Essays on the Economic Consequences of Mandatory IFRS Reporting Around the World

Ulf Brüggemann Essays on the economic consequences of mandatory IFRS reporting around the world

GABLER RESEARCH Quantitatives Controlling Herausgegeben von Professor Dr. Carsten Homburg, Universität zu Köln

Die Schriftenreihe dient als Forum für hervorragende Forschungsergebnisse auf dem Gebiet des Controlling. Ihr liegt ein weites Controllingverständnis zugrunde, das über Problemstellungen der traditionellen internen Unternehmensrechnung hinaus geht und beispielsweise auch Aspekte der Verhaltenssteuerung einschließt. Der Schwerpunkt der Reihe liegt auf quantitativen Analysen aktueller Controllingfragen. Hierbei werden formal-analytische ebenso wie empirisch ausgerichtete Arbeiten in Betracht gezogen.

Ulf Brüggemann

Essays on the economic consequences of mandatory IFRS reporting around the world With a preface by Prof. Dr. Carsten Homburg

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 zu Köln, 2011

1st Edition 2011 All rights reserved © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011 Editorial Office: Marta Grabowski | Stefanie Loyal Gabler Verlag ist eine Marke von Springer Fachmedien. Springer Fachmedien ist Teil der Fachverlagsgruppe 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, photocopying, 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 publication 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. Umschlaggestaltung: KünkelLopka Medienentwicklung, Heidelberg Printed on acid-free paper Printed in Germany ISBN 978-3-8349-3169-6

Geleitwort Ein Großteil der börsennotierten Unternehmen in der Europäischen Union ist gemäß EU-Regulierung 1606/2002 verpflichtet, seit dem Geschäftsjahr 2005 konsolidierte Abschlüsse nach International Financial Reporting Standards aufzustellen. Ziel dieser Umstellung ist es, durch gemeinsame Rechnungslegungsstandards ökonomische Verbesserungen auf dem Kapitalmarkt zu erreichen. Dies soll insbesondere durch die erhöhte Transparenz der Fall sein. Eine Vielzahl an Literatur beschäftigt sich mit der Frage, ob die durch die IFRS-Umstellung erwünschten ökonomischen Konsequenzen tatsächlich eingetreten sind. In diesen Literaturstrang ist die vorliegende Dissertationsschrift einzuordnen. Neben einer Einleitung und einem Schlussteil besteht sie aus drei Arbeitspapieren. Der erste Hauptteil bespricht die empirische Literatur zu den ökonomischen Konsequenzen der verbindlichen Umstellung auf IFRS. Ziel ist es, die bestehende Literatur kritisch zu beleuchten und Anregungen für zukünftige Forschungsarbeiten zu liefern. Die Analyse zeigt eine erstaunliche Diskrepanz in den bisherigen Forschungsergebnissen auf. Einerseits gibt es kaum empirische Evidenz für Veränderungen in der Transparenz der länderübergreifenden Vergleichbarkeit von Geschäftsberichten. Anderseits findet die überwältigende Mehrheit aller Studien Hinweise auf positive Effekte in Kapitalmärkten sowie auf makroökonomischer Ebene. Der Verfasser argumentiert, dass das sog. Identifikationsproblem zu dieser Diskrepanz beiträgt, d.h. die Schwierigkeit, einen möglichen IFRS-Effekt von gleichzeitigen Veränderungen zu trennen, die nicht mit der Rechnungslegung in Zusammenhang stehen. Der zweite Hauptteil widmet sich den Konsequenzen der IFRS-Einführung im Rahmen einer empirischen Analyse. Im Gegensatz zur bestehenden Literatur, die sich auf Portfolioentscheidungen institutioneller Investoren oder aggregierte Kapitalmarktreaktionen konzentriert, wird hier das Anlageverhalten von Privatanlegern untersucht. Konkret wird analysiert, ob Privatanleger in Reaktion auf die IFRS-Einführung verstärkt in ausländische Aktien investieren. Der Open Market der Frankfurter Börse bietet hier eine ideale Plattform, um diese Fragestellung zu behandeln. Es wird zunächst der institutionelle Rahmen des Open Markets beschrieben. Anschließend wird der Einfluss der verbindlichen IFRS-Umstellung auf Handelsvolumina im Open Market untersucht. Der Verfasser argumentiert, dass im Open Market besonders kompetente und erfahrene Anleger handeln, für die RechnungslegungsV

informationen eine wichtige Rolle spielen. Auf dieser Grundlage wird die Hypothese aufgestellt, dass die Einführung von IFRS grenzüberschreitende Aktieninvestments der Anleger im Open Market anregt, weil Rechnungslegungsinformationen unter international einheitlichen Standards besser zu vergleichen sind. Die Ergebnisse liefern überzeugende empirische Evidenz für die Hypothese, dass die verbindliche Umstellung auf IFRS grenzüberschreitende Aktieninvestments durch Privatanleger anregt. Dieses Ergebnis ist neu in der Literatur und dürfte insbesondere für Regulierer von Interesse sein, die sich dem Schutz von Privatanlegern verschrieben haben und eine zukünftige Einführung von IFRS in Erwägung ziehen. Im Oktober 2008 hat das International Accounting Standards Board unter großem politischen Druck Änderungen an IAS 39 beschlossen. Diese Änderungen räumen Banken die Option ein, finanzielle Vermögenswerte, die zuvor erfolgswirksam zum Zeitwert bilanziert wurden, unter bestimmten Bedingungen in Kategorien zu reklassifizieren, die eine Bilanzierung zu fortgeführten Anschaffungskosten erfordern. Der dritte Hauptteil analysiert empirisch, welche Banken aus welchen Gründen die Reklassifizierungsoption in Anspruch nehmen und wie der Aktienmarkt hierauf reagiert. Zunächst werden ausführlich die neuen Reklassifizierungsregeln sowie die Ereignisse auf politischer Ebene, die zu der Regeländerung führten, beschrieben. Dann wird die relevante Literatur diskutiert. Es wird die Hypothese aufgestellt, dass die Reklassifizierungen kurzfristigen Nutzen stiften, da sie finanziell bedrohten Banken die Möglichkeit bieten, durch eine reine Bilanzierungsmaßnahme das regulatorische Kapital zu erhöhen. Der Verfasser vermutet allerdings auch, dass dem Aktienmarkt durch die Reklassifizierungen Informationen entzogen werden, was sich langfristig in höheren Geld-Brief-Spannen niederschlägt. Es folgen die empirischen Analysen. Die Untersuchung basiert auf einer umfassenden globalen Stichprobe von 302 börsennotierten Banken und erfolgt in drei Schritten. Im ersten Schritt wird die deskriptive Evidenz zu Umfang und Art der Reklassifizierungsentscheidung präsentiert. Im zweiten Schritt wird gezeigt, dass das Risiko einer Verletzung der regulatorischen Kapitalvorschriften eine wesentliche Determinante der Reklassifizierungsentscheidung darstellt. Im dritten Schritt der empirischen Analyse werden kurz- und langfristige Aktienmarktreaktionen auf die Reklassifizierungsentscheidungen betrachtet. Insgesamt wird empirische Evidenz dafür geliefert, dass die Änderungen an IAS 39 zwar kurzfristig Nutzen stiften, langfristig allerdings auch Kosten verursachen. Auf der einen Seite erleichtert die Reklassifizierungsoption finanziell bedrohten Banken, die regulatorischen Kapitalanforderungen VI

einzuhalten. Auf der anderen Seite hält sich die Mehrheit der reklassifizierenden Banken nicht an Offenlegungspflichten, die mit den Reklassifizierungen einhergehen. Es wird durch die Analyse nicht nur ein Beitrag zur IFRS-Literatur geleistet, sondern es werden auch interessante Einsichten für das IASB als internationalen Standardsetter geliefert. Bei der vorliegenden Dissertationsschrift handelt es sich um einen hervorragenden Beitrag, der zu einem aktuellen Thema zahlreiche neue Einsichten liefert. Auch methodisch befindet sich die Arbeit auf äußerst hohem Niveau.

Prof. Dr. Carsten Homburg

VII

Vorwort Die vorliegende Arbeit wurde im Sommersemester 2011 von der Wirtschafts- und Sozialwissenschaftlichen Fakultät der Universität zu Köln als Dissertation angenommen. In den folgenden Zeilen möchte ich den Personen danken, die zum Gelingen dieser Arbeit beigetragen haben. Ein besonderer Dank gilt meinem Doktorvater, Herrn Prof. Dr. Carsten Homburg. Er KDWPLFKDQGDV'LVVHUWDWLRQVWKHPDKHUDQJHIKUWXQGPLUJURȕH)UHLKHLWHQEHLGHU$XVJHstaltung meiner Doktorarbeit eingeräumt. Ich danke ihm für seine freundliche und geduldige Unterstützung. Weiterhin danke ich Herrn Prof. Dr. Christoph Kuhner für die Übernahme des Zweitgutachtens sowie Herrn Prof. Dr. Heinrich Schradin für den Vorsitz bei der Disputation. Die Grundlagen meiner Dissertation wurden im Rahmen des Graduiertenkollegs Risikomanagement an der Universität zu Köln gelegt. In diesem Zusammenhang danke ich der Deutschen Forschungsgemeinschaft für die finanzielle Unterstützung. Bei Herrn Prof. Dr. Alexander Kempf bedanke ich mich für die Leitung und Organisation des Graduiertenkollegs. Ein ganz herzliches Dankeschön geht an meine Mit-Kollegiaten der zweiten Kohorte für die wunderbare Zeit und hervorragende Verpflegung: Sanela CeljoHörhager, Achim Dahlbokum, Manuela Ender, Martin Honal, Peter Kosater, Thomas Moosbrucker und Tanja Thiele. Den Kern meiner Dissertation habe ich als Forschungsstipendiat des INTACCT Research Training Networks an der Lancaster University verfasst. Ich danke der Europäischen Kommission für die Finanzierung von INTACCT. Den Kollegen des Department of Accounting and Finance in Lancaster danke ich für die anregende und lXȕHUVW QHWWH $UEHLWVDWPRVSKlUH ,17$&&7 KDW HV PLU HUP|JOLFKW DQ VHKU QW]OLFKHQ Doktorandenkursen teilzunehmen und meine Ko-Autoren Jannis Bischof, Holger Daske, Jörg-Markus Hitz, Peter Pope und Thorsten Sellhorn kennenzulernen. Bei ihnen bedanke ich mich für die angenehme und lehrreiche Zusammenarbeit. Mein ausdrücklicher Dank gilt Peter, der mich in das INTACCT-Netzwerk aufgenommen hat, und Holger, der mir gezeigt hat, dass man Forschungsprojekte niFKW QXU EHJLQQHQ VRQGHUQ DXFK DEVFKOLHȕHQ kann. Ein besonderer Dank gilt auch den Kollegen des Seminars für Allgemeine Betriebswirtschaftslehre und Controlling an Universität zu Köln, die mich während meiner Promotionszeit begleitet haben: Daniel Baumgarten, Max Berens, Marcus Berghäuser, IX

Ute Bonenkamp, Cordula Ebeling, Sebastian Gell, Dominika Gödde, Stefan Henschke, Tanja Klettke, Michael Lorenz, Christian Müller, Julia Nasev, Philipp Plank, Kristina Reimer, Peter Scherpereel, Peter Stebel, Jörg Stephan, 0DWWKLDV :HLȕ XQG 1LNRODXV Wrede. Ich habe mich am Lehrstuhl immer wohl und willkommen gefühlt, was nicht zuletzt an Elisabeth Tokarski-Eich liegt, die mit ihrer Herzlichkeit den Betrieb zusammenhält und mich nebenbei immer wieder rettete, wenn ich eine Frist verpasst hatte. Vielen lieben Dank. 6FKOLHȕOLFKEHGDQNHLFKPLFKEHLPHLQHQ(OWHUQXQGPHLQHP%UXGHUGLHPLULQJXWHQ und in schwierigen Phasen meiner Promotionszeit ein verlässlicher Rückhalt waren. Danke für alles.

Ulf Brüggemann

X

Table of Contents 1 Introduction ................................................................................................................... .1 2 Intended and unintended consequences of mandatory IFRS adoption .................... 5 2.1 2.2

Introduction ............................................................................................................ 5 Conceptual underpinnings ...................................................................................... 7

2.2.1 2.2.2 2.3

Economic consequences .................................................................................. 7 Intended versus unintended consequences of mandatory IFRS adoption ....... 9

Intended consequences of mandatory IFRS adoption .......................................... 11

2.3.1 2.3.2

Theoretical background ................................................................................. 11 Empirical evidence ........................................................................................ 13

2.3.2.1 2.3.2.2 2.3.2.3 2.3.3 2.4

Individual contracts ....................................................................................... 23 Collective contracts ....................................................................................... 25

Economic consequences in anticipation of mandatory IFRS adoption ................ 31

2.5.1 2.5.2 2.5.3 2.5.4 2.6 2.7

Discussion ...................................................................................................... 20

Unintended consequences of mandatory IFRS adoption ..................................... 22

2.4.1 2.4.2 2.5

Accounting effects .................................................................................. 13 Capital market effects ............................................................................. 16 Macroeconomic effects ........................................................................... 19

Ex-ante stock market reactions ...................................................................... 31 Avoidance strategies ...................................................................................... 32 Lobbying activities ........................................................................................ 34 Timing transactions ....................................................................................... 37

Summary and suggestions for future research ..................................................... 37 Tables ................................................................................................................... 43

2.7.1

Sample size comparison ................................................................................ 43

3 The impact of mandatory IFRS adoption on cross-border equity investments of individual investors ............................................................................. 45 3.1 3.2

Introduction .......................................................................................................... 45 The Open Market.................................................................................................. 48

3.2.1 3.2.2 3.2.3

Institutional background ................................................................................ 48 Data and descriptive statistics ....................................................................... 51 Determinants of Open Market inclusion ....................................................... 54 XI

3.3

Mandatory IFRS adoption and trading activities in the Open Market ................. 55

3.3.1 3.3.2

Predictions ..................................................................................................... 55 Open Market inclusion analysis .................................................................... 59

3.3.2.1 3.3.2.2 3.3.3

Open Market trading volume analysis ........................................................... 61

3.3.3.1 3.3.3.2 3.3.4

Frankfurt floor versus XETRA ............................................................... 66 Other German exchanges ........................................................................ 66

Conclusions .......................................................................................................... 67 Tables ................................................................................................................... 69

3.5.1 3.5.2 3.5.3 3.5.4 3.5.5 3.5.6 3.5.7 3.6

Research design ...................................................................................... 64 Empirical findings................................................................................... 65

Auxiliary analyses ......................................................................................... 66

3.3.5.1 3.3.5.2 3.4 3.5

Research design ...................................................................................... 61 Empirical findings................................................................................... 62

Cross-sectional analysis of the Open Market reaction .................................. 64

3.3.4.1 3.3.4.2 3.3.5

Research design ...................................................................................... 59 Empirical findings................................................................................... 60

Sample composition by country and year ..................................................... 69 Descriptive statistics ...................................................................................... 71 Institutional characteristics of IFRS and non-IFRS adoption countries ........ 72 Determinants of Open Market inclusion ....................................................... 74 Mandatory IFRS adoption and Open Market inclusion ................................ 75 Mandatory IFRS adoption and Open Market trading activity ....................... 76 Cross-sectional analysis of the Open Market reaction .................................. 78

Appendices ........................................................................................................... 79

3.6.1 3.6.2

Comparison of transaction costs: Open Market versus home markets ......... 79 Google search results and trading volume..................................................... 81

4 The economic consequences of fair value reclassifications under IFRS ................. 83 4.1 4.2

Introduction .......................................................................................................... 83 Institutional background ....................................................................................... 88

4.2.1 4.2.2

Events preceding the reclassification amendments ....................................... 88 The reclassification amendments .................................................................. 91

4.2.2.1 4.2.2.2 4.2.2.3 XII

Amendments to IAS 39: recognition and measurement ......................... 91 Amendments to IFRS 7: disclosures....................................................... 92 Consequences for the regulatory capital of banks .................................. 93

4.2.3 4.3

Hypotheses development and related literature ................................................... 96

4.3.1 4.3.2 4.4

Initial reactions to the amendment decision .................................................. 94 Short-term benefits of the reclassification choice ......................................... 96 Long-term costs of the reclassification choice ............................................ 100

Empirical analysis .............................................................................................. 101

4.4.1 4.4.2

Sample selection and data sources .............................................................. 101 Descriptive evidence.................................................................................... 103

4.4.2.1 4.4.2.2 4.4.3

Economic and political drivers of the reclassification choice ..................... 107

4.4.3.1 4.4.3.2 4.4.4

Research design .................................................................................... 112 Empirical findings................................................................................. 116

Long-term effects of reclassifications on information asymmetry ............. 118

4.4.5.1 4.4.5.2 4.5 4.6

Research design .................................................................................... 107 Empirical findings................................................................................. 110

Stock price reactions to reclassification announcements ............................ 112

4.4.4.1 4.4.4.2 4.4.5

Accounting effects of the reclassification amendments ....................... 103 Footnote disclosures of reclassifications .............................................. 106

Research design .................................................................................... 118 Empirical findings................................................................................. 119

Conclusions and implications............................................................................. 120 Tables ................................................................................................................. 122

4.6.1 4.6.2 4.6.3 4.6.4 4.6.5 4.6.6 4.6.7 4.6.8

Reclassification data and selected variables ................................................ 122 Reclassification effects ................................................................................ 123 Descriptive statistics .................................................................................... 124 Reclassification disclosures ......................................................................... 125 Determinants of reclassification choice....................................................... 126 Stock market reactions to regulatory events ................................................ 127 Stock market reactions to reclassification announcements ......................... 129 Long-term effects of reclassifications on information asymmetry ............. 130

5 Summary and Conclusions ........................................................................................ 133 References ....................................................................................................................... 135

XIII

1 Introduction Regulation EU No. 1606/2002 requires most listed companies in the European Union (EU) to prepare their consolidated accounts according to International Financial Reporting Standards (IFRS) from fiscal year 2005 onwards (EC, 2002; henceforth the IAS Regulation).1 This regulation is part of an unprecedented accounting experiment that has seen IFRS reporting being currently mandated in almost 100 countries around the world.2 Even the US Securities and Exchange Commission (SEC) is actively considering to incorporate IFRS into the financial reporting system for US issuers.3 Regulators justify the move towards IFRS by the expectation that collective adoption of a single set of global accounting standards will trigger desirable economic consequences in capital markets and at the macroeconomic level. For example, the IAS Regulation explicitly aims ³WRHQVXUHDKLJKGHJUHHRIWUDQVSDUHQF\DQGFRPSDUDELOLW\RIILQDQFLDOVWDWHPHQWV and hence an efficient functioning of the Community capital market and the Internal 0DUNHW´ (&  $UW ). There is an emerging stream of empirical literature that evaluates whether these objectives have been met. This thesis contributes to this stream of literature by providing three essays on the economic consequences of mandatory IFRS reporting around the world. While the first two essays focus on the response to the adoption process, the third essay analyzes the effects of a subsequent change in IFRS accounting rules. The first essay is based on a working paper with Jörg-Markus Hitz (University of Göttingen) and Thorsten Sellhorn (WHU ± Otto Beisheim School of Management) and reviews empirical literature on the economic consequences of the mandatory IFRS adoption in the EU. We use the term economic consequences to denote any effects of financial reporting on firm values and on the wealth of those who make decisions based on accounting information or are affected by such decisions. We define economic consequences to be intended (unintended) if they can (cannot) be reconciled with the explicitly stated objectives of the IAS Regulation. Our literature review reveals that empirical research on the intended consequences of mandatory IFRS adoption generally fails to find an increase in the comparability or transparency of financial statements. In contrast, there is rich and almost unanimous evidence of positive reactions in capital 1

International Accounting Standards (IAS) were renamed to IFRS in 2001. In this thesis, the terms IAS and IFRS are used interchangeably.

2

For details, see e.g. http://www.iasplus.com/country/useias.htm.

3

For details, see e.g. http://www.sec.gov/rules/other/2010/33-9109.pdf. 1

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1_1, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

markets and at the macroeconomic level. We argue that research design challenges (e.g., the difficulty of separating a potential IFRS effect from concurrent changes unrelated to financial reporting) are likely to contribute to this apparent mismatch in research findings. Literature on the unintended consequences is still in its infancy. However, extant empirical evidence and insights from non-IFRS settings suggest that mandatory IFRS adoption has a material impact on contractual outcomes and stimulates opportunistic anticipatory actions. In summary, our analysis of the extant empirical evidence suggests that both the intended and the unintended consequences merit further scrutiny to establish a balanced view on the overall impact of mandatory IFRS adoption. We therefore conclude the essay by, first, providing suggestions on how to address research design challenges in order to identify intended consequences more precisely and, second, proposing several avenues for future research on the unintended consequences of mandatory IFRS adoption. The second essay is based on a working paper with Holger Daske (University of Mannheim), Carsten Homburg (University of Cologne) and Peter F. Pope (Lancaster University) and focuses on the intended consequences of mandatory IFRS adoption. We predict that mandatory IFRS adoption enhances cross-border equity investments of individual investors. The underlying assumption of this prediction is that collective IFRS adoption removes entry barriers to investments in foreign stocks by replacing unfamiliar local accounting standards with familiar IFRS. We test our prediction by examining trading activity in the Open Market at Frankfurt Stock Exchange. The Open Market is a trading segment designed for German individual investors to trade a large selection of foreign (i.e., non-German) stocks. This large selection allows us to compare IFRS adopters with non-adopting firms within a difference-in-differences approach. Using a sample of 4,869 firms from 31 countries around the world, we find that stocks experience a substantial and statistically significant increase in Open Market trading activity following mandatory adoption of IFRS. For example, percentage trading volume in stocks of mandatory IFRS adopters increases by more than 20% relative to non-US control firms. If we include US firms in the benchmark, the effect on Open Market trading activity is even stronger. These findings are robust to the inclusion of a battery of control variables (e.g., media coverage in Germany based on Google News archive search results) and to several sensitivity analyses (e.g., a two-step approach to address potential sample selection issues). Taken together, our results are consistent with the idea that collective IFRS adoption reinforces foreign equity investments of individual investors. 2

The third and final essay is based on a working paper with Jannis Bischof (University of Mannheim) and Holger Daske (University of Mannheim). At the peak of the financial crisis in October 2008, the International Accounting Standards Board (IASB) succumbed to severe political pressure by EU leaders to issue an emergency amendment to IAS 39. This amendment grants companies the option to reclassify selected financial assets from fair value into categories that require measurement at amortized costs. In the first part of the essay, we provide a detailed description of the events that led to the amendment to IAS 39. In the next step, we examine the economic consequences that this change in fair value rules entails. Using a comprehensive global sample of 302 publicly listed IFRS banks, we show that 124 banks take the reclassification option in the first annual report following the amendment, thus increasing profits by a total of 22.7 billion Euros and firm-specific profits by 44% on average. Accounting choice tests provide evidence that banks are more likely to reclassify if they are close to violating regulatory capital restrictions. The most troubled of these banks experience positive abnormal stock returns around the announcement of the reclassification decision. Analyses of accompanying footnote disclosures reveal that two thirds of the reclassifying banks do not fully comply with the requirements of IFRS 7. This lack of compliance is associated with a significant long-term increase in information asymmetry among investors, especially if the reclassifications have a strong impact on net income. Our findings suggest that the emergency amendment to IAS 39 provides economically weak banks with effective means of regulatory arbitrage. However, a high degree of non-compliance with disclosure requirements also leads to an unintended decrease in transparency of financial statements. Each essay in this thesis makes a distinct contribution to academic literature. The first essay adds to the emerging stream of research on mandatory IFRS adoption by providing a comprehensive review of extant evidence that encompasses both intended and unintended consequences. Moreover, we point out the apparent disagreement in empirical findings on the intended consequences. In contrast, prior literature reviews on mandatory IFRS adoption are confined to early and less ambiguous evidence on the intended consequences (e.g., Hail and Leuz, 2007; Hail, Leuz, and Wysocki, 2010). The second essay contributes to the empirical literature on the intended consequences of mandatory IFRS adoption. Prior research focuses on aggregate capital market reactions (e.g., Daske, Hail, Leuz, and Verdi, 2008) or asset allocation decisions of institutional investors (e.g., Florou and Pope, 2009; Yu, 2009). We add to this literature by providing the first and, thus far, only empirical analysis of LQGLYLGXDO LQYHVWRUV¶ UHDFWLRQ WR mandatory IFRS adoption. In addition, we introduce the Open Market as a useful setting to study trading 3

behaviour of a large and homogeneous group of individual investors, i.e. German individual investors trading in foreign stocks. The third essay provides new empirical evidence on the link between fair value measurement and regulatory capital restrictions in an international setting. Further, we show a substantial degree of non-compliance with fair value disclosures that is associated with detrimental capital market effects. This result confirms the concerns by IFRS critics that uniform accounting standards are not sufficient to ensure homogenous and, thus, comparable reporting practices (e.g., Ball, 2006). Finally, our detailed description of the events around the emergency amendment to IAS 39 provides compelling anecdotal evidence how vested national interests can threaten the whole IFRS experiment (House of Commons, 2008). The remainder of the thesis is organized as follows. Section 2 contains the first essay with the literature review on mandatory IFRS adoption. The third section presents the second essay on the impact of mandatory IFRS adoption on cross-border equity investments of individual investors. Section 4 comprises the third essay on the economic consequences of the fair value reclassifications following the emergency amendment to IAS 39. The fifth and final section summarizes the main findings of this thesis and concludes.

