parties has an
incentive to deviate. One possible and widely used strategy here is the grim trigger
strategy. We assume that the host country plays a different MNE each round, but
MNEs in later rounds observe the outcomes of prior rounds. The grim trigger strategy
for a given investor is then:


(1) Play invest in the first round.
(2) Continue to choose invest in the following rounds, unless the HC expropriated in
the prior round, in which case play the strategy of the stage game Nash Equilibrium
(Don't Invest/Expropriate) forever.


It is straightforward to show that the host country will always play accommodate and
therefore the equilibrium (invest/accommodate) will prevail when:

             CH
G t 1                                                                                           (3.1)
             WH

where is the discount rate of the host country with 0< <1.115


This very simple model is already able to map some interesting features concerning
the incentives to expropriate in a dynamic setting. Cooperation can be sustained
through repeated interaction, even when there is a different investor in each round. It is
crucial to note that this conclusion depends on the assumption that MNEs that play the
HC in later rounds can observe the outcome of earlier games. One might refer to this
as an effect of reputation. However, most game theoretic accounts, including the
authoritative Fudenberg and Tirole (1991), refer to reputation mainly in the case of
incomplete information: "modelling reputations as complete-information strategies



113
      Please note that the Folk Theorem does not work when the number of repetitions is finite, which is
      the conclusion of Selten's famous chain-store paradox.
114
      Holler and Illing (2006), p.140.
115
      See Appendix A for the derivation of this result.
Economic and Legal Protection of FDI                                                                35

cannot capture the idea that a player's reputation corresponds to something that his
opponents have learned about him."116
Equation (3.1) shows the condition for cooperation. If the ability of the host state to
use the investment efficiently, that is WH, is high, then expropriation becomes more
likely, while the ability to extract more of the rents of cooperation in the cooperative
outcome (CH) reduces the incentives to expropriate. Furthermore, and a standard result
of models of this sort, is the fact that cooperation is more likely when the host state
values future payments highly, namely when is big. A shortcoming of these kinds of
models that are based on the Folk Theorem is the strong dependence on the type of
trigger strategy. In fact, an indefinite number of other equilibria are possible and
accurate predictions based on the attained equilibrium are therefore impossible to
sustain.

3.2.2.2.2         Reputation
In the repetition model, the assumption of an infinite time horizon is unrealistic,
especially given the rather short time horizons of political decision-makers.117 Also,
investors might not be perfectly informed about the payoffs of decision-makers. Some
of these issues can be addressed through a model of a finitely repeated game of
asymmetric information or, put differently, a reputation model. The functioning of
reputation in FDI settings has also been described by Veugelers (1993) and Thomas
and Worrall (1994). The model by Veugelers is an application of the seminal game-
theoretic description of reputation by Kreps and Wilson (1982).118 The application of
game-theoretic models to the issue of reputation is not unproblematic and requires the
use of mixed-strategy equilibria in games of incomplete information. The logical
structure of the game in our context, which is also followed by Veugelers (1993), is as
such: the basic assumption is that there are two types of states, a "good" (or reliable)
and a "bad" (or unreliable) host country. While the host country knows if it is "good"
or "bad", the MNE does not have this information. However, as the game proceeds,
each MNE observes the prior moves. The MNE updates its beliefs concerning the type
of the host country according to Bayes' Rule. Similar to the grim trigger strategy used
in section 3.2.2.2.1, it can be assumed that MNEs stop investing once they have
observed the expropriation outcome (or use a similar punishment strategy). For certain


116
      Fudenberg and Tirole (1991), 367. See also Gibbons (1992), p.225. The following section will
      address the topic of reputation directly.
117
      Introducing limited time-horizons of political decision-makers might be considered as a deviation
      from the assumption of a social welfare maximizing government. However, another interpretation
      could be that this reflects merely limited time-horizons of the constituents.
118
      For a useful description of the Kreps and Wilson reputation game, see Holler and Illing (2006),
      p.168ff.
36                                                             Economic and Legal Protection of FDI

values of the parameters and the prior beliefs, it is possible to define a subgame perfect
equilibrium where the bad host country mimics the behaviour of the good type. More
specifically, the bad host country will randomise between its strategies. As soon as the
randomised outcome is expropriate, the MNEs will stop investing in the future. As
long as the randomised outcome is accommodate, the MNEs update, as before, their
beliefs using Bayes' Rule and increase their belief that the MNE is actually of the good
type. As the host country takes the beliefs of the MNEs into account, it will in turn
increase its probability of playing accommodation. The reputational capital rises until
the last round, where the bad type expropriates with certainty. Veugelers (1993) builds
a rich model based on this logic and finds that the reputation outcome will prevail for
intermediate values of the discount factor. In addition, much depends on the initial
probability that the MNE places on the possibility that the HC is reliant. The more
pessimistic these priors are with regard to the realibilty of the HC, the weaker are the
incentives of the bad HC to accommodate – with the result that the (likely)
expropriation will happen earlier.
The fact that the players should play mixed strategies might be confusing at first
glance. If mimicking the behaviour of the good host country is worthwhile for the bad
type, why not play the good type's strategy all the time? This cannot be an equilibrium
for the following reason: the incentive for a bad host country to play accommodation is
based on the idea that his behaviour induces the MNE to invest in the following
period. However, if the bad host country always plays accommodate, the MNEs will
not receive new information through this behaviour. Thus, playing accommodate does
not influence the action of the MNE in the following period. Consequently, the bad
host country has no incentive to play accommodate in the first place. On the other
hand, if the host country never plays accommodate, he cannot build any reputation at
all.119 This can also be highlighted in a two-period setting using the basic trust game
with only a few modifications. Most importantly, there must be the possibility that the
HC is not of the (bad) type as described in section 3.1.1, but of the good type as
exemplified in figure 3.5.




119
      A similar argument can be made for the MNEs playing mixed strategies. See Appendix B.
Economic and Legal Protection of FDI                                                        37


                                        (0,0)
                         Don't Invest


                          M                             (CM,CH)
                                        Accommodate
                              Invest

                                          H


                                          Expropriate
                                                                LOW
                                                        (-LM , WH )




Figure 3.5: Time Inconsistency and the Good Host Country (HCgood)



with WHLOW  WH and WHLOW  CH


The "good" host country always prefers the cooperative over the uncooperative
outcome. A second deviation from the trust game as introduced in section 3.1.1 is that
the expropriation payoff for the MNE has been set specifically to -LM with LM>0.120
The game proceeds as described above, yet limited to two periods. The MNE is
uncertain about the host country’s payoffs. It assigns a certain probability to the
possibility that payoffs are either those specified in figure 3.1 (with -LM) or those in
figure 3.5. The investment decision in the second period depends on the observation
made in the first period. In this simple two period setting, a subgame perfect Bayesian
equilibrium where the bad type builds a reputation can be described as follows:121


Notation
 : a-priori probability that the host country is of the good type (prior)
q2: probability that the host country is of the good type as assessed by the MNE in
period two (after observing the history of the game based on Bayes' Rule)
z: probability that the bad type will play accommodation in the first period


120
      This assumption does not alter the game but only the notation. This change simplifies the
      derivation of the equilibrium. See Appendix B.
121
      For a derivation of this result, see Appendix B.
38                                                         Economic and Legal Protection of FDI

v: probability that the MNE plays invest


q2 is computed the following way:
(a) if there is no investment in the first stage, then q2 =
(b) if there is investment in the first stage and this investment is accommodated, and
>0, then the MNE updates his belief according to Bayes' Rule,
(c) if there is investment in the first stage and this investment is expropriated, then q2
=0


                                                                            L2
The proceeding of the game can be described as follows: for T               M
                                                                                     , the MNE
                                                                       ( LM  CM ) 2

will not invest in the first round. The HC consequently cannot build a reputation. Thus,
                                                                                 LM
the MNE will also not invest in the second round (as in that case, T                  has to
                                                                               LM  CM

hold; please note that the “required prior” for investment in the second round needs to
be larger, because in the second round, the bad type would definitely expropriate and
                                                                           L2
never mimic the strategy of the good type). However, for T !                M
                                                                                    , there is a
                                                                      ( LM  CM ) 2

perfect Bayesian equilibrium (“PBE”) where the MNE will invest in the first round. In
the reputation equilibrium (that we are interested in), the bad HC will randomise and
                                               T ª CM  LM º
play accommodate with probability z                  «     1» . If the outcome of this
                                            (1  T ) ¬ LM    ¼
randomisation process was accommodate, the MNE will update his belief according to
                                               WH  CH
Bayes' Rule and invest with probability v              .
                                                G WH

 What is of interest here is the behaviour of the bad HC in the first round. Indeed,
under certain parameter specifications, a pooling (or mimicking) equilibrium occurs.
Pooling equilibrium means that the bad HC plays accommodate, which is the same
strategy the good type would play as well. Whether he mimics in the first period
depends crucially on the priors ( T or prior beliefs) of the MNE. If the priors are
sufficiently high, the bad HC randomises. If, as a result of the randomizing, the bad
Economic and Legal Protection of FDI                                                                 39

HC nevertheless plays expropriate, there will be no more investment by the MNE in
the second round.
So, can reputation really solve the time inconsistency problem? Some qualifications
must be made with respect to reputation models (as presented here), but also with
regard to reputation in general. It has been highlighted that the reputation equilibrium
depends on the fact that players are using mixed-strategies, which not only
considerably complicates the analysis, but may also cause interpretation problems.
Nevertheless, it appears to be common sense in economic theory that the mixed
strategies need not be interpreted as strategies in the strict sense but rather as a
probability distribution over identical players.122 A second criticism concerns the fact
that the prior beliefs are exogenous and not determined within the model. Therefore, if
one only assumes high enough priors, cooperation will occur with certainty. However,
as Drazen (2002) asserts in a similar context, this is rather "a description of reality".123
Of course, countries will rather cooperate when the disposition of investors, as
embodied in their priors, to invest is high – there will be more to gain in the future.
Lastly, a common problem (and comparable to the repetition case) in these models is
the multiplicity of equilibria.124 This might also simply be a description of reality
rather than a problem as such. Naturally, it must be acknowledged that the existence of
multiple equilibria limits the predictive power of these sorts of models. Please note
that the main purpose of the model in this context is to show that bad types may
cooperate – even when there is the endgame problem. Another crucial point is the
importance of the priors. In this context, this is also a shortcoming since the HC can
only build a reputation if the MNE invested in the first place; if the priors are low, the
HC does not even have the chance to build a reputation.
In sum, reputation may support cooperation, but is no panacea: the behaviour of the
HC depends on the time horizon and reputation can only play a role when MNEs
cannot differentiate between host countries (put differently, the existence of
asymmetric information is a prerequisite – which is an assumption that could be
challenged in the case of international investment law). Also, as in the repetition
model, MNEs must use a punishment strategy to induce cooperation. This, in turn,
demands that MNEs are able to observe the host country's behaviour.




122
      Rasmusen (2001), p.69.
123
      Drazen (2002), p.182.
124
      See also Drazen (2002), p.187. This problem is not as evident in the two-period example used here,
      but is prevalent for example in the model presented by Veugelers (1993).
40                                                             Economic and Legal Protection of FDI

3.2.3             Discussion
It has been argued that there is a time inconsistency problem in the area of foreign
direct investment and that this problem is relevant to investors and policy-makers.
From an economic point of view, several options are conceivable to mitigate this
problem and have been discussed. While all of these options might have the potential
to make FDI more secure, each has its own limitations.125 Hostages, collateral and
hands-tying might not always be available. In addition, hostages suffer from the
inherence of non-credible threats and collateral from the potential of opportunistic
behaviour. Unions are also prone to opportunistic behaviour and a suitable partner
with whom to build a union is not necessarily always at hand. Other forms of
protection, like insurance, the devaluation of assets and lobbying, are affected by high
costs and uncertain impact. Repetition and reputation can be undermined by limited
time-horizons and high discount rates. In sum, none of these mechanisms can
completely remedy the time inconsistency problem. In the next section, the legal
protection of FDI, especially through international law and, more specifically,
Bilateral Investment Treaties, will be described. These treaties will then, in chapter 4,
be integrated into the context of the economics of FDI protection.

3.3               Legal FDI Protection
Foreign direct investment is legally protected on the national (domestic) level and on
the international level. The protection of FDI in international law includes customary
international law, multilateral treaties and bilateral treaties. Bilateral treaties can be
directly concluded between an investor and a state (investor-state contracts) and
between two states (for instance through Bilateral Investment Treaties, BITs). The
multitude of instruments and treaties that relate to FDI protection is remarkable. Sauvé
and Zampetti (2007) rightly note that providing "a concise but reasonably
comprehensive summary account of the international legal framework for foreign
investment is thus a difficult task, because there are so many international rules that
have a clear impact on foreign investment, spanning such fields as taxation,
intellectual property, trade in services, antitrust, labour-relations and corporate social
responsibility."126 Therefore, this chapter will describe only the most important of the
legal mechanisms with a focus on Bilateral Investment Treaties.




125
      A discussion on the major ex-ante mechanisms with regard to their specific problems can also be
      found in Eger (1995), p.187. Efficient incentives cannot be achieved under all circumstances.
126
      Beviglia Zampetti and Sauvé (2007), p.211.
Economic and Legal Protection of FDI                                                                   41

3.3.1             Domestic Regulation
States that are recipients of foreign direct investment are obviously free to adopt
domestic regulations intended to protect foreign investment and have done so with
varying strictness. In 2006, according to the UNCTAD database on national laws and
regulations, 93 countries made a total of 184 changes in their national regulations with
regard to FDI.127 Of these changes, 147 were favourable to FDI – which constitutes
almost 80%. The importance of adequate national regulations as a tool to attract
foreign investment has obviously not escaped policy-makers. To give an example, in
the newly released Investor’s Guide to São Tomé and Príncipe, the positive national
regulatory framework with regard to FDI has been strongly emphasised.128 It should be
noted that home countries might also have national regulations aimed at influencing
the activities of their constituents abroad. These measures can have either promoting
or rather restricting effects depending on the motivation of the home state. For
example, states may promote foreign investment as part of their development policy,
such as through insurance schemes. Alternatively, states may also restrict capital
outflows on grounds of balance-of-payments concerns.129

3.3.2             Customary International Law
Customary International Law (CIL) can be defined as a "general and consistent
practice of states followed by them from a sense of legal obligation".130 With regard to
the treatment of foreign investment, disagreement prevails as to whether CIL indeed
exists and, if so, what standard of treatment it demands.131 This disagreement has a
long history, reflected by the disagreement between developing and developed
countries regarding the amount of compensation in the case of expropriation
throughout the past century.132 Regardless of the standards set by CIL, it offers foreign
investors little effective protection on its own, as it lacks a binding mechanism to settle
investment disputes.133




127
      UNCTAD (2007a), p.14.
128
      Guide available under http://vcc.columbia.edu/pubs/documents/SaoTome-sept11eng.pdf (October
      11, 2009).
129
      UNCTAD (2004), p.7.
130
      Restatement (Third) of the Foreign Relations Law of the United States §102(2), quote taken from
      Goldsmith and Posner (1999), p.1113.
131
      See Sornarajah (2004), p.205, rejecting the notion that there is CIL in investment matters.
      Herdegen (2005), p.203, alternatively claims that certain minimum standards as to the treatment of
      foreign investment can be identified. A short illustration of the relationship of BITs to CIL can be
      found in section 3.3.3.9.
132
      See, e.g., Lowenfeld (2002), p.391 ff. for an historical account of this disagreement.
133
      Salacuse (1990), p.659f.
42                                                                 Economic and Legal Protection of FDI

Prior to World War I, there was a general acceptance between nations that takings of
private property would trigger an obligation to pay prompt and adequate compensation
and that aliens were entitled to equal treatment compared to nationals of the host state.
The revolutionary movements in Russia and especially Mexico, which involved the
expropriation of foreign held assets, introduced the approach that the protection of
private property was to be subordinated to the common welfare. The Western nations
protested these uncompensated takings. Subsequently, Mexico and other countries of
Latin America advocated the so-called Calvo doctrine. This doctrine traces its roots to
the writings of 19th century Argentine jurist Carlos Calvo. In its narrowest
interpretation, the Calvo doctrine postulated simply that aliens should have no greater
rights than the citizens of the host country. However, some countries adopted a
broader interpretation, stipulating that foreign investors could obtain no diplomatic
protection from their home states at all. The classical Western view, especially
promoted by the United States, called for an international minimum standard for the
protection of foreign investment. In this context, the so-called Hull formula gained
importance, dating back to a diplomatic exchange between US Secretary of State
Cordell Hull and the Mexican Ambassador to the United States in the 1930s as a
consequence of unresolved expropriations of property owned by American citizens.
This formula claimed that compensation should be prompt, adequate and effective.
The quarter century following World War II saw a wave of expropriations and
nationalisations in all parts of the world, for example in Eastern Europe and Latin
America. These expropriations were typically followed by negotiations between the
relevant states and usually resulted in some type of compensation. However, while the
Hull formula (also called Hull rule) and the Calvo doctrine played a role in the judicial
discourse, no approach gained general acceptance.134 Attempts to create an
international investment law through multinational agreements, such as through the
United Nations and its organs, remained without effect. The 1962 UN Resolution on
Permanent Sovereignty reflected only a short-lived compromise that did not align the
opposing views on the appropriate protection of foreign investment.135 Another
important event that reflects the controversy over the protection of FDI was the call for
the New International Economic Order in the United Nations General Assembly in
1974. One of its cornerstones was the focus on national remedies in cases of
nationalisation, expropriation or transfer of ownership of foreign property.136 This
focus, which reflected the view expressed in the Calvo doctrine, was however
undermined by the growing spread of Bilateral Investment Treaties. These treaties,



134
      Neither could the Soviet position that no compensation should be paid at all.
135
      Lowenfeld (2002), p.415.
136
      Dolzer and Schreuer (2008), p.15.
Economic and Legal Protection of FDI                                                         43

which will be described in section 3.3.5, generally contained a compensation standard
in accordance with the Hull rule. The trend towards increased economic liberalism,
starting in the late 1980s and reflected in the Washington Consensus, supported the
protection standards as expressed in BITs.137
In summary, the question on the existence and extent of protection of foreign
investment in customary international law remains unsettled. There has been new
research that deals with the question of how far the aforementioned large amount of
BITs contributes to the emergence of CIL.138 Nevertheless, in the absence of other
international agreements that cover FDI, the protection of these assets through CIL can
safely be described as weak.

3.3.3            Multilateral Treaties
Until now, all attempts to establish a comprehensive multilateral agreement have
failed. Consequently, no truly international multilateral treaty on the regulation of
foreign direct investment exists. Promising attempts in recent history were launched
through the Organisation for Economic Co-operation and Development (OECD) and
the World Trade Organization (WTO). The draft of the OECD was referred to as the
Multilateral Agreement on Investment (MAI).139 The project was launched in 1994
and mirrored the protection standards that were at that time long established in
Bilateral Investment Agreements. Intensive negotiations were held between 1995 and
1998, when the process finally came to a halt. Dolzer and Schreuer (2008) state a
number of reasons for this, including the lack of support from the United States and
France and intense criticism by NGOs over the lack of public involvement in the
negotiations. The focus then shifted to the efforts within the WTO. However,
discussions on a multilateral investment treaty, which were put on the agenda during a
meeting in Singapore in 1996, were abandoned in 2004. According to Dolzer and
Schreuer, the negotiations failed mainly due to the resistance of a number of
developing countries, especially India and Brazil. Despite having concluded a big
number of BITs with comparable content, these countries feared that a multilateral
agreement under the auspices of the WTO would limit their regulatory sovereignty to
an undue extent.
Although a comprehensive multilateral treaty on foreign investment has never come
into existence, a number of important regional investment agreements have been
concluded. Also, various regional trade agreements and a number of sectoral


137
      Dolzer and Schreuer (2008), p.15f.
138
      This question will be covered in greater detail in section 3.3.3.9.
139
      On the failure of the MAI and the discussion within the WTO, see e.g., Dolzer and Schreuer
      (2008), p.26.
44                                                             Economic and Legal Protection of FDI

agreements cover investment issues. The Energy Charter Treaty and the North
American Free Trade Agreement (“NAFTA”) are probably the most important two
examples in this respect. Before turning to these treaties, the investment regulations
that form a part of international trade law will first be discussed.

3.3.3.1           WTO
In spite of the failure of the WTO negotiations with regard to a multilateral treaty,
certain aspects of foreign investment interfere with trade matters and are thus covered
on the margin of WTO regulations. As the most prominent codification in this respect,
the Agreement on Trade Related Investment Measures (TRIMs Agreement) should be
mentioned. The TRIMs Agreement was enacted during the Uruguay round of trade
negotiations and applies to investment measures related to trade in goods only.140 Prieß
and Berrisch (2003) emphasise that the TRIMs Agreement has not yet played an
important role in the WTO dispute settlement procedures.141 The TRIMs Agreement
bans performance requirements, such as local content or trade balancing requirements,
when they are considered to distort trade.142 Also the General Agreement on Trade in
Services (“GATS”) and the Agreement on Trade-Related Aspects of Intellectual
Property Rights (“TRIPS”) have certain links to matters of foreign direct
investment.143

3.3.3.2           Energy Charter Treaty
The Energy Charter Treaty (ECT), along with the Protocol on Energy Efficiency and
Related Environmental Aspects, was signed in Lisbon in December 1994. With the
ratification by the first thirty members in April 1998, the ECT entered into force. The
Energy Charter Treaty has now been ratified by 52 countries, but is limited to
investments in the energy sector only.144 More specifically, part III of the treaty covers
the promotion and protection of investment and contains standards of treatment like
national and most-favoured-nation treatment. Disputes between an investor and a
contracting state can be settled through international arbitration. A recent case that has
drawn some attention to the ECT in Germany is Vattenfall v. Germany 145 concerning
environmental restrictions on the construction of a power plant imposed by the city of
Hamburg.



140
      See Article 1 of the Agreement on Trade-Related Investment Measures.
141
      Prieß and Berrisch (2003), p.334.
142
      Sornarajah (2004), p.73.
143
      Sornarajah (2004), p.297ff.
144
      Herdegen (2005), p.224.
145
      ICSID Case No. ARB/09/6.
Economic and Legal Protection of FDI                                                              45

3.3.3.3           NAFTA
The North American Free Trade Agreement (“NAFTA”) is an agreement between the
United States, Canada and Mexico that was formed on January 1st, 1994. As the name
implies, it is mainly a free trade agreement between the three states. Nevertheless,
chapter 11 of NAFTA covers investment issues. In short, chapter 11 includes roughly
the same provisions with regard to investment that can be found in BITs.146 These
enclose, among other aspects, national treatment, most-favoured-nation treatment, the
prohibition of performance requirements and compensation for expropriation. NAFTA
is an important treaty in international investment law as a number of arbitrations under
NAFTA have garnered significant public attention and the arbitration awards have
significantly influenced the interpretation of clauses also within the context of BITs.

3.3.4             Investor-State Contracts
Bilateral agreements on investment matters can be made either between the investor
and the host state or between two states. The first are referred to as Investor-State
Contracts.147 These contracts are not limited to questions of investment protection, but
usually encompass clauses on these issues. Obviously, not every contract between
investors and states is considered to be part of international law. However, through the
use of choice-of-law and stabilisation clauses it has been attempted to
“internationalise” investor-state contracts. Whether an internationalisation of investor-
state contracts is actually possible under international law remains an unresolved
question.148

3.3.5             Bilateral Investment Treaties
The modern legal landscape in international investment law is mainly shaped by the
impressive amount of bilateral agreements between states on investment matters.
These Bilateral Investment Treaties (“BITs”) are the main source of international
investment law and constitute the main focus of this work. Generally, the main
function of a BIT is reflected by its name. For example, the Bilateral Investment
Treaty between the US and Argentina is called: “Treaty between the United States of
America and the Argentine Republic concerning the encouragement and reciprocal
protection of investment.”149 Though obligations in BITs are reciprocal, they are in
practice generally concluded between industrialised, capital-exporting states and


146
      See section 3.3.5.
147
      See Comeaux and Kinsella (1997), p.133, for the nature and content of these kinds of contracts.
148
      Herdegen (2005), p.214.
149
      This treaty was concluded in 1991 and is available under http://www.unctad.org/
      sections/dite/iia/docs/ bits/argentina_us.pdf (October 12, 2009).
46                                                                Economic and Legal Protection of FDI

developing, capital-importing states.150 Clearly, for the first group of countries, it is the
aspect of the protection of the investments by their nationals that is the main objective
for concluding a BIT, while the latter countries focus on the encouragement of
investors to invest in their country.

3.3.5.1           Overview and History
Bilateral Investment Treaties are the cornerstone of the protection of foreign direct
investment in international law. By the end of 2009, 2750 BITs had been concluded
(with 82 new treaties concluded in 2009).151 UNCTAD counts 179 countries that have
signed at least one BIT.152 The significance of BITs can best be understood through a
short review of the historical development of international investment law.153 The
emergence of bilateral investment agreements took place within the controversial
setting outlined in section 3.3.2. Customary international law offered no consensual
standards on the extent of protection of FDI and the different positions on this matter
showed no sign of convergence. The legal landscape changed when, in 1959, the first
BIT was concluded between Germany and Pakistan.154 This event was followed by a
slow but steady growth of the number of BITs during the 60s, 70s and 80s. In early
1991, about 440 of these treaties had been concluded. The 15 years that were to follow
saw the strongest increase in BITs, leading to the substantial number of BITs in effect
currently. During the expansion of BITs, treaties were also signed by countries that
were traditionally sceptical of strong investment protection, e.g., many Latin American
countries. While the number of BITs continues to rise, the speed of growth seems to be
declining recently.155 In the recent past, the spread of BITs has differed between
regions. While many Asian countries are actively expanding their BIT network, Latin
American countries concluded only 4 BITs in 2007 and some countries in that region
took steps towards withdrawal from the modern system of FDI protection.156 Most
prominently, Bolivia withdrew from the International Centre for the Settlement of


150
      This pattern seems to be changing recently. But even in a South-South relationship, one country
      might still be the “capital-importer”, while the other is the “capital-exporter”.
151
      UNCTAD (2010), p.81. However, it must be noted that not all of these treaties are actually in force.
      See Sauvé and Zampetti (2007), p.215.
152
      UNCTAD (2008c), p.2.
153
      The following paragraph is based on Lowenfeld (2002), p.391 ff. and Sornarajah (2004), 37 ff.
154
      It must be noted that the modern BITs had predecessors in the form of Treaties of Friendship,
      Commerce and Navigation (FCN) which were concluded from the 18th century onwards. See
      Sornarajah (2004), p.209. However, these treaties covered a wide range of issues apart from
      investment matters and offered only little effective protection to investors. FCNs were thus no
      longer considered an appropriate instrument for the protection of foreign investment. Dolzer and
      Stevens (1995), p.11.
155
      See UNCTAD (2006), p.26 and UNCTAD (2008c), p.2.
156
      UNCTAD (2008c), p.6.
Economic and Legal Protection of FDI                                                               47

Investment Disputes (ICSID) in 2008. Ecuador denounced 9 of its 25 BITs and has
initiated renegotiations with regard to its remaining BITs.157 Equally noteworthy is the
trend towards the conclusion of BITs between developing countries themselves (often
called South-South BITs), as formerly BITs were mainly concluded between a
developed and a developing country. More specifically, in 2007, 26% of all new BITs
were concluded between developing countries.158

3.3.5.2           Treaty Practice and Treaty Interpretation
BITs vary in the degree of protection they grant to foreign investors, but resemble each
other in their underlying structure. The following section describes some of the main
contractual clauses of a typical BIT.159 As Karl (2008) notes, recent treaties have
become more complex in their contents and start to reflect public concerns, such as the
environment and labour rights.160
However, before examining these treaty clauses in greater detail, it is worthwhile to
take a brief look at the methods usually employed by tribunals to interpret these
clauses. Starting point for most tribunals is Article 31 of the Vienna Convention on the
Law of Treaties (VCLT).161 Article 31 of the VCLT determines the general rules of
treaty interpretation. These rules include, among other things, that the treaty shall be
interpreted in good faith in accordance with the ordinary meaning to be given to the
terms of the treaty in their context and in the light of its object and purpose.162
Preambles and Annexes naturally have to be taken into account,163 as well as any
subsequent agreements between the parties regarding the interpretation of the treaty164,
subsequent practice in the application of the treaty that establishes the agreement of
the parties regarding its interpretation165 and any relevant rules of international law
applicable in the relations between the parties.166 In some cases, and in accordance
with Article 31 of the VCLT, the preparatory work (Travaux Préparatoires) of the
ICSID Convention have been consulted as a supplementary means of interpretation.167


157
      UNCTAD (2008c), p.6.
158
      UNCTAD (2008c), p.4.
159
      For a comprehensive survey about treaty practice including a large number of representative
      clauses, see Dolzer and Stevens (1995). A more recent overview is Dolzer and Schreuer (2008).
160
      Karl (2008), p.229.
161
      See Dolzer and Schreuer (2008), p.31 and the examples given therein.
162
      Article 31 (1) VCLT.
163
      Article 31 (2) VCLT.
164
      Article 31 (2(a)) VCLT.
165
      Article 31 (2(b)) VCLT.
166
      Article 31 (2(c)) VCLT.
167
      As Dolzer and Schreuer (2008) note, the travaux préparatoires of BITs are usually not available.
      See Dolzer and Schreuer (2008), p.33.
48                                                               Economic and Legal Protection of FDI

With regard to BITs, and unlike in arbitration cases under NAFTA, interpretative
statements by one of the parties usually do not play a crucial role in treaty
arbitrations.168 Concerning precedent, it must be noted that tribunals are not bound at
all by previous cases, which prevents the development of a coherent body of case law.
Nevertheless, tribunals have frequently taken earlier arbitration awards into account. A
citation analysis considering 207 publicly available decisions and awards from 1972
onwards reveals that ICSID tribunals cite earlier decisions "in much the same manner
than common law courts do".169 Especially since 2001, tribunals have increasingly
based their decisions on prior arbitration awards.170 An increase of the number of
citations of earlier tribunals can also be found in cases that were arbitrated under the
ICSID Additional Facility and other ad hoc arbitrations, such as those governed by the
UNCITRAL rules.171 Tribunals have also taken into account the numerous decisions
rendered by the Iran-US Claims Tribunal. However, the precedential value of these
decisions and arbitral awards in general must be considered carefully as the
circumstances and the underlying treaty of the respective investment disputes may
differ strongly and these differences naturally must be taken into consideration.172

3.3.5.3           Preamble and Definitions
Bilateral Investment Treaties usually start with a statement of intention, articulating
the aim to protect and foster reciprocal investment, followed by a definition of certain
important expressions used in the treaty.173 Most importantly, this part of BITs defines
what kind of investor and what kind of investment is protected.

3.3.5.3.1         Investor
Investors can either be natural or judicial persons.174 The right of an investor to invoke
protection under a BIT naturally depends on his nationality. For natural persons,
nationality is determined by the regulations of the home state. For judicial persons like
corporations, the determination of nationality is often more complex.175 The German

168
      Dolzer and Schreuer (2008), p.34f. In the case of NAFTA, the Free Trade Commission (FTC) can
      and has issued binding interpretations of the treaty. It must be noted that this body consists of
      representatives of all three member countries.
169
      Commission (2007), p.5.
170
      Commission (2007), p.6.
171
      Commission (2007), p.6. On the different fora of arbitration, see section 3.3.3.8.
172
      Sornarajah (2004), p.96.
173
      See the 2004 US Model BIT, available under http://www.state.gov/documents/
      organization/117601.pdf (September 1st, 2010) or the German Model BIT, available under
      http://www.unctad.org/sections/dite/iia/docs/Compendium//en/201%20volume%207.pdf
      (September 1st, 2010).
174
      See German Model BIT, Article I(3) (a).
175
      Dolzer and Schreuer (2008), p.49.
Economic and Legal Protection of FDI                                                                 49

Model BIT covers the investment of "any judicial person as well as any commercial or
other company or association with or without legal personality having its seat in the
territory of the Federal Republic of Germany, irrespective of whether or not its
activities are directed at profit".176 In other treaties, the place of incorporation, rather
than the seat, is relevant.177 Some treaties go beyond these formal definitions and
employ criteria like the effective control over the corporation by nationals of the state
or, alternatively, of genuine economic activity of the company in the state.178
Obviously, interpretation of these rather ambiguous criteria may give rise to
conflicting interpretations.

3.3.5.3.2         Investment
The scope of the application of the treaty is not only limited by the definition of the
term investor, but also by the definition of what exactly constitutes an investment.
According to the US Model BIT, an investment "means every asset that an investor
owns or controls, directly or indirectly, that has the characteristics of an investment,
including such characteristics as the commitment of capital or other resources, the
expectation of gain or profit, or the assumption of risk."179 The Model BIT presents a
non-exhaustive list of examples, including enterprises, shares, loans, intellectual
property rights, licenses and permits. Two questions have occupied arbitral tribunals
with respect to shareholding: first, whether a shareholder can pursue claims under a
BIT on the basis of its own nationality (if the nationality of the company as such does
not qualify for BIT protection) and second whether minority shareholding is sufficient
to invoke protection through a BIT.180 As Dolzer and Schreuer (2008) note, arbitral
practice has answered both of these questions positively. Even when the company
itself does not meet the nationality qualifications, the foreign investor is entitled to
pursue the claim in its own name. This also holds true for minority shareholders. In the
case CMS v Argentina, the claimant CMS held only 29.42% of TGN, a company
incorporated in Argentina that was negatively affected by a government regulation.181
The high extent of BIT protection may, according to Dolzer and Schreuer (2008),
cause future problems as multiple claimants have a right to sue and therefore, different
tribunals will be concerned with the same case at the same time.182


176
      German Model BIT, Article I (3) a ii.
177
      Dolzer and Schreuer (2008), p.49.
178
      Dolzer and Schreuer (2008), p.51.
179
      US Model BIT, Section A, p.3.
180
      See Dolzer and Schreuer (2008), p.56ff.
181
      Dolzer and Schreuer (2008), p.58. Please note that 29.42% is still well above the threshold of the
      way portfolio investment is defined. So far there appears to be no case law on whether portfolio
      investment is covered or not.
182
      Dolzer and Schreuer (2008), p.59.
50                                                     Economic and Legal Protection of FDI

3.3.5.4           Admission
Admission relates to the right of investors to buy or start a business in a foreign
country. It is normally acknowledged that entry depends on the relevant host state’s
legislation (that is, there is no absolute right of entry for foreign investors). The
German Model BIT reflects this perception and states: "Each Contracting State shall in
its territory promote as far as possible investments by investors of the other
Contracting State and admit such investments in accordance with its legislation."183
Governments may be opposed to FDI for social and environmental but also
protectionist reasons.
The US treaty practice substantially differs in this respect from the approach of other
countries, as US BITs specifically demand national treatment on the question of
admission. Article 3 (1) of the US Model BIT reads: "Each Party shall accord to
investors of the other Party treatment no less favourable than that it accords, in like
circumstances, to its own investors with respect to the establishment, acquisition,
expansion, management, conduct, operation, and sale or other disposition of
investments in its territory." This strategy has also been adopted by Japan and
Canada.184 Nonetheless, it must be noted that exceptions may be applied to certain
sectors and subsectors.185 Consequently, completely unrestricted access is rarely found
in practice. In addition to the right of admission, recent US treaties also included the
prohibition of certain performance requirements, such as contents requirements, as a
prerequisite for admission.186

3.3.5.5           Standards of Treatment
With regard to treatment, once investment has been admitted, there are a number of
standards that can be found in different formats in almost all investment agreements.
The most important ones are fair and equitable treatment, most-favoured-nation and
national treatment as well as full protection and security.

3.3.5.5.1         Fair and Equitable Treatment
Article 2 (2) of the German Model BIT demands that each "Contracting State shall in
its territory in any case accord investments by investors of the other Contracting State
fair and equitable treatment as well as full protection under the treaty". Currently, this
standard, referred to as the Fair and Equitable Treatment (FET) standard, appears to



183
      German Model BIT, Article 2 (1).
184
      Dolzer and Schreuer (2008), p.81.
185
      Dolzer and Schreuer (2008), p.81.
186
      Lowenfeld (2002), p.474.
Economic and Legal Protection of FDI                                                                51

be the most important standard in investment disputes.187 Dolzer and Schreuer (2008)
remark that successful claims in international investment arbitration are in most cases
grounded on the FET standard. Although clauses with regard to fair and equitable
treatment have been part of the vast majority of treaties since the 1970s,188 the
importance of the FET standard only manifested itself rather recently, namely in the
year 2000 with the cases Metalclad v. Mexico and Maffezini v. Spain.
The exact meaning of the fair and equitable treatment standard is controversial. This
also holds true for most of the standards in BITs, apparently as a result of the
vagueness of these standards. An ongoing discussion continues over whether the fair
and equitable treatment standard merely reflects the protection offered through CIL or
whether it constitutes an autonomous standard on its own that goes beyond the
minimum standard in international law.189 Arbitral decisions on disputes relating to
Bilateral Investment Treaties have shown a tendency to treat the fair and equitable
standard as an autonomous standard whose interpretation depends on the treaty
context.190 Schreuer (2005), reviewing the case law, developed a non-exhaustive list of
criteria that are of relevance to the application of the fair and equitable standard. These
principles include transparency and the protection of the investor’s legitimate
expectations, freedom from coercion and harassment, procedural propriety and due
process as well as good faith.191 Particularly the notion of legitimate expectations has
been taken up in a number of important arbitration awards, including Saluka
Investment v. Czech Republic, Azurix Corp. v. Argentina and CMS v. Argentina. The
tribunal in Saluka Investment v. Czech Republic describes the notion of legitimate
expectations as the dominant element of that standard.192 The tribunal, in line with
Schreuer (2005), asserts that the "expectations of foreign investors certainly include
the observation by the host State of such well-established fundamental standards as
good faith, due process, and nondiscrimination."193 In that respect, the standard of due
process contains, among other things, the obligation to grant a hearing before an
independent tribunal and the timely transmission of relevant information.194 Based on
the recent case law, Westcott (2007) developed a working definition of fair and
equitable treatment. This definition reflects the main elements also emphasised by
Schreuer (2005). The working definition by Westcott (2007) contains the regard for


187
      See Schreuer (2005), p. 357. For a recent overview see also Westcott (2007).
188
      UNCTAD (1999), p.9.
189
      Westcott (2007), p.409.
190
      Schreuer (2005), p.364. In the context of NAFTA, however, the official interpretation is that the
      fair and equitable standard simply reflects the international minimum standard.
191
      Schreuer (2005)p.373-374.
192
      Saluka v. Czech Republic, para 302.
193
      Ibid., para 303.
194
      Choudhury (2005), p.305.
52                                                          Economic and Legal Protection of FDI

due process rights and demands treatment that does not affect an investor's legitimate
expectations. States are also prohibited from acting in an arbitrary or discriminatory
manner. In addition, the state must respect the good faith principle. The FET standard
has now been applied to a diverse set of circumstances. This wide scope of application
and lack of precision may cause some criticism, but can also be considered a virtue.195
Dolzer and Schreuer (2008) argue that, as possible infringements of the investment are
impossible to anticipate, the FET standard "allows for independent and objective third
party determination of this type of behaviour on the basis of a flexible standard".196

3.3.5.5.2        Most-Favoured-Nation Treatment
A standard that bears similarity to fair and equitable treatment is the most-favoured-
nation (MFN) treatment. The MFN standard demands of the host state treatment of
investors that is not less favourable than the treatment it grants to nationals and
companies of any other country. Most BITs combine this standard with a national
treatment standard.197 National treatment calls for treatment that is not less favourable
than the treatment of nationals of the host state. The fundamental difference between
the MFN standard and the fair and equitable treatment standard is that the former is a
relative standard, depending on the treatment accorded to other investors, while the
latter is an absolute standard that “provides a fixed reference point.”198 Just like in the
case of the fair and equitable treatment standard, the scope of the MFN clause is
subject to debate and has been interpreted by arbitral tribunals in different ways.199
The MFN standard has been applied to both substantive and procedural aspects of
investment treaties.200 It appears that the application of the MFN standard to
substantive aspects is less controversial compared to its application to procedural
aspects.201 With regard to substantial rights, the MFN clause was, inter alia,
successfully combined with the fair and equitable treatment standard (e.g., MTD v.
Chile) and with the question of compensation for expropriation (Bayindir v. Pakistan).
As Dolzer and Schreuer (2008) hold, the "weight of authority clearly supports the view
that an MFN rule grants a claimant the right to benefit from substantive guarantees
contained in third treaties."202




195
      See especially chapter 5 on this issue.
196
      Dolzer and Schreuer (2008), p.148.
197
      Dolzer and Stevens (1995), p.65.
198
      Schreuer (2005), p.367.
199
      Teitelbaum (2005) and Hsu (2006) provide overviews.
200
      Chukwumerije (2007), p.598.
201
      Chukwumerije (2007), p.599.
202
      Dolzer and Schreuer (2008), p.191.
Economic and Legal Protection of FDI                                                                   53

With regard to procedural aspects, the main question relates to the access to dispute
resolution procedures that may be more favourable in some treaties than in others. An
arbitration ruling from the year 2000, Maffezini v. Spain, especially attracted a lot of
attention and caused some criticism.203 In the given case, the investor, originally
protected under the Argentina-Spain BIT, invoked the MFN clause to avoid the duty to
resort to the host state's domestic courts for 18 months before he could initiate
arbitration. This duty was part of the Argentina-Spain BIT, but another Spanish BIT,
the Chile-Spain BIT, contained no equivalent obligation. Relying on this Chile-Spain
BIT, the investor successfully argued that the duty to wait did not apply to him. A
number of tribunals, including Siemens v. Argentina and S.A. v. Argentina, followed
this line of reasoning. However, other tribunals have interpreted the MFN clause more
narrowly and have stressed its limited applicability to procedural rights.204 At this
point, the case law has not converged towards a reliable and generally accepted
interpretation of the MFN standard for procedural rights. What is clear is that
contracting parties can explicitly exclude certain areas from the application of the
MFN clause in the treaty.
As Dolzer and Schreuer (2008) emphasise, the straightforward application of the MFN
standard, though common in trade law, is more problematic in international investment
law. More precisely, investment treaties contain clauses specifically negotiated by two
parties. Consequently, the question arises whether the MFN clauses are supposed to
alter the results of these specific negotiations. Kurtz (2004) has also expressed
scepticism on this matter. The author's concern is mainly based on the wider scope of
operation of an MFN clause in investment law as compared to trade law, as in the
latter, a potentially large number of regulatory areas are concerned. In addition,
economic theory provides a rather uncontroversial welfare-enhancing case for tariff
reduction in international trade law. Yet, according to Kurtz (2004), in international
investment law, no similar unequivocal economic case can be made.

3.3.5.5.3         Additional Standards
BITs usually include a diverse number of additional standards. Among these are "full
protection and security", the "umbrella clause" and the "transfer of funds".205 The “full
protection and security” clause is a frequent standard that relates to the amount of
protection a host state must provide for the investment vis-à-vis third parties. As a


203
      ICSID Case No. ARB/97/7, award available under http://icsid.worldbank.org/ICSID/Index.jsp
      (September 1st, 2010).
204
      For instance, Plama v. Bulgaria or Telenor v. Hungary.
205
      Please note that this overview on standards in bilateral investment is not exhaustive. Treaties may
      also address the access to justice, armed conflicts or arbitrary measures. For a recent and complete
      overview, see Dolzer and Schreuer (2008).
54                                                          Economic and Legal Protection of FDI

practical matter, this clause requires the host state to exercise due diligence in the
protection of foreign investment.206
The so-called "umbrella clause" defines the relationship between investor-state
contracts and Bilateral Investment Treaties. Article 8 (2) of the German Model BIT
requires that each "Contracting State shall observe any other obligation it has assumed
with regard to investments in its territory by investors of the other Contracting State."
On the other ahdn, the 2004 US Model BIT does not contain an umbrella clause. In
sum, about 1000 BITs are estimated to incorporate an umbrella clause.207 The exact
wording of umbrella clauses differs to a certain extent among treaties concluded by
different countries.208 The umbrella clause did not receive much attention until the
interpretation of the clause by the tribunal of SGS v. Pakistan in 2003. Since then, two
different lines of jurisprudence regarding the interpretation of the clause have
emerged: a narrow and a wide interpretation. The narrow interpretation rests on the
classical distinction between violations of international and contractual breaches.209
Yet, as Kunoy (2006) remarks, "bearing in mind that this conceptual distinction is not
a peremptory norm, and in the event that the clause is clear and unambiguous, ICSID
tribunals are under the obligation to confer an effet utile to the umbrella clause."210 In
practice, the matter is unresolved and some but not all tribunals have been reluctant to
internationalise contractual obligations of host states vis-à-vis investors.
The "transfer of funds" relates to the import of funds to initiate the business and to the
repatriation of profits. Small countries may especially be concerned about their current
account balance and therefore impose restrictions on the in- and outflow of capital.
Although virtually all treaties deal with this matter and there is an inherent conflict of
interest between investors and states, the transfer of funds is rarely the subject of
arbitral proceedings. Dolzer and Schreuer (2008) attribute this to the high specificity
of regulation in most treaties.211

3.3.5.6           Expropriation and Compensation
Virtually all Bilateral Investment Treaties contain provisions concerning the
expropriation and nationalisation of property held by foreign investors. The treaties
formulate the conditions under which a taking is considered lawful. An expropriation
is accordingly considered legitimate if it serves a public purpose, is conducted in a


206
      Dolzer and Stevens (1995), p.61.
207
      See Dolzer and Schreuer (2008), p.153.
208
      See OECD (2006) for an overview on treaty language.
209
      Kunoy (2006), p.299.
210
      Kunoy (2006), p.299.
211
      Dolzer and Schreuer (2008), p.192.
Economic and Legal Protection of FDI                                                      55

non-discriminatory manner, is followed by prompt, adequate and effective
compensation and is in accordance with due process of the law.212 It is evident that as
to the question of compensation, when drafting treaty provisions, the contracting
parties have adhered to the Hull-Formula, although the specific determination of
compensation remains opaque.213 Even though outright nationalisations and
expropriations, as in the case of Bolivia, have received a lot of public attention, it is
mainly the problem of indirect expropriation with which arbitral tribunals must deal.

3.3.5.6.1          Indirect Expropriation
Indirect expropriation, also referred to as regulatory expropriation, describes a
situation where the host state deprives the owner of his property rights through
regulatory measures. The differentiation between a regulatory taking (implying
compensation for the asset-owner) and a regulation (implying no such compensation)
is a problem persistent also in the domestic sphere. The economic discussion focuses
primarily on the question whether takings justify the payment of compensation at all214
- with regard to the differentiation in practice, the standards applied obviously differ
from country to country.215 In international investment law, BITs take account of the
danger of indirect expropriation by extending the cover to measures “equivalent” or
“tantamount” to expropriation.216 Finding criteria that distinguish compensable
expropriation from justified, non-compensable regulation poses a great challenge to
arbitrators and international legal scholars. Evidently, the concept of indirect
expropriation is problematic as it affects the regulatory space of a country.
Various arbitral tribunals have approached the topic from different angles, some
focusing solely on the effects of the regulation, others placing emphasis on the purpose
of the measure and some taking the middle ground between these two approaches.217
Undoubtedly, the variance in the interpretation of BITs and other treaties regarding the
question of indirect expropriation has created some legal uncertainty. Kriebaum
(2007b) distinguishes between the Sole Effects doctrine, the Radical Police Powers
doctrine and the Moderate Police Powers doctrine. As the name implies, the Sole
Effects doctrine only considers one factor decisive as to whether a regulatory measure
constitutes expropriation: the effect of the measure on the investment. The motivation
and purpose of the measure is irrelevant. If the effect of the measure exceeds a
threshold, it is expropriation. The case of Metalclad v. Mexico is often cited as an


212
      See Dolzer and Stevens (1995), p.99-101, for different variations of this clause.
213
      On the dispute with regard to the Hull formula, see section 3.3.3.1.
214
      Miceli and Segerson (1999)
215
      See also section 5.4.3 on this issue.
216
      Dolzer and Stevens (1995), p.98 ff.
217
      See Fortier and Drymer (2004).
56                                                                Economic and Legal Protection of FDI

example for this reasoning. Alternatively, the Radical Police Powers doctrine does not
look at the effects on the investment but at the purpose of the measure. Put differently,
no expropriation will have occurred when the government measure was in the public
interest and non-discriminatory.218 The problem with this approach, as Kriebaum
(2007b) notes, is that it leads to a fragmentation of international investment law.
Compensation will be required in case of a direct expropriation (even when in the
public interest), but a regulatory measure with the same effect will not lead to a
compensation payment. An intermediate position is adopted by the Moderate Police
Powers doctrine. This doctrine is mainly based on the effect of the measure, but takes
considerations like the purpose of the measure and the existence of legitimate
expectations into account.219
New treaty practice appears to abandon the Sole Effects doctrine. The 2004 US Model
BIT emphasises three criteria that should be considered in identifying indirect
expropriation:220
"the economic impact of the government action, although the fact that an action or
series of actions by a Party has an adverse effect on the economic value of an
investment, standing alone, does not establish that an indirect expropriation has
occurred;
the extent to which the government action interferes with distinct, reasonable
investment-backed expectations; and
the character of the government action."
Further, the Model BIT clarifies: "Except in rare circumstances, non-discriminatory
regulatory actions by a Party that are designed and applied to protect legitimate public
welfare objectives, such as public health, safety, and the environment, do not
constitute indirect expropriations."221 As mentioned, this wording seems to be closer to
the Police Power doctrine than to the Sole Effects doctrine. However, as Kriebaum
(2008) notes, the legitimacy of measures will strongly depend on the interpretation of
the term "rare circumstances".222

3.3.5.6.2         Compensation
Once a tribunal has decided that a state measure has been of an expropriatory nature,
the investor is entitled to compensation. It has been mentioned before that this


218
      Kriebaum (2007b), p.726.
219
      Kriebaum (2007b), p.727.
220
      2004 US Model BIT, Annex B, Article 4. (a) (i-iii).
221
      2004 US Model BIT, Annex B, Article 4. (b).
222
      Kriebaum (2008), p.267. There are no arbitral awards based on this formulation yet.
Economic and Legal Protection of FDI                                                       57

compensation must usually be prompt, adequate and effective. With regard to the
amount of compensation, some treaties stipulate the "fair market value".223 Sornarajah
(2004) lists three common methods of valuation: the book value, the market value and
the discounted cash flow (DCF) method. The difference between the last two remains
unclear as the DCF should equal the market value. Consequently, Marboe (2006) lists
the DCF method as one possible method to determine the fair market value.
Alternative methods are the comparative method, multiples and the liquidation value.
The comparative method simply looks at the market value of comparable objects. The
method of multiples combines the DCF method and the comparative method by
multiplying key figures of the enterprise (e.g., EBIT) with certain factors (which are
derived from sales prices of a comparable enterprise). The liquidation value is self-
explanatory. Obviously, the appropriate method depends on the specific characteristic
of the case and the availability of the relevant data. It is also clear that tribunals cannot,
as has happened in the past, add the expenses incurred and the lost profits as this
would be double counting.224

3.3.5.7          Public Concerns
Public concerns regarding to the host country's right to regulate have played an
increasing role in recent BIT negotiations.225 Norway, for example, published a new
draft Model BIT that includes an article on the "Right to Regulate" (Article 12).226 The
relevant article reads: "Nothing in this Agreement shall be construed to prevent a Party
from adopting, maintaining or enforcing any measure otherwise consistent with this
Agreement that it considers appropriate to ensure that investment activity is
undertaken in a manner sensitive to health, safety or environmental concerns." More
specifically, international arbitration as the most important instrument to settle
disputes in investment matters has in the past been described as "ill-equipped" in
protecting the public interest.227 How public concerns can be incorporated in Bilateral
Investment Treaties will certainly be of importance in future BIT negotiations. For
example, environmental considerations have come up in the BIT negotiations between
China and Canada.228 It is yet unclear if this trend will continue to expand as public
concerns have been recognised already by arbitral tribunals based on the existing
standards of treatment.


223
      Kriebaum (2008), p.528.
224
      Wells (2003).
225
      UNCTAD (2008c), p.5. See also Karl (2008), p.229.
226
      Available under www.regjeringen.no/upload/NHD/Vedlegg/hoeringer/Utkast%20til%20modellav
      tale2.doc (September 1st, 2010).
227
      See, e.g., Gruner (2003), p.924. On dispute settlement, see the following section.
228
      UNCTAD (2008c), p.5.
58                                                     Economic and Legal Protection of FDI

3.3.5.8           The Settlement of Disputes
Disputes may arise between the contracting parties (inter-governmental disputes) and
between the investor and the host state (investor-state disputes or investment disputes).
Bilateral Investment Treaties usually address both issues.229 In both cases, treaties
generally provide for arbitration. While this is standard for disputes between
governments, it is remarkable in the context of international law that private parties
can initiate arbitration under an inter-governmental treaty. As the investors are actually
not contracting partners to the treaty, this type of arbitration has been termed
“arbitration without privity”.230 The main advantage for the investor here is that he
does not have to rely on the diplomatic protection of his home state to take legal
action. In the case of diplomatic protection, the home state of the investor pursues the
claim of its national against the host state in its own name. Although diplomatic
protection was the major remedy in traditional international investment law, its
usefulness is limited.231 The main problem is that the investor depends on the
willingness of his own government to pursue the claim.
Obviously, the use of domestic courts is the natural first step if a conflict cannot be
resolved through negotiations. As a matter of fact, states may make the exhaustion of
local remedies a condition for the consent to arbitration. As Dolzer and Schreuer
(2008) point out, this possibility is not used very frequently.232 In addition, BITs
sometimes require the investor to decide between the use of local remedies and
international arbitration. This requirement has been termed a "fork in the road
provision". As with diplomatic protection, the use of domestic courts is often not very
promising. A lack of impartiality and competence may be the two most important
problems in many jurisdictions. These shortcomings of domestic courts and diplomatic
protection have made international arbitration a valuable tool to foreign investors.
There are different types of arbitration rules that contracting states employ in their
BITs. Most common are references to arbitration under the institutional framework of
the International Centre for the Settlement of Investment Disputes (ICSID). However,
reference to UNCITRAL Rules or other forms of arbitration, as well as a combination
thereof, is also possible. Consequently, the wording of the relevant BIT determines
which type of arbitration an investor can access. The number of investor-state
arbitrations has seen a sharp increase in recent years. By the end of 2009, the total
number of known treaty arbitrations reached 357 cases.233 In 2009 alone, at least 32


229
      Dolzer and Stevens (1995), p.119 ff.
230
      Paulsson (1995).
231
      Dolzer and Schreuer (2008), p.211.
232
      Dolzer and Schreuer (2008), p.215.
233
      UNCTAD (2010), p.83.
Economic and Legal Protection of FDI                                                                59

new cases were filed.234 No universal public registry of claims exists; consequently,
the actual number is likely higher. As mentioned before, the drafters of a BIT have a
number of choices concerning the forum of arbitration. Most common is the reference
to ICSID. According to UNCTAD (2009b), out of the 317 known treaty-based
arbitrations by the end of 2008, 201 were handled under the institutional framework of
ICSID, while 83 were initiated under the UNCITRAL Rules, 17 under the rules of the
Stockholm Chamber of Commerce (SCC), 5 under the International Chamber of
Commerce (ICC), 5 using ad-hoc arbitration and the remaining cases under the Cairo
Regional Centre for International Commercial Arbitration. The Permanent Court of
Arbitration (PCA) in The Hague also administered one case.235 Evidently, ICSID is
and can be expected to remain the most frequently used arbitration facility in the
context of investment disputes. The following section will therefore describe the
arbitration process under ICSID in greater detail.236

3.3.5.8.1         ICSID
The International Centre for the Settlement of Investment Disputes was established in
1965 under the auspices of the World Bank through the so-called ICSID or
Washington Convention.237 So far, 143 states have signed and ratified the convention.
ICSID is one of the five international organisations that make up the World Bank
Group. ICSID is not an arbitral tribunal itself but instead a secretariat that assists in the
arbitral process and provides a set of rules for arbitration.238 These rules include, for
example, provisions about the appointment of the arbitrators, the procedure itself and
the costs of proceeding. The jurisdiction of the Centre is outlined in article 25 of the
ICSID Convention: "The jurisdiction of the Centre shall extend to any legal dispute
arising directly out of an investment, between a Contracting State (or any constituent
subdivision or agency of a Contracting State designated to the Centre by that State)
and a national of another Contracting State, which the parties to the dispute consent in
writing to submit to the Centre. When the parties have given their consent, no party
may withdraw its consent unilaterally."239



234
      UNCTAD (2010), p.83.
235
      The rules applied by the PCA are, in turn, the UNCITRAL Arbitration Rules. See Dolzer and
      Schreuer (2008), p.229. For the remaining four cases, UNCTAD (2009b) provided no information
      on the venue.
236
      For a description of the ICC and UNCITRAL Rules, see Collier and Lowe (1999), p.45ff. The
      different arbitration rules tend to have very similar procedural rules.
237
      Blackaby, Paulsson et al. (2004), p.1ff.
238
      Arbitration is not the only form of dispute settlement that is supported by ICSID. The ICSID
      convention also provides guidelines for conciliation.
239
      ICSID Convention, Article 25, available under http://www.worldbank.org/icsid/highlights/04-10-
      06.htm. Even if the investor is not from a contracting state or the host state has not signed the
60                                                             Economic and Legal Protection of FDI

In the context of Bilateral Investment Treaties, it is important to note that ICSID
accepts arbitrations that arise from indirect consent, such as in a BIT or a multilateral
investment treaty.240 In other words, when an investor files a claim against a state, the
ratification of the ICSID Convention is taken as a written consent. Consequently, the
state cannot frustrate the arbitration process unilaterally. Once a valid request for
arbitration has been filed, an arbitration tribunal is constituted. The choice of
arbitrators and size of the panel is supposed to be made cooperatively. However,
ICSID provides default rules in cases where the parties cannot find a consensus on
these issues. Usually, an arbitral tribunal consists of three members. The actual
proceedings of the arbitration are regulated in Articles 41 to 47 of the convention and
in the Arbitration Rules 13 to 38. The arbitration process consists of written and oral
procedures. According to Article 45, failure of a party to appear before the tribunal
will not frustrate the proceedings.241
As to which legal grounds the arbitral tribunal must use to decide the case, article 42
(1) of the ICSID Convention states: "The Tribunal shall decide a dispute in accordance
with such rules of law as may be agreed by the parties. In the absence of such
agreement, the Tribunal shall apply the law of the Contracting State party to the
dispute (including its rules on the conflict of laws) and such rules of the international
law as may be applicable." Obviously, the underlying investment agreement is the
starting point of the legal analysis of the investment dispute at hand. BITs usually refer
to the provisions in the treaty itself, the laws of the contracting party involved in the
conflict, specific agreements concluded in relation to the relevant investment and
applicable principles of international law.242 Regarding interpretation, the general rules
of the Vienna Convention on the Law of Treaties, especially article 31, apply.243 While
BITs often do not specifically prescribe how the different sources of law are to be
combined, arbitral tribunals have given priority to the application of international law
as compared to the domestic law of the contracting state.244 Within 120 days of the
closure of proceedings, the tribunal must render an award. As to the costs of
proceedings, if the parties do not agree otherwise, the tribunal decides in its final
award on the allocation of costs.




      Washington Convention, arbitration may still be undertaken under the auspices of ICSID by using
      the ICSID Additional Facility.
240
      See Blackaby, Paulsson et al. (2004), chapter 3.
241
      Collier and Lowe (1999), p.70, call this “perhaps the most important provision”.
242
      Redfern, Hunter et al. (2004), paragraph 11-19.
243
      Dolzer and Stevens (1995), p.15. See also section 3.3.3.2.
244
      Comeaux and Kinsella (1997), p.206 and Redfern, Hunter et al. (2004), paragraph 11-19.
Economic and Legal Protection of FDI                                                       61

3.3.5.8.2        Remedies
ICSID awards are binding and are not subject to any appeal or remedy other than those
provided for in the convention itself.245 In other words, awards under ICSID cannot be
reviewed by national courts. The ICSID convention offers only three post-award
remedies: interpretation, revision and annulment. The first two appear to be of limited
practical importance: as of February 2010, ICSID has, according to the cases listed on
its homepage, only received five requests for interpretation (Marvin Roy Feldman
Karpa v. United Mexican States in 2003, Wena Hotels Limited v. Arab Republic of
Egypt in 2004, Archer Daniels Midland Company and Tate & Lyle Ingredients
Americas, Inc. v. United Mexican States in 2008, Tanzania Electric Supply Company
Limited v. Independent Power Tanzania Limited in 2008 and Corn Products
International, Inc. v. United Mexican States in 2009) and only three requests of
revision (American Manufacturing & Trading, Inc. v. Democratic Republic of the
Congo in 1999, Víctor Pey Casado and President Allende Foundation v. Republic of
Chile and Siemens A.G. v. Argentine Republic, both in 2008).246 One of the requests of
revision was settled before the tribunal made a decision; the other two cases are still
pending. The third available remedy, annulment of the award, is only possible on very
limited grounds, for instance if the tribunal was not properly constituted or manifestly
exceeded its powers.247 However, there have been a number of annulment cases in the
past.248 It is important to note that ICSID awards are not subject to annulment (or
review) by domestic courts. Upon request of a party, the chairman of ICSID's
Administrative Council will appoint an ad hoc committee consisting of three persons.
The enforcement of the award may be deferred until the annulment procedure is
completed. The main point is that the ad hoc committee may evaluate only the
legitimacy of the process of the original arbitration and does not review the substance
of the award. If the legitimacy of the original arbitration process was frustrated, the
case should be submitted to a new tribunal. Article 52 (1) of the ICSID Convention
lists the five exhaustive criteria for annulment: (a) that the Tribunal was not properly
constituted; (b) that the Tribunal manifestly exceeded its powers; (c) that there was
corruption on the part of a member of the Tribunal; (d) that there was a serious
departure from a fundamental rule of procedure; or (e) that the award failed to state the
reasons on which it was based.




245
      ICSID Convention, article 53.
246
      See http://icsid.worldbank.org/ICSID/Index.jsp (September 1st, 2010).
247
      ICSID Convention, article 52.
248
      Out of the 205 concluded cases (through settlement or award) listed on the ICSID website
      (November 2nd, 2010), 25 cases triggered an annulment procedure.
62                                                            Economic and Legal Protection of FDI

In non-ICSID arbitrations, an award may be challenged through national courts. The
procedure depends on whether the defendant is a party to the United Nations
Convention on the Recognition and Enforcement of Foreign Arbitral Awards, the so-
called New York Convention. The New York Convention was adopted in 1958 and
deals, as the name implies, with the recognition and enforcement of international
arbitral awards. A total of 138 states have signed the convention. The New York
Convention allows domestic courts to review an arbitral award on certain limited
grounds and is less severe than the ICSID Convention as it allows, for instance, the
review of an award if the "recognition or enforcement of the award would be contrary
to the public policy of that country".249 If the country is not party to the New York
Convention, domestic arbitration laws apply. Many of these laws are modelled on the
UNCITRAL Model Law.250 The UNCITRAL Model Law is, in turn, based on Article
V of the New York Convention.

3.3.5.8.3         Enforcement and Execution of Arbitral Awards
The ICSID Convention contains an automatic recognition and enforcement
mechanism. An award must be enforced by each contracting state “as if it were a final
judgment of a court in that state.”251 Consequently, investors can enforce awards not
only in the host state, but in any state that is a signatory to the ICSID Convention.
Nevertheless, enforcement of the award and execution are distinct procedures.
Domestic courts have refused to execute arbitral awards if the assets in question were
considered as being protected by sovereign immunity.252 Sovereign immunity is
generally the only argument that can prevent the execution of international arbitral
awards. A waiver of sovereign immunity included in the investment agreement could
solve this problem, but states will be reluctant to agree to such a waiver.253 As relative
recent research shows, attempts to enforce ICSID awards in domestic courts have been
rare so far. As of 2006, there are only four known cases.254 This implies a high
compliance rate. However, the number of cases may obviously increase due to the
rising number of arbitrations under ICSID. As with remedies, the execution and
enforcement of non-ICSID awards depend on the regulations of the New York
Convention (if applicable) and on national laws.



249
      United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards, also
      called the New York Convention, Article V 2. (b).
250
      Dolzer and Schreuer (2008), p.278.
251
      ICSID Convention, article 54 (1).
252
      Blackaby, Paulsson et al. (2004), p.107ff. Which assets actually fall under the category of
      sovereign immunity depends on national legislation and varies between countries.
253
      Comeaux and Kinsella (1997), p.209.
254
      Baldwin, Kantor et al. (2006), p.5.
Economic and Legal Protection of FDI                                                          63

As mentioned earlier, the New York Convention leaves more room for appeal than the
ICSID convention. Nevertheless, in principle, the New York Convention also requires
the contracting state to recognise arbitral awards as binding and enforce them.255

3.3.5.8.4        Costs
The costs associated with arbitration proceedings can be substantial. This does not
concern so much the fees charged by arbitration institutions like ICSID.256 Instead, the
cost of legal representation is by far the largest cost for the parties and may amount to
millions of USD.257 The cost sharing rule depends on the applicable arbitration rules.
Article 61 (2) of the ICSID Convention places the decision on the distribution of costs
in the hands of the tribunal. The UNCITRAL Arbitration Rules, however, specifically
place the burden of the costs on the losing party by default.258 Arbitral practice in this
respect is not uniform.259 Tribunals have in some cases required the parties to carry
their own expenses, in other cases the losing party must cover the expenses of the
winning party as well.

3.3.5.8.5        Empirical Aspects of Arbitration
Scholars have recently begun to collect data on the different aspects of international
investment treaty arbitration. In addition, the importance of empiricism in international
law and international investment law has been emphasised.260 Franck (2007) has
compiled a number of figures on investment treaty arbitration to evaluate the merits of
frequent claims on investment dispute resolution. A total of 102 awards from 82
separate cases formed part of the examination. Of these awards, 52 finally resolved the
case's treaty claims. The analysis showed that investors who filed claims were
predominantly from developed countries, in most cases from the United States. The
situation looks different with regard to government respondents. Based on World Bank
classifications, in 18% of the cases the respondent government was a high-income
country, in 45% an upper middle income country, in 28% a lower middle income
country and in 8.5% a low income country. This qualifies earlier assertions that only
very few cases are brought against non-developing countries. As Franck (2007) states,
this "suggests that 'middle income' countries, particularly those with a higher income,



255
      United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards,
      Article III.
256
      The schedule of fees for ICSID (as of January 1, 2008) stipulates administrative charges of
      20.000,00 USD per case and 3.000,00 USD payments to arbitrators per day.
257
      Dolzer and Schreuer (2008), p.276.
258
      Article 40, UNCITRAL Arbitration Rules.
259
      Dolzer and Schreuer (2008), p.276.
260
      See Franck (2008).
64                                                                Economic and Legal Protection of FDI

were at the greatest risk of arbitration."261 Only three cases concern countries usually
classified as Least Developed Countries (“LDCs”). Karl (2008) uses slightly different
figures and asserts that only about 20% of all known investment disputes were filed
against a developed country.262 It is obvious that much depends on how the line
between developed and developing countries is drawn.
An interesting point concerns whether investors win more often than governments.
The data collected by Franck (2007) are rather surprising in this respect. Out of the 52
cases that finally resolved a treaty claim, 30 cases were decided in favour of the
government. Put differently, in 57.7%, the government had to pay nothing. In 38.5%,
the investor was awarded damages, and two awards embodied settlement agreements.
The claim that tribunals are biased towards investors accordingly could not be verified
by Franck. But if investors do win, do they recover large amounts? The average
amount in the 21 cases where investors actually were awarded damages was 10.4
million USD. In most cases, investors received less than 5 million USD. To quote
Franck: "When investors did win, they did not win big."263 Franck also finds a striking
difference between the amounts claimed and the amounts awarded. The high claims
often put forward by investors may have fostered the perception that governments are
facing high financial risks through investment treaties (in one case, Generation
Ukraine v. Ukraine, the investor claimed 9.4 billion USD264). Concerning the total
number of investment arbitrations, Karl (2008) emphasises that given the number of
BITs (approximately 2.700) and the number of potential claimants (78.000
transnational companies could in theory invoke arbitration procedures), the number of
arbitrations is still very small.265

3.3.5.9           BITs and Customary International Law
With almost 2700 BITs signed by more than 170 countries, the question arises whether
BITs are evidence of customary international law, that is, if they are “applicable even
when a given situation or controversy is not explicitly governed by a treaty”.266 There
is no unified answer to this question. Some writers reject the notion that BITs
represent CIL by stating that BITs merely constitute a lex specialis between the



261
      Franck (2007), p.32. This does, of course, not control for the magnitude of investment flows and
      the number of investment treaties. Also, there were two huge outliers: Argentina faced fifteen cases
      alone and Mexico faced nine.
262
      Karl (2008), p.233.
263
      Franck (2007), p.61.
264
      Franck (2007), p.57.
265
      Karl (2008), p.230.
266
      Lowenfeld (2002), p.486.
Economic and Legal Protection of FDI                                                     65

contracting states.267 To quote Sornarajah (2004): “It would be wrong to subscribe to
the thesis that the treaties stabilised customary international law. If there was a definite
conviction as to the existence of customary law in the area, there would have been
little need for such frenetic treaty-making activity on investment protection.”268 Other
writers argue that common principles have nevertheless passed into customary
international law.269 These principles include, for instance, that the host state is under
the obligation to treat foreign investors fairly or that disputes between foreign
investors and host states should be subject to impartial adjudication or arbitration.270
However, it must be noted that “there are particular provisions of BITs that, although
widely used, are applicable only between the parties.”271

3.3.5.10       Summary
Studying the realm of international investment law reveals that this is a dynamic area.
Many specific questions are the subject of controversy. The historical debate on the
appropriate amount of compensation for expropriations has calmed with a changing
attitude of most developing countries towards FDI. Meanwhile, the growing number of
BITs has brought up a whole set of more specific questions and controversial issues
related to the rightful application and interpretation of certain treaty clauses. This
includes for instance the fair and equitable treatment standard, the most-favoured-
nation clause and the question of when a regulation is of an expropriating nature. This
vagueness in some areas of international investment law does not imply that
international investment law is useless and ineffective. From the investor’s point of
view, Bilateral Investment Treaties offer a number of advantages: the investor does not
have to rely on the diplomatic protection of his government; the level of protection in
BITs is usually higher and better specified than the rules in customary international
law; the investor can expect that his case will be handled by an unbiased panel of
international experts; Bilateral Investment Treaties are backed up by multilateral
treaties on the recognition of arbitral awards, etc. It must be noted that BITs have
received considerable attention in the past years. Some concerns by commentators
relating to the (possibly negative) impact of dispute resolution on developing countries
have not (yet) been verified empirically. The large amount of attention is also
surprising in another respect: BITs are obviously international law and international
law is often rather perceived as (too) weak. The logical next question is therefore to



267
    See, e.g., Kishoiyian (1994) or Guzman (1998), p.686f.
268
    Sornarajah (2004), p.213.
269
    Hindelang (2004), Lowenfeld (2002), p.486ff.
270
    Hindelang (2004), p.799.
271
    Hindelang (2004), p.809.
66                                                   Economic and Legal Protection of FDI

ask if BITs actually make a difference and if so, why. The next section will deal with
this question.
4                 The Economics of BITs

Bilateral Investment Treaties are the cornerstone of international investment law.
While empirical research indicates that BITs can increase foreign investment to the
signatory state, the economic analysis of law has so far only barely been applied to
BITs. The decisive question in this context is why BITs function given that
international law is weak and lacks a supranational authority with coercive power.
Further, how do litigation costs, damages, reputational concerns etc. impact the
economic functioning of BITs? Which countries can actually utilise BITs to attract
investment? This chapter will attempt to answer these and similar questions and
proceeds as follows: after discussing the perceived weakness of international law
(section 4.1), the empirical literature on the impact of BITs on FDI will be reviewed
(section 4.2). The burgeoning research approach of international law and economics
will be introduced (section 4.3) and the main findings of this method will be applied to
international investment law (section 4.4). Special focus will be given to the
commitment and signalling properties of BITs. These properties will be checked
against the existing empirical research on the topic (section 4.5). Section 4.6 discusses
the main results and concludes.

4.1               The (Perceived) Weakness of International Law
Because international law lacks the coercive enforcement that is present on the
domestic level, international law may be perceived as weak and ineffective. One might
conclude that states only comply with international law when it coincides with their
general policy goals. Put differently, international law as such does not affect state
behaviour. This idea can be exemplified using a simple game-theoretic example:272
two countries have agreed on a treaty (in the example of Guzman (2002), the treaty
concerns a ban on satellite-based weapons) and must make a compliance decision. We
are considering a one-shot game with the following payoffs:

                                                Country 2

                                                Comply       Violate

                                   Comply       5,5          2,6
                   Country 1
                                   Violate      6,2          3,3



272
      This example is taken from Guzman (2002), p.1841 ff.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_4,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
68                                                                                The Economics of BITs

Using the satellite-based weapon example, both countries have a preference for
obtaining the weapon while the other country does not have it. In addition, both
countries not having the weapon is preferred to both having the weapon. In other
words, both countries prefer a situation where the other country complies while
violating itself. The well-known equilibrium of this prisoner's dilemma is
(violate/violate). While this situation might be called the "bad state of nature", payoffs
can alternatively be modelled as follows (the good state of nature):

                                                   Country 2

                                                   Comply            Violate

                                     Comply        10,10             6,8
                   Country 1
                                     Violate       8,6               4,4



The satellite system might be unreliable and a waste of resources. Still, each country
prefers that the other does not even have an unreliable weapon system. In this case, for
both countries, compliance is the dominant strategy. If the payoffs are frequently as
depicted in the second case, states will usually comply with their international legal
commitments. This behaviour satisfies the famous observation that "almost all nations
observe almost all principles of international law and almost all of their obligations
almost all of the time".273 At the same time, international law itself did not make a
difference. The decisions of the states were guided by the underlying fundamental
payoffs. However, the example shows immediately that the theory that international
law does not make a difference is based on the assumption that the law does not affect
the payoffs of states in any meaningful way.274
In this spirit, it could reasonably be argued that BITs - as being part of international
law - do not affect state behaviour and can, consequently, also not help attract
additional FDI. Put differently, the (potential) costs that are generated by a BIT may
not influence the decision of states or governments.275 The following sections will


273
      This is a quote from Louis Henkin cited by Guzman (2002), FN 79.
274
      If, however, the existence of a treaty between the parties reduces the payoff of game 1 in the case
      of violation by 2, compliance would also be the dominant strategy.
275
      A second argument against the efficacy of BITs could be that MNEs do not consider the additional
      legal protection to be important. As Sornarajah (2004) notes on whether investment treaties
      facilitate investment flows: "This is an untested hypothesis. […] Stability and other factors have a
      greater influence on investment flows than do investment treaties." Sornarajah (2004), p.215, FN
      32. This might be true, but does not refute the hypotheses that BITs influence FDI flows at the
      margin. See section 4.2.
The Economics of BITs                                                                  69

analyse the economic case for (or against) the efficacy of BITs using a commitment
and a signalling framework. However, before evaluating the functioning of BITs from
a theoretical point of view, it is worthwhile to review the empirical literature in this
respect and thereby to shed some light on the impact of BITs on FDI.

4.2               The Effect of BITs on FDI
Several empirical studies have tried to evaluate whether BITs fulfil their aim of
attracting FDI.276 These studies differ in their methodology, complexity, sample
selection and, most notably and rather unfortunately, in their results. While some
studies affirm a robust link between the conclusion of investment agreements and FDI,
other studies find only weak or no correlation at all. Put differently, a number of
studies could not confirm a positive effect of BITs.

4.2.1             Empirical Studies on the Effect of BITs on FDI
The first empirical analysis concerning the impact of BITs on foreign direct
investment was carried out by UNCTAD in 1998.277 The authors analyse time-series
data for 200 BITs (of the approximately 1.300 BITs in existence at that time) from the
period 1971 to 1994. More specifically, they test the bilateral investment flows
between two countries that have signed a BIT with each other for different time frames
using T-statistics. The study finds a weak effect of BITs on FDI. A consistent time lag
between the signature of the BIT and the effect on FDI cannot clearly be defined, but
is most likely two years. The authors also conduct a cross-sectional regression for 133
countries for the year 1995. Dependent variables are FDI flows and stocks,
independent variables include the number of BITs, GDP, GDP growth, population, etc.
The overall result is that BITs play - compared to market size (GDP and population) -
only a weak, secondary role in attracting FDI.
Quite similar are the results by Hallward-Driemeier (2003): the author analyses dyadic
(bilateral pairs) investment flows from OECD to developing countries in the years
1980 to 2000 using panel data in a fixed-effects model. The independent variable of
interest is the existence of a BIT between the country pair (using a dummy variable),
while other independent variables control for the size of the relevant states, macro-
economic stability, trade openness, education, etc. She detects only slight evidence
that BITs have induced additional investment. At first, the coefficients on the BIT
variable are negative and insignificant, only after five years appears to be an extremely
weak positive correlation. Furthermore, BITs are found to be more of a complement


276
      Sauvant and Sachs (2009) provide a recent overview of the relevant literature.
277
      UNCTAD (1998).
70                                                                               The Economics of BITs

than a substitute for domestic institutions. In other words, countries that actually profit
from investment agreements already possess reasonably strong domestic institutions.
This rather pessimistic view on the efficacy of BITs is further supported by Aisbett
(2007). The author tests whether BITs have stimulated the flows of FDI from OECD
countries to 28 low- and middle-income countries between 1980 and 2005 using panel
data with fixed-effects (including year effects and country-pair fixed effects). While
the BIT indicator in the study is positive and significant, Aisbett (2007) attributes this
not to a positive impact of BITs on FDI, but to a problem of endogeneity mainly based
on reverse causality and omitted variable bias. Put differently, according to a simple
model developed by Aisbett (2007), the benefit of signing a BIT depends on the
magnitude of the existing stock of FDI. The underlying reason is that in the proposed
model, the signature of BITs reduces the maximum tax rate the host state can impose
on new and existing investment.278 Controlling for this relationship in the econometric
model reveals that the positive correlation between BITs and FDI might be driven by
endogeneity.
Rose-Ackerman and Tobin (2003) use panel data from 1980 to 2000 in a fixed effects
as well as random effects specification. The dependent variable are FDI inflows as a
percentage of World-FDI. In testing the impact of cumulative BITs on FDI, the
authors control for GDP, political risk, population, growth, natural resources and the
interaction of BITs with risk. BITs turn out to have a positive impact on FDI, though
not a statistically significant one. Statistical significance (in the expected directions)
can be shown for GDP, political risk and population. The authors further find that –
surprisingly - for high risk countries, a higher number of BITs lowers FDI inflows.
Only for low risk countries does the flow of FDI increase. Specifically, studying the
effects of US-BITs, the authors find no significant effects on FDI.
These rather pessimistic findings on the impact of BITs on FDI are counterbalanced by
a number of studies that confirm a robust and positive relationship between the two
variables. Salacuse and Sullivan (2005) run cross-country regression of US-BITs,
OECD-BITs and other BITs on aggregate FDI inflows to more than 100 countries
controlling for GDP, GDP per capita, inflation, real effective exchange rate,
population and rule of law. In addition, the authors analyse US FDI inflows to 31
countries over a ten year period using panel data in a fixed-effects model. With both
techniques, the authors find that US-BITs are correlated with higher FDI inflows. The
same strong result cannot be established for other BITs. Thus, the authors conclude
that "1. A U.S. BIT is more likely than not to exert a strong and positive role in
promoting U.S. investment. 2. A U.S. BIT is more likely than not to exert a strong and


278
      Which is a rather arbitrary and therefore in our view not a convincing assumption.
The Economics of BITs                                                               71

positive role in promoting overall investment. 3. A U.S. BIT is likely to exert more of
an impact than other OECD BITs in promoting overall investment."279
Two studies limit their attention to the analysis of BITs and FDI in specific regions.
Grosse and Trevino (2005) investigate the impact of institutional factors on FDI flows
in Central and Eastern Europe (CEE countries). The authors use a data set for 13 CEE
countries from 1990 to 1999 and employ a random effects model. They find that the
presence of Bilateral Investment Treaties was positively and significantly related to
inward FDI. Similar results are obtained by Gallagher and Birch (2006) who use fixed
effects regressions with panel data on 24 countries in Latin America for the time
period 1980 to 2003. According to this study, the total number of BITs has a positive
and independent effect on FDI flows. No independent effect could be found for US-
BITs on FDI inflows from the United States. Remarkably, the authors discover
regional differences. While BITs could be shown to augment FDI in South America,
this was not true for BITs in Mesoamerica.
Swenson (2005) runs separate negative binomial regressions for the time frames 1990
to 1994 and 1995 to 1999 to account for the hypothesis that investments may be
characterised by lags or leads. The author controls in her regression for previous
investment flows, country risk and regional differences (through region dummies). In
addition, she differentiates between US-BITs and non US-BITs. The data set includes
80 countries for the first time frame (1990-1994) and 84 countries for the second time
frame (1995-1999). The author finds that signing a BIT with the US was correlated
with a much larger stimulus to investment flows than were BITs with other countries.
Further, BITs fostered investment in the second half of the 1990s, but not in the first
half.
Robust results indicating a positive of impact of BITs on FDI have been published by
Egger and Pfaffermayr (2004). The authors use a panel of OECD data on outward FDI
and include fixed country-pair effects as well as fixed time effects. The study covers
the time period from 1982 to 1997 and includes 19 home and 54 host countries.
Control variables include market size (in GDP), ratio of skilled to unskilled labour,
distance as well as EU and NAFTA membership. The authors differentiate between
the signing and the implementation (ratification) of BITs. Signing a BIT exerts only a
low, in most cases statistically insignificant, positive influence on FDI. However, the
ratification (and thus implementation) of a BIT provokes a significant rise of
investment. The study estimates a relative increase of FDI of up to 30%, with 15% as a
lower bound.



279
      Salacuse and Sullivan (2005), p.111.
72                                                                               The Economics of BITs

The affirmative results by Egger and Pfaffermayr (2004) are supported by Neumayer
and Spess (2005). The latter use both random effects and fixed effects estimations on
panel data for 91 countries (120 in some specifications) covering the years 1984 to
2001 (1970 to 2001 in some specifications). To account for the difference in
importance among BITs, the main explanatory variable is not simply the number of
cumulative BITs with OECD countries, but each BIT is weighted by the share of
outward FDI flows the OECD country accounts for relative to total world outward FDI
flow. Control variables include GDP, population, growth, trade agreements, different
measures of institutional quality (country risk) etc. Robust to sample size and model
specification, the authors find very strong evidence for a significant rise in FDI caused
by the conclusion of Bilateral Investment Treaties.
In a recent working paper, Busse, Königer et al. (2008) provide similar evidence. The
authors use a gravity type model and find that BITs indeed increase investment flows
to developing countries. It is important to note that they choose a dyadic approach, that
is, the authors look at bilateral flows and thus show that BITs promote FDI even
absent signalling effects. In addition, the (preferred) dependent variable “is the share
of FDI attracted by a specific host country in total FDI flows from the source country
under consideration to all developing host countries included in our sample.”280
Consequently, the authors measure the attractiveness of a particular developing
country relatively to other developing countries and not absolute investment flows.
Finally, the authors find some evidence that BITs may serve as a substitute for weak
domestic institutions, but not for unilateral liberalisation regarding FDI.

4.2.2             Discussion
The myriad of different results on the empirical relevance of BITs with regard to FDI
is rather unsettling. Can any conclusion on the efficacy of BITs be drawn at all? Some
of the studies have significant shortcomings that should be taken in account before
forming an opinion. UNCTAD (1998) is certainly not up-to-date with regard to its
methodology.281 The study by Hallward-Driemeier (2003) may well be criticised by its
restricted sample size (31 countries). In addition, the study has been criticised for its
focus on dyadic relationships only, thus ignoring possible signalling effects of BITs.282


280
      Busse, Königer et al. (2008), p.11.
281
      Neumayer and Spess (2005), going as far as calling the UNCTAD study “garbage can” modeling
      (FN5). However, the authors cannot explain the differing results as compared to the study of Tobin
      and Rose-Ackerman, as both studies use high sample sizes and sophisticated methodology.
282
      Neumayer and Spess (2005), p. 1572. If BITs have a signalling effect, this would imply that
      signing a BIT with country A would also increase FDI inflows from country B, C, etc. A dyadic
      estimation can obviously not account for this. The question on the signalling properties of BITs is
      explored in greater detail in section 4.4.4.
The Economics of BITs                                                                73

Egger and Pfaffermayr (2004), Busse, Königer et al. (2008) as well as Neumayer and
Spess (2005) all have larger sample sizes and are up-to-date in their methodology. All
three studies find a positive effect of BITs on FDI, and Busse, Königer et al. (2008)
find one even with a dyadic approach. With regard to sample size and methodology,
the same can be said about Rose-Ackerman and Tobin (2003) - finding, on the other
hand, no robust link between the two variables. Possibly, the affirmative studies may
suffer from endogeneity problems as Aisbett (2007) suggests. Omitted variable bias
could be problematic in the sense that BIT conclusion and higher FDI flows may both
be caused by changes in domestic policy. In addition, endogeneity problems may be
caused by reverse causality. This would imply that higher FDI flows lead to more
BITs in later periods. However, Aisbett (2007) fails to provide a convincing theory on
this possibility. It could be theorised that MNEs lobby for stronger protection once
they have made their investment. However, if the influence of MNEs on governments
was strong enough to lobby them into concluding a BIT – why would MNEs need a
BIT in the first place? In addition, Busse, Königer et al. (2008) control for this
potential endogeneity problem.
Clearly, the matter on the impact of BITs on FDI flows is not closed yet. At this point,
the more sophisticated studies confirm a positive link between BITs and FDI. Given
the vast number of potential factors that influence investment decisions by MNEs, this
result is indeed remarkable. In addition, no positive impact of BITs on the "investment
climate" would imply that governments all over the world have wasted and still waste
time and resources to negotiate and conclude Bilateral Investment Treaties. This
massive irrationality is possible, yet unlikely. In sum, the hypothesis of a positive
impact of BITs on FDI on the margin appears to be reasonable.

4.3               Law and Economics of International Law
In section 4.1, it was noted that BITs, as part of international law, could be perceived
as ineffective. The notion that international law is irrelevant has mainly been put
forward by the neorealist school of international relations.283 Obviously, most scholars
of international law and international relations (IR) do not share the view that
international law is irrelevant. At the same time, disunity persists as to why states
create and follow rules and norms of international law. The approach used in this book
is based on the economic analysis of law and could be termed the law and economics
of international law. This approach, based on rational choice theory, has recently
especially been put forward by Guzman (2008a). The following section briefly
discusses the key assumptions of this approach and the main differences to other


283
      See the discussion of this approach in Guzman (2002).
74                                                                   The Economics of BITs

prominent theories of compliance with international law. Only by understanding
compliance with (and breach of) BITs, can one understand why states conclude BITs
in the first place.

4.3.1            Theories of International Law
Why do states obey international law? Legal scholars have given multiple answers to
this question.284 Traditional legal theories include the managerial model, consent-based
theories, legitimacy theory and the transnational legal process. The managerial school
was developed by Chayes and Chayes (1995) and includes the idea that states find it
costly to violate international law because it has the status of law. In other words,
states have a propensity to comply with international law. Non-compliance is mainly
rooted in informational ambiguity and limited capacity and can thus be 'managed'
through increased transparency, fewer ambiguities in rules and greater capacity. The
consent-based theory is, as the name implies, based on the notion of consent. A state’s
consent to a treaty creates a legal obligation that leads to compliance. This logic
obviously lacks explanatory power. Another theory advanced by Thomas Franck has
been termed legitimacy theory. States obey rules that were created in "accordance with
the right process"285 which, in turn, depends on four factors: determinacy, symbolic
validation, coherence, and adherence. If these factors are present, states perceive the
respective rule as legitimate and are inclined to comply with it. A theory called
transnational legal process has been advanced by Koh (1996). Through the multi-
layered interaction of state and non-state actors in different fora "patterns of behaviour
and norms emerge and are internalised, leading to their incorporation within the
domestic legal institutions of states and, in turn, compliance".286

4.3.2            Rational Choice Approach
This book is based methodologically on the economic analysis of law or, put
differently, law and economics. Consequently, the underlying assumptions of the book
are based on rational choice theory and in accordance with Guzman (2008a): "States
are assumed to be rational, self-interested, and able to identify and pursue their
interests. Those interests are a function of state preferences, which are assumed to be
exogenous and fixed. States do not concern themselves with the welfare of other states
but instead seek to maximise their own gains or payoffs. States, therefore, have no
innate preference for complying with international law, they are unaffected by the
'legitimacy' of a rule of law (Franck 1995), past consent to a rule is insufficient to


284
      The following overview is based on Guzman (2002)
285
      See Franck (1988), p.706.
286
      See Guzman (2002), p.1835.
The Economics of BITs                                                                                75

ensure compliance, and there is no assumption that decision-makers have internalised
a norm of compliance with international law (Koh 1997)."287

4.3.3             Methodological Individualism
Guzman (2002) deviates from the standard assumptions of economic theory in one
important respect: he treats states as unitary actors and thus departs in a certain respect
from the assumption of methodological individualism. Put differently, the domestic
political process is treated as a “black box”. This mainly concerns the principal-agent
problem between governments and citizens, the interaction among policy-makers as
well as the influence of certain interest groups on government behaviour. Law and
economics of international law as such is not incompatible with methodological
individualism. For example, in the realm of international trade law, Sykes (1991)
successfully explains the GATT escape clause as the result of domestic interest group
politics. Also, when an international treaty, for example a human rights treaty, is used
by governments to signal properties to its own citizens, certain elements of the treaty
cannot be analysed without recourse to the domestic political landscape. Nevertheless,
it can often be useful to leave the "black box" state closed, namely where incorporating
the domestic principal-agent problem and interest group dynamics does not yield
additional explanatory power.
To give an example: when analyzing the way governments structure dispute settlement
mechanisms in treaties, it can be very useful to assume government preferences as
fixed since it is not the focus of analysis whether these preferences were formed as a
process of maximizing social welfare or as the result of successful lobbying. When
analyzing the specific rules within the same treaty, however, incorporating the interest
and incentives of governments can deliver fruitful additional insights.
The same logic applies to BITs. Of course, the distinction between mechanisms and
contents is to some extent artificial. Governments may choose different content when
they know that only weak mechanisms will be employed in the treaty. Keeping that in
mind, this book will nevertheless assume government preferences as exogenous when
the formation of these preferences is unlikely to yield additional insights (especially in
chapter 4). Yet, this assumption will be abandoned where certain phenomena related to
BITs need further elaboration based on the principal-agent problem between
governments and domestic constituents (especially in chapter 5).




287
      Guzman (2008a), p.17. Parentheses and quotation marks in original. Please note that this book will
      deviate from the assumption of state having exogenous preferences in chapter 5.
76                                                                           The Economics of BITs

4.3.4             The Three R's: Reputation, Reciprocity and Retaliation
Based on the preceding assumptions, why do states create international law and
comply with it? At first glance, the assumptions – and here especially the emphasis on
the self-interest of states – appear to be hostile to cooperation. According to Guzman
(2008a), compliance with international law can be sustained through the three R's:
reputation, reciprocity and retaliation.288
Reputation refers to the ability and willingness of a state to honour its international
commitments. Consequently, reputation can only be valuable when there is
asymmetric information regarding these characteristics of a state. The value of
reputation is rooted in a state's ability to conclude agreements. Put differently, the
greater a state's reputation, the more credibly it can commit itself to a certain course of
action, which makes cooperation in the future more feasible. Thus, reputation can only
be valuable in a situation of repeated interactions. Consider the following simple
example: imagine two states have already concluded a treaty. Now, they must decide
whether to comply with it. Three situations are possible: states comply with the treaty
because they would have chosen the respective action even in absence of the treaty.
Second, states may find the action associated with non-compliance most attractive and
decide not to comply. Third, states may decide to comply because of the existence of
the treaty. As this paper attempts to understand why international law matters, the third
case is of most interest here. When would a state choose to comply when the actual
payoffs of non-compliance are higher? This will only be the case when there are
additional costs associated with the breach of international law. These costs accrue
when the treaty partner and/or third parties observe the breach and adjust their
assessment on the state's (government's) internal discount rate or its payoffs, that is, its
willingness and/or ability to forgo short-term profits for future gains of cooperation.
States that are perceived, for example, to have a low discount rate will find it harder to
establish cooperation in the future. The advantage of this form of punishment is that it
occurs rather automatically as the punishment is not costly to the punisher.
Why, then, do we need the presence of asymmetric information? As the theory of
repeated games implies, cooperation is possible even in PD situations with complete
information when games are repeated infinitely and discount rates are low. In that
case, international law would indeed be useless. Every party would already know what
the other party will do. Treaties and other instruments of international law would only
serve to define a focal point and would not affect compliance decisions. It seems clear
that this might be the case sometimes, but is not in line with reality on a more general


288
      Guzman (2008a), p.33ff. The following section is a short summary of Guzman (2008a), especially
      p. 33-49.
The Economics of BITs                                                                   77

basis. The assumption of the existence of asymmetric information is therefore
reasonable.
As mentioned, international law can only affect compliance when it changes the
payoffs of parties. The mechanism of reputation is one way to affect these costs.
Another mechanism is the concept of reciprocity. Put simply, if a partner to a treaty
decides to breach, he may face the costs that the other partner also breaches. This
mechanism is related to reputation, but yet a distinct concept. First, reputation applies
to all third parties that form expectations on the characteristics of a breaching party,
whereas reciprocity refers only to the relevant treaty. The termination of reciprocity
may therefore very well be a stronger compliance factor than reputation, especially in
bilateral treaties. On the other hand, the termination of reciprocity might be a weak
factor compared to reputation in other circumstances. Consider the example of a
human rights treaty between country A and country B that commits both states to
refrain from torture with regard to their citizens. If country A violates this treaty,
country B is not likely to react by torturing its own citizens. Reciprocity in this setting
will not help to induce compliance, while reputation might play a crucial rule. More
generally, reciprocity does not enhance compliance if the threat of termination of
reciprocity is not credible or if the termination of reciprocity does not impose costs on
the original violator.
The last of the enforcement mechanisms of the three R's is retaliation. Retaliation or
retaliatory sanctions describe actions by an aggrieved party to punish the violator of
international law. These actions are usually costly, which differentiates them from
reputational sanctions and the withdrawal of reciprocity (which are, in an immediate
sense, beneficial to the aggrieved party). These costs, however, constitute the main
problem with retaliation. States may refrain from retaliatory sanctions to avoid these
costs, making the threat of retaliation non-credible in the first place. In that case, why
would a state retaliate at all? Again, reputational concerns play a role, yet this time not
with regard to the violating party but with regard to the retaliating party. In line with
the theory advanced here, the aggrieved party will only retaliate when the benefits
outweigh the costs. This might very well be the case when the retaliating state wants to
build or maintain a reputation of dealing harshly with states that violate their treaties.
As a result, other states will come to expect that violating a treaty with the retaliating
state is costly – and thus make cooperation for the retaliating state cheaper in the
future. Of course, the efficacy of retaliation depends not only on the payoffs of the
(potential) retaliator (which determines the credibility of the threat), but also on the
question of whether retaliatory sanctions are available. For instance, if imports from
country A to country B are virtually non-existent, country B will not have the option to
impose retaliatory tariffs on imports from country A.
78                                                                  The Economics of BITs

In summary, according to the rational choice approach, states will honour their
international legal commitments only when benefits outweigh costs. They do not feel
any form of internal compliance pull. Consequently, international law can only make a
difference when it creates costs that would otherwise be absent. These costs may
consist of reputational losses, the withdrawal of reciprocity or retaliatory sanctions by
the treaty partners. The costs and their consequences with regard to BITs will be
discussed in the following chapter.

4.4          The Functioning of BITs
Where do BITs fit in the outlined theory of international law? Clearly, if BITs are
supposed to promote FDI and protect investor's rights, a breach must impose costs on
the violating party. The framework of the analysis is again set by the trust game as
introduced in section 3.1. The Pareto-optimal solution would be that the host country
credibly commits to play accommodate whenever the MNE invests. In domestic law,
this could easily be done through a contract. However, the inherent problem with
international law is the lack of a supranational enforcement agency. Simply speaking,
making expropriate more costly compared to accommodate will, at the margin,
increase the attractiveness of investment. The starting point of the analysis of BITs is
the assumption that BITs must affect the payoffs of the involved parties in some way.
The second important feature of BITs is that due to the lack of supranational
enforcement, their impact can ex-post only be mainly informational. Yet, the possible
enforcement of arbitration awards in third countries may impose costs that go beyond
the costs usually suffered by violators of international law according to the three R's.
In addition, not only the ex-post costs of breach play role with regard to the
functioning of BITs, but also the costs of concluding BITs. Also, the merits of BITs
might not be found in making certain actions more costly, but in the benefits of
additional information. Yet, we have seen from the discussion on reputation (section
3.2.2.2.) that in the case of FDI, asymmetric information may be at the core of the
problem. The BIT may then rather be understood as signal. The following sections will
discuss each of these possibilities in turn, starting with a more general description of
the costs of BITs.

4.4.1        The Costs of BITs
Many different costs are associated with BITs which slightly complicates the analysis.
Directly and evidently, the negotiation and conclusion of a BIT creates costs. A second
class of costs results from the breach of BITs. To define these costs, it is first
necessary to lay out what constitutes breach. These costs are mainly reputational, but
also include non-reputational factors. The costs related to BITs have also been referred
to as sovereignty costs. Elkins, Guzman et al. (2006) describe these sovereignty costs
The Economics of BITs                                                                             79

as the "costs any government pays when it negotiates, ratifies, and complies with an
investment treaty. We would include here the political costs of assembling a coalition
in support of foreign investors' rights, as well as the costs associated with giving up a
broad range of policy instruments relevant to domestic social or developmental
purposes (taxation, regulation, performance requirements, property seizure, and
currency and capital restrictions). Most striking are the sovereignty costs associated
with the delegation of adjudicative authority: virtually any legal change or rule that
affects foreign investors is potentially subject to review by a foreign tribunal."289 The
following sections will focus on the costs of concluding and breaching the BIT, but not
on the compliance costs. The reason is that this book is primarily interested in the
question of when a BIT can deter breach and the compliance costs are implicitly
already part of the cooperation payoff (CH) of the host country.290

4.4.1.1           Concluding BITs
As with any international agreement, concluding BITs generates a number of direct
costs to a country. First, states must invest resources into negotiations. However, given
the existence of Model BITs used by many capital-exporting states and the brevity of
BITs, these negotiations are probably less complex than the negotiations of other
international agreements. In addition, UNCTAD provides developing countries with
technical and advisory assistance and has also hosted BIT negotiations in the past.
Second, the BIT must be ratified according to the procedures of the respective state.
Obviously, this process strongly differs from country to country. Yet, the overall direct
costs appear to be manageable even for countries with limited resources.

4.4.1.2           Breaching BITs: The Three R's Revisited
Section 4.3 outlined a general theory of international law based on the rational choice
approach. Reputation, reciprocity and retaliation were identified as the main
enforcement mechanisms. These mechanisms increase the costs of a violation of
international law. The following sections will evaluate the three R's in relation to BITs.

4.4.1.2.1         Reciprocity and Retaliation
Reciprocity implies that a party to a treaty will terminate behaviour that is in
accordance with the treaty as a reaction to the breach by the other party. It has been
pointed out that reciprocity does not enhance compliance if the threat of termination of
reciprocity is not credible or if the termination of reciprocity does not impose costs on

289
      Elkins, Guzman et al. (2006), p.825.
290
      Therefore, in the model that will be introduced below, adding these costs would only complicate
      the analysis without yielding additional insights.
80                                                                    The Economics of BITs

the original violator. Unfortunately, this is exactly the case with BITs. It appears
improbable that, for example, Germany would decide to refrain from the protection of
investment from a particular country when that country is in breach of a BIT with
Germany. In addition, investment flows are very often asymmetric, so that the
revocation of reciprocal investment protection would not affect a substantial amount of
investment flows and stocks.
Retaliation, or punishment, of a violator of a BIT might in theory indeed have a strong
impact on the violator. Potential retaliatory actions can take numerous forms,
including economic sanctions or even, in the extreme case, military actions. All
retaliatory measures have in common that they are costly to the party that retaliates. It
appears that retaliation only plays a minor role in investment law. It may be that the
stakes of individual investment are not high enough to induce governments to act in a
retaliatory manner. In addition, the costs of retaliation seem to be rather high. For
example, economic sanctions like tariffs as a reaction to expropriation would trigger
proceedings under the WTO and are thus not an attractive option. Military actions to
solve economic conflicts are certainly not extinct, yet it seems that the stakes in the
case of FDI are not high enough to prompt military sanctions. In addition, please also
note that retaliation (as well as reciprocity) requires an action from the home
government of the investor. The home government is unlikely to be a perfect agent of
the investor and may thus not consider it important to acquire a reputation of acting
tough on BIT violators. It is exactly the independence from the home government that
makes BITs a valuable tool for investors. In sum, reciprocity and retaliation are
unlikely to play a major role in the realm of international investment law.
Consequently, it is reputation that seems to matter the most in disputes regarding
international investment. In this respect, it is important to differentiate between
reputation vis-à-vis other states and reputation vis-à-vis investors.

4.4.1.2.2     Reputation vis-à-vis Other States
A breach of an international agreement will induce other states to update their beliefs
with regard to the breaching state. This is also true for BITs. As pointed out earlier,
reputational sanctions require the existence of asymmetric information. Consequently,
the costs of a breach when a BIT is in force are limited to the reputational loss the
country faces in the eyes of other countries concerning its ability and willingness to
keep (other) international commitments. These commitments may consist of other
BITs, but also of commitments unrelated to the investment sector. The reputational
losses therefore differ among governments. Governments that already have a very low
reputation or that are not interested in international cooperation in the future will face
only small reputational sanctions and vice versa. The reputational losses vis-à-vis other
states are limited in an important respect relating to the disposition of third countries to
The Economics of BITs                                                                                  81

make inferences concerning the willingness and ability of the country to honour
international commitments based on the breach of a BIT. With regard to other BITs, it
seems reasonable to assume a certain connection. If a state repeatedly fails to comply
with the treaty clauses, other states may very well infer that the respective government
is inclined to forgo long-term for short-term profits or that domestic checks and
balances against unjustified interferences with the property rights of investors are
weak. Combined with the fact that BIT arbitrations are lengthy and costly, other
governments might therefore rightfully conclude that investment agreements with that
country are not a sensible investment protection strategy. Yet, third countries might
not make any inferences at all from the breach of BITs relating to host countries’
reliability in other important realms of international law, like environmental treaties or
non-proliferation treaties.291 To summarise, costs in the form of reputational sanctions
with regard to third countries certainly must be taken into account; nonetheless, these
costs are limited to the extent a breach of a BIT actually impacts the beliefs of other
states.

4.1.1.2.3          Reputation vis-à-vis Investors
To see how a BIT can be costly in terms of reputation vis-à-vis investors, it seems
reasonable to recapitulate the repeated trust game from section 3.2.2.2.1. It has been
demonstrated that repetition can play a role in fostering cooperation and preventing
expropriation (section 3.2). Yet, such an equilibrium depends on a trigger mechanism.
As Kreps (1990) points out: "The point is that when one player cannot observe directly
that the agreement is carried out, and when this player can only rely on noisy, indirect
observations, the problem of finding self-enforcing arrangements is vastly more
complicated".292
Remember that the payoff for the host country when accommodating was CH, while it
was WH when expropriating (with CH < WH). Given that the host country discounts
future profits with G , the condition for cooperation was:

 CH
     t WH                                                                          (4.1)
1 G

              CH
œ G t 1                                                                           (4.2)
              WH




291
      Guzman (2008a) refers to this as the "compartmentalizing of reputation", p.100ff. This
      compartmentalizing may not only relate to issue areas, but also to regimes and dyadic relationships.
292
      Kreps (1990), p.105.
82                                                                   The Economics of BITs

The left-hand side of equation 4.1 was the discounted payoff in the case of infinite
cooperation; the right-hand side was the one-shot payoff for the uncooperative
behaviour. This calculation was based on a grim trigger strategy where the MNEs stop
investing once they observe an expropriation. Now imagine that the investor can
observe a deviation of the HC only with a certain likelihood, say s. With probability
(1-s), the deviation goes unpunished, that is, the MNEs keep playing the cooperative
strategy:

 CH                   C
     t sWH  (1  s ) H                                                (4.3)
1 G                 1 G

                CH
œ G t 1                                                               (4.4)
           sWH  (1  s)CH


The right-hand side of this equation is increasing in s. This indicates that increasing
the probability of observation relaxes the requirement on the minimal discount factor
that ensures cooperation (remember: a higher discount factor means that the HC
discounts future profits less strongly). Consequently, if the BIT increases observability
of defective behaviour, breach of the BIT will induce a reputational sanction in the
sense of losing future profits. It is obvious that in reality governments care about their
reputation vis-à-vis investors. Section 3.2.2.2.2 described a situation where
governments adapt their behaviour to build up or protect a reputation for being
investor-friendly. Irrespective of the existence of a BIT, investor will reduce,
discontinue or hold back their investment when the government acts in a manner that
derogates the investment. The question is whether there is an additional reputational
cost vis-à-vis the investors that is generated through investment agreements. Two
mechanisms seem possible: investors may learn about an arbitration proceeding
through the media or through postings on relevant websites, such as the ICSID website
itself, and draw inferences from that. Secondly, investors may consider the award as a
reliable and neutral source of information and will adjust their behaviour accordingly.
Is it reasonable to assume that the breach of a BIT usually does not convey any
information to the investment community additional to the act of expropriation (or any
other derogation of the assets) itself? To put it another way: is the reputational harm
that may induce potential investors not to invest mainly caused by the expropriation
itself or does the breach of international law add to it? One should think that investors
expend money and resources to inform themselves about past behaviour of potential
host governments before making investment decisions. In addition, the ability of
international investment arbitration to do so is questionable, as arbitration is often kept
private and as such is not a reliable source of information. Claims might be rejected on
The Economics of BITs                                                                             83

formal grounds although the act of the government would actually be considered as
undue by investors. On the other hand, investors will use all publicly available
information including information generated through an investment dispute.
Consequently, BITs may indeed trigger an additional reputational loss vis-à-vis MNEs,
albeit with limited impact. At this point, it is worth noting that BIT arbitrations are
certainly not specifically designed to enhance the information flow on host state
behaviour to (potential) investors.