4

2 Intended and unintended consequences of mandatory IFRS adoption 2.1 Introduction The IAS Regulation (EC, 2002) requires most listed EU firms to prepare their consolidated accounts according to IFRS since 2005. This essay complements an earlier review by Soderstrom and Sun (2007) on voluntary IFRS adoption: We summarize and discuss the extant empirical literature on the economic consequences of mandatory IFRS adoption and provide suggestions for future research. Following Zeff (1978) and Holthausen and Leftwich (1983) we use the term economic consequences to denote any effects of financial reporting on firm values and on the wealth of those who make decisions based on accounting information or are affected by such decisions. We define economic consequences to be intended (unintended) if they can (cannot) be reconciled with the explicitly stated objectives of the IAS Regulation. These objectives emphasize positive capital-market as well as macroeconomic effects resulting from enhanced transparency and cross-country comparability of financial statements. However, they do not refer to the impact of accounting on contractual relationships. Our distinction between intended and unintended consequences therefore corresponds to that between the informational and the contracting (or stewardship) roles of accounting. We document that the rich empirical literature on the intended consequences of mandatory IFRS adoption generally fails to find an increase in the comparability or transparency of financial statements. In contrast, evidence of positive reactions in capital markets and at the macroeconomic level is plentiful and almost unanimous. Since these capital-market and macroeconomic reactions are assumed to be achieved through IFRS adoption making financial statements more comparable and/or transparent, these two sets of research findings appear to be in disagreement. We argue that this disagreement is likely due to the difficulty of separating a potential IFRS effect from concurrent changes that are unrelated to financial reporting, such as regulatory or institutional changes that impact on the microstructure of capital markets. In addition, we show that extant evidence on the capital-market effects of mandatory IFRS adoption relies on commercial databases that suffer from a systematic bias towards large companies. To the extent that large companies are more likely to benefit from IFRS (e.g., Christensen, Lee, and Walker, 2007), this database bias contributes towards finding positive effects of mandatory IFRS adoption. 5

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1_2, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

Empirical literature on the unintended consequences suggests that mandatory IFRS adoption has material effects on contractual outcomes. For example, Christensen, Lee, and Walker (2009) find evidence that mandatory IFRS adoption leads to significant wealth transfers between lenders and shareholders through its impact on debt covenants that are based on rolling GAAP. Other studies and insights from non-IFRS settings indicate that mandatory IFRS adoption triggers similar unintended wealth transfers due to its influence on individual (e.g., management compensation plans) and collective contracts (e.g., dividend payouts, regulatory restrictions in the bank industry). However, direct evidence on the unintended consequences of mandatory IFRS adoption is still scant at present. We expect this literature to grow as researchers have started to recognize the importance of the stewardship function of accounting in the context of IFRS adoption (e.g., Wu and Zhang, 2009a). The IAS Regulation was accepted and published in 2002. This early announcement gave firms (represented by their management) and investors time to conduct a costbenefit analysis in anticipation of mandatory IFRS adoption and react accordingly. Empirical literature finds evidence of significant stock market reactions to events that changed the likelihood or scope of mandatory IFRS adoption. Cross-sectional variation in these reactions is associated with firm-level or country-level proxies for the net benefits of IFRS. Extant literature also indicates that firms had incentives to engage into opportunistic anticipatory actions, e.g., delisting from EU-regulated trading segments to avoid mandatory IFRS adoption or manage its impact through lobbying efforts. In summary, our analysis of the extant empirical evidence suggests that both the intended and the unintended consequences merit further scrutiny to establish a balanced view on the overall impact of mandatory IFRS adoption. We therefore conclude our discussion by, first, providing suggestions on how to address research design challenges in order to identify intended consequences more precisely. We also argue that progress is likely to come from disclosure and compliance studies that rely on manually collected data, as well as from studies of smaller firms. Second, we propose several avenues for future research on the unintended consequences of mandatory IFRS adoption, including the use expert accounting knowledge to identify beneficial settings for studying these effects as well as the use of actual contracting variables or, where absent, more precise proxies. Prior literature reviews on mandatory IFRS adoption are confined to early evidence on the intended consequences (e.g., Hail and Leuz, 2007; Hail, Leuz, and Wysocki, 2010). 6

This paper adds to this stream of research by providing a comprehensive review of present literature that encompasses both the intended and the unintended consequences of mandatory IFRS adoption. Considering the most recent evidence allows us to point out the mismatch in research findings on the intended consequences and identify the unintended consequences of mandatory IFRS adoption as a fruitful area for future research. Our analysis is based on the objectives of the IAS Regulation and primarily focuses on empirical evidence from the EU. However, since the objectives of mandatory IFRS adoption are similar around the world (see, e.g., AGFRC 2002), we also draw on literature from other settings in case EU-based evidence is absent. The findings in this paper are therefore likely to extend beyond the EU setting and hopefully prove to be useful for researchers and regulators alike. The rest of this essay is organized as follows. In section 2.2, we develop the conceptual underpinnings of our distinction between intended and unintended economic consequences of mandatory IFRS adoption. Section 2.3 is a review of extant evidence on the intended consequences. Section 2.4 focuses on empirical studies on the unintended consequences. In section 2.5, we discuss the literature on evasive actions that firms take in anticipation of the mandated introduction of IFRS. Section 2.6 summarizes our findings and provides suggestions for future research. Section 2.7 contains the tables. 2.2 Conceptual underpinnings 2.2.1 Economic consequences Academic literature on the impact of financial reporting typically refers to economic consequences, a concept introduced by Zeff (1978:   DV ³WKH LPSDFW RI DFFRXQWLQJ reports on the decision-making behavior of business, government, unions, investors and FUHGLWRUV´ ,Q D VLPLOar vein, Holthausen and Leftwich (1983: 77) view accounting as having HFRQRPLF FRQVHTXHQFHV ³LI FKDQJHV LQ WKH UXOHV XVHG WR FDOFXODWH DFFRXQWLQJ QXPEHUVDOWHUWKHGLVWULEXWLRQRIILUPV¶FDVKIORZVRUWKHZHDOWKRISDUWLHVZKRXVHWKRVH numbers for contraFWLQJ RU GHFLVLRQ PDNLQJ´ We follow these definitions and use the term economic consequences to denote any effects of financial reporting on firm values and on the wealth of those who make decisions based on accounting information or are affected by such decisions. Financial reporting can trigger economic consequences through its two fundamental roles. The first fundamental role of financial reporting is to provide current and potential investors with decision-useful information on a compDQ\¶V EXVLQHVV activities (e.g., 7

IASCF 2001, paras. 10, 12). From this information perspective, financial reporting helps decrease information asymmetries between more and less well-informed investors as well as between management and investors. As a consequence, investors arrive at less noisy and/or less biased predictions of future cash flows. Thus, financial reporting potentially affects firm values by influencing the information set of current and potential investors. The second fundamental role of financial reporting is to hold management accountable for the resources entrusted to it (e.g., IASCF 2001, para. 14). Consistent with this stewardship function of accounting, contracts between the firm and its stakeholders are frequently based on financial accounting numbers. These contracts are either set at the firm level (individual contracts) or determined for multiple firms, e.g., by means of regulation at the jurisdictional level or through collective private arrangements (collective contracts). Individual contracts include performance-based management compensation plans to mitigate shareholder-manager conflicts (e.g., Jensen and Meckling, 1976; Healy, 1985) and lending agreements that contain financial covenants to mitigate conflicts between lenders and shareholders (e.g., Smith and Warner, 1979; Leftwich, 1983). Examples of collective contracts are dividend payout restrictions tied to accounting income (e.g., Leuz, Deller, and Stubenrath, 1998) and the determination of taxable income based on financial statements (e.g., Shackelford and Shevlin, 2001). Positive accounting theory focuses on this contracting perspective and emphasizes that financial reporting in general and accounting choices in particular affect contractual outcomes (Watts and Zimmermann, 1986; for a synthesis and review of early evidence, see Holthausen and Leftwich, 1983). These RXWFRPHVDIIHFWVWDNHKROGHUV¶decisions, which in turn influences firm values. It is important to point out that the two fundamental roles of financial reporting are not necessarily compatible with each other. Indeed, recent research suggests that the information role of accounting information is negatively related to its contracting role. For example, Gassen (2008) shows that stewardship-related demand for and properties of accounting information vary inversely with the valuation usefulness of that information. In a similar vein, N. Li (2010) documents that debt contracts tend to exclude transitory components from US GAAP earnings in order to increase contracting usefulness. To the H[WHQW WKDW 86 *$$3 HDUQLQJV UHIOHFW DFFRXQWLQJ¶V information role, these adjustments are indicative of incompatibilities between the two roles of accounting.

8

2.2.2 Intended versus unintended consequences of mandatory IFRS adoption This paper explores the economic consequences of mandatory IFRS adoption. Building on the discussion in section 2.2.1, we now distinguish between two types of economic consequences: intended consequences and unintended consequences. Intended consequences We refer to the intended consequences of mandatory IFRS adoption as those that feature in the official pronouncements of regulators and the objectives explicitly stated therein. The IAS Regulation expresses its pursued objectives in Article 1:4 ³7KLV 5HJXODWLRQ KDV DV LWV REMHFWLYH WKH DGRSWLRQ DQG XVH RI LQWHUQDWLRQDO accounting standards in the Community with a view to harmonising the financial information presented by the companies ... in order to ensure a high degree of transparency and comparability of financial statements and hence an efficient functioning of the Community capital market and of the Internal 0DUNHW´(&$UW In addition to the stated goals in the key source document, standard setter representatives and EU officials also communicated their expectations. For example, the chairman of the IASB, Sir David Tweedie, announced: ´A common financial language, applied consistently, will enable investors to compare the financial results of companies operating in different jurisdictions more easily and provide more opportunity for investment and diversification. The removal of a major investment risk ± the concern that the nuances of different national accounting regimes have not been fully understood ± should reduce the cost of capital and open new opportunities for diversification and improved investment returns.´ (Tweedie, 2006) Similarly, the European Commissioner for the Internal Market and Services, Charlie McCreevy, proclaimed: ´$VXVHUVEHFRPHPRUHIDPLOLDUDQGFRQILGHQWZLWK,)56WKHFRVWRIFDSLWDO for companies using IFRS should fall. It should lead to more efficient capital 4

)RU H[DPSOH WKH $XVWUDOLDQ )LQDQFLDO 5HSRUWLQJ &RXQFLO )5&  ³IXOO\ VXSSRUWV WKH *RYHUQPHQW¶V view that a single set of high quality accounting standards which are accepted in major international capital markets will greatly facilitate cross-border comparisons by investors, reduce the cost of capital, DQGDVVLVW$XVWUDOLDQFRPSDQLHVZLVKLQJWRUDLVHFDSLWDORUOLVWRYHUVHDV´$*)5&  9

allocation and greater cross-border investment, thereby promoting growth and HPSOR\PHQWLQ(XURSH´ (McCreevy, 2005) In summary, European regulators expect mandatory IFRS adoption to increase the transparency and comparability of financial statements, triggering desirable economic consequences in equity and debt markets (e.g., lower cost of capital on average 5) which in turn are expected to translate into positive effects at the macroeconomic level (e.g., increase in growth and employment). Unintended consequences We define the unintended consequences of IFRS adoption to be those that are not an H[SOLFLWSDUWRIUHJXODWRUV¶VWDWHGREMHFWLYHV:HDUJXHWKDWWKHDGRSWLRQRI,)56PD\QRW only fail to meet these intended objectives (or even lead to the opposite of what was expected) ± rather, its impact on financial statements may also extend to areas that regulators did not explicitly emphasize ex ante. As pointed out in section 2.2.1, these unintended consequences are largely associated with the contracting role of accounting. Mandatory IFRS adoption changes accounting numbers that underlie certain contracts. The contractual parties may attempt to mitigate the impact of these accounting changes by adjusting the contracts accordingly. However, such adjustments require renegotiations that are costly and sometimes even impossible (e.g., when debt covenants are based on rolling GAAP). Moreover, it is doubtful whether renegotiations result in contractual arrangements that fully offset the effects of IFRS adoption on accounting numbers, especially if there is information asymmetry about the potential impact of IFRS. Accounting-based contracts that evolved under local GAAP may therefore turn out to be less efficient subsequent to mandatory IFRS adoption. This loss in efficiency leads to redistributions of cash flows if one party, for example a lender in a debt covenant setting, can control the impact of IFRS on the contractual arrangement. It also leads to detrimental wealth effects to both parties if the new arrangement creates adverse incentives that result in deadweight agency costs. In addition to these unintended economic consequences of IFRS adoption, various other consequences are conceivable. For example, Sunder (2010) discusses potential adverse effects of globally uniform accounting standards on the accounting profession 5

10

Note that a change in capital allocation efficiency, by construction, leads to relative winners and losers. Hence, even if IFRS increases the efficiency of capital allocation, some firms may end up KDYLQJ ORZHU PDUNHW YDOXDWLRQV +RZHYHU WKH UHJXODWRUV¶ VWDWHPHQWV PDNH FOHDU WKDW WKH average effect on firm value is expected to be positive.

and on accounting academia. Biondi and Suzuki (2007) summarize research on the socioeconomic impacts of IFRS. While these aspects are certainly important, it is beyond the scope of this paper to pursue them in detail. Interim conclusion It is noteworthy that the stated objectives of the IFRS mandate focus on the information role of accounting and do not refer to the contracting perspective. Arguably, this fact owes to the supranational character of the regulatory process. Supranational accounting regulation seeks to exploit network externalities by enhancing the comparability of financial statement information across all jurisdictions involved (e.g., Leuz, 2010). While these network externalities create obvious benefits for investors, it is less clear how they affect accounting-based contracts that are typically set at the firm level or at the jurisdictional level (see section 2.2.1).6 Our conceptual distinction between the intended and unintended consequences of mandatory IFRS adoption therefore largely corresponds to that between the information and contracting roles of accounting. 2.3 Intended consequences of mandatory IFRS adoption In this section, we review literature that addresses the intended consequences of mandatory IFRS adoption. We begin by providing a short discussion of the underlying theoretical arguments (section 2.3.1), continue with a summary of the empirical literature (section 2.3.2), and conclude with a critical evaluation of the extant evidence (section 2.3.3). 2.3.1 Theoretical background The objectives of the IAS Regulation rely on two crucial assumptions. First, mandating IFRS is sufficient to improve (transparency argument) and/or level (comparability argument) reporting practices across countries. Second, improved and/or leveled reporting practices yield positive economic consequences, i.e. trigger financial statement user responses that ultimately lead to higher firm values on average. It is debatable whether these assumptions are justified. Based on the observation that accounting regimes generally provide firms with considerable reporting discretion,7 recent accounting research concludes that reporting 6

Consequently, the IAS Regulation passes regulatory power in this area on to the EU member states.

7

It is not clear whether IFRS provide financial statement preparers with more or less discretion than other sets of accounting standards. Rather, any set of accounting standards inevitably requires 11

SUDFWLFHV DUH WR D ODUJH H[WHQW GHWHUPLQHG E\ ILUPV¶ UHSRUWing incentives, and that the underlying accounting standards act as mere side conditions (e.g., Ball, Kothari, and 5RELQ%XUJVWDKOHU+DLODQG/HX] )LUPV¶UHSRUWLQJLQFHQWLYHVDUHVKDSHG by jurisdiction-level institutional factors such as legal systems, enforcement regimes, and capital-market forces, as well as by firm-level compensation arrangements, ownership structures, and governance mechanisms. Hence, the impact of IFRS adoption on UHSRUWLQJSUDFWLFHVLVOLNHO\WREHOLPLWHGLIDILUP¶V institutional environment and firmlevel incentives remain unchanged (e.g., Ball, 2006; Soderstrom and Sun, 2007; Hail, Leuz, and Wysocki, 2010). These insights stand in contrast to statements made by regulators that seem to imply the expectation of a homogenizing impact of IFRS adoption in isolation on reporting practices across different firms and jurisdictions. Even if reporting practices become more transparent and/or more comparable subsequent to IFRS adoption, it is not clear whether and what kind of economic consequences will follow. Recent theoretical literature shows that the directional impact of high-quality accounting information on the cost of capital is ambiguous (e.g., Lambert, Leuz, and Verrecchia, 2007). This finding may help explain why empirical evidence on the relation between disclosure or earnings properties and the cost of capital is inconclusive (for an overview, see Leuz and Wysocki, 2008). Similarly, it is unclear whether more comparable financial statements enhance cross-border investment if underinvestment in foreign equities is due to behavioral biases or rational tendencies to invest in geographically proximate countries (e.g., Beneish and Yohn, 2008; for a theoretical model on detrimental effects of accounting harmonization in equity markets, see Barth, Clinch, and Shibano, 1999). This discussion shows that the objectives of the IAS Regulation rely on assumptions that are not necessarily supported by prior research. It is important to keep this caveat in mind when interpreting the empirical evidence on the intended consequences of mandatory IFRS.

preparers to exercise reporting discretion and make subjective judgments (e.g., Leuz, 2006; Nobes, 2006). 12

2.3.2 Empirical evidence This section reviews empirical evidence on the intended consequences of mandatory IFRS adoption.8 We structure this section by distinguishing between accounting effects (section 2.3.2.1), capital-market effects (section 2.3.2.2), and macroeconomic effects (section 2.3.2.3). Research on accounting effects of mandatory IFRS adoption analyzes the immediate impact of the change in accounting standards on reporting practices. Empirical literature on capital-market and macroeconomic effects evaluates financial VWDWHPHQW XVHUV¶ UHVSRQVHV WR PDQGDWRU\ ,)56 DGRSWLRQ &DSLWDO-market effects are measured by firm-level variables related to stock and debt markets. Studies on macroeconomic effects, in contrast, rely on industry- and/or country-level statistics that may or may not include capital-market measures.9 2.3.2.1 Accounting effects We classify the empirical literature on the accounting effects of mandatory IFRS adoption into three categories: compliance and accounting choice studies, studies analyzing the properties of accounting numbers, and value relevance studies.10 Compliance and accounting choice studies Cascino and Gassen (2010) provide one of the few studies that investigate compliance with disclosure rules after mandatory IFRS adoption. Using a matched sample of 153 German and 153 Italian firms they find that, even under the same set of accounting standards, predictable country- and firm-level incentives remain important determinants of reporting practices. In a similar vein, Gaeremynck, Thornton, and Verriest (2009) examine the properties of mandatory IFRS adoption using three different measures: early application of IAS 39, transparency of disclosures made in the notes to the financial statements, and rigor of initial application of IFRS in the income statement. Based on a sample of 152 European firms, they document that better governed firms adopt IFRS in a 8

Given the abundance of empirical literature on this subject, we focus on cross-country studies. Exceptions from this approach are made for single-country studies that we consider to be particularly innovative in terms of the research question addressed and/or the research design employed.

9

Some studies reviewed in this section analyze measures that could be affected by both intended and unintended consequences of mandatory IFRS adoption. For example, the cost of debt is likely to be influenced by the informational as well as the contracting effects of IFRS. This section focuses on literature that exclusively refers to the informational effects in interpreting the results.

10

We categorize value relevance studies with the literature on accounting effects, but acknowledge that changes in value relevance around mandatory IFRS adoption may also reflect capital-market effects, e.g., in the form of IFRS-induced changes in the market value of equity. 13

more transparent and less opportunistic way. Kvaal and Nobes (2010) analyze several accounting choices under IFRS using a sample of 232 companies from four European countries and Australia. They find strong evidence that pre-IFRS methods continue to be used where this is allowed under IFRS. Similarly, de las Heras, Moreira, and Teixeira (2010) provide evidence that accounting choices under IFRS differ considerably across Europe and that institutional factors such as the tax, financial, and enforcement systems are important determinants of these differences. The results presented in this subsection are in line with those from non-academic surveys (e.g., KPMG and von Keitz, 2006; Ernst & Young, 2007) and call into question whether IFRS adoption alone facilitates the comparability of financial statements across countries. Studies analyzing the properties of accounting numbers While compliance studies typically rely on hand-collected data and thus are confined to relatively small samples, most analyses of the properties of accounting numbers allow for large-sample evidence as relevant information can easily be obtained from commercial databases.11 Adopting the latter approach, Garcia Osma and Pope (2009) examine earnings management attributes across 30 countries before and after mandatory ,)56DGRSWLRQ7KHLUUHVXOWVLQGLFDWHWKDWLQVWLWXWLRQVVKDSLQJILUPV¶UHSRUWLQJLQFHQWLYHV such as the enforcement regime and investor protection rules, dominate accounting standards in determining the properties of earnings. These findings are generally in line with those from other cross-country studies that apply a similar research design (e.g., Jeanjean and Stolowy, 2008; Ahmed, Neel, and Wang, 2009; Cascino and Gassen, 2010; Chen, Tang, Jiang, and Lin, 2010). Christensen, Lee, and Walker (2008) analyze the impact of incentives versus standards on accounting properties by comparing voluntary and mandatory IFRS adoption. In contrast to the studies mentioned above, they focus on German companies, thereby holding the institutional environment constant. The authors observe accounting property improvements for firms that voluntarily switched to IFRS, but fail to find such effects for mandatory adopters. Since voluntary adopters choose to switch accounting standards and therefore are more likely to have reporting incentives

11

14

Not all studies on the properties of accounting numbers are able to retrieve appropriate information from commercial databases. For example, Gebhardt and Novotny-Farkas (2010) use manually collected data for 90 banks from the EU to analyze how IFRS affects accounting quality metrics through its impact on loan loss provisions. They find mixed results as both income smoothing and timely loss recognition appear to decrease following mandatory IFRS adoption.

that support high-quality application of IFRS, these results again suggest that incentives are more important determinants of accounting properties than standards.12 Lang, Maffett, and Owens (2010) analyze the impact of mandatory IFRS adoption on two measures of accounting comparability. They find that while cross-country earnings co-movement increases following mandatory IFRS adoption, the comparability of the mapping of earnings into returns decreases. The increase in earnings co-movement is associated with negative effects on analyst forecast properties and bid-ask spreads. Taken together, these results cast doubt that mandatory IFRS adoption enhances cross-country comparability of financial statements. Value relevance studies Morais and Curto (2009) provide evidence that the value relevance of book value of equity and net income has increased among EU companies that mandatorily adopted IFRS. However, they also document that even after the introduction of IFRS the value relevance of accounting information differs substantially across countries. Similar evidence is reported for French and German firms by Liao, Sellhorn, and Skaife (2009), while Clarkson, Hanna, Richardson, and Thompson (2010) are able to attribute IFRSinduced increases in value relevance mostly to code law countries. Aharony, Barniv, and Falk (2010) find that the value relevance of three accounting items (goodwill, research and development expenses, and revaluation of property, plant, and equipment) increases following mandatory IFRS adoption in the EU. This effect is positively associated with the deviations of local GAAP numbers from their corresponding IFRS values. Morricone, Oriani, and Sobrero (2009) examine a sample of 267 Italian firms and find that mandatory IFRS adoption has generally not increased the association of recognized intangible assets and equity market values. Wu and Zhang (2009b), in contrast, focus on the credit relevance of accounting information. They show that voluntary adoption of IFRS or US GAAP is associated with 12

The Christensen, Lee, and Walker (2008) analysis illustrates the different methodological challenges faced by voluntary versus mandatory IFRS adoption studies. While voluntary IFRS adoption is not the focus of our review, we note that this literature generally finds it easier to isolate the effect of a change in accounting standards (as opposed to a contemporaneous change in incentives) on accounting properties because voluntary adoption settings usually allow for the construction of a control sample to achieve a difference-in-differences research design. In contrast, there usually is no obvious control sample in mandatory adoption settings because all firms in the jurisdiction under study switch accounting standards at the same time. Possible ways to nonetheless achieve a difference-indifferences design include matching mandatory adopters to voluntary adoption firms that switched to IFRS earlier, or firms from non-IFRS jurisdictions. Regardless of the setting under study, selfselection bias is a potential challenge in most IFRS adoption studies. 15

significant increases in the sensitivity of credit ratings to the accounting default factor. Mandatory IFRS adoption triggers the same effect, but only in countries with a strong rule of law.13 Similarly, Bhat, Callen, and Segal (2010) analyze the impact of mandatory IFRS adoption on the ability of earnings to convey credit risk information. Using Credit Default Swap spreads as a proxy for credit risk they provide evidence that IFRS earnings are more informative determinants of credit risk than earnings under local GAAP. However, this effect is confined to companies in code law countries that are internationally oriented (i.e., have high foreign sales) and that appear to be more transparent (i.e., trigger more accurate and less dispersed analysts forecasts) post-IFRS. 2.3.2.2 Capital market effects The empirical literature on capital-market effects to mandatory IFRS adoption can be classified into two broad categories: studies providing indirect evidence by examining capital-market perceptions of accounting quality, and studies that directly analyze economic consequences in capital markets. Indirect evidence on economic consequences in capital markets The first category of literature involves measures such as the information content of earnings announcements, stock return synchronicity, equity analyst forecast properties, or credit ratings. These measures do not directly capture economic consequences in the sense of value changes, but serve as proxies for the quality of financial statement information as perceived by capital-market participants. These perceptions, in turn, are assumed to influence firm values. Wang, Young, and Zhuang (2008) as well as Landsman, Maydew, and Thornock (2010) investigate the information content of earnings announcements, as measured by abnormal return volatility, around the mandatory introduction of IFRS. Both studies find mild evidence of an increase in the information content of earnings announcements postIFRS with firms from code-law countries experiencing the largest effects. Kim and Li (2010) focus on financial reporting externalities by analyzing how investors react to earnings announcements of other firms in the same industry. Their results suggest that 13

16

Wu and Zhang (2009b) are careful to point out that their results allow two alternative interpretations. First, IFRS/US GAAP financial statements provide useful information that was previously unavailable to credit rating agencies under local GAAP. Second, IFRS/US GAAP numbers better reflect the information that was already available to credit rating agencies through other channels. For a discussion on the interpretation of value relevance studies and their relevance for accounting standard setting, see Holthausen and Watts (2001) and Barth, Beaver, and Landsman (2001).

such intra-industry information transfers increase after mandatory IFRS adoption. This effect is stronger when announcing firms are from countries with lower pre-adoption earnings quality or larger differences between their local standards and IFRS. Beuselinck, Joos, Khurana, and van der Meulen (2010) examine the impact of mandatory IFRS adoption on stock price informativeness using stock return synchronicity. Consistent with IFRS reports revealing new firm-specific information in the adoption period and subsequently lowering the surprise of future disclosures, they find a V-shaped pattern in stock return synchronicity around the mandatory introduction of IFRS. This result is primarily driven by firms domiciled in countries with large differences between local GAAP and IFRS. Horton, Serafeim, and Serafeim (2009) investigate responses to mandatory IFRS adoption by equity analysts instead of analyzing stock price reactions directly. Using forecast accuracy, disagreement between analysts and information precision of individual forecasts as proxies for transparency, they document a general improvement in the information environment of firms that adopt IFRS. Their results are more pronounced for firms with greater differences between local GAAP and IFRS earnings. 14 Similarly, Preiato, Brown, and Tarca (2009) find a significant reduction in forecast errors and disagreement between equity analysts following mandatory IFRS adoption in the EU. 3UR[LHV IRU WKH VWUHQJWK RI D FRXQWU\¶V HQIRUFHPHQW V\VWHP DUH RQO\ PLOGO\ DVVRFLDWHG with analyst forecast properties, leading the authors to conclude that institutional differences may be less important than is commonly believed. Tan, Wang, and Welker (2009) use a unique database that enables them to identify the location of equity analysts. They find evidence that mandatory IFRS adoption attracts analysts from abroad and improves the accuracy of forecasts made by foreign analysts. Their results are particularly pronounced for analysts who are located in countries that are simultaneously adopting IFRS. Panaretou, Shackleton, and Taylor (2009) use a sample of 665 nonfinancial firms from the UK and find that the positive impact of mandatory IFRS DGRSWLRQRQDQDO\VWV¶IRUHFDVWHUURUVDQGGLVSHUVLRQLVVLJQLILFDQWO\VWURQJHUIRUILUPVWKDW use derivatives. Their findings suggest that, at least in this specific setting, the observed improvement in forecasts by equity analysts is mainly driven by new hedge accounting rules under IFRS.