4.4.1.3           Breaching BITs: Non-Reputational Costs
Section 3.4 pointed out that the enforcement mechanisms in international law are
reputation, reciprocity and retaliation. Although part of international law, Bilateral
Investment Treaties appear to possess an additional form of enforcement. More
specifically, in the case of BITs, enforcement is sometimes also possible through
domestic courts. The ICSID Convention, for example, contains an automatic
recognition and enforcement mechanism. An award must be enforced by each
contracting state “as if it were a final judgment of a court in that state.”293
Consequently, investors can enforce awards not only in the host state, but also try to
seize assets in any state that is a signatory to the ICSID Convention. It was already
pointed out that the enforcement of awards in third countries is a very tedious and
costly strategy for the investor and often not very promising.294 Obviously, non-
reputational sanctions can only play a role when the legal prerequisites for an
application of the ICSID or, respectively, the New York convention apply. This, in
turn, can only be the case when the relevant country is a member of one of these
conventions. Also, trying to enforce an arbitral award only makes sense when the
country holds certain assets in a member state which can be seized (and sold) to satisfy
the claim. Especially poor countries will not possess substantial assets in third
countries that could readily be seized. In any case, a government that is being sued will
have to expend substantial resources to protect its foreign assets. Consequently, even if
the incentive of the MNE to initiate arbitration is moderate, the high costs induced by
arbitration might affect the behaviour of states (especially if deterrence is understood
as the product of the likelyhood of being sued and the potential costs). To sum up,
although the possibility of enforcing arbitral awards through the ICSID or the New
York Convention might not always exist, in cases where it does, this possibility might
have a certain deterrence effect. Potentially, it is also the threat of domestic



293
      ICSID Convention, article 54 (1). See section 3.3.3.8.3.
294
      For a recent and longsome case from Germany, see Sedelmayer against the Russian Federation, Az.
      OLG Köln 22 U 98/07.
84                                                                    The Economics of BITs

enforcement that induces governments to comply with arbitration awards, thus keeping
the number of known enforcement procedures so low.295
The possibility of enforcing arbitral awards through domestic courts might contradict
the theory that reputation, reciprocity and retaliation are the only enforcement
mechanisms in international law. However, it should be noted that this possibility is in
itself based on an international agreement which can be denounced by the member
state. Article 71 of the ICSID convention reads: "Any Contracting State may denounce
this Convention by written notice to the depositary of this Convention. The
denunciation shall take effect six months after receipt of such notice." As article 72
further points out, the denunciation does not affect cases initiated before such notice
was received by the depositary. Consequently, in the short run, the investor has under
certain conditions a legal tool at his disposal that goes beyond the usual enforcement
mechanism in international law. Yet, in the middle and long run, reciprocity,
retaliation and reputation are indeed the only available enforcement mechanisms in
international law.

4.4.1.4            Summary
Bilateral Investment Treaties are associated with numerous costs. It seems reasonable
to assume that the actual direct costs from negotiating and concluding a BIT are rather
low. When it comes to breach, (termination of) reciprocity and retaliatory sanctions do
not appear to be forceful factors. Non-reputational costs like the seizure of assets
through domestic courts might under certain circumstances play a certain role. Yet, the
most important factors seem to be reputational. Here, one must differentiate between
reputation vis-à-vis investors and reputation vis-à-vis other states. While arbitration
proceedings are not explicitly designed to help third parties to update their beliefs, it is
still plausible to assume that investors as well as governments will use all available
information, including that generated through arbitration proceedings. However, the
potential reputational losses are not uniform among countries but may vary greatly. To
understand the functioning of BITs, the following sections will illuminate the
influence of this diversity of costs on the strategic interaction between host
governments and investors.

4.4.2              Commitment and Signalling
The main question regarding the functioning of BITs is whether investment
agreements rather solve a commitment problem or a problem of adverse selection



295
      Unfortunately, there is no way to evaluate this claim.
The Economics of BITs                                                                                    85

through signalling. Both possibilities have been suggested in the literature,296 albeit
without providing a systematic analysis of this question yet. This is an important
distinction as it relates not only to the question of which countries may find a BIT to
be useful, but also to what potential investors can learn from BITs. With regard to
signalling, there can be different theories relating to what is actually being signalled.
Also, understanding the functioning of BITs might give some guidance to the effects
of arbitration awards on the incentives of states. However, one might suggest that BITs
go beyond signalling and commitment and are rather related to the question of
allowing for flexibility in long-term contracts or the question of development aid. Yet,
as will be argued, these two points can be subsumed under the theory that BITs are a
commitment device or a signal.
The analysis regarding commitment on one side and signalling on the other side rests
on slightly different methodological grounds. The time inconsistency problem that
causes what has been termed "hold-up" is rather related to transaction cost theory as
pioneered by Williamson (1990). More specifically, the focus is "the comparative
efficacy with which alternative governance structures manage transactions during
contract execution."297 The foundation of this theory is the idea that contracts (here:
treaties) that are unsupported by credible commitments will suffer from opportunism.
The signalling theory, on the other hand, goes more into the direction of agency
theory. Only one type of state is prone to opportunistic behaviour – the problem then is
that the MNE lacks the information to differentiate between states. Taking this
approach does not contradict transaction cost theory – in the end, the lack of
information might be interpreted as the result of transaction costs in terms of
information gathering.

4.4.3              Commitment
The starting point of the commitment model developed here is the assumption that
both parties have symmetric information. As the game structure introduced earlier
exemplifies, once the MNE has sunk its resources, it is prone to opportunistic
behaviour by the host state. Knowing that under these circumstances the MNE will not
invest, both parties have ex-ante an incentive to write a contract, or, in our case, a
treaty.298 As the host state has the incentive to violate the contract as soon the
investment is made anyway, an enforcement mechanism must be put in place that

296
      See, e.g., on commitment Elkins, Guzman et al. (2006), p.823, on signalling see also Grosse and
      Trevino (2005), p.129, or Neumayer and Spess (2005), p.1571.
297
      Williamson (1990), p.67.
298
      For the analysis at this moment, it does not matter that the contract (treaty) is not really concluded
      between the state and the MNE, but between two states. The important point is that the MNE is
      protected under the treaty.
86                                                                      The Economics of BITs

places costs on the non-cooperative behaviour. We refer to the commitment case when
these costs are big enough to change the incentives for the host state.

4.4.3.1       Repetition Revisited
A very simple first step to analyse the problem could be to refer to the repeated
baseline model from section 3.2.2.2.1. The condition for the HC to choose a strategy
of accommodating the investment was established as:

         CH
G t 1                                                                          (4.5)
         WH


Cooperation emerges if the MNE values future payoffs highly (high ) and/or has high
payoffs from accommodation as compared to expropriation. Modelling BITs simply as
an addition of costs in the case of playing the defective strategy (here denoted, as in
the case of hands-tying, as D), the condition for cooperation changes to:

          CH
G t 1                                                                          (4.6)
         WH  D


Obviously, increasing the costs of defection at the margin makes investment c.p. more
likely. Unfortunately, this approach cannot shed light on the incentives of MNEs to
actually initiate arbitration. In addition, it ignores the peculiarities of international law,
including the fact that states always have an exit option. The following section will
therefore describe the commitment game as a one-shot game played between the
(representative) investor (MNE) and the host country.

4.4.3.2       The Commitment Game
Section 4.4.1 has identified a number of costs related to BITs. Some costs were ex-
ante in the sense that they were directly related to the conclusion of the BIT. For the
moment, we consider these costs as sunk. Put differently, we are looking at a situation
where the government has already concluded a BIT, thus not taking the costs of
conclusion into account anymore. This approach can be justified by the fact that we
are trying to understand under what conditions the investment decision of the MNE
will be influenced by the BIT. Only when this is the case, the host country has an
incentive to conclude a BIT in the first place and carry the costs of negotiation and
ratification associated with BITs. The analysis is based on the trust game as introduced
in section 3.1.1. Consequently, the game starts as before: the MNE has to decide
between Invest and Don't Invest. If the MNE decides not to invest, the game ends –
The Economics of BITs                                                                                 87

payoffs in this case are normalised to zero. If the MNE invests, the host country can
accommodate or expropriate. In the case of accommodation, the MNE will receive its
return on investment CM, the HC will receive taxes, spillover benefits etc., denoted as
CH. Until this point, the game is identical to the baseline game of time inconsistency.
However, if a BIT is in place, the investor can bring a claim before an arbitral tribunal
(Arbitration) or not (Keep Still). Theoretically, he can do that even if there was no
expropriation in the first place. Keeping still will yield the prior expropriation payoff
(LM and WH with LM <0) or, respectively, the cooperation payoff (CH and WH with CH
< WH). If the MNE starts arbitration, a tribunal will decide on the case. The decision of
the tribunal is not explicitly modelled, but can be understood as a move of nature (we
are not interested in strategic tribunals); the tribunal finds with probability y a violation
of the law in the case of the defective behaviour and with probability (1-y) no violation
(even if there was defective behaviour).299 If the tribunal finds a violation, then the
host country must pay damages to the MNE amounting to D. In addition, there will be
reputational sanctions based on the fact that the HC has breached its obligations under
international law, denoted as R. Please note that reputational costs include especially
the loss of future cooperation. Therefore, the repeated game as presented in the
preceding section is implicitly reflected in these costs. Arbitration is assumed to be
costly to both the MNE and the host country. These transaction costs include legal
fees, costs of using arbitration facilities and paying the arbitrators etc., and are denoted
by T. For simplicity, it is assumed that the costs are equally high for both parties,
consequently TM=TH=T. The game can be presented in extensive form as such:




299
      y can therefore be understood as the ability of the tribunal to make the "right" decision or as the
      probability of error. In that sense, the probability of being convicted although being not guilty
      amounts to (1-y). A more sophisticated approach at this stage would be to differentiate between
      type 1 and type 2 errors of courts. This would then imply an inherent trade-off between the error
      types. However, for our purposes, this approach would only complicate the analysis without
      yielding further insights.
88                                                                           The Economics of BITs


                                                         (CM +(1-y)D-T, CH –(1-y)(R+D)-T)
     Don't Invest      (0,0)               Arbitrate


                                          M
     M
                       Accommodate       Keep Still

         Invest                                          (CM, CH)
                         H
                                                         (LM +yD-T, WH -y(R+D)-T)
                                           Arbitrate
                         Expropriate

                                          M


                                           Keep Still
                                                         (LM, WH)



Figure 4.1: The Commitment Game



Notation
CM       =          Cooperation gain for the MNE (profit)
CH       =          Cooperation gain for the host country (taxes, spillover etc.)
LM       =          Loss for the MNE in the case of expropriation (Loser's payoff)
WH       =          Value of the assets to the host country if expropriated (Winner's payoff)
y        =          Probability of error (ability of the tribunal to detect violation)
D        =          Damages/Compensation
T        =          Litigation costs
R        =          Reputational loss


This is a game of complete information; the possible equilibria can therefore be
derived straightforwardly using backwards induction, starting at the last decision node
of the game.
The Economics of BITs                                                                  89

4.4.3.2.1          The Tribunal's Decision
The decision of the tribunal is assumed to be non-strategic and can be understood as a
move of nature (although not explicitly modelled). Put differently, the tribunals are not
assumed to follow an agenda of their own. It could be argued that this assumption
deviates too strongly from reality. Yet, arbitrators are chosen jointly by the parties. An
arbitrator that consistently fails to act neutrally and makes biased decision will soon be
lacking mandates. The tribunal here, in line with Guzman (2008b), is defined "as a
disinterested institution to which the parties have delegated some authority and that
produces a statement about the facts of a case and opines on how those facts relate to
relevant legal rules".300 The assumption is that the tribunal rules in favour of the MNE
only with a probability y if there was uncooperative behaviour in the first place. The
idea is that in some cases the uncooperative behaviour of the host country cannot be
verified with regard to third parties. In addition, the value y depends on the clarity of
the underlying legal rule, which is limited in our case as the standards in BITs are
often subject to wide interpretation. Lastly, y is also determined by the competence of
the relevant tribunal. Despite these qualifications, it should be expected that y>1-y,
which is true for any value of y that is above 0.5, indicating that the tribunal has some
ability to differentiate between the defective and the cooperative behaviour. The ruling
of the tribunal will not only define the amount of compensation D, but also impose the
reputational sanction R on the host country, as it is then verified by a third, expert
party that the country has violated the terms of the BIT.301 This reputational sanction
relates not only to investors, but also to other states who might refrain from
cooperation (also in other areas of international law) in the future.

4.4.3.2.2          The Arbitration Decision
At the final stage of the game, the MNE will have to decide between initiating
arbitration or keeping still. This stage of the game will only be reached if the MNE
has invested in the first round. The decision of the MNE to file for arbitration depends
on the behaviour of the host state in the preceding round. Therefore, we have to
differentiate between two cases:




300
      Guzman (2008b), p.185.
301
      Unless the proceedings were not held in private. See chapter 7 on this issue.
90                                                                     The Economics of BITs

                  Case 1: HC accommodated in Case 2: HC expropriated in
                  the prior round            the prior round

 The MNE will C + (1-y)D-T > C                         LM + yD-T > LM
                   M           M
 initiate
 arbitration if: rearranges to                         rearranges to
                 T < (1-y)D                            T < yD

The MNE will initiate arbitration if the cost of arbitration is smaller than the expected
compensation payment of the arbitration award. The two conditions are
straightforward. The decision to initiate arbitration depends on the possibility to prove
the wrongdoing to the tribunal and the expected damage payments. The MNE does not
care about the reputational losses of the HC. Equally, the payoff from the investment
(LM or CM)is sunk at that point. It is simple to see that a low value of y may give room
for opportunistic behaviour on the part of the MNE who might be tempted to initiate
arbitration even if there was no defective HC-behaviour in the first place. If, however,
the ability of the tribunal to differentiate between defective and cooperative behaviour
is reasonably high, this case is rather unlikely.
Depending on the payoffs of the MNE, we can identify three cases:
(1) the MNE will always initiate arbitration: (1-y)D > T,
(2) the MNE will only initiate arbitration if the HC expropriates (but not otherwise):
yD > T > (1-y)D
(3) the MNE will never initiate arbitration: T > yD.
A theoretical fourth case is that the MNE would only initiate arbitration when the HC
accommodates but not otherwise. This can be ruled out as y > (1-y) which implies that
T < (1-y)D and T > yD cannot both be true.

4.4.3.2.3    The Expropriation Decision
Again, reaching this stage of the game presupposes that the MNE has invested in the
first round. The decision of the host country then depends on the subsequent choice of
the MNE. The HC will anticipate the behaviour of the MNE in the subsequent stage
and adapt his strategy accordingly. The three possible cases are summarised in the
following table (the assumption y>(1-y) holds):
The Economics of BITs                                                                   91

                                    Case 2: MNE will
               Case 1: MNE will                              Case 3: MNE will
                                    only            initiate
               always      initiate                          never       initiate
                                    arbitration if the HC
               arbitration                                   arbitration
                                    expropriates

               CH -(1-y)(R+D)-T <
 HC     will WH -y(R+D)-T   CH < WH -y(R+D)-T
 expropriate rearranges to  rearranges to                         CH < WH
 if
             WH - CH > (2y- WH - CH > y(R+D)+T
             1)(R+D)



Starting with case 3: obviously, if the MNE will never initiate arbitrations, the game is
practically similar to the basic trust game and the HC will, in a one-shot setting,
expropriate as long as CH < WH. With regard to case 1, we can see that the HC will
expropriate if the expropriation value is high compared to the cooperation payoff. The
assumption y > (1-y) ensures that the higher the sanction (R+D), the smaller the
incentive to expropriate. Equally, with growing ability of the tribunal to detect the
uncooperative behaviour, the incentive to expropriate naturally declines. However, it
has been mentioned in the previous section that the prerequisite for this case, namely
(1-y)D > T, is unlikely to be true. Clearly, the interesting case for our purpose is case 2
where T takes an intermediate value. Only if the cost of the BIT can counterbalance
the difference in payoffs between defection and cooperation (WH - CH ) will the BIT be
able to serve as a commitment device. This will be true for high values of y, high
reputational costs (R) or damages (D) and also high litigation costs (T). Noteworthy, at
the same time, high litigation costs will diminish the incentive of the MNE to seek
arbitration in the first place.

4.4.3.2.4     The Investment Decision
The investment decision marks the first stage of the discussed model and relates to the
crucial question: when can the BIT induce the MNE to invest? As before, the
subsequent choices of the players shape the strategy of the MNE in the first round
(including the choices of the MNE itself). Remember that the payoff in the case where
no investment takes place is normalised to zero. The choice to invest may be followed
by the HC to expropriate or to accommodate, which may, in turn, be followed by the
binary choice to seek arbitration or not. This leaves four cases to look at:
92                                                                  The Economics of BITs

              MNE keeps still and               MNE seeks arbitration and

              Case 1: HC Case 2: HC Case 3: HC Case 4: HC
              accommodates expropriates accommodates expropriates

 MNE
              CM>0             LM>0             CM+(1-y)D-T>0       LM+yD-T>0
 invests if



By assumption, we have LM < 0, which rules out case 2. Consequently, the MNE will
either invest if the host country does not expropriate or if the remedy D compensates
for the losses despite litigation costs.

4.4.3.3       Equilibria
First, one point must be remembered that has been neglected in the preceding sections:
the fact that concluding the BIT itself is associated with costs. These costs consist
mainly of negotiations and the ratification process. Only if the existence of the BIT has
a positive impact on the investment decision of the MNE will a host country ever
bother to incur these costs. Therefore, we are interested in the equilibria where the
MNE actually invests. Four of these investment equilibria (also exemplified by the
final nods in figure 4.1) exist and without making any assumptions on the values of the
variables, each of these final nodes is a possible equilibrium of the game. Yet, if the
assumptions of the baseline game are used, the equilibrium relating to the strategy set
(invest, expropriate, keep still) can be eliminated. If the remedies of the BIT are not
used, the HC will not have to adapt its behaviour. Therefore, the situation is identical
to the baseline game – consequently, this cannot be an equilibrium of the game as long
as LM < 0. Put differently, if the MNE would prefer keeping still despite expropriation,
the HC would anticipate this and expropriate. This, in turn, would be anticipated by
the MNE leading to no investment in the first place. If this is the case and the
conclusion of a BIT is associated with costs, the HC never has an incentive to
conclude a BIT. The remaining three equilibria cannot be ruled out in a similar
fashion. However, two of these equilibria include the use of arbitration and require
strong assumptions. First, the equilibrium relating to (invest, accommodate, arbitrate)
presupposes that the MNE has a high chance of success even if there was no
uncooperative behaviour by the HC in the first place (T< (1-y)D). This is certainly
possible and may sometimes be the case in reality, but certainly does not reflect the
intentions when concluding a BIT. Second, the equilibrium based on (invest,
expropriate, arbitrate) presumes that the MNE would invest knowing that there will
be expropriation, which will be the case if LM+yD-T>0. Especially as LM<0, this
The Economics of BITs                                                                  93

equilibrium appears removed from reality and has no explanatory power with regard to
the functioning of BITs.
For the functioning of BITs, the interesting equilibrium is therefore the equilibrium
where the MNE invests and the HC accommodates because of the BIT (invest,
accommodate, keep still). Yet, for the BIT to make a difference, the MNE must have
the incentive to initiate arbitration when the investment was expropriated, that is T <
yD. This will only be the case when arbitration costs are low and/or the compensation
and the probability of winning the case are high. It is important to note that the
reputational losses to the HC do not play a role in the decision-making of the MNE.
However, these costs play a role with regard to the decision of the host country to
expropriate or not. The BIT will only induce cooperation if the arbitration threat is
credible and the expected costs of arbitration offset the gain of non-cooperation
compared to cooperation (which can be understood as the opportunity costs of non-
cooperation), namely WH - CH < y(R+D)+T. Consequently, if the payoffs in the cases
of accommodation CH and expropriation WH differ too strongly relative to the costs of
BITs, the BIT will not be able to induce investment. A BIT can consequently not be a
commitment device for countries that have a very strong incentive to play expropriate.
An interesting point also relates to the litigation costs T. As long as parties must carry
their own legal costs, high arbitration costs, such as those arising through lengthy
proceedings, will deter MNEs from seeking remedies under the BIT. At the same time,
the very same high arbitration costs may deter HCs from expropriating in the first
place. Any call to make tribunals more “efficient” in the sense of reducing litigation
costs must take these opposing deterrence effects into account.

4.4.3.4       Hostages and Collateral
Section 3 discussed how MNEs protect their investment in the absence of legal
protection. Based on the analysis in section 4.4.3.3, we can now subsume BITs in the
economics of FDI protection. First, the functioning of BITs requires an action of the
affected investor – therefore, a BIT does not imply hands-tying in the economic sense
of the term. In fact, a BIT institutionalises the exchange of hostages and/or collateral.
The difference between hostages and collateral is the value the asset has to the hostage
or collateral taker, in our case the MNE. While a collateral is valuable to both parties,
a hostage is usually primarily only valuable to the hostage giver. How does this relate
to a BIT? In essence, through the enforceability of arbitration awards, BITs
collateralise the foreign assets of the host country (or convert them into hostages). At
the same time, through the ascertainment of wrongdoing by a neutral body and the
publication of awards, the HC places its reputation as a hostage or, depending on the
ease with which this reputation can be converted into a compensation for the MNE, a
collateral. As a consequence, with regard to BITs, the same incentive problems are
94                                                                   The Economics of BITs

encountered as before: if destruction of the hostage is too costly (high litigation costs,
small expected awards), the investor may have no incentive to do so. On the other
hand, high expected awards (such as through the lack of tribunals to differentiate
between cooperative and uncooperative behaviour) and small litigation costs may raise
the incentive to "collect the collateral" even if no uncooperative behaviour took place.
As Eger (1995) notes, the perfect incentive structure in all cases using ex-ante
guarantees cannot be achieved.302 The functioning of BITs as a commitment device
suffers inevitable from the same shortcomings.

4.4.3.5           Extensions
The commitment game is obviously a stylised game and, in reality, the host country
and the MNE have many choices that could not be incorporated here. The basic
intention was to underline the similarity to the hostage/collateral case and to discuss
the interplay of the main variables. In this section, three directions into which the game
can be extended are analysed. For example, the above model assumes that the HC
would comply with a ruling and pay the damages D to the MNE. If the host country
fails to comply, the MNE may seek to enforce the arbitration award using domestic
courts. This aspect is discussed in section 4.4.3.5.1. Also, at virtually every stage of
the game and certainly prior to arbitration, parties can decide to settle. This will be
discussed in section 4.4.3.5.2. Lastly and inherent to international law, states always
have an exit option in the sense that they may leave the system altogether if it is no
longer beneficial to them. This will be discussed in section 4.4.3.5.3.

4.4.3.5.1         Compliance and Enforcement in Third Countries
An interesting observation can be made that relates to the potential strength of
international adjudication and especially BITs: the host country always has the option
to ignore the award. Note that the costs of ignoring the award are only marginally
influenced by the tribunal, but rather depend on the reputational costs of ignoring and
the possibilities of enforcement of the award in a third country – two factors lying
outside the scope of the tribunal. A consequence is that higher compensation payments
determined by tribunals will not necessarily induce the HC to accommodate, but may
simply lead the HC to ignore the award. This will be exactly the case when the
reputational sanction of non-compliance is low or the likelihood of a loss of assets that
may be enforced through domestic courts, e.g., under the ICSID convention, is
negligible (either because such assets do not exist or the enforcement in third countries
is too costly for the MNE). At this point, the country already faced a reputational loss



302
      Eger (1995), p.187.
The Economics of BITs                                                                                        95

when the tribunal declared the country had breached its obligations.303 So, if the
additional reputational loss for not complying is small, ignoring the award can be a
rational strategy. Argentina is an interesting case in this respect, facing more than 25
pending cases under ICSID304, mainly due to measures taken during its financial crisis.
It remains to be seen how the payment behaviour of Argentina develops if many of
those awards find the defendant guilty. A noteworthy aspect is that the smaller the
damages the tribunals awards, the more the HC will, ceteris paribus, be inclined to pay
the damages specified in the award and thus avoid further reputational losses and
lawsuits in third countries. Whether the HCs comply with their obligations as
expressed in BIT arbitration awards is not monitored. Yet, it appears that the
enforcement of awards in third countries does not occur very often.305 In addition, and
supported by the empirical research by Franck (2007), the amount of the damages
awarded might - after all - on average be negligible for states.306

4.4.3.5.2          Settlement vs. Trial
Settlement relates to the possibility of the party finding a consensual solution to any
conflict before or after arbitration has been initiated.307 Settlement is, at first glance,
preferable to arbitration in the sense that it is consensual and saves litigation costs. It is
straightforward to see that settlement is facilitated through the shadow of arbitration.
As a first step, this section will sum up the discussion on settlement vs. litigation in the
domestic context, as there are many similarities to the context of international
investment law. Nevertheless, the analysis of international investment law must
consider the potentially high importance of reputational concerns. Frequently, cases
where a settlement was reached after the arbitration proceeding was initiated can be
observed. Unfortunately, no information exists on the frequency of claims settled even
before arbitration is initiated, but as arbitration imposes costs on the parties, this
number should considerably exceed the number of known arbitrations.



303
      This is, of course, only true when the award was public. On the transparency of international
      investment law, see chapter 7.
304
      According to the ICSID website as of September 1st, 2010.
305
      See section 3.3.3.8.3.
306
      Remember that at this point, the arbitration process has already been initiated. The question of
      compliance therefore relates to the award, not the treaty. The point that the amount of damages is
      often not extremely high is therefore consistent with the discussion in section 4.4.1.3 where the
      non-reputational costs of treaty breach are discussed (which include, in addition to the awarded
      damages, considerable legal costs).
307
      Please note that this is not part of the commitment game as such. In the commitment game, there
      will usually be no trial. Nevertheless, a conflict is already persistent at this point. The reason for the
      conflict is not relevant here; possibilities include asymmetric information and/or different
      estimations on the interpretation of the relevant legal rule.
96                                                                              The Economics of BITs

In the domestic context, the question of trial vs. settlement has been analysed
extensively.308 The standard results from the domestic context are summarised by
Posner (2007), Shavell (2004) and Cooter and Rubinfeld (1989). If parties to a trial do
not differ in their subjective estimation on the outcome of the trial or the settlement
costs and the stakes are zero-sum in the sense that the penalty for the defendant equals
the compensation for the plaintiff, settlement will only occur when the transaction
costs for settlement are smaller than the litigation costs.309 In addition, the space for
settlement is affected by the relative difference of the parties' beliefs on the outcome of
the trial (e.g., due to private information). Specifically, settlement is feasible when the
plaintiff's estimate of his probability of winning is lower than the defendant's estimate
(of the plaintiff winning). Such is the case, for example, when the parties are
pessimistic regarding their subjective chances of winning. Settlement may also occur
when both parties are optimistic – "as long as the plaintiff's estimate of the expected
judgement does not exceed the defendant's estimate by more than the sum of their
costs of trial".310 As a consequence, when the stakes of the lawsuit are high (and
symmetric), the more likely it is that the case will be litigated.311 A similar logic
applies to a divergence of the stakes of the case. If the defendant’s stakes are high
compared to the plaintiff’s, litigation is less likely compared to when stakes are
equally high.312 In addition, if the parties are risk-averse, the likelihood of litigation
will ceteris paribus be lower.313
These points are well-established in the economics of the legal process and equally
true for international investment law. There is consequently no need to scrutinise them
in more detail here. However, specific to international investment is that especially the
difference in stakes can be of high relevance and this aspect shall thus be analysed in
more detail. The approach used here is similar to that introduced by Cooter and
Rubinfeld (1989). The assumptions of the commitment model, most importantly risk
neutrality, still hold. In addition, it is assumed that settlement imposes costs (e.g., for
negotiations) S, which do not differ between parties. Further, both parties have the
same estimation on the outcome of the trial, denoted by the probability y that the
tribunal finds the defendant guilty. Finally, we assume that the HC will comply with
the ruling of the tribunal. A natural prerequisite for settlement is the existence of a



308
      See Cooter/Rubinfeld (1989), p.1075 or the overviews in Posner (2007), p.597ff or Shavell (2004),
      p.401ff.
309
      Cooter and Rubinfeld (1989), p.1075.
310
      Shavell (2004), p.403.
311
      Shavell (2004), p.405. Notably, a countervailing effect is that higher stakes may also induce more
      effort by the parties. See Posner (2007), p.599f.
312
      Posner (2007), p.599.
313
      Posner (2007), p.599.
The Economics of BITs                                                                                    97

range of payoffs that is acceptable to both parties. The range is defined by the threat
values of each party. The threat value is here defined as the value the party receives in
the case of non-cooperation (proceeding with the trial and paying damages instead of
settlement). The notation is similar to the commitment game, where y denotes the
likelihood the defendant (HC) is found guilty, T the litigation costs, D the
compensation payment and R the reputational damages.


Threat value of the HC (payoff in the case of non-cooperation)314
y(- R - D) - T                                                                               (4.5)
Threat value of the MNE
yD - T                                                                                       (4.6)
The non-cooperative value is the sum of the subjective threat values
y(- D - R) - T+ yD - T
= -yR -2T                                                                                    (4.7)
The cooperative value is (the net transfer between the parties amounts to zero)
-2S                                                                                          (4.8)
The surplus of cooperation is therefore the difference between cooperation and non-
cooperation. If this surplus is positive, the parties will prefer settlement over trial.
-2S -(-yR -2T)> 0 Ù                                                                          (4.9)
yR - 2 (S-T)>0 Ù                                                                             (4.10)
yR > 2 (S-T)                                                                                 (4.11)
This result is in principle equivalent to the situation in the domestic context as
summarised above. Imagine that there were no reputational sanctions (R =0) in excess
to the zero-sum damages. In that case, the choice between settlement and trial would
simply depend on which procedure created fewer costs. However, the point
emphasised here relates to the impact of asymmetrical stakes as a result of reputational
sanctions. Please note that the threat to seek arbitration must be credible
notwithstanding the reputational sanctions. There will neither be arbitrations nor
settlement if yD - T <0, as otherwise the threat of the MNE to bring a suit will not be
credible. The important aspect highlighted by equation 4.11 is that high reputational
sanctions will weaken the threat point of the HC and thereby make settlement more


314
      Please note that the payoffs for expropriation (WH) and the losses of the MNE (LM) are at that point
      sunk and do not influence the decision of the parties. The payoffs are taken directly from figure 4.1.
98                                                                            The Economics of BITs

likely. In that case, settlement is preferable for both parties. It is noteworthy that the
settlement decision may have distributional implications. Lacking further information,
it is reasonable to assume that the surplus to the parties will be equally divided. Put
differently, a reasonable settlement agreement will yield a payoff for each party that
matches the respective threat value plus an equal share of the surplus.315 The surplus
has been derived in equation (4.10) as yR - 2 (S-T). The expected payoffs for
settlement vs. trial are summarised in the following table:

 Settlement
                       Payoff MNE                           Payoff HC
 vs. Trial

                       yD - T + 0.5 (y R - 2 (S-T))         y(-D - R) - T+ 0.5 (yR - 2 (S-T))
 Settlement
                       = y(D + 0.5 R)                       = y(-D - 0.5 R)

 Trial                 yD - T                               y(- D - R) - T



The important point is that the MNE profits from high reputational sanctions for the
HC only in the settlement case. In that case, the MNE can obtain a share of the
reputational capital (in this example one half). The reputation is lost in the case of trial.
This is exactly the negative-sum feature inherent in reputation as an enforcement
mechanism.
The analysis differs with regard to the settlement surplus when settlement after the
case has been officially filed and made public is considered. While the structure is
essentially equal to settlement before registration, two points must be considered here:
(1) any reputational sanctions from the registration of the case (assuming it is an
ICSID case) will already be sunk. (2) Settlement itself may create reputational
sanctions in the sense that the willingness to settle may be interpreted by third parties
as a negative signal. If T is rather understood as the cost for a continuation of the
arbitration, and R' < R (as part of the reputational loss is sunk) and S’ denotes the
reputational losses for settlement, equation 5, the settlement surplus, can be modified
to:
-2S – S’ -(-y R'-2 T)> 0                                                              (4.12)
Two effects affect the settlement surplus compared to the analysis prior to registration
of the case (initiating arbitration). Higher reputational sanctions for settlement, R', will
reduce the settlement surplus and make trial more likely. In addition, part of the


315
      Cooter and Ulen (2004), p.415, use the same assumption.
The Economics of BITs                                                                   99

reputational losses to the HC are already sunk (R' < R), thereby also tipping the
calculation in favour of trial. Consequently, it can be expected that the registration of
the application reduces the settlement surplus. This observation raises an important
point, namely the question of whether there will always be a settlement solution as
long as the value of cooperation (settlement) exceeds the value of non-cooperation
(trial). One of the possible reasons for the failure of (early) settlement may be
asymmetric information. In the given example, the HC may not be sure whether the
condition for going to trial (yD - T >0) holds for the MNE and the MNE may not be
able to signal this information. Filing for arbitration may provide exactly such a signal
and consequently a settlement solution can be reached (which is, however, more costly
to the parties). Moreover, settlements may fail as a result of the potential inability of
parties to divide the surplus316 and/or principal-agent problems between lawyer and
client.317
The importance of reputation for the settlement process may raise concerns on the
existence of nuisance suits. A nuisance suit has been defined as "a lawsuit with low
probability of success at trial, brought even though the plaintiff knows that his
probability of prevailing would not justify his costs if the judicial process were to be
completed instantly."318 It might be argued that the high potential reputational
sanctions and the fact that the MNE will obtain some of the reputational capital in the
case of settlement will lead to a high frequency of nuisance suits. However, if yD - T
<0, the MNE’s threat of going to trial is not credible. Knowing this, the (risk-neutral)
HC will never agree to a settlement – irrelevant of the magnitude of the reputational
sanctions. Exceptions may apply, inter alia, in cases of sunk costs (e.g., the plaintiff
pays his lawyers up front to increase the credibility of his claim), different timing of
costs (e.g., the defendant incurs costs before the plaintiff does) and asymmetric
information (the defendant cannot evaluate the merits of the claim ex-ante).319 None of
these points appears to be particularly important with regard to BITs. For example,
while there may be cases of asymmetric information, it is not reasonable to assume
that states frequently lack a reasonable estimation on the merits of a claim. In addition,
it is noteworthy that the ICSID Convention provides a safeguard against frivolous
suits. Article 41 (5) of the ICSID Arbitrations Rules states: "Unless the parties have
agreed to another expedited procedure for making preliminary objections, a party may,
no later than 30 days after the constitution of the Tribunal, and in any event before the
first session of the Tribunal, file an objection that a claim is manifestly without legal


316
      See Cooter, Marks et al. (1982).
317
      See Miller (1987).
318
      Rasmusen (1998), p.690.
319
      See Rasmusen (1998), p.691 for a short description of the different approaches.
100                                                                           The Economics of BITs

merit." If the tribunal decides that a claim is without legal merit, it can render an award
in favour of the defendant and the arbitration proceeding will be discontinued. To the
knowledge of the author and based on the list of cases as presented by ICSID, this
clause has only been invoked once, namely in the case Trans-Global Petroleum Inc. v.
Hashemite Kingdom of Jordan. The request of the defendant in this case was only
partially successful.320 In sum, it appears that nuisance suits do not pose a major
problem in international investment law.
What can be inferred from the analysis? First, it is straightforward to identify the
factors that facilitate settlement, as these are in principle similar to the domestic
context: high litigation costs and pessimistic parties will encourage a settlement
solution. Naturally, high settlement costs will tip the outcome in favour of trial. High
stakes will favour settlement solutions if the plaintiff is more optimistic than the
defendant. An important point is that higher reputational sanctions will increase the
likelihood for settlement. In addition, the reputational costs have a distributional
dimension in the settlement case as MNEs can capture some of the benefits that accrue
if the HC can avoid trial. These points are equally true for settlement after registration
of the trial. The difference in this case, however, is that the surplus of settlement will
be lower. Nuisance suits do not appear to play an important role, neither in theory nor
in practice.
While reputational sanctions and compensation payments influence the question of
settlement vs. trial, another important point should not be neglected: when reputation
and compensation costs become excessive, HCs will be tempted to "leave" the BIT
system altogether, e.g. in the short term by ignoring awards, and in the long term by
not concluding any BITs in the future. This problem will be discussed in the following
section.

4.4.3.5.3         The Perils of Success
Section 4.4.3.5.1 highlighted an important point: states always have the option of not
paying the award. In addition, states always have a more fundamental exit-option of
leaving the BIT system altogether. If tribunals consistently interpret the terms of the
BIT too broadly, the BIT might become too costly for countries trying to commit
themselves. The result is that these countries might exit their existing obligations and
abstain from entering new ones and thus weaken the system as such. This tendency has
been observed by Van Aaken (2008a), who gives a number of examples of what she
refers to as "perils of success". One aspect concerns the problematic balancing by


320
      See Trans-Global Petroleum Inc. v. Hashemite Kingdom of Jordan, ICSID Case No. ARB/07/25,
      Decision on the Respondent’s Objection under Rule 41(5) of the ICSID Arbitration Rules (May 12,
      2008).
The Economics of BITs                                                                 101

tribunals between ex-ante commitment and ex-post flexibility. Often, the problem is
therefore not rooted in verifiability, but in the fact that BITs bear similarity to
incomplete long-term contracts and that the standards in BITs require interpretation.
The risk exists, for example, that tribunals will interpret the standards of the BIT too
narrowly. This point will be discussed in more detail in section 4.4.5.1. In any case,
the danger persists that even reliable countries might refrain from signing BITs in the
future or quit their existing obligations if these become too expensive. To provide a
recent example, Ecuador decided to denounce the ICSID Convention effective January
2010.321

4.4.3.6         Summary
Based on the analysis above, it is possible to summarise the main findings and
hypotheses of the commitment model as such:
      x Crucial for the functioning of the commitment mechanism is that the threat of
        the MNE to initiate arbitration is credible. However, using the enforcement
        mechanism places costs on the MNE. Consequently, the enforcement costs
        must be small enough to induce the MNE to seek enforcement. High expected
        compensation can achieve the same effect.
      x For the host country, commitment is only valuable if the potential additional
        costs of breach exceed the opportunity costs of non-cooperation. It is, prima
        facie, irrelevant whether these additional costs are the result of high damages or
        high reputational sanctions.
      x Structurally, the BIT features elements of hostages and collateral. Only if these
        elements are valuable to the HC can the BIT induce cooperation. Put
        differently: only when the HC has a reputation to lose or foreign assets to be
        confiscated can a BIT credibly commit the country.
      x Higher (in terms of compensation) awards are not necessarily helpful for the
        MNE as the host country can always choose to ignore the ruling and also make
        commitment too costly altogether (as MNEs seek arbitration although no
        expropriation occurred in the first place).
      x Further, excessive awards will not only shift the incentive of the HC towards
        ignoring the award, but, if tribunals expand the scope of application of BITs,
        may lead countries to leave the BIT game altogether. However, if the expected
        costs of breach are too low, the BIT may not induce commitment and, if


321
      See ICSID NEWS RELEASE: Ecuador Submits a Notice under Article 71 of the ICSID
      Convention, July 9, 2009.
102                                                                    The Economics of BITs

          expropriation already occurred, incentives for settlement solutions in the
          "shadow" of arbitration might vanish. This is the same structural problem in the
          hostages/collateral case in general.
      x Using the BIT as a commitment device is therefore only interesting for
        countries that have sufficient benefits through spillover and have foreign assets
        that can be placed as collateral or a reputation vis-à-vis investors or third parties
        that can be placed as a hostage. However, the reputational loss of the HC is not
        a gain for the MNE. Therefore, even potentially high reputational losses will
        not induce investment if the MNE has no incentive to "destroy the hostage".

4.4.4           Signalling
The preceding section analysed Bilateral Investment Treaties in their function as a
commitment device. The analysis rested on the, arguably problematic, assumption of
complete and symmetric information. However, the problem addressed by BITs might
not be so much the hold-up caused by time inconsistency as such, but a problem of
asymmetric information in the sense that investors cannot differentiate between states
(or governments) that have the willingness and ability to protect the (sunk) investment
or not. The following section quickly reviews the reputation model that is equally
characterised by the existence of asymmetric information. It will be argued that the
framework of the reputation model is not suitable for the analysis of BITs in a
situation of asymmetric information. The more fruitful approach is to understand the
BIT as a signal (section 4.4.4.2 following).

4.4.4.1         Reputation Revisited
The discussion in section 3.2.2.2.2 on reputation in international investment law was
based on the premise of asymmetric information between the MNE and the host
country. More specifically, there were two types of states, a good (reliable) and a bad
(unreliable) type. The MNE was assumed to be unable to differentiate between these
two types. How can we analyse BITs in the environment of this kind of asymmetric
information? A natural starting point for this analysis would be to insert the costs of
BITs (or rather of breach) into the existing reputation framework that has been
outlined before. However, as this section will argue, an analysis of this kind is not
instrumental for two reasons: first, it has been shown that BITs cannot induce
investment of the MNE in the first period of the game. Investment will occur when
The Economics of BITs                                                                103

          L2
T!         M
                   , which is independent of WH. In other words: investment will occur
     ( LM  CM ) 2

when the payoffs to the MNE itself are high (high CM as compared to LM).
Consequently, a country facing low prior beliefs concerning its reliability cannot, in
this framework, use BITs to induce the investment in the first place. Put differently,
modelling BITs as additional costs in the case of breach cannot overcome the inherent
weakness of the reputation model in the trust game that the countries cannot start
building up a reputation for good behaviour when prior beliefs do not lead to
investment in the first place.

                                                           L2
Second, and even more problematic, when T                  M
                                                                    , the effect of BITs
                                                      ( LM  CM ) 2

represented as additional costs in the case of breach has an adverse impact on the
                                                                          WH  CH
likelihood of investment. The MNE is playing invest with probability              . We
                                                                           G WH
are interested in the change of the probability when we change WH. Therefore we
calculate:

 d § WH  CH ·      CH
    ¨        ¸            ! 0 (because CH ! 0 ).
dWH © G WH ¹       G WH
                      2




The probability that the MNE will invest is increasing in WH. Consequently, reducing
WH by subtracting a (reputational) sanction reduces the probability that the MNE will
invest in the market. This result is counterintuitive and not realistic. The reasons for
this result lie in the mixed strategy properties of the equilibrium. Reducing WH makes,
all other things being equal, accommodation for the HCbad in the present round more
attractive (note that the HCbad will expropriate in the last round for sure). To keep the
HC indifferent between expropriation and accommodation again (which is necessary
for a reputation equilibrium), the MNE must reduce his probability of investing.
The result that reducing the benefits from expropriation in a reputation game based on
Kreps and Wilson (1982) leads to lower probability of investment is troublesome. In
the original game itself, where a weak monopolist tries to deter market entrance, a
similar observation arises: increasing the payoff for the monopolist when there is no
market entry makes entry even more likely (while one should assume that the
monopolist fights even harder when he has more to gain). As before, this is the
consequence of the mixed strategy equilibrium where one player calculates his
104                                                                               The Economics of BITs

strategy using the payoffs of the other player in a way to keep the other party
indifferent between his options. However, if there were only pure-strategy equilibria,
for example a pooling equilibrium where both (the bad and the good type) play
accommodation, no reputation building could occur. The same is true for a separating
equilibrium in pure strategies. The bad type reveals his type right away and the MNE
reacts accordingly. A conclusion could be that it is unreasonable to assume that players
play their perfect Bayesian mixed strategies and that our focus should be shifted to
different, more reasonable strategies.322 In any case, the bottom line is that dealing
with the effect of BITs on asymmetric information cannot fruitfully be analysed within
the framework of the reputation model outlined above. Therefore, for the analytical
purpose of this book, it appears reasonable to address the question of asymmetric
information in international investment law from a signalling perspective. This
approach will be taken in the following sections. Put differently, the BIT will be
interpreted as a signal and thus as an attempt to induce investment. The main points
will be highlighted using a simple signalling model (section 4.4.4.4). Before that, the
basic functioning of signals will be described (section 4.4.4.2) and also the idea of
rights as signals will be presented (section 4.4.4.3).

4.4.4.2            Signalling Theory
Signalling theory goes mainly back to Spence (1973).323 Signalling games are
concerned with the strategic submission of information. The starting point is the
assumption that producers have private information on the quality of the good they are
selling that consumers cannot observe. The consumer's willingness to pay increases
with the quality of the product. Since consumers are not able to assess the quality of
the good before purchase, they will only be willing to pay a price based on the
(perceived) average quality of all goods. This, in turn, incites producers of high quality
goods to invest into signals that convey to the consumers the true (high) quality of the
good. However, low quality producers might be able to imitate these signals. The
original signalling model as developed by Spence (1973) dealt with the question of
how applicants (or employees) can signal their productivity to employers through
education.
A necessary condition for a signalling model to function relates to the costs of the
signal. The activity that is supposed to serve as a signal must generate different
(marginal) costs for different economic agents (single crossing condition). In the
education example, acquiring education must be less costly for the more productive



322
      See Holler (1993) for an example.
323
      The following introduction to signalling follows Holler and Illing (2006), p.177ff.
The Economics of BITs                                                                  105

employees as compared to the unproductive employees. There can be two types of
pure strategy equilibria: a separating equilibrium and a pooling equilibrium. In the
separating equilibrium, the different types send different signals, while in the pooling
equilibrium, both types invest into the same signal (or send no signal at all). If a
separating equilibrium can be established, the good type (e.g., the type with better
quality or higher productivity) can reveal his true nature and, as in the context of
symmetric information, obtain a higher price on the market. In a pooling equilibrium,
however, the consumer (e.g., employer) still cannot differentiate between the types.
Thus, the price will be formed based on the average quality on the market and the
good type would get a smaller price as compared to the symmetric information context
while the bad type would get a higher price. A separating equilibrium can only emerge
when it is, apart from different marginal costs, not profitable for the bad type to imitate
the signal of the good type.

4.4.4.3           Rights and Treaties as Signals
The idea of rights and treaties as signals is based on the signalling theory described in
the preceding section. The basic concept is simple: states or governments have
characteristics that cannot be observed by third parties. A treaty might be more costly
to one group of states as compared to another group in relation to that hidden
characteristic – therefore, states might use treaties as a signal.
Farber (2002) uses a signalling framework to analyse the connection between human
rights, the rule of law and international investment. At the core of his argument lies the
assumption "that legal reform is a good signal of being truly committed to economic
reform."324 In addition, investors can also infer important conclusions from the
observation of the enforcement of human rights generally. Farber (2002) considers the
adoption of constitutionalism as a signal that the ruling coalition has a low discount
rate. The idea is that constitutionalism benefits the economy in the long run, but has
short term costs to the ruling coalition, embodied especially in a loss of power. In
addition, an independent judiciary that enforces human rights issues signals that the
investor can also rely on this judiciary with regard to expropriation issues. This
enforces the investor's belief that the ruling coalition does not discount long-term
profits too heavily. Hathaway (2002) also looks at human rights and ascribes them a
dual role: a functional and an expressive one. The functional role relates to the
commitment function of the treaty, while the expressive role may relate to signalling;
as Hathaway (2002) asserts: "Yet treaties also have an expressive function that arises




324
      Farber (2002), p.84.
106                                                                               The Economics of BITs

from what membership in a treaty regime says about the parties to the treaties."325 If
the treaty provides for strong monitoring or enforcement mechanisms, the two
functions of the treaty usually are indivisible. Yet, if there are no monitoring or
compliance mechanisms in place, the two functions may divorce. This disjuncture
might even have the adverse effect that the treaty will "serve to relieve pressure for
real change in performance in countries that ratify the treaty".326 As a result of non-
existing compliance mechanisms (or imperfect information) and potential reputational
benefits of simply expressing commitment to a (human rights) treaty, signing such a
treaty might, paradoxically, be especially interesting for countries that are ex-ante in
non-compliance with the treaty terms.327 These countries then have no incentive to
make the costly domestic policy changes necessary to come into compliance. Clearly,
this interpretation of human rights treaties is at odds with the aforementioned
interpretation by Farber (2002).328 With regard to international law in general,
Simmons (2000) describes international legal commitments, especially those relating
to the International Monetary Fund, as a "signaling device to convince private market
actors as well as other governments of a serious intent to eschew the proscribed
behavior."329 In addition, the author empirically analyses the compliance decisions
with regard to IMF obligations and finds that international commitments have a
distinctive effect on the behaviour of governments, probably because of reputational
concerns. Martin (2005) links the idea of treaties as signalling devices to the domestic
decision of the US President between different forms of international commitment. In
the United States, the government can construct an international commitment either as
an executive agreement or as a treaty. The idea is that treaties are more costly (because
they require a two-thirds majority in the Senate to be ratified) than executive
agreements. Therefore, the choice between these devices is a strategic one: the
President can be of two types, either reliable or unreliable. The agreement he offers
may signal something to the potential treaty partner with regard to the President's type.
In addition, the action depends on other parameters, as for example the potential
benefits of the agreement. An empirical analysis underpins the author's assertion that
the probability that the agreement is a treaty is negatively correlated with the reliability
of the president.
The idea that BITs are a signal is mentioned in virtually every publication on BITs, yet
is never developed and tested on functionality. With regard to BITs, Sauvé and
Zampetti (2007) assert that the “trend accelerated anew in the 1990s, albeit in a


325
      Hathaway (2002), p.2005. Italics in original.
326
      Hathaway (2002), p.2007.
327
      In addition, these countries might have less reputational capital to lose. Hathaway (2002), p.2013.
328
      Hathaway (2002), p.2012.
329
      Simmons (2000), p. 324.
The Economics of BITs                                                                                    107

markedly changed policy and ideological environment, as host country (i.e. capital-
importing) governments in both developing and transitions countries sought to exploit
the putative signalling properties of BITs".330 The question therefore is: are the
signalling properties of BITs only putative or realistic?

4.4.4.4            The Signalling Game
The signalling function of Bilateral Investment Treaties can be exemplified and
discussed using a simple model. As in the reputation model, there are two types of host
governments: a good (or reliable) type and a bad (unreliable) type. The type is
determined by nature: with probability , the host government (or host county, HC) is
reliable. The two types differ in the following way: the unreliable type will always
expropriate the investment once it is made, while the reliable type will always
accommodate. The HC knows its type and has the first choice of action: the HC can
sign a BIT or not. The investor (MNE) does not know the government's type, but can
observe the action of the HC, namely its decision to conclude a BIT or not. The
interesting question with regard to the signalling game is what an investor can infer
from the action of the HC. Therefore, it is assumed that it is irrelevant whether the
country with which the HC concludes the BIT is the home country of the MNE or not.
After observing whether the host country concluded a BIT, the MNE decides whether
to invest or not. Because of the asymmetric information, the MNE at this point does
not know whether it is facing the reliable or the unreliable type. After the investment
decision, the game ends.
The payoffs are as such: concluding the BIT comes with two different costs - the direct
costs of concluding a BIT and the expected costs or compliance costs of the BIT once
it is in force. The costs of concluding the BIT include negotiation and ratification costs
and should not differ substantially between reliable and unreliable types. These costs
are denoted as A. The compliance costs of the BIT once it is in force should, on the
other hand and crucially for a signalling model, differ between the two types and are
denoted by B. The reliable type will always accommodate the investment and should
therefore not face costs of arbitrations (unless there are no systematic errors of
tribunals or nuisance suits). Put diffently, the standards of the BIT reflect the general
behaviour of the reliable type, only minor changes to the institutional order might be
necessary331 and/or a small arbitration risk exists (therefore BR>0). On the other hand,


330
      Sauvé and Zampetti (2007), p.215.
331
      One might think of the sovereignty costs as discussed in section 4.4.1. Nevertheless, the MNE does
      not care about the costs to the host country, but about the arbitration (expropriation risk). This will
      of course be lower when the sovereignty costs are low – in that sense, the two aspects are closely
      related.
108                                                                                 The Economics of BITs

the unreliable type will expropriate and thereby expose itself to high arbitration risk. It
will have to face reputational costs, pay legal fees and may be forced to pay
compensation. The assumption is therefore BR < BU.332 Obviously, the arbitration risk
will also depend on the amount of investment covered by the BIT in question, meaning
that, for example, a BIT with a major capital-exporting country like Germany may
have a bigger effect on BU than a BIT with a small, non capital-exporting country.
If there is no investment and no BIT is in force, payoffs for all parties are zero. If there
is no investment and a BIT is in force, the MNE has a payoff of zero, while both the
reliable and the unreliable type must carry the cost of concluding the BIT. If there is
investment and no BIT in force, payoffs correspond to the baseline game (see section
3.1.1) with CH for the reliable type and WH for the unreliable type. In that case, the
MNE receives CM if he is facing a reliable type and LM if he is facing the unreliable
type (as before LM <0 holds). Finally, if there is a BIT in force and the MNE invests,
the HCs receive the respective gains (CH or WH) minus the costs of conclusion and the
compliance costs of the BIT. The extensive form of the game can be illustrated as
follows:




332
      The subscript R identifies the reliable host country, the subscript U the unreliable host country.
The Economics of BITs                                                                                              109



                                                        Nature


                                    Reliable ( )                       Unreliable (1- )




                               HC                                                         HC


                    BIT                 No BIT                             BIT            No BIT

              MNE                            MNE                        MNE                             MNE



                     Don’t                         Don’t                         Don’t                        Don’t
 Invest              Invest         Invest         Invest     Invest             Invest        Invest         Invest



 CH- A - BR               -A          CH            0       WH -A- BU            -A              WH            0
   CM                     0           CM            0         LM                 0               LM            0


Figure 4.2: The Signalling Game




Notation
CM        =         investment return for the MNE
CH        =         benefit of investment for the HC: taxes and spillover for the reliable type
WH        =         expropriation value of the asset for the unreliable type
LM        =         expropriation value of the asset for the MNE
          =         probability that the HC is of the reliable type
A         =         costs of concluding the BIT
BR        =         compliance costs of the BIT for the reliable type (arbitration risk)
BU        =         compliance costs of the BIT for the unreliable type (arbitration risk)
with BU > BR
110                                                                   The Economics of BITs

4.4.4.5           Equilibria
As the MNE cannot observe the type of the host country, but can observe the action
the host country takes, the reliable type might use BITs as a signal. Three types of
equilibria are possible: separating, semi-separating and pooling equilibria. In the
separating equilibria, the optimal strategies of the reliable and the unreliable type
differ. Typically, the reliable type will sign a BIT and the unreliable type will not sign
a BIT. In a semi-separating equilibrium, at least one type uses mixed strategies. In
pooling equilibria, both types will take the same action, that is, both will either sign a
BIT or refrain from doing so. Depending on the parameter specifications, all of these
equilibria are possible. Therefore, it appears reasonable to make further assumptions
on the feasible and realistic parameter space.
First, the existence of many BITs can be observed in reality. If the signalling model
has any explanatory power, a pooling equilibrium where nobody signs BITs can be
ruled out. Such an equilibrium exists, inter alia, when the expected payoff of the MNE
is positive despite the absence of a BIT – the MNE would invest in any case.
Consequently, and considering that a BIT creates costs for both types of HC, no HC
would ever bother to sign a BIT. This is exactly the case when T CM  (1  T ) LM ! 0 .
This equation will be especially true when the probability of facing an unreliable type
is very low. In the following, it is assumed therefore that T CM  (1  T ) LM  0 . Put
differently, in the absence of a BIT, the MNE will not invest. Yet, as CM>0 and LM<0,
the MNE would invest if certain that it faces the reliable and not invest if certain that it
faces the unreliable type.
The costs of concluding a BIT (A) do not differ between reliable and unreliable types.
The MNE will choose its strategy taking the incentives of the respective HC into
account. The prevailing equilibria therefore depend on the payoffs for the HC and can
be summarised as follows:333




333
      See Appendix C for a derivation of these results.
The Economics of BITs                                                            111


 Payoff            Equilibria
 host country

                   Pooling on No BIT (1)
                   Signing a BIT is not worthwhile for any type of HC. The MNE
 CH  A  BR  0
                   will not invest.
 and
 WH  A  BU  0   Strategy HC: (No BIT/ No BIT)    Strategy       MNE:      (Don't
                   Payoff HC: 0                     invest/Don't Invest)
                                                    Payoff MNE: 0

                   Separating Equilibrium
                   Signing a BIT is worthwhile for the reliable type but not the
 CH  A  BR ! 0   unreliable type. The MNE will invest if it observes a BIT and not
 and               invest otherwise.