14

In addition to the information content of earnings announcements, Wang, Young, and Zhuang (2008) also analyze analyst forecast properties. Their results are generally consistent with those presented by Horton, Serafeim, and Serafeim (2009). 17

Shahzad (2010) examines how credit rating agencies react to mandatory IFRS adoption. Using a sample of 788 bond issues by European financial firms he provides evidence that, post-IFRS, firms receive higher credit ratings and credit agencies are less likely to disagree when assigning bond ratings. These results are consistent with the view that the introduction of IFRS results in financial statements that are more informative to debt market participants. In contrast, Pagratis and Stringa (2009) find a negative impact of IFRS adoption on credit ratings for a global sample of 293 banks covering the period 1999-2006. They attribute these results to higher volatility in IFRS accounting numbers, but also acknowledge the alternative explanation that the sample period ends before potential benefits of IFRS adoption materialize in credit ratings. Direct evidence on economic consequences in capital markets The second category of literature on the capital-market effects of mandatory IFRS adoption provides direct evidence on economic consequences by using measures that are strongly associated with firm values. These measures include the implied cost of equity capital, bid-ask spreads, stock liquidity, foreign ownership, and the cost of debt. S. Li (2010) finds that in EU countries with strong enforcement systems the implied cost of equity capital decreases after mandatory IFRS introduction. In contrast, Daske, Hail, Leuz, and Verdi (2008) provide only weak evidence of a cost of equity capital effect using a large sample that includes firms from 26 IFRS-adopting countries around the world. On the other hand, they document a significant increase in stock liquidity postIFRS in countries where firms have incentives to be transparent and where legal enforcement is strong. To the extent that stock liquidity is reflected in stock returns (e.g., Amihud and Mendelson, 1986), this result is consistent with an IFRS-induced increase in firm values.15 In a more specific setting, Muller, Riedl, and Sellhorn (2010) provide evidence that mandatory IFRS adoption reduces, but does not eliminate, pre-IFRS information asymmetry differences (measured by the bid-ask spread) across European real estate firms. Gkougkousi and Mertens (2010) focus on European banks and insurance companies and find that the implied cost of equity capital decreases and stock liquidity increases following mandatory IFRS adoption. Research that focuses on emerging markets provides inconclusive evidence. LagoardeSegot (2009) analyzes market microstructures in 28 emerging equity markets, but fails to 15

18

In a concurrent research study, Hail and Leuz (2007) focus their analysis on the effects in EU countries. Their results are similar to those presented in Daske, Hail, Leuz, and Verdi (2008).

document an impact of IFRS adoption on stock liquidity. In contrast, Bova and Pereira (2010) confine their sample to Kenyan firms and find that higher IFRS compliance is associated with capital-market benefits in form of higher foreign ownership and increased share turnover. DeFond, Hu, Hung, and Li (2009) examine changes in equity ownership by foreign institutional investors. Their results suggest that stocks of mandatory IFRS adopters experience an increase in demand from abroad if proxies for comparability benefits and the credibility of IFRS implementation are sufficiently high. Similar evidence is presented by Yu (2009) and Florou and Pope (2009). In the second part of this thesis, we document that German individual investors increase trading in foreign, i.e. non-German, stocks of mandatory IFRS adopters. Taken together, these findings are consistent with the notion that the mandatory introduction of IFRS enhances cross-border investment in stock markets by both institutional and individual investors. Ceteris paribus, this increase in cross-border investment increases the market values of IFRS-adopting firms. Florou and Kosi (2009) focus on debt markets instead of equity markets. They predict that mandatory IFRS adopters are more likely to issue public bonds and experience a decrease in bond yield spreads, because IFRS enhances the quality and comparability of their financial statements. Their results are consistent with these predictions, but only in countries with more developed enforcement mechanisms, higher control of corruption and lower financial risk. These findings suggest that mandatory IFRS adoption has the potential to impact on firm values by reducing the cost of public debt. 2.3.2.3 Macroeconomic effects To date, evidence on the macroeconomic effects of mandatory IFRS adoption is relatively scarce. Beneish, Miller, and Yohn (2010) analyze, separately for equity and GHEW PDUNHWV WKH SHUFHQWDJH RI D FRXQWU\¶V PDUNHW FDSLWDOL]DWLRQ WKDW UHIOHFWV IRUHLJQ investment. They show that the introduction of IFRS has no discernable effect on foreign equity investment, but find evidence for a significant increase in cross-border debt investment. Amiram (2009) uses bilateral data on foreign equity portfolio holdings at the country level. He finds that foreign investors have significantly higher equity holdings in countries that introduced IFRS. This relation is stronger if the foreign investors are from countries that also mandate IFRS, and if the institutional environment of the investee country ensures low corruption and high investor protection. Finally, Marquez-Ramos (2008) reports evidence that the mandatory introduction of IFRS reinforces international 19

trade flows and foreign direct investments. These effects are particularly pronounced in transition countries. 2.3.3 Discussion Extant empirical evidence on the intended consequences of mandatory IFRS adoption can be broadly summarized as follows. First, the accounting effects of mandatory IFRS adoption seem to be limited as reporting practices continue to be dominated by incentives at the firm and jurisdictional levels. Second, there is plenty of evidence that the mandatory introduction of IFRS coincides with capital-market and macroeconomic benefits. There are two potential non-mutually exclusive explanations for these contradictory research findings. The first explanation is that the literature on the accounting effects of IFRS applies measures that are irrelevant to users of financial statements. Our literature review shows that most accounting effects studies rely on aggregate numbers that can be retrieved from commercial databases. It is beyond the scope of this paper to discuss whether accounting property measures based on such aggregate numbers capture precisely what financial statement users are interested in. Yet, insights from prior literature that suggest a limited impact of mandatory IFRS adoption on reporting practices (see section 2.3.1) rely on the same measures. Financial statement users are certainly also interested in information beyond aggregate numbers. Empirical evidence on how such information (e.g., additional disclosures) changes following mandatory IFRS adoption, however, is scarce and ambiguous. We conclude that extant evidence on the accounting effects of mandatory IFRS adoption is in line with prior literature, but that more insights on the impact beyond aggregate numbers is warranted (see also section 2.6). The second explanation for the mismatch in research findings is that the studies on capital-market and macroeconomic effects capture something that is not related to mandatory IFRS adoption. This explanation touches on the identification problem, i.e. the challenge of disentangling the potential impact of mandatory IFRS adoption from other concurrent changes that influence the outcome under study. These concurrent changes either affect financial statements as well (e.g., the introduction of more rigid enforcement mechanisms) or are outside the realms of financial reporting (e.g., regulatory or institutional changes that impact on the microstructure of capital markets). Only concurrent changes with no impact on financial reports have the potential to explain the contradictory findings from the IFRS literature. If these non-accounting changes are not 20

appropriately controlled for in the research design, the results suffer from low internal validity and cannot be solely attributed to changes in financial reporting, let alone to mandatory IFRS adoption. Extant empirical literature typically addresses the identification problem by using firms that are not required to adopt IFRS as a control group. While certainly better than having no benchmark at all, this approach relies on the strong assumption that non-adopting firms reflect the counterfactual, i.e. what would have happened to adopting firms in the absence of mandatory IFRS adoption. Most studies also report cross-sectional heterogeneity in the IFRS effect related to countrylevel institutional factors (e.g., a rule of law variable to proxy for the strength of the enforcement system). This strategy would allow evaluating the internal validity if the variation in the IFRS effect could be predicted unambiguously. However, this is typically not the case. While institutional factors certainly contribute to the quality of financial reporting in general (e.g., Leuz, Nanda, and Wysocki, 2003), it is less clear whether and how they are associated with changes in reporting practices following mandatory IFRS adoption (e.g., Holthausen, 2009). Hence, present evidence on cross-sectional heterogeneity in the IFRS effect is mainly descriptive. We conclude that extant literature on the capital-market and macroeconomic effects of mandatory IFRS adoption suffers from an identification problem that urges caution in interpreting the results, but also offers opportunities for future research. We discuss these opportunities in the section 2.6. In addition to the identification problem, the empirical literature on capital-market effects of mandatory IFRS adoption also faces the challenge of having to deal with potentially biased samples. In 2007, the Committee of European Securities Regulators (CESR) published a review of the implementation and enforcement of IFRS in the EU (CESR, 2007). Based on information provided by national enforcement institutions across Europe, CESR identified a total of 5,323 equity issuers that had met the requirements of the IAS Regulation and consequently prepared consolidated IFRS accounts for the fiscal year 2005. Table 2.7.1 compares this number, which we believe to be a precise estimate of the actual population of IFRS adopters, with the sample sizes used in four representative cross-country studies discussed in the previous sections (Daske, Hail, Leuz, and Verdi, 2008; DeFond, Hu, Hung, and Li, 2009; Horton, Serafeim, and Serafeim, 2009; Landsman, Maydew, and Thornock, 2010). This comparison illustrates that sample sizes in most academic studies are substantially smaller than the actual number of IFRS adopters. The main reason for this gap is that commercial database coverage outside the US suffers from a systematic bias towards large companies (e.g., Garcia Lara, Garcia Osma, and Gill de Albornoz Noguer, 2006). 21

To the extent that large companies are more likely to benefit from IFRS (e.g., Christensen, Lee, and Walker, 2007; see also the literature on voluntary IFRS adoption, e.g., Soderstrom and Sun, 2007), this database bias distorts research findings towards overstating the positive effects of mandatory IFRS adoption. Finally, we would like to stress the importance of separating the source of an economic consequence from the construct that is used to measure it empirically. Specifically, economic consequences that stem from the information and contracting roles of financial reporting, respectively, may materialize in the same empirical measures, e.g., capital-market outcomes. The literature on mandatory IFRS adoption offers a prime example of this potential conflict. Horton and Serafeim (2009) as well as Christensen, Lee, and Walker (2009) observe significant stock market reactions to IFRS reconciliations in the UK, but provide opposite interpretations for this phenomenon. While the former study interprets its evidence as being consistent with IFRS reconciliations containing value-relevant information for shareholders, the authors of the latter study argue that the stock market reactions ultimately reflect wealth transfers between lenders and shareholders. Christensen, Lee, and Walker (2009: 1168) conclude WKDW ³IDLOXUH WR FRQVLGHU DFFRXQWLQJ¶V GHEW-contracting role risks attributing IFRS UHFRQFLOLDWLRQPDUNHWUHDFWLRQVWRLQIRUPDWLRQDERXWDJLYHQILUP¶VIXWXUHRSHUDWLQJFDVK flows rather than information about the likelihood of violating covenants. This, in turn, means that one of the costs of mandatory IFRS (i.e., the effect on existing contracts) may EHLQFRUUHFWO\LGHQWLILHGDVDEHQHILW´ We do not comment on the relative merits of these conflicting positions. Rather, we stress that a lesson for future research on mandatory IFRS adoption is that observed capital-market outcomes warrant further analysis to ascertain whether the underlying cause is more consistent with an information explanation (i.e., an intended consequence according to our definition) or a contracting explanation (i.e., an unintended consequence). 2.4 Unintended consequences of mandatory IFRS adoption In section 2.2, we defined the unintended consequences of mandatory IFRS adoption DVWKRVHHIIHFWVWKDWDUHQRWDQH[SOLFLWSDUWRIUHJXODWRUV¶VWDWHGREMHFWLYHV,QWKHFDVHRI EU-wide adoption of IFRS, regulators have been particularly silent on the contracting uses of IFRS and their potential economic consequences. In this section, we identify these potential consequences and review extant evidence. We distinguish between individual contracts (section 2.4.1) and collective contracts (section 2.4.2). Individual 22

contracts are set at the firm level and, as most prominent examples, include management compensation plans and lending agreements. Collective contracts are determined for groups of firms. We focus on collective contracts that use IFRS financial statements as a basis for dividend payouts, regulatory restrictions, taxation, and employee benefits. 2.4.1 Individual contracts Management compensation plans Management compensation plans regularly include performance targets that are stated in terms of accounting numbers such as earnings per share or return on total assets. These SODQV DWWHPSW WR DOLJQ PDQDJHPHQW¶V LQWHUHVW ZLWK WKDW RI VKDUHKROGHUV E\ DZDUGLQg managers a share of profits if the performance targets are attained or exceeded. Changes in accounting rules affect the distribution of wealth between managers and shareholders if compensation plans are not adjusted to offset the rule changes. In response, managers may engage in potentially value-decreasing investment and financing decisions to compensate for inefficient compensation schemes (Holthausen and Leftwich, 1983). The previous discussion suggests that mandatory IFRS adoption can trigger economic consequences through costly adjustments of compensation plans (before or after the adoption) or opportunistic management reactions to unadjusted compensation plans. Chen and Tang (2009) provide the only related empirical study so far that we are aware of.16 Using a sample of 70 property firms from Hong Kong, they show that the gains from revaluation of investment property are positively associated with executive cash compensation after mandatory IFRS adoption, but not before. In contrast, no association is found between revaluation losses and compensation. These effects are more pronounced in companies where agency problems between managers and shareholders are more severe (e.g., firms with lower founding family ownership). Since revaluation gains and losses did not flow through the income statement before IFRS adoption, these results suggest that compensation contracts have not been adjusted to offset the change in accounting rules. Hence, managers benefited from mandatory IFRS adoption by receiving higher salaries through more attractive compensation plans. 16

Wu and Zhang (2009a) show that voluntary adoption of IFRS or US-GAAP is associated with increases in the sensitivities of CEO turnover and employee layoffs to accounting earnings. Wu and Zhang (2010) provide evidence that CEO turnover is more sensitive to the accounting performance of foreign peers following mandatory IFRS adoption. These results confirm that accounting changes such as a mandatory switch to IFRS have the potential to influence internal performance evaluation processes for managers and other employees. 23

Lending agreements 0DQDJHUV DUH W\SLFDOO\ DVVXPHG WR DFW LQ WKH VKDUHKROGHUV¶ LQWHUHVW HJ WKURXJK performance-based compensation plans (see the previous subsection). Under this assumption, management has incentives to engage in activities that benefit shareholders to the detriment of lenders (e.g., Jensen and Meckling, 1976). Such activities include the payment of liquidating dividends to the shareholders or the substitution of low-risk for high-risk investments. Corporate lending agreements therefore frequently include FRYHQDQWV WKDW LPSRVH UHVWULFWLRQV RQ PDQDJHUV¶ DFWLRQV WR FRQWURO IRU WKH FRQIOLFW between lenders and the shareholders of the borrowing firm. Such covenants regularly include restrictions based on accounting numbers, for example creditor rights to terminate the agreement upon violation of an accounting leverage ratio (e.g., Smith and Warner, 1979). Hence, similar to compensation contracts, changes in accounting rules can result in wealth transfers between lenders and shareholders if debt covenants are not adjusted to accommodate the rule changes (Holthausen and Leftwich, 1983). Since such adjustments are costly, debt covenants often include provisions that determine how to deal with future accounting rule changes by mandating to use either GAAP at the date of calculation (rolling GAAP) or GAAP in force when the contract was set up (frozen GAAP) (e.g., Leftwich, 1983; Citron, 1992). Based on this contracting perspective, we conclude that mandatory IFRS adoption has the potential to induce economic consequences through costly adjustments of debt contracts (before or after the adoption) or a redistribution of wealth between lenders and shareholders. Ormrod and Taylor (2004) predict the latter effect to dominate in the UK, because debt covenants in this market are typically based on rolling GAAP. Christensen, Lee, and Walker (2009) provide empirical evidence in line with this prediction. Their analysis is based on a sample of 137 UK firms and proceeds in two steps: First, they show that reconciliations between earnings under IFRS and UK GAAP for 2004 predict future IFRS earnings. Since rolling GAAP is prevalent in UK debt contracts, IFRS reconciliations therefore contain information on the likelihood of covenant violations. Specifically, a positive difference between IFRS and UK-GAAP earnings reduces the likelihood of covenant violations, and vice versa. In the second step, Christensen, Lee, and Walker (2009) find that the reconciliation difference between IFRS and UK-GAAP earnings is positively related to abnormal equity returns on the announcement day, suggesting that the stock market did not anticipate the impact of IFRS on earnings. This effect is more pronounced among companies that are expected to have greater contracting 24

and monitoring costs (e.g., smaller firms and firms with lower interest cover). Taken together, Christensen, Lee, and Walker (2009) provide indirect evidence that mandatory IFRS adoption leads to wealth transfers between lenders and shareholders through its impact on debt covenants. We are not aware of any other study that explicitly examines the effect of IFRS adoption on lending agreements.17 2.4.2 Collective contracts Dividend payouts Unintended consequences of mandatory IFRS adoption may also stem from the interrelation of accounting earnings and dividend payouts. For example, according to the (8¶V VHFRQG &RPSDQ\ /DZ 'LUHFWLYH¶V &DSLWDO 'LUHFWLYH18  ³EDODQFH VKHHW WHVW´ WKH maximum amount of distributable profit of EU corporations is restricted to accumulated accounting earnings. These accounting earnings have traditionally been calculated in FRPSDQLHV¶ XQFRQVROLGDWHG OHJDO-entity financial statements under domestic accounting rules. However, group earnings may be viewed as a de facto basis for distributable income because investors of the parent primarily observe the (IFRS) group accounts UDWKHU WKDQ WKH SDUHQW¶V OHJDO-entity financial statements.19 As a result, investors may claim a portion of group earnings as dividends, possibly because group earnings are SHUFHLYHGDVDEHWWHUSHUIRUPDQFHLQGLFDWRUWKDQWKHSDUHQWILUP¶VOHJDO-entity earnings. In DGGLWLRQ VRPH LQYHVWRUV PD\ EH XQDZDUH WKDW WKH SDUHQW ILUP¶V OHJDO-entity earnings commonly serve as the legal basis for dividend payouts. In both scenarios, IFRS adoption LQWKHJURXSDFFRXQWVLVOLNHO\WRLQIOXHQFHLQYHVWRUV¶GLYLGHQGFODLPVZKLFKLQWXUQPD\ DIIHFWILUPV¶GLYLGHQGSROLFLHV We are not aware of empirical research that explores these issues directly in the context of mandatory IFRS adoption. However, studies that exploit related settings reveal potentially interesting insights. For example, Goncharov and van Triest (2009) provide 17

De Jong, Rosellón, and Verwijmeren (2006) provide evidence that firms incur costs to alter their financing structures in an effort to avoid certain predictable effects of IFRS adoption on their financial statement ratios, including those assumed to be used in debt covenants. We discuss this paper in the context of IFRS avoidance strategies (section 2.5.2).

18

Second Council Directive 77/91/EEC of 13 December 1976 on coordination of safeguards which, for the protection of the interests of members and others, are required by Member States of companies within the meaning of the second paragraph of Article 58 of the Treaty, in respect of the formation of public limited liability companies and the maintenance and alteration of their capital, with a view to making such safeguards equivalent, Official Journal L 026, 31/01/1977: 1-13.

19

Pellens, Gassen and Richard (2003) report German survey evidence broadly consistent with this notion. 25

case study evidence on how the largest quarterly profit in world corporate history, recorded largely due to IFRS-style fair value gains in fiscal 2006 by Russian energy giant United Energy System (UES), resulted in an omission of dividends. Due to the legally required use of accounting earnings to calculate mandatory minimum dividends to preferred shareholders, UES saw itself under pressure to pay preferred dividends on large unrealized fair value gains. The only option to avoid this payment was to set dividends to zero for all its shareholders, common and preferred. The case study suggests that the interaction of accounting changes (e.g., IFRS adoption) with prevailing legal requirements has the potential to cause unintended redistributions of wealth (in this case, between common shareholders and preferred shareholders) if it is too costly or impossible to contract around the accounting change. Goncharov and van Triest (2011) follow up with large-sample evidence on whether upward fair-value adjustments of financial assets lead to increased dividend payouts, resulting in the distribution of unrealized earnings. Contrary to concerns commonly voiced by regulators, the evidence suggests that upward-revaluing firms actually decrease dividend payouts, even when the decision to upward revalue is treated as an endogenous choice, thus controlling for potential self-selection bias. The authors discuss two possible explanations for this finding. First, managers may use a large increase in transitory earnings to opportunistically reduce dividends, which are typically thought of as a certain fraction of persistent earnings. This explanation is consistent with wealth transfers occurring between managers and shareholders. Second, fair value gains may be correlated with an unobservable response by managers to high growth. Growth could be perceived as unsustainable and dividends are reduced towards a certain percentage of expected persistent earnings, or high growth expectations lead to expanded investment which in turn reduces free cash flow. Goncharov and van Triest (2011) are unable to empirically distinguish between these conflicting explanations. However, their findings suggest that more frequent use of fair value as a measurement basis has the potential to increase the proportion of transitory earnings (see also Hitz, 2007) and thus upsets longstanding relations between accounting earnings and dividends. To the extent that a switch from domestic accounting standards to IFRS typically increases the importance of fair value accounting, mandatory IFRS adoption may therefore cause changes in dividend policies.

26

Regulation Regulatory actions that hinge on accounting data occur in various industries and with different objectives (Watts and Zimmerman, 1978). For instance, bank regulation requires banks to maintain minimum levels of equity in relation to total assets in order to restrict default risk of financial institutions and thus provide stability to the financial system. In rate-regulated industries, e.g., electricity utilities, regulators attempt to restrict PRQRSRO\ SURILWV WR SURWHFW FRQVXPHUV¶ LQWHUHVWV DQG contribute to overall economic welfare. Rate regulation is typically based on accounting data, e.g., through restricting reported profitability to levels that are deemed acceptable. Finally, accounting information is also used in the regulation of trade flows. For example, the United States International Trade Commission officially considers the profitability of an industry as reflected in the published financial statements to determine import relief actions such as tariff increases or quota reductions (Jones, 1991). These examples demonstrate that mandatory IFRS adoption has the potential to induce unintended consequences if regulators pursue costly adjustments of contracts or review processes when accounting rules change. In the absence of such adjustments, firms have incentives to manage IFRS financial statements opportunistically to avoid (yield) unfavorable (favorable) regulatory actions. For instance, IFRS do not cater to the needs of single industries and their regulators by providing industry-specific accounting guidance. Given increased demand for such guidance, however, the IASB recently embarked on a project on rate-regulated activities that is expected to result in the issue of a new standard in the second half of 2011. To the extent that this standard deviates from national GAAP on which rate regulation is based in IFRS-adopting jurisdictions, costly revision processes are expected. This expectation is illustrated by the recent proposal issued by the Canadian Accounting Standards Board (AcSB) to grant regulated entities an option to postpone mandatory IFRS adoption from 2011 to 2013 (AcSB, 2010). Similar concerns also exist in the US where the introduction of IFRS is currently contemplated (e.g., PWC, 2008). With respect to financial-sector regulation, a report by the Committee of European Banking Supervisors (CEBS, 2007) documents material impacts of the transition from GRPHVWLF*$$3WR,)56RQEDQNV¶HTXLW\FDSLWDOPRVWO\UHODWHGWRILUVW-time application effects (pertaining e.g. to IAS 19 on employee benefits, or to IAS 40 on investment property assets). These effects, in turn, had to be addressed at the banking supervision level via specific adjustments to regulatory capital. Such IFRS-related adjustments to 27

EDQNV¶ UHSRUWHG HTXLW\ QRWDEO\ ZLWK Uespect to financial instruments (in particular to equity securities categorized as available-for-sale under IAS 39), continue to be of high significance for the calculation of regulatory capital. While these adjustments are harmonized at the European level by concurrent CEBS guidance, they are supplemented by varying adjustments at the country level (CEBS, 2007). Bushman and Landsman (2010) provide a specific example from Spain where mandatory IFRS adoption may have resulted in unintended effects on regulatory capital requirements for banks. Spanish banks were forced to switch from dynamic provisioning mandated by domestic GAAP to an incurred loan-loss provisioning model prescribed by IAS 39. Bushman and Landsman (2010: 25-  SRLQW RXW WKDW ³to the extent thDW 6SDQLVK EDQNV¶ PRYHPHQW DZD\ IURP D dynamic provisioning model affects their ability to assess their own capital needs, 6SDQLVKEDQNUHJXODWRUVFDQQRWUHO\DVKHDYLO\RQWKHEDQN¶VLQWHUQDOULVNDVVHVVPHQWDQG therefore have to expend more resources to make their own assessment of each member EDQN¶V ULVN SURILOH´ VHH DOVR %DUWK DQG /DQGVPDQ  -18). Similar scenarios in other countries and/or industries are conceivable. However, we are not aware of any research evidence that empirically documents the interrelation between mandatory IFRS adoption and regulatory actions. Taxation Few countries cede tax legislation authority to the IASB by determining taxable income directly on unadjusted IFRS financial statements. Still, tax accounting is traditionally linked to financial accounting in EU member states.20 For instance, both the UK and Germany use unconsolidated reporting as the starting point for calculating taxes. While Germany mandates unconsolidated legal-entity financial statements to be prepared under domestic GAAP, UK companies are allowed to choose between IFRS and UK GAAP for unconsolidated reporting (Schön, 2005; Ng, 2009). Hence, the direct influence of tax accounting on consolidated financial statements is limited (Gee, Haller, and Nobes, 2010). However, in some cases current national tax legislation explicitly refers to IFRS financial statements. One pertinent example is the interest deduction ceiling rule (Zinsschranke) introduced as part of the German Corporate Tax Reform Act of 2008 (Blaufus and Lorenz, 2009). This rule limits the amount of tax-deductible net interest expense to 30% of earnings before interest, taxes, depreciation and amortization 20

28

)RU DQ RYHUYLHZ RI (8 PHPEHU VWDWHV¶ UXOHV IRUWKHGHWHUPLQDWLRQ RIFRUSRUDWH WD[DEOHLQFRPH VHH Endres, Oestreicher, Scheffler, and Spengel (2007: 159-68).

(EBITDA). The reference to IFRS occurs by means of an escape clause: The interest deduction ceiling does not apply if the firm is part of a consolidated group and its OHYHUDJHUDWLRLVKLJKHUWKDQRUHTXDOWRWKHJURXS¶VOHYHUDJHUDWLR7KHOHYHUDJHUDWLRVDUH measured based on consolidated financial statements prepared under IFRS. As a result, the interested deduction ceiling rule may trigger incentives to manage IFRS leverage ratios leading to potentially distorted operating, investing, and financing decisions. We conclude that tax considerations may have generated unintended consequences around mandatory IFRS adoption through an indirect impact on consolidated financial statements. Corresponding research evidence is, however, still absent. Employee benefits In section 2.4.1, we discussed management compensation plans as one prominent example of individual accounting-based contracts that have the potential to trigger unintended economic consequences when the underlying accounting rules change due to mandatory IFRS adoption. Other forms of employee benefit arrangements can serve as examples of colOHFWLYHFRQWUDFWVWKDWPD\FDXVHDILUP¶VHTXLOLEULXPUHSRUWLQJVWUDWHJ\WR be upset when accounting rules change. While in the former case economic consequences VWHPIURPPDQDJHUV¶LQFHQWLYHWRPDQLSXODWHWKHDFFRXQWLQJQXPEHUVRQZKLFKWKHLURZQ compensation arrangements are based, collective employee benefit arrangements create other types of incentives and unintended consequences. Defined benefit pension plans are a prime example of collective employee benefit arrangements that are based on accounting information. Dixon and Monk (2009) discuss the effects of the introduction of IFRS on the extent to which defined benefit pension plans are used in the UK and the Netherlands. They argue that the move towards fair value accounting for pension plans, in particular its volatility effects on earnings and equity numbers, has worked in conjunction with other, non-accounting factors21 to provide firms with incentives to move away from defined benefit pension plans towards arrangements that transfer the retirement-income risk to the individual. For the Netherlands, this finding is empirically corroborated by Swinkels (2006). On a broader scale, the literature reviewed by Kiosse and Peasnell (2009) provides evidence consistent with accounting rules affecting the allocation of pension plan assets (for IFRS in the UK: $PLU *XDQ DQG 2VZDOG   DV ZHOO DV ILUPV¶ GHFLVLRQV WR IXQG IRU ,)56 LQ 21

These include demographic changes (e.g., greater longevity), regulatory changes intended to mitigate employee risk, as well as increasing and volatile demand for cash contributions into defined benefit plans. 29

Germany: Stadler and Lobe, 2010), terminate, freeze, curtail or convert their defined benefit plans. However, the authors also stress the importance of non-accounting factors DQG FRQFOXGH WKDW ³DFFRXQWLQJ PDWWHUV WKRXJK SHUKDSV QRW DV PXFK DV LV VRPHWLPHV FODLPHG´.LRVVHDQG3HDVQHOO 22 Rixon and Faseruk (2009) study a different type of post-employment benefit arrangement and another aspect of unintended consequences. They argue that IFRS which are primarily devised for private-sector entities may be unfit for public-sector DJHQFLHV EHFDXVH WKH\ KDUP FRPSDUDELOLW\ 7KH DXWKRUV H[DPLQH &DQDGLDQ ZRUNHUV¶ compensation boards (WCBs). These are governmental organizations representing the twelve Canadian provinces/territories. Similar to pension funds, WCBs invest revenue levied from employers in financial assets to fund employee benefits in the case of injury on the job. WC%V¶ ILQDQFLDO UHSRUWLQJ LV H[SOLFLWO\ LQWHQGHG WR IXOILOO D VWHZDUGVKLS function and to achieve cross-jurisdictional comparability. This comparability arguably suffered when WCBs adopted IFRS-style accounting for financial instruments in 2004 as a basis for calculating key performance indicators such as funded ratios that are used by employers (employees) to lobby for premium rate and worker benefit decreases (worker EHQHILW LQFUHDVHV  :&%V¶ IXQGHG UDWLRV FKDQJHG EHFDXVH   WKH WZHOYH :&%V categorized different portions of their holdings as available-for-sale investments or heldfor-trading securities, respectively, and (2) different versions of the funded ratio were used with respect to the treatment of unrealized gains/losses on available-for-sale financLDODVVHWV7KHVHFKDQJHVLPSRVHGUHDOFRVWVRQ:&%V¶VWDNHKROGHUVE\PDNLQJWKH premium rates levied by WCBs more volatile and introducing additional financial UHSRUWLQJ FRPSOH[LW\ WKDW RYHUEXUGHQHG :&%V¶ VWDNHKROGHUV &RQVLVWHQW ZLWK RXU concept of unintended consequences, Rixon and Faseruk (2009: 26) argue that adoption of IFRS-W\SH UXOHV ³KDV ZLGHU LPSOLFDWLRQV WKDQ OLNHO\ HQYLVLRQHG E\ VWDQGDUG VHWWHUV´ DQG WKDW ³WKH DGRSWLRQ RI ,)56 E\ JRYHUQPHQW DQG >QRW-for-profit] organizations will likely introduce challenges in providing stakeholders with information that will enable WKHPWRHYDOXDWHSHUIRUPDQFHDQGDFFRXQWDELOLW\´

22

30

Similarly, Dixon and Monk (2009: 625) coQFOXGHWKDW³JOREDOFRQYHUJHQFHWRZDUGIDLUYDOXHSHQVLRQ DFFRXQWLQJ KDV FRPSRXQGHG SUREOHPV IRU >GHILQHG EHQHILW SHQVLRQ SODQV@ « $FFRXQWLQJ FKDQJH LV WKHUHIRUH DQRWKHU µOD\HU¶ LQ WKH SHQVLRQ SDUDGLJP FUHDWLQJ D µWKUHVKROG HIIHFW¶ LQ D UHODWLYHO\ ORQJ caXVDOSURFHVVRILQVWLWXWLRQDOWUDQVIRUPDWLRQ´

2.5 Economic consequences in anticipation of mandatory IFRS adoption The IAS Regulation was accepted and published in 2002 but did not come into effect before fiscal years beginning on or after January 1, 2005. Hence, the introduction of IFRS was not unexpected in the EU. We therefore conjecture that firms (represented by their management) and investors conducted a cost-benefit analysis in anticipation of mandatory IFRS adoption and reacted accordingly. Section 2.5.1 discusses evidence on LQYHVWRUV¶UHVSRQVHVWRHYHQWVWKDWFKDQJHGWKHOLNHOLKRRGRI,)56EHLQJLPSOHPHQWHGLQ the EU. The subsequent three subsections deal with anticipatory actions at the firm level. Firms that expect net benefits of IFRS adoption will adjust their contractual setup within the IFRS regime. In contrast, firms that expect net costs of IFRS adoption have incentives to self-select out of the group of firms for which IFRS application is made mandatory. Some firms, while adopting IFRS overall, may attempt to avoid certain individual IFRS rules that are anticipated to be particularly costly. We discuss such avoidance strategies in section 2.5.2. Other, more pro-active approaches include lobbying with the IASB to avoid mandatory adoption of individual IFRS provisions, or with jurisdictional regulators (e.g., the European Commission) to prevent IFRS as a whole from becoming mandatory (section 2.5.3). Finally, firms could also initiate certain transactions before the IFRS mandate that are expected to be more costly afterwards (section 2.5.4). 2.5.1 Ex-ante stock market reactions Armstrong, Barth, Jagolinzer, and Riedl (2010) use event study methodology to assess stock market reactions to 16 events during the period 2002 through 2005 that are hypothesized to change the likelihood and scope of mandatory IFRS introduction in Europe. The results are consistent with investors expecting net information quality benefits as well as convergence benefits from mandatory IFRS adoption. However, LQYHVWRUV¶ UHDFWLRQV DSSHDU WR UHIOHFW FRQFHUQV ZLWK WKH SURSHU LPSOHPHQWDWLRQ DQG enforcement of IFRS in countries that have weak enforcement environments. In a related paper, Comprix, Muller, and Stanford (2003) examine a set of eleven earlier events, occurring between 2000 and 2002, which are associated with the introduction of IFRS in the EU. Their evidence is consistent with short-window stock price reactions being related to expected implementation costs, the quality of various corporate governance mechanisms and the extent to which IFRS recognition, measurement, and disclosure rules differ from the domestic rules they are expected to replace. Christensen, Lee, and Walker (2007) perform both short- and long-window event tests for a sample of UK firms to 31

H[DPLQH WKH HFRQRPLF FRQVHTXHQFHV RI WKH (8¶V GHFLVLRQ WR LPSRVH PDQGDWRU\ ,)56 7KH\ ILQG WKDW D SUR[\ IRU D ILUP¶V ZLOOLQJQHVV WR adopt IFRS predicts cross-sectional variation in the immediate as well as in the long-run stock market response to events related to this decision. Pae, Thornton, and Welker (2008) provide long-window evidence that investors expect mandatory IFRS adoption to reduce agency costs associated with ownership concentration and information asymmetry in EU firms. In summary, this stream of research suggests that, by and large, investors in European VWRFN PDUNHWV VKDUHG UHJXODWRUV¶ H[SHFWDWLRQV RI QHW HFRQRPLF EHQHfits resulting from mandatory IFRS adoption. However, the evidence indicates that investors expected these consequences to vary in the cross-section resulting in relative winners and losers from mandatory IFRS adoption. The studies presented in this subsection face the same challenges of low internal validity and biased samples as the research on intended consequences of mandatory IFRS adoption (see section 2.3.3).23 2.5.2 Avoidance strategies Avoiding IFRS adoption The literature on the economic consequences of the Sarbanes-Oxley Act (SOX) in the US provides evidence that firms pursue avoidance strategies if the introduction of mandatory financial reporting rules is expected to yield negative net benefits. For example, it is documented that firms went dark, i.e. withdrew from the SEC reporting system while continuing to trade on Pink Sheets (Marosi and Massoud, 2007; Leuz, Triantis, and Wang, 2008), or deliberately remained small (Gao, Wu, and Zimmerman, 2009) in order to avoid the costs imposed by compliance with SOX. It appears likely that firms also have taken (and are still taking) active measures to avoid compliance with IFRS. The IAS Regulation only applies if certain conditions are fulfilled: (1) the firm prepares consolidated accounts, (2) the firm is governed by the law of an EU country; DQG   WKH ILUP¶V VHFXULWLHV DUH DGPLWWHG WR WUDGLQJ RQ DQ (8 UHJXODWHG PDUNHW ,$6 Regulation, Art. 4). Hence, a company that met these conditions before the IAS Regulation became effective could have avoided IFRS adoption by (1) switching from 23

32

The sample in Armstrong, Barth, Jagolinzer, and Riedl (2010) comprises 3,265 firms from 18 European countries. Comprix, Muller, and Stanford. (2003) (Pae, Thornton, and Welker (2008)) focus on 2,152 (1,211) companies from 14 (11) EU countries. Hence, the sample sizes of the cross-country studies presented in this section also fall well short of the numbers provided by CESR (2007); see table 2.7.1.

consolidated to individual accounts,24 (2) moving its domicile outside the EU or (3) delisting from the regulated market they were trading on. Since the requirement to SUHSDUH FRQVROLGDWHG DFFRXQWV LV D FRQVHTXHQFH RI D ILUP¶V JURXp structure, the first strategy is likely to induce prohibitive costs. The same applies to the second strategy as FKDQJLQJDILUP¶VGRPLFLOHLVDIXQGDPHQWDOGHFLVLRQZLWKIDU-reaching consequences that go beyond mere accounting issues. The third strategy is implemented by either going private, i.e. an entire delisting from the stock exchange, or by going dark, i.e. switching from the regulated market to an unofficial trading segment. Unofficial trading segments that permit local GAAP for consolidated accounts exist, for example, at the Frankfurt Stock Exchange (Open Market including the Entry Standard), Euronext (Alternext) and the Nordic Stock Exchange (First North).25 Since these unofficial trading segments differ substantially from their regulated counterparts, they provide benefits (e.g., fewer disclosure requirements), but also impose costs (e.g., lower media coverage, lower stock liquidity). Therefore, the option of going dark is not simply an option between IFRS and local GAAP, but represents a weighing of the net benefits of switching markets with the net costs of adopting IFRS. If a company expects IFRS to induce high costs, unofficial trading segments may provide an attractive alternative way of accessing equity capital. The mandatory introduction of IFRS, therefore, may have affected and may still be influencing the listing choices of companies that seek to avoid IFRS application. To our knowledge, Vulcheva (2009) provides the only study to date that examines whether firms engaged in IFRS avoidance strategies. Using data from four countries (Germany, Ireland, Italy, United Kingdom) her analysis shows an abnormal increase in the delisting probability in the year of IFRS introduction. The delisting probability is greater for companies that have high reporting incentives and operate in jurisdictions with strict enforcement. This finding is consistent with the prediction that the costs of IFRS application are higher for such companies. Vulcheva (2009) concludes that

24

This strategy would not have worked throughout the EU as some member states (e.g., Greece, Belgium and Italy for some firms) also require IFRS for individual accounts.

25

Since both Alternext and First North have been established only in 2005, these trading segments could only have served as hiding places from IFRS after the IAS Regulation became effective. Note also that the Alternative Investment Market (AIM) at the London Stock Exchange gave up its status as an EUregulated market in 2003, but soon afterwards announced that it would mandate IFRS from 2007 onwards (LSE, 2005). Hence, avoiding IFRS by switching to AIM has never been an option. For more information on unofficial trading segments in the EU, see FESE (2008b). 33

standardization of accounting information might push high-quality firms out of stock markets leading to an overall decrease in accounting quality. Avoiding individual IFRS Short of avoiding IFRS adoption outright, firms also have the possibility to sidestep or at least dampen the impact of individual IFRS rules they expect to be particularly costly. 6XFKDYRLGDQFHEHKDYLRULVFRQVLVWHQWZLWKDFFRXQWLQJUXOHVKDYLQJUHDOHIIHFWVRQILUPV¶ operating, investing, and financing decisions (see, for example, Beatty, 2007). The study by de Jong, Rosellón, and Verwijmeren (2006) provides evidence suggesting that such behavior occurs in the context of IFRS adoption as well. Under IAS 32, Financial Instruments: Presentation, most preference shares are classified as liabilities while in most EU jurisdictions they had been treated as equity instruments under pre-IFRS domestic GAAP. The authors show that 71% of Dutch firms with preferred stock that under IFRS would have been classified as liabilities initiate preference share buy-backs or restructure their preference shares so as to maintain the classification as equity. The strongest predictor of these circumventive measures is the anticipated impact of the UHFODVVLILFDWLRQRQDILUP¶VGHEWUDWLR7KHDXWKRUVLQWHUSUHWWKHLUILQGLQJVDVHYLGHQFHWKDW the economic consequences being avoided are those that stem from the anticipated effect of IFRS adoption on the tightness of debt covenants (see section 2.4.1). De Jong, RosellóQ DQG 9HUZLMPHUHQ    FRQFOXGH WKDW ³,)56 « FKDQJHV ILUPV¶ UHDO capital structurH´ GLVFRXUDJLQJ ILUPV IURP XVLQJ ILQDQFLDO LQVWUXPHQWV WKDW RWKHUZLVH would have been advantageous for firms and investors. 2.5.3 Lobbying activities Managers have incentives to lobby on accounting issues if accounting rule changes affect contracts that rely on numbers reported in financial statements (Watts and Zimmerman, 1978). Thus, anticipatory cost-benefit considerations may have enticed firms (represented by their management) to lobby against individual IFRS provisions or mandatory IFRS adoption as a whole. Preventing of individual IFRSs There is an established empirical literature that documents constituent lobbying in the context of accounting standard setting (for a review, see for example Zeff, 2006). Two strands of this research are usually distinguished (e.g., Königsgruber, 2009). First, WUDGLWLRQDO OREE\LQJ VWXGLHV HPSLULFDOO\ H[DPLQH FRQVWLWXHQWV¶ SDUWLFLSDWLRQ FRDOLWLRQ EXLOGLQJ DUJXPHQWV XVHG DQG VXFFHVV LQ D VWDQGDUG VHWWHU¶V GXH SURFHVV 7KHVH VWXGLHV 34

typically analyze comment letters that are submitted in response to due process documents such as discussion papers or exposure drafts. While empirical evidence is DEXQGDQW RQ OREE\LQJ ZLWK WKH ,$6%¶V SUHGHFHVVRU WKH ,QWHUQDWLRQDO $FFRXQWLQJ Standards Committee (IASC) (e.g., Larson, 1997) and its US counterpart, the Financial Accounting Standards Board (FASB) (e.g., Armstrong, 1977), fewer studies exist for the period after the IASB constituted itself (e.g., Jorissen, Lybaert, and Van de Poel, 2006; Larson, 2007). Controversial IFRS topics studied include business combinations (e.g., Soonawalla and Ireland, 2010), share-based payment (Jorissen, Lybaert, and Van de Poel, 2006; Giner and Arce, 2004), financial instruments (Dewing and Russell, 2008; Zeff, 2006), as well as emission rights and segment reporting (e.g., Königsgruber, 2009). However, much of this evidence is anecdotal. *LYHQWKHPXOWLWXGHRIZD\VLQZKLFKFRQVWLWXHQWV¶YLHZVFDQSRWHQWLDOO\EHEURXJKWWR WKH ,$6%¶V DWWHQWLRQ HJ )OELHU +LW] DQG 6HOOKRUQ   DVVHVVLQJ WKH H[WHQW and success of lobbying based on comment letters alone is problematic. Since it is costly for constituents to reveal their true preferences in publicly visible comment letters, the salient influencing is likely to take place behind the scenes. A second stream of more recent studies therefore focuses on lobbying efforts towards the main players in the political environment. These studies are predominantly based on settings in the US. For example, Ramanna (2008) uses data on political contributions to provide evidence linking US Congress members to firms that had a vested interest in a proposed reform of goodwill accounting under US GAAP. Königsgruber (2009) provides a game-theoretical model of ILUPV¶WUDGH-RIIEHWZHHQSDUWLFLSDWLQJLQWKHVWDQGDUGVHWWHU¶VGXe process and engaging in political lobbying. This model suggests that political lobbying is more likely to occur in the EU than in the US. However, empirical evidence on this prediction is still lacking. To summarize, there are several anecdotes suggesting that lobbying is pervasive and IUHTXHQWO\ VXFFHVVIXO LQ WKH ,)56 FRQWH[W )RU H[DPSOH WKH ,$6%¶V 2FWREHU  emergency amendment to IAS 39, Reclassification of Financial Assets, can be viewed as one such case.26 However, systematic empirical findings, especially those allowing inferences regarding the specific unintended consequences that constituents strive to 26

The amendment grants companies the option to abandon fair value accounting for selected financial DVVHWV DQG ZDV WKH UHVXOW RI VWURQJ SROLWLFDO SUHVVXUH ,$6% &KDLUPDQ 6LU 'DYLG 7ZHHGLH ³QHDUO\ UHVLJQHG RYHU WKH UXOH FKDQJH GHPDQGHG E\ (XURSHDQ SROLWLFLDQV µ, ZDV VR IUXVWUDWHG E\ WKH ZKROH thing. All the time when we are trying to build a global accounting system, and we are pretty close to LW DQG WKHQ VXGGHQO\ RXWRI OHIWILHOGWKLV WKLQJ DSSHDUV ,W V MXVW DEVROXWHO\ H[DVSHUDWLQJ¶´ .HVVOHU 2008). For details on the events leading up to the reclassification amendment, see e.g. Fiechter (2010) or the third essay in this thesis. 35

avoid, is still scarce. Meanwhile, lobbying in the IFRS context is expected to intensify as numerous controversial and fundamental accounting issues are presently (re-) considered by the IASB27 and IFRS are becoming applicable in more jurisdictions around the globe. Preventing wholesale IFRS adoption In addition to analyses of lobbying behavior on individual IFRSs, insights into the expected economic consequences of IFRS adoption can potentially be gleaned from two other types of investigation: studies of lobbyism against IFRS adoption as a whole and, PRUHJHQHUDOO\VWXGLHVRIMXULVGLFWLRQV¶GHFLVLRQVWRPDQGDWHRUSHUPLW,)56LQVWHDGRI domestic GAAP. With regard to the first type, we are not aware of any systematic analysis of constituent lobbying against national or supranational (e.g., EU) plans to introduce IFRS. However, anecdotal evidence suggests that such lobbying does take place. For example, in July 2003, the French president Jacques Chirac sent an official OHWWHUWR(XURSHDQ&RPPLVVLRQSUHVLGHQW5RPDQR3URGLYRLFLQJ³FRQFHUQWKDWDGRSWLQJ ,)56SDUWLFXODUO\,$6ZRXOGQRWEHLQWKHEHVWLQWHUHVWRI(XURSH&KLUDF¶VLQWHUHVWLQ the debate arose at least in part because French banks were among the most critical of ,$6  &KLUDF¶V LQYROYHPHQW VKRZHG ,)56-related concern at the highest level of JRYHUQPHQW´$UPVWURQJ%DUWK-DJROLQ]HUDQG5LHGO  A second growing body of literature examines the determinants of country-level decisions to adopt IFRS. Hope, Jin, and Kang (2006) find that countries with a commitment to opening their capital markets to foreign investors but with relatively weak investor protection systems use IFRS adoption as a bonding mechanism to attract foreign FDSLWDO :LWKUHVSHFWWR FRXQWULHV¶ JRYHUQDQFH LQVWLWXWLRQV5DPDQQD DQG 6OHWWHQ   provide similar evidence. In addition, their results suggests that IFRS adoption at the country level is associatHG QHJDWLYHO\ ZLWK D FRXQWU\¶V SRZHU FRQVLVWHQW ZLWK PRUH powerful countries being less willing to surrender standard-setting authority to an international body) and positively with perceived network effects. Johnson (2009) uses 5DPDQQD DQG 6OHWWHQ¶V 9) sample of 102 non-EU countries to provide evidence suggesting that IFRS are adopted later in countries where more persons or groups are capable of blocking legislative change. These veto players are expected to slow down the process of IFRS adoption. Overall, the budding literature on country-level IFRS adoption LVEDVHGRQWKHDVVXPSWLRQWKDW³D sweeping change in accounting standards is likely to XSVHW WKH EDODQFH RI SRZHU EHWZHHQ WUDQVDFWLQJ SDUWLHV´ -RKQVRQ   ZKLFK LV 27