 WH  A  BU  0   Strategy HC: (BIT/No BIT)        Strategy MNE: (Invest/ Don't
                   Payoff HC: T (CH  A  BR )      Invest)
                                                    Payoff MNE: CM

                   Pooling on No BIT (2)
                   Signing a BIT is not profitable for the reliable type and
 CH  A  BR  0   consequently also not for the unreliable type. The MNE will not
 and               invest.

 WH  A  BU ! 0   Strategy HC: (No BIT/No BIT)     Strategy MNE: (Don't Invest
                   Payoff HC: 0                     /Don't invest)
                                                    Payoff MNE: 0
112                                                                          The Economics of BITs


                   Pooling on No BIT (3)
                   The unreliable type has the incentive to mimic the behaviour of the
                   reliable type. Sending the signal is therefore not profitable for the
                   reliable type. The MNE will not invest.

                   Strategy HC: (No BIT/No BIT)              Strategy MNE: (Don't Invest
                   Payoff HC: 0                              /Don't invest)
                                                             Payoff MNE: 0

                   If
 CH  A  BR ! 0   CH  A  BR ! WH  A  BU

 and               in addition:
 WH  A  BU ! 0   Semi-Separating Equilibrium
                   The reliable type signs a BIT, the unreliable type signs a BIT with
                   probability q and the MNE invests with probability s.

                   Strategy HC:                              Strategy MNE:
                   (BIT/BIT with prob.                       (Invest/Invest with prob.
                          CM T                                      A
                    q                )                       s           )
                        LM (1  T )                             CH  BU


                   Payoff HC:                                Payoff MNE: 0
                    T > s (CH  A  BR )  (1  s )( A) @



Before discussing the equilibria as such, a few words seem to be in order concerning
the prior belief of the MNE with regard to the reliability of the HC, in the signalling
game denoted as . The prior belief can be interpreted twofold: either, there is a
population of host countries and the MNE cannot differentiate between these countries
at all but knows the respective fraction of reliable ( ) and the fraction of unreliable (1-
 ) HCs in the basic population. The MNE uses this knowledge to form a belief about
the HC he is playing, simply taking the known values for                and (1- ) as the
probabilities of facing a reliable or, respectively, an unreliable HC. A different – and
more realistic – interpretation of the prior belief would be that the MNE has a certain
amount of information with regard to a specific HC and has, based on this information,
formed a certain belief with regard to the reliability of the HC, namely . If following
this interpretation, the value of does not denote the fraction of reliable HCs in the
The Economics of BITs                                                                  113

basic population of HCs, but is an individual value formed for each HC separately. As
a basic prerequisite for the ability of BITs to function as a signal, the condition
T CM  (1  T ) LM  0 was introduced. If the prior belief of the MNE that the relevant HC
is reliable is high, the host country has no incentive to sign a BIT. Using this insight,
we can make a point that has so far been neglected when applying signalling theory to
international law: a signal must be seen in perspective with the prior estimation of the
signal-taker on the characteristic that is being signalled. More specifically and
concerning the problem at hand: a signal will not necessarily be sent by the most
reliable HCs. If the prior estimation that a relevant HC is reliable is already high, these
countries especially will not bother to send a signal. A linear positive relationship
between institutional quality and the existence and/or number of BITs cannot be
inferred based on signalling theory.
A second important point is that the condition that the signal is more expensive to the
unreliable than to the reliable type is certainly not enough to ensure a separating
equilibrium. Especially in the investment context, the benefits of the investment (be it
through spillover and taxes on the one hand or expropriation on the other) need to be
taken into account. If the benefits to the reliable type are low, it may not bother to send
the signal at all and a pooling equilibrium without BITs and investment will result.
The same is true if the signal is very expensive to the reliable type. In addition, if the
benefit of the investment to the unreliable type is very high, a pooling situation
without a BIT and without investment would result again as the unreliable type would
try to mimic the behaviour of the reliable type. The reliable type will anticipate this
and not bother to send the signal in the first place. Looking at the separating
equilibrium more closely reveals the following interrelations: an equilibrium where
BITs fulfil a signalling function is more likely when the benefits to the reliable host
country are high (high CH) or the costs of the signal to the reliable type are modest
(low BR). Consequently, if BR rises above a certain threshold, which might be the case
when tribunals inflict unexpected costs on actually reliable states, the reliable type will
refrain from sending the signal. This concern is not unfounded.334 MNEs pursuing
illegitimate claims and errors by tribunals could be reasons for an increase in BR. On
the other hand, if the benefits to the unreliable state are also substantial (high WH) or
the costs of the signal negligible (low BU), a separating signalling equilibrium is ceteris
paribus less likely. Under the reasonable assumption that conclusion costs do not differ
between types, it is evident that these costs are unable to help MNEs differentiate
between types; ergo only ratified BITs can be a signal. However, if conclusion costs
rise above a certain threshold, a separating equilibrium might shift into a pooling



334
      See also section 4.4.3.4.4.
114                                                                            The Economics of BITs

equilibrium without BITs. Nevertheless, in reality, these costs appear to be rather
moderate.
Yet, the existence of BITs is observed in the real world and, if the signalling function
plays a role in reality, the answer probably lies in the semi-separating equilibrium. A
BIT is not too costly for either type of HC, yet the unreliable HC will sign a BIT only
with a certain probability. This could explain the strategy of some countries to sign
BITs, while others, seemingly very similar countries, almost completely abstain from
doing so. In the example presented here, the mixed-strategy equilibrium rests on a
number of conditions. For both types, the investment must be profitable despite the
costs of the BIT, namely CH  A  BR ! 0 and WH  A  BU ! 0 . For the bad type, this
might be the case if the chances of being punished for uncooperative behaviour are
low. Nevertheless, the payoff for the reliable type must be higher than for the
unreliable type, denoted by the equation CH  A  BR ! WH  A  BU . Obviously, if this
wasn't the case, a pooling equilibrium would emerge. High conclusion costs A may
destroy the semi-separating equilibrium (just as the separating equilibrium) altogether.

4.4.4.6           Hidden Characteristics and Hidden Intentions
The analysis of the signalling properties of BITs is further complicated by the fact that
the asymmetric information may take different forms. So far, the set-up of the
signalling model was simply based on the separation between two types: a reliable
type that always accommodates the investment and an unreliable type that prefers
expropriation. The question arises as to how the two types specifically differ from each
other and why it should be impossible for MNEs to differentiate between types. At this
point, the connection to the commitment game as presented in section 4.4.3.2 can be
seen. In essence, the reliable type is the type that would accommodate the investment
anyway or that would do so induced by the BIT335 - this, inter alia, implies a low
discount or strong political constraints. Consequently, the asymmetric information can
consist of many different elements – some more reasonable than others. Broadly, these
may be divided into hidden characteristics and hidden intentions.
At first glance, the basic problem addressed by BITs appears to be a factual or
perceived lack of institutional quality. The question of whether BITs are a signal or a
commitment device is therefore often translated into the question of whether BITs
complement the given institutional quality or substitute for the lacking institutional
quality, serving thus as a signal in the former and a commitment device in the latter
case.336 Yet, the information on the institutional quality of countries is abundant. For


335
      Therefore, the commitment and the signalling function can never completely separated.
336
      See e.g. Ginsburg (2005).
The Economics of BITs                                                                                 115

example, one of the most widely used set of indicators of institutional quality, the
Worldwide Governance Indicators, published by the World Bank, even divides
institutional quality into six parameters: Voice and Accountability, Political Stability
and Absence of Violence, Government Effectiveness, Regulatory Quality, Rule of
Law and Control of Corruption. This indicator is only one among various, easily
accessible sources of information. A signalling model, however, requires the existence
of asymmetric information.
The idea that a bilateral investment treaty can convey yet unknown information in this
regard is therefore, at best, questionable.337 However, concerning the discount factor
and the payoffs of a specific country, asymmetric information might indeed persist.
Namely, the value of CH and WH may depend on political characteristics not easily
observable. In addition, the discount factor represents the willingness of a country to
forgo current for future profits. This might depend on objective factors like the
political system as such, but certainly also on the plans and objectives (or better,
preferences) of the constituents (represented by the government) and might therefore
better be understood as hidden intentions than hidden characteristics. The problem is
that hidden intentions cannot be revealed through the publication of information, e.g.,
by an investment promotion agency. A signal in the form of an international treaty
might therefore indeed be suitable to reduce information asymmetries. Nevertheless,
institutional quality is not unimportant: if the quality is high, the short-run intentions of
the constituents become less important as the government (representing the
constituents) is more strongly limited in its actions by domestic safeguards – also with
regard to foreign investment.338
The conclusions from this discussion are twofold: first, if BITs function as signals,
their occurrence is not necessarily correlated with high institutional quality, at least not
in a linear way. If the prior is understood not as distribution over types, but as a
specific prior belief on the reliability of the host country, the effect might even be the
reverse: the signal is sent by countries that are perceived as having high political risk
or low institutional quality, especially with regard to their judicial system. Countries
that are perceived as "safe" will not bother to incur the costs of the signal. Second, if
the signal relates to the current preferences of the constituents, "old" BITs cannot have
a signalling function. The logic here is immediately obvious: a BIT that has been



337
      Please note that the point here is not to refute the existence of asymmetric information. Rather the
      ability of BITs to address the type of asymmetric information that actually exists is questioned.
      Notably, if there was no asymmetric information from the start, the commitment model would also
      not be backed up by reputational concerns and therefore be useless.
338
      This does not presuppose that governments are non-benevolent. It has been explained before that
      time inconsistency problems are also relevant for benevolent governments.
116                                                                   The Economics of BITs

concluded in the early 1980s will not provide much information on the preferences and
planned policies of the current government.

4.4.4.7         Summary
International law and international treaties have been described as signals in a number
of issue areas including the area of human rights and the provisions established by the
IMF. Many authors have alluded to a potential signalling function of BITs without
analyzing this function in detail. As with any economic signal, the signalling function
of BITs presupposes the existence of asymmetric information. In addition, different
types of players need to have different costs associated with the signal. The analysis of
BITs as a signal in the preceding sections highlighted the following points:
      x Signed but not ratified BITs cannot be a signal as long as negotiation and
        conclusion costs do not differ between types.
      x The investment decisions of MNEs depend on their prior belief regarding the
        reliability of a country: if this prior does not reflect the distribution in the
        general population of host countries, but is formed specifically for each
        country, a BIT can be talked about as a signal only in "perspective": the signal
        will therefore not necessarily be sent by "low risk"-countries (as a result,
        positive correlations between institutional quality and BITs cannot necessarily
        be expected based on signalling theory)
      x If the unreliable type can obtain significant benefits from investment, it might
        try to mimic the signal despite higher signalling costs. This will undermine the
        ability of the reliable type to separate himself. The result may then be a semi-
        separating equilibrium or pooling without BITs. In addition, if the potential
        gains of the investment to the reliable host country are low, the signal will
        equally not be sent.
      x Increasing the costs of the signal to the reliable type will make the reliable type
        exit the game. This concern is not unfounded. Alternatively, increasing the costs
        for the unreliable type can support a separating equilibrium. Consequently,
        BITs with "big" states should have a stronger signalling function (if at all).
      x A mixed-strategy equilibrium can exist where the unreliable type signs a BIT
        with a specified probability. This equilibrium aligns signalling theory best with
        the uneven spread of BITs in reality.
      x The nature of the asymmetric information is problematic: the asymmetric
        information is more likely found in the intentions (current preferences) of the
        host countries and less in their institutional quality (as the latter is public
The Economics of BITs                                                                 117

           information). Consequently, the signalling function of older BITs is certainly
           rather limited.
Before turning to the question of how far BITs are really capable of serving as a signal
or a commitment device, conceivable additional functions of BITs will be discussed as
well as the relevant empirical literature.

4.4.5              Beyond Commitment and Signalling
It can be argued that a BIT is neither primarily a device to commit host countries nor a
signal, but instead serves a different function. To start with, BITs might be best
understood as tools to ensure an optimal trade-off between flexibility and commitment.
A different theory categorises BITs rather as a form of development aid. In addition,
BITs may be designed to deal with the risk-aversion of investors. Each of these ideas
will be discussed in turn.

4.4.5.1            Flexibility vs. Commitment
Any form of effective commitment comes naturally at the cost of less flexibility at a
later stage. This is not only true for BITs, but for any issue area where states optimally
pre-commit to a certain course of action, e.g., in the area of monetary policy.339 The
potential for international investment protection law to balance these diverging
interests has recently been analysed by Van Aaken (2009). As the author notes, "states
trade credibility for sovereignty, as international investment law not only restricts
regulatory conduct of states to an unusual extent but also subjects it to control through
compulsory international adjudication mechanisms."340 The question of how BITs and
especially tribunals are coping with uncertainty and unforeseen contingencies from an
economic point of view is certainly an important one. Indeed, tribunals have dealt with
cases where investors were not subject to opportunistic measures, but where the
respective host country reacted to unforeseen contingencies in good faith.341
In this sense, BITs and arbitration might be less of a commitment device but rather an
attempt to efficiently deal with necessary adaptions of a long-term contract, thereby
saving transaction costs when drafting the contract.342 In that case, the problem is not
the hold-up situation as such. The idea is not, in deviation to the other two approaches
(signalling and commitment), the observation that international law lacks
supranational enforcement and that, at the same time, the host country faces a time


339
      See for example Drazen (2002), p.126ff.
340
      Van Aaken (2009), p.3.
341
      Van Aaken (2009), p.15.
342
      Chapter 5 discusses this point in more detail.
118                                                                 The Economics of BITs

inconsistency problem. The idea is rather that the BIT as a contract is incomplete and
needs neutral interpretation and specification along the way. The functioning and
importance of mechanisms that support the adaption of long-term contracts to
unforeseen contingencies is described by Eger (1995).343 Arbitral tribunals can then be
a less expensive and better suited solution as compared to national courts. It is clear
that BITs and the system of international investment law in general need to deal with
the problem of unforeseen contingencies. Obviously, if the ex-post costs of
inflexibility rise above a certain threshold, states will no longer use BITs to commit to
an investor-friendly course of action. However, in the case of BITs, arbitration is not
mainly an alternative instrument for the adaption of contracts. First, international
arbitration is unlikely to be less expensive than national courts. Second, if the main
issue were simply contract adaption, BITs would also exist between developed
countries and thus cover a more substantial amount of FDI. Put differently, even
though the BIT and the tribunals need to address the problem of ex-post flexibility, the
main issue is nevertheless a perceived (signal) or factual (commitment) lack of
reliability and/or a functioning domestic adjudication. The decisive point is that the
relationship between investors and states is hierarchical in the sense of Kreps (1990).
There are unforeseen contingencies that one cannot contract on but, in theory, the host
country could solve this problem through its own courts (or other mechanisms, e.g., by
building up a reputation for dealing with unforeseen contingencies in a fair and
consistent manner). Therefore, while balancing the interests of host countries and
investors is certainly important, the existence of a BIT must nevertheless either be
based on commitment or signalling properties of the BIT.

4.4.5.2          BITs as Development Aid
Domestic investors profit from the BITs concluded by their governments. The
incentives and motives of capital-exporting countries to conclude BITs therefore
appear to be obvious. Neumayer (2006) explores the question of with whom capital-
exporting countries conclude BITs in more detail using econometric tools. The self-
interest of capital-exporting countries as presumed in this book could be confirmed by
Neumayer (2006). If a developing country is economically attractive (by having a
larger market and by being more open to trade), it will indeed have more BITs.
Nevertheless, in addition, the author finds that "poorer countries have more, not fewer,
BITs, which suggests that developing countries needs also play a role in developed
countries' BIT programs. Good governance in the form of either democracy or human
rights protection does not matter."344 The author concludes that BIT programs fulfil -


343
      Eger (1995), p.146.
344
      Neumayer (2006), p.261.
The Economics of BITs                                                                  119

similar to aid programs - the dual functions of self-interest and foreign need.345 While
this research sheds some light on the patterns of BIT conclusion, it should be noted
that self-interest is substantially more important than using BITs as an indirect aid to
development. In addition, the aid aspect does not contradict the analysis of the
functioning of BITs as presented in the preceding sections. Quite to the contrary: the
results underline that countries seem to consider BITs as a suitable means to overcome
commitment problems.

4.4.5.3            Risk Aversion
The main motive for the existence of BITs may neither be commitment nor signalling,
but the existence of risk aversion. The idea works similarly to the function of
warranties. Consumers may perceive the warranty either as a quality signal or
consumers may be interested in the general protection of the warranty in the sense of
insurance.346 In sales contracts, such insurance is a Pareto-improvement if the buyer is
risk-averse while the seller is risk-neutral – the buyer is willing to pay a premium if the
seller takes on the risk that the product is defective. The seller will accept this if the
risk premium is higher than the expected value of replacing defective products. Can
this theory be applied to BITs? In the preceding sections, MNEs were assumed to be
risk-neutral. This assumption must be abandoned if insurance comes into play. While
abandoning this assumption is unproblematic, the application of the insurance
rationale to BITs is problematic for another reason: the price system. The host country
can certainly not charge a premium to investors that are protected by a BIT. The
remuneration for assuming the risk would therefore be indirect: a higher amount of
FDI. With regard to product warranties, the buyer may choose between a cheaper
product without the warranty and a more expensive product that has a warranty. The
quality of the product and specifically its default rate is known to the buyer. If this
default rate is rising, the seller can only charge a lower price and will have to charge
more for a warranty (as the revenue from the warranty must at least cover the expected
replacement costs). This mechanism can certainly not be applied to BITs. A higher
default rate would correspond to higher political risk in the respective host country.
While more political risk may well lead to less investment ("lower the price"), we do
not see that high-risk countries offer stronger protection in their BITs (and receiving
thereby additional investment that covers the cost of the "warranty" BIT). Put
differently, the price mechanism that adequately reflects the prices for higher risks
does not work in international investment law. Consequently, while investors and



345
      Neumayer (2006), p.262.
346
      See, e.g., Schäfer and Ott (2005), p.478ff.
120                                                                  The Economics of BITs

MNEs may very well be risk-averse, the insurance rationale is unlikely to be the
foundation for the existence of BITs.

4.5              Empirical Studies and the Functioning of BITs
Section 4.2.1 summarised the relevant empirical studies on the effects of BITs on FDI.
Some of these studies also touch (indirectly) on the functioning of BITs. Other
empirical research addresses the question of which country characteristics are related
to the conclusion of BITs. Four interesting points can be analysed using the existing
empirical research.
The first relates directly to correlations between institutional quality or political risk
and the number of signed BITs. Two papers address this question directly. Ginsburg
(2005) finds that countries that have adopted a BIT have higher levels of governance
(as measured by the World Bank governance indicators) than those that have not
adopted a BIT.347 Swenson (2005), on the contrary, asserts that those countries that
were viewed as risky exhibited a propensity for entering into a larger number of
BITs.348 The inconsistency of these results is not surprising as the analysis of this
chapter suggests that a linear correlation between BITs and institutional quality can
neither be expected with regard to signaling theory, nor with regard to commitment.
The second interesting question is how BITs and institutional quality interact with
regard to FDI. Note that this does not concern how FDI reacts to political risk or
institutional quality. Empirical research rather clearly indicates that there is a positive
relationship between FDI and good institutions349. The question is rather: how does
political risk in reference to a BIT influence FDI. Econometrically, the tool to answer
this question is the use of an interaction term. This implies inserting the product of
BITs and institutional quality or risk into the regression as an independent variable
(e.g., BIT*institutional quality). Finding a positive and significant relationship
between the interaction term BIT*institutional quality and FDI would indicate that
BITs have a higher impact when institutional quality is high, while a negative
relationship would obviously indicate the opposite. Two papers employ such a
variable: Neumayer and Spess (2005) find that "there is some limited evidence that
BITs function as substitutes for institutional quality, as in a few estimations the
interaction term between the accumulated number of BIT variable and institutional
quality is negative and statistically significant."350 In a similar direction, Busse,



347
      Ginsburg (2005), p.120.
348
      Swenson (2005), p.141f.
349
      Grosse and Trevino (2005).
350
      Neumayer and Spess (2005), p.1582.
The Economics of BITs                                                                121

Königer et al. (2008) state: "BITs may even substitute for weak local institutions"351 as
the interaction term between BITs and institutional development, proxied by the
political constraints on the executive branch, is negative and significant.
The third question relates to the research design and can be divided into two parts:
first, whether the research design is dyadic and second, whether a distinction is made
between signed and non-ratified BITs on the one hand and signed and ratified BITs on
the other hand. Hallward-Driemeier (2003) uses a dyadic research design, which
means that she is only looking at the effects of BITs on the specific investment inflows
from the respective treaty partner and not on the investment flows from all other
countries. As her study finds no or only marginal effect of BITs on FDI, it has been
criticised as ignoring the signalling effect.352 Aisbett (2007) addresses the question of
the existence of a signalling effect by comparing a dyadic with a cumulative research
design and finds no signalling effect. However, studies that use a cumulative approach,
like Neumayer and Spess (2005), did find a positive relationship between BITs and
FDI, which indicates the existence of a signalling effect. However, this effect could
alternatively be explained by the fact that investment from one source (that is
protected by the BIT) attracts additional investment from third countries (that may not
be protected), for example through suppliers, but also through the existence of
knowledge-spillover and agglomeration rents.
With regard to the distinction between signed and non-ratified as compared to signed
and ratified BITs, Egger and Pfaffermayr (2004) find that a positive effect of just
signing, which they call the anticipation effect, is not significant in most
specifications, leading them "to conclude that the advantages to simply signing a BIT
are inconsequential."353 Considering the discussion on the commitment game and the
signalling game, this result is not surprising.

4.6               Discussion and Conclusions
Despite the perceived weakness of international law, Bilateral Investment Treaties
appear to attract FDI at the margin. This chapter analysed the economic foundations of
this effect by using the recent insights of the literature on international law and
economics. Prerequisite for international law (including BITs) to influence the
behaviour of the relevant actors is an impact on the payoffs. Three factors have been
identified to do so, namely reciprocity, retaliation and reputation. While the former
two certainly play an important role in many fields of international law, for example in



351
      Busse, Königer et al. (2008), p.17.
352
      Neumayer and Spess (2005), p.1582.
353
      Egger and Pfaffermayr (2004), p.79.
122                                                                  The Economics of BITs

trade law, it was argued that, in the case of international investment law, reputation
plays the most important role. However, and particular to international law, in the case
of BITs, the reputation mechanism is “backed up” in the short run by non-reputational
costs, more specifically the possibility of enforcing arbitration awards against assets in
third countries.
Two explanations have been put forward in the literature to describe the function of
BITs, namely commitment and signalling. Each of these approaches has been
discussed on the basis of a simple game-theoretic model. Structurally, the two
approaches differ in the existence of complete vs. asymmetric information.
Unfortunately, the empirical research did not shed much light on the question of
whether we are rather dealing with a signal or a commitment device. The problem is
that institutional quality will not necessarily be related to the commitment or the
signalling function. An exception could be an independent judiciary or political
constraints in the following sense: if either of the two exist (and MNEs know that or
attach a high probability to this), expropriating is not an option to start with –
consequently, a BIT will neither be used as a commitment device nor as a signal (as
MNEs will also invest absent the signal). This hypothesis readily explains why highly-
developed countries, despite extensive FDI flows, do not conclude BITs with each
other and also with the empirical observations by Neumayer and Spess (2005) and
Busse, Königer et al. (2008) that countries with lower institutional quality can profit
more strongly from a BIT. Yet, below a certain threshold, the connection between
institutional quality and BIT-signing remains empirically dubious. This is not
surprising as in neither of the two models (signalling or commitment) do countries
with high institutional quality have the incentive to sign BITs.
How do the commitment and the signalling theory relate to each other? A signal can
only work when (apart from other factors like observability, low prior beliefs etc.)
there are different costs associated with BITs for the reliable and the unreliable type.
The costs will only differ when arbitration creates costs which will, inter alia, only be
the case when the MNE has an incentive to initiate arbitration in the case of
expropriation. This is exactly the place where the two approaches overlap. Put
differently, it is the countries which can use the BIT as a commitment device or
happen to be reliable from the start (but are not perceived as such by third parties) that
could use BITs as a signal. Consequently, it is important to note that the approaches do
not exclude each other, but in a way condition each other (and can never be completely
separated). The commitment model is therefore a useful tool to analyse the interaction
between MNEs and host countries, as the mechanism is also a prerequisite for the
signalling model. The hypotheses developed in section 4.4.3.6 summarised the main
points under which HCs must anticipate sanctions in the case of expropriation and are
hence in principle valid even with regard to the signalling function of BITs. The
The Economics of BITs                                                                123

question then is which of the two groups use the mechanism more –the countries that
can commit themselves or the countries that already are committed - assuming the
existence of asymmetric information.
However, it is precisely the existence of asymmetric information that raises concerns
on the validity of the signalling model: clearly, if the relevant asymmetric information
could be easily measured, it would not be asymmetric information in the first place.
This makes the signalling of political risk and institutional quality unlikely as
information on these two aspects is readily available. Further, the overall pattern of
BIT conclusion makes the BIT a very noisy signal. Most countries in the world have
concluded at least one BIT, while the actual number of BITs varies vastly between
countries. In addition, the signalling strength of a BIT should vary depending on the
treaty partner. If BITs are mainly a signal, why not conclude a BIT with only the three
or four major capital-exporting countries in the world? It nevertheless seems likely that
a certain signalling function of BITs exists, especially with regard to the hidden
intentions of host countries. Obviously, MNEs will use all information they can obtain
to make informed investment decision. Yet, BITs as signals must be interpreted with
care and particularly BITs signed and ratified a long time ago might not convey much
information. To sum up, while BITs may have a certain signalling function, a
separating equilibrium that clearly reveals the type is not a good approximation of
reality. Finally, the signalling function may only have a temporary effect, while the
commitment function will last. In that sense, a BIT will be a very costly signal in the
long run.
5                 BITs and Institutional Competition

Countries are competing for capital. Bilateral Investment Treaties are a tool in this
competition. The use of laws and regulations to attract foreign investment has been
termed institutional competition. Bilateral Investment Treaties have so far not been
directly discussed in light of the different theories of institutional competition.
However, the famous prisoner’s dilemma thesis as put forward by Guzman (1998) and
the literature based on this article can be subsumed under the theory of institutional
competition. The literature on institutional competition suggests that this competition
can be either beneficial or harmful for the involved countries. The important question
asked here is therefore: are BITs better understood as part of a ruinous competition for
capital or, to the contrary, can BITs help overcome domestic inefficiencies and foster
the positive effects of institutional competition? This question will be addressed with a
focus on developing countries.
The chapter proceeds as follows: section 5.1 describes the notion of efficiency in the
context of international investment law to provide a useful frame of reference. Section
5.2 describes the prisoner’s dilemma hypothesis and subsumes this hypothesis under
the theory of institutional competition. The examples of tax competition and
environmental competition will be discussed to highlight the potential benefits and
costs of institutional competition. Section 5.3 illustrates to what extent international
investment law may interact with the concept of institutional competition. This implies
the consideration of political economy elements. As international investment law has
thus far rarely been analysed from a political economy point of view, this paper takes
recourse to the well-established political economy of international trade law. This
approach entails a deviation from the unitary actor assumptions as employed in the
previous chapters. It will be argued that BITs can help overcome domestic
inefficiencies, but may at the same time pose a threat of an underprovision of domestic
regulation. Taking this into consideration, section 5.4 analyses the relevant provision
of BITs and reviews the case law focusing on the specific example of environmental
regulation. Section 5.5 discusses the main results and provides a conclusion.

5.1               Efficiency and BITs
The economic analysis of law employs the notion of efficiency as a normative
criterion. Two concepts of efficiency are usually applied: Pareto efficiency and
Kaldor-Hicks efficiency.354 A situation can be called Pareto-efficient when no one can



354
      See Posner (2007), p.10ff.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_5,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
BITs and Institutional Competition                                                    125

be made better off without making somebody else worse off. Similarly, the notion of a
Pareto-improvement is employed if the relevant transaction or reform is beneficial to
the parties involved. The Kaldor-Hicks criterion, on the other hand, does not require
that there are only winners to a reform: it suffices that the winners could compensate
the losers (which does not necessitate that the winners actually do).355 Even without
compensation, the use of the Kaldor-Hicks criterion can be justified on the grounds
that in the long run, everyone in society will be better off as compared to a society that
only employs the Pareto criterion. This section argues that the Kaldor-Hicks criterium
is not well-suited to analyse international investment law. Further, the application of
the Pareto principle also raises a number of concerns that forces us to limit our
attention to Pareto improvements within a subgroup of countries, namely the
developing countries.
When applying the concept of efficiency to the area of international investment law,
one crucial question relates to the level of consideration, namely: what parties will be
included in the efficiency considerations? To start with, Bilateral Investment Treaties
are treaties between two states. Assuming that states act in the best interests of their
citizens and that no party threatens to use (military) force, the treaty, similar to a
contract, should be Pareto-efficient for the two states concerned. If a capital-importing
and a capital-exporting state are assumed, the former must weigh the benefits and costs
of higher investment protection. Idealy, the BIT will help overcome the commitment
problem and therefore attract additional FDI. The additional investment will increase
taxes, create spillover-effects and therefore contribute to the development of the
country. The capital-importing state should consequently offer a level of investment
protection where the marginal benefits equal the marginal costs. The capital-exporting
state should accept the offer when the treaty implies higher investment protection to its
MNEs as compared to a situation without the treaty. In sum, both parties are better off.
The issue is then to find the right balance between commitment and flexibility (in the
presence of contracting imperfections). This approach is taken by Van Aaken (2009)
who argues, with regard to BITs, that "(t)he point of optimality between commitment
and flexibility still needs to be found."356 Possible reasons are that the treaties lack
adequate mechanisms that ensure flexibility or that tribunals interpret the terms of the
treaty too strictly.357
Therefore, it appears that a BIT is virtually by definition Pareto-optimal as long as
only the welfare of the two contracting parties is concerned. However, this argument



355
      Schäfer and Ott (2005), p.32.
356
      Van Aaken (2009), p.31.
357
      See Van Aaken (2009), p.16.
126                                                                     BITs and Institutional Competition

would neglect the fact that a BIT may have consequences for third parties.358
Consequently, this approach is unsatisfying as the whole strategic interaction between
developing countries is based on this external effect. In this respect, it is also important
to remember that Pareto efficiency is actually a measure of social welfare. The
question must therefore be whether BITs are efficient when looking at the global
welfare or a larger subgroup of states and parties. Yet, defining boundaries on the unit
of analysis is difficult.359 An alternative approach would therefore be to look at global
welfare and apply the Kaldor-Hicks instead of the Pareto criterion. In the case of
investment protection, the BIT regime strengthens property rights and is in this sense
very likely efficient. However, the efficiency concept applied here cannot be the
Kaldor-Hicks criterion for the following reasons: investment flows are rather stable
over time and generally flow from MNEs in developed countries to developing
countries. The basic justification of the Kaldor-Hicks criterion, a compensation of the
loser in the long run, will therefore never occur.
Given these myriad of problems with the efficiency concept in international
investment law, the following approach is chosen: first, the specific problem is indeed
that the BIT has an external effect; the additional investment will fully or partly be a
diversion of investment from another developing country. Compensation in the long
run as implicit in the Kaldor-Hicks criterion is unlikely. Therefore, this section is
concerned with whether the BIT regime is beneficial for developing countries as a
group. Second, governments may not always act in the best interest of their citizens.
Hence, where possible, the utility of subgroups of society will be taken into account.
Finally, the Pareto criterion will not be applied in the sense of Pareto optimality (as
this would also imply a global perspective) but rather in the sense of a Pareto
improvement or Pareto superiority. The modest question to be asked is consequently:
is the BIT regime as such a Pareto-improvement for developing states?360


358
      The following sections are basically a discussion as to whether this external effect is negative or
      positive. However, the discussion here rather relates to the question whether to consider this effect
      at all. A different aspect could be emphasized here: first, think about FDI as flows: the decision
      where to invest is taken every period new and the fact that a country has received this investment in
      one period entails no entitlement to receive investment in the following period. From this point of
      view, there is actually no effect on the third party of the investment decision. Alternatively, one
      might focus on FDI as stocks: increasing the stock in one country usually goes together with a
      decreasing stock in another country. In that sense, any investment decision (and indirectly also the
      BIT) has an external effect on third parties that needs to be considered.
359
      Which is probably the reason that most approaches in international law and economics have so far
      been positive and not normative in nature. A global perspective is for example chosen by Posner
      (2006).
360
      Obviously, a different approach would be to disregard the notion of efficiency given its lack of
      definition for the international level and simply ask whether the BIT regime is “beneficial” to all or
      a subgroup of countries. Nevertheless, this would only solve a semantic problem and the definition
BITs and Institutional Competition                                                                   127

5.2               Institutional Competition

5.2.1             BITs and the Prisoner’s Dilemma
Probably one of the most widely cited articles in the area of international investment
law is entitled 'Why LDCs Sign Treaties That Hurt Them: Explaining the Popularity of
Bilateral Investment Treaties' by published Andrew Guzman in 1998.361 The article
focuses on why least developed countries (“LDCs”) sign Bilateral Investment Treaties
even though these treaties may have a detrimental effect on their welfare. The starting
point of the analysis by Guzman is the observation that developing countries have
frequently rejected committing themselves to multilateral rules with high protection
standards (as embodied by the Hull Rule), but have, at the same time, concluded
bilateral agreements (BITs) that reflect exactly these high standards (or even higher
ones). The reason lies, according to Guzman, in the prisoner's dilemma situation in
which developing countries find themselves. The author claims that "although an
individual country has a strong incentive to negotiate with and offer concessions to
potential investors – thereby making itself a more attractive location relative to other
potential hosts – developing countries as a group are likely to benefit from forcing
investors to enter contracts with host countries that cannot be enforced in an
international forum, thereby giving the host a much greater ability to extract value
from the investment."362 Notably, Guzman’s argument does not imply a loss of
efficiency through BITs. Guzman writes: "In global terms, the most efficient outcome
is achieved if investment takes place where it will earn the greatest total return."363
Therefore, if BITs help to mitigate the time inconsistency problem (see section 3.3.1)
and the related issue of underinvestment, BITs do not pose a problem of efficiency and
may even enhance efficiency. The PD situation is consequently foremost a
distributional problem. The point Guzman emphasises is that the increase in
investment through BITs comes at the expense of other developing countries.364 In
addition, the concessions the BIT makes will shift the surplus of the additional FDI to
the investors (leaving the respective country with low or zero economic profit). In
sum, developing countries would be better off without any international legal
investment protection - yet, as in the cases of cartels, countries have the incentive to
defect from this strategy by signing BITs. This result rests on a number of
assumptions. One assumption, as Guzman acknowledges, concerns the elasticity of


      of the parties included in the considerations would require a discussion similar to the one in this
      section.
361
      Guzman (1998).
362
      Guzman (1998), p.643.
363
      Guzman (1998), p.666.
364
      Guzman (1998), p.670.
128                                                        BITs and Institutional Competition

investment with regard to legal protection: "Specifically, the above theoretical claims
are true only if the flow of investment into LDCs as a group is relatively insensitive to
the terms on which that investment is made as compared to the flow of investment into
a single developing country".365 In addition, the BIT must indeed be able to constrain
the ability of the host country to extract surplus from the investment.
The thesis that BITs are a tool in the competition for capital was later backed up
empirically by Elkins, Guzman et al. (2006). Evidence for alternative theories of BIT
conclusion, like coercion, learning or cultural explanations based on emulation, was
less significant. More specifically, the authors find that countries sign more BITs when
their competitors have also signed BITs. Three alternative measures are used to define
competitors: countries with the same export trade relationships, countries that export
the same basket of goods and countries that have similar educational and
infrastructural resources. All three measures turn out to be significant. Coercion is
proxied through the use of IMF credits and can be shown to have some influence on
the decision of host states. Cultural factors (such as language or religion) seem to play
only a minor role with regard to BIT conclusion. Learning from success is modelled
by Elkins, Guzman et al. through the inclusion of a variable that captures the
relationship of the number of BITs of a specific country in the preceding five years
and its FDI inflows. This variable had some explanatory power, yet does not seem to
play an equally crucial role as the competitive pressures to catch a share of FDI in light
of the behaviour of potential competitors. While the results by Elkins, Guzman et al.
clearly indicate that BITs are a tool in a competitive market, the distributional effects
are not elucidated.
The argument of Guzman has been addressed, formalised and critiqued by Bubb and
Rose-Ackerman (2007). The authors argue that a prisoner’s dilemma situation is not
necessary to understand the behaviour of LDCs. The authors present a model that
"makes clear that poor countries may be caught in a prisoner's dilemma where they
have incentives to sign BITs but are made worse off with a BITs regime. However it
also illustrates that such a result is by no means necessary. It depends on the trade-off
between FDI creation and sharing of gains between investors and host countries."366
Consequently, as compared to Guzman (1998), the authors are much more critical with
regard to the hypothesis that investment to LDCs as a group is inelastic. On one point,
Bubb and Rose-Ackerman (2007) and Guzman (1998) clearly agree: "The net effect of
FDI on developing countries under a BIT regime compared with traditional customary
law is an empirical question."367 The prisoner's dilemma hypothesis is also analysed


365
      Guzman (1998), p.674.
366
      Bubb and Rose-Ackerman (2007), p.302.
367
      Guzman (1998), p.305.
BITs and Institutional Competition                                                           129

further by Engel (2008) in a comment to Elkins, Guzman et al. (2006). The author
hypothesises a situation where policy makers maximise internal political support, not
wealth. Benefits to the country stem from being ahead of similar countries (while the
sovereignty cost of the BIT is expressed as an absolute value). Based on these
assumptions, it is possible to lay out different structures of the game that are different
from the prisoner's dilemma game (e.g., the chicken game, asymmetric games where
states have different preferences, two period games) that result in equilibria where
only a fraction of countries sign BITs– without having to give up the hypothesis that
investment flows to LDCs as a group are inelastic. Yet, if the relative position of the
country enters the welfare function, it is difficult to make a prediction on the actual
outcome of the game.
To sum up, the validity of the argument by Guzman (1998) strongly depends on the
elasticity of investment flows. Put differently, the prisoner's dilemma hypothesis is
certainly not the only possible explanation for the behaviour of LDCs and may not
necessarily be the best description of the strategic situation that LDCs face vis-à-vis
each other. As Sykes (2007) notes: "An alternative hypothesis, of course, is that
developing countries came to realise that their strategy in the United Nations was a
mistake, and that retaining an opportunity to expropriate raised their cost of capital by
more than the gains from any expected expropriation (perhaps because of the risk
aversion of investors). Although Guzman’s thesis is surely an intriguing one, it likely
does not represent the final word on the matter."368

5.2.2            The Economics of Institutional Competition
The discussion on the strategic situation between host countries fits into the broader
framework of the discussion on the merits of institutional competition. Institutional
competition is a widely discussed phenomenon in (institutional) economics and legal
studies.369 Different terms have been attributed to this phenomenon, emphasizing
slightly different aspects, for example regulatory competition370 and (new) systems
competition.371 There exists neither consensus among scholars with regard to the
efficiency and welfare effects of institutional competition nor a well-established
definition of the term. A short but very useful definition was put forward by Siebert
and Koop (1990): "Institutional Competition is the competition of the immobile for the
mobile factor."372 The history on the economic discussion of institutional competition,

368
      Sykes (2007), p.811f.
369
      A lucid and balanced recent overview on the topic with special attention to the question of
      regulatory harmonisation vs. centralisation is provided by Behrens (2009).
370
      Romano (2005).
371
      Sinn (2003).
372
      Siebert and Koop (1990), p.4.
130                                                                   BITs and Institutional Competition

which started mainly with the famous Tiebout model in 1956, shall not be reiterated
here.373 It suffices to note that there are different forms of institutional competition
and, as already mentioned, different assessments on the impact of institutional
competition. At least four manifestations of institutional competition can be
distinguished: (1) yardstick competition, (2) indirect institutional competition through
trade, (3) locational competition and (4) institutional competition through the choice of
law. Yardstick competition relates to the ability of constituents to compare the
performance of their domestic governments with that of foreign governments and
pressure for changes.374 Indirect institutional competition through trade describes the
fact that the consumer's decision for or against a specific good is always also an
implicit decision for or against the regulatory framework under which the good was
produced.375 Locational competition evidently describes the choice of location of
mobile factors where institutions may play an important role. Lastly, institutional
competition through the choice of law does not require factor mobility and describes a
situation where users can choose between different legal arrangements without
changing their location. The form of institutional competition with regard to BITs can
be subsumed under the label of locational competition. Countries use the "institution"
BIT to compete for capital.
Economists that have analysed the functioning of institutional competition diverge in
their estimation on the usefulness of the concept. A rather negative assessment on
institutional competition has been developed by Sinn (2003). The author develops a
concept which he terms the Selection Principle: “The Selection Principle says that
governments have taken over all those activities which the private market has proved
to be unable to carry out. Because the state is a stopgap which fills the empty market
niches and corrects the failures of existing markets, it cannot be expected that the
reintroduction of the market by the back door of systems competition will lead to a
reasonable allocation result. Instead, it must be feared that the failures that originally
caused the government to take action will show up again at the higher level of
government competition.”376
Other authors have emphasised the positive effect of market forces of institutional
competition that may lead to overall better results and reduce inefficiencies of the
domestic political process. Representative of this view is the evolutionary approach
advanced by Kerber and Vanberg (1994). This view emphasises the knowledge-


373
      Tiebout (1956).
374
      For an explanation of the term and application to the EU, see Salmon (2003).
375
      This type of competition is especially relevant in the European Union in relation to the country of
      origin principle.
376
      Sinn (2003), p.6. See also Eger and Wagener (2009), p.361f.
BITs and Institutional Competition                                                   131

creating process of competition in the sense of Schumpeter and Hayek.377 A key
difference to the more sceptical approaches as exemplified by Sinn (2003) is, apart
from the importance of knowledge-creation, the assumption on the nature and
incentives of governments. The public choice literature highlights the prevalence of
deficiencies in the domestic political process. Based on the influential writings of
James Buchanan and Gordon Tullock378 is the argument that governments do not
necessarily maximise the social welfare, but rather concede to the rent-seeking
activities of subgroups of society. A major advantage of institutional competition is
consequently a control function: as the mobile factors prefer efficient regimes over
inefficient ones, their location decision will induce governments to offer an efficient
set of regulations. This logic is clearly analogous to markets where the competition
forces companies to act efficiently. In the legal literature, the market analogy has
especially found its way into the discussion on the regulatory competition in corporate
law and securities regulation. For example, with regard to the latter, Romano (2001)
lists as the benefits of regulatory competition that it "reduces the possibility that a
regulator will be able to transfer wealth across different regulated entities or
redistribute wealth from the regulated sector to preferred individuals or
organizations"379, that it "more quickly corrects for policy mistakes than a single
regulator can"380 and that it "can be expected to foster innovation, since the feedback
mechanism of interjurisdictional flows that reduces the possibility of regulatory error
also provides an incentive for regulators to improve their regimes".381 This view is also
influential among the forerunners of law and economics. As Gary Becker states:
"Competition among nations tends to produce a race to the top rather than a race to the
bottom by limiting the ability of powerful and voracious groups and politicians in each
nation to impose their will at the expense of the interests of the vast majority of their
populations."382

5.2.2.1           Tax Competition
Institutional competition may affect a wide array of issue areas. Put differently,
governments may use different institutional parameters to compete for FDI. Most
evident and probably most widely discussed is the use of tax rates (or subsidies) to
attract FDI. Another example may be the lowering of environmental standards. Both
discussions reflect the general discussion on institutional competition. With regard to


377
      Kerber and Vanberg (1994), p.198.
378
      See, e.g., Buchanan and Tullock (1962).
379
      Romano (2001), p.5.
380
      Romano (2001), p.5.
381
      Romano (2001), p.6.
382
      Becker (1998), p.22.
132                                                           BITs and Institutional Competition

tax competition, a number of authors have raised the concern that it might be
detrimental for the involved countries. Oates summarised this concern as early as
1972: “The result of tax competition may well be a tendency toward less than efficient
levels of output of local services. In an attempt to keep taxes low to attract business
investment, local officials may hold spending below those levels for which marginal
benefits equal marginal costs, particularly for those programs that do not offer direct
benefits to local business.”383 This concern has been formalised, most notably by
Zodrow and Mieszkowski (1986). The authors conclude that tax competition will lead
to an underprovision of public goods. The claim is logically based on a Bertrand-like
competition. As Wilson (1999) emphasises: "A central message of the tax competition
literature is that independent governments engage in wasteful competition for scarce
capital through reductions in tax rates and public expenditure levels."384 At the core of
the argument lies the fiscal externality that arises because countries do not consider the
effect of their tax rate on the tax base of other countries.
A different view has recently been expressed by Baldwin and Krugman (2004). The
authors depart from the traditional models of tax competition by introducing a model
where the investment in certain regions produces agglomeration rents as a result of
good infrastructure, accumulated experience etc. These agglomeration rents are
valuable to investors and as a result, tax competition may not lead to an inefficient race
to the bottom but "may lead same-sized nations to have different equilibrium tax
rates".385 Also, more optimistic views on tax competition may prevail when states face
internal inefficiencies which may be overcome through institutional competition. In
general, this may be the case when governments have interests other than the
maximisation of social welfare. The economic foundation of this argument is usually
attributed to Brennan and Buchanan (1980). The basic idea is that Leviathan
governments act to maximise tax revenues and that it is beneficial to constrain the
government to do so. If the government does not face any domestic fiscal constraints
on the taxing power, intergovernmental competition may be a substitute for these
restraints.386 This argument is also applied by McLure (1986) who argues that the
“likely benefits of reducing tax competition are relatively slight”, while “the benefits
of tax competition are potentially quite important”.387 The benefits he identifies in
favour of tax competition are the same as the benefits of mitigating the problems posed
through Leviathan governments. Different specifications of the Leviathan model have
subsequently been used to analyse tax competition. In general, the introduction of non-


383
      Oates (1972), p.143 as cited in Wilson (1999), p.269.
384
      Wilson (1999), p.269.
385
      Baldwin and Krugman (2004), p.19.
386
      Brennan and Buchanan (1980), p.184.
387
      McLure (1986), p.346.
BITs and Institutional Competition                                                    133

benevolent interests of policy-makers leaves a welfare-enhancing role for tax
competition. Some contributions predict an ambiguous net effect on welfare. Edwards
and Keen (1996) argued that decision-makers that are not welfare maximizing face a
trade-off between social welfare (interests of voters) and creating rents for themselves.
The introduction of tax competition will in this setting have two different effects on
welfare: while the rent-seeking behaviour of the politicians is mitigated (implying
positive welfare effects), providing the public good will equally become more costly
(implying negative welfare effects). A recent contribution by Eggert and Sørensen
(2008) goes into a similar direction: absent tax competition, politicians have the
incentive to create efficiency-distorting rents for the public sector employees.
Introduction of tax competition will mitigate this problem, but may at the same time
cause an underprovision of public goods (similar to the fiscal externality argument).
The authors note that tax competition is "a double edged sword" that may reduce rent-
seeking, but "that also tends to distort the supply of public goods."388 The conclusion is
that "to a certain point tax competition may play a useful efficiency-enhancing role,
but if it becomes too intense it is likely to be welfare-reducing."389
The analysis underlines that the literature cannot be clearly divided into two strands
with the assumption of benevolent governments leading necessarily to inefficient
outcomes of tax competition on the one hand and the assumption of self-interested or
Leviathan governments leading inevitably to efficient results of tax competition on the
other hand. Janeba (1998) is an example of where inefficiencies (in the form of
subsidies to domestic firms) are persistent as a result of imperfect competition, but
these inefficiencies can be overcome through tax competition when governments
cannot discriminate between domestic and foreign firms. In this case, the government
is benevolent in nature and yet, there is a welfare-enhancing role for tax competition.
Nevertheless, the control function is at the core of optimistic accounts of institutional
competition and it is clear that the control function loses its impact when ignoring the
control problem through the assumption of benevolent governments.

5.2.2.2           Environmental Competition
With regard to environmental competition, the theoretical research is likewise
inconclusive. In general, the focus of the literature is less on taming Leviathan
governments but more on the internal inefficiencies or inefficiencies of the
competitive process. Nevertheless, the basic functioning is similar to the case of tax
competition. A number of scholars have expressed the concern that institutional



388
      Eggert and Sørensen (2008), p.1157.
389
      Eggert and Sørensen (2008), p.1157.
134                                                        BITs and Institutional Competition

competition may lead to a race to the bottom on environmental standards. The logic of
the race to the bottom hypothesis is simple: as strict environmental policies raise costs
for companies, governments will opt for lax standards to attract capital, resulting in, in
the extreme case, a race to the bottom.390 According to Revesz (1992), this idea
became commonplace following two articles by Richard B. Stewart that were used by
the US Congress to justify federal environmental statutes.391 In one of these articles,
Stewart (1977) argues: “Given the mobility of industry and commerce, any individual
state or community may rationally decline unilaterally to adopt high environmental
standards that entail substantial costs for industry and obstacles to economic
development for fear that the resulting environmental gains will be more than offset by
movement of capital to other areas with lower standards.”392
An influential model by Oates and Schwab (1988) asserts that the competition among
jurisdictions will lead governments to set environmental standards where the trade-off
between additional investment and environmental protection is optimal. However,
Kunce and Shogren (2005) posit that the efficiency result of Oates and Schwab (1988)
is driven by the assumption that the pollution rents (the economic rents for using the
environment as a production factor) are completely captured by the domestic wage
earners. It is also noteworthy that the efficiency result does not hold when capital taxes
are (inefficiently) high or when the government does not maximise social welfare, but
instead maximises its budget.
The model introduced by Wellisch (1995) accounts for the fact that pollution rents are
not fully captured by domestic constituents but partly captured by the polluting
industry, usually owned by non-residents (as MNEs locate where pollution rents are
the highest). Using certain policy instruments (namely direct controls) the result may
actually be overprotection of the environment. The following quote by Wellisch
(1995) explains the intuition behind the model: "If direct controls are used, then the
implicit factor reward on pollution is left with the firms which are not (entirely) in the
hands of local residents. Residents do not share the rents from environmental pollution
but have to bear the entire burden. Therefore, the environmental agency has too strong
an incentive to protect the environment. If emission taxes are used and the tax
revenues are distributed among local residents, the implicit rent from local pollution is
also internalised and the environmental agency has incentives to choose a socially
efficient pollution level."393 The approach of Markusen, Morey et al. (1995) yields a
similar result. The authors find that under some circumstances, namely only a small


390
      See Wilson (1996) for a discussion of this issue.
391
      Revesz (1992), p.1210f.
392
      Stewart (1977), p.1211f.
393
      Wellisch (1995), p.292.
BITs and Institutional Competition                                                    135

disutility from pollution, two competing jurisdictions will try to attract FDI by
lowering environmental standards. Quite to the contrary, if disutility of pollution is
substantial, the jurisdictions may even increase their environmental standards in an
attempt to drive production of the good into the other jurisdiction (while still being
able to consume the good through exports). The authors refer to what is known as the
NIMBY-strategy: not in my backyard.