36

7KH,$6%¶VZRUNSODQFDQEHDFFHVVHGDWZZZLDVERUJ

consistent with our notion that mandatory IFRS adoption triggers both intended and unintended economic consequences. However, the high degree of aggregation of the data used in this research makes it difficult to pinpoint the exact nature of these consequences. 2.5.4 Timing transactions As mentioned above, the IAS Regulation was accepted and published in 2002 ± almost three years before coming into effect. This early decision allowed companies to prepare for the implementation of IFRS. However, the long implementation window also enabled managers to act opportunistically in anticipation of certain IFRS effects. Wang and Welker (2008) is the only study we are aware of that provides evidence of such opportunistic behavior. Their analysis is based on data from 14 EU countries as well as Australia and proceeds in two steps. First, they show that IFRS reconciliations during the transition year reveal price-sensitive information that triggers predictable stock market reactions. Specifically, stock markets react negatively if net income under IFRS is significantly lower than net income under local GAAP, and vice versa. In the second step of their analysis, Wang and Welker (2008) find that firms with larger IFRS-induced decreases in net income are more likely to issue equity and issue a larger volume of equity during the three years leading up to mandatory IFRS adoption. These results are consistent with the hypothesis that management possessed and exploited private information about the effects of mandatory IFRS adoption to decrease the costs of equity financing at the expense of outside investors. 2.6 Summary and suggestions for future research This paper reviews extant literature on the economic consequences of mandatory IFRS adoption in the EU. We define economic consequences as intended (unintended) if they can (cannot) be reconciled with the explicitly stated objectives of the IAS Regulation. According to this definition, intended consequences refer to the informational effects of mandatory IFRS adoption, while unintended consequences relate to its impact on contractual outcomes. Economic consequences also arise in anticipation of the introduction of IFRS. In this concluding section, we summarize the main findings of our literature review and provide suggestions for future research.28 28

We note that the mandatory IFRS adoption literature, while having undergone rapid development during the past five years, is still relatively immature. This leads to several caveats: First, many of the studies summarized here represent unpublished work still awaiting rigorous academic peer review. It is an open question to what extent the findings will stand the scrutiny of the publication process. 37

Intended consequences of mandatory IFRS adoption Our literature review on the intended consequences of mandatory IFRS adoption reveals that empirical research generally fails to document an increase in comparability or transparency of financial statements. In contrast, there is rich and almost unanimous evidence on positive reactions to mandatory IFRS adoption in capital markets and at the macroeconomic level. In the following, we present three areas for future research to address this apparent mismatch in extant empirical evidence. First, we know little about how mandatory IFRS adoption affects financial statements beyond the aggregate numbers that can be retrieved from commercial databases. Evidence on whether firms actually comply with the IFRS guidance is also limited. It is therefore still an open question whether financial statements are more informative following mandatory IFRS adoption. To address this important issue, we advocate more disclosure and compliance studies that rely on manually collected and thus finer data. Second, we encourage researchers to develop more convincing identification strategies when analyzing capital-market or macroeconomic effects of mandatory IFRS adoption. The challenge is to disentangle a potential IFRS effect from other concurrent changes which are (e.g., the implementation of stricter enforcement mechanisms) or are not (e.g., changes in the microstructure of capital markets) related to financial reporting. These different forces are difficult to separate in cross-country studies. Focusing on more specific settings (e.g., a single country or trading segment) may therefore be a useful starting point to better understand and control for simultaneous non-IFRS effects and thus increase the internal validity of results. Another strategy to deal with the identification problem is to confirm internal validity by analyzing whether the potential IFRS effect varies in the cross-section according to theory-based predictions. For example, future research could directly link potential capital-market effects of mandatory IFRS adoption to changes in reporting practices at Second, most of the papers discussed here rely on short time series to assess the economic consequences of mandatory IFRS adoption. The documented effects are likely to change as those involved in financial reporting move up the learning curve in terms of preparing, auditing, and using IFRS financial statements. Third, during the first few years under IFRS, accounting numbers are likely to be tainted by the effect of IFRS 1, First-time adoption of IFRSs. This standard, while in general requiring firms to apply IFRS retrospectively (i.e., as if they had always been doing so), provides for several mandatory exceptions and voluntary exemptions from this principle. As a result, the transition WR ,)56 UHSUHVHQWV D VWUXFWXUDO EUHDN LQ WKH WLPH VHULHV RI ILUPV¶ DFFRXQWLQJ QXPbers that will take several years to wash out. Finally, the financial crisis may have affected capital-market variables and, thus, studies using such data. 38

the firm level. Potential proxies for such changes in reporting practices include the difference between local GAAP and IFRS earnings as reported in the reconciliation statements of the first IFRS report or changes in firm-specific earnings management scores around IFRS adoption.29 The underlying assumption of this approach is that capital-market reactions stem from mandatory IFRS adoption having a material impact on financial statements. As a second approach, we suggest relating potential capitalmarket effects to proxies for firm-level benefits of mandatory IFRS adoption. Christensen, Lee, and Walker (2007), for example, estimate these benefits with a counterfactual proxy for a compaQ\¶VZLOOLQJQHVVWRDGRSW,)567KLVSUR[\LVEDVHGRQ firm characteristics of voluntary IFRS adopters in Germany and applied to companies in the UK where voluntary IFRS adoption was not allowed. Future research could apply a similar methodology to other settings to evaluate whether potential IFRS effects are indeed positively associated with its expected benefits. Alternatively, proxies for the benefits of IFRS adoption could be based on firm-specific stock market reactions to events that changed the likelihood of mandatory IFRS introduction in Europe (see section 2.5.1. for empirical studies that analyze these events). To be sure, expected crosssectional variation in the potential IFRS effect is only useful in evaluating internal validity if the underlying assumptions and measures are convincing. While the approaches advocated above are no panacea, we are confident that they mark an improvement over the strategies applied in most of the existing research (see section 2.3.3. for a discussion). Finally, we suggest addressing the potential sample bias in extant evidence on the capital-market effects of mandatory IFRS adoption. As our analysis of the current literature shows, this stream of research relies on commercial databases that suffer from a systematic bias towards large companies (see table 2.7.1). To the extent that large firms are more likely to benefit from IFRS adoption (see the literature on voluntary IFRS, e.g., Soderstrom and Sun, 2007), this database bias works in favor of finding positive capitalmarket reactions to mandatory IFRS adoption. We therefore encourage researchers to manually gather data on smaller companies and evaluate whether these experienced

29

Recent empirical studies started to use reconciliation data to evaluate cross-sectional differences in the capital-market effects of mandatory IFRS adoption (Beuselinck, Joos, Khurana, and van der Meulen, 2010; Horton, Serafeim and Serafeim, 2009). Changes in firm-specific earnings management scores, in contrast, have so far only been used in the context of voluntary IFRS adoption (Daske, Hail, Leuz, and Verdi, 2009). 39

systematically different capital-market reactions to mandatory IFRS adoption than their larger counterparts. Unintended consequences of mandatory IFRS adoption Insights from related settings suggest that mandatory IFRS adoption triggers material wealth transfers due to its impact on contractual outcomes. Our literature review on these unintended consequences distinguishes between individual contracts (management compensation plans and lending agreements) and collective contracts (dividend payouts, regulation, taxation, and employee benefits). While there is some evidence that individual contracts are affected by mandatory IFRS adoption (e.g., Chen and Tang, 2009; Christensen, Lee, and Walker, 2009), empirical literature related to collective contracts is confined to analyses of the impact of IFRS on employee benefits schemes (Rixon and Faseruk, 2009; for pension accounting issues, see e.g., Kiosse and Peasnell, 2009). Evidence on how the introduction of IFRS influences dividend payout policies, regulatory actions and review processes (e.g., in the banking industry), or taxable corporate income is missing entirely. Two challenges contribute to the evident gap in the literature on the unintended consequences of mandatory IFRS adoption. First, it is difficult to obtain information on relevant contractual arrangements. For example, publicly available data on lending agreements is very limited in the EU. Christensen, Lee, and Walker (2009) therefore use financial statement data to construct proxy variables (e.g., firm size, interest coverage) for the existence of covenants and the likelihood and costs of covenant violation. We suggest researchers to follow a similar strategy in case the data at hand do not allow for direct tests. The second challenge is that some unintended consequences of mandatory IFRS adoption may only apply to specific settings and, thus, are difficult to identify in a large-sample analysis. We encourage researchers to gain and exploit expert knowledge about such settings. While this strategy may not always lead to insights that extend easily to more general settings, it has the potential to provide important small-sample or case study evidence of high internal validity (e.g., de Jong, Rosellón, and Verwijmeren, 2006). Economic consequences in anticipation of mandatory IFRS adoption Since the IAS Regulation was accepted and published three years before it came into effect, firms and investors had plenty of time to conduct cost-benefit analyses in anticipation of mandatory IFRS adoption and react accordingly. Several empirical studies document evidence of significant stock market reactions to events that changed the 40

likelihood or scope of mandatory IFRS adoption. These reactions exhibit cross-sectional heterogeneity that is related to firm-level or country-level proxies for the net benefits of ,)566LQFHDQDO\VHVRILQYHVWRUV¶UHDFWLRQVLQDQWLFLSDWLRQRIPDQGDWRU\,)56DGRSWLRQ face the same challenges of low internal validity and biased samples as capital-market studies on the intended consequences, we refer to the above paragraph for suggestions for future research. Extant evidence and insights from related settings indicate that anticipatory actions also occurred at the company level. If firms perceive mandatory IFRS adoption as too costly, they have incentives to circumvent the impact of individual standards or avoid the adoption process entirely. Vulcheva (2009) shows that delisting probabilities in four EU markets increase abnormally in the year of IFRS introduction. This finding is consistent with firms avoiding IFRS adoption through a delisting. Future research could examine whether firms also engaged in other avoidance strategies, e.g., switching to unofficial trading segments (e.g., Open Market at Frankfurt Stock Exchange) where the IAS Regulation does not apply (see section 2.5.2. for details). More pro-active approaches to managing the impact of mandatory IFRS adoption include direct lobbying with the IASB and regulators or indirect lobbying through politicians. Evidence on lobbying efforts to influence or prevent the mandatory introduction of IFRS is still missing. We encourage researchers to fill this void in the literature through in-depth analysis of comment letters submitted to the IASB and by examining public statements by politicians in anticipation of the IAS Regulation. Such research is more likely to be relevant if it explicitly takes into account the specific political environment. In terms of two extremes: Where IFRS financial statements have tax effects and are strictly enforced, lobbying with the IASB can be expected to be more forceful than where such effects do not exist and enforcement is lax. Researchers should therefore be mindful of the most fruitful lobbying avenues in their settings, which are likely to include national regulators. Conclusions and Outlook This paper reviews the abundant literature on the economic consequences of mandatory IFRS adoption in the EU and provides suggestions for future research. While extant evidence provides useful insights, plenty of research opportunities on both the intended and the unintended consequences remain. We hope that this paper stimulates future research on these issues to establish a balanced view of the overall impact of mandatory IFRS adoption. 41

The impact of IFRS adoption is also of growing interest in an area that this paper has deliberately omitted, and in which the IAS Regulation passes regulatory power on to the (8 PHPEHU VWDWHV ILUP¶V VHSDUDWH OHJDO-entity financial statements. In many jurisdictions, there is an intense linkage of separate financial statements to contractual/legal consequences, e.g., in German taxation and dividend payout regulation. Existing contractual and regulatory equilibriums are therefore likely to be disturbed if MXULVGLFWLRQV SHUPLW RU UHTXLUH ,)56 WR EH DSSOLHG LQ ILUPV¶ VHSDUDWH financial statements.30 We hope that this paper provides a useful starting point for future research on this emerging topic.

30

42

For a discussion of IFRS implementation policies for unconsolidated financial statements in the EU, see e.g. Sellhorn and Gornik-Tomaszewski (2006).

2.7 Tables 2.7.1 Sample size comparison Country

Year of EU accession

CESR (2007)

Daske et al. (2008)

DeFond et al. (2009)

Horton et al. (2009)

Landsman et al. (2010)

Austria

1995

72

17

28

32

0

Belgium

1957

144

56

39

68

57

Denmark

1973

140

62

49

62

52

Finland

1995

135

99

82

88

61

France

1957

680

370

270

261

227

Germany

1957

768

216

232

266

153

Greece

1981

356

150

52

59

61

Ireland

1973

43

36

22

34

0

Italy

1957

288

79

159

118

135

Luxembourg

1957

35

13

0

7

0

Netherlands

1957

165

129

72

104

72

Portugal

1986

50

37

22

25

0

Spain

1986

190

91

51

80

68

Sweden

1995

350

168

162

128

87

United Kingdom

1973

953

418

378

530

492

Bulgaria

2007

369

0

0

0

0

Cyprus

2004

141

0

0

0

0

Czech Republic

2004

66

5

0

5

0

Estonia

2004

16

0

0

0

0

Hungary

2004

34

3

0

10

0

Latvia

2004

13

0

0

0

0

Lithuania

2004

43

0

0

0

0

Malta

2004

15

0

0

0

0

Poland

2004

197

55

0

21

0

Romania

2007

0

0

0

0

0

Slovakia

2004

0

0

0

0

0

Slovenia

2004

60

0

0

0

0

1957-1995

4,369

1,941

1,618

1,862

1,465

2004-2007

954

63

0

36

0

Total

1957-2007 5,323 2,004 1,618 1,898 1,465 Notes: This table compares the number of IFRS adopters in the 27 EU countries as identified by the Committee of European Securities Regulators (CESR) with the sample sizes of five academic studies. CESR (2007) reports the number of equity issuers that were admitted to trading on a regulated market and prepared a consolidated IFRS account in fiscal year 2005. These numbers are based on information provided by national enforcement institutions across Europe. Daske, Hail, Leuz, and Verdi (2008) consider all companies during fiscal years ± ³WKDW have sufficient financial data from Worldscope and price/volume data from Datastream to estimate « Model 3 for Zero 5HWXUQV´ (see their table 1). The sample of mandatory IFRS adopters in Daske, Hail, Leuz, and Verdi (2008) is confined to firms with fiscal year end in December 2005. The number of voluntary IFRS adopters analyzed by Daske, Hail, Leuz, and Verdi (2008) is not directly retrievable from their tables and therefore not considered here. DeFond, Hu, Hung, and Li (2009) analyze mandatory IFRS adopters that covered in the TFS international mutual fund database and have sufficient data in Compustat Global (stock performance, accounting standards followed) and IBES (analyst following). Countries with few public firms are excluded from their sample. Horton, Serafeim, and Serafeim (2009) select all voluntary and mandatory IFRS adopters covered in IBES that have data both before and after IFRS adoption (sample period: fiscal years 2003-2007). Landsman, Maydew and Thornock (2010) focus on mandatory IFRS adopters and require IBES earnings announcement dates as well as returns and volume data from Datastream to be available. They also impose size (Total Assets > US$ 100MM) and liquidity restrictions (proportion of zero return days < 80%) and delete countries with less than 200 firm-years. In this table, we present the composition of their sample for fiscal year 2005.

43

3 The impact of mandatory IFRS adoption on cross-border equity investments of individual investors 3.1 Introduction IFRS reporting is currently accepted in over 100 countries around the world. Regulators justify the move towards IFRS by the expectation that collective adoption of IFRS will enhance transparency and comparability of financial statements across countries and thus, among other benefits, one single accounting language will reinforce cross-border equity investments (e.g., EC, 2002). In this essay, we evaluate this claim by analyzing IFRS-related changes in cross-border equity investments of individual investors.31 Despite the focus on institutional investors in prior literature, individual investors play a vital role in financial markets. At the end of 2007, domestic individuals directly owned 14% of the market value of listed stocks in Europe (FESE, 2008a). In the United States, more than 20% of equity is held directly by individual investors (French, 2008). Anecdotal evidence suggests that individual investors are more likely to pursue long-term objectives than their institutional counterparts. Companies therefore make great efforts to attract individual investors, e.g. via corporate websites and investor relations departments (Vogelheim, Schoenbachler, Gordon, and Gordon, 2001). The relevance of individual investors is also recognized by regulators. For example, one of the explicitly stated SXUSRVHV RI WKH 6(& LV WR ³H[WHQG LQGLYLGXDO LQYHVWRU SURWHFWLRQ´ 6(& 2008). Mary 6KDSLUR WKH FXUUHQW 6(& FKDLU HPSKDVL]HV WKDW ³ZLWKRXW UXOHV WR Srotect [individual] investors, financial systems will not raise capital and thH HFRQRP\ ZLOO QRW JURZ´ (FINRA, 2008). A key challenge for any study of individual investors is that data on their investing and trading behavior is not publicly available. Prior studies have adopted several strategies to DGGUHVV WKLV SUREOHP LQFOXGLQJ DQDO\VLV RI ³VPDOO´ WUDGHV XVLQJ LQWUD-day transactions data (e.g., Lee, 1992), surveys of individual investor opinion (e.g., Elliott, Hodge, and Jackson, 2008) and examination of a proprietary dataset provided by a US online broker

31

We use the term individual investors to refer to non-institutional investors. Retail investors and private investors are synonymous expressions used in prior studies. It is interesting to note that some of the HDUOLHVW UHVHDUFK H[DPLQLQJ LQGLYLGXDO LQYHVWRUV¶ XVH DQG XQGHUVWDQGLQJ RI ILnancial statement information was provided by Sir David Tweedie, the current Chairman of the IASB (e.g., Lee and Tweedie, 1977). 45

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1_3, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

in the 1990s (e.g., Odean, 1998). None of these approaches are capable of providing large-sample evidence on the impact of IFRS adoption on individual investor decisions. Our proxies for cross-border equity investments by individual investors are based on trading activity in the Open Market at the Frankfurt Stock Exchange (FSE). The Open Market is an unofficial trading segment designed for German individual investors to trade foreign (i.e. non-German) stocks.32 This trading segment provides an ideal setting to study our research question for the following reasons. First, the Open Market comprises nearly one-third of all global stocks (excluding Germany). We are not aware of any other trading segment worldwide that offers such a diversity of foreign stocks. Descriptive statistics show that the Open Market sample favors more visible and transparent companies. To the extent that these characteristics complement high-quality adoption of IFRS, this increases the likelihood of finding an effect of IFRS adoption (e.g., Ball, 2006; Daske, Hail, Leuz, and Verdi, 2009; Christensen, Lee, and Walker, 2007). Second, the majority of listed companies in Germany adopted IFRS well before these accounting standards were introduced in other countries around the world. IFRS adoption by German companies was either voluntary (e.g., Leuz and Verrecchia, 2000) or due to exchange regulation of the former New Market (e.g., Leuz, 2003). In fact, much of the discussion on the benefits and costs of IFRS in the mid to late 1990s took place in Germany. This enabled German investors to familiarize themselves at an early stage with IFRS through their investments in domestic stocks.33 To the extent that lack of familiarity with IFRS creates costs for IFRS financial statement users, we would expect that these costs are lower in Germany. Therefore we expect that IFRS-related investment effects are more likely to be observable in the Open Market. We predict that IFRS adoption removes entry barriHUV WR LQYHVWPHQW LQ IRUHLJQ VWRFNV E\ UHSODFLQJ XQIDPLOLDU ORFDO *$$3¶V ZLWK familiar IFRS. Hence, IFRS adoption moves foreign stocks into the choice set of German investors and is expected to reinforce demand and ultimately trading activity in IFRSadopting stocks traded in the Open Market.

32

The only trading segment outside Germany that is remotely comparable to the Open Market is the Grey Market (or Other OTC) in the United States. However, Grey Market securities are not traded or TXRWHG RQ DQ H[FKDQJH RU LQWHUGHDOHU TXRWDWLRQ V\VWHP 6LQFH LQYHVWRU¶V ELGV DQG RIIHUV DUH QRW collected in a central spot, market transparency is very low. For more information, see (http://pinksheets.com/pink/otcguide/ investors_market_tiers.jsp).

33

Open Market investors engage in stock picking of foreign shares. Therefore they may be characterized as individual investors that exhibit greater financial literacy, and thus are more likely to actually utilize accounting information (e.g., Bailey, Kumar, and Ng, 2008).

46

Using a sample of 4,869 firms from 31 countries around the world, we find that stocks experience an economically and statistically significant increase in Open Market trading activity following the mandatory introduction of IFRS. For example, percentage trading volume in stocks of mandatory IFRS adopters increases by more than 20% relative to non-US control firms. Voluntary IFRS adopters also experience enhanced trading activity, but in contrast to mandatory IFRS adopters the effect is not always significant. If we include US firms in the benchmark, the IFRS effect on Open Market trading activity is even stronger, illustrating that US stocks lose trading volume in comparison to firms from IFRS adopting countries. These findings are robust to inclusion of a battery of control variables (e.g. media coverage in Germany based on Google News archive search results) and sensitivity analyses (e.g. a two-step approach to address potential sample selection issues). Further tests show that the trading reaction to mandatory IFRS adoption LQWKH2SHQ0DUNHWGRHVQRWGHSHQGRQDFRXQWU\¶VLQVWLWXWLRQDOHQYLURQPHQWFRQWUDU\WR the findings of most other studies on the economic consequences of IFRS adoption (e.g., Daske, Hail, Leuz, and Verdi, 2008). This finding, taken at face value, may point to a naïve reaction of individual investors towards IFRS adoption. However, we offer and present evidence for the alternative explanation that this result is attributable to the observation that lead brokers select stocks for trading in the Open Market when the ILUP¶V UHSRUWLQJ LQFHQWLYHV ILW ZLWK KLJK-quality IFRS adoption. Taken together, our analyses provide strong evidence consistent with the idea that IFRS as one global accounting language enhances cross-border equity investments by individual investors. Prior literature that examines the economic consequences of mandatory IFRS adoption focuses on aggregate capital market reactions (e.g., Daske, Hail, Leuz, and Verdi, 2008), macroeconomic effects (e.g., Beneish, Miller, and Yohn, 2010) or asset allocation decisions by institutional investors (e.g., Florou and Pope, 2009). Our study contributes to the literature in a number of ways. First, to the best of our knowledge, we provide the first analysis of individuDO LQYHVWRUV¶ UHDFWLRQ WR PDQGDWRU\ IFRS adoption. Thus, we provide evidence on how a single accounting language impacts individual investors, who are an explicit clientele of regulators such as the SEC. Second, we present evidence that Open Market lead brokers play a significant role in screening and filtering-out firms with low quality reporting incentives and opaque financial reporting. We therefore also add to the growing literature studying differential reactions to heterogeneity in reporting incentives and practice. Third, we introduce a novel and unique setting into the currently highly dynamic academic literature on individual investors. This setting allows us to 47

directly observe the aggregate trading activities of a large and homogeneous group of individual investors, i.e. German individual investors trading in foreign stocks. The remainder of the essay is organized as follows. Section 3.2 describes the Open Market. In section 3.3, we examine Open Market reactions to mandatory IFRS adoption. We provide concluding remarks in section 3.4. Section 3.5 comprises the tables and section 3.6 contains two appendices. 3.2 The Open Market 3.2.1 Institutional background The Open Market (³)UHLYHUNHKU´in German) is an unofficial trading segment at FSE. In contrast to official stock market segments in Europe (e.g. Prime and General Standard at FSE, Main Market at London Stock Exchange), the Open Market is not subject to regulations and directives of the EU, but is exclusively governed by stock exchange rules. It covers a variety of financial instruments such as stocks (both from Germany and abroad), bonds, certificates and warrants. The stocks segment is structured into the First Quotation Board and the Second Quotation Board. The First Quotation Board contains companies with a primary listing in the Open Market. 34 Companies whose stocks are already listed at another domestic or foreign trading venue (home market) are included in the Second Quotation Board. Since the Open Market is an unofficial trading segment, the EU regulation mandating IFRS is not applicable to companies in the First Quotation Board. In contrast, many companies in the Second Quotation Board are obliged to prepare their financial statements in accordance with IFRS due to regulation in the respective home markets. In this study, we focus on the foreign (i.e. non-German) stocks in the Second Quotation Board. For simplicity, we refer to this sub-segment as the Open Market.35 34

In 2005, FSE introduced the Entry Standard as a sub-segment of the First Quotation Board. Transparency requirements in the Entry Standard are higher than in the rest of the Open Market, but considerably lower than in official FSE stock market segments. While the Entry Standard is open to all companies, it is specifically targeted at small- and mid-caps that seek low cost access to the capital market. The Entry Standard is marketed as an alternative to the Alternative Investment Market at London Stock Exchange (e.g., Sudmeyer, Rückert, and Kuthe, 2005; Schlitt and Schäfer, 2006). At the end of 2008, 104 (11) German (foreign) companies were listed in the Entry Standard (FSE, 2008).

35

Official resources and the academic literature provide only little information on the Second Quotation Board of the Open Market. Much of the following description is based on the insights we gained from interviews with FSE staff and brokers. For more general information on the Open Market see, e.g., Müller-Michaels and Wecker, 2005; Harrer and Müller, 2006; or the website of FSE: www.deutscheboerse.com.

48

Established in 1987, the Open Market has become increasingly popular with German investors in recent years. At the end of 2000, a total of 4,544 foreign stocks were traded in the Open Market. This number doubled to 9,113 by the end of 2008. For comparison, the number of domestic stocks traded at FSE increased by only 16% from 905 to 1,054 during the same period (FSE, 2000; FSE, 2008). The remarkably high number of foreign stocks available for trading in the Open Market is a consequence of its unique set of rules.36 These rules permit eligible brokerage houses, i.e. those that are accredited for trading at FSE, to include securities in the Open Market on their own initiative (§2 (3) AGB). The stock issuing company need not be informed, nor need it approve inclusion of its securities in the Open Market. For the brokerage house, the inclusion process involves two basic requirements. First, it has to guarantee orderly fulfillment of transactions by acting as a lead broker (§14 (1) AGB). Second, it has to pay a non-recurring fee of 750 Euro (§35 AGB). Follow-up obligations of the lead broker are confined to informing the )6( DERXW HVVHQWLDO FRPSDQ\ QHZV FRQFHUQLQJ WKH LVVXHU WKDW FDQ EH DFTXLUHG ³E\ JHQHUDOO\DFFHVVLEOHLQIRUPDWLRQVRXUFHVLQDUHDVRQDEOHZD\´†$*% /HDGEURNHUV are authorized to exclude securities from the Open Market at any time with a notice SHULRGRIIRXUZHHNVRUZLWKRXWQRWLFH³XSRQJRRGFDXVH´† $*%  37 In summary, brokerage houses face very few constraints or institutional barriers relating to inclusion or exclusion of securities in the Open Market. Once a security has been included, the lead broker holds the exclusive right to set bid and ask quotes.38 Although officially non-binding (§79 FSE Exchange Rules), these quotes are de-facto tradable up to a size the lead broker specifies (Freihube, Kehr, Krahnen, and Theissen, 1999). When an investor places an order to trade on the bid (ask) quote, the lead broker buys (delivers) the agreed number of stocks. The resulting position would then be immediately closed with an offsetting order to entirely eliminate inventory risk. However, due to market illiquidity (see below for supporting evidence), perfect 36

We refer to this set of rules as AGB in the following. $*% VWDQGV IRU ÄAllgemeine Geschäftsbedingungen für den Freiverkehr an der Frankfurter Wertpapierbörse³*HQHUDO7HUPVDQG Conditions for the Regulated Unofficial Market).

37

Order books of Open Market securities can also be terminated by FSE (§33 (1) AGB). For example, in December 2005, FSE suspended trading in Turkish stocks until further notice because of unanswered questions about a planned tax on Turkish equities (Greil, 2005).

38

In case more than one party applies to be the lead broker for a particular stock, the allocation of the order book is decided by lot. Baader Bank AG, mwb fairtrade Wertpapierhandelsbank AG and Wolfgang Steubing AG Wertpapierdienstleister are the leading brokerage houses in the Open Market (Hiller von Gaertringen, 2006), but there are a number of other competitors. Detailed information on the allocation of Open Market order books is not publicly available. 49

offsetting may not be possible for Open Market stocks. The lead broker is then forced to carry out inventory rebalancing countertrades in another market, typically the home market where liquidity is generally much higher. Hence, in setting bid and ask quotes the lead broker faces a trade-off. On one hand, she has an incentive to offer low bid-ask spreads to generate trades and earn brokerage fees. On the other hand, there is the risk of losing out on trades if countertrades at the home markets are carried out at unfavorable prices. The resulting Open Market quotes are therefore likely to be determined by the home market bid-ask spread as a lower bound plus a premium that reflects the price risks faced by lead brokers in executing inventory rebalancing and due to currency risk exposure during trade execution. Other factors with a potential impact on the bid-ask spread premium in the Open Market include trading volume (i.e. the likelihood that the OHDG EURNHU LV DEOH WR PDWFK RIIVHWWLQJ RUGHUV  DQG FRPSHWLWLRQ WR WKH OHDG EURNHU¶V services. Competition may arise from other German exchanges, e.g. in Berlin, Stuttgart or Munich, where similar but much smaller trading segments exist. Within the FSE, trading volume can shift from floor trading where the lead broker operates to the fully electronic platform XETRA where quotes are automatically determined by an open limit order book. We analyze the Open Market competitors in more detail in section 3.3.5. Despite high bid-ask spreads, the Open Market provides a cost-efficient alternative to the home markets under certain circumstances. German retail banks and brokers pass on high, mostly fixed order fees when local clients choose to trade directly on a foreign exchange, whereas fixed charges for trading at FSE are considerably lower. Hence, for small trade sizes the higher bid-ask spreads (i.e. variable transaction costs that increase with trade size) are outweighed by lower order processing costs (i.e. mostly fixed transaction costs that are independent of trade size) in the Open Market. Appendix A provides an illustrative example on these links. The combination of low fixed and high variable fees in the Open Market is likely to be particularly attractive for German individual investors who trade small sizes of foreign stocks.39 Put differently, the Open

39

50

Here and in the following, we refer to Open Market investors as German individual investors for two reasons. First, the economics of the Open Market show that Open Market investors trade through German retail banks and brokers (see Appendix A). Due to institutional barriers most German retail banks and brokers require their clients to be domiciled in Germany. Second, German retail banks and brokers typically target their services at investors who speak German. For example, the web-portal of comdirect bank (www.comdirect.de), the leading online broker in Germany, provides information in German language only.

Market lead broker provides German individual investors with cheaper access to foreign stocks.40 3.2.2 Data and descriptive statistics In this section, we describe the quantitative characteristics of the Open Market. The analysis is based on two main samples: 1) a main sample defined as the Datastream (DS) Universe, and 2) the Open Market sample - a subset of the DS Universe. The DS Universe spans fiscal-years 2001 to 2007 and includes all firms covered by Datastream that meet the following requirements: At the country-level, we focus on firms domiciled in countries that either introduced IFRS in 2005 (the treatment group) or mandated domestic accounting standards throughout the sample period (the control group).41 At the firm-level, the DS Universe is restricted to companies that have their primary listing on the main exchange of their country of domicile (home markets)42 and for which sufficient data on trading volume, stock returns and required accounting data is available for both fiscal-years 2004 and 2005. We only include companies in the treatment group that switched from local GAAP to IFRS in 2005 (mandatory IFRS adopters) or before 2005 (voluntary IFRS adopters).43 The control group consists of companies that used domestic accounting standards throughout the sample period. The Open Market sample covers all firm-years within the DS Universe during which the respective stock could be traded on the FSE. We identify this sample using a proprietary dataset from FSE containing daily trading volume data (both in Euros and in number of shares traded) as well as the number of ticks for every stock traded in the 40

Note that institutional investors typically 1) have preferred and cheaper access to foreign markets through their lead brokerage houses and 2) trade in volumes well above the break-even where the foreign market turns into the more cost-efficient trading alternative.

41

We deliberately exclude Germany (our focus is on foreign, i.e. non-German, stocks), New Zealand (IFRS introduction in 2007), Singapore (IFRS introduction in 2003) and Turkey (IFRS introduction in 2006) from the DS Universe. Other countries such as China or Argentina are not included, because requirements on the firm-level are not met, usually because of missing information on accounting standards followed for firms listed on the main domestic exchange.

42

The main exchange is defined as the trading venue with the largest number of companies listed. We consider only one exchange per country except for the United States where firms from both New York Stock Exchange (NYSE) and NASDAQ are included. By focusing on the main exchange(s) in each country, we exclude companies listed at less regulated trading venues (such as the OTC Bulletin Board in the US) and thus ensure a minimum level of transparency among sample firms.

43

Thus, in order to obtain a clean sample, companies from the treatment group countries that did not switch to IFRS during the sample period (e.g. firms that need not prepare consolidated financial statements), did so after 2005 (e.g. firms listed in the Alternative Investment Market at London Stock Exchange) or applied US-GAAP (e.g. due to a cross-listing in the US) are not considered. 51

Open Market during the sample period.44 The FSE trading volume dataset spans the period January 2002 to June 2008. For consistency, we confine capital market data from Datastream (e.g. trading volume for the home markets) to the same period. Depending on the fiscal-year end and the resulting measurement period (-6 to +6 months relative to the fiscal-year end), coverage for fiscal-year 2001 and/or fiscal-year 2007 can therefore be incomplete for some firms. We only include firm-years into the DS Universe (Open Market sample) with at least 100 daily observations of home market trading volume (FSE trading volume) during the measurement period to ensure consistency across fiscal-years. Table 3.5.1 presents details on the composition of the DS Universe and the Open Market sample. Panel A focuses on the treatment group, i.e. countries that introduced IFRS in 2005. The DS Universe consists of 363 (5,121) voluntary (mandatory) IFRS adopters from 22 countries. The Open Market sample covers 55% (29%) of all voluntary (mandatory) IFRS adopters in the DS Universe. Panel B shows that the DS Universe comprises 4,198 (5,487) US (non-US) firms within the control group. 55% (16%) of these companies are part of the Open Market sample.45 Open Market coverage differs substantially across countries and firms. For example, while the majority of Austrian and US stocks are tradable in the Open Market, some countries (e.g., Poland, South Korea or Morocco) are not represented at all. At the firm-level, the considerable difference in coverage rates across accounting standards and IFRS adopter types gives a first indication that the lead brokers do not randomly choose the securities they offer in the Open Market. From panel C, we learn that the number of Open Market firms covered in the DS Universe increases over time, both in absolute as well as in relative terms. The total number of unique Open Market firms (Open Market share) climbs from 2,229 (21%) in fiscal-year 2001 to 4,060 (32%) in fiscal-year 2007. In total, the Open Market

44

Information on Open Market trading volume at FSE is also available via Datastream. However, using Datastream trading data is less reliable in this particular setting. First, Datastream does not allow identification whether a stock is traded in the Open Market or in an official FSE stock market segment (Prime Standard, General Standard). Second, Open Market coverage in Datastream is incomplete. Trading volume information is missing for nearly 20% of all firm-years in the Open Market sample. Finally, Datastream data is rounded, i.e. information on the number of stocks traded is reported in thousands rounding to one decimal place. This method inherently induces a measurement error which is more significant when trade sizes are small like in the Open Market. We compare daily Open Market trading volume data from Frankfurt Stock Exchange and Datastream for those stocks both databases cover. After taking the rounding errors in Datastream into account, both databases provide consistent numbers on 99.51% of all trading days.

45

US firms account for approximately 50% of the Open Market sample. Here and in the following, we therefore distinguish between US and non-US firms within the control group.

52

sample (DS Universe) comprises 4,869 (15,169) unique firms and 23,366 (94,797) firmyears.46 Table 3.5.2 shows descriptive statistics of various firm characteristics for the Open Market sample (panel A) as well as for the rest of the DS Universe (panel B). Panel A presents trading volume, number of trades, trade size and bid-ask spread statistics from FSE and the respective home markets for the same set of firm-years. Panel B is naturally confined to data from the home markets, because the covered sample (DS Universe excluding the Open Market sample) is not traded at FSE. In addition to liquidity measures, both panels show statistics on variables that are independent of the trading venue (other variables). Liquidity in stocks from the Open Market sample is quite low at FSE. During the average firm-year, trading occurs on slightly less than 25% of all trading days - during 2,307 firm-years (nearly 10% of the Open Market sample) there is no trading at all. On average, each stock is traded little more than twice per day, average daily trading volume is about 18,000 Euro. In contrast, daily trading volume in the home markets averages nearly 30 Million Euro. These massive liquidity differences across exchanges are hardly surprising. While trading at FSE is confined to a small subset of individual investors (i.e. German individual investors), institutional investors who account for most trading volume will prefer to trade in the respective home market.47 Average trade size at FSE is about 2,700 Euro which is well below the threshold typically used in the prior literature to distinguish between individual and institutional investors (e.g., Bhattacharya, 2004; Malmendier and Shantikumar, 2007). Comparison of bid-ask spreads across exchanges indicates that the variable fee the lead broker charges on Open Market transactions is in fact substantial: the median bid-ask spread is 3.17% at FSE compared to only 0.29% in the respective home markets. 48 Taken together, these 46

Differences in the size of the Open Market sample and the official numbers from the FSE Factbooks stem from the data requirements we impose on the DS Universe.

47

Untabulated statistics show that trading volume at FSE aggregated over the whole Open Market sample varies between 10 and 20 billion Euros per year. Hence, despite its relative lack of liquidity the Open Market offers substantial income opportunities for its participants. For example, with an average brokerage fee of 0.08% of the order volume (see Appendix A) Open Market lead brokers earn a total of 8 to 16 million Euros per year for their services.

48

For FSE, we use bid-ask spreads from the floor, because this is where most Open Market takes place (see section 3.3.5.1). We retrieve spread data from Datastream (CRSP) for non-US (US) exchanges. Datastream started coverage of bid and ask quotes for US exchanges in March 2006. Comprehensive comparison of daily spread data for the US shows that Datastream and CRSP provide the same information on 98.57% (62.82%) of all trading days for NASDAQ (NYSE) stocks. While Datastream/CRSP offer information on home market bid-ask spreads for large parts of the Open Market sample, coverage for FSE is limited. To ensure comparability, we demand that bid-ask spreads 53

descriptive statistics confirm that the Open Market at FSE is a trading segment that is specifically designed for individual investors. 3.2.3 Determinants of Open Market inclusion In this section we analyze the characteristics of stocks that are tradable in the Open Market. Stocks are tradable in the Open Market if they have been included by the lead EURNHU7KLVGHFLVLRQGHSHQGVRQDVWRFN¶VSRWHQWLDOWRJHQHUDWHVXIILFLHQW2SHQ0DUNHW trading volume and thus brokerage fees. Potential trading volume in the Open Market is XOWLPDWHO\GHWHUPLQHGE\LQGLYLGXDOLQYHVWRUV¶GHPDQGIRUDSDUWLFXODUVWRFN Detailed descriptions of all variables used in the analysis are provided in table 3.5.2 (firm-level variables) and table 3.5.3 (country-level variables).49 Table 3.5.4 presents results from probit regressions relating the likelihood of inclusion of a stock in the Open Market to various firm- and country-specific variables. When compared with stocks that are not tradable at FSE, firms in the Open Market feature higher volatility (Return Variability), higher trading volume in the home market (Home Trading Volume), more analyst activity (Up-/Downgrades), greater media coverage in Germany (Google Ratio)50 and higher market values (Market Value (Firm Level)). These results are consistent with the observation that individual investors are net-buyers of attention-grabbing stocks (Barber and Odean, 2008). At the country-level, Open Market companies are more likely to come from countries with a developed capital market (MCAP/GDP)51, domestic accounting standards that are similar to IFRS (Distance to IFRS)52, the Euro as the are available for both the home market and FSE. This reduces the number of firm-years with spread data to 13,589 (58.16% of the Open Market sample). 49

Note that most country- and firm-level variables are highly skewed. We transform highly skewed variables using natural logarithms to mitigate the influence of outliers.

50

Our proxy for media coverage is the number of search results in Google News archive. See Appendix B for details.

51

Measures of capital market development are also viewed as useful proxies for the quality of the enforcement system (Jackson, 2008). This view is based on the argument that capital markets cannot develop without a legal system that effectively protects investors. Alternative measures for capital market development (Turnover/GDP) or the strength of the enforcement system (Control of Corruption, Rule of Law, Regulatory Quality) are similarly significant determinants of Open Market inclusion.

52

Alternative measures for the (dis-)similarity between domestic accounting standards and IFRS (Absence and Divergence) yield similar results. Due to multicollinearity the coefficient estimate on the IFRS distance measure flips signs when all country-level variables are included. For example, the correlation between Distance to IFRS and EM Measure (Country Level) is 0.66 measured at the firmyear level.

54

national currency (Euro), a main exchange that is geographically close to FSE (Distance btw Exchanges)53 and economies that are closely linked to Germany (Imports + Exports). Finally, transparent reporting practices are significant determinants for Open Market inclusion, both at the country- and at the firm-level (EM Measure). To the extent that a strict enforcement system and generally transparent reporting practices are a prerequisite for high-quality application of IFRS (e.g., Daske, Hail, Leuz, and Verdi, 2008; Garcia Osma and Pope (2009)), our findings indicate that the lead brokers prefer serious over label IFRS adopters for Open Market inclusion. Taken together, our results provide strong evidence that the Open Market sample is a non-representative subset of the DS Universe. Specifically, the Open Market sample is significantly tilted towards more visible and transparent companies with reporting incentives that support high-quality adoption of IFRS. Thus, it seems that the lead broker acts as a gatekeeper to the Open Market, either proactively acting as a screening intermediary or explicitly responding to demand from individual investors. These findings have implications for our analysis of Open Market trading activities following mandatory IFRS adoption, because we are dealing with a selected sample. 3.3 Mandatory IFRS adoption and trading activities in the Open Market 3.3.1 Predictions In this section, we form predictions as to how mandatory IFRS adoption around the world affects trading activities in the Open Market. The previous section showed theoretical and empirical support for viewing the Open Market as a trading segment designed for German individual investors to trade foreign stocks. Our dataset described above does not allow us to directly observe if and how accounting information influences LQGLYLGXDOV¶WUDGLQJEHKDYLRULQWKH2SHQ0DUNHW+RZHYHUSULRUOLWHUDWXUHDQG RIILFLDO statistics provide useful insights. Official statistics show that similar to evidence from other countries the majority of German individuals do not actively invest in stocks and that those who do exhibit a considerable degree of home bias when selecting individual stocks. For example, in 2004 53

The results are unchanged if we use the distance between the respective capitals (Distance Berlin ± Capital) as a proxy for geographic proximity. The coefficient estimate on the geographic distance measure flips signs when all country-level variables are included. The reason for this change of sign again is multicollinearity. For example, the correlation between Log(Distance btw Exchanges) and Log(Imports + Exports) is -0.43 measured at the firm-year level. 55

the number of individual shareholders over 14 years in Germany amounted to only 4.6 Million or 6.5% of the whole population (DAI, 2008). Despite the well known benefits of international diversification, German individual shareholders invested a total of 145,495 Million Euro in domestic, but only 36,873 Million in foreign stocks (20.2% of all investments in stocks; Deutsche Bundesbank, 2005). These macro-level statistics suggest that the small subset of German individuals that actively trades individual foreign stocks possesses more financial literacy, on average, than those who focus on domestic equity or entirely refrain from stock picking. Consistent with this observation, recent research shows that individual investors are more likely to have internationally diversified portfolios if they are highly educated, wealthy and/or have more trading experience (e.g., Abreu, Mendes, and Santos, 2009; Bailey, Kumar, and Ng, 2008; Graham, Harvey, and Huang, 2009). German individual investors are also more likely to trade in foreign stocks if they consult financial advisors (e.g., Bluethgen, Gintschel, Hackethal, and Müller, 2008; Gerhardt and Hackethal, 2009). Comprehensive survey evidence by Ernst, Gassen, and Pellens (2009) reveals that German individual investors use business media and financial statements as central information sources. While usage of accounting information varies considerably, more experienced investors tend to analyze financial reports more carefully. We argue that it is likely that Open Market investors engaging in active picking of individual foreign stocks exhibit greater financial literacy, and therefore have a higher likelihood of actually utilizing financial statement information in their investment decisions as compared to the average individual investor in the economy. Taken together, these insights suggest that accounting information affects Open Market trading through two non-mutually exclusive channels: either Open Market investors utilize financial statements by themselves and/or they consult other information sources such as the business media or financial advisors that in turn rely on information disclosed in financial statements. Individual investors are presumably most familiar with their local GAAP. In Germany, however, the majority of listed companies voluntarily switched from local GAAP to IFRS or US-GAAP around the turn of the millennium (e.g., Leuz and Verrecchia, 2000; Daske, 2006).54 In addition, exchange regulation required companies listed at the New 54

56

We confirm this observation by analyzing the accounting standard strategy of all German firms that match the firm-level sample selection criteria outlined in the previous section. We find 322 (177) voluntary (mandatory) IFRS adopters and 122 firms that used US-GAAP at some point before fiscalyear 2005. In fiscal-year 2001, already around 60 percent of all German firms in our sample prepared financial statements according to IFRS or US-GAAP.

Market, a former IPO platform at FSE for high-tech growth firms, to report under either IFRS or US-GAAP (e.g., Leuz, 2003). These combined developments enabled German investors to become familiar with IFRS and US-GAAP through their investments in domestic stocks well before mandatory IFRS adoption around the world took off. We argue that German individual investors are more likely to invest in foreign firms that apply accounting standards with which they are familiar. This argument is supported by the observation that voluntary IFRS adopters as well as US-GAAP users (i.e. US stocks) are overrepresented in the Open Market sample (see table 3.5.1). Conversely, unfamiliar accounting standards impose prohibitive (true or perceived) information processing costs on individual investors and thus deter them, among other reasons, from trading in foreign stocks. We predict that by replacing unfamiliar loFDO *$$3¶V ZLWK IDPLOLDU ,)56 the mandatory switch to IFRS removes the entry barrier to foreign stocks that is caused by cross-country accounting diversity (see also Beneish and Yohn, 2008, for a conceptual discussion). Hence, stocks of IFRS adopting companies are pushed into the choice set of Open Market investors, triggering more demand and ultimately more trading activity in these stocks. More trading in IFRS stocks may affect both the supply as well as the demand side in the Open Market. On the supply side, the lead broker is more (less) likely to include (exclude) foreign IFRS stocks after mandatory IFRS introduction, either because she anticipates the increased demand or Open Market investors explicitly express their interest in such stocks. On the demand side, Open Market trading in stocks that had been included pre-IFRS is likely to increase after the adoption of IFRS, because the lower (true or perceived) information gathering costs of individual investors decrease after IFRS adoption. The inclusion decision involves low immediate costs and little follow-up obligations for the lead broker. In contrast, the wealth of individual investors is substantially influenced by their stock trading decisions. We therefore believe that the demand side offers a more powerful test of the impact of mandatory IFRS adoption around the world on Open Market trading. Stated in alternative form, our first two predictions are as follows: P1:

Mandatory IFRS adoption enhances the likelihood of Open Market inclusion for foreign IFRS stocks.

P2:

Mandatory IFRS adoption enhances Open Market trading activity in foreign IFRS stocks that had been included pre-IFRS. 57

These predictions imply Open Market reactions regardless of how mandatory IFRS adoption affects actual reporting practices. However, there are reasons to expect a more differentiated response. Prior literature provides strong theoretical arguments (e.g., Ball, 2006; Hail, Leuz, and Wysocki, 2010) as well as empirical evidence that mandating IFRS changes very little in reporting practices in case of adverse reporting incentives (e.g., Christensen, Lee, and Walker, 2008; Garcia Osma and Pope, 2009). Thus, if reporting practices remain unchanged following a switch from local GAAP to IFRS, it is questionable whether true information processing costs of investors actually decrease. In fact, the information processing costs might even increase if discretion in reporting standards and lack of enforcement make it difficult for investors to figure out the extent to which firms are serious about their IFRS adoption. Assuming full financial literacy and rationality of individual investors, we would therefore expect a heterogeneous Open Market reaction where trading increases only in those stocks where IFRS fits with a ILUP¶V LQVWLWXtional environment and can really be expected to decrease the true information processing costs of such investors (Hail, Leuz, and Wysocki, 2010). However, it could well be the case that there are limits to the extent to which individual investors have the time or the ability to understand and unravel the details of the accounting regulation and practice (e.g., Bartlett and Chandler, 1997). Given the vast OLWHUDWXUH RQ LQGLYLGXDO LQYHVWRUV¶ QDLYHW\ HJ, Bhattacharya, 2004), and their usage of simple heuristics in decision making (e.g., De Bondt, 1998), it is also conceivable that individual investors simply perceive information processing costs to be lower following IFRS adoption, even if reporting practices might not have changed more than in providing a common format and familiar style of the annual report. If so, a homogeneous Open Market response as implied by our second prediction is to be expected. In addition, recall that lead brokers mainly consider stocks of visible and transparent companies for Open Market inclusion (see section 3.2.3). If Open Market lead brokers do not broadly add stocks from IFRS mandating countries, but go one step further and add only stocks from those firms which are likely to seriously adopt IFRS, and therefore really increase their transparency, the Open Market universe does not exhibit the same cross-sectional variation as the overall universe of stocks. Again, a homogeneous Open Market response as implied by our first two predictions is to be expected. Given that the above mentioned conflicting arguments both have merit, we test the following hypothesis, but leave the prediction open: 58

P3: The effect of mandatory IFRS adoption on Open Market trading activity is homogeneous in reporting incentives. In the next sections, we successively test our three predictions with empirical data. 3.3.2 Open Market inclusion analysis 3.3.2.1 Research design In this section, we test our first prediction, i.e. we investigate whether mandatory IFRS adoption enhances the likelihood of a stock being included in the Open Market.55 The dependent variable Inclusion is a binary variable that equals one (zero) for all firmyears in the Open Market sample (the rest of the DS Universe). The key independent variable is Post-IFRS, a dummy variable that takes on value one (zero) for fiscal-years 2005-2007 (2001-2004). To test the impact of mandatory IFRS adoption, we define nonUS control firms as our benchmark group and interact Post-IFRS with binary variables that indicate voluntary IFRS adopters (Voluntary), mandatory IFRS adopters (Mandatory) and US firms (USA), respectively. These interaction terms capture the average effect of mandatory IFRS adoption on the likelihood of Open Market inclusion for the respective group of companies relative to the benchmark (i.e. non-US control firms). Combining these variables results in the following basic regression specification: Inclusion = ȕ0 ȕ1 Post-IFRS ȕ2 Post-IFRS*Voluntary ȕ3 Post-IFRS*Mandatory + ȕ4 Post-IFRS*USA Ȉȕj Controlsj İ (3.1) where Controlsj denotes the set of control variables. The set of control variables includes non-IFRS determinants of Open Market inclusion as used in section 3.2.3. We estimate regression specification (1) at the firm-year level using a fixed effects linear probability model, a pooled probit model and a random effects probit model. The fixed effects linear probability model has the desirable property that potential biases in coefficient estimates caused by unobserved firm characteristics are mitigated, because firm effects are estimated explicitly. One of the downsides of this approach is that it imposes strong restrictions on the estimated firm effects that may result in inconsistent estimates of the other coefficients. Moreover, the predicted values for the dependent 55

Untabulated statistics show that a total of 2,027 (69) firms have been included into (excluded from) the Open Market sample during fiscal-years 2002 to 2007. Due to the infrequency of Open Market exclusions and for simplicity, we refer to Open Market inclusions here and in the following when, more precisely, in- and exclusions are analyzed. 59

variable may lie outside the interval [0, 1]. In contrast, predicted values from probit models are bounded between 0 and 1 by construction. However, probit models do not allow controlling for unobserved firm effects because of the incidental parameters problem. The difference between the pooled and the random effects probit model is that the latter makes an explicit assumption about the correlation structure between unobserved firm effects and covariates, while the former is silent on this issue.56 Since each approach has its merits, we report estimates from all three models using different combinations of control variables. 3.3.2.2 Empirical findings Table 3.5.5 presents regression results. Consistent with the descriptive evidence in table 3.5.1, panel C, we find that the likelihood of Open Market inclusion is significantly higher after mandatory IFRS adoption across all groups of companies. The coefficient estimates on the interaction terms indicate that the post-IFRS increase is strongest for voluntary IFRS adopters followed by US firms, mandatory IFRS adopters and non-US control firms. This order closely resembles the order of DS Universe shares across the four groups at the beginning of the sample period (see table 3.5.1, panel C) suggesting that mandatory IFRS adoption contributes at best modestly in explaining the likelihood of Open Market inclusion over and above the (mostly cross-sectional) determinants identified in section 3.2.3. Our results provide mild support for our prediction that mandatory IFRS adoption enhances the likelihood of Open Market inclusion for foreign IFRS stocks. The results also suggest that there are forces other than IFRS that we cannot control for and that drive the likelihood of Open Market inclusion over time. High competition among lead brokers and the low costs of Open Market inclusion create an environment that supports the dramatic growth of the Open Market. This environment makes it complicated to model the time-series properties of Open Market inclusion. In contrast, the coefficient estimates on the control variables confirm that the cross-sectional determinants of Open Market inclusion as identified in section 3.2.3 remain significant factors.