5.2.3            Discussion
The articles and models discussed constitute by no means an exhaustive account of the
literature, but rather are a representative cross-section. The theoretical discussion on
institutional competition is voluminous and diffuse. General conclusions with regard
to institutional competition are therefore difficult to draw. Most commentators share
this view. Kunce and Shogren (2005) state in this respect that "the setting of tax policy
and environmental standards remains a central but unresolved public policy
question"394 while Oates (2001) diagnoses with regard to the issue of environmental
protection that the “theoretical literature is, in my view, inconclusive on this issue"395
and Wilson (1996) declares that "neither the theoretical nor empirical work points
clearly in one direction."396
The indecisiveness of the matter may call for more empirical research, but the
empirical literature on the topic is equally inconclusive.397 The problem here is that the
arguments of each side with regard to social welfare are hardly verifiable. Empirical
evidence may find that tax competition leads to a reduction of tax rates. However,
these findings cannot shed light on the welfare question as it is questionable whether
tax rates were at the efficient level in the first place. With regard to standards, like
environmental standards, this evaluation is likely even more complicated. One point
that can nevertheless be derived from the diverse literature on institutional competition
is that a positive view on institutional competition usually requires some form of
domestic inefficiency: this inefficiency is mostly the result of a non-benevolent
government. However, even the assumption of a self-interested government does not
necessarily lead to an efficiency-enhancing effect of institutional competition. The
distortions of institutional competition identified by many scholars may nevertheless
arise and outweigh the positive control effects of institutional competition. It is also
important to understand that the assumption of a self-interested (or Leviathan)
government is not a necessary assumption to reject the race to the bottom hypothesis.


394
      Kunce and Shogren (2005), p.212.
395
      Oates (2001), p.8.
396
      Wilson (1996), p.408.
397
      See Oates (2001), p.13ff.
136                                                        BITs and Institutional Competition

Other factors, like agglomeration rents, may enable countries to set tax rates that are
not subject to a downward pressure by institutional competition. It should also be
noted that the term race to the bottom relates to an extreme case and therefore any
model that suggests such a condition requires very strong assumptions, mainly
regarding the sensitivity of investment to policy choices. Therefore, in the following
sections, the potential underprovision of public goods will be discussed instead of the
extreme case of a race to the bottom.
Finally, it seems that none of the approaches is clearly true while ruling out the others.
Rather, institutional competition has countervailing effects that need to be balanced.
This idea becomes especially apparent in the models of Edwards and Keen (1996) and
Wellisch (2000). The latter points out regarding tax competition: “The gains derived
so far of an increased supply of local public goods must be contrasted with the
disadvantage of less competition among self-interested policy makers wasting tax
revenues.”398

5.3                BITs in the Context of Institutional Competition
How do BITs relate to the different theories of institutional competition? To begin, it
is evident that neither the discussion on tax competition nor ecological competition can
fully capture the use of BITs in the competition for capital, as BITs cover potentially
almost every area of domestic policy-making. The purpose of the preceding sections
was rather to illustrate that the theoretical literature on institutional competition is
divided. This is even true without giving up the assumption of benevolent
governments (and ignoring the importance of learning effects).
The equilibria where states choose inefficiently low taxes or lax regulation as a result
of institutional competition for capital are in essence prisoner’s dilemma situations for
the involved states.399 In that sense, Guzman (1998) basically subsumes the existence
of a detrimental institutional competition under a prisoner's dilemma scheme.
Guzman’s argument certainly holds as long as investment to developing countries is
indeed quite inelastic. Obviously, if investment is inelastic, any measure that limits the
behaviour of host states must lead to a prisoner’s dilemma. In that case, it is even
irrelevant if governments do not intend to maximise social welfare. What is important
is simply if the measure (in our case, the BIT) limits in any way the ability of




398
      Wellisch (2000), p.74.
399
      For the tax case, see Wellisch (2000), p.64.
BITs and Institutional Competition                                                              137

developing countries to redistribute rents at all to domestic constituents (that are part
of the welfare function).400
The following will deviate from these assumptions and assume that additional
investment and/or agglomeration rents may theoretically leave a surplus of the
transaction that may be captured by the host country. The implication of this
assumption is not necessarily that there will be no prisoner’s dilemma. The
consequences of the competition for capital may still lead (developing) countries to
end up in a prisoner’s dilemma. Yet, the opposite may also be true and institutional
competition in the form of BITs may help overcome domestic inefficiencies and
countries will, after all, profit from the competition for capital. The question is
consequently: can BITs be understood as a manifestation of an institutional
competition that is harmful to developing countries? Or do BITs, on the other hand,
help overcome internal deficiencies, for example, by aligning the interests of
governments and constituents? A plausible hypothesis is therefore that the competition
for capital reduces the ability to redistribute rents from the MNE to preferred groups,
but at the same time reduces the ability of governments to take measures that are
welfare-enhancing.
A number of aspects can be examined in this respect. First, it will be discussed
whether domestic inefficiencies regarding investment protection can be expected in the
first place. A useful frame of reference here is the (well-established) political economy
of international trade. Second, the underprovision hypothesis regarding investment
protection will be illuminated. Finally, the content of BITs and the existing case law
(using the example of environmental regulation) will be discussed in light of these
theories.

5.3.1             BITs and Self-Interested Governments

5.3.1.1           Political Economy and International (Trade) Law
The simplifying assumption that states act as unitary actors was a useful
approximation while analyzing the economic functioning of BITs. It may, however, be
problematic for explaining the outcomes of international investment policy making.
With regard to regulation in general, a deviation of the interests of governments from




400
      Think about the extreme case where investment protection does not exist at all and developing
      countries would always expropriate 100% of the investment inflows, but these inflows remain
      stable despite the lack of protection. Any commitment that limits the discretion of developing
      countries would then be a prisoner’s dilemma.
138                                                              BITs and Institutional Competition

the social welfare has been observed by the theory of regulation for a long time.401 For
example, when investment is sunk, governmental opportunism may result from a
shorter time horizon or the need to redistribute rents in favour of powerful interest
groups to ensure re-election.402 An interesting empirical underpinning of a similar idea,
namely that governments use regulation as a rent-extracting device, has been provided
by Djankov, La Porta et al. (2002). The authors analyse the behaviour of 85
governments concerning (domestic) regulation of firm entry (start-ups) and find that
"heavier regulation of entry is generally associated with greater corruption and a larger
unofficial economy, but not with better quality of private or public goods".403 The
authors conclude that these findings are hard to reconcile with the public interest
theories of regulation: "The principal beneficiaries appear to be the politicians and
bureaucrats themselves."404
Theories based on the political economy approach have especially been applied to
international trade law. The starting point of the analysis is the observation that the
economic case for trade barriers is weak.405 Many economists have long argued that a
unilateral abolishment of trade barriers is welfare-enhancing. Nevertheless, many if
not most countries engage in protectionist policies. The reason is that while free trade
may be welfare-enhancing, it also has distributional implications.406 Some interest
groups may consequently lose under a free trade regime. If politicians do not maximise
welfare, but rather their chances to be re-elected, the economic rents for some interest
groups may play a more important role in the domestic political process than the rents
for other groups. In the case of international trade, it appears that free trade may
especially hurt the interests of well-organised groups, for example producers, while the
winners of free trade, for example consumers, are less-well organised and face a
collective action problem that is exacerbated by the fact that the individual gains of
free trade may be low. As a result, there may be protectionist policies in favour of
certain producers despite their detrimental consequences for the overall social welfare.
Interest cleavages among interest groups regarding the trade regime have also been
identified between capital and labour and between different firms or different industry
sectors.407 A problem with that line of reasoning is that it can explain the existence of
trade barriers, but not why countries seek to simultaneously abolish the trade barriers


401
      According to Ekelund and Tollison (2001), p.357, the approach can already be found in the
      writings of Adam Smith and James Madison. Walter and Sen (2009) provide an overview of this
      approach with regard to international economics.
402
      See, for example, Spiller and Tommasi (2008).
403
      Djankov, La Porta et al. (2002), p.35.
404
      Djankov, La Porta et al. (2002), p.35.
405
      See Walter and Sen (2009), chapter 3.
406
      See, for example, WTO (2008), p.123.
407
      Walter and Sen (2009), p.70ff.
BITs and Institutional Competition                                                                  139

on the international level. Shouldn't the very same special interest groups prevent
governments from advancing international cooperation? The standard answer to this
question is the emphasis of the importance of international negotiation and the
reciprocity these negotiations entail.408 The idea is that governments know that a
unilateral tariff reduction is not feasible exactly because of the opposition of the
aforementioned interest groups, but at the same time, all governments are in a similar
position. By concluding a trade agreement, the governments can rally political support
from other groups of society, thereby making the tariff reduction possible in the first
place. As the WTO puts it: “Reciprocity is the key because it will convert each
nation’s exporters from bystanders in the tariff debate to opponents of protection
within their own nation.”409
A different political economy argument can be put forward concerning trade
agreements. This argument relates to time inconsistency problems between
governments and the electorate.410 The idea is that governments use trade agreements
to commit themselves to domestic interest groups in cases where optimal government
plans may be time inconsistent. Different specifications of this argument have been
developed.411 The core of the argument is that governments seek to avoid distortions
that may arise in the future. An illustrative example has been provided by Matsuyama
(1990).412 In his example, the government intends to liberalise a protected industry, but
the (representative) domestic firm in the protected industry is at the present time
inefficient. The threat to liberalise the industry in the future is supposed to induce the
domestic firm to invest. However, this threat may not be credible as, if the domestic
firm does not invest, the government may find it not optimal to liberalise at the pre-
announced, future date.413 One possible solution is that the "government might be able
to sign a contract with a third party (perhaps, the GATT) to make the cost of
postponement of liberalisation (or renewal of temporary protection) prohibitively
high."414

5.3.1.2           Political Economy and International Investment Law
The questions to be answered are whether these problems may also emerge in the
context of the regulation of FDI and whether institutional competition (in the form of


408
      Krugman and Obstfeld (2009), p.228ff. A very good overview on the issue can also be found in
      WTO (2007), p.55ff.
409
      WTO (2007), p.57.
410
      For a description of the approach see Bagwell and Staiger (2002), p.32ff.
411
      WTO (2007), p.59f.
412
      See also WTO (2007), p.59.
413
      Obviously, the structure here is similar to the time inconsistency game laid out in chapter 3.1.1.
414
      Matsuyama (1990), p.490.
140                                                       BITs and Institutional Competition

Bilateral Investment Treaties) can provide a remedy for these inefficiencies. The
literature on international trade and the theory of regulation have underlined that
governments may favour the interests of certain interest groups over others, that
governments may use regulation mainly to extract rents and that governments may
have problems committing themselves to domestic interest groups.
In the absence of a (domestic) time inconsistency problem, the interpretation of BITs
as solely the outcome of a complex interplay of different interest groups is
problematic. Certainly, it is reasonable to argue that different interest groups have
different interests concerning the regulation of FDI. A point in case is market-seeking
FDI. Domestic producers of the same or similar products will be strongly opposed to
the admission of MNEs. Even if the MNE is admitted, domestic producers might
nevertheless lobby for regulation that will ensure a competitive advantage. On the
other hand, consumers and potential employees that might profit from the presence of
the MNE are not a strong interest group. However, domestic suppliers can also gain
through the production facility of an MNE. In sum, the situation is as complicated as
the one of trade law and indeed, the interactions of interest groups may sometimes
result in strong protection of domestic industries and in other cases not. But why
should governments limit their discretion through a BIT? In the case of trade, the key
answer to this question was reciprocity. Of course, BITs are also reciprocal in nature,
but due to the asymmetric nature of investment flows in most cases, governments can
hardly use this reciprocity to convert the indifference of some interest groups into
support for FDI protection. Lacking reciprocity, the interest-group case for BITs in the
absence of a time inconsistency problem is weak. If the government is able to extract
rents from the political process, it is probably better off in maintaining its political
discretion than relinquishing it under a BIT.
Consequently, the time inconsistency problem cannot be ignored when analyzing the
political economy of BITs. This problem was examined in detail in chapter 3. Yet,
analogous to international trade law, the commitment problem might apply to the
private sector more than to foreign investors. Similar to the case in trade, a company or
sector may require protection at the present date, but will not enhance its efficiency if
the threat to withdraw the protection is not credible. A logical legal equivalent to
tariffs in the case of FDI regulation would be entry requirements. However, the pre-
establishment stage is usually not addressed by most BITs (with the notable exception
of US BITs) and has, as a practical matter, not been relevant to international
investment arbitrations. Indeed, BITs are mainly concerned with post-establishment
issues. Therefore, a BIT is not well suited to influence the present behaviour of the
private sector through making its threats credible. To be clear, governments may of
course be concerned with future rent-seeking of the private sector. Yet, it is precisely
the government that will profit from these rents – again, there will technically be no
BITs and Institutional Competition                                                                 141

time inconsistency problem and the situation is similar as described above. The
government will probably be better off not to limit its discretion.
Nevertheless, the time inconsistency problem vis-à-vis investors (as mentioned in
section 3.1.1) can indeed be exacerbated through the influence of interest groups and
short time horizons of political decision-makers. Of course, the time inconsistency
problem as such is the result of sunk investments and thus affects any host country.
Yet, we observe that that developed countries do not use international agreements to
protect their (substantial) investment flows among themselves. This is a strong
indicator that the time inconsistency problem has an important political economy
dimension (assuming that the patience of citizens and the value of spillover effects do
not differ systematically and substantially between populations). The simple repetition
model introduced in section 3.2.2 can highlight this point. The condition for a
cooperative equilibrium in a repeated game was derived as:

          CH
G t 1
          WH


If the interests of the government deviate from social welfare goals in the sense that
governments are less patient (lower ), value the expropriated asset higher (higher WH,
e.g. because the asset can be given to cronies or special interest groups) or profit less
from the investment (lower CH, e.g. because the positive spillover are widely diffused
and can therefore not easily be converted into political support), the BIT can be a
suitable tool to tie the government behaviour more closely to social welfare objectives
by inducing cooperative behaviour.415 Indeed, especially the non-discrimination
clauses in BITs seem to be suited to prevent governments from taking distributional
actions that would forgo future investments. These provisions will be considered in
more detail in section 5.4. Apart from the substantial ability of BITs to prevent self-
interested governments from acting opportunistically (disregarding social welfare), it
should nevertheless not be forgotten that the BIT will only work if the sanctioning
mechanism works. In addition, the expected sanction to the government must be high
relative to the benefits of the opportunistic behaviour.
To sum up: in contrast to the political economy of international trade law, BITs are
neither well-suited to reap gains of reciprocal commitment nor to mitigate domestic
commitment problems. Nevertheless, interest group politics and self-interested


415
      Obviously, the same mechanisms depicted in the preceding chapter determine the incentives of
      governments to actually sign a BIT. Specifically, the BIT will only be useful if it is a credible
      commitment.
142                                                                 BITs and Institutional Competition

politicians potentially play an important role for BITs. The institutional competition
through BITs may therefore indeed cure a domestic inefficiency and be welfare-
enhancing, especially if it can prevent governments from taking distributional actions
that forgo future spill-over effects for the immediate gain to well-organised interest
groups to the detriment of social welfare.

5.3.2             BITs and the Prisoner's Dilemma Revisited
Section 5.2.1 described the prisoner's dilemma hypothesis that was developed in the
context of BITs. In light of the discussion on institutional competition, two aspects can
be identified that may foster the occurrence of a PD situation. The first relates to the
underprovision hypothesis, which is the concern by some commentators that
institutional competition will lead to an underprovision of public goods. The second
aspect relates to the distributional implications of BITs and analyses to what extent
BITs are able to influence the distribution of surplus between countries and investors.

5.3.2.1           BITs and the Underprovision Hypothesis
The underprovision hypothesis embodies the potential detrimental effects of
institutional competition. With regard to BITs, the public good that is potentially
underprovided is regulation. Governments may refrain from welfare enhancing
regulation for fear of being sanctioned by international law - the structural similarity to
environmental competition is immediately apparent. Indeed, environmental issues
have been of concern in a number of arbitration proceedings. Yet, a BIT covers in
principle any kind of regulatory measure. An underprovision of regulation
consequently means over-deterrence with regard to enacting regulation.416 Is over-
deterrence a real concern in international law? After all, is it not the perceived
weakness of international law that arouses the most criticism? The answer is yes and
no. Obviously, in areas like disarmament, it would be far-fetched to worry about over-
deterrence. However, a different picture might be drawn in international investment
law. Indeed, a number of critics have pointed out the dangers of what has been termed
regulatory chill, especially in the area of environmental regulation.417 This concern has
also been mentioned by a number of international organisations, such as the OECD,
the World Bank and UNCTAD.418




416
      The extreme case of this theory could be, as before, a race to the bottom. However, the author is
      not aware of any commentator expressing this extreme view and will thus focus on the potential
      regulatory chill.
417
      See, e.g., IISD (2001).
418
      OECD (2001), p.8, World Bank (2001), p.16 and UNCTAD (2007b), p.85.
BITs and Institutional Competition                                                                 143

The discussions on this issue in the area of international investment law seem to have
begun with a number of investment cases under NAFTA Chapter 11, namely Ethyl
Corporation v. Canada, Methanex v. U.S., Metalclad Corp. v. Mexico, S.D. Myers v.
Canada and Pope & Talbot Inc. v. Canada.419 For example, the Methanex v. U.S. case
concerned the ban of a certain additive, called MTBE, in gasoline by the state of
California. Methanex, a Canadian company that produced an important component of
MTBE, filed for arbitration and claimed 900 million USD in damages. This case
sparked a controversial discussion on the influence of international investment law on
the state’s regulatory powers. Environmental and social interest groups as well as
scholars raised strong concerns about the incentives of states to pass environmental
and social reforms in the presence of investment treaties.420 Would states abstain from
sensible and important (environmental) regulation to avoid being sued by an
international arbitration tribunal? The alleged reluctance to raise legal standards for
fear of losing capital would then be aggravated by an additional international legal
sanction. With regard to the NAFTA cases mentioned above, Wälde and Kolo (2001)
as well as Neumayer (2001) have argued that the concerns for a regulatory chill are
unfounded. As Neumayer (2001) concludes "(h)ardly any concrete evidence is
available on whether states might be scared away from enacting stringent
environmental standards due to the threat of a potential dispute challenge."421
However, the (recent) impact of BITs might differ substantially compared to the
NAFTA cases at the beginning of the century. Indeed, legal commentators have
observed an expansion of the scope of some BIT provisions by tribunals.422
The concern that BITs may aggravate the problem of regulatory chill rests on a
number of assumptions. To start with (and similar to the case of controlling self-
interested governments), the sanctioning mechanism must "function" in the sense that
the benefits of regulation must be low compared to the expected sanctions. Most
importantly, BITs must theoretically be able to challenge welfare-enhancing regulation
and there should be cases where this has actually happened. The existing evidence
with regard to these points will be discussed in section 5.4.




419
      Wälde and Kolo (2001), p.812.
420
      See, e.g., Mann and Moltke (1999), Banks (1999) or Ganguly (1999).
421
      Neumayer (2001), p.87. Nevertheless, legal scholars criticizing international investment law for
      posing dangers to environmental regulation still seem to base their claims mainly on these cases.
      As an example, see Miles (2008), p.3, fn 6.
422
      Picherack (2008).
144                                                                 BITs and Institutional Competition

5.3.2.2           Distribution of the Benefits of FDI
Even if the strengthened investment protection can generate additional investment
flows to developing countries as a group, countries will find themselves in PD
situations when the benefits of these investment flows are mainly or fully captured by
the MNEs. This point has a direct and an indirect aspect: the former is the specific
distribution of profits that may be agreed upon in an investor-state contract and that
may be "stabilised" through the BIT. The latter aspect relates to the ability of
governments to regulate FDI in a way that helps maximise the indirect profits of the
investment. The specific contents of investor-state contracts may vary substantially
and can obviously not be analysed here. If the competition between states for capital
would be perfect competition (and countries would therefore be price-takers), the
country could only "charge" marginal costs and the surplus would end up completely
with the MNEs. However, it seems unlikely that the bargaining position of the host
country is always weak compared to the MNE. If, for example, the market access is
valuable to the MNE, because the MNE can profit through agglomeration rents, etc., a
host country may very well have a certain amount of bargaining power vis-à-vis the
MNE. In addition, many indirect (and immeasurable) benefits to the host state cannot
be captured by the MNE. Nevertheless, even if it is assumed that investor-state
contracts have a strong distributional bias in favour of MNEs, do Bilateral Investment
Treaties strengthen these contracts via stabilisation provisions? This question relates to
the umbrella clause and has already been addressed in section 3.3.3.5.3. As the OECD
notes, the umbrella clause was not evaluated by an arbitral tribunal until the 2003 case
SGS Société Générale de Surveillance, S.A. v. Pakistan.423 In that case, the tribunal
concluded that breaches of investor-state contract clauses are not automatically
breaches of international treaty law. A number of subsequent decisions employed a
wider approach to the interpretation of the umbrella clause. For example, the tribunal
in LG&E v. Argentina found that the umbrella clause "creates a requirement by the
host State to meet its obligations towards foreign investors, including those that derive
from a contract; hence such obligations receive extra protection by virtues of their
consideration under the bilateral treaty."424 The case law at this point is not
conclusive.425 Yet, tribunals have frequently granted BIT protection to investor-state
contracts. Consequently, there is indeed a certain risk that contracts that have an
inherent distributional bias in favour of investors are covered by international law.


423
      OECD (2006), p.15.
424
      ICSID ARB/02/1, Decision on Liability (10/03/2006), paras. 169-175. Similar interpretations seem
      to have been applied in Noble Ventures v. Romania and SGS v. Philippines. See Dolzer and
      Schreuer (2008), p.156.
425
      Dolzer and Schreuer (2008) refer to “disturbingly divergent lines of jurisprudence.” Dolzer and
      Schreuer (2008), p.155.
BITs and Institutional Competition                                                    145

However, it has also been remarked that it is unclear whether this bias exists in the
first place. Further, it has been estimated that fewer than half of all BITs actually
contain an umbrella clause.426

5.4               Provisions and Evidence
A number of treaty provisions appear to be relevant in the context of institutional
competition. Obviously, especially the non-discrimination clause may address
preferential treatment of (domestic) interest groups and will be discussed in section
5.4.1. On the other hand, the scope of the fair and equitable treatment clause and the
topic of indirect expropriation may have repercussions on the propensity of
governments to enact regulation and are addressed in sections 5.4.2 and 5.4.3.
Notwithstanding the actual treaty clauses, a threat of a regulatory chill requires that
relevant regulatory policies have been challenged through BIT arbitrations. The
problem of the regulatory chill may concern any issue area, but has so far been mainly
discussed in the context of environmental regulation. Therefore – and also to keep the
problem traceable – the discussion on the relevant case law will be limited to the cases
that have so far been put forward in the environmental realm. This will be done in
section 5.4.4.

5.4.1             Non-Discrimination
Evidently, especially the provisions in BITs that include standards of non-
discriminatory behaviour may impede the redistribution of rents from MNEs to
domestic interest groups. The non-discrimination principle can be considered as an
integral part of BITs. In many BITs, non-discrimination represents a standard of its
own and is often combined with the fair and equitable treatment standard.427 The non-
discrimination principle is also implicit in a number of other typical BIT clauses. For
example, the fair and equitable treatment certainly has a strong non-discriminatory
connotation.428 In addition, non-discrimination is insured through a most-favoured-
nation clause, which can be found in virtually all BITs. Most BITs combine the MFN
clause with the national treatment standard.429 Also, the expropriation clause itself
usually defines an act of expropriation as lawful if it was conducted in a non-
discriminatory manner. It has been mentioned before that the non-discrimination
standard is a relative standard in the sense that it compares the treatment of the foreign




426
      See section 3.3.3.5.3.
427
      Dolzer and Stevens (1995), p.62.
428
      This will be discussed in the following section.
429
      See for example Article 3 of the German Model BIT.
146                                                                 BITs and Institutional Competition

investor to the treatment of domestic (or other foreign) investors.430 The verification of
discriminatory behaviour in reality is certainly not a trivial task. For example,
questions may arise as to the correct group of reference. Put differently, can a
comparison only be drawn within an industry or also across industries and even
sectors? In addition, Van Harten (2007) emphasises with regard to national treatment
that to some extent, regulators must differentiate between the subjects of regulation.431
Furthermore, different interpretations exist concerning the relevance of discriminatory
intent.432 Nevertheless, a number of cases with regard to NAFTA and also BITs have
found the defendant in breach of some form of the non-discrimination principle.
Examples include Feldman v. Mexico433, Lauder v. Czech Republic434 and Occidental
v. Ecuador.435

5.4.2             Fair and Equitable Treatment (FET)
The fair and equitable treatment provision is among those most likely to be relied upon
in an investment dispute with an investor.436 At the same time, the content and
interpretation of the standard is still disputed.437 Section 3.3.3.5.1 has summarised the
main criteria that have been employed in the context of the FET standard, including
transparency and the protection of the investor’s legitimate expectations, freedom from
coercion and harassment, procedural propriety and due process as well as good faith.
Certainly, the FET standard also contains a non-discrimination element, such as that
inherent in the good faith requirement. As the tribunal in Saluka v. Czech Republic
underlined: “A foreign investor protected by the treaty may in any case properly
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.”438 The language of the FET
standards is broad enough to apply to virtually any adverse circumstance involving an
investment. As Picherack (2008) emphasises, recent, more stringent requirements that


430
      See Schreuer (2005), p.367.
431
      Van Harten (2007), p.86.
432
      Dolzer and Schreuer (2008), p.183.
433
      ICISD Case No. ARB(AF)/99/1, Award (December 16, 2002).
434
      Lauder v. Czech Republic, UNCITRAL, Final Award (3 September 2001), available under
      http://ita.law.uvic.ca/documents/LauderAward.pdf (September 5, 2010).
435
      Occidental Exploration and Production Company v. The Republic of Ecuador, LCIA Case No.
      UN3467,        Award      available   under     http://ita.law.uvic.ca/documents/Oxy-EcuadorFinal
      Award_001.pdf (September 5, 2010).
436
      Schreuer (2005), p.375.
437
      Picherack (2008), p.255.
438
      Saluka Investments v. The Czech Republic, Partial Award (17 March 2006), available under
      http://ita.law.uvic.ca/documents/Saluka-PartialawardFinal.pdf (September 5, 2010), para 307.
BITs and Institutional Competition                                                                  147

tribunals have applied with regard to the FET standard create "greater risks that
traditionally legitimate, good faith governmental regulatory, legislative, judicial, or
administrative actions will be impugned simply because they have a negative impact
on a foreign investment."439 The requirements the author criticises are the
consideration of the legitimate expectations of the investor, a stable and predictable
business and regulatory environment and total transparency of governmental
regulations. Arguably, regulatory measures can nevertheless be structured so as not to
be in breach with these requirements. According to Fauchald (2008) “(i)t is hard to
argue that public authorities may need to take “unfair” or “inequitable” measures in
order to achieve environmental protection."440 Nevertheless, a strict interpretation of
the fair and equitable treatment clause may very well lead to a breach of the standard
by measures taken in good faith, but lacking other criteria as defined by arbitral
tribunals, such as due process.

5.4.3             Indirect Expropriation
The problem of indirect or regulatory expropriation has been introduced in section
3.3.3.6.1. Obviously, capital-exporting states are aware of the dangers of regulatory
expropriation and include provisions in their BITs that demand compensation in these
cases. For example, article 6 of the 2004 US Model BIT on expropriation and
compensation explicitly covers indirect expropriations as well. These statutes are
evidently not very precise and arbitral tribunals have struggled to find the right
distinction between indirect expropriation and non-compensable regulation.441 The
problem at hand is not unique to the international domain. Domestic courts have a long
history of trying to define the appropriate border between regulation and regulatory
taking.442 As summarised before, the different approaches used by tribunals are the
Sole Effects doctrine, the Radical Police Powers doctrine and the Moderate Police
Powers doctrine. The Sole Effects doctrine constitutes the most stringent interpretation
on the occurrence of indirect expropriation, focusing exclusively on the effect of the
measure on the investment. According to Fortier and Drymer (2004), enrichment of
the host state and the deliberate targeting of the investor have in the past been taken as


439
      Picherack (2008), p.272.
440
      Fauchald (2008), p.18.
441
      See Behrens (2006) for a recent overview.
442
      In German law, the so called “Naßauskiesungsbeschluss” of the Bundesverfassungsgericht
      (BVerfGE 58, 300) based on Art. 14 GG serves as the reference case. See Maurer (1999), p.678.
      Important US case law is discussed in Miceli and Segerson (1999). The problem is that the
      discussion cannot be simply transferred to the international domain. As has been emphasised in the
      case of land by Blume, Rubinfeld et al. (1984), compensation for expropriation in the domestic
      context is economically not necessarily justified. However, non-compensation of expropriation
      would fail to reflect the trade-offs countries face on the international level.
148                                                                BITs and Institutional Competition

signs of a regulatory taking by arbitral tribunals, while the promotion of the general
welfare, as also contained in the US Model BIT, has been an indicator of a non-
compensable regulation. Evidently, with regard to a regulatory chill, the Sole Effects
doctrine can be problematic as it does not consider the intentions and purpose of the
regulatory measure but only the effect on the investment. Tribunals have applied this
doctrine, for example, in Biloune v. Ghana, Southern Pacific Properties v. Egypt and
Metalclad v. Mexico.443 However, according to the OECD, the balanced (that is, Police
Powers) approach is the predominant approach in practice.444

5.4.4             Evidence

5.4.4.1           Case Law
The discussions on non-discrimination, the FET standard and regulatory expropriation
have underlined that BIT clauses have at least the potential to chill the regulatory
efforts of developing countries. Much depends on the interpretation by tribunals.
Nevertheless, considering the initial example of environmental regulation, the actual
number of cases involving BITs is rather limited. As Fauchald (2008) notes: "It can be
argued that BITs have existed for approximately fifty years and that few cases have
come up in which there have been serious conflicts with environmental interests."445
Cases regularly cited by scholars expressing fears about a regulatory chill include
especially the NAFTA cases mentioned in section 5.3.2.1. (Ethyl Corporation v.
Canada, Methanex v. U.S., Metalclad Corp. v. Mexico, S.D. Myers v. Canada and
Pope & Talbot Inc. v. Canada) and, with regard to BITs, the cases Azurix Corp. v.
Argentine Republic (Azurix)446, Técnicas Medioambientales Tecmed, S.A. v. United
Mexican States (Tecmed)447 and Compañía del Desarrollo de Santa Elena, S.A. v.
Costa Rica (Santa Elena)448. It has been argued by Neumayer (2001) as well as Wälde
and Kolo (2001) that the NAFTA cases do not frustrate efforts of governments to
implement reasonable environmental regulations. For example, with regard to the
Ethyl Case (which was actually settled before the tribunal reached a ruling), Neumayer
(2001) points out that the case does not prove that governments cannot enact
(environmental) regulation, but rather that "discriminatory measures can be
successfully challenged under the investor-to-state dispute settlement provisions if



443
      See Kriebaum (2007b), p.724.
444
      OECD (2004), p.15.
445
      Fauchald (2008), p.35.
446
      Miles (2008), p.13 and Fauchald (2008), p.18.
447
      See Tienhaara (2008a), p.140, fn 4.
448
      See Miles (2008), p.18 and Tienhaara (2008a), p.140, FN 4.
BITs and Institutional Competition                                                               149

they are justified with dubious reasons, environmental or not."449 Wälde and Kolo
(2001) conclude with regard to the NAFTA cases mentioned: "It is only when the
environment becomes a pretext for domestic protectionism and when elements of
discrimination, of breach of governmental commitments or of use of regulation to
extract benefits unrelated to the legitimate purpose of the regulation can be detected
that a regulatory taking would, and should, be found."450 Can a similar claim be made
regarding the relevant BIT cases mentioned above?
The case Tecmed was filed with ICSID on July 28, 2000.451 In 1996, Tecmed, a
Spanish company, acquired a landfill for hazardous industrial waste in the
municipality of Hermosillo (which it operated under the name Cytrar). This landfill
had, prior to the purchase, an unlimited licence to operate the landfill issued by the
relevant regulatory agency, which was, after the purchase, replaced with a limited
licence. The limited licence had a duration of only one year, but could be extended
every year for an additional year. In 1998, the extension was not granted. Tecmed
invoked arbitration under the BIT between Spain and Mexico based on numerous BIT
provisions including expropriation, fair and equitable treatment and national treatment.
As Mexico is not a signatory state of the ICSID convention, the case was handled
under the ICSID Additional Facility. The refusal to grant the extension was seemingly
based on the community opposition to the landfill.452 On May 29, 2003, the tribunal
found the Mexican government's behaviour expropriatory and in breach of the fair and
equitable standard and awarded 5,533,017.12 USD to Tecmed. Regarding fair and
equitable treatment, the tribunal underlined that bad faith is not a prerequisite for a
violation of this standard.453 In addition, the tribunal emphasised that the regulatory act
was not directly driven by environmental considerations, but rather by strong
community opposition to the landfill. In this regard, the tribunal states that "it is
irrefutable that there were factors other than compliance or non-compliance by Cytrar
with the Permit's conditions or the Mexican environmental protection laws and that
such factors had a decisive effect in the decision to deny the Permit's renewal. These
factors included 'political circumstances'."454




449
      Neumayer (2001), p.89.
450
      Wälde and Kolo (2001), p.846. The Methanex case was at that time not decided, but the tribunal
      dismissed the claim in August 2005. See Methanex Corporation v. United States of America, Final
      Award (3 August 2005), available under http://ita.law.uvic.ca/documents/MethanexFinal
      Award.pdf (October 13, 2009).
451
      ICSID Case No. ARB (AF)/00/2, Award (May 29, 2003).
452
      ICSID Case No. ARB (AF)/00/2, Award (May 29, 2003), para 127ff.
453
      ICSID Case No. ARB (AF)/00/2, Award (May 29, 2003), para 153.
454
      ICSID Case No. ARB (AF)/00/2, Award (May 29, 2003), para 127. Original emphasis.
150                                                               BITs and Institutional Competition

In Azurix, the dispute related to a long-term contract as part of the privatisation process
of water utilities between the province of Buenos Aires and the US company Azurix in
1999.455 Apparently, the contractual relationship was problematic from the start, but
escalated in April 2000. Due to an algae bloom in the water reservoir, the water quality
was greatly reduced. The authorities issued warnings to the population, prohibited
Azurix from invoicing residents for the water and imposed a fine on Azurix.
Subsequently, the company terminated its concession agreement and filed for
bankruptcy. Azurix claims that the algae bloom was caused by a failure of the local
authorities to provide the necessary infrastructure prior to Azurix taking over the
responsibility for the algae removal works. The company filed for arbitration under the
US-Argentina BIT, claiming, among other things, that Argentina's behaviour
constituted indirect expropriation and violated the FET standard. The tribunal did not
find indirect expropriation, but a breach of the FET standard. The interesting point is
that the tribunal confirmed the view that failure of fair and equitable treatment does
not require bad faith on the part of the defendant.456 Azurix was consequently awarded
165,240,753.00 USD. In September 2009, a tribunal repudiated the defendant’s
attempt to have the award annulled.
The Santa Elena case is remarkable as it concerns a direct expropriation that was not
even denied by the defendant.457 The case was based on the US-Costa Rica BIT and
concerned the expropriation of land that was located next to a National Park and had
been acquired by the claimant to build a tourist resort. Aiming to preserve biodiversity,
the government of Costa Rica expropriated the land in 1978. The claimant did not
object to this act as such, but rather considered the amount of compensation to be
insufficient. After long negotiations and legal proceedings in the domestic realm, the
claimant filed for ICSID arbitration on June 2, 1995. As the lawfulness of the
expropriation (and its nature of being for a public purpose) was not disputed, the sole
function of the tribunal was to determine the amount of compensation. The tribunal
argued that "(w)hile an expropriation or taking for environmental reasons may be
classified as a taking for a public purpose, and thus may be legitimate, the fact that the
Property was taken for this reason does not affect either the nature or the measure of
the compensation paid for the taking."458 The award was finally set at 16,000,000.00
USD.




455
      ICSID Case No. ARB/01/12, Award (July 14, 2006).
456
      ICSID Case No. ARB/01/12, Award (July 14, 2006), para 372.
457
      ICSID Case No. ARB/96/1, Final Award (February 17, 2000).
458
      ICSID Case No. ARB/96/1, Final Award (February 17, 2000), para 71.
BITs and Institutional Competition                                                                      151

5.4.4.2            Analysis
It has been mentioned that serious conflicts based on BITs with regard to
environmental issues are rare. In addition, the cases discussed in the previous section
appear to have only limited potential to actually impact the regulatory behaviour of
governments in the area of environmental regulation. For example, the Santa Elena
case did not directly affect a regulatory measure but an outright expropriation. The
approach of the tribunal is in principle in line with the eminent domain principles of
most countries. It is therefore unclear how this case could induce policy-makers to
abstain from environmental regulation in the future. In the Tecmed case, the behaviour
of the government was strongly driven by political pressures, while a solid proof of
actual environmental damages through the MNE activity was not provided. In
addition, the award was, from a financial point of view, rather small (yet, reputational
damages may be substantial in these cases)459. The Azurix case, although not strictly
dealing with an environmental issue but rather an issue of public health, may indeed be
problematic due to the high sanction and the explicit negligence of the good faith
principle by the tribunal. In sum, the concern that Bilateral Investment Treaties may
cause a regulatory chill cannot be completely rejected. However, the existing evidence
in the realm of environmental regulation does not provide strong support for this
thesis. If governments act in good faith and in a non-discriminatory manner, the risk of
facing arbitrations against environmental regulation under BITs is low (but certainly
not zero). The problem, however, may be more persistent in other issue areas.460 It is
noteworthy that some developed countries have already taken specific steps towards
strengthening the importance public purpose considerations, which underlines that the
concern is taken seriously by policy-makers. The most prominent example in this
regard is the 2004 US Model BIT.461 The amendments of the US Model BIT have been



459
      High reputational sanctions may cause strong deterrence even if the probability of arbitration is
      low. A possible solution would be to increase confidentiality. This will be discussed in chapter 7.
460
      Argentina and the measures taken during the financial crisis are still a strongly-disputed example.
      However, as the caseload resulting from the financial crisis is inconclusive and not representative
      of "regular" BIT cases, Argentina's financial crisis was not discussed here. For an overview, see
      Burke-White (2008).
461
      The 2004 US Model Bit, Annex B, specifies: “(a) The determination of whether an action or series
      of actions by a Party, in a specific fact situation, constitutes an indirect expropriation, requires a
      case-by-case, fact-based inquiry that considers, among other factors: (i) the economic impact of the
      government action, although the fact that an action or series of actions by a Party has an adverse
      effect on the economic value of an investment, standing alone, does not establish that an indirect
      expropriation has occurred; (ii) the extent to which the government action interferes with distinct,
      reasonable investment-backed expectations; and (iii) the character of the government action. (b)
      Except in rare circumstances, non-discriminatory regulatory actions by a Party that are designed
      and applied to protect legitimate public welfare objectives, such as public health, safety, and the
      environment, do not constitute indirect expropriations.”
152                                                                     BITs and Institutional Competition

preceded by similar developments in NAFTA, probably as a result of the United States
and Canada finding themselves unexpectedly in the position of the defendant in a
number of treaty claims.462

5.5                Discussion and Conclusions
Bilateral Investment Treaties are used by countries that compete for capital. However,
while a BIT would not be concluded if it were not a Pareto-improvement for the two
signatory states, the BIT regime as such may very well be Pareto-inferior for
developing countries. If the elasticity of FDI flowing to developing countries is low,
these countries find themselves inevitably in a prisoner's dilemma situation. This
prisoner's dilemma situation can be interpreted as a result of a detrimental form of
institutional competition. This competition may lead to sub-optimal results even if the
assumption of inelastic investment is abandoned. The literature on institutional
competition is vast, but contains two recurring concepts. One is the potential for
institutional competition to act as a control mechanism when domestic inefficiencies
prevent welfare-optimal policies; the other is the threat of an underprovision of public
goods. In addition, the competitive situation may lead to a distributional bias of the
surplus of the investment in favour of MNEs. If the latter two effects dominate and are
reinforced by BITs, the BIT regime will leave developing countries worse off despite
additional investment to developing countries as a group. On the other hand, a Pareto-
improvement can be expected if BITs successfully exert a control function. That the
incentives of governments may indeed deviate from the goal of welfare-maximisation
is an assumption that is well accepted in the political economy of international trade
law. However, BITs cannot be understood as reaping the benefits of the political
support due to the reciprocal nature of the treaty. Yet, governments are likely to have
higher discount rates and a propensity to redistribute rents to interest groups to the
detriment of social welfare. Therefore, the time inconsistency problem, which
theoretically concerns all countries that receive FDI, will be exacerbated if
governments are unconstrained by domestic safeguards. In this case, a BIT will help
align the incentives of governments and citizens.
The content and also the interpretation of BITs certainly display elements of both
manifestations of institutional competition.463 This is not surprising as richer models of
institutional competition equally contain elements with countervailing welfare effects.
While the (theoretical) ability of BITs to prevent governments from redistributing
rents to domestic interest groups is obvious (yet difficult to apply in reality), the
content and interpretation of BITs also have the ability to raise concerns regarding the

462
      Van Aaken (2009), p.32.
463
      The answer to the question raised in the introduction of the chapter is therefore: both.
BITs and Institutional Competition                                                                       153

ability of countries to enact welfare-enhancing regulation. Finally, the distributional
dimension, which is mainly part of investor-state contracts and not BITs, is difficult to
evaluate. The assumption that MNEs always have more market power is nevertheless
unreasonable. In addition, the extent to which these agreements are actually covered
by BITs is still disputed and additionally restricted by the limited spread of the
umbrella clause. With regard to the specific example of environmental issues, the risk
of an underprovision of the "public good regulation" was argued as being low.
Nevertheless, some states have taken steps towards a specific emphasis of the public
policy goals of regulatory measures. This implies that policy-makers take the threat
seriously.
This chapter argues therefore that, as far as the welfare of developing countries is
concerned, BITs should foster the positive elements of institutional competition while
mitigating the harmful effects. The question is therefore not if BITs should limit state
sovereignty or national regulatory discretion as such - the question is rather how.464
More specifically, the BIT regime should rather focus on the abolition of internal
inefficiencies without inducing a regulatory chill. Put differently, discriminatory
behaviour requires sanctioning, while the public purpose and good faith of any
measure should certainly be taken into account. The concept of legitimate expectations
might be a good approach that captures this idea.
This conclusion is perfectly in line with the judgement of Van Aaken (2009), yet the
argument presented here emphasises a different aspect. In Van Aaken (2009), the
optimal trade-off between commitment and flexibility was compromised by the lack of
adequate measures in BITs or the far-reaching interpretations by tribunals. The
argument presented here is rather that this trade-off is not necessarily distorted as
under certain circumstances, even the Sole Effects doctrine may be Pareto-efficient for
the treaty partners.465 The problem stems from the competitive interactions of the
involved countries and is therefore a systematic problem rather than a problem of
treaty design. At the same time, this competition may have beneficial effects as long as
FDI flows to developing countries are not completely inelastic. That developing
countries find themselves in a prisoner's dilemma is therefore questionable. In any
case, improvements of the BIT regime along the lines outlined above would certainly


464
      If BITs did not limit regulatory discretion at all, they would be useless in the first place.
465
      Namely, if the Sole Effects doctrine induced a lot of additional FDI and the benefits of this FDI
      offset the opportunity costs of regulation. In addition, the argument here does not rest on a trade-off
      in the sense of a relationship along a single variable. The problem of investment protection has
      several dimensions and the point is rather, as explained below, that it is beneficial to strengthen
      certain aspects of commitment, while loosening others. This would not be possible if investment
      protection only had one dimension. Nevertheless, the notion of a trade-off is not completely
      misplaced.
154                                                        BITs and Institutional Competition

be beneficial under the development aspect. Nevertheless, as Van Aaken (2009)
rightly emphasises, it is mainly the few countries where reciprocity (e.g., in the context
of NAFTA) has played a role that have adapted their treaty design to ensure the
lawfulness of measures taken in support of the public purpose. Whether this step will
be followed by capital-exporting countries that face a low risk of being subjected to
investment arbitrations remains to be seen.
6                BITs and Institutional Quality

Bilateral Investment Treaties are designed to increase investment flows to the
signatory states. Section 4.2.1, reviewing the empirical evidence regarding BITs,
revealed that some studies have found robust evidence that BITs indeed have a
positive impact on investment flows. In addition, BITs might not only fulfil their
stated purpose of attracting FDI but may also have external effects on domestic
institutional quality. These external effects have been suggested to be either of a
positive or negative nature. As the importance of institutional quality on economic
growth has been emphasised by a number of authors,466 the question of an interaction
between BITs and domestic institutional quality is of high relevance. This chapter
addresses the question of the interplay between BITs and institutional quality using
panel data in a fixed effects model. Specifically, we run regressions using the World
Governance Indicators as the dependent variable over the time period 1996 to 2007 for
non-OECD countries. The number of BITs and changes in domestic regulation in the
field of investment are utilised as independent variables. The analysis further controls
for GDP per capita, FDI inflows, trade openness, development aid and membership of
the ICSID convention.
The chapter will proceed as follows: Section 6.1 introduces the competing theories,
explaining why one might expect BITs to have a negative or positive impact on
domestic institutions. Section 6.2 describes the model and the data that are used to test
this impact and displays the results. Section 6.3 presents potential problems with
regard to the results of the regressions. Section 6.4 discusses the results and presents a
conclusion.

6.1              The Effect of BITs on Institutional Quality

6.1.1            Some Theory
What are the benefits of Bilateral Investment Treaties? The stated purpose of BITs is
the "encouragement and reciprocal protection of investment."467 A number of
empirical studies suggest that BITs accomplish this purpose.468 While FDI is, by now,
generally associated with positive effects on economic development, it has been
suggested that BITs might be detrimental to the welfare of developing countries who



466
      See, e.g., Rodrik, Subramanian et al. (2004).
467
      See, e.g., the US Model BIT or the German Model BIT.
468
      See section 4.2.1.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_6,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
156                                                                        BITs and Institutional Quality

find themselves in a prisoner's dilemma situation.469 This point was discussed in the
previous chapter. The argument was that individually, developing countries can attract
additional investment through the conclusion of a BIT. However, this investment
constitutes merely a diversion from one developing country to another. As a group,
developing countries would be better off without any BITs. This argument falls into
the discussion of the problems and merits of institutional competition. Chapter 5 also
pointed out that it has been suggested that FDI as such may have a negative or positive
impact on certain government policies in the sense of a race to the bottom or a race to
the top, e.g., in the area of tax policy.470 The preceding chapter tried to answer whether
BITs support the negative or the positive aspects of institutional competition.
This chapter deals with a different, yet closely related issue, namely the specific
impact of BITs on institutional quality. The analysis goes beyond the question of
whether BITs can have immediate effects like a regulatory chill or the prevention of
discriminatory government behaviour, but rather asks what effects BITs may have on
institutional quality in general. Certainly, in the discussion presented below on the
interplay between BITs and domestic institutions, arguments related to the theory of
institutional competition will also play a role. However, the primary thrust of this
chapter is a general and fundamental discussion on the causes of the quality of
institutions. The BITs are in this context not analysed as being a direct manifestation
of institutional competition, but rather as exerting a (mainly unintended) external
effect on institutional quality.
The idea that the effect of BITs on institutional quality may be negative was first put
forward by Daniels (2004).471 The author argues that the BIT "has systematically
diverted the interests of potentially influential foreign investors from demanding the
creation of good generalised laws and legal institutions, and has further encouraged
them to enter into long term arrangements that impair the state's capacity to regulate
effectively in the public interest and which further increase the risk of corruption and
abuse."472 The argument rests on two pillars. The first relates to the well known
importance of exit and voice in the political process as put forward by Hirschman
(1970). Daniels asserts that in "Hirschmann's terms, the BIT enclave enables foreign
investors to exit from domestic legal regimes and this, in turn, implies a withdrawal of
their voice from the domestic debate over the need for, and character of, good laws


469
      Guzman (1998).
470
      See Basinger and Hallerberg (2004) and the literature mentioned therein. The authors themselves
      develop a model that implies that a race to the bottom will in many specifications not occur. See
      also Sinn (2003).
471
      While the article of Daniels (2004) is only a working paper, it has been noticed and cited by other
      authors in the field. See, e.g., Franck (2007) and Dammann and Hansmann (2008).
472
      Daniels (2004), p.37.
BITs and Institutional Quality                                                         157

and legal institutions."473 This logic can obviously be taken one step further: if the
institutional quality does not matter to investors that are protected by a BIT, the
relevant state has no incentive to improve the quality to attract investment in the first
place. The second pillar relates to the influence of domestic political elites. In this
respect, Daniels argues that "the BIT allows domestic political elites to access foreign
capital but without having to abide any restrictions on the exercise of their political
power by independent courts that would be required were a more general set of rule of
law reforms enacted."474 In other words, the author suggests that domestic elites profit
from BITs and, as a consequence of the existence of BITs, have an incentive to
frustrate legal reform. Ginsburg (2005) follows a line of reasoning similar to Daniels
(2004). He argues that, under some circumstances, BITs can lead to reductions in
governance quality. The reason is that BITs allow foreign investors to bypass domestic
courts. This may reduce the courts' incentives to improve performance. In addition, the
existence of BITs allows "the government to segment its reputation among domestic
and foreign actors."475 A third point that might lead to a reduction of institutional
quality is that, under certain circumstances, foreign investors may "exacerbate
corruption and poor governance".476
A problem with this line of reasoning is that it implicitly assumes that foreign
investors have a significant influence on domestic processes, either through their mere
participation in the market or their own lobbying efforts. This assumption might be
challenged on the grounds that the very reason BITs exist is that investors are not very
powerful when operating in foreign counties. As Ginsburg (2005) acknowledges, an
argument can also be made that the adoption of BITs leads to higher (perceived)
institutional quality.477 First, BITs contain specific provisions to prevent behaviour that
might be considered as bad governance (e.g., fair and equitable treatment standards)
and also provisions that may cause a reduction of rent-seeking efforts (e.g., national
treatment clauses). As Dolzer (2005) notes: "[…], reformers in developing countries
nevertheless see these treaties as powerful tools for the modernisation of the domestic
administrative legal system, providing effective external checks and discipline on
deficiencies and shortcomings which may be difficult to agree upon and to implement
at the domestic level."478 Second, BITs may induce domestic courts to engage in some
kind of regulatory competition with international arbitration.479 A plausible argument


473
      Daniels (2004), p.25.
474
      Daniels (2004), p.25.
475
      Ginsburg (2005), p.119.
476
      Ginsburg (2005), p.119, referring to Hellman, Jones et al. (2003).
477
      Ginsburg (2005), p.119.
478
      Dolzer (2005), p.972.
479
      Ginsburg (2005), p.119.
158                                                                       BITs and Institutional Quality

might also be that domestic investors lobby against the discrimination that they might
face vis-à-vis foreign investors and thus pressure for an improvement of domestic
institutional quality.
The view that BITs can help promote the rule of law in developing countries seems to
enjoy - hardly surprising - some sympathy in the arbitration community. In a speech at
the University of Richmond, prominent arbitration lawyer Jan Paulsson asserts in
relation to investment protection treaties: "Others write that such treaties are false
promises because it is not demonstrable that they actually lead to more foreign
investment. But what if they 'merely' lead to better justice – say in Mexico, where in
the wake of NAFTA we are told that officials have developed the salutary instinct of
avoiding conduct which might be criticised in an international forum: a direct case of
compliance pull to the benefit of local citizens. And I confess impatience with some
writers – apparently chained to their desks – who criticise arbitration because, they
posit, the availability of decent justice in international arbitration stunts the
development of good practices in national courts!"480

6.1.2 Empirical Literature
The empirical question that results from this discussion is: can we identify an
influence of investment treaties on institutional quality? Ginsburg (2005) runs
empirical tests to verify this hypothesis. He finds that for the rule of law, as reported in
the Worldwide Governance Indicators, BIT adoption leads to a subsequent decline of
quality. More specifically, Ginsburg uses the changes in government effectiveness,
regulatory quality, rule of law and corruption control for the years 1998, 2000 and
2002 as the dependent variable. The independent variable is a dummy variable for
each country that adopted a BIT in 1995 or 1996. The author controls for GDP per
capita, democracy and political stability. However, the empirical results are rather
weak. Apart from Ginsburg, this question has not been addressed empirically. Also,
while studies of the effects of institutional quality on economic development are
legion481, less effort has been spent to evaluate the (empirical) determinants of
institutional quality. Nevertheless, a number of factors have been identified that shape
domestic institutions.
Especially in the field of corruption, scholars have tried to discover the causes of this
phenomenon. One study addresses the effect of FDI on corruption directly: Kwok and
Tadesse (2006) find that, controlling for a number of factors and checking for reverse
causality, corruption is lower in countries that experienced high FDI inflows in the


480
      Paulsson (2007), p.12.
481
      See, e.g., Knack and Keefer (1995), Rodrik, Subramanian et al. (2004) or Barro (2003).
BITs and Institutional Quality                                                                        159

past. In addition, the authors "find strong evidence that the harmful effects of culture
on corruption are lower and the beneficial effects of education on corruption are higher
in countries with higher FDI in the past."482 In a well-known cross-national study,
Treisman (2000) analyses several indexes of perceived corruption. He identifies six
factors that influence the level of corruption, finding that federalism leads to more
corruption. In addition, countries with Protestant traditions, histories of British rule,
more developed economies, and higher imports were perceived to be less corrupt.
Surprisingly, perceived corruption is not positively related to the current level of
democracy. Only a long tradition of democracy could be shown to have a positive (that
is, reducing) impact on corruption. Serra (2006) provides a sensitivity analysis on the
empirical determinants of corruption as identified by the literature. He confirms the
robustness of five variables: political instability leads to more corruption, while a long
democratic tradition, economic development and a high fraction of Protestants among
the population as well as British colonial history are associated with lower corruption.
Persson, Tabellini et al. (2003) examine the influence of electoral rules on corruption.
Analyzing the variation across countries, they find that certain aspects of the electoral
system have an impact on corruption. Even more interesting, the authors also look at
the time variation of the data using a panel analysis with country fixed effects. It
appears that increasing the legislators individually accountable under plurality rule as
well as increasing district magnitude reduces corruption.
Other studies have not limited their focus on corruption, but include other measures of
institutional quality like indexes of the rule of law or expropriation risk. Berkowitz,
Moenius et al. (2005) find that signing the New York Convention, especially without
reservations, has a positive and statistically significant influence on the rule of law.483
Further, signing the convention reduces expropriation risk as well as corruption. An
influential paper on the quality of institutions is La Porta, Lopez-de-Silanes et al.
(1999). In a cross-country study, the authors examine a wide array of performance
indicators for governments, including corruption and property rights indexes.484 The
authors find that ethnolinguistic heterogeneity, legal origins and religion have a
significant influence on government performance. In sum, countries that are poor,
closer to the equator, ethnolinguistically heterogeneous, use French socialist law and



482
      Kwok and Tadesse (2006), p.781.
483
      As mentioned before, the New York Convention, officially known as the Convention on the
      Recognition and Enforcement of Foreign Arbitral Awards, is an international treaty that requires all
      Member States to enforce arbitration awards issued outside their own jurisdiction without
      reviewing these awards through their own courts. The treaty was signed in 1958 and currently has
      142 signatory states.
484
      But also more objective indicators of government performance like the size of the public sector, the
      infant mortality rate or the illiteracy rate.
160                                                                       BITs and Institutional Quality

have large proportions of Catholics or Muslims tend to have worse governments.
Comparable results have been obtained by Hall and Jones (1999). Looking primarily at
the question of why some countries are more productive than others, the authors find
that differences are mainly driven by government policies and institutions, which they
call social infrastructure. This social infrastructure is, in turn, positively associated
with distance from the equator, trade share, the fraction of the population speaking
English and, alternatively, the fraction of the population speaking a European language
other than English. Another variable that has been identified as influencing the quality
of institutions is development aid. Knack (2001) shows that aid has a negative impact
on governance quality (measured as a combined index of corruption in government,
bureaucratic quality, and the rule of law). In a similar vein, Djankov, Montalvo et al.
(2005) test the influence of aid on democracy and find a negative impact of aid.