56

60

For a more detailed discussion on how to deal with unobserved firm characteristics in a panel when the dependent variable is binary, see e.g. Nini, Smith, and Sufi, 2009.

3.3.3 Open Market trading volume analysis 3.3.3.1 Research design In this section, we test our second prediction. Specifically, we analyze whether stocks that had been included into the Open Market pre-IFRS experience an increase in trading activity following the introduction of IFRS. The basic regression specification is the same as in the Open Market inclusion analysis (see section 3.3.2.1): Trading = ȕ0 ȕ1 Post-IFRS ȕ2 Post-IFRS*Voluntary ȕ3 Post-IFRS*Mandatory + ȕ4 Post-IFRS*USA Ȉȕj Controlsj İ (3.2) where Trading denotes a measure of Open Market trading activity and the other variables are defined as above. We use percentage trading volume (FSE Trading Volume (%)) and the proportion of non-zero trading volume days (FSE Trading Days (%)) as measures for trading activity in the Open Market.57 Following related IFRS literature (e.g. Daske, Hail, Leuz, and Verdi, 2008; Florou and Pope, 2009) we estimate regression specification (2) using a firm fixed effects model that controls for time trends.58 The goal of this differences-in-differences approach is to identify a causal relationship between a treatment (IFRS introduction) and an endogenous YDULDEOH2SHQ0DUNHWWUDGLQJDFWLYLW\ E\FRPSDULQJWKHWUHDWPHQW¶VLPSDFWRQDIIHFWHG firms (treatment group) with its impact on unaffected firms (control group). To ensure that OLS estimation produces consistent standard errors we use standard errors clustered by firm as suggested by Bertrandt, Duflo, and Mullainathan (2004). Since we estimate a firm fixed effects model, our set of control variables is confined to variables that capture firm-specific changes over time. Similar to the inclusion analysis we predict that, ceteris paribus, Open Market trading activity increases in stocks that experience an increase in volatility (our control variable: Return Variability), home market trading volume (Home Trading Volume), media coverage in Germany (Google Ratio) and/or analyst activity

57

Since FSE Trading Volume is highly skewed (see table 3.5.2), we use the natural logarithm to mitigate the influence of outliers. To preclude computing the natural logarithm of zero we add a small constant (0.000001%) to the raw values. The results are very similar if we use different values for the constant or if we eliminate firm-years with no trading volume in the Open Market.

58

This approach is similar to the fixed effects linear probability model in the Open Market inclusion analysis. The difference is that in the trading volume analysis the dependent variables are continuous instead of binary. Continuous variables on the left-hand side allow for consistent estimates under much weaker assumptions (e.g., Nini, Smith, and Sufi, 2009). 61

(Up//Downgrades). Again, we use the natural logarithm of the control variables if the raw values are highly skewed. Note that in our model, the estimated effect of mandatory IFRS adoption is exclusively determined by firms that are part of the Open Market sample both before and after IFRS introduction. Hence, stocks that are added to the Open Market after IFRS introduction (see the inclusion analysis) do not directly influence the results of the trading volume analysis. 3.3.3.2 Empirical findings In table 3.5.6, we present regression results of the Open Market trading volume analysis. Panel A (Panel B) reports results with percentage trading volume (the proportion of non-zero trading volume days) as dependent variable. The coefficient estimates on Post-IFRS show a significant increase in Open Market trading activity for the benchmark group (i.e. non-US control firms) following the introduction of IFRS. This increase is even more pronounced for mandatory IFRS adopters. Depending on the specification, the coefficient estimates on Post-IFRS*Mandatory suggest an increase in percentage trading volume (in the proportion of non-zero trading volume days) of 20% to 39% (by 3 to 4 percentage points) relative to non-US control firms. Voluntary IFRS adopters also benefit from enhanced trading activity, but in contrast to mandatory IFRS adopters the effect is not always significant. 59 On the other hand, post-IFRS Open Market trading in US firms declines sharply relative to non-US control firms. Hence, if we include US firms in the benchmark, the IFRS effect on Open Market trading activity is even stronger. The results suggest that individual portfolio investments shifted from both outside and within the Open Market towards IFRS stocks. Within the Open Market, shifts from US stocks seem to clearly dominate those from non-US control stocks.60 Taken

59

In contrast to e.g. Daske, Hail, Leuz, and Verdi (2008), we do not estimate the immediate impact of voluntary IFRS adoption on the dependent variable, because the Open Market sample includes a mere 29 companies that 1) switched from local GAAP to IFRS between 2002 and 2004 and 2) were tradable in the Open Market both before and after voluntary adoption. Note that the effect of voluntary adoption is captured by the firm fixed effects if these two requirements are not fulfilled. In unreported analysis, we follow the strategy by Daske, Hail, Leuz, and Verdi (2008) and find a mildly significant increase in Open Market trading activity after voluntary IFRS adoption. However, the main findings with regard to the impact after global IFRS adoption remain unchanged.

60

One explanation for this sharp difference may lie in the fact that Open Market investors are familiar with US-GAAP from their domestic investments, while trading in non-US control stocks requires them to familiarize with the respective foreign GAAP (see section 3.3.1). Hence, Open Market investors that trade US stocks may be more likely to switch to stocks from IFRS adopters, because

62

together, the regression results provide compelling evidence consistent with our prediction that mandatory IFRS adoption enhances Open Market trading in IFRS stocks. We acknowledge that there may be forces other than IFRS that drive our strong results. For example, mandatory IFRS adoption may have led financial advisors and/or the business press to promote IFRS stocks. However, we try to control for such effects by including proxies for analyst activity (Up-/Downgrades) and media coverage in Germany (Google Ratio). The coefficient estimates on these control variables have the predicted sign, are statistically significant throughout all specifications, but affect the IFRS coefficients of primary interest only mildly. The coefficient estimates on the remaining two controls (Return Variability and Home Trading Volume) are also consistent with expectations but again do not affect the coefficients of interest, supporting the argument that the IFRS effect in the Open Market is not simply due to a general increase in the visibility of IFRS stocks. The increase in Open Market trading activity may also be triggered by a concurrent decrease in order fees that was confined to IFRS stocks. In unreported analyses, we address this issue by including the bid-ask spread difference between FSE and the respective home market as additional control. Similar to the other control variables, the coefficient estimate on the spread difference is highly significant, but has little impact on the main findings. We perform a variety of unreported sensitivity analyses. First, we include fiscal-year dummies to model the time trend of the benchmark group (i.e. non-US control stocks) more precisely than with the Post-IFRS dummy. The results are very similar. Second, we examine the persistence of the IFRS effect by running the same regressions without the observations from the transitional period (fiscal-year 2005). The results turn out to be slightly more pronounced for mandatory IFRS adopters, indicating that the impact of IFRS on Open Market trading activity is not merely a temporary phenomenon. Third, we apply a supplemental procedure suggested by Bertrandt, Duflo, and Mullainathan (2004) to correct for serial correlation in the panel. Specifically, we ignore all time-series effects by averaging data across two periods: pre-IFRS (fiscal-year 2001 to 2004) and post-IFRS (2005 to 2007). We then run the OLS estimation on the resulting two-period panel dataset with robust standard errors. The results remain largely unchanged. Finally, we address potential sample selection issues. While firm fixed effects render the difference-indifferences approach robust to time-invariant sample selection bias, systematic they have incurred fewer costs in understanding their previous investments and thus are more attracted by low-cost alternatives. 63

differences in the likelihood of Open Market inclusion across time may induce a sample selection bias that is time-variant. To correct for such a bias we apply a two-step approach based on the methodology introduced by Heckman (1979) and further developed by Wooldridge (1995). In the first step, we estimate a cross-sectional probit model to explain the likelihood of Open Market inclusion separately for each fiscal-year and compute the value of the Inverse Mills Ratio (IMR). In the second step, we apply the difference-in-differences approach with IMR as additional control. The coefficient estimate on IMR turns out to be insignificant. The main findings hardly change. In summary, our results are robust to a variety of different specifications. 3.3.4 Cross-sectional analysis of the Open Market reaction 3.3.4.1 Research design In this section, we examine cross-sectional variation in the Open Market reaction around mandatory IFRS introduction. The analysis aims at understanding whether Open Market reaction to mandatory IFRS adoption is conditional on its impact on reporting practices. Similar to Daske, Hail, Leuz, and Verdi (2008) we interact the key independent variables in regression specification (2) with a binary variable Conditional that partitions the treatment sample. The idea is to identify those companies that are more likely to change their reporting practices following IFRS adoption than the rest of the treatment sample in order to test whether the Open Market reaction is systematically different between both groups. This approach translates into the following basic regression specification: Trading = ȕ0 ȕ1 Post-IFRS + ȕ2 Post-IFRS*Voluntary ȕ3 PostIFRS*Voluntary*Conditional + ȕ4 Post-IFRS*Mandatory ȕ5 Post-IFRS*Mandatory*Conditional ȕ6 Post-IFRS*USA Ȉȕj Controlsj İ

(3.3)

We use the following country-level factors for the variable Conditional: (i) Rule of Law, (ii) MCAP/GDP, (iii) Distance to IFRS, (iv) EM Measure (Country Level) and (v) LNW EM Aggregate. Rule of Law is the Worldwide Governance Indicator (WGI) for the year 2005 provided by World Bank (see Kaufmann, Kraay, and Mastruzzi (2009) for details). Higher values represent countries with a stricter enforcement regime. MCAP/GDP is the ratio of a FRXQWU\¶VVWRFNmarket capitalization to its Gross Domestic Product (GDP). We use this 64

measure of capital market development as an additional proxy for the quality of enforcement system following the suggestion by Jackson (2008). The underlying argument is that capital markets cannot develop without a legal system that effectively protects investors. Distance to IFRS is the Bae, Tan, and Welker (2008) summary score RIKRZWKHFRXQWU\¶VORFDO*$$3GLIIHUVIURP ,)56RQ NH\DFFRXQWLQJGLPHQVLRQV Higher values stand for more discrepancies between local GAAP and IFRS. We presume that the likelihood of IFRS induced changes in reporting practices is increasing with Distance to IFRS. LNW EM Aggregate is the aggregate earnings management score as taken from Leuz, Nanda, and Wysocki (2003), EM Measure (Country Level) is a selfconstructed earnings management score based on the same methodology. The advantage of LNW EM Aggregate is that its measurement period is confined to the pre-IFRS era (1990-1999) and therefore does not capture any potential IFRS effects. EM Measure (Country Level) in contrast is based on accounting numbers from both the pre- and the post-IFRS period. However, it has the benefit of being more up-to-date and more comprehensive in terms of countries covered. We assume that IFRS is more likely to improve reporting practices in countries where the accounting system is more transparent, i.e. if the aggregate earnings management scores are low. If the relevant variable is continuous we transform it into binary variable splitting it at the median computed over the treatment sample countries. 3.3.4.2 Empirical findings Table 3.5.7 presents regression results. For brevity, we report only one regression specification per conditional variable, namely the regression model with FSE Trading Volume as dependent variable and all controls included (equivalent to model (6) in table 3.5.6, panel A). We discuss other specifications if the tenor of the results changes. The regression results provide strong evidence that the Open Market reaction to mandatory IFRS adoption is homogeneous in reporting incentives across countries. The coefficient estimates on the interacted conditional variables are insignificant throughout all models with the exception of Post-IFRS*Mandatory*Distance to IFRS. However, even the Distance to IFRS measure yields insignificant results if home market trading volume is not included as a control variable. Unreported sensitivity analyses show that the results do not change if we use Absence and/or Divergence as alternative proxies for the distance between local GAAP and IFRS. The results also remain unchanged if we distinguish between countries that have adopted the Euro as their national currency and 65

countries that have not. Taken together, the results are consistent with the prediction of a homogeneous Open Market reaction. 3.3.5 Auxiliary analyses 3.3.5.1 Frankfurt floor versus XETRA There are two trading platforms at FSE that work in parallel: 1) the floor where the lead brokers operate and 2) the fully electronic XETRA where quotes are automatically determined by an open limit order book. Order processing costs are lower in XETRA, but its anonymity induces higher costs arising from adverse selection. Since the adverse selection component becomes more important when trading volume is low, the floor is more attractive for less liquid stocks (Theissen, 2002). Consistent with this evidence, we find that trading activity in the illiquid Open Market usually takes place in the floor, but shifts to XETRA if liquidity is high. Specifically, trading activity in the floor is higher in 97.79% of all firm-years in the Open Market sample. In contrast, the difference in mean trading volume (Trading Volume (Euro)) is less pronounced between both systems (floor: 11,379 Euro, XETRA: 6,714 Euro). For our main analyses, we use trading volume from both the floor and from XETRA. We perform the same set of tests using only trading volume from the floor. The results remain largely unchanged. 3.3.5.2 Other German exchanges Trading segments similar to the Open Market exist at the regional German exchanges in Berlin, Stuttgart and Munich. Datastream data indicates that these segments are smaller and less liquid than the Open Market in Frankfurt. Applying the selection criteria described in section 3.3.2, we identify a sample of 8,019 firm-years (8% of the DS Universe) for Berlin, 1,576 firm-years (2%) for Stuttgart and 249 firm-years (0%) for Munich Stock Exchange. Hence, the Open Market sample (23,366 firm-years) is more than twice as big as the combined samples of the other three German exchanges. The mean proportion of non-zero trading volume days is lower than 5% in Berlin, Stuttgart and Munich. We conclude that trading is a rare phenomenon at these exchanges. To address potential quality issues with Datastream data (see section 3.3.2), we analyze a further proprietary dataset from FSE that contains monthly trading volume in all Open

66

Market stocks separately for FSE and, if applicable, for other German exchanges.61 Descriptive statistics show that FSE combines more than 80% of the aggregate trading volume in these stocks. Taken together, our results suggest that the Open Market lead broker at FSE faces little competition from other German exchanges. We perform the same tests as in the main IFRS analyses using trading volume at other German exchanges instead of FSE. The results are inconsistent with our main findings when using Datastream data. This result is hardly surprising given the small sample sizes, extreme illiquidity, and thus low power of those tests. In contrast, the results are very similar to our main findings if we use monthly trading volume from the proprietary FSE dataset. Hence, there is evidence of IFRS induced trading activity at the regional German exchanges in Berlin, Stuttgart and Munich. 3.4 Conclusions This study examines the impact of mandatory IFRS adoption on trading activities in the Open Market. The Open Market is a trading segment at FSE designed for German individual investors to trade foreign (i.e. non-German) stocks. We find strong evidence that stocks experience an increase in Open Market trading activity following mandatory adoption of IFRS. This finding is consistent with the idea that collective IFRS adoption as one global accounting language reinforces cross-border investments by individual investors who might use financial statement information. As such, our results support the efforts by the IASB and standards setters around the world to foster a single global set of financial reporting standards. Despite the consistency of our results, we urge caution in interpreting this study. First, since our dataset does not allow us to directly observe if and how accounting information influences individual investors decision making and trading in the Open Market, we cannot rule out the possibility that our results may reflect other coincident institutional changes rather than the impact of mandatory IFRS adoption per se. However, given our extensive efforts to control for other determinants of trading activity, and the strength and robustness of our results, we are confident that the Open Market reaction at least partly echoes a positive response to mandatory IFRS adoption. Second, although increased trading activities in foreign stocks will, ceteris paribus, translate into a reduction in the 61

Note that the FSE dataset is confined to Open Market stocks, i.e. stocks that are tradable at FSE. However, analysis of Datastream data shows that some stocks are tradable in Berlin, Stuttgart and/or Munich that are not included in the Open Market. 67

equity home bias, our data does not allow for clean testing of the impact of mandatory IFRS adoption on the extent of investors equity home bias. In fact, such an analysis UHTXLUHV DFFHVV WR LQGLYLGXDO LQYHVWRU¶V SRUWIROLR level holdings over time. We consider such an analysis to be a fruitful area of future research. Third, we recognize that individual Open Market investors taking active positions in individual foreign stocks provide a high-power setting for our tests, but they are not necessarily representative of WKH XQLYHUVH RI LQGLYLGXDO LQYHVWRUV LQ WKH JOREDO HFRQRP\ +RZHYHU WKH ,$6%¶V DQG RWKHU VWDQGDUG VHWWHUV¶ HIIRUWV DUH QDWXUDOO\ WDUJHWHG WRZDUGV WKRVH LQYHVWRUV ZKR XWLOL]H financial statement information in their investment decisions. Individual investors engaging in cross-border investments are likely to be an important subset of this group.

68

3.5 Tables 3.5.1 Sample composition by country and year Panel A: Treatment Group IFRS Voluntary

IFRS Mandatory

Unique Firms IFRS Adoption Countries

Open Market

Australia

DS Universe

Local GAAP

Unique Firms Share

Open Market 279

DS Universe

US-GAAP

Unique Firms Share

Open Market

1,268

22%

-

DS Universe

Total

Unique Firms DS Universe

Firm-Years

Share

Open Market

Share

Open Market

DS Universe

-

-

-

-

-

1,080

7,786

Share

5

15

33%

Austria

37

44

84%

3

11

27%

-

-

-

-

-

-

188

344

Belgium

13

24

54%

21

76

28%

-

-

-

-

-

-

129

629

21%

Czech Republic

5

7

71%

4

9

44%

-

-

-

-

-

-

32

95

34%

Denmark

9

16

56%

19

110

17%

-

-

-

-

-

-

136

829

16%

Finland

8

10

80%

34

117

29%

-

-

-

-

-

-

241

844

29%

France

3

12

25%

156

517

30%

-

-

-

-

-

-

821

3,407

24%

Greece

3

9

33%

26

243

11%

-

-

-

-

-

-

161

1,519

11%

Hong Kong

6

16

38%

277

825

34%

-

-

-

-

-

-

1,195

5,271

23%

Hungary

15

16

94%

3

6

50%

-

-

-

-

-

-

68

136

50%

Ireland

0

0

-

19

36

53%

-

-

-

-

-

-

68

226

30%

Italy

2

3

67%

84

225

37%

-

-

-

-

-

-

470

1,489

32%

Luxembourg

0

2

0%

1

10

10%

-

-

-

-

-

-

1

74

1%

Netherlands

4

6

67%

46

111

41%

-

-

-

-

-

-

193

761

25%

Norway

3

3

100%

33

130

25%

-

-

-

-

-

-

143

820

17%

Poland

0

8

0%

0

59

0%

-

-

-

-

-

-

-

422

0%

Portugal

3

5

60%

12

40

30%

-

-

-

-

-

-

89

276

32%

South Africa

14% 55%

11

25

44%

35

236

15%

-

-

-

-

-

-

133

1,623

8%

Spain

0

0

-

71

106

67%

-

-

-

-

-

-

351

701

50%

Sweden

2

4

50%

82

250

33%

-

-

-

-

-

-

412

1,637

25%

63

115

55%

9

30

30%

-

-

-

-

-

-

353

924

38%

8

23

35%

270

706

38%

-

-

-

-

-

-

1,163

4,550

26%

200

363

55%

1,484

5,121

29%

-

-

-

-

-

-

7,427

34,363

22%

Switzerland United Kingdom Total Panel B: Control Group

Non-IFRS Adoption Countries

IFRS Voluntary

IFRS Mandatory

Local GAAP

US-GAAP

Total

Unique Firms

Unique Firms

Unique Firms

Unique Firms

Firm-Years

Share

Open Market

Share

Open Market

Share

Open Market

Brazil

Open Market -

DS Universe -

-

-

-

-

10

173

6%

-

-

-

25

999

3%

Canada

-

-

-

-

-

-

355

935

38%

-

-

-

1,187

5,063

23%

Chile

-

-

-

-

-

-

1

154

1%

-

-

-

2

892

0%

Indonesia

-

-

-

-

-

-

54

190

28%

-

-

-

221

1,188

19%

Israel

-

-

-

-

-

-

6

66

9%

-

-

-

24

403

6%

Japan

-

-

-

-

-

-

436

2,255

19%

-

-

-

2,711

14,849

18%

Korea

-

-

-

-

-

-

0

633

0%

-

-

-

0

4,093

0%

Mexico

-

-

-

-

-

-

9

90

10%

-

-

-

21

543

4%

Morocco

-

-

-

-

-

-

0

11

0%

-

-

-

0

59

0%

Taiwan

-

-

-

-

-

-

2

676

0%

-

-

-

4

4,397

0%

Thailand

-

-

-

-

-

-

0%

-

-

-

1

1,852

0%

United States

-

-

-

-

-

-

-

-

-

2,311

4,198

55%

11,743

26,096

45%

Total

-

-

-

-

-

-

874

5,487

16%

2,311

4,198

55%

15,939

60,434

26%

Share

Open Market

DS Universe

1

DS Universe

304

DS Universe

DS Universe

Share

Panel C: Treatment and Control Group IFRS Voluntary

IFRS Mandatory

Local GAAP

US-GAAP

Total

Unique Firms

Unique Firms

Unique Firms

Unique Firms

Unique Firms

Year

Open Market

2001

71

DS Universe 223

Share 32%

Open Market 594

DS Universe 3,576

Share

Open Market

17%

392

DS Universe 3,705

Share

Open Market

DS Universe

11%

1,172

3,072

Share

Open Market

DS Universe

38%

2,229

10,576

Share 21%

2002

87

275

32%

673

4,379

15%

444

4,423

10%

1,424

3,544

40%

2,628

12,621

21%

2003

107

309

35%

777

4,810

16%

544

5,136

11%

1,560

3,875

40%

2,988

14,130

21% (continued)

69

2004

148

363

41%

928

5,121

18%

612

5,487

11%

1,779

4,198

42%

3,467

15,169

23%

2005

170

363

47%

1,072

5,121

21%

684

5,487

12%

1,944

4,198

46%

3,870

15,169

26%

2006

172

344

50%

1,174

4,872

24%

756

5,313

14%

2,022

3,926

52%

4,124

14,455

29%

2007

161

289

56%

1,293

4,318

30%

764

4,787

16%

1,842

3,283

56%

4,060

12,677

32%

Total Firm-Years

916

2,166

42%

6,511

32,197

20%

4,196

34,338

12%

11,743

26,096

45%

23,366

94,797

25%

Total Unique Firms

200

363

55%

1,484

5,121

29%

874

5,487

16%

2,311

4,198

55%

4,869

15,169

32%

Notes: This table presents the sample composition by country and by fiscal-year. The Datastream Universe (DS Universe ) comprises a total of 94,797 firm-years from 34 countries between fiscal-year 2001 and 2007 with sufficient data on trading volume, stock returns and accounting standards followed. We split the DS Universe into two groups: (1) countries that introduced IFRS in fiscal-year 2005 (treatment group), and (2) countries that required domestic accounting standards throughout the sample period (control group). We include only companies in the treatment group that switched from local GAAP to IFRS before (IFRS Voluntary ) or in fiscal-year 2005 (IFRS Mandatory ). The control group consists of companies that used domestic accounting standards (Local GAAP or US-GAAP ) throughout the sample period. For simplicity, we refer to Hong Kong as a country in our analyses, although it has the status of a Special Administrative Region of the 3HRSOH¶V Republic of China. Using proprietary data from Frankfurt Stock Exchange (FSE) we identify all firms within the DS Universe whose stocks are traded in the Open Market. The Open Market sample (Open Market ) is a subset of the DS Universe and consists of 23,366 firm-years. Share indicates the proportion of the Open Market sample relative to the DS Universe. Panel A (Panel B) reports the number of unique firms and the number of firm-year observations by country for the treatment (control) group. Panel C shows the number of unique firms by fiscal-year for treatment and control group combined.

70

3.5.2 Descriptive statistics Panel A: Open Market Sample Variables

Firm-Years

Mean

Std.Dev.