6.2           Model and Data

6.2.1         Fixed-Effects Model
To assess the relationship between institutional quality and BITs, a panel-data model
with fixed effects is used. Specifically, the following regression is estimated:


Institutional Qualityit=   0   +   1BITsit   +   2Controlsit   + countryi + yeart+   it




where i denotes country i, and t denotes year t. The variables countryi and yeart are
dummy variables for each country and each year respectively. Expressed verbally,
annual frequency data is used to examine the within-country effects of BITs on
institutional quality. In most specifications of the model, the independent variables are
lagged by one period (t-1).
BITs and Institutional Quality                                                                         161

Fixed-effects panel models are now widely used in economics.485 A fixed-effects
model allows for different intercepts for different groups of observation (in this case,
different countries) and thus controls for unobserved (time-invariant) country effects.
The fixed-effects model is especially suited to analyse dynamic adjustment. One
advantage for our purpose is that non-time varying omitted variables that jointly
influence the amount of BITs and the quality of governments do not pose any
problems in this set-up. The reason is that the very nature of the fixed-effects models
eliminates any non-time varying variables from the regression. The logic is quite
simple: in a fixed-effects model, a dummy variable for each individual country is
included. This dummy does not vary over time and must therefore be perfectly
collinear with any non-time varying characteristic of the country, e.g., a dummy for
civil law tradition or British colonial past. Nevertheless, the country dummy is needed
to allow for different intercepts for each country. In essence, fixed-effects models do
not explain differences or deviations between groups (here, countries) but deviations
of variables from their mean within groups. Put differently, a fixed-effects model
cannot so much tell us why the level of corruption is high in India and low in Finland,
but it can shed light on the question if a change in corruption in India was associated
with a change of some other variable (here, Bilateral Investment Treaties) in India. Put
differently: did the fact, that India had 5 ratified BITs in 1996, but 47 in 2006 have any
impact on or relation to institutional quality?486 In most specifications, a time-fixed
effect is also included to control for events that affected all recorded values of
institutional quality in a given year. All regressions use robust standard errors
(clustered by country), allowing for serially correlated disturbances.487




485
      An alternative approach would be to use a random-effects model. The random-effects model also
      allows for different intercepts, yet interprets them as being randomly drawn from a normal
      distribution. The advantage compared to the fixed-effects model is that the random-effects model
      uses fewer degrees of freedom and produces a more efficient estimator. However, the random
      effects estimator is biased when its composite error term is correlated with the explanatory
      variables. In the current case, there is a strong theoretical case for the fixed effects: as we are
      dealing with countries, it is very likely that there are unobserved country-specific effects that can
      be controlled for using the fixed-effects model. In addition, the fixed-effects approach is usually
      appropriate when the data exhausts the whole population or there is selection bias as opposed to
      drawing a sample from a large population (see Kennedy (2003), p.312). Especially, the fixed-
      effects approach primarily uses the variation within groups (countries), which is exactly the
      information of interest here. Finally, virtually all new panel data studies analyzing panel data in
      relation to different countries employ the fixed-effects model. See, e.g., Tobin and Rose-Ackerman
      (2006), Rodrik and Wacziarg (2005) or Büthe and Milner (2008). See Kennedy (2003), p.301ff, for
      a good comparison of fixed- and random-effects models.
486
      Figures are based on own counting using the country lists of BITs as provided by UNCTAD.
487
      This appears to now be the conventional approach in fixed-effects panel studies. See also Bertrand,
      Duflo et al. (2004). For the exact method employed here, see Cameron and Trivedi (2009), p.233
162                                                                      BITs and Institutional Quality

The short literature review on the empirical determinants of institutional quality
showed that many if not most empirical determinants as identified by the literature are
time-invariant. Geography (e.g., distance to the equator), culture (e.g., language or
religion), history (e.g., colonial past) and legal tradition (e.g., common law) do not or
only marginally change over time. Consequently, these determinants cannot be
included in the fixed-effects model.488 The focus is therefore on the explanatory
variables that have at least some variation over time. These variables are discussed in
the following section. Before that, however, some words with regard to sample
selection are in order. The following regressions only consider countries that are not
rich.489 Similar to the World Bank classification, a rich country is defined as having
more than $ 11,115 of GDP per capita.490 The reason is that BITs remain to be a tool
for developing countries. Although they are no longer only concluded between
developed states on the one side and developing states on the other side, BITs are
never concluded between two developed states. Bilateral Investment Treaties are
obviously also binding for rich countries as they are reciprocal in nature. Nevertheless,
in practice the focus is on the encouragement and protection of investment in
developing countries. This is also reflected in the research design of the main studies
analyzing the effects of BITs on FDI.491 Therefore, it does not seem unreasonable to
assume that the effect of BITs on institutional quality, if existent, mainly affects low
and middle income countries.

6.2.2             Variables

6.2.2.1           Dependent Variables
For the quality of institutions, a set of perception-based indicators called the
Worldwide Governance Indicators (WGI) is used. The WGI are collected by World
Bank economists Daniel Kaufmann, Aart Kray and Massimo Mastruzzi. The notion of
"perception-based" refers to the fact that a score is given to countries depending on
one or more surveys that ask for the subjective estimation by investors, experts,
development agencies etc., on the situation in a given country. The indicators are
divided into six subcategories: Voice and Accountability, Political Stability and


      and p.251ff. Indeed, running test statistics as described in Drukker (2003) revealed serial
      correlation in the error terms.
488
      And could reveal nothing about the source of changes in institutional quality anyway. For a good
      overview on panel data, see Kennedy (2003), chapter 17 or Verbeek (2000), chapter 10.
489
      A similar approach is taken by Büthe and Milner (2008) albeit excluding OECD countries.
490
      Actually, the World Bank uses GNI per capita. However, these data were not available and the
      difference in sample selection should be negligible.
491
      See, e.g., Neumayer and Spess (2005) or Hallward-Driemeier (2003), both focussing on the FDI
      flow to developing countries only.
BITs and Institutional Quality                                                                          163

Absence of Violence, Government Effectiveness, Regulatory Quality, Rule of Law
and Control of Corruption.492 Regressions are run for four of these categories that
seem most relevant, namely the Government Effectiveness index, the Regulatory
Quality index, the Rule of Law index and the Control of Corruption index. The data
are available for up to 212 countries for the years 1996 to 2008 (depending on the
country).493 Data are coded from –2.5 (worst) to +2.5 (best).494

6.2.2.2            Independent Variables
The independent variables of interest are Bilateral Investment Treaties and domestic
regulations in investment law. The BIT variable is simply the number of BITs of a
given country in a given year.495 We use the BITs that are ratified and not only signed
because a non-signed BIT is comparable to "cheap talk" in game theory: it is
communication (a signal) between players that does not affect payoffs.496 The number
of BITs is logged. A similar approach is taken by Tobin and Rose-Ackerman (2005).
Logs have the advantage that the results are not driven by a few influential data
points.497 In another specification of the variable, the sum of ratified BITs in three
preceding years (respectively) is calculated. This variable is supposed to account for
the possibility that BITs only exert short term or signalling effects (as the dependent
variable is perceived institutional quality) of statistical significance. The BIT variables
were constructed for the years 1992 to 2006.
The domestic regulation variable is the sum of changes in investment law with an
expected favourable impact on FDI minus the changes with an expected unfavourable
impact on FDI. These data are collected by UNCTAD and submitted to the relevant
state for verification. They are regularly used as an indicator for trends in FDI policies
in UNCTAD's World Investment Report.498 Eight areas of national FDI policies are
incorporated in this statistic: foreign ownership, sectoral restrictions, approval
procedures, operational conditions, foreign exchange, promotions (including
incentives), protection (including guarantees) and corporate regulation (stock


492
      These subsets are also used in Ginsburg (2005).
493
      No values are available for 1997, 1999 and 2001.
494
      With a few outliers scoring less than –2.5.
495
      The numbers for each country were compiled using the country lists of BITs available on
      www.unctad.org
496
      See the discussion in chapter 4.
497
      This also accounts for the idea that the marginal effect of an additional BIT/regulation is
      decreasing. A general disadvantage is losing data points with the value of zero. However, in the
      given data set, this is not a concern as virtually all countries have a positive number of BITs in the
      first year included. The transformation employed here is that one BIT was added to the cumulative
      number of BITs before taking the natural log.
498
      See, e.g., UNCTAD (2007a), p.14.
164                                                               BITs and Institutional Quality

exchange/ financial market laws). The major drawback for analytical purposes of these
figures as well as for the BIT variable is that the numbers are unweighted: one major
change in domestic regulation might have the same impact as a number of minor
changes. Equally, a BIT with a large, rich country might have a much bigger impact on
institutional quality than a BIT with a small developing country. This makes the
interpretation of coefficients difficult if not impossible. Nevertheless, it can still be
observed if these variables have an impact at all and if this impact is of a negative or
positive nature. In any case, the use of this variable is not so much to identify the
impact of these regulatory changes on institutional quality, but to have a control for the
domestic reform activities of the relevant country. The UNCTAD figures appear to be
a good proxy for that. The number is logged or summed over three years, depending
on specification, for the same reason and using the same technique as the BIT variable.
Data were available from UNCTAD (upon request) for the time period of 1992 to
2006.
The control variables are essentially the time-varying variables that have been shown
to influence the quality of institutions. GDP per capita is used as a measure of a
country's economic performance, aid as a percentage of GNI as a country's
dependency on foreign aid and the value of imports of goods and services as a share of
GDP as a measure of a country's trade openness. Further, it has been argued that the
amount of FDI may have an impact on the quality of government.499 Clearly, a
reversed effect can be expected: countries with higher institutional quality attract more
FDI. Following the convention in empirical studies using these data, the values for
GDP per capita and FDI are used as natural logs. All the figures are taken from the
World Bank's World Development Indicators. Finally, a dummy for the ratification of
the ICSID Convention is included to control for the possibility that not the BITs
themselves but the legal enforcement through ICSID is important for the (perceived)
institutional quality. In most specifications, the independent variables are lagged by
one period (year) to account for a lagged adjustment of the perception-based indicator
(or, in the case of BITs, that the treaty is ratified at the end of the year).




499
      See Jensen (2006), chapter 4 and Kwok and Tadesse (2006).
BITs and Institutional Quality                                                                           165


6.2.2.3          Overview
           Variable Name                          Description                             Source
                                 the ability of the government to formulate and
                                                                                        Worldwide
                                 implement sound policies and regulations that
          Regulatory Quality                                                            Governance
                                       permit and promote private sector
                                                                                         Indicators
                                                   development

                                  the quality of public services, the quality of
                                      the civil service and the degree of its
                                                                                        Worldwide
                                   independence from political pressures, the
     Government Effectiveness                                                           Governance
                                        quality of policy formulation and
                                                                                         Indicators
                                    implementation, and the credibility of the
                                  government’s commitment to such policies


                                  the extent to which agents have confidence in
                                  and abide by the rules of society, in particular      Worldwide
              Rule of Law        the quality of contract enforcement, the police,       Governance
                                    and the courts, as well as the likelihood of         Indicators
                                                crime and violence

                                  the extent to which public power is exercised
                                                                                        Worldwide
                                   for private gain, including petty and grand
          Corruption Control                                                            Governance
                                 forms of corruption, as well as “capture” of the
                                                                                         Indicators
                                       state by elites and private interests

                                     Number of ratified Bilateral Investment          Own Composition
                 BITs
                                        Treaties (logs and averages)                 based on UNCTAD

                                    Differential of positive (FDI friendly) and
                                                                                      Own Composition
          Domestic Regulation      negative (FDI hostile) changes in domestic
                                                                                     based on UNCTAD
                                   investment regulation (logs and averages)

                                 Gross Domestic Product divided by population World Development
            GDP per capita
                                                    (logs)                        Indicators

                                                                                     World Development
                 FDI              Inflows of Foreign Direct Investment (logs)
                                                                                         Indicators

                                     Development Aid (% of Gross National            World Development
                 Aid
                                                  Income)                                Indicators

                                 Measure of Trade Openness: value of imports World Development
                Trade
                                  of goods and services as a share of GDP        Indicators

                                   Dummy Variable for the ratification of the
           ICSID Convention                                                                ICSID
                                            ICSID Convention
166                                                                  BITs and Institutional Quality




6.2.3             Summary Statistics
              Variable Name                 Mean                SD     Min             Max
             Regulatory Quality             -0,282          0,762      -3,175          1,505
         Government Effectiveness           -0,356          0,665      -2,109          1,491
                Rule of Law                 -0,403          0,704      -2,274          1,545
             Corruption Control             -0,392          0,639      -2,084          1,616
                 BITs (log)                  1,815          1,204        0             4,489
         Domestic Regulation (log)           1,302          1,047        0             4,934
            GDP per capita (log)             7,036          1,230      4,034           9,315
              FDI Inflow (log)              18,823          10,667       0            25,654
             Development Aid                7,233           40,631       0            124,349
              Trade Openness                83,818          2,553     10,831          316,740
             ICSID Convention               0,696           0,460        0               1




6.2.4             Regulatory Quality and the Rule of Law
The following table displays the regression results for regulatory quality from the
World Development Indicators as the dependent variable:500




500
      All regressions are calculated using STATA, version 10.
BITs and Institutional Quality                                                                               167

                                   Dependent Variable: Regulatory Quality
                                       -I-         -II-        -III-       -IV-         -V-         -VI-
                BITs                -0.0538
                (log)               (0.0463)
             BITs in t-1                        -0.0737*     -0.0681      -0.0620                 -0.149***
               (log)                            (0.0424)     (0.0428)    (0.0431)                 (0.0430)
                BITs                                                                  0.00369
            (sum 3 years)                                                            (0.00371)
        Domestic Regulation         0.108***
              (log)                 (0.0408)
      Domestic Regulation in t-1                0.0992***                0.133***                 -0.00658
                (log)                           (0.0374)                 (0.0392)                 (0.0299)
        Domestic Regulation                                                           0.00559
          (sum 3 years)                                                              (0.00773)
           GDP per Capita           0.991***
               (log)                 (0.193)
        GDP per Capita in t-1                   0.847***     0.803***    0.738***     0.845***    0.552***
               (log)                             (0.179)     (0.168)      (0.180)     (0.179)      (0.155)
          Development Aid          0.00862***
             (%GNI)                 (0.00288)
       Development Aid in t-1                   0.00632**   0.00743***   0.00596**   0.00766***   0.00592**
             (%GNI)                             (0.00260)    (0.00223)   (0.00271)    (0.00239)   (0.00232)
                                    -0.000885
          Trade Openness
                                    (0.00116)
                                                -0.000620   -0.000678    -5.94e-05   -0.000290    -0.00127
       Trade Openness in t-1
                                                (0.00135)   (0.00129)    (0.00139)   (0.00136)    (0.00137)
             FDI Inflow             0.0471***
               (log)                 (0.0124)
          FDI Inflow in t-1                     0.0336***    0.0320**                 0.0302**    0.0337***
                (log)                            (0.0118)    (0.0126)                 (0.0122)     (0.0121)
                                    -0.0233
         ICSID Convention
                                    (0.0826)
                                                 0.00420      0.0103      0.0453      -0.0138      -0.0115
      ICSID Convention in t-1
                                                 (0.0844)    (0.0870)     (0.0808)    (0.0837)     (0.0857)
                                   -8.409***    -6.626***   -6.202***    -5.794***   -6.609***    -4.364***
              Constant
                                     -1.431       -1.242      -1.201       -1.306      -1.259       -1.026
          Observations                 932         1045        1124         1105        1063         1045
      Number of Countries              133          133         140          133        133           133
           Time Effects                Yes          Yes         Yes          Yes         Yes          No
       R-squared (within)             0.212        0.171      0.132         0.152      0.148        0.091
  *** p<0.01, ** p<0.05, * p<0.1
  Robust standard errors in parentheses

The number of countries and the number of observations is not constant as a result of
missing data. In general, data were used for the time frame of 1996 to 2007.501 Here


501
      No data for the dependent variable are available for 1997, 1999 and 2001, so these years are
      excluded from the regression. The control variable domestic regulation was only available until
      2006. Therefore, the first, non-lagged specification covers the years 1996 to 2006, the lagged
      specifications cover the years 1996 to 2007.
168                                                                   BITs and Institutional Quality

and in the following tables, robust standard errors (clustered by country) are reported
in parentheses below the coefficients. The first column reports the values for the non-
lagged variables, while one-year lags are used in all other columns. Columns 2 to 5
vary with regard to the selection of control variables. Column 6 differs from the rest of
the estimations as it does not include time effects, which are by now standard in
empirical panel studies.502
The first observation that can be made is that the control variables are often significant
and mostly point in the expected directions. Domestic legislation and GDP per capita
are robustly and positively related to institutional quality. The same is true for FDI
inflows (if lagged) and development aid. The latter might be surprising as some
commentators have emphasised the negative impact of aid on institutional quality.503
Trade Openness is significant only in some of the specifications. The negative sign is
not surprising as trade and FDI may often substitute each other.504 Signing the ICSID
convention has a positive coefficient but lacks statistical significance.
The second important observation is that the variable of interest, the BIT variable,
indeed has negative coefficients. However, the variable is not significant in most
specifications of the model. Statistical significance can only be observed in two of the
regressions. First, significance at the (low) 10% level can be observed with the full set
of controls in the lagged model.505 Yet, this result proves to be very dependent on the
structure of the controls: omitting either Domestic Regulation or FDI Inflows from the
regression leads to a disappearance of significance. Second, the BIT variable is
statistically relevant at the 1% level when the time effects (time dummies) from the
regression (column VI) are excluded. However, year dummies are standard in modern
panel data studies as they have an important function: they eliminate common effects
that affect all data points in a given year. Especially in the example of perception
based indicators, year dummies appear to be important: the compilation of the indices
may differ between years for technical reasons. For example, compilation weights of
the used surveys might have been adapted; certain surveys may have been taken into
account while others were removed etc. In addition, political factors that affect all or
most countries in a given year which cannot be controlled for may have influenced the
overall perceptions of institutional quality.
Finally, it can be acknowledged that the average number of ratified BITs in the
preceding three years (column V) has a positive sign but is not significant at all. It


502
      Examples include Neumayer and Spess (2005) and Dreher and Voigt (2008).
503
      See, e.g., Djankov, Montalvo et al. (2005).
504
      See section 2.2.
505
      Sometimes, significance at the 10% level is not even reported as the 5% threshold is usually
      considered to indicate statistical significance.
BITs and Institutional Quality                                                         169

appears therefore that the ratification of many BITs has no immediate signalling effect
in the sense that it positively affects the perception about the institutional quality of a
country.506
The estimation results for government effectiveness are very similar to the results for
regulatory quality and are reported in Appendix D. The following table displays the
results for the rule of law index:




506
      This also holds also true when using a two year average.
170                                                                             BITs and Institutional Quality



                                Dependent Variable: Rule of Law
                                    -I-          -II-        -III-       -IV-           -V-         -VI-
               BITs               -0.0405
               (log)              (0.0317)
            BITs in t-1                        -0.0575*    -0.0554*     -0.0528                  -0.0848***
              (log)                            (0.0325)    (0.0321)    (0.0326)                   (0.0284)
               BITs                                                                   0.00202
           (sum 3 years)                                                             (0.00220)
       Domestic Regulation         0.0423
             (log)                (0.0265)
  Domestic Regulation in t-1                    0.0339                 0.0580**                   0.00539
            (log)                              (0.0249)                (0.0254)                   (0.0255)
       Domestic Regulation                                                            0.00125
         (sum 3 years)                                                               (0.00242)
         GDP per Capita          0.458***
             (log)                (0.144)
       GDP per Capita in t-1                   0.353**       0.231      0.290**       0.334**     0.274**
              (log)                            (0.141)      (0.143)     (0.121)       (0.141)      (0.108)
         Development Aid         0.00352**
            (%GNI)               (0.00146)
      Development Aid in t-1                  0.00435**    0.00462**   0.00400**    0.00597***   0.00455***
            (%GNI)                            (0.00188)    (0.00211)   (0.00162)     (0.00225)    (0.00172)
                                -0.00239***
         Trade Openness
                                 (0.000828)
                                              -0.00206**   -0.00158    -0.00151      -0.00172*   -0.00224**
       Trade Openness in t-1
                                              (0.000971)   (0.00101)   (0.00103)     (0.00100)   (0.000970)
            FDI Inflow             0.0166
              (log)               (0.0106)
         FDI Inflow in t-1                    0.0264***    0.0291***                 0.0232***   0.0252***
               (log)                          (0.00892)    (0.00845)                 (0.00849)   (0.00826)
                                   0.0684
        ICSID Convention
                                  (0.0653)
                                                0.0363      0.0383      0.0653        0.0226       0.0340
      ICSID Convention in t-1
                                               (0.0550)    (0.0559)     (0.0544)      (0.0576)    (0.0547)
                                  -3.679***   -3.249***    -2.324**    -2.416***     -2.994***   -2.485***
             Constant
                                    -1.045      -1.091      -1.027       (0.893)       -1.032      (0.750)
         Observations                 926        1039        1115         1099          1057        1039
     Number of Countries              133         133         140          133          133          133
          Time Effects                Yes         Yes         Yes          Yes          Yes          No
      R-squared (within)             0.090       0.088       0.065        0.061        0.078        0.072
 *** p<0.01, ** p<0.05, * p<0.1
 Robust standard errors in parentheses
BITs and Institutional Quality                                                       171

As before, the BIT variable has a negative sign, but shows only a very weak
significance that is not robust to changes in the specification. Again, truly significant
results (at the 1% level) can only be obtained without year dummies. The other control
variables mostly display the same patterns as in the other regressions (government
effectiveness, regulatory quality).
172                                                                               BITs and Institutional Quality

6.2.5                  Corruption


                                Dependent Variable: Corruption Control
                                      -I-        -II-        -III-        -IV-            -V-          -VI-
               BITs               -0.0295
               (log)              (0.0338)
            BITs in t-1                        -0.0279      -0.0125      -0.0220                     -0.0595*
              (log)                            (0.0349)     (0.0350)    (0.0345)                     (0.0332)
               BITs                                                                    0.00357*
           (sum 3 years)                                                               (0.00208)
       Domestic Regulation       0.0748***
             (log)                (0.0285)
  Domestic Regulation in t-1                    0.0478                  0.0659**                      0.0169
            (log)                              (0.0304)                 (0.0290)                     (0.0230)
       Domestic Regulation                                                              0.00119
         (sum 3 years)                                                                 (0.00337)
         GDP per Capita           0.375**
             (log)                  (0.150)
       GDP per Capita in t-1                   0.342**      0.289**       0.167         0.347**      0.263**
              (log)                            (0.152)      (0.138)      (0.145)        (0.149)      (0.118)
         Development Aid        0.00478***
            (%GNI)               (0.00176)
      Development Aid in t-1                  0.00395**    0.00491**    0.00324*      0.00566***    0.00370**
            (%GNI)                            (0.00185)    (0.00201)    (0.00181)      (0.00212)    (0.00162)
                                -0.00223***
         Trade Openness
                                 (0.000710)
                                              -0.00198**   -0.00151     -0.00175*       -0.00150    -0.00211**
       Trade Openness in t-1
                                              (0.000946)   (0.000944)   (0.00100)      (0.000975)   (0.000925)
            FDI Inflow           0.0204**
              (log)              (0.00978)
         FDI Inflow in t-1                     0.00592      0.00693                     0.00417      0.00540
               (log)                           (0.0110)     (0.0103)                    (0.0108)     (0.0112)
                                  0.197**
        ICSID Convention
                                  (0.0784)
                                               0.159**      0.163**     0.168**         0.155**      0.151**
      ICSID Convention in t-1
                                               (0.0750)     (0.0768)    (0.0704)        (0.0764)     (0.0736)
                                  -3.334***    -2.795**    -2.573**      -1.647        -2.850***    -2.188***
             Constant
                                    -1.064      -1.079      -1.044       -1.072          -1.063      (0.802)
         Observations                 912        1026        1101         1085            1044         1026
     Number of Countries              133         133         140          133            133           133
          Time Effects                Yes         Yes         Yes          Yes             Yes          No
      R-squared (within)             0.102       0.063       0.051        0.047          0.059        0.048
 *** p<0.01, ** p<0.05, * p<0.1
 Robust standard errors in parentheses
BITs and Institutional Quality                                                                         173

Also in the case of corruption, no robust impact of Bilateral Investment Treaties can be
detected. In this regression, we see a positive impact of the average number of BITs
concluded in the preceding three years, however significant only at the 10% level. As
before, changes in domestic regulation show a relatively robust relation to changes in
the corruption score. Remarkably and in opposition to the other regressions, the
corruption variable displays a significant and robust relation to the ratification of the
ICSID convention.

6.2.6              Outliers, Economic Freedom and OECD BITs
Different specifications of the model were tested to examine the validity of the
results.507 First, outliers with regard to excessive BIT conclusion did not affect the
results. This problem was addressed from the start by using logged values for the BIT
variable. Three countries have nevertheless been excluded: China (88 ratified BITs in
2006), Egypt (71 ratified BITs in 2006) and Romania (77 ratified BITs in 2006).
Leaving these countries out of the regression did not alter the results. Second, an
alternative measure of government quality, the Index of Economic Freedom, has been
used as the dependent variable. The index is compiled by the Heritage Foundation and
is a composite index accounting for the following aspects: business freedom, trade
freedom, fiscal freedom, government size, monetary freedom, investment freedom,
financial freedom, property rights, freedom from corruption and labour freedom. BITs
did not manifest a positive impact on this index (and neither did the control variables
except for GDP per capita). A third specification relates to the choice of BITs: it may
be argued that some BITs have a stronger impact than others. Especially BITs with
developing countries that do not cover large amounts of FDI flows may be less
influential than others. To account for this possibility, regressions have been run using
only BITs with OECD countries. Again, the results are mainly similar to the results
presented above. The full results of these regressions are displayed in Appendix D.

6.3                Caveats
The regressions presented above can contribute to place the discussion on the effects
of BITs on domestic institutions on a more objective foundation. Quite novel for
studies of this kind is the inclusion of a control variable for the domestic reform
propensity. Yet, before discussing the results, a few problems with these empirical
results must be addressed.




507
      The tables for the tests mentioned in this section as well as the correlation matrix are presented in
      Appendix D.
174                                                                     BITs and Institutional Quality

6.3.1             Selection and Composition of Variables
The lack of significance for the number of BITs might be a consequence of the fact
that the BITs (as well as the domestic regulatory changes) are unweighted. One
possible approach would be to weigh the BIT of a specific country with the investment
flows from that country.508 The problem was indirectly addressed through the use of
BITs from OECD countries only. Yet, for future research, more sophisticated weights
might be in order. These weights might go beyond the amount of FDI, but also
consider the legal differences within BITs regarding the level of protection.

6.3.2             Perception Based Indicators
Strong criticism could be expressed regarding the use of perception based indicators as
the dependent variable. Among others, Kurtz and Schrank (2007) have criticised the
use of perception based indicators as good indicators for the quality of institutions. The
authors suggest that the perception based indexes suffer, among other things, from
perceptual biases and adverse selection in sampling. Recently, the criticism on the
measuring of institutions has been analysed by Voigt (2009). The author discusses,
inter alia, the thesis as presented by Glaeser, La Porta et al. (2004) that the Worldwide
Governance Indicators and other, comparable indicators (e.g., the International
Country Risk Guide), measure policy outcomes rather than institutions. A better
approach might therefore be to focus on more objective criteria that indicate the
quality of governance.509 However, it is not clear which variables could constitute a
good proxy for the impact BITs have on the quality of institutions.
A related criticism might be that the changes over time are too small to allow for
statistical inferences or that they reflect technical rather than factual changes of
institutional quality. The latter point is addressed by the composers of the data
themselves who claim that changes in the behaviour of governments are mainly
factual.510 Also, technical changes that affect all countries should be accounted for
through the year dummies. As to the former point, it can indeed be observed that the
statistical power is low. Certainly, institutions as such suffer from inertia and path
dependency. Yet, at the same time, the control variables move in the expected
directions and often with statistical significance. Consequently, the data can certainly
contribute to point in the right direction. Obviously, more research in this direction and
especially the identification of more useful indicators is needed. In sum, the criticism
on perception based indicators is manifold and in many cases justified. Nevertheless,



508
      Neumayer and Spess (2005) use that approach when testing the effect of BITs on FDI.
509
      Voigt (2009), p.3.
510
      Kaufmann, Kraay et al. (2008), p.20f.
BITs and Institutional Quality                                                                  175

given the research question at hand and the prior research in the field, the Worldwide
Governance Indicators constituted the natural choice.511

6.4               Discussion and Conclusions
It has been argued in the literature that BITs influence the quality of institutions in a
positive as well as in a negative manner. In addition, empirical evidence has been
presented in the literature that supports the latter thesis. In this chapter, a simple set of
regressions in a fixed-effects framework have been presented to approach the problem.
Using this standard approach, the data showed no signs at all of a potential positive
impact of BITs on institutional quality. On the other hand, a negative impact of BITs
was only detected in few regressions and was virtually never robust to the inclusion of
time fixed effects and/or model specification. Therefore, the prior empirical findings
that BITs do have a negative influence on institutional quality might be exactly driven
by the lack of control for time effects and the set of chosen controls. Nevertheless, the
consistent negative sign of the BIT variable and the (yet small) significance in some
specifications is troubling and calls for further research.
A strength compared to earlier studies has been that the efforts of domestic reform
could be controlled for. On the other hand, the use of perception-based indicators is
not unproblematic. Remarkably, despite the problems with these indicators, the
relevant control variables are usually significant and point in the expected directions.
Namely, changes of GDP per capita are strongly and significantly associated with
changes in institutional quality in most of the regressions. A possible explanation
might be that more GDP results in higher tax revenue – money that could be invested
in improving institutions. Also, richer or growing countries might simply be perceived
as having better institutions. The causality here most likely goes both ways and
improved institutions lead to higher GDP per capita. A relatively stable connection
between most institutional quality indicators and changes in domestic investment
regulation can be shown. The unweighted nature of this variable makes quantitative
predictions impossible. Also, regulation in investment law is usually part of a larger
reform program that influences the institutional quality more directly. The domestic
regulation variable should therefore rather be understood as a proxy for the propensity
of domestic reform and not as an accurate predictor of institutional change.
Nevertheless, it can be relatively safely said that domestic reform can influence the
(perceived) institutional quality. This is a remarkable result as most other empirical
determinants of institutional quality, like history, religion and geography, cannot be
influenced by governments.

511
      See Ginsburg (2005) or Berkowitz, Moenius et al. (2005), who use comparable data from the ICRG
      (International Country Risk Guide) as the dependent variable.
176                                                          BITs and Institutional Quality

In summary, while optimistic expectations on the effect of BITs on institutional
quality are evidently unfounded, neither could a robust negative impact be established.
The negative sign of the BIT covariates is nevertheless striking and calls for further
research on the issue. This is especially true as the perception-based indicators
employed here suffer from a number of shortcomings. However, plausible proxies
based on objective indicators will be difficult to identify.
7                 BITs and Transparency

The current system of international investment law has been criticised for a lack of
transparency.512 The main concern is that topics of public interest should not be
decided without the knowledge and influence of citizens who are potentially affected.
In domestic law, it seems obvious that, for example, constitutional courts hear all cases
in public and publish the verdicts. In contrast, international investment arbitrations are
characterised by a high amount of confidentiality, despite the fact that these
arbitrations are usually concerned with issues similar to those heard in constitutional
courts. Often, cases that are not administered by ICSID are not registered anywhere.
Consequently, the exact number of investment arbitration cases is not even known.513
While the proponents of increased transparency have very strong arguments, this
chapter argues that transparency comes at a cost. At the core of the argument lies the
functioning of international law and the fact that reputational losses are not zero but
negative-sum to the parties. It will become clear why parties, apart from the desire to
protect business secrets, often prefer confidentiality and under which circumstances it
might be reasonable to allow for confidentiality. The aim of this chapter is in no way
to disavow the merits of transparency and indeed, in many cases if not most, a
balancing of arguments will result in favour of transparency. Yet, also in these cases, it
will be seen that transparency comes with costs that should not be neglected.
This chapter is organised as follows: the starting point is the legal background of
transparency in the different investment arbitration fora (section 7.1). The arguments
for more transparency with regard to international law and international investment
law will be summarised in section 7.2. Section 7.3 presents the case for confidentiality
emphasizing the incentive to comply with arbitration awards, the issue of efficient
breach and the incentive to conclude agreements. Section 7.4 discusses the results and
concludes.

7.1               Legal Background
The rules regarding transparency differ depending on the arbitration rules and fora.
Arbitration proceedings based on BITs are usually ICSID arbitrations, UNCITRAL
arbitrations or ad hoc-arbitrations.514 Obviously, in ad hoc-arbitrations, parties can


512
      See, e.g., Knahr (2007) and Van Harten (2007), p.159. See also the Model International Agreement
      on Investment developed by the International Institute for Sustainable Development (IISD),
      available under http://www.iisd.org/pdf/2005/investment_model_int_agreement.pdf (October 13,
      2009).
513
      Karl (2008), p.226.
514
      See section 3.3.3.8.


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_7,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
178                                                                         BITs and Transparency

agree on many different modalities regarding confidentiality. This includes the
possibility of not disclosing the existence of the conflict at all. With regard to the
UNCITRAL arbitration, the publication of awards is governed by article 32 (5). The
article states: "The award may be made public only with the consent of both parties."
With regard to public or third party access to hearings, article 25(4) stipulates that
“(h)earings shall be held in camera unless the parties agree otherwise”. The
UNCITRAL arbitration rules do not require parties to publicise the existence of an
arbitration proceeding. Yet, as article 32(7) establishes, the tribunal must register or
file the award if this is required by the national laws of the country where the award is
made. Quite similar to the UNCITRAL rules, the Convention on the Settlement of
Investment Disputes between States and Nationals of Other States, namely article 48
(5), prohibits the publication of awards by the centre without the consent of the parties.
Often, the awards of concluded cases are published on the ICSID website. However,
parties have also frequently chosen not to do so. Recent examples for non-disclosure
of the award include Europe Cement Investment and Trade S.A. v. Republic of Turkey
and Bernardus Henricus Funnekotter and others v. Republic of Zimbabwe. The
existence of a dispute as such (including the names of the plaintiff and the defendant)
is usually not confidential and consequently published on the ICSID website.515 It
must be remarked that many ICSID cases are settled before an award is issued. The
terms of settlement are normally not made public.
The same rules apply in principle for NAFTA Chapter 11 Proceedings as these
proceedings are also based on the ICSID (Additional Facility) or UNCITRAL
arbitration rules. Nevertheless, the NAFTA Free Trade Commission (FTC) has
recently established that “nothing in the NAFTA imposes a general duty of
confidentiality on the disputing parties to a Chapter Eleven arbitration, and, subject to
the application of Article 1137(4), nothing in the NAFTA precludes the Parties from
providing public access to documents submitted to, or issued by, a Chapter Eleven
Tribunal."516
Evidently, confidentiality constitutes a prominent feature in the realm of international
investment law. Regularly, not even the existence of a proceeding is made public. At
the same time, observers and scholars have called for more transparency in
international investment law. With regard to amicus curiae, ICSID and also NAFTA
(as well as the WTO in international trade law) have taken steps towards more
transparency.517


515
      OECD (2005), p.3.
516
      Notes of Interpretation of Certain Chapter 11 Provisions (NAFTA Free Trade Commission, July
      31, 2001), Article 1a, available under http://www.international.gc.ca/ (November 2nd, 2010).
517
      See Knahr (2007).
BITs and Transparency                                                                         179

The publication of documents during the proceedings and of the award has been
discussed in a few cases. With regard to NAFTA Chapter 11 proceedings, Loewen
Group, Inc. v. U.S. is often cited in the context of transparency.518 The tribunal had to
consider a request by the United States to release documents related to the arbitration
proceedings. The tribunal decided that there is no general duty of confidentiality in
NAFTA arbitrations or imposed through the rules of the ICSID Additional Facility.
The tribunal argued that especially arbitrations involving a state party should not be
confidential as this would deprive the public of knowledge and information concerning
government and public affairs.519 Another NAFTA case touching on this issue is
Methanex Corporation v. United States of America. This case is especially of
relevance with regard to amicus curiae.520 Nevertheless, the tribunal also stresses the
importance of the public interest in investment arbitrations in general, which goes
beyond regular commercial arbitrations. Noteworthy, the tribunal emphasised that the
public interest is not based on the fact that one party of the arbitration is a state. The
tribunal notes: "There is an undoubtedly public interest in this arbitration. […] This is
not merely because one of the Disputing Parties is a State: there are of course disputes
involving States which are of no greater general public importance than a dispute
between private persons. The public interest in this arbitration arises from its subject-
matter […]."521 NAFTA cases have gone as far as allowing hearings to be public and
broadcasted live. Examples are United Parcel Services of America v. Canada and
Canfor v. United States of America. In both cases, the disputing parties agreed to this
procedure.522
The case Biwater Gauff v. Tanzania is the most prominent case regarding
confidentiality based on a BIT and is discussed in detail by Knahr and Reinisch
(2007). Two points made by the tribunal appear to be of particular importance. First,
the tribunal asserted that the coverage of the case in the media or the publication of
documents "may aggravate or exacerbate the dispute and may impact upon the
integrity of the procedure."523 Second, the tribunal "argued that the tensions between
increasing transparency and safeguarding procedural integrity were only pertinent as
long as proceedings were pending."524 Evidently, the concern of the tribunal was that


518
      See e.g. Buys (2003), p.132 and OECD (2005), p.6.
519
      Fracassi (2001), p.218.
520
      OECD (2005), p.7.
521
      Methanex Corporation v. United States of America, Decision of the Tribunal on Petitions from
      third persons to intervene as "Amici Curiae" (January 15, 2001), para. 49, available under
      http://ita.law.uvic.ca/documents/Methanex-AmiciCuriae.pdf.
522
      OECD (2005), p.7ff.
523
      ICSID Case No. ARB/05/22, procedural note 3, para 136, available under
      http://ita.law.uvic.ca/documents/Biwater-PONo.3.pdf (October 13, 2009).
524
      Knahr and Reinisch (2007), p.108.
180                                                                          BITs and Transparency

extensive discussions of the case in public has a negative impact on the proceeding
itself especially through the aggravation of the conflict between the parties. The
tribunal referred to comparable statements made by other tribunals, namely Loewen
Group, Inc. v. U.S. and Metalclad Corporation v. Mexico.525 A key term in this line of
reasoning is therefore the procedural integrity. According to Knahr and Reinisch
(2007), the term appears “to comprise the entire set of circumstances necessary for the
efficient conduct of proceedings of which confidentiality seems to be just one, albeit a
crucial aspect.”526
In a recent decision, the tribunal in Giovanna a Beccara and others v. Argentine
Republic had to decide on a number of issues with regard to confidentiality, including
the right of the parties to engage in public discussions of the case and the publication
of the award as well as of minutes and records of hearings.527 As a general rule, the
tribunal stated: “In conclusion, the Tribunal deems that the ICSID Convention and
Arbitration Rules do not comprehensively cover the question of the
confidentiality/transparency of the proceedings. Thus, in accordance with Article 44 of
the ICSID Convention and Rule 19 of the ICSID Arbitration Rules, unless there exist
an agreement of the Parties on the issue of confidentiality/transparency, the Tribunal
shall decide on the matter on a case by case basis and, instead of tending towards
imposing a general rule in favour or against confidentiality, try to achieve a solution
that balances the general interest for transparency with specific interests for
confidentiality of certain information and/or documents.”528 For the specific case, the
tribunal decided, inter alia, that public discussion of the case “is restricted to what is
necessary”, that the award shall be published (as that had been agreed by the parties in
the first place) and that minutes and records of hearings “shall be restricted unless the
Parties otherwise agree, or the Tribunal otherwise directs”.529
The case law revealed that tribunals have recognised the tensions between
transparency and confidentiality. The arguments for transparency will be discussed in
the following section. It will become obvious that these arguments are certainly well
founded. Indeed, openness and transparency have a natural and understandable appeal.
A wide range of virtues are promoted by transparency, including honesty,
predictability, accountability, and legitimacy. Nothing in the argument presented
below contradicts or resists these advantages of transparency. The modest point that
will be developed in section 7.4 is that under certain circumstances, confidentiality has

525
      Knahr and Reinisch (2007), p.106.
526
      Knahr and Reinisch (2007), p.106.
527
      ICSID Case No. ARB/07/5, Procedural Order on Confidentiality (January 27, 2010).
528
      ICSID Case No. ARB/07/5, Procedural Order on Confidentiality (January 27, 2010), para 73. Bold
      in original.
529
      ICSID Case No. ARB/07/5, Procedural Order on Confidentiality (January 27, 2010), para 153.
BITs and Transparency                                                                181

its own advantages which should be weighed against the benefits of openness when
deciding upon the rules of a dispute settlement system. In the next section the
conventional case for transparency in international law in general and in international
investment law specifically will be laid out. Then, in section 7.4, the circumstances in
which confidentiality might be beneficial will be considered.

7.2                The Case for Transparency

7.2.1              International Law
Transparency has become a key concept in international law. Scholars with different
backgrounds have identified transparency as a means to induce compliance with
international legal norms. As Chayes and Chayes (1995) note: "Transparency
influences strategic interaction among parties to the treaty in the direction of
compliance:
      x It facilitates coordination converging on the treaty norms among actors making
        independent decisions.
      x It provides reassurance to actors, whose compliance with the norms is
        contingent on similar action by other participants that they are not being taken
        advantage of.
      x It exercises deterrence against actors contemplating non-compliance."530
Scholars using the rational choice approach have emphasised the importance of
transparency in international law from a game-theoretic point of view. Abbott (1993)
has described international treaties as prisoner's dilemmas, repeated prisoner's
dilemmas or stag hunt games. In these games, information is crucial for a cooperative
outcome to emerge. Therefore, the costs of the acquisition of information should be
kept low. As Mock (2000) implies, games of perfect information lead more easily to
cooperation than games with imperfect information, where means such as signalling
need to be applied to achieve a cooperative outcome.
Other scholars have taken the assertion that transparency is crucial to the effectiveness
of international regimes as given and focus on how to reach more transparency.531 In
addition, arguments for transparency include the positive external effects of awards in
the sense that every award clarifies the contents and meaning of international law.532
Also, transparency might reduce corruption and gives third parties that are affected by



530
      Chayes and Chayes (1995), p.22.
531
      Mitchell (1998).
532
      For the example of bilateral investment treaties, see, e.g., Franck (2005).
182                                                                         BITs and Transparency

the issue at hand, for instance NGOs or other public institutions, the possibility to
express their views and participate in the decision process.
The calls for more transparency have not been limited to international law in general,
but have also been specifically applied and extended to the area of dispute settlement
and arbitration.533 However, most commentators acknowledge that confidentiality has
its place in private commercial arbitration. Nevertheless, there is a strong request for
more transparency when at least one party to the arbitration is a state. The main
argument here is that when a state is party to an arbitration proceeding, the public
interest is concerned.534 NGOs like CIEL (The Center for International Environmental
Law) and IISD (International Institute for Sustainable Development) therefore call for
a revision of the UNCITRAL Arbitration rules.535 The public interest obviously results
from the fact that residents are naturally interested in how their governments behave in
international arbitration, awards of compensation might affect the budget of the
country, arbitrations often concern the misconduct of governments (and residents
should be informed about that misconduct) and arbitrations penetrate deeply into
domestic decision making processes. More generally, transparency is seen as a central
aspect of good governance as such.536 Also, most authors emphasise the importance of
transparency for a consistent body of case law to emerge.537 This, in turn, will increase
predictability (and avoid unnecessary disputes) and thus increase the confidence in the
whole arbitration system. Another side effect that Mistelis (2005) mentions is that the
publication of awards assists scholars and practitioners in studying the topic.538
As Buys (2003) notices, these arguments have also been used by the US and the EU to
call for increased transparency of the WTO dispute resolution, as the "(1) WTO
disputes deal with sensitive issues that affect large segments of civil society; (2)
Transparency would assist other Members in understanding the issues involved and
would permit more widespread and effective participation in the system; and (3)
Transparency would increase public confidence in the WTO dispute settlement
system, which in turn could facilitate compliance with DSB recommendations."539




533
      See, e.g., Gruner (2003).
534
      See Mistelis (2005), Buys (2003) and Knahr and Reinisch (2007).
535
      See http://www.iisd.org/pdf/2007/investment_revising_uncitral_arbitration.pdf (November 2nd,
      2010).
536
      Knahr and Reinisch (2007), p.110.
537
      Knahr and Reinisch (2007), p.111, Mistelis (2005), p.172.
538
      Mistelis (2005), p.172.
539
      Buys (2003), p.134.
BITs and Transparency                                                                  183

7.2.2             International Investment Law
The advantages of transparency in international law and dispute settlement naturally
also apply to international investment law and investment arbitration. Arbitral
tribunals frequently base their decision on earlier awards.540 Consequently, more
transparency could very well contribute to the growing body of (informal) precedent in
international investment law. As predictability increases, unnecessary disputes may be
avoided. In addition, confidence in the system of international investment arbitration
may be strengthened.541 It appears, prima facie, reasonable to assume that transparency
has the potential to make the system more effective.
Nevertheless, the most important issue of transparency with regard to international
investment arbitration is the involvement of state parties as defendants and thus the
presence of a general public interest. As Mistelis (2005) points out: "Increased public
visibility and the emerging 'transparency' evident in investment disputes are not only
justified by the participation of states, but is also an underlying public expectation held
by citizens in many countries. The subject matter of many investment disputes affects
the daily life of citizens, and some could even have an impact upon on the provision
and cost of 'public' services such as water, waste management, electricity, gas etc." Put
differently, transparency is not only a value in itself, but also enables citizens to make
informed decisions such as voting or the consumption of public goods. As Knahr and
Reinisch (2007) find regarding transparency: "Against the background of this host of
strong policy reasons in favour of transparency, i.e., publication of awards, it is
difficult to see any reason why the outcomes of investment arbitration should remain
confidential."542
The argument presented in the following does not seek to invalidate these strong
arguments for transparency in international law, international investment law and
international dispute settlement. It will be shown that confidentiality in a certain
respect, namely with regard to the proceeding of international tribunals, can be
beneficial – mainly as an incentive to comply with awards. In addition, it will be
pointed out that increased transparency will strengthen the negatives-sum nature of
sanctions - with potentially problematic effects for the parties. The idea that the parties
to a dispute have a strong interest in confidentiality, while the benefits of increased
transparency accrue to third parties might reduce the incentive of using the dispute
settlement mechanism in the future. The net effect of increased transparency is
therefore not unambiguously positive.



540
      See Commission (2007).
541
      See Knahr and Reinisch (2007), p.110.
542
      Knahr and Reinisch (2007), p.115.
184                                                                             BITs and Transparency

7.3               The Case for Confidentiality
To understand the benefits of confidentiality in international dispute resolution, it is
worthwhile to recall how dispute settlement in international law works and, more
importantly, what it seeks to achieve.543 The main mechanisms of dispute settlement in
international investment law are, apart from direct negotiations, international tribunals.
It has been emphasised that, although international tribunals generate rulings, there is a
contrast to the domestic context as there is no centralised system of enforcement in
place. In this sense, the tribunal cannot prevent a losing defendant from ignoring the
ruling.
The functioning of international tribunals despite the lack of coercive enforcement
powers was already described in chapter 4. However, it is worthwhile to repeat that the
key (and only) action performed by an international arbitration tribunal is the
rendering of a judgement. Whatever the tribunal can achieve must be achieved by the
ruling itself (or the threat of a ruling). In other words, the function of a tribunal is
mainly informational.544 In addition and specific to international investment law, there
frequently exists (depending on the case) the option of enforcement of the award in
third countries. Given that this is a very tedious path not used very often in practice,
the focus of the remainder of this chapter is on the informational impact of the
tribunal’s ruling. As mentioned, the information provided by the tribunal has the
potential to provide its own form of enforcement. This enforcement is, to be sure,
weaker than the coercive enforcement of domestic courts, but if international tribunals
are effective in any way it is because their rulings have the potential to impose costs on
losing defendants. Out of the Three R’s of Compliance introduced in section 4.3 -
namely retaliation, reciprocity and reputation – the latter has been identified as the
main enforcement mechanism in international investment law. The tribunal clarifies
the facts of the case (if there are ambiguities) and states whether the defendant has
breached international law or not. Consequently, the conflicting party as well as
observing states and parties can update their beliefs concerning the willingness (and
ability) of the defendant to comply with norms of international law. Since observing
parties have no inside information on the facts of the conflict, they therefore need a
neutral body like a tribunal to state the facts and render a ruling to allow them to
understand the situation. As observing states adjust their beliefs based on the
information provided by the tribunal, the tribunal’s ruling directly affects the


543
      The functioning of international law and dispute settlement from the rational choice perspective
      were previously discussed in section 4.3 and will only be briefly repeated here. The standard
      assumptions of rational choice theory equally apply in this section. See, e.g., Guzman (2008a),
      p. 17. These assumptions include that states are rational, self-interested and have exogenous and
      fixed preferences.
544
      The following description is based on Guzman (2008b) and Guzman (2008a).
BITs and Transparency                                                                  185

defendant's ability to conclude beneficial agreements in the future. This applies to
other states but also to potential investors. In other words, the existence of an
arbitration tribunal increases the reputational costs of a violation of international law.
One important point to emphasise in this context is that the sanctions in international
law are usually not zero-sum. That means, for example, that the reputation loss of the
defendant does not accrue to the country that wins the case or to anyone else (except
maybe through some kind of belief updating). As will be seen shortly, the negative-
sum nature of the sanctions can be problematic. Before turning to the structural
arguments in favour of confidentiality based on the enforcement mechanism described
here, the conventional arguments for confidentiality in international arbitration will be
restated.

7.3.1             Conventional Arguments for Confidentiality
The conventional arguments for confidentiality have become partly evident in the
discussion of the case law in section 7.2 and are well summarised by Buys (2003).
First, it may be necessary to protect sensitive business information and trade secrets
from disclosure. This is an argument usually put forward for confidentiality in private
commercial arbitration, but might in some cases also be important when one party is a
state. In addition, in private proceedings parties may take positions that they would not
take publicly. This concerns the bargaining space. Parties to the arbitration “may not
wish to expose certain allegations to the public, e.g., allegations of bad faith,
misrepresentation, incompetence, lack of adequate financial resources, etc.”545 In a
similar vein, parties “may not want a “loss” publicised, especially if a party is involved
in other cases with similar claims and defenses”.546 The arguments presented below are
also concerned with the reputational losses of defendants, but link these losses to the
enforcement structure of international investment law. A different point in favour of
confidentiality does not only concern the publication of awards but the publication of
documents during the trial. The point here is that publication of documents during the
hearings might frustrate or politicise the arbitration process. Confidentiality can in that
case facilitate settlement.547 Lastly, often confidentiality will be cheaper in terms of
litigation costs. In other words, more transparency might come with prolonged and
more complicated proceedings. Knahr (2007) makes this point with regard to the
admittance of amicus curiae.548




545
      Buys (2003), p.123.
546
      Buys (2003), p.123.
547
      Knahr and Reinisch (2007), p.110.
548
      Knahr (2007), p.352.
186                                                                               BITs and Transparency

7.3.2             Structural Arguments for Confidentiality
As mentioned, the argument presented here goes beyond these arguments for
confidentiality and focuses more on the systematic way international tribunals
function. In that sense, these arguments might be termed structural arguments.
International arbitration may come about for a number of different reasons. First, there
may be ambiguities about the interpretation of the legal rule. The defendant may come
off as a winner or a loser in this trial. Second, the legal rule may be clear and the state
deliberately chooses to breach as it deems the benefits of the breach higher than the
costs. In some circumstances, this situation may constitute what is known as efficient
breach in contract law and economics.549 In any case, arbitration in relation to
transparency might lead to reputational costs that are zero-sum by nature and may
therefore negatively affect the incentive of states to conclude treaties in the future.
Each of these cases will be discussed in the following sections.
Before discussing the benefits of confidentiality in more detail, a simple
confidentiality mechanism will be introduced as a frame of reference to the case of
transparency. In discussing the merits of confidentiality, the mechanism shall be such
that the existence and the result of a dispute resolution proceeding should be kept
secret from others, including other states.550 Furthermore, and critically, this
confidentiality is conditional on compliance by the losing party with the ruling of the
tribunal. If the losing defendant fails to comply with the ruling, the existence of the
dispute should be publicised along with the tribunal ruling and the defendant’s failure
to comply. The way the term is used here, confidentiality does not mean that the
parties hide information from each other or the tribunal. Nor does it imply that the
tribunal’s ruling is kept secret from one of the parties. Confidentiality is limited to
keeping the proceeding and its results confidential among the parties as long as a
losing defendant complies with the tribunal’s ruling.

7.3.2.1           Informational Ambiguities and the Incentive to Comply
As a first step, it is assumed that there is uncertainty about the content of the
underlying legal rule or a clause in the respective BIT. For example, the exact extent
of the fair and equitable treaty standard is still not undisputed.551 A country may then
be sued under international law for actions that it may perceive as lawful. If the
proceedings are not secret, there are reputational sanctions by the time a country is


549
      Obviously, a combination of these two reasons is also possible and probably closer to reality.
      Nevertheless, for analytical purposes, it makes sense to consider these two points separately.
550
      It has already been mentioned that this is a common feature in international investment arbitration,
      especially with regard to ad hoc tribunals.
551
      See section 3.3.3.5.1.
BITs and Transparency                                                                  187

being sued. Third countries will consider the defendant as a less reliable partner
concerning its compliance with international treaties. Consequently, the defendant will
find it harder to conclude mutually beneficial agreements with his international
partners in the future. Even if the final ruling clearly asserts that the defendant was not
in breach of international law, his initial reputation may not be completely recovered.
At a minimum, the defendant will face negative reputational sanctions during the time
the case is pending. These costs can be avoided through confidentiality.
Now it is assumed, in contrast to before, that the tribunal finds the behaviour of the
defendant unlawful and the defendant thus guilty. The defendant now has to decide
whether to comply with the ruling.552 Compliance will impose direct costs on the
defendant as he has to pay damages. These costs will be the same, no matter whether
the proceedings have been held secretly or not. Nevertheless, confidentiality as
described above changes the compliance decision in an important way: it makes non-
compliance relatively more expensive and therefore increases the incentive for
compliance. To understand the impact on the compliance incentive, it is useful to first
take a look at public proceedings: as before, the defendant faces a loss of reputation
simply for being sued. In addition, the defendant faces a reputational loss for breaking
his commitments (as he is found guilty). Put differently, in the case of transparency,
the public at that point already knows that the losing state is a violator or international
investment law. All of these costs are sunk at the moment of the compliance decision.
Consequently, the defendant will weigh the direct costs of the damages against the
costs or the reputational loss of being perceived as a country that does not comply with
the rulings of the arbitral tribunal. Under confidentiality, there will be a different
payoff structure: when the case is filed, the defendant does not face a reputational
sanction. If there is a conviction and the state complies, there is also no reputational
sanction either. The state faces, however, as before, direct compliance costs, such as
expectation damages. The state will weigh these costs against the reputational loss of
non-compliance which, in this scenario, includes being sued, breaking the law in the
first place and also ignoring the subsequent award. Consequently, while the costs of
compliance remain the same, at the very moment the compliance decision must be
made, non-compliance is relatively more expensive when the proceedings have been
kept secret. In other words, in the case of confidentiality, compliance with the
tribunal’s ruling is more likely. Second, there are no reputational losses where the
tribunal finds the defendant guilty, but the defendant complies with the ruling. The
argument is in essence simply that confidentiality changes the cost-benefit analysis of
the compliance decision of a violator of international law in favour of compliance. The
reason is that if the arbitration was confidential before and non-compliance will trigger

552
      On the compliance decision, see also section 4.4.3.2.1.
188                                                                         BITs and Transparency

the publication of the award, the reputational losses for the violation of international
law are not sunk yet and will enter the cost-benefit analysis of the state when it comes
to compliance with the ruling. In this sense, transparency will, contrary to popular
opinion, not increase the effectiveness of international law, at least not with regard to
compliance with arbitral awards.
International investment arbitration, it may seem, does not suffer from severe
compliance problems. Indeed, cases where investors have tried to enforce an award in
national courts, which would be a reasonable reaction to non-compliance, appear to be
rare.553 The problem is, however, that this strategy is costly and the chances of success
are low. Therefore, it is not possible to be certain that the seemingly low frequency
indicates high compliance. In addition, it can be observed that many arbitration awards
were recently followed by annulment proceedings. The latter part is especially
remarkable as the drafters of the ICSID convention deliberately chose only narrow
grounds on which annulment proceedings could be based.554 Argentina is a case in
point here: it appears that the country is attempting to annul all unfavourable awards
relating to the proceedings initiated as a result of the measures taken during the
financial crisis.555 Annulment is clearly an attempt to avoid compliance without
incurring the risk of the award being enforced in national courts.
An important point has not been addressed yet: if there was only uncertainty about the
rule, why do the parties not simply enter into negotiations? Why start (costly and
lengthy) arbitration proceedings at all? Despite the lack of reliable data on this issue, it
is reasonable to assume that these pre-arbitration negotiations indeed take place
between investors and governments. The problem is that negotiations can easily fail
for various, well-known reasons, often related to asymmetric information or the
inability of parties to find a distribution scheme for the surplus.556
The area of human rights shows that the mechanism as described above is also being
used in practice, yet in a different context. The so-called Inter-American System of
Human Rights provides a good example. This system consists of the Inter-American
Commission on Human Rights and the Inter-American Court of Human Rights. Both
bodies are organs of the Organization of American States (OAS).557 The former
resumed its work in 1960 and bases its legitimacy on a resolution of the General


553
      However, there appears to be no database collecting information on this issue, consequently,
      accurate statements on the frequency and significance of this problem cannot be made.
554
      See section 3.3.3.8.2.
555
      Examples are Siemens A.G. v. Argentine Republic (ICSID Case No. ARB/02/8), LG&E Energy
      Corp., LG&E Capital Corp. and LG&E International Inc. v. Argentine Republic (ICSID Case No.
      ARB/02/1) and Sempra Energy International v. Argentine Republic (ICSID Case No. ARB/02/16).
556
      See, e.g., Cooter and Rubinfeld (1989), 1078f.
557
      See Harris (1998) for an overview.
BITs and Transparency                                                                                      189

Assembly of the OAS. The rules of its organisation and functions are laid down in the
American Convention on Human Rights. The tribunal consists of seven members, is
headquartered in Washington, D.C., and has two main functions: the preparation of
country reports on the human rights situation in specific member states and the
examination of individual petitions. These petitions follow a procedure as laid down in
the "Rules of Procedure of the Inter-American Commission on Human Rights".
Petitions can be submitted by "any person or group of persons or nongovernmental
entity legally recognised in one or more of the Member States of the OAS" (Article
23). The assessment of the petition includes an initial review by the secretariat of the
commission (Article 26), a decision of admissibility by the commission (Article 37),
on-site investigations (Article 40) and the encouragement of a friendly settlement
(Article 41). If no friendly settlement is reached, the Commission will compose an
intermediary report with a decision on the merits. The process is – in short – as such
(Article 42 to 45). If the Commission finds no violation, the report will be published
with the annual report of the OAS Assembly. If the Commission finds one or more
violations, it will include proposals and recommendations into the intermediary report
that aim to ensure compliance with the relevant human rights regulations. This
intermediary report is transmitted to the state in question. In that case, the intermediary
report also contains a deadline by which the state in question must adopt the measures
proposed. If the state fails to do so, the case will, under certain conditions, be referred
to the Inter-American Court of Human Rights. If the case is not referred to the court,
Article 45 of the Rules of Procedure of the Inter-American Commission on Human
Rights describes the next steps:558




558
      See also Article 50 and Article 51 of the American Convention on Human Rights:
      Article 50
      1. If a settlement is not reached, the Commission shall, within the time limit established by its
      Statute, draw up a report setting forth the facts and stating its conclusions. If the report, in whole or
      in part, does not represent the unanimous agreement of the members of the Commission, any
      member may attach to it a separate opinion. The written and oral statements made by the parties in
      accordance with paragraph 1.e of Article 48 shall also be attached to the report.
      2. The report shall be transmitted to the states concerned, which shall not be at liberty to publish it.
      3. In transmitting the report, the Commission may make such proposals and recommendations as it
      sees fit.

      Article 51
      1. If, within a period of three months from the date of the transmittal of the report of the
      Commission to the states concerned, the matter has not either been settled or submitted by the
      Commission or by the state concerned to the Court and its jurisdiction accepted, the Commission
      may, by the vote of an absolute majority of its members, set forth its opinion and conclusions
      concerning the question submitted for its consideration.
190                                                                             BITs and Transparency

"1. If within three months from the transmittal of the preliminary report to the State in
question the matter has not been solved or, for those States that have accepted the
jurisdiction of the Inter-American Court, has not been referred by the Commission or
by the State to the Court for a decision, the Commission, by an absolute majority of
votes, may issue a final report that contains its opinion and final conclusions and
recommendations.
2. The final report shall be transmitted to the parties, who, within the time period set
by the Commission, shall present information on compliance with the
recommendations.
3. The Commission shall evaluate compliance with its recommendations based on the
information available, and shall decide on the publication of the final report by the
vote of an absolute majority of its members. The Commission shall also make a
determination as to whether to include it in the Annual Report to the OAS General
Assembly, and/or to publish it in any other manner deemed appropriate."
It is important to understand that cases before the Inter-American Court of Human
Rights are usually heard in public559 and judgements are publicised.560 Put differently
(and keeping in mind that reputation is one of the main enforcement mechanisms in
international law), non-compliance with the preliminary report will either trigger
publication of the case as it is referred to the Inter-American Court of Human Rights,
or if it is not referred, a final report will be drafted. Non-compliance with that final
report will usually trigger a publication of the report. The main point here is that the
state can avoid being publicly denounced as a violator of human rights and
international law by complying with the measures proposed by the Inter-American
Commission on Human Rights. This mechanism is structurally comparable to the
confidentiality mechanism described with regard to international investment law as it
uses the potential costs of reputational sanctions to induce compliance.

7.3.2.2           Efficient Breach and Settlement Problems
The preceding section focussed on compliance with the arbitration award. This does
not necessarily imply that confidentiality always improves compliance with the



      2. Where appropriate, the Commission shall make pertinent recommendations and shall prescribe a
      period within which the state is to take the measures that are incumbent upon it to remedy the
      situation examined.
      3. When the prescribed period has expired, the Commission shall decide by the vote of an absolute
      majority of its members whether the state has taken adequate measures and whether to publish its
      report.
559
      See Article 14 of the court's rules of procedure (http://www.corteidh.or.cr/reglamento.cfm).
560
      See Article 30 of the court's rules of procedure (http://www.corteidh.or.cr/reglamento.cfm).
BITs and Transparency                                                                  191

underlying legal rule. Quite to the contrary, confidentiality of the proceedings under
certain circumstances may make breach with the underlying rule easier – to the benefit
of both parties. It should be noted that in many instances in international law, legal
rules are akin to clauses in domestic contracts between private parties. These clauses
have a simple function: to maximise the joint surplus for the contracting parties. In
some cases, a contractual breach may be beneficial to both parties. In other words, the
benefit of the breach is sufficiently high that the breaching party can fully compensate
the other party and still be better off. Situations of this kind have been termed efficient
breach.561
Efficient breach in international law is problematic, mainly because sanctions are not
zero-sum. In addition, the breaching parties will face reputational sanctions that are
impossible to assess and most likely exceed the benefits of the breach. The concept of
efficient breach in relation to international arbitrations has been addressed by Wells
(2003) who reviews the award in Kahara Bodas Company (KBC) v. Pertamina and
PLN with reference to the methods the tribunal used to calculate the award. Please
note that this arbitration was actually not based on a BIT and also not administered by
ICSID. The claim was based on an investor-state contract and governed by the
UNCITRAL arbitration rules (Pertamina is a state-owned company). Nevertheless, the
issues the author raises when discussing the award are equally relevant for arbitrations
based on BITs. The main point that Wells emphasises is that the tribunal in the case
issued an excessive award due to an economically unjustified double counting. More
specifically, the tribunal awarded KBC with 111.1 million USD for "lost expenditures"
and 150 million USD for "loss of profits".562 The author asserts that the aggregation of
the two types of losses is economically not justified as the expenditures are a
prerequisite for the profits.563 Put differently, if "the project was expected to generate
'normal' rates of return for the business, then the amount of investment itself provides a
reasonable starting point for determining FMV."564 The conclusion is that the award
was excessive and that this may have problematic implications: "[…] excessive awards
discourage government takings, or breach of contract, when such actions are in fact
efficient and thus desirable."565 The author quotes in this context Richard Posner, who
asserts: "Notice how careful the law must be not to exceed compensatory damages if it
doesn't want to deter efficient breaches."566



561
      On efficient breach, see Schäfer and Ott (2005), p. 461f.
562
      Wells (2003), p.472.
563
      Wells (2003), p.474.
564
      Wells (2003), p.475. FMV represents the fair market value of the project.
565
      Wells (2003), p.478.
566
      See Wells (2003), p. 478, fn 23 referring to Posner (1986), p.108.
192                                                                               BITs and Transparency

Marrela and Marboe (2007) apply these arguments to international investment law.
The problem with international investment law is that in addition to the damages
awarded, the state in question will also face reputational sanctions. As the authors
point out: "The 'cost' of a treaty breach then will turn out to be much higher than the
amount of payable damages or compensation. The State might face a decline in foreign
investment, higher political risk rating and thus insurance premiums and a higher
threshold hurdle rate of return to compensate for such risk."567 The authors conclude
that, as these costs of a breach are not only virtually impossible to forecast, but may
also be excessively high, efficient breach does not seem to be an available option in
international investment law. The authors discuss some potential remedies for this
problem. This includes the fact that efficient breach is, in a certain way, built into
BITs, namely through the fact that the requirements of an expropriation are included in
the treaty.568 The authors point out: "If the payment of compensation for a lawful
expropriation (usually the fair market value of the expropriated property rights) in the
eyes of the host State seems to be more sensible than to perform the contract (or to pay
contract damages) it may well opt for a lawful (preferably direct) expropriation."569
Another option would simply be negotiations. While the latter would suffer from the
same problems as identified in the preceding section (particularly asymmetric
information and high transaction costs), a lawful expropriation is in many cases not an
available option for at least two reasons: first, a regulation that may be at odds with a
BIT cannot always be structured as a lawful (and compensated) expropriation. Second,
there may be valuation problems with regard to the lost profits of the MNE that need
to be solved through a neutral forum.570 More specifically, as interests are strongly
opposed once the expropriation has taken place, an arbitral tribunal might be necessary
as a neutral forum to clarify the facts of the case and evaluate if, for example, the
expropriation was lawful in the first place. To sum up, there may be over-deterrence
because host countries may abstain from efficient breach for fear of high reputational
damages if the fragile process of negotiations fails.571 Marrela and Marboe rightfully
conclude that the concept of efficient breach is difficult to adapt to the specific
investor-state relationship and that, given the high and hard to estimate costs, "can
therefore, also from an economic point of view, hardly be recommended."572



567
      Marrela and Marboe (2007), p.13.
568
      Therefore, it is actually technically incorrect to talk about breach here.
569
      Marrela and Marboe (2007), p.16.
570
      The case Santa Elena, which was discussed in chapter 5, is a good example.
571
      If the BIT puts strong limits on regulations (as, for example, the sole effects doctrine does), this
      might even lead to over-deterrence in regulatory matters. This problem has been discussed in more
      detail and with regard to institutional competition in chapter 5.
572
      Marrela and Marboe (2007), p.19.
BITs and Transparency                                                                 193

The problem with regard to settlement and efficient breach is exacerbated by the fact
that settlement has a distributional dimension. The decision of settlement vs. trial has
been analysed in section 4.4.3.5.2. It was pointed out that a settlement agreement
divides the surplus of the settlement. The assumption made in section 4.4.3.5.2 was an
equal division of the surplus between the parties. The surplus itself was defined as the
difference in payoffs between cooperation (settlement) and non-cooperation (trial).
The payoff to the host country (“HC”) in the case of settlement therefore reflected the
potential avoidance of reputational sanctions. However, this saving for the HC is partly
captured by the investor as part of the settlement payment. In the example presented in
section 4.4.3.5.2, the payoff to the HC was defined as y(- D - 0.5R). Assuming the
behaviour of the HC was unquestionably a breach, that is y=1, the HC will have to pay
0.5R in addition to the damages potentially awarded by a tribunal (presumably
expectation damages). Put differently, even when the host state can negotiate a
settlement agreement, it will face costs higher than expectation damages as a result of
the potential of reputational sanctions.
In a similar vein, Cooter and Rubinfeld (1989) emphasise with regard to the domestic
context that “[...], a defendant who wants to avoid the publicity of a trial will settle
cases that he has a high probability of winning”.573 As explained, this is especially also
true for the international investment context where reputation is a, if not the most,
valuable asset. Please note that the payment from the HC to the MNE is obviously not
a net loss to the parties. Nevertheless, it must be noted that efficient breach becomes
ceteris paribus less likely if the potential for reputational losses exists (because the
gain to the HC must be larger to cover the higher settlement sum). In other words,
while the system of investment arbitration is justified to threaten to abstain from
unlawful behaviour, the existence of high reputational sanctions may render this threat
too strong. This also holds true, as has been argued, if a settlement can be reached.
Even if efficient breach is nevertheless worthwhile, there will be a possibly unintended
transfer of welfare from the HC to the MNE (which may be especially problematic if
the HC is a developing state).
Confidentiality could play an important role here as the reputational sanctions may be
avoided. This would make efficient breach, which is by definition value-enhancing, a
more available option. The logic is simple: if neither the existence of the case nor the
final award (as long as the HC complies) is made public, there will be no reputational
sanctions in the case of efficient breach. This will also have an impact on any
settlement agreement as the MNE can no longer capitalise on the reputational losses of
the HC. The advantages for the HC are immediately evident. The important question is


573
      Cooter and Rubinfeld (1989), p.1075.
194                                                                          BITs and Transparency

therefore: will the MNE have a unilateral incentive to ignore the confidentiality
agreement (or threaten to do so to obtain the reputational capital)? Would this be a
credible threat? In other words, is confidentiality stable to unilateral deviation? Before
answering these questions from a theoretical point of view, it should be noted that in
practice, many ad hoc arbitrations are indeed often settled confidentially. Apparently,
in these cases, the plaintiff has neither before, during or after trial the incentive to
disclose the existence of his claim. The main reason is possibly that the threat of
disclosure is not credible due to the classic hostage problem. The reputation as a
hostage is not as such valuable to the MNE while destruction of the hostage is not only
costly but deprives the MNE of any benefit the hostage might have. Second, the
compliance effect as discussed in the preceding section could also exercise a pull
towards confidentiality. Specifically, the MNE has no interest in taking actions that
might frustrate or reduce the likelihood of compliance with the final award. Finally, as
long as third countries cannot verify the existence of the conflict, the reputational
damage (and consequently also the threat potential) appears to be weak.574

7.3.2.3           The Incentive to Conclude an Agreement
The risk of facing high reputational sanctions in case of informational ambiguities and
the difficulties related to efficient breach have repercussions on the incentives to
conclude agreements. In particular, the costs associated with transparency may
undermine the incentives of countries to conclude new treaties or encourage them to
search for alternatives that are less costly like, for example, ad hoc arbitrations. This
point is related to Van Aaken (2008a) who demonstrated the tendency of states to
“exit” their commitments in international investment law when they become too
costly. Ecuador, to give a recent example, denounced the ICSID convention on July 6,
2009.575 In accordance with Article 71 of the ICSID Convention, the denunciation will
take effect six months after notification. With regard to the choice between ICSID and
ad hoc-arbitrations, it is impossible to find out if the number of (confidential) ad hoc
arbitrations is rising.
The crucial point concerning the issue of transparency is the negative-sum character of
reputational sanctions. The problems that arise from the non zero-sum (negative)
character of sanctions in international agreements have been discussed by Guzman
(2005). The underlying logic is simple (and closely related to the preceding sections):
reputational sanctions are a loss to the state that has violated the law that is not offset
by a gain to the treaty partner (in the present case, the investor). Although high



574
      I thank Andrew Guzman for pointing that out to me.
575
      See the ICSID News Release from July 9, 2009, available under www.worldbank.org/icsid.
BITs and Transparency                                                                          195

(reputational) sanctions are positive in the sense that they increase compliance with the
underlying legal rule, states are aware that they will sometimes have to face these
sanctions (even if they do not intend to act unlawfully). In that sense, the sanctions are
a net loss to the parties. As Guzman (2005) asserts, the “desire to increase the
credibility commitment, then, is tempered by a desire to avoid this loss.”576 This is
exactly the reason why countries do not always choose the strongest type of
commitment available to them.
With regard to international investment law, it can be argued that a higher potential for
reputational losses through increased transparency might incentivise parties to avoid
BITs or use different, more confidential, means of conflict settlement like ad hoc
arbitrations. When the dispute is public, this gain flows to all states rather than being
limited to the parties of the dispute. From the perspective of the parties, then, the harm
to the defendant is not matched by a gain to the complainant – the sanction imposes
net costs on the parties. This “negative sum” feature might not only be a problem in
the instant the conflict is prevalent, but it could also be a more systematic problem
because it reduces the potential joint surplus from the agreement. This may make the
signing of the particular treaties in question less attractive. As mentioned, states may
refrain to employ arbitration as a means of dispute resolution in their international
agreements, while a regime where proceedings are kept secret could avoid these costs.
International cooperation or the use of international arbitration would in the end be
more likely. This problem might be exacerbated by the fact that the conventional
benefits of transparency as cited in section 7.4.1 do not mainly flow to the parties of
the agreement but rather to third parties (for example the value of precedence). As a
result, arbitration might not be chosen in the future if it loses its valuable feature of
confidentiality.
The question is, however, whether this logic holds for both partners to a treaty. The
provisions of the treaty itself obviously apply to both partners and developed countries
have already had costly experiences with the reciprocal nature of investment rules.577
Nevertheless, an asymmetry in investment flows between treaty partners still persists
in many cases. If one country is capital-importing while the other country is capital-
exporting, the latter, acting as an agent of the domestic MNEs, may have a certain
interest in increased transparency in the sense that in settlements, transparency
influences the distribution of the surplus in favour of MNEs. Further, benefits of


576
      Guzman (2005), p.582. See also section 4.1.1.
577
      A recent case concerns the German government and the energy company Vattenfall, albeit with
      regard to the investment provisions under the Energy Charter Treaty. See ICSID Case No.
      ARB/09/6. A tribunal was constituted on August 06, 2009. At the moment, the case is suspended
      (September 5, 2010).
196                                                                      BITs and Transparency

precedent and information are also captured by the domestic MNEs of capital-
exporting states. However, these benefits can only be captured partially and, in
addition, a countervailing effect exists, namely the risk of the domestic governments
themselves being the defendants in arbitration proceedings.

7.3.3            More Transparency?
Section 7.4.2 introduced a simple benchmark mechanism to highlight the problems
and costs associated with transparency. Another mechanism to avoid unjustified
reputational (negative-sum) sanctions could go in the opposite direction and may
imply even more transparency.578 For example, it has been argued that many cases are
based on different interpretation of the rules of the treaty. Upon registration of the case
the state may face reputational sanctions even though its behaviour may later be
deemed lawful. More transparency, e.g., through open hearings, might be valuable in
the sense that third parties will be better able to draw informed conclusions. This may
mitigate the potential loss that arises when an arbitration proceeding is registered -
especially when the host country can credibly ascertain that its behaviour was not
opportunistic and that it intends to comply with the ruling of the tribunal. However,
this credibility may exactly be the problem in the absence of neutral ruling. In
addition, the problems associated with efficient breach and over-deterrence will
probably not be solved through increased transparency.

7.4              Discussion and Conclusions
The striking argument against a confidentiality mechanism as presented above is
certainly the public policy argument. Indeed, in many cases, the underlying issues and
conflicts are of paramount public interest and should therefore not receive confidential
treatment. On the other hand, some cases involving a state party are actually more akin
to commercial arbitrations. A good example in this respect is the case Malaysian
Historical Salvors, SDN, BHD v. Malaysia. The subject matter of this case is a
contract between a marine salvage company (the claimant) and Malaysia (the
respondent) to salvage a British vessel that sank off the coast of Malacca in 1817.579
The recovered items were at a later stage auctioned for a total value of approximately
2.98 million USD. Subsequently, a conflict arose on the division of the proceeds and
whether the respondent withheld valuable items from the auction. Clearly, this case
does not concern vital issues of public regulation such as social or environmental
policy. In addition, the financial stakes are low while reputational sanctions are


578
      I thank Thomas Eger for pointing that out to me.
579
      Malaysian Historical Salvors, SDN, BHD v. Malaysia, ICSID Case No. ARB/05/10, Award (May
      17, 2007).
BITs and Transparency                                                                                197

potentially huge. Consequently, in cases that do not concern major public policy issues
and high stakes, greater confidentiality might indeed be a feasible mechanism. That,
however, would require tribunals to decide at the beginning of the arbitration if
transparency is warranted by public policy concerns on a case-by-case basis. Taking
this kind of decision would not be a novel one for tribunals: they employ the same
case-by-case evaluation when they decide on the admission of amicus curiae580 and
recently the tribunal in Giovanna a Beccara and others v. Argentine Republic has also
emphasized in principle the legitimacy of a case-by-case approach with regard to
issues of confidentiality.581
The confidentiality mechanism as presented here was mainly used as a benchmark to
highlight the costs that are induced through (increased) transparency. However, it is
noteworthy that this mechanism exhibits strong similarities to confidential ad hoc
arbitrations. Ad hoc arbitrations are by definition not registered and often operate
under confidentiality. If the losing party fails to comply, the winning party may try to
enforce the award in domestic courts – thereby making the case, the verdict and the
non-compliance decision of the losing party public. Consequently, this form of
mechanism is in principle available to the parties.
Drawing inferences about efficiency and social welfare is problematic in international
investment law. As host countries and investors choose confidentiality in practice with
regard to their arbitration proceedings, this arrangement must therefore in fact under
certain circumstances be Pareto-superior compared to other arrangements (considering
only the two treaty partners).582 For the treaty partners, the net benefits for the capital-
exporting state are unclear. However, capital-importing states should profit from
confidentiality as compared to transparency as they are primarily affected by the
potential loss of reputation. This is also true when settlement solutions are frequently
feasible. Consequently, if only the welfare of capital-importing (mainly developing)
countries was of concern, the avoidance of reputational losses through confidentiality
would probably be welfare enhancing. If, however, social welfare on a global scale is
concerned, the welfare effects of more confidentiality are impossible to calculate or
forecast.
Calls for more transparency in international investment law are legion. Definitely,
issues of great public concern cannot be kept secret from the public. In addition,


580
      Knahr (2007), p.336.
581
      However, as pointed out in section 7.1, the tribunal did not decide on the question whether the
      dispute as such should be made public, but rather on minor issues concerning confidentiality as for
      example the right of the parties to discuss the case in public.
582
      Keeping in mind the problems with using the Pareto criterion in this context. See also section 5.1
      on this issue.
198                                                                 BITs and Transparency

increased transparency, as has been argued in the literature, might strengthen the
legitimacy of the system of international investment arbitration. Therefore, it has been
emphasised that the merits of transparency should not be neglected. However, it was
also pointed out that increased transparency will not inevitably lead to increased
effectiveness. If effectiveness is understood as compliance with final awards,
transparency may actually weaken effectiveness. Transparency, as some authors have
claimed, has a deterrence effect against breach.583 This may be true, but if failure to
comply with the ruling does induce publication, deterrence would not be undermined.
At the same time, over-deterrence will be reduced and efficient breach enabled. Put
differently: through confidentiality, it is possible to avoid the reputational losses
without reducing compliance. This may, in turn, lead to more use of arbitration in the
first place. Nevertheless, it might be argued that the reputational loss of the defendant
is zero-sum to the treaty partners, but still valuable to third parties, as it helps them
update their beliefs about the defendant. In other words, there is a positive external
effect of precedence. This argument, as well as the public interest argument, is
undoubtedly true, yet both must be weighed against the potential benefits of
confidentiality. The existence of a public interest is not per se sufficient to disregard
the potential benefits of confidentiality. After all, issues in international law by
definition touch the public interest, and yet international law contains areas where
confidentiality is crucial.
In summary, a comprehensive discussion on the matter of transparency vs.
confidentiality should not neglect that transparency also creates costs. These costs are
mainly rooted in the negative-sum of reputational sanctions that may lead to over-
deterrence when it comes to efficient breach as well as to disincentives to conclude
new treaties (or incentives to exit existing ones). Furthermore, confidentiality can
make compliance with a ruling more likely once the violation has occurred. A
remarkable point is that developing or capital-importing countries might profit from
confidentiality more than developed countries. Yet, the overall welfare effects of the
trade-off between transparency and confidentiality remain unclear. Consequently, this
chapter does not argue in favour of increased confidentiality. Obviously, some
benefits of transparency are undeniable and in some cases of paramount importance.
The modest aim of this chapter is simply to introduce a new perspective based on
rational choice arguments on the issue of transparency vs. confidentiality that has not
yet been part of the discussion on this matter and that may highlight the incentives of
the relevant parties. This allows an understanding of why parties may choose
confidential arbitrations and raises some pessimism with regard to the likelihood and
feasibility of moves towards increased transparency in the future.

583
      Chayes and Chayes (1995), p. 151.
8             Summary and Outlook


8.1           Summary
Foreign direct investment is of considerable importance in bringing goods and services
across national borders to consumers and also constitutes the most important source of
external finance for developing countries. However, FDI is vulnerable to opportunistic
behaviour once the MNE has sunk the investment. This phenomenon can be
understood as a time inconsistency problem countries face or can, in game-theoretic
terms, be translated into a trust game. The international realm is mainly characterised
by its lack of supranational enforcement mechanisms. Consequently, economic actors
that care about the time inconsistency problem cannot simply “contract around” the
problem. Nevertheless, economic and legal scholars have discussed a number of
mechanisms that can support the transaction even in the absence of contract
enforcement. These mechanisms include the exchange of hostages or collateral, hands-
tying and union. Additionally, in the context of FDI, investors may buy insurance,
actively devaluate the assets to avoid expropriation and engage in lobbying. Forceful
mechanisms to achieve cooperation in a trust game are, under certain circumstances,
also what have been called here dynamic devices, most importantly repetition and
reputation. All of these mechanisms were discussed with regard to FDI and none
turned out to be a perfect substitute for domestically enforceable contract law.
Depending on the respective mechanism, problems exist concerning availability, high
transaction costs or an imperfect incentive structure. Dynamic devices may face the
endgame-problem, insufficient observability of past behaviour or high discount rates.
Following the analysis of the economic mechanisms to protect FDI, an overview on
the legal landscape in the realm of international investment was provided. It was
pointed out that Bilateral Investment Treaties are the cornerstone of modern
international investment law. By now, there are more than 2700 BITs in force and the
number of (known) treaty arbitrations has been steadily rising.
More important than the actual number of BITs is probably the fact that a couple of
empirical studies provide robust evidence that BITs can indeed attract FDI. To critics
of international law, the spread and success of BITs may be surprising – how can BITs
work despite a lack of supranational enforcement? International law and economics
has identified mainly retaliation, reciprocity and reputation as the relevant mechanisms
that ensure compliance in international law. In the case of BITs, it appears that
reciprocity and retaliation do not play significant roles. It is therefore mainly the
potential loss of reputation (also vis-à-vis treaty partners) that will make uncooperative
behaviour more costly when a BIT is in place. In addition, arbitration is backed up by



J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3_8,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
200                                                                   Summary and Outlook

multilateral treaties that in many cases allow the enforcement of awards against assets
in third countries if the (losing) defendant fails to comply with the award. Although
the actual enforcement in third countries appears to be costly and tedious, this option
will certainly raise the expected cost of breach. The (expected) costs of the BIT can be
“used” in two ways: countries may credibly commit themselves by ratifying a BIT or
countries may ratify BITs with the intention to signal hidden, positive characteristics.
Both approaches were discussed using simple, game-theoretic examples. Noteworthy,
the two approaches do not strictly exclude each other. In a sense, the commitment
model that was introduced is also at the heart of the signalling function. Exactly the
countries that can use the BIT as a commitment device or happen to be reliable from
the start might use BITs as a signal. As the models are interrelated, similar conclusions
can be drawn about the importance of credible threats (e.g., low litigation costs,
sufficiently high damages) and the availability of hostages or collateral. With regard to
signalling, it is worth mentioning that, contrary to common perception, the prior belief
on the reliability of the country needs to be low – consequently, the signal will not
necessarily be sent by low-risk countries. It must be noted that the signalling approach
suffers from a convincing explanation on the nature of asymmetric information.
Institutional quality, as may be argued, cannot be at the core of the information
asymmetry as information in this area is readily accessible. Asymmetric information
may exist if the current preferences of a country (represented by its government) are
not known to MNEs. If this is the case, it should be noted that the signal function of
BITs is rather short-lived. On the other hand, the commitment function suffers from
the incentive problems that are well-known in the context of hostages and collateral.
While the strategic interaction between (potential) MNEs and host countries is
certainly at the core of BITs, the strategic interaction between countries to attract the
MNE is also of high importance. Put differently, countries are using BITs to compete
for capital, a claim that can be empirically validated. Yet, the consequences of this
competition on welfare are dubious. Given the complicated nature of efficiency and
welfare considerations in public international law, the focus was set on developing
countries. When the treaty policy of these countries has an external effect, a treaty will
still be a Pareto-improvement for the two treaty partners, but may result in a Pareto-
inferior situation for the developing countries as group. In the case of investment
protection, it has therefore been argued in the literature that developing countries will
end up in a prisoner’s dilemma situation – a claim that is certainly true if stronger
investment protection will not lead to higher investment flows to developing countries
as a group. If this assumption is relaxed, countries may still bring themselves in a PD
situation with BITs. However, the opposite (in the sense of a Pareto-improvement)
may also be true. This discussion reflects the more general discussion on the merits
and dangers of institutional competition, a discussion which touches on a number of
Summary and Outlook                                                                   201

different issue areas. The main contributions in tax as well as ecological competition
have been reviewed, concluding that the benefits of increased government control (and
commitment) and the threat of underproviding public goods need to be balanced.
Certainly, BITs are commitment devices to start with, but additional positive welfare
effects can be expected when government incentives do not align with social welfare
goals, e.g., as a result of the influence of private interest groups. A deviation from the
social welfare goal has been part of the discussion on international trade for a long
time. Although it was argued here that the insights of the political economy of
international trade cannot be transferred one-to-one to international investment law,
redistributive pressures to the detriment of social welfare are probably nevertheless
existent. On the other hand, the underprovision hypothesis has unquestionably been
part of the discussions on international investment law under the catchword of
regulatory chill. Therefore, BITs were analyzed in light of this discussion and, to keep
the problem traceable, the case law was reviewed in a specific issue area, namely
environmental regulation. The relevant provisions in BITs certainly show extensive
elements of the non-discrimination principle. On the other hand, the importance of
good faith and the public purpose of a regulatory measure may be and has been
neglected by tribunals applying a strict interpretation of the standards of treatment and
the provisions regarding indirect expropriation. The case law regarding the specific
issue of environmental regulation indicated that the threat of a widespread regulatory
chill may be unfounded. Nevertheless, strict interpretations by tribunals that do not
take account of the good faith of the government and/or the public purpose of a
measure give rise to concern. Consequently, if the development aim of BITs is to be
taken seriously, policy-makers and tribunals should underline the lawfulness of
measures taken in the public interest. This argument does not imply a relaxation of the
commitments as such: governmental measures in favour of interest groups to the
detriment of the investor and indirectly also to the detriment of social welfare need to
be strictly sanctioned.
Closely related to the discussion on institutional competition is the question whether
BITs may have an external effect on institutional quality. This effect may either be
positive (e.g., domestic investors demand equal protection or BITs serve as a “good
example”) or negative (e.g., investors bypass domestic courts or governments
“segment” their reputation). This question was approached empirically using panel
data from 1996 to 2007 in a fixed-effects model. As the dependent variable, different
dimensions of the World Governance Indicators were used, namely regulatory quality,
government effectiveness, rule of law and corruption control. As the relevant
independent variable, the cumulative number of BITs was utilised. Control variables
included domestic investment regulation, GDP per capita and trade openness. Clearly,
the regression results show that a positive impact of BITs on institutional quality
202                                                                   Summary and Outlook

cannot be expected. With regard to the opposite, the regressions produced negative
coefficients for the BIT variable, albeit in most cases not significant and not at all
robust to model specifications. A negative effect of BITs on institutional quality
cannot therefore be entirely excluded, but also cannot be confirmed. In any case, this
result calls for further research, especially as the subjective measures of institutional
quality have aroused a lot of criticism in the literature. It might therefore be attempted
to find a better, more objective dependent variable. Interesting is the strong and
positive relation between (lagged) domestic reform (proxied by the number of
regulatory changes favourable to FDI) and institutional quality. This is an indicator
that countries can shape the quality of domestic institutional quality (or at least the
perception thereof) through regulatory measures. Given the somewhat static
explanatory variables like religion, colonial and legal heritage, language etc. that have
been identified in empirical research to shape the quality of institutions, this result
provides some optimism on the potential impact of domestic reform in this regard.
Calls for increased transparency are not only standard in international law and
international governance, but also in international investment law. The traditional
confidentiality that is actually a key feature of commercial arbitration has also shaped
investment arbitration to a certain degree. As arbitration based on BITs (or other
treaties that incorporate investment provisions like NAFTA or the Energy Charter
Treaty) frequently touches upon issues of great public interest, the calls for more
transparency have therefore a solid ground. Nevertheless, increased transparency will
also come with costs that have so far been neglected in the literature, namely the
opportunity costs of the non-availability of efficient breach and a potentially reduced
incentive to comply with an award. That the latter idea can play a role in the realm of
international law has been underlined with the example of the OAS. In addition, the
high reputational threat may have a distributional implication in the sense that in
settlement negotiations, the host country is disadvantaged by the threat of reputational
sanctions. Nevertheless, this book did not make a univocal argument in favour of more
confidentiality. The public interest argument probably outweighs the potential benefits
of confidentiality. An exception may be arbitration proceedings that primarily concern
the commercial activity of countries. Another point has also been highlighted: the
costs of more transparency are primarily carried by the parties to a dispute (who may
have the option to choose confidentiality). On the other hand, the benefits of increased
transparency, like a higher legitimacy of the system or the value of precedence, accrue
to all players in the system. This observation raises concerns as to whether moves
towards more transparency are actually feasible.
Summary and Outlook                                                                   203

8.2           Outlook
The law and economics approach has only recently started to systematically analyse
issues of international law. International investment law is no exception. This thesis
attempted to shed light on the economic functioning of BITs and other economic and
legal aspects of Bilateral Investment Treaties. Nevertheless, much work remains to be
done to better understand the structure, content and impact of international investment
law. Two potential areas for further research will shortly be sketched. The first relates
to empirical issues and the second to theoretical aspects regarding the overall
landscape of international investment law.
With regard to empirical research, it has already been mentioned that the results
derived in chapter 6 call for more research. In addition, the growing number of
investment arbitrations may open the opportunity for interesting empirical studies, for
example with regard to the relationship between characteristics of the defendant and
the number of investment arbitrations. At this point, the number of cases is too low to
obtain any meaningful results using regression techniques. However, it would certainly
enhance the understanding of international investment law if it were known what kind
of countries are prone to higher arbitration risk. A different area for empirical research
would be to use event studies to measure the impact of the arbitration mechanism on
share prices of the involved companies. This might give a hint as to how market
participants estimate the value and efficacy of the arbitration mechanism. Certainly,
the area of international investment law and economics offers opportunities for
empirical research going beyond the question of whether or not BITs attract FDI.
A different potential area for future research rather touches the general structure of
international investment law. In the mid-1990s, discussions on the necessity of a
multilateral investment treaty peaked with the attempt of the OECD to establish a
Multilateral Agreement on Investment (MAI). With the failure of this initiative, the
discussion on this topic has calmed. However, the legal landscape of international
investment law may change with the ratification of the Lisbon Treaty by the member
states of the European Union. The treaty will subsume the area of investment
protection under the Common Commercial Policy (article 188 C of the Lisbon Treaty).
It is doubtful to what extend this might influence the BIT regime of the member states.
Nevertheless, this step may trigger a new discussion on the question of a multilateral
treaty on investment. Whether a multilateral regime is indeed superior to bilateral
solutions is a topic where economic analysis can definitely make a useful contribution.
Appendices


Appendix A
The grim trigger strategy is defined as:
(1) Cooperate (here: invest (MNE) or accommodate (HC)) in the first round
(2) Continue to choose invest or accommodate in the following rounds, unless the
other player deviates, in which case play the stage game Nash Equilibrium (Don't
Invest/Expropriate) forever.


This strategy constitutes a subgame-perfect Nash-Equilibrium when both players have
no incentive to deviate from the strategy, assuming the other player plays the strategy.
So, if the investor plays the grim trigger strategy and the host state always
accommodates, payoff for the host state is:

                                         CH
                         CH ¦ t 0 G t
                               f
CH ª1  G  G 2  ...º
   ¬                 ¼                                                        (A1)
                                        1G
with
G 1/(1  i )
where G is the discount rate of the host state.


What is the payoff when the host state deviates? Because of the stationary character of
the game and the fact that future payoffs are discounted, deviation will be most
attractive in the first round of the repeated game. If the host state plays expropriation
in the first round, it will receive the opportunistic profit in the first round and, as the
investor plays the grim trigger strategy, a payoff of 0 in all following rounds. In
mathematical terms:
WH  0G  0G 2  ... WH                                                       (A2)
Unilateral deviation from the grim trigger strategy does therefore NOT pay off when:

 CH               C
    t WH œ G t 1  H                                                          (A3)
1G               WH




J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
206                                                                        Appendix A

The case for the investor is trivial. He cannot gain an opportunistic profit through
unilateral deviation. His profit will be CM in every round the host state accommodates
(which is the highest payoff he can achieve).
Appendix B
We want to show that the strategies and beliefs presented in section 3.2.2.2.2 are
indeed a perfect Bayesian equilibrium.584 The following notation holds:


 : a-priori probability that the host country is of the good type
q2: (a-posteriori) probability that the host country is of the good type as assessed by the
MNE in period two (after observing the history of the game)
z: probability that the bad type will play accommodation in the first period
v: probability that the MNE plays invest
r: probability that the HC plays accommodate in the first round




The expected profit of the MNE in the second round amounts to:
–2       q2CM  (1  q2 )( LM )                                                            (B1)
If – 2 ! 0 , the MNE will invest in the second round.
q2CM  (1  q2 )( LM ) ! 0                                                                 (B2)
solving for q2:

          LM
q2 !                                                                                        (B3)
        CM  LM

                              LM
Consequently, if T !                , there will be investment for sure in the first stage.
                            CM  LM


Let z denote the probability with which the bad host country accommodates in the first
period. The MNE in round two updates his beliefs according to Bayes' Rule. If he
observed accommodating behaviour, his belief will be:

             T
q2                                                                                          (B4)
        T  (1  T ) z




584
      This description draws on Holler and Illing (2006), p.168ff. and Feess (2004), p.624ff.
208                                                                            Appendix B

                                                                    LM
From equation (B3) we know that the MNE will invest if q2 !               . Combining
                                                                  CM  LM

equation (B3) with (B4) yields:


  LM         T
       !                                                                   (B5)
CM  LM T  (1  T ) z


It has been explained in section 3.2.2.2.2 that the only way to build a reputation is to
play mixed strategies. Therefore, the HC in round 1 will set z to make the MNE
indifferent between investment and non-investment in the following round. The MNE
will be indifferent if his expected profit is zero:
q2CM  (1  q2 )( LM ) 0                                                     (B6)
Rearranging

         LM
q2                                                                            (B7)
       CM  LM


Combining (B4) and (B7)

     T             LM
                                                                           (B8)
T  (1  T ) z   CM  LM


This is the indifference condition for the MNE in round two. Solving for z:

        T      ª CM  LM    º
z              «          1»                                                 (B9)
      (1  T ) ¬ LM         ¼

z is the probability that the bad HC must choose in round one to make the MNE
indifferent between investment and no investment in round 2. Also the MNE must play
mixed strategies to make the bad HC indifferent in the preceding round (otherwise, the
HC wouldn't play mixed strategies). Let v denote the probability with which the MNE
invests in the second period. The bad HC will be indifferent if the following equation
holds:
CH  vG WH  1  v  G 0 WH                                                  (B10)
Appendix B                                                                          209

Note that the HC could expropriate in period 1 and thus reveal his type. It would
receive its expropriation payoff, but the MNE would not invest in the second period
(right-hand side of the equation). Alternatively, the HC could accommodate and
receive CH in period 1 and expropriate in the second round. Its expected payoff then
depends on the probability with which the MNE invests in the second round.
Solving for v:

     WH  CH
v                                                                           (B11)
      G WH

This is the probability the MNE needs to play that makes the bad HC indifferent. The
HC will react with playing z as specified in equation (B9).


Until now, it was assumed that the MNE has invested in the first period so that there
could be Bayesian Updating. When will the MNE indeed invest? We denote this
probability by r. The MNE cares about the probability that the HC will play
accommodation. Because not only the good type plays accommodation, but also the
bad type with probability z, the probability r is bigger than the a-priori probability ;
specifically:


r T  (1  T ) z                                                            (B12)
Probability z in equation (B9) has already been calculated. Consequently:

                    T ª CM  LM º
r T  (1  T )            «     1»                                         (B13)
                 (1  T ) ¬ LM    ¼

simplifying to

      CM  LM
r T                                                                         (B14)
        LM


The profit for the MNE in round 1 is
S1    rCM  (1  r )( LM )                                                 (B15)
The MNE will only invest if
210                                                                                 Appendix B

rCM  (1  r )( LM ) ! 0                                                         (B16)
This simplifies to:

        LM
r!                                                                                (B17)
      LM  CM


Inserting (B14)

    CM  LM     LM
T           !                                                                     (B18)
      LM      LM  CM


rearranging

           L2
T!          M
                                                                                  (B19)
      ( LM  CM ) 2



Consequently, it can be shown that the equilibrium described in section 3.2.2.2.2 is
                                               L2
indeed a PBE of this game. If T                M
                                                        , the MNE will not invest in the first
                                          ( LM  CM ) 2

                       L2
round. If T !           M
                                , the MNE will invest in the first round as its expected profit
                  ( LM  CM ) 2

is positive. The bad HC will randomise and play accommodate with probability
         T ª CM  LM º
z              «     1» . This is exactly the probability that makes the MNE indifferent
      (1  T ) ¬ LM    ¼
between investment and no investment in the following round. If the outcome of this
randomisation process was accommodate, the MNE will update its belief according to
                                                  WH  CH
Bayes' Rule and invest with probability v                 . This probability makes sure that
                                                   G WH
the bad type is indifferent between expropriation and accommodation in the preceding
round – consequently, none of the players has an incentive to deviate from this
equilibrium.
Appendix C
We are looking for perfect Bayesian equilibria (PBE). The MNE has a prior belief ( )
on the reliability of the HC and updates this belief using Bayes' Rule, depending on the
action (BIT or No BIT) of the HC. We can differentiate between pooling, separating
and semi-separating equilibria. The following assumptions hold:


(a) The (expected) ex-post costs of a BIT are lower for the reliable type: BR<BU
(b) The investment outcome absent BITs is worthwhile to the reliable and the
unreliable type, CH>0 and WH>0.
(c) The MNE would invest in the case of the reliable HC, CM>0, but not invest in the
case of the unreliable type, LM<0.
(d) The MNE will not invest absent a BIT: T CM  (1  T ) LM  0
(e) MNEs update their beliefs using Bayes' Rule.
(f) Equilibria not fulfilling the intuitive criterion as introduced by Cho and Kreps
(1987) shall be eliminated.


1) Pooling on No BIT (1)
Imagine that CH  A  BR  0 and WH  A  BU  0 . These conditions imply that for both
types, signing a BIT is not rational even if it would induce investment. If no type signs
a BIT, no information will be transmitted to the MNE. The MNE will calculate its
payoff using the prior probability that the HC is reliable. The payoff if the MNE
invests is T CM  (1  T ) LM , not investing will generate a payoff of zero. Assumption (d)
postulates that T CM  (1  T ) LM  0 . Consequently, the MNE will not invest. Therefore,
we have a pooling equilibrium where no party signs a BIT, the MNE will not invest,
and the payoff to both parties is zero. As playing BIT is not rational for any HC with
any positive probability, equilibria in mixed strategies can also be ruled out.


2) Separating Equilibrium
Imagine that CH  A  BR ! 0 and CH  A  BU  0 and assume that the MNE will invest
when it observes a BIT, but not invest otherwise. The reliable type will sign a BIT and
the unreliable type will not. Given the MNE strategy, does any type of HC have the
incentive to deviate? No, as CH  A  BR ! 0 assures that the reliable type is better off
with a BIT and investment, while CH  A  BU  0 assures that the unreliable type
prefers no investment compared to a BIT with investment. The MNE has no incentive
212                                                                           Appendix C

to deviate from the strategy as its expected payoff T CM  (1  T )0 T CM is the highest
payoff it can obtain. As a matter of fact, any HC strategy where the unreliable type
plays BIT with a positive probability can be ruled out.


However, there could be a PBE where the strategy for both types of HC is No BIT. If
no BIT is signed, the MNE has no incentive to deviate from its earlier strategy of
never investing because it cannot differentiate between types and assumption (d)
holds. This constitutes a PBE with pooling on No BIT. Please note that this
equilibrium can be eliminated using the intuitive criterion. The problem with PBE is
that there are no restrictions on the beliefs parties have after moves that are out-of-
equilibrium. The intuitive criterion postulates that unexpected moves will not be
chosen by parties that can only lose as compared to the relevant equilibrium.585 In our
case, if the MNE observed a BIT, this would obviously be the reliable HC signing it –
therefore, the reliable type can "destroy" this pooling equilibrium – consequently, this
equilibrium can be eliminated using the intuitive criterion.


3) Pooling on No BIT (2)
Imagine CH  A  BR  0 and WH  A  BU ! 0 . In this situation the unreliable type
actually prefers a situation with the investment and a BIT, while the BIT is too costly
for the reliable type. As BR<BU, it must be the case that WH>CH, that is, the benefits of
the expropriated asset to the unreliable type must be high. Nevertheless, a situation
where only the unreliable type signs a BIT and consequently induces the MNE to
invest cannot be an equilibrium as the MNE would have an incentive to deviate (as (1-
 )LM<0). In turn, the unreliable HC will have no incentive to sign the BIT in the first
place (as –A <0). The only pure-strategy equilibrium is where both parties refrain from
signing a BIT and no investment takes place. The MNE cannot update its belief,
assumption (d) assures that it will not invest and thus not deviate from this strategy.
The reliable type has no incentive to sign a BIT and "destroy" the pooling equilibrium
as CH  A  BR  0 , the unreliable type has no incentive to sign a BIT and reveal itself
as the unreliable type as concluding a BIT is costly (–A <0) and the MNE would
optimally react by not investing.


4) Pooling on No BIT (3)
Given the following two conditions


585
      Holler and Illing (2006), p.185.
Appendix C                                                                           213

CH  A  BR ! 0                                                             (C1)
and
WH  A  BU ! 0                                                             (C2)
even the unreliable type would sign a BIT if it can induce investment. However, the
only pure strategies equilibrium is a pooling equilibrium on No BIT. As a first step,
the two other potential pure-strategies can be eliminated. First, imagine a separating
equilibrium where only the reliable types sign BITs and the MNE invests if it observes
a BIT, but does not observe otherwise. The unreliable type has an incentive to deviate
from its strategy as it has the incentive to "bluff" as long as WH  A  BU ! 0 . For the
same reason there cannot be a separating equilibrium where only the unreliable type
signs a BIT.
Now, imagine a pooling equilibrium where both types sign a BIT. If both types sign a
BIT, there cannot be Bayesian Updating. The MNE will calculate its payoff using the
prior probability that the HC is reliable. Yet, according to assumption (d), the MNE
will not invest. Consequently, no HC will bother to send the costly signal.
The only pure strategies equilibrium given conditions (C1) and (C2) can be the
equilibrium where the HCs both do not sign a BIT and the MNE does not invest. The
strategy of the MNE is never to invest, while the strategy of the HCs is never to sign a
BIT. No party has an incentive to deviate from this equilibrium – equally, this
equilibrium cannot be eliminated using the intuitive criterion.


5) Semi-Separating Equilibrium
Please note that under the assumptions (C1) and (C2), there can be a PBE in mixed
strategies under the following additional condition:
CH  A  BR ! WH  A  BU                                                   (C3)
In a semi-separating equilibrium, a reliable type will sign a BIT while an unreliable
type randomises between signing and non-signing. The MNE will not be able to
differentiate between the types and randomise itself unless the unreliable clearly
identifies itself by not signing a BIT.


Let q with 0<q<1 be the probability with which the unreliable type randomises
between signing or not signing a BIT and s with 0<s<1 the probability with which the
MNE randomises its decision to invest. Both players set their probabilities in order to
make the other player be indifferent between its choices. Therefore, the payoff of the
unreliable type depends on the MNEs randomisation decision:
214                                                                            Appendix C

s (WH  A  BU )  (1  s)( A) 0 œ


         A
s                                                                            (C4)
      WH  BU


Equally, the unreliable HC will choose q to make the MNE indifferent. The MNE
updates its belief using Bayes' Rule. If the MNE observes a BIT, its updated belief
given the prior probability is:

       T
                                                                             (C5)
(T  q(1  T ))


In words: the probability that the HC is reliable - after observing a BIT - is the prior
probability that the HC is reliable ( ) multiplied with the probability that a reliable HC
signs a BIT (1), divided by the prior probability of a BIT being signed (T  q(1  T )) .
The converse probability, that is the probability that the HC is unreliable despite
having signed a BIT, is therefore

   q(1  T )
                                                                             (C6)
(T  q(1  T ))


The expected investment payoff of the MNE after observing a BIT can be written as

       T                       q(1  T )
                  (CM )                    ( LM )                           (C7)
(T  q(1  T ))             (T  q(1  T ))


If no investment takes place, the MNE will receive a payoff of zero. It is consequently
indifferent between investing and not investing if

       T                       q(1  T )
                  (CM )                    ( LM ) 0                         (C8)
(T  q(1  T ))             (T  q(1  T ))


solving for q yields (given 0< <1)

         CM T
q                                                                            (C9)
       LM (1  T )
Appendix C                                                                              215

                                                              CM T
The unreliable type will sign a BIT with probability q                   . If the unreliable
                                                            LM (1  T )

                                                              A
type signs a BIT, the MNE will invest with probability s           . If no BIT is being
                                                           CH  BU

signed, the MNE will not invest. For this equilibrium to be stable, the reliable type
must also not have an incentive from deviating from its strategy of always signing a
BIT. Theoretically, he may be tempted to refrain from always doing so and saving the
costs of signing a BIT. Will this increase his expected payoff? When the reliable type
doesn't sign a BIT, the MNE will not invest. The payoff to the reliable HC would be
zero. Imagine that the reliable HC would also randomise and sign a BIT with
probability u, where 0 d u d 1 . Given the equilibrium strategy of the MNE, the expected
payoff would be

u ( s (CH  A  BR )  (1  s )( A))  (1  u )0 .                          (C10)
It is known from (C3) that CH  A  BR ! WH  A  BU and also
s (WH  A  BU )  (1  s)( A) 0 .                                          (C11)
and therefore
s (CH  A  BR )  (1  s)(  A) ! 0                                         (C12)
It follows that u=1 is the optimal strategy for the reliable HC as the payoff of the
reliable type is increasing in u. The reliable type has no incentive to refrain from
signing a BIT in the presented equilibrium. Please note that without condition (C3),
the reliable type could have an incentive to destroy this mixed equilibrium by playing
No BIT.
But can a "reverse" mixed strategies equilibrium exist in the sense that the unreliable
type signs a BIT with probability 1, while the reliable type randomises with 0<q<1 (if
we also change condition (C3) to CH  A  BR  WH  A  BU )? The answer must be no
because any strategy where the unreliable type plays BIT with a probability 1 can
never increase the MNE's payoff compared to not investing at all. Put differently, if
the unreliable type played BIT for sure, while the reliable type played BIT only with
probability 0<q<1, the MNE would know for sure that he is facing a reliable type
when it observes No BIT. The strategy of the reliable HC to play BIT only with a
certain probability is therefore, given the beliefs of the MNE, not optimal.
Appendix D
Correlation Matrix
Appendix D                                                                                                    217


                               Dependent Variable: Government Effectiveness
                                    -I-          -II-         -III-       -IV-          -V-          -VI-
             BITs                -0.0482
             (log)               (0.0341)
          BITs in t-1                          -0.0570*     -0.0415      -0.0437                  -0.0730***
            (log)                              (0.0324)     (0.0344)     (0.0324)                   (0.0270)
             BITs                                                                     0.00150
      (averaged 3 years)                                                             (0.00267)
     Domestic Regulation        0.0825***
           (log)                 (0.0292)
  Domestic Regulation in t-1                   0.0564**                 0.0744***                   0.0225
           (log)                               (0.0267)                  (0.0269)                  (0.0222)
     Domestic Regulation                                                              0.00245
      (averaged 3 years)                                                             (0.00344)
       GDP per Capita            0.747***
           (log)                  (0.128)
    GDP per Capita in t-1                      0.569***     0.553***    0.441***      0.542***     0.438***
           (log)                               (0.114)      (0.109)      (0.134)      (0.115)      (0.105)
       Development Aid         0.00492***
          (%GNI)                (0.00151)
    Development Aid in t-1                    0.00448***   0.00586***   0.00390**    0.00577***   0.00432***
          (%GNI)                               (0.00149)    (0.00177)   (0.00153)     (0.00173)    (0.00147)
                                -0.00176**
       Trade Openness
                                (0.000701)
                                              -0.00161**   -0.00162*    -0.000801     -0.00129    -0.00198**
    Trade Openness in t-1
                                              (0.000778)   (0.000827)   (0.000762)   (0.000817)   (0.000769)
          FDI Inflow              0.0140
            (log)                (0.0108)
       FDI Inflow in t-1                      0.0241**     0.0205**                  0.0235**     0.0217**
             (log)                            (0.00951)    (0.00941)                 (0.00984)    (0.00955)
                                   0.112
      ICSID Convention
                                 (0.0848)
                                                 0.132        0.143       0.163*        0.119        0.128
   ICSID Convention in t-1
                                               (0.0906)     (0.0911)     (0.0887)     (0.0922)     (0.0887)
                                  -5.990***   -4.858***    -4.606***    -3.632***    -4.700***    -3.661***
           Constant
                                    (0.886)     (0.774)      (0.765)      (0.935)      (0.786)      (0.640)
         Observations                 931        1044         1123         1104         1062         1044
     Number of Countries              133         133          140          133          133          133
          Time Effects                Yes         Yes          Yes          Yes          Yes          No
      R-squared (within)             0.148       0.127        0.102        0.093        0.111        0.105
 *** p<0.01, ** p<0.05, * p<0.1
 Robust standard errors in parentheses
218                                                                                                  Appendix D


                                    Dependent Variable: Economic Freedom
                                         -I-         -II-        -III-       -IV-        -V-         -VI-
                  BITs                -0.0921
                  (log)               (0.697)
               BITs in t-1                          -0.0333     -0.0848     -0.0897                -0.170
                 (log)                              (0.642)     (0.618)     (0.641)                (0.571)
                  BITs                                                                  0.0697
           (averaged 3 years)                                                          (0.0523)
          Domestic Regulation           0.440
                (log)                  (0.568)
       Domestic Regulation in t-1                    0.552                   0.561                 0.0461
                (log)                               (0.526)                 (0.547)                (0.453)
          Domestic Regulation                                                           -0,0244
           (averaged 3 years)                                                           -0,0409
            GDP per Capita            15.38***
                (log)                  -3.002
         GDP per Capita in t-1                     15.78***    14.41***    15.44***    16.67***    12.19***
                (log)                               -2.854      -2.836      -2.531      -2.779      -2.236
            Development Aid           -0.0451
               (%GNI)                 (0.0552)
         Development Aid in t-1                    -0.0265     -0.0407     -0.0420     -0.0225     -0.0314
               (%GNI)                              (0.0468)    (0.0426)    (0.0485)    (0.0463)    (0.0396)
                                      0.00561
            Trade Openness
                                      (0.0162)
                                                    0.0118      0.0197      0.0122      0.0165     0.00447
         Trade Openness in t-1
                                                   (0.0171)    (0.0165)    (0.0164)    (0.0169)    (0.0177)
               FDI Inflow               0.186
                 (log)                 (0.217)
            FDI Inflow in t-1                      0.331**     0.374**                  0.318*     0.283*
                  (log)                            (0.163)     (0.169)                  (0.163)    (0.145)
                                       1.467
           ICSID Convention
                                       -1.223
                                                    0.841       0.762        1.060       0.810       0.664
        ICSID Convention in t-1
                                                    -1.186      -1.204      -1.174      -1.254      -1.213
                                       -57.57***   -62.41***   -54.97***   -54.53***   -68.26***   -35.12**
                Constant
                                        (20.43)      (19.50)     (20.29)     (17.40)     (19.12)    (14.59)
              Observations               1222         1322        1380        1393        1339       1322
          Number of Countries             123          123         127         124         123        123
               Time Effects               Yes          Yes         Yes         Yes         Yes        No
           R-squared (within)            0.216        0.235       0.226       0.228       0.243      0.202
      *** p<0.01, ** p<0.05, * p<0.1
      Robust standard errors in parentheses
Appendix D                                                                                                219


                               Dependent Variable: Regulatory Quality (OECD BITs)
                                   -I-          -II-        -III-       -IV-         -V-         -VI-
        OECD BITs               -0.0513
          (log)                 (0.0472)
     OECD BITs in t-1                        -0.0804*     -0.0762      -0.0752                 -0.153***
         (log)                               (0.0459)     (0.0479)    (0.0466)                 (0.0490)
            BITs                                                                   0.00472
        (sum 3 years)                                                             (0.00732)
    Domestic Regulation         0.107***
          (log)                 (0.0404)
  Domestic Regulation in t-1                 0.0988***                0.133***                  -0.0192
           (log)                             (0.0372)                 (0.0392)                 (0.0293)
    Domestic Regulation                                                            0.00352
      (sum 3 years)                                                               (0.00532)
       GDP per Capita           0.990***
           (log)                 (0.193)
    GDP per Capita in t-1                    0.851***     0.805***    0.740***     0.835***    0.539***
           (log)                              (0.180)     (0.168)      (0.180)     (0.182)      (0.156)
      Development Aid          0.00862***
         (%GNI)                 (0.00291)
   Development Aid in t-1                    0.00624**   0.00733***   0.00586**   0.00765***   0.00581**
         (%GNI)                              (0.00263)   (0.00225)    (0.00274)   (0.00239)    (0.00233)
                                -0.000916
      Trade Openness
                                (0.00115)
                                             -0.000668   -0.000704    -9.86e-05   -0.000341    -0.00138
   Trade Openness in t-1
                                             (0.00133)   (0.00127)    (0.00137)   (0.00136)    (0.00136)
         FDI Inflow                          0.0318***    0.0306**                 0.0315**    0.0305**
           (log)                             (0.0118)     (0.0125)                 (0.0125)    (0.0122)
      FDI Inflow in t-1        0.0461***
            (log)               (0.0124)
                                -0.0226
     ICSID Convention
                                (0.0834)
                                              0.00725      0.0131      0.0490      -0.0126      -0.0135
  ICSID Convention in t-1
                                              (0.0847)    (0.0874)     (0.0811)    (0.0838)     (0.0863)
                                 -8.411***   -6.631***   -6.195***    -5.811***   -6.554***    -4.241***
          Constant
                                   -1436       -1249       -1205        -1306       -1274        -1035
        Observations                932         1045        1124         1105        1063         1045
    Number of Countries             133          133         140          133         133          133
        Time Effects                Yes          Yes         Yes          Yes         Yes          No
     R-squared (within)            0.211        0.171       0.132        0.154       0.147        0.083
*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses
Bibliography

Abbott, K. W. (1993). ""Trust But Verify": The Production of Information in Arms
Control Treaties and Other International Agreements." Cornell International Law
Journal 26: 1-58.

Aisbett, E. (2007). "Bilateral Investment Treaties and Foreign Direct Investment:
Correlation versus Causation." MPRA Paper No. 2255.

Aitken, B. J. and A. E. Harrison (1999). "Do Domestic Firms Benefit from Direct
Foreign Investment? Evidence from Venezuela." American Economic Review 89(3):
605-618.

Akerlof, G. (1970). "The Market for "Lemons": Qualitative Uncertainty and the
Market Mechanism." Quarterly Journal of Economics 47: 448-500.

Alchian, A. A. and H. Demsetz (1972). "Production, Information Costs and Economic
Organization." American Economic Review 62: 777-795.

Alfaro, L., C. Areendam, et al. (2003). "FDI and Economic Growth: The Role of Local
Financial Markets." Journal of International Economics 61(1): 512-533.

Bagwell, K. and R. W. Staiger (2002). The Economics of the World Trading System.
Cambridge, Massachusetts, The MIT Press.

Balasubramanyam, V. N., M. Salisu, et al. (1996). "Foreign Direct Investment and
Growth in EP and IS Countries." Economic Journal 106(434): 92-105.

Baldwin, E., M. Kantor, et al. (2006). "Limits to Enforcement of ICSID Awards."
Journal of International Arbitration 23(1): 1-24.

Baldwin, R. E. and P. Krugman (2004). "Agglomeration, Integration and Tax
Harmonisation." European Economic Review 48: 1-23.

Banks, K. (1999). "NAFTA's Article 1110 - Can Regulation be Expropriation?" Law
and Business Review of the Americas 5: 499-521.

Barro, R. J. (2003). "Determinants of Economic Growth in a Panel of Countries."
Annals of Economics and Finance 4: 231-274.

Basinger, S. J. and M. Hallerberg (2004). "Remodeling the Competition for Capital:
How Domestic Politics Erases the Race to the Bottom." American Political Science
Review 98(2): 261-276.

Becker, G. (1998). What's Wrong with a Centralized Europe? Plenty. Business Week
(June 29, 1998).


J. P. Sasse, An Economic Analysis of Bilateral Investment Treaties,
DOI 10.1007/ 978-3-8349-6185-3,
© Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011
222                                                                       Bibliography

Behrens, P. (2006). Wirtschaftsverfassungsrechtliche Ansätze im völkerrechtlichen
Investitionsschutz. Recht und spontane Ordnung. C. Engel and W. Möschel. Baden-
Baden, Nomos: 55-81.

Behrens, P. (2009). Economic Law Between Harmonization and Competition: The
Law & Economics Approach. Economic Law as an Economic Good. K. M. Meessen,
M. Bungenberg and A. Puttler. München, Sellier European Law Publisher: 45-60.

Berkowitz, D., J. Moenius, et al. (2005). "Legal Institutions and International Trade
Flows." University of Michigan International Law Review 26(1): 163-198.

Bertrand, M., E. Duflo, et al. (2004). "How Much Should We Trust Differences-in-
Differences Estimates?" The Quarterly Journal of Economics 199(1): 249-275.

Beviglia Zampetti, A. and P. Sauvé (2007). International Investment. Research
handbook in international economic law. A. Guzman and A. O. Sykes. Cheltenham,
Edward Elgar: 211-270.

Blackaby, N., J. Paulsson, et al. (2004). Guide to ICSID Arbitration. The Hague,
Kluwer Law International.

Blonigen, B. A. (2001). "In Search of Substitution Between Foreign Production and
Exports." Journal of International Economics 53: 81-104.

Blume, L., D. L. Rubinfeld, et al. (1984). "The Taking of Land: When Should
Compensation Be Paid?" Quarterly Journal of Economics 99: 71-92.

Borensztein, E., J. De Gregorio, et al. (1998). "How Does Foreign Direct Investment
Affect Economic Growth?" Journal of International Economics 45: 115-135.

Brennan, G. and J. M. Buchanan (1980). Analytical Foundations of a Fiscal
Constitution. Cambridge, Cambridge University Press.

Bubb, R. and S. Rose-Ackerman (2007). "BITs and Bargains: Strategic Aspects of
Bilateral and Multilateral Regulation of Foreign Investment." International Review of
Law and Economics 27(3): 291-311.

Buchanan, J. M. and G. Tullock (1962). The Calculus of Consent: Logical
Foundations of Constitutional Democracy. Ann Arbor, University of Michigan Press.

Burke-White, W. W. (2008). "The Argentine Financial Crisis: State Liability Under
BITs and the Legitimacy of the ICSID System." Working Paper, available under
http://ssrn.com/abstract=1088837 (November 2nd, 2010).

Busse, M. and J. L. Groizard (2008). "Foreign Direct Investment, Regulations and
Growth." The World Economy: 861-886.
Bibliography                                                                   223

Busse, M. and C. Hefeker (2007). "Political Risk, Institutions and Foreign Direct
Investment." European Journal of Political Economy 23: 397-415.

Busse, M., J. Königer, et al. (2008). "FDI Promotion through Bilateral Investment
Treaties: More Than a Bit?" Kiel Working Paper No. 1403.

Büthe, T. and H. V. Milner (2008). "The Politics of Foreign Direct Investment into
Developing Countries: Increasing FDI through Policy Commitment via Trade
Agreements and Investment Treaties?" American Journal of Political Science 52(4):
741-762.

Buys, C. G. (2003). "The Tensions Between Confidentiality and Transparency in
International Arbitration." The American Review of International Arbitration 14(1):
121-138.

Cameron, A. C. and P. K. Trivedi (2009). Microeconometrics Using Stata. College
Station, Stata Press.

Carkovic, M. and R. Levine (2005). Does Foreign Direct Investment Accelerate
Economic Growth? Does Foreign Direct Investment Promote Development? T. H.
Moran, E. M. Graham and M. Blomström. Washington, Institute for International
Economics: 195-220.

Cass, R. A. (1997). Introduction: Economics and International Law. Economic
Dimensions in International Law. J. S. Bhandari and A. O. Sykes. Cambridge,
Cambridge University Press: 1-42.

Chayes, A. and A. H. Chayes (1995). The New Sovereignty - Compliance with
International Regulatory Agreements. Cambridge, Harvard University Press.

Cho, I.-K. and D. M. Kreps (1987). "Signaling Games and Stable Equilibria." The
Quarterly Journal of Economics 102(2): 179-221.

Choudhury, B. (2005). "Evolution or Devolution? Defining Fair and Equitable
Treatment in International Investment Law." The Journal of World Investment and
Trade 6(2): 297-320.

Chukwumerije, O. (2007). "Interpreting Most-Favoured-Nation Clauses in Investment
Treaty Arbitrations." The Journal of World Investment and Trade 8(5): 597-646.

Coase, R. (1937). "The Nature of the Firm." Economica 1: 386-405.

Collier, J. and V. Lowe (1999). The Settlement of Disputes in International Law.
Oxford, Oxford University Press.
224                                                                       Bibliography

Comeaux, P. E. and N. S. Kinsella (1997). Protecting Foreign Investment Under
International Law: Legal Aspects of Political Risk. New York, Oceana Publications,
Inc.

Commission, J. P. (2007). "Precedent in Investment Treaty Arbitration: The Empirical
Backing." Transnational Dispute Management 4(5): 1-7.

Cooter, R., S. Marks, et al. (1982). "Bargaining in the Shadow of the Law: A Testable
Model of Strategic Behavior." The Journal of Legal Studies 11(2): 225-251.

Cooter, R. and D. L. Rubinfeld (1989). "Economic Analysis of Legal Disputes and
Their Resolution." Journal of Economic Literature XXVII: 1067-1097.

Cooter, R. and T. Ulen (2004). Law and Economics. Boston, Pearson Addison Wesley.

Dammann, J. and H. Hansmann (2008). "Globalizing Commercial Litigation." Cornell
Law Review 94(1): 1-72.

Daniels, R. J. (2004). "Defecting on Development: Bilateral Investment Treaties and
the Subversion of the Rule of Law in the Developing World." Working Paper.

De Backer, K. and L. Sleuwaegen (2003). "Does Foreign Direct Investment Crowd
Out Domestic Entrepreneurship?" Review of Industrial Organization 22: 67-84.

Dixit, A. K. (2004). Lawlessness and Economics: Alternative Models of Governance.
Princeton, Princeton University Press.

Djankov, S., R. La Porta, et al. (2002). "The Regulation of Entry." The Quarterly
Journal of Economics CXVII(1): 1-37.

Djankov, S., J. G. Montalvo, et al. (2005). "The Curse of Aid." World Bank Working
Paper, available under http://ideas.repec.org/p/upf/upfgen/870.html (November 2nd,
2010).

Dolzer, R. (1985). Eigentum, Enteignung und Entschädigung im geltenden
Völkerrecht. Berlin, Springer Verlag.

Dolzer, R. (2005). "The Impact of International Investment Treaties on Domestic
Administrative Law." New York University Journal of International Law and Politics
37(4): 953-972.

Dolzer, R. and C. H. Schreuer (2008). Principles of International Investment Law.
Oxford, Oxford University Press.

Dolzer, R. and M. Stevens (1995). Bilateral Investment Treaties. The Hague, Kluwer
Law International.
Bibliography                                                                     225

Drazen, A. (2002). Political Economy in Macroeconomics. Princeton, Princeton
University Press.

Dreher, A. and S. Voigt (2008). "Does Membership in International Organizations
Increase Governments' Credibility? Testing the Effects of Delegating Powers." KOF
Working Papers No. 193.

Drukker, D. M. (2003). "Testing for Serial Correlation in Linear Panel-data Models."
The Stata Journal 3(2): 168-177.

Dunning, J. H. and S. M. Lundan (2008). Multinational Enterprises and the Global
Economy. Cheltenham, Edward Elgar.

Dunoff, J. L. and J. P. Trachtman (1999). "Economic Analysis of International Law."
Yale Journal of International Law 24: 1-59.

Eaton, J. and M. Gersovitz (1983). County risk: Economic Aspects. Managing
International Risk. R. J. Herring. New York, Cambridge University Press: 75-108.

Eaton, J. and M. Gersovitz (1984). "A Theory of Expropriation and Deviations from
Perfect Capital Mobility." The Economic Journal 94: 16-40.

Edwards, J. and M. Keen (1996). "Tax Competition and Leviathan." European
Economic Review 40: 113-134.

Eger, T. (1995). Eine ökonomische Analyse von Langzeitverträgen. Marburg,
Metropolis-Verlag.

Eger, T. and H.-J. Wagener (2009). Europäische Integration - Wirtschaft und Recht,
Geschichte und Politik. München, Verlag Vahlen.

Egger, P. and M. Pfaffermayr (2004). "The Impact of Bilateral Investment Treaties on
Foreign Direct Investment." Journal of Comparative Economics 32(4): 788-804.

Eggert, W. and P. B. Sørensen (2008). "The Effects of Tax Competition When
Politicians Create Rents to Buy Political Support." Journal of Public Economics 92:
1142-1163.

Ekelund, R. B. J. and R. D. Tollison (2001). The Interest-group Theory of
Government. The Elgar Companion to Public Choice. W. F. Shugart II and L.
Razzolini. Cheltenham, Edward Elgar: 357-378.

Elkins, Z., A. Guzman, et al. (2006). "Competing for Capital: The Diffusion of
Bilateral Investment Treaties, 1960-2000." International Organization 60(4): 811-846.

Engel, C. (2008). "Governments in Dilemma - A Game Theoretic Model for the
Conclusion of Bilateral Investment Treaties - A Comment on Competing for Capital."
University of Illinois Law Review 1: 305-318.
226                                                                       Bibliography

Ethier, W. J. and J. R. Markusen (1996). "Multinational Firms, Technology Diffusion
and Trade." Journal of International Economics 41: 1-28.

Farber, D. A. (2002). "Rights as Signals." Journal of Legal Studies XXXI: 83-98.

Fauchald, O. K. (2008). International Investment Law and Environmental Protection.
Yearbook of International Environmental Law 2007. O. K. Fauchald and D. Hunter.
Oxford, Oxford University Press: 3-47.

Feess, E. (2004). Mikroökonomie: eine spieltheoretisch- und anwendungsorientierte
Einführung. Marburg, Metropolis.

Fischer, S. (1980). "Dynamic Inconsistency, Cooperation and the Benevolent
Dissembling Government." Journal of Economic Dynamics and Control 2: 93-107.

Fortier, Y. L. and S. L. Drymer (2004). "Indirect Expropriation in the Law of
International Investment: I Know It When I See It, or Caveat Investor." ICSID
Review: Foreign Investment Law Journal 19: 293-327.

Fracassi, F. (2001). "Confidentiality and NAFTA Chapter 11 Arbitrations." Chicago
Journal of International Law 2: 213-222.

Franck, S. D. (2005). "The Legitimacy Crisis in Investment Treaty Arbitration:
Privatizing Public International Law Through Inconsistent Decisions." Fordham Law
Review 73: 1521-1625.

Franck, S. D. (2007). "Empirically Evaluating Claims About Investment Treaty
Arbitration." North Carolina Law Review 86: 1-79.

Franck, S. D. (2007). "Foreign Direct Investment, Investment Treaty Arbitration, and
the Rule of Law." Global Business & Development Law Journal 19: 337-372.

Franck, S. D. (2008). "Empiricism and International Law: Insights for Investment
Treaty Dispute Resolution." Virginia Journal of International Law 48: 767-815.

Franck, T. M. (1988). "Legitimacy in the International System." American Journal of
International Law 82(4): 705-760.

Fudenberg, D. and J. Tirole (1991). Game Theory. Cambridge, Massachusetts, The
MIT Press.

Gallagher, K. P. and M. B. L. Birch (2006). "Do Investment Agreements Attract
Investment? Evidence from Latin America." Journal of World Investment and Trade
7(6): 961-974.

Ganguly, S. (1999). "The Investor-State Dispute Mechanism (ISDM) and a
Sovereign's Power to Protect Public Health." Columbia Journal of Transnational Law
38: 113-168.
Bibliography                                                                       227

Gibbons, R. (1992). A Primer in Game Theory. London, Prentice Hall.

Ginsburg, T. (2005). "International Substitutes for Domestic Institutions: Bilateral
Investment Treaties and Governance." International Review of Law and Economics
25(1): 107-123.

Glaeser, E. L., R. La Porta, et al. (2004). "Do Institutions Cause Growth?" Journal of
Economic Growth 9(3): 271-303.

Goldsmith, J. L. and E. A. Posner (1999). "A Theory of Customary International
Law." University of Chicago Law Review 66: 1133-1177.

Goldsmith, J. L. and E. A. Posner (2005). The Limits of International Law. Oxford,
Oxford University Press.

Grosse, R. and L. J. Trevino (2005). "New Institutional Economics and FDI Location
in Central and Eastern Europe." Management International Review 45(2): 123-145.

Gruner, D. M. (2003). "Accounting for the Public Interest in International Arbitration:
The Need for Procedural and Structural Reform." Columbia Journal of Transnational
Law 41: 923-963.

Guzman, A. (1998). "Why LDCs Sign Treaties that Hurt Them: Explaining the
Popularity of Bilateral Investment Treaties." Virginia Journal of International Law 38:
639-688.

Guzman, A. (2002). "A Compliance-Based Theory of International Law." California
Law Review 90: 1823-1887.

Guzman, A. (2005). "The Design of International Agreements." The European Journal
of International Law 16(4): 579-612.

Guzman, A. (2008a). How International Law Works - A Rational Choice Theory.
Oxford, Oxford University Press.

Guzman, A. (2008b). "International Tribunals: A Rational Choice Analysis."
University of Pennsylvania Law Review.

Hall, R. E. and C. I. Jones (1999). "Why Do Some Countries Produce So Much More
Output Per Worker Than Others?" The Quarterly Journal of Economics 114: 83-116.

Hallward-Driemeier, M. (2003). "Do Bilateral Investment Treaties Attract FDI? Only
a Bit...and They Could Bite." World Bank Policy Research Working Paper No. 3121.

Harris, D. J. (1998). Regional Protection of Human Rights: The Inter-American
Achievement. The Inter-American System of Human Rights. D. J. Harris and S.
Livingstone. Oxford, Clarendon Press: 1-29.
228                                                                          Bibliography

Hathaway, O. A. (2002). "Do Human Rights Treaties Make a Difference?" The Yale
Law Journal 111(8): 1935-2042.

Hellman, J. S., G. Jones, et al. (2003). "Far from Home: Do Foreign Investors Import
Higher Standards of Governance in Transition Economies?" Working Paper, available
under http://papers.ssrn.com/sol3/papers.cfm?abstract_id=386900 (November 2nd,
2010).

Hennart, J.-F. (2001). Theories of the multinational enterprise. Oxford Handbook of
International Business. A. M. Rugman and T. L. Brewer. Oxford, Oxford University
Press: 127-149.

Herdegen, M. (2005). Internationales Wirtschaftsrecht: ein Studienbuch. München,
Beck.

Hindelang, S. (2004). "Bilateral Investment Treaties, Custom and a Healthy
Investment Climate - The Question of Whether BITs Influence Customary
International Law Revisited." The Journal of World Investment and Trade 5(5): 789-
809.

Hirschman, A. O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms,
Organizations, and States. Cambridge, Harvard University Press.

Holler, M. J. (1993). "Fighting Pollution When Decisions Are Strategic." Public
Choice 76: 347-356.

Holler, M. J. and G. Illing (2006). Einführung in die Spieltheorie. Berlin, Springer.

Hsu, L. (2006). "MFN and Dispute Settlement - When the Twain Meet." The Journal
of World Investment and Trade 7(1): 25-37.

IISD     (2001).    Private    Rights,     Public   Problems,     available        under
http://www.iisd.org/pdf/trade_citizensguide.pdf (November 2nd, 2010).

Jain, S. C. and S. Vachani (2006). The Role of MNCs in Alleviating Global Poverty.
Multinational Corporations and Global Poverty Reduction. S. C. Jain and S. Vachani.
Cheltenham, Edward Elgar: 3-28.

Janeba, E. (1998). "Tax Competition in Imperfectly Competitive Markets." Journal of
International Economics 44: 135-153.

Jensen, N. M. (2006). Nation-States and the Multinational Corporation. Princeton,
Princeton University Press.

Karl, J. (2008). International Investment Arbitration: A Threat to State Sovereignty?
Redefining Sovereignty in International Economic Law. W. Shan, P. Simons and D.
Singh. Oxford and Portland, Oregon, Hart Publishing.
Bibliography                                                                       229

Kaufmann, D., A. Kraay, et al. (2008). "Governance Matters VII: Aggregate and
Individual Governance Indicators, 1996-2007." World Bank Policy Research Working
Paper No. 4654.

Kekic, L. and K. Sauvant (2007). World Investment Prospects to 2011 - Foreign Direct
Investment and the Challenge of Political Risk, Economist Intelligence Unit.

Kennedy, P. (2003). A Guide to Econometrics. Cambridge, Massachusetts, The MIT
Press.

Keohane, R. O. (2002). "Rational Choice Theory and International Law: Insights and
Limitations." Journal of Legal Studies 31(1): 307-319.

Kerber, W. and V. Vanberg (1994). "Institutional Competition Among Jurisdictions:
An Evolutionary Approach." Constitutional Political Economy 5(2): 193-219.

Kirstein, R. (1999). Imperfekte Gerichte und Vertragstreue: eine ökonomische Theorie
richterlicher Entscheidungen. Wiesbaden, Deutscher Universitätsverlag.

Kishoiyian, B. (1994). "The Utility of Bilateral Investment Treaties in the Formulation
of Customary International Law." Northwestern Journal of International Law &
Business 14: 327-375.

Knack, S. (2001). "Aid Dependence and the Quality of Governance: Cross-Country
Empirical Tests." Southern Economic Journal 68(2): 310-329.

Knack, S. and P. Keefer (1995). "Institutions and Economic Performance: Cross-
Country Tests Using Alternative Institutional Measures." Economics and Politics 7(3):
207-227.

Knahr, C. (2007). "Transparency, Third Party Participation and Access to Documents
in International Investment Arbitration." Arbitration International 23(2): 327-355.

Knahr, C. and A. Reinisch (2007). "Transparency versus Confidentiality in
International Investment Arbitration - The Biwater Gauff Compromise." The Law and
Practice of International Courts and Tribunals 6: 91-118.

Kobrin, S. (1979). "Political Risk: A Review and Reconsideration." Journal of
International Business Studies 10(1): 67-80.

Kobrin, S. (2005). "The Determinants of Liberalization of FDI Policy in Developing
Countries: A Cross-sectional Analysis, 1992-2001." Transnational Corporations 14(1):
67-104.

Koh, H. H. (1996). "Transnational Legal Process." Nebraska Law Review 75: 181-
207.
230                                                                      Bibliography

Kreps, D. M. (1990). Corporate Culture and Economic Theory. Perspectives on
Positive Political Economy. J. E. Alt and K. A. Shepsle. Cambridge, Cambridge
University Press: 90-143.

Kreps, D. M. and R. Wilson (1982). "Reputation and Imperfect Information." Journal
of Economic Theory 27: 253-279.

Kriebaum, U. (2007b). "Regulatory Takings: Balancing the Interests of the Investor
and the State." The Journal of World Investment and Trade 8(5): 717-744.

Kriebaum, U. (2008). Eigentumsschutz im Völkerrecht - Eine vergleichende
Untersuchung zum internationalen Investitionsrecht sowie zum Menschenrechtsschutz.
Berlin, Duncker & Humblot.

Kronman, A. T. (1985). "Contract Law and the State of Nature." The Journal of Law,
Economics and Organization 1(1): 5-32.

Krugman, P. and M. Obstfeld (2009). International Economics: Theory and Policy.
Boston, Addison-Wesley.

Kunce, M. and J. F. Shogren (2005). "On Interjurisdictional Competition and
Environmental Federalism." Journal of Environmental Economics and Management
50: 212-224.

Kunoy, B. (2006). "Singing in the Rain - Developments in the Interpretation of
Umbrella Clauses." The Journal of World Investment and Trade 7(2): 275-300.

Kurtz, J. (2004). "The MFN Standard and Foreign Investment - An Uneasy Fit." The
Journal of World Investment and Trade 5(6): 861-886.

Kurtz, M. J. and A. Schrank (2007). "Growth and Governance: Models, Measures, and
Mechanisms." The Journal of Politics 69(2): 538-554.

Kwok, C. C. and S. Tadesse (2006). "The MNC as an Agent of Change for Host-
country Institutions: FDI and Corruption." Journal of International Business Studies
37: 767-785.

Kydland, F. E. and E. C. Prescott (1977). "Rules Rather than Discretion: The
Inconsistency of Optimal Plans." Journal of Political Economy 85(3): 473-491.

La Porta, R., F. Lopez-de-Silanes, et al. (1999). "The Quality of Government." The
Journal of Law, Economics and Organization 15(1): 222-279.

Li, X. and X. Liu (2005). "Foreign Direct Investment and Economic Growth: An
Increasingly Endogenous Relationship." World Development 33(3): 393-407.

Lipsey, R. E. (2002). "Home and Host Country Effects of FDI." NBER Working Paper
9239, available under http://www.nber.org/papers/w9293 (November 2nd, 2010).
Bibliography                                                                    231

Lipsey, R. E. and M. Y. Weiss (1984). "Foreign Production and Exports of Individual
Firms." The Review of Economics and Statistics 66: 304-308.

Lowenfeld, A. F. (2002). International Economic Law. Oxford, Oxford University
Press.

Lowenfeld, A. F. (2008). International Economic Law. Oxford, Oxford University
Press.

Mailath, G. J. and L. Samuelson (2006). Repeated Games and Reputations. Oxford,
Oxford University Press.

Mann, H. and K. v. Moltke (1999). NAFTA's Chapter 11 and the Environment -
Addressing the Impacts of the Investor-State Process on the Environment. Winnipeg,
International   Institute   for    Sustainable    Development,    available under
http://www.iisd.org/pdf/nafta.pdf (November 2nd, 2010).

Marboe, I. (2006). "Compensation and Damages in International Law - The Limits of
“Fair Market Value”." The Journal of World Investment and Trade 7(6).

Markusen, J. R., E. R. Morey, et al. (1995). "Competition in Regional Environmental
Policies When Plant Locations are Endogenous." Journal of Public Economics 56: 55-
77.

Marrela, F. and I. Marboe (2007). "'Efficient Breach' and Economic Analysis of
International Investment Law." Transnational Dispute Management.

Martin, L. L. (2005). "The President and International Commitments: Treaties as
Signaling Devices." Presidential Studies Quarterly 35(3): 440-465.

Matsuyama, K. (1990). "Perfect Equilibria in a Trade Liberalization Game." American
Economic Review 80(3): 480-492.

Maurer, H. (1999). Allgemeines Verwaltungsrecht. München, Verlag C.H. Beck.

McLure, C. E. (1986). "Tax Competition: Is What's Good for the Private Goose Also
Good for the Public Gander?" National Tax Journal 39: 341-348.

McTaggart, D. and D. Salant (1989). "Time Consistency and Subgame Perfect
Equilibria in a Monetary Policy Game." Journal of Macroeconomics 11(4): 575-588.

Meyer, K. E. (2004). "Perspectives on Multinational Enterprises in Emerging
Economies." Journal of International Business Studies 35(4): 259-276.
232                                                                       Bibliography

Miceli, T. J. and K. Segerson (1999). Takings. Encyclopedia of Law and Economics.
B. Bouckaert and G. De Geest, Edward Elgar and the University of Ghent, available
under http://encyclo.findlaw.com/ (November 2nd, 2010).

Miles, K. (2008). "International Investment Law and Climate Change: Issues in the
Transition to a Low Carbon World." Working Paper 27/08, available under
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1154588 (November 2nd, 2010).

Miller, G. P. (1987). "Agency Problems in Settlement." The Journal of Legal Studies
16(1): 189-215.

Mistelis, L. A. (2005). Confidentiality and Third Party Participation - UPS vs. Canada
and Methanex Corp v. United States. International Investment Law and Arbitration:
Leading Cases from the ICSID, NAFTA, Bilateral Treaties, and Customary
International Law. T. Weiler. London, Cameron May: 169-199.

Mitchell, R. B. (1998). "Sources of Transparency: Information Systems in
International Regimes." International Studies Quarterly 42: 109-130.

Mock, W. B. T. (2000). "An Interdisciplinary Introduction to Legal Transparency: A
Tool for Rational Development." Dickinson Journal of International Law 18: 293-304.

Monaldi, F. (2001). "Sunk-Costs, Institutions, and Commitment: Foreign Investment
in the Venezuelan Oil Industry." Manuscript, Department of Political Science,
Stanford University (on file with the author).

Monaldi, F. (2002). "Government Commitment Using External Hostages: Attracting
Foreign Investment to the Venezuelan Oil Industry." available under
http://www.allacademic.com/meta/p65158_index.html (November 2nd, 2010).

Moran, T. H. (2006). Harnessing Foreign Direct Investment for Development.
Washington, Center for Global Development.

Neumayer, E. (2001). Greening Trade and Investment: Environmental Protection
Without Protectionism. London, Earthscan Publications Ltd.

Neumayer, E. (2006). "Self-Interest, Foreign Need, and Good Governance: Are
Bilateral Investment Treaty Programs Similar to Aid Allocation?" Foreign Policy
Analysis 2: 245-267.

Neumayer, E. and L. Spess (2005). "Do Bilateral Investment Treaties Increase Foreign
Direct Investment to Developing Countries?" World Development 33(10): 1567-1585.

Oates, W. E. (1972). Fiscal Federalism. New York, Harcourt Brace Jovanovich.
Bibliography                                                                         233

Oates, W. E. (2001). "A Reconsideration of Environmental Federalism." Discussion
Paper, available under http://www.rff.org/documents/RFF-DP-01-54.pdf (November
2nd, 2010).

Oates, W. E. and R. M. Schwab (1988). "Economic Competition Among Jurisdictions:
Efficiency Enhancing or Distortion Inducing?" Journal of Public Economics 35: 333-
354.

OECD (1996). Detailed Benchmark Definition of Foreign Direct Investment, available
under http://www.oecd.org/dataoecd/10/16/2090148.pdf (November 2nd, 2010).

OECD    (2001).    "Regulatory   Investment   Incentives."    available         under
www.oecd.org/dataoecd/32/17/2510459.pdf (November 2nd, 2010).

OECD (2002). "Foreign Direct Investment for Development - Maximising Benefits,
Minimising Costs." available under http://www.oecd.org/dataoecd/47/51/1959815.pdf
(November 2nd, 2010).

OECD (2004). ""Indirect Expropriation" and the "Right to Regulate" in International
Investment Law." Working Papers on International Investment 2004/4.

OECD (2005). "Transparency and Third Party Participation in Investor-State Dispute
Settlement Procedures." Statement by the OECD Investment Committee.

OECD (2006). "Interpretation of the Umbrella Clause in Investment Agreements."
Working Papers on International Investment(3).

Paulsson, J. (1995). "Arbitration Without Privity." ICSID Review: Foreign Investment
Law Journal 10(2): 232-257.

Paulsson, J. (2007). "Enclaves of Justice." Transnational Dispute Management 4(5).

Persson, T., G. Tabellini, et al. (2003). "Electoral Rules and Corruption." Journal of
the European Economic Association 1(4): 958-989.

Picherack, J. R. (2008). "The Expanding Scope of the Fair and Equitable Treatment
Standard: Have Recent Tribunals Gone Too Far?" The Journal of World Investment
and Trade 9(4).

Posner, E. A. (2006). "International Law: A Welfarist Approach." The University of
Chicago Law Review 73(2): 487-543.

Posner, R. (1986). Economic Analysis of Law. Boston, Little Brown.

Posner, R. (2003). Economic Analysis of Law. New York, Aspen Publishers.

Posner, R. (2007). Economic Analysis of Law. New York, Aspen Publishers.
234                                                                      Bibliography

Prieß, H. J. and G. M. Berrisch (2003). WTO-Handbuch. München, Verlag C.H. Beck.

Rasmusen, E. (2001). Games and Information. Oxford, Blackwell Publishing Ltd.

Rasmusen, E. B. (1998). Nuisance Suits. The New Palgrave Dictionary of Economics
and the Law. P. Newman. New York, Stockton Press. 2: 690-693.

Redfern, A., M. Hunter, et al. (2004). Law and Practice of International Commercial
Arbitration. London, Sweet & Maxwell.

Revesz, R. L. (1992). "Rehabilitating Interstate Competition: Rethinking the "Race-
To-The-Bottom" Rationale for Federal Environmental Regulation." New York
University Law Review 67: 1210-1254.

Rodrik, D., A. Subramanian, et al. (2004). "Institutions Rule: The Primacy of
Institutions Over Geography and Integration in Economic Development." Journal of
Economic Growth 9: 131-165.

Rodrik, D. and R. Wacziarg (2005). "Do Democratic Transitions Produce Bad
Economic Outcomes?" American Economic Review 95(2).

Romano, R. (2001). "The Need For Competition in International Securities
Regulation." Yale ICF Working Paper No. 00-49, available under
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=278728 (November 2nd, 2010).

Romano, R. (2005). "Is Regulatory Competition a Problem or Irrelevant for Corporate
Governance?" Oxford Review of Economic Policy 212(2): 212-231.

Romer, P. M. (1993). "Two Strategies of Economic Development: Using Ideas and
Producing Ideas." Proceedings of the World Bank Annual Conference on
Development Economics 1992: 63-91.

Rose-Ackerman, S. and J. Tobin (2003). "Foreign Direct Investment and the Business
Environment in Developing Countries: the Impact of Bilateral Investment Treaties."
William Davidson Institute Working Paper Number 587.

Saggi, K. (2002). "Trade, Foreign Direct Investment and International Technology
Transfer: A Survey." World Bank Research Observer 17(2): 191-235.

Salacuse, J. W. (1990). "BIT by BIT: The Growth of Bilateral Investment Treaties and
Their Impact on Foreign Investment in Developing Countries." The International
Lawyer 24(3): 655-675.

Salacuse, J. W. and N. P. Sullivan (2005). "Do BITs Really Work?: An Evaluation of
Bilateral Investment Treaties and Their Grand Bargain." Harvard International Law
Journal 46: 67-130.
Bibliography                                                                     235

Salmon, P. (2003). Assigning Powers in the European Union in the Light of Yardstick
Competition among Governments. Jahrbuch für neue politische Ökonomie. M. J.
Holler, H. Kliemt, D. Schmidtchen and M. Streit. Tübingen, Mohr-Siebeck. 22: 197-
216.

Sauvant, K. (2008). The Rise of International Investment, Investment Agreements and
Investment Disputes. Appeals Mechanism in International Investment Disputes. K.
Sauvant and M. Chiswick-Patterson. New York, Oxford University Press, available
under          http://www.cpii.columbia.edu/pubs/documents/KarlP.Sauvant-Appeals
Mechanism.pdf (October 11, 2009).

Sauvant, K. P. and L. E. Sachs, Eds. (2009). The Effect of Treaties on Foreign Direct
Investment: Bilateral Investment Treaties, Double Taxation Treaties, and Investment
Flows. Oxford, Oxford University Press.

Sauvé, P. and A. B. Zampetti (2007). International Investment. Research Handbook in
International Economic Law. A. Guzman and A. O. Sykes. Cheltenham, Edward
Elgar: 211-270.

Schäfer, H.-B. and C. Ott (2005). Lehrbuch der ökonomischen Analyse des
Zivilrechts. Berlin, Springer.

Schmidtchen, D. (1994). Ökonomik des Vertrauens. Recht und Verhalten. H. Hof, H.
Kummer and P. Weingart. Baden-Baden, Nomos Verlagsgesellschaft: 129-163.

Schreuer, C. H. (2005). "Fair and Equitable Treatment in Arbitral Practice." The
Journal of World Investment and Trade 6(3): 357-386.

Serra, D. (2006). "Empirical Determinants of Corruption: A Sensitivity Analysis."
Public Choice 126: 225-256.

Shavell, S. (2004). Foundations of Economic Analysis of Law. Cambridge, The
Belknap Press of Harvard University Press.

Siebert, H. and M. J. Koop (1990). "Institutional Competition. A Concept for Europe?"
Kiel Working Paper No. 440.

Simmons, B. (2000). "Money and the Law: Why Comply with the Public International
Law of Money?" The Yale Journal of International Law 25: 323-362.

Sinn, H.-W. (2003). The New Systems Competition. Oxford, Blackwell.

Snidal, D. (2002). Rational Choice and International Relations. Handbook of
International Relations. W. Carlsnaes, T. Risse and B. Simmons. London, Sage: 73-94.

Sornarajah, M. (2004). The International Law on Foreign Investment. Cambridge,
Cambridge University Press.
236                                                                      Bibliography

Spence, M. (1973). Market Signaling: Information Transfer in Hiring and Related
Processes. Cambridge, Harvard University Press.

Spiller, P. T. and M. Tommasi (2008). The Institutions of Regulation. An Application
to Public Utilities. Handbook of New Institutional Economics. C. Ménard and M. M.
Shirley. Berlin, Springer: 515-544.

Stewart, R. B. (1977). "Pyramids of Sacrifice? Problems of Federalism in Mandating
State Implementation of National Environmental Policy." Yale Law Journal 86: 1196-
1272.

Swenson, D. L. (2004). "Foreign Investment and the Mediation of Trade Flows."
Review of International Economics 12(4): 609-629.

Swenson, D. L. (2005). "Why Do Developing Countries Sign BITs?" U.C. Davis
Journal of International Law and Policy 12(1): 131-155.

Sykes, A. O. (1991). "Protectionism as a "Safeguard": A Positive Analysis of the
GATT "Escape Clause" with Normative Speculations." University of Chicago Law
Review 58: 255-303.

Sykes, A. O. (2007). International Law. Handbook of Law and Economics, Volume 1.
A. M. Polinsky and S. Shavell. Amsterdam, Elsevier: 759-826.

Teitelbaum, R. (2005). "Who's Afraid of Maffezini? Recent Developments in the
Interpretation of Most Favored Nation Clauses." Journal of International Arbitration
22(3): 225-238.

Thomas, J. and T. Worrall (1994). "Foreign Direct Investment and the Risk of
Expropriation." Review of Economic Studies 61: 81-108.

Tiebout, C. M. (1956). "A Pure Theory of Local Expenditures." The Journal of
Political Economy 64(5): 416-424.

Tienhaara, K. (2008a). Unilateral Commitment to Investment Protection: Does the
Promise of Stability Restrict Environmental Policy Development? Yearbook of
International Environmental Law 2007. O. K. Fauchald and D. Hunter. Oxford,
Oxford University Press: 139-167.

Tobin, J. and S. Rose-Ackerman (2005). "Foreign Direct Investment and the Business
Environment in Developing Countries: The Impact of Bilateral Investment Treaties."
Yale Law School, Center for Law, Economics and Public Policy, Research Paper No.
293.

Tobin, J. and S. Rose-Ackerman (2006). "When BITs Have Some Bite: The Political-
Economic Environment for Bilateral Investment Treaties." Working Paper.
Bibliography                                                                   237

Treisman, D. (2000). "The Causes of Corruption: A Cross-national Study." Journal of
Public Economics 76: 399-457.

UNCTAD (1997). World Investment Report 1997. New York and Geneva, United
Nations.

UNCTAD (1998). Bilateral Investment Treaties in the Mid-1990s. New York and
Geneva, United Nations.

UNCTAD (1999). Fair and Equitable Treatment. Geneva, United Nations
Publications.

UNCTAD (2003). World Investment Report 2003. New York and Geneva, United
Nations.

UNCTAD (2004). Home Country Measures. New York and Geneva, United Nations.

UNCTAD (2006). World Investment Report 2006. New York and Geneva, United
Nations.

UNCTAD (2007a). World Investment Report 2007: Transnational Corporations,
Extractive Industries and Development. New York and Geneva, United Nations.

UNCTAD (2007b). Investor-State Dispute Settlement and Impact on Investment
Rulemaking. New York and Geneva, United Nations.

UNCTAD (2007c). "Recent Developments in International Investment Agreements
(2006 - June 2007)." IIA Monitor(3): 1-10.

UNCTAD (2008a). World Investment Report 2008. New York and Geneva, United
Nations.

UNCTAD (2008c). "Recent developments in International Investment Agreements
(2007 - June 2008)." IAA Monitor, available under http://www.unctad.org/en/docs/
webdiaeia20081_en.pdf (November 2nd, 2010).

UNCTAD (2009a). World Investment Report 2009. New York and Geneva, United
Nations.

UNCTAD (2009b). "Latest Developments in Investor-State Dispute Settlement." IAA
Monitor, available under http://www.unctad.org/en/docs/webdiaeia20096_en.pdf
(November 2nd, 2010).

UNCTAD (2010). World Investment Report 2010: Investing in a Low-Carbon
Economy. New York and Geneva, United Nations.

Van Aaken, A. (2008a). "Perils of Success? The Case of International Investment
Protection." European Business Organization Law Review 9(1): 1-27.
238                                                                       Bibliography

Van Aaken, A. (2008b). How to do Constitutional Law and Economics: A
Methodological Proposal. Internationalization of the Law and its Economic Analysis,
Festschrift for Hans-Bernd Schäfer zum 65. Geburtstag. T. Eger, J. Bigus, C. Ott and
G. v. Wangenheim. Wiesbaden, Gabler.

Van Aaken, A. (2009). "International Investment Law Between Commitment and
Flexibility: A Contract Theory Analysis." Journal of International Economic Law.

Van Harten, G. (2007). Investment Treaty Arbitration and Public Law. Oxford, Oxford
University Press.

Verbeek, M. (2000). A Guide to Modern Econometrics. Chichester, Wiley.

Veugelers, R. (1993). "Reputation as a Mechanism Alleviating Opportunistic Host
Government Behavior Against MNEs." Journal of Industrial Economics 41(1): 1-17.

Voigt, S. (2009). "How (Not) to Measure Institutions." Working Paper, available under
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1336272 (November 2nd, 2010).

Wälde, T. W. and A. Kolo (2001). "Environmental Regulation, Investment Protection,
and 'Regulatory Taking' in International Law." International and Comparative Law
Quarterly 50(4): 811-848.

Walter, A. and G. Sen (2009). Analyzing the Global Political Economy. Princeton,
Princeton University Press.

Wellisch, D. (1995). "Locational Choices of Firms and Decentralized Environmental
Policy with Various Instruments." Journal of Urban Economics 37: 290-310.

Wellisch, D. (2000). Theory of Public Finance in a Federal State. Cambridge,
Cambridge University Press.

Wells, L. T. (2003). "Double Dipping in Arbitration Awards? An Economist Questions
Damages Awarded to Karaha Bodas Company in Indonesia." Arbitration International
19(4): 471-481.

Westcott, T. J. (2007). "Recent Practice on Fair and Equitable Treatment." The Journal
of World Investment and Trade 8(3): 409-430.

Weston, J. F. and B. W. Sorge (1972). International Managerial Finance. Homewood,
Irwin.

Williamson, O. E. (1983). "Credible Commitments: Using Hostages to Support
Exchange." American Economic Review 73: 519-540.

Williamson, O. E. (1985). The Economic Institutions of Capitalism. New York, Free
Press.
Bibliography                                                                     239

Williamson, O. E. (1990). "A Comparison of Alternative Approaches to Economic
Organization." Journal of Institutional and Theoretical Economics 146(1): 61-71.

Wilson, J. D. (1996). Capital Mobility and Environmental Standards: Is There a
Theoretical Basis for a Race to the Bottom? Fair Trade and Harmonization:
Prerequisites for Free Trade? J. Bhagwati and R. E. Hudec. Cambridge, MIT Press.

Wilson, J. D. (1999). "Theories of Tax Competition." National Tax Journal 52(2): 269-
304.

World Bank, T. (2001). Global Economic Prospects and the Developing Countries.
Washington, The World Bank.

World Bank, T. (2007). World Development Indicators. Washington D.C.,
International Bank for Reconstruction and Development.

WTO      (2007).     "World      Trade     Report     2007."    available      under
http://www.wto.org/english/res_e/reser_e/wtr_e.htm (November 2nd, 2010).

WTO (2008). "World Trade Report 2008 - Trade in a Globalizing World." available
under http://www.wto.org/english/res_e/reser_e/wtr_e.htm (November 2nd, 2010).

Zodrow, G. R. and P. Mieszkowski (1986). "Pigou, Tiebout, Property Taxation, and
the Underprovision of Local Public Goods." Journal of Urban Economics 19: 356-370.

								
To top