P1

P25

Median

P75

P99

Frankfurt Stock Exchange (FSE) Trading Volume (%)

23,366

0.00252%

0.03633%

0.00000%

0.00001%

0.00005%

0.00033%

Trading Volume (Euro)

23,366

18,093

130,592

0

38

273

1,767

0.03834% 442,479

Trading Days (%)

23,366

23.19%

30.69%

0.00%

1.60%

7.87%

32.81%

100.00%

No. of Trades

23,366

2.32

10.79

0.00

0.04

0.17

0.86

45.73

Trade Size (Euro)

21,059

2,675

5,131

78

925

1,800

3,230

15,179

Bid-Ask Spread

13,589

4.84%

7.22%

0.38%

1.99%

3.17%

4.98%

40.00%

Home Markets Trading Volume (%)

23,366

3.05354%

152.75050%

0.03880%

0.37450%

0.72660%

1.40605%

7.45414%

Trading Volume (Euro)

23,366

28,200,000

83,800,000

5,712

660,367

4,929,413

22,700,000

390,000,000

Trading Days (%)

23,366

96.33%

6.21%

65.87%

96.89%

97.63%

97.67%

100.00%

Bid-Ask Spread

13,589

0.67%

1.34%

0.03%

0.13%

0.29%

0.68%

5.71%

Other Variables Return Variability

23,366

11.71%

7.54%

2.78%

6.70%

9.70%

14.51%

39.55%

Market Value (m Euro)

23,366

4,792.91

15,483.78

6.60

186.15

823.46

3,319.06

64,205.58

Up-/Downgrades

23,366

4.66

6.04

0.00

0.00

2.00

7.00

26.00

Google Hits (Germany)

23,366

234

3,846

0

0

3

26

2,670

Google Hits (Worldwide)

23,366

9,573

124,674

0

60

226

1,210

122,000

Google Ratio

23,366

5.57%

14.76%

0.00%

0.00%

1.17%

3.85%

96.17%

EM Aggregate (Firm Level)

20,606

0.45

0.21

0.07

0.28

0.45

0.61

0.90

Panel B: Datastream Universe (excluding Open Market Sample) Variables

Firm-Years

Mean

Std.Dev.

P1

P25

Median

P75

P99

Home Markets 71,431

0.66579%

2.13348%

0.10859%

0.27663%

0.66732%

5.72978%

Trading Volume (Euro)

71,431

2,516,486

21,900,000

122

30,318

184,302

1,102,041

33,500,000

Trading Days (%)

Trading Volume (%)

71,431

85.99%

21.50%

0.00258% 6.69%

87.84%

94.86%

97.28%

100.00%

Bid-Ask Spread

58,335

2.75%

6.11%

0.08%

0.48%

1.03%

2.59%

27.04%

Panel B: Datastream Universe (excluding Open Market Sample) Variables

Firm-Years

Mean

Std.Dev.

P1

P25

Median

P75

P99

Other Variables 71,431

11.66%

8.34%

2.25%

6.38%

9.45%

14.39%

42.72%

Market Value (m Euro)

71,431

556.93

2,964.76

1.73

31.66

102.52

348.78

6,749.23

Up-/Downgrades

Return Variability

71,431

1.50

3.63

0.00

0.00

0.00

1.00

17.00

Google Hits (Germany)

71,431

160

4,296

0

0

0

0

1,710

Google Hits (Worldwide)

71,431

8,938

134,907

0

1

15

106

129,000

Google Ratio

71,431

1.46%

7.63%

0.00%

0.00%

0.00%

0.00%

33.33%

EM Aggregate (Firm Level) 57,129 0.52 0.22 0.07 0.36 0.52 0.68 0.94 Notes: This table reports descriptive statistics of all relevant variables across firm-years for the Open Market sample (panel A) and for the rest of the DS Universe (panel B). The total Datastream (DS) Universe comprises a total of 94,797 firm-years from 34 countries between fiscal-year 2001 and 2007. The Open Market sample is a subset of the DS Universe and consists of 23,366 firm-years. Each panel is subdivided into groups of variables that depend (Frankfurt Stock Exchange (FSE) and/or Home Markets ) or do not depend (Other Variables ) on the trading venue. For all capital market variables the measurement period spans months -6 to +6 relative to the ILUP¶V fiscal-year end date. Trading Volume (%) is the annual stock turnover, i.e. annual trading volume in number of stocks divided by the average number of free floating stocks, divided by the annual number of exchange trading days. Trading Volume (Euro) is the annual trading volume in Euro divided by the annual number of exchange trading days. Trading Days (%) is the number of exchange trading days with non-zero trading volume divided by the annual number of exchange trading days. No. of Trades is the annual number of ticks divided by the annual number of exchange trading days. Trade Size (Euro) is the annual trading volume in Euro divided by the annual number of trades. FSE trading volume information is based on a proprietary dataset from Frankfurt Stock Exchange. We use trading volume from both the floor and from XETRA to compute FSE trading volume variables. Trading volume data for the home markets is retrieved from Datastream. No. of Trades and Trade Size (Euro) are available for FSE only, because Datastream does not provide information on number of trades. Bid-Ask Spread is the yearly median quoted spread (i.e. the difference between the closing bid and the closing ask price divided by the midpoint). For FSE, we use bid-ask spreads from the floor. Information on closing bid and ask prices is obtained from Datastream (CRSP) for non-US (US) exchanges. Return Variability is the annual standard deviation of monthly stock returns (Datastream). Market Value (m Euro) is the market value of outstanding equity in Million Euro at the ILUP¶V fiscal-year end (Datastream). We compute return variability and market value based on data from the home markets, because liquidity is much higher compared to FSE. Up-/Downgrades is the annual number of changes in DQDO\VW¶V recommendations (IBES). This variable is set to zero for firm-years not covered by IBES. Google Hits (Germany) (Google Hits (Worldwide) ) is the annual number of search results in the Google News archive in German (any) language. Google Ratio is Google Hits (Germany) divided by Google Hits (Worldwide) . In case the number of worldwide Google hits is zero, Google Ratio is set to zero. For details on the Google variables, see Appendix B. EM Aggregate (Firm Level) is the average within-country rank of two earnings management (EM) measures: (1) the ratio of the firm-level standard deviations of operating earnings and operating cash-flow (both scaled by lagged total assets), and (2) the ratio of the absolute value of accruals and the absolute value of operating cash-flows. Accounting data is obtained from Worldscope.

71

3.5.3 Institutional characteristics of IFRS and non-IFRS adoption countries Panel A: Macroeconomic Variables Country

MCAP/GDP

Australia

107.38

Turnover/GDP

Control of Corruption

Rule of Law

Regulatory Quality

Euro

German Imports 2004

German Exports 2004

Distance Berlin - Capital

Distance btw Exchanges 16,498

72.35

1.97

1.73

1.62

0

1,302

4,663

16,082

Austria

19.85

4.78

1.98

1.82

1.56

1

24,020

40,244

524

Belgium

100.55

16.10

1.46

1.43

1.29

1

26,525

40,308

653

313

Brazil

41.13

12.68

-0.23

-0.45

0.05

0

4,668

4,644

9,602

9,797

Canada

100.85

60.70

1.92

1.75

1.54

0

2,485

4,915

6,135

6,346

604

Chile

99.90

8.08

1.34

1.16

1.43

0

1,276

919

12,543

12,078

Czech Republic

18.70

7.80

0.44

0.74

1.06

0

16,493

17,766

280

420

56.73 116.58

36.48 125.18

2.24 2.4

1.94 1.9

1.71 1.76

0 1

9,669 6,124

11,358 7,322

353 1,104

684 1,527

81.50

67.20

1.4

1.33

1.1

1

51,535

74,360

880

471

62.28 397.35

23.60 183.85

0.4 1.68

0.65 1.47

0.88 1.83

1 0

1,613 2,041

6,291 4,049

1,804 8,759

1,798 9,287

Hungary Indonesia

20.17 20.08

9.50 43.83

0.61 -0.88

0.71 -0.86

1.11 -0.48

0 0

13,412 2,296

12,816 1,695

690 10,784

836 11,122

Ireland

60.00

25.40

1.69

1.59

1.59

1

14,772

4,311

1,318

1,089

Israel Italy

61.05 44.43

39.05 47.35

0.77 0.39

0.73 0.52

0.86 0.89

0 1

1,269 35,677

2,467 51,479

2,853 1,187

2,955 508 9,369 8,586

Denmark Finland France Greece Hong Kong

Japan Korea (South) Luxembourg Mexico Morocco Netherlands Norway Poland Portugal South Africa Spain Sweden Switzerland Taiwan Thailand United Kingdom United States Source

64.48 52.73

52.63 124.33

1.25 0.5

1.35 0.78

1.17 0.79

0 0

21,583 7,728

12,719 6,397

8,928 8,140

132.38

1.50

1.84

1.9

1.79

1

2,294

3,684

604

176

20.23 26.97

5.18 2.03

-0.39 -0.17

-0.51 -0.08

0.33 -0.33

0 0

1,634 460

4,912 984

9,739 2,615

9,569 2,266

104.63

143.73

1.99

1.72

1.7

1

46,204

46,730

578

365

43.85 19.00

35.55 4.53

2.05 0.19

1.94 0.33

1.47 0.79

0 0

12,270 15,973

5,168 18,776

838 516

1,110 916

40.08 134.60

19.13 64.07

1.15 0.54

1.08 0.18

1.2 0.53

1 0

4,530 3,289

6,720 6,073

2,314 8,829

1,880 8,650

83.00

132.55

1.34

1.1

1.23

1

17,426

36,249

1,873

1,434

95.88 216.55

109.43 185.38

2.1 2.12

1.79 1.97

1.54 1.46

0 0

10,197 21,445

15,730 27,917

810 755

1,203 297

41.98

35.88

0.7

0.85

1.08

0

5,641

4,276

8,959

9,411

50.27 135.03

45.37 149.45

-0.19 1.93

0.1 1.63

0.41 1.57

0 0

2,465 34,466

2,022 59,986

8,612 936

8,978 634

128.40

210.05

World Bank FSDI

World Bank FSDI

1.56

1.52

World Bank WGI

World Bank WGI

1.54 World Bank WGI

0

40,709

64,860

ECB

Federal Stat. Office

Federal Stat. Office

6,718

6,215

CEPII

indo.com

Panel B: Accounting Variables Country Australia Austria Belgium Brazil Canada Chile Czech Republic Denmark Finland France Greece Hong Kong Hungary Indonesia Ireland Israel Italy Japan Korea (South) Luxembourg Mexico Morocco Netherlands Norway Poland

EM1

EM2

EM3

EM4

0.840 0.444 0.585 0.670 0.790 0.642 0.623 0.663 0.681 0.566 0.371 0.762 0.691 0.585 0.839 0.654 0.609 0.596 0.560 0.308 0.716 0.269 0.629 0.790 0.668

-0.488 -0.794 -0.768 -0.754 -0.527 -0.839 -0.826 -0.759 -0.687 -0.783 -0.901 -0.634 -0.794 -0.810 -0.644 -0.697 -0.795 -0.802 -0.781 -0.783 -0.709 -0.876 -0.740 -0.623 -0.752

0.398 0.636 0.620 0.589 0.509 0.475 0.666 0.639 0.587 0.695 0.851 0.623 0.602 0.732 0.441 0.641 0.739 0.668 0.685 0.589 0.548 0.483 0.614 0.661 0.697

2.08 0.86 2.56 2.63 1.47

3.50 1.80 2.99 2.51 2.67

10.00 2.00 1.96 2.50

1.00 4.17 2.00

# Small Profit 77 6 23 21 103 0 0 7 27 206 168 24 0 0 10 4 104 5 0 0 3 0 25 8 5

# Small Loss 37 7 9 8 70 0 0 2 15 69 67 9 0 0 1 2 53 2 0 0 3 0 6 4 0

rEM1

rEM2

rEM3

3 30 25 14 6 18 21 16 13 27 32 8 12 26 4 17 23 24 28 33 10 34 19 5 15

1 24 18 16 3 30 29 17 10 20 34 6 23 27 7 12 25 26 19 21 13 33 14 4 15

1 21 19 16 7 4 26 22 14 29 34 20 17 31 3 23 32 27 28 15 9 5 18 25 30

EM Aggregate 1.7 25.0 20.7 15.3 5.3 17.3 25.3 18.3 12.3 25.3 33.3 11.3 17.3 28.0 4.7 17.3 26.7 25.7 25.0 23.0 10.7 24.0 17.0 11.3 20.0

LNW EM Aggregate

Distance to IFRS

4.8 28.3 19.5

4 12 13 11 5 13 14 11 15 12 17 3 13 4 1 6 12 9 6 9 1

5.3

16.0 12.0 13.5 28.3 19.5 18.3 5.1 24.8 20.5 26.8

16.5 5.8

4 7 12

Absence Divergence 22 34 22

21 36 32

4

25

31 22 21 40 14

21 31 34 28 15

12 0

12 34

27 18 15

37 22 11

10 7

25 17 (continued)

72

Portugal South Africa Spain Sweden Switzerland Taiwan Thailand United Kingdom United States

0.503 0.705 0.419 0.874 0.767 0.624 0.617 0.750 0.854

-0.863 -0.686 -0.842 -0.625 -0.694 -0.813 -0.791 -0.666 -0.520

0.741 0.430 0.587 0.525 0.553 0.646 0.562 0.549 0.500

2.50 1.00 7.50 6.00 2.75

1.93 1.38

20 1 30 6 22 14 2 239 695

8 1 4 1 8 0 0 124 502

29 11 31 1 7 20 22 9 2

32 9 31 5 11 28 22 8 2

33 2 13 8 11 24 12 10 6

31.3 7.3 25.0 4.7 9.7 24.0 18.7 9.0 3.3

25.1 5.6 18.6 6.8 22.0 22.5 18.3 7.0 2.0

13 0 16 10 12 6 4 1 4

29 7 28 10

22 1 29 26

19 29 0 6

23 7 35 23

Ding et al. Leuz et al. Bae et al. self-constructed (2007) (2003) (2008) Notes: This table presents institutional characteristics for all countries used in our analyses. Panel A (Panel B) contains macroeconomic (accounting) variables. MCAP/GDP is the ratio of a FRXQWU\¶V stock market capitalization to its Gross Domestic Product (GDP). Turnover/GDP is the ratio of a FRXQWU\¶V stock market trading volume to its GDP. Yearly ratios are obtained from the World Bank (www.fsdi.org). We compute both variables as the mean ratio over the period 2001-2004. Control of Corruption , Rule of Law and Regulatory Quality are Worldwide Governance Indicators (WGI) for the year 2005 provided by World Bank (see Kaufmann, Kraay, and Mastruzzi (2009) for details). Higher values represent countries with less corruption, higher quality legal enforcement and higher regulatory quality, respectively. Euro distinguishes between countries that have adopted the Euro as their national currency (variable equals one) and countries that have not (variable equals zero). We collect this information from the European Central Bank (www.ecb.int). German Imports 2004 (German Exports 2004 ) is the total amount of German imports from (German exports to) the respective country during year 2004 in Million Euro. Data is obtained from the Federal Statistical Office of Germany (www.destatis.de). Distance Berlin ± Capital is the distance between Berlin, capital of Germany, and the capital of the respective country in kilometers. Data source is the French research centre in international economics (CEPII). Distance btw Exchanges is the kilometer distance between Frankfurt, the home of the Open Market, and the main stock exchange in the respective country. We obtain this information from www.indo.com. Following Leuz, Nanda, and Wysocki (2003) we construct updated country-specific earnings management (EM) scores using all firmyears in the Datastream Universe and our sample period with sufficient accounting data in Worldscope. EM1 is the FRXQWU\¶V median ratio of the firm-level standard deviations of operating earnings and cash flow from operations (both scaled by lagged total assets). Operating cash flow equals operating earnings minus accruals with accruals = ǻWRWDO current assets ± ǻFDVK ± ǻWRWDO current liabilities ± ǻVKRUWWHUP debt ± ǻWD[HV payable) ± depreciation expense. EM2 is the FRXQWU\¶V Spearman correlation between the change in accruals and the change in cash flow from operations (both scaled by lagged total assets). EM3 is the FRXQWU\¶V median ratio of the absolute value of accruals and the absolute value of the cash flow from operations. EM4 is the number of small profits (# Small Profit) divided by the number of small losses (# Small Loss ). A firm-year observation is classified as a small profit (loss) if net earnings scaled by lagged total assets are in the range [0, 0.01] ([-0.01, 0)). Higher (lower) values of EM1 and EM2 (EM3 and EM4) correspond to less earnings management. EM4 is missing for some countries due to lack of small losses during our sample period. In addition, it is not clear whether EM4 is an appropriate proxy to compare earnings management (Durtschi and Easton (2005)). We therefore rank countries using the first three earnings management measures only (rEM1 ± rEM3 ) and compute EM Aggregate as the average rank. LNW EM Aggregate is the aggregate earnings management score from Leuz, Nanda, and Wysocki (2003). Distance to IFRS is the Bae, Tan, and Welker (2008) summary score of how the FRXQWU\¶V local GAAP differs from IFRS on 21 key accounting dimensions. Higher values stand for more discrepancies between local GAAP and IFRS. Absence (Divergence ) is obtained from Ding, Hope, Jeanjean, and Stolowy (2007) and represents how many out of 111 IFRS accounting issues are not covered (differ from IFRS) in the local GAAP of the respective country.

Source

73

3.5.4 Determinants of Open Market inclusion Dependent Variable: Open Market Inclusion (0 = No, 1 = Yes)

Independent Variables Log(Return Variability) Log(Home Trading Volume (%)) Log(Up-/Downgrades) Google Ratio Log(Market Value (Firm Level))

Pooled Probit 0.39

0.38

0.40

0.41

0.42

0.47

0.45

0.36

(20.98)***

(20.35)***

(21.76)***

(22.46)***

(22.82)***

(25.33)***

(23.88)***

(16.68)***

0.16

0.18

0.18

0.19

0.19

0.13

0.14

0.11

(17.44)***

(19.04)***

(19.30)***

(20.60)***

(20.54)***

(14.26)***

(14.01)***

(9.43)***

0.10

0.12

0.13

0.13

0.12

0.14

0.11

0.11

(8.01)***

(9.92)***

(10.72)***

(10.98)***

(10.56)***

(12.03)***

(9.17)***

(8.49)***

1.73

1.65

1.69

1.60

1.55

1.61

1.56

1.50

(16.87)***

(15.12)***

(16.57)***

(15.94)***

(15.44)***

(16.02)***

(15.05)***

(12.50)***

0.32

0.31

0.31

0.30

0.30

0.29

0.31

0.30

(37.20)***

(36.60)***

(36.68)***

(36.90)***

(36.90)***

(33.69)***

(35.36)***

(30.13)***

EM Measure (Firm Level)

-0.70 (11.35)***

EM Measure (Country Level)

-0.03

-0.03

-0.04

(21.29)***

(15.43)***

(16.92)***

Log(MCAP/GDP)

0.43 (21.18)***

Distance to IFRS

0.05

(4.35)***

(9.56)***

(8.99)***

0.18

0.39

0.40

(4.57)***

(6.23)***

(5.90)***

Log(Distance btw Exchanges)

-0.07 (5.84)***

Log(Imports + Exports) Year

Year

0.31 (9.23)***

0.05

Euro

Fixed Effects

0.29 (9.56)*** -0.02

Year

Year

Year

0.07

0.11

(4.45)***

(5.94)***

0.21

0.15

0.20

(19.45)***

(12.38)***

(15.08)***

Year

Year

Year

0.27 0.26 0.24 0.24 0.25 0.26 0.29 0.31 (Pseudo) R2 Observations 94,797 94,797 94,738 94,797 94,797 94,797 94,738 77,704 Notes: This table presents results from regressions that relate the likelihood of inclusion into the Open Market to various country- and firm-specific variables. The DS Universe comprises a total of 94,797 firm-years from 34 countries between fiscal-year 2001 and 2007. The Open Market sample is a subset of the DS Universe and consists of 23,366 firm-years. The dependent variable in all regressions is a binary variable that equals one (zero) for all firm-years in the Open Market sample (the rest of the DS Universe). Market Value (Firm Level) is a ILUP¶V mean market value over the sample period. EM Measure (Country Level) is the self-constructed earnings management score EM Aggregate (see table 3.5.3). Imports (Exports) is the total amount of German imports from (German exports to) a ILUP¶V country of domicile during year 2004 (see table 3.5.3). For a description of the remaining variables see table 3.5.2 (firmlevel) and table 3.5.3 (country-level). The table reports coefficient estimates and (in parentheses) t-statistics. The t-statistics are based on standard errors that are clustered by firm. We use the natural logarithm of the raw values (plus a small constant in case of Up-/Downgrades ) where indicated in the panels. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed) respectively.

74

3.5.5 Mandatory IFRS adoption and Open Market inclusion Independent Variables Post-IFRS Post-IFRS * Voluntary Post-IFRS * Mandatory Post-IFRS * USA

Dependent Variable: Open Market Inclusion (0 = No, 1 = Yes) Firm Fixed Effects

Pooled Probit

0.03

0.20

0.20

0.88

0.87

(17.23)***

(16.41)***

(9.77)***

(9.39)***

(19.62)***

(17.99)***

0.13

0.12

0.51

0.53

0.91

0.92

(7.11)***

(6.52)***

(6.40)***

(5.85)***

(6.77)***

(6.08)***

0.05

0.05

0.15

0.13

0.43

0.40

(12.28)***

(11.54)***

(4.77)***

(4.08)***

(7.48)***

(6.34)***

0.08

0.08

0.14

0.14

0.53

0.55

(16.59)***

(16.47)***

(5.33)***

(4.91)***

(8.99)***

(8.40)***

Log(Return Variability) Log(Home Trading Volume (%)) Log(Up-/Downgrades) Google Ratio

0.01

0.44

0.35

0.40

0.27

(3.04)***

(25.04)***

(17.42)***

(14.38)***

(8.71)***

0.01

0.15

0.11

0.29

0.23

(9.07)***

(14.97)***

(10.22)***

(18.83)***

(13.50)***

0.02

0.11

0.11

0.22

0.23

(11.03)***

(9.44)***

(8.70)***

(11.62)***

(10.84)***

0.11

1.50

1.46

1.93

2.06

(7.84)***

(14.64)***

(12.22)***

(14.15)***

(13.08)***

Log(Market Value (Firm Level))

0.31

0.30

0.78

0.76

(36.39)***

(31.06)***

(52.24)***

(46.00)***

EM Measure (Firm Level) EM Measure (Country Level) Log(MCAP/GDP) Euro Distance to IFRS Log(Distance btw Exchanges) Log(Imports + Exports) 2

(Pseudo) R

Observations

Random Effects Probit

0.04

-0.70

-1.73

(-11.33)***

(-13.78)***

-0.03

-0.03

-0.09

-0.10

(-14.16)***

(-15.73)***

(-22.42)***

(-23.43)***

0.27

0.29

0.45

0.47

(8.88)***

(8.62)***

(9.26)***

(8.69)***

0.37

0.39

0.94

1.02

(5.88)***

(5.59)***

(7.45)***

(7.31)***

0.05

0.05

0.09

0.08

(9.15)***

(8.58)***

(7.95)***

(6.99)***

0.09

0.12

0.13

0.23

(5.24)***

(6.52)***

(3.87)***

(5.90)***

0.14

0.19

0.25

0.41

(11.20)***

(13.72)***

(9.44)***

(13.85)***

0.87

0.87

0.29

0.31

0.24

0.37

94,797

94,797

94,738

77,704

94,738

77,704

Notes: This table presents results from regressions that relate the likelihood of inclusion into the Open Market to IFRS adoption. The DS Universe comprises a total of 94,797 firm-years from 34 countries between fiscal-year 2001 and 2007. The Open Market sample is a subset of the DS Universe and consists of 23,366 firm-years. The dependent variable in all regressions is a binary variable that equals one (zero) for all firm-years in the Open Market sample (the rest of the DS Universe). Post-IFRS equals one (zero) for fiscal-years 20052007 (2001-2004). Voluntary (Mandatory ) indicates companies that switched from local GAAP to IFRS before (in) fiscal-year 2005. USA equals one for US firms and zero otherwise. For a description of the control variables see table 3.5.4. The table reports coefficient estimates and (in parentheses) t-statistics. The t-statistics are either based on robust standards errors (random effects probit) or on standard errors that are clustered by firm (all other specifications). We use the natural logarithm of the raw values (plus a small constant in case of Up-/Downgrades ) where indicated in the panels. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed) respectively.

75

3.5.6 Mandatory IFRS adoption and Open Market trading activity Panel A: Percentage Trading Volume as Dependent Variable Independent Variables Post-IFRS Post-IFRS * Voluntary Post-IFRS * Mandatory Post-IFRS * USA

Dependent Variable: Log(FSE Trading Volume (%)) (1)

(2)

(3)

(4)

(5)

(6)

0.50

0.50

0.28

0.49

0.50

0.30

(9.73)***

(9.90)***

(5.54)***

(9.72)***

(9.75)***

(5.90)***

0.18

0.21

0.19

0.13

0.16

0.17

1.49

(1.76)*

1.64

1.11

1.32

1.44

0.19

0.22

0.33

0.18

0.19

0.32

(2.76)***

(3.21)***

(4.86)***

(2.56)**

(2.70)***

(4.77)***

-0.80

-0.72

-0.45

-0.81

-0.81

-0.45

(-13.34)***

(-12.15)***

(-7.57)***

(-13.53)***

(-13.46)***

(-7.61)***

Log(Return Variability)

0.42

0.19

(15.51)***

(6.95)***

Log(Home Trading Volume (%))

0.69

0.64

(29.85)***

(26.35)***

Log(Up-/Downgrades)

0.21

0.09

(12.55)***

(5.37)***

Google Ratio

0.55

0.49

(5.97)***

(5.59)*** Firm

Fixed Effects

Firm

Firm

Firm

Firm

Firm

R2 Observations

0.82

0.83

0.84

0.83

0.82

0.84

23,366

23,366

23,366

23,366

23,366

23,366

76

Panel B: Proportion of Non-Zero FSE Trading Volume Days as Dependent Variable Independent Variables Post-IFRS Post-IFRS * Voluntary Post-IFRS * Mandatory Post-IFRS * USA

Dependent Variable: FSE Trading Days (%) (1)

(2)

(3)

(4)

(5)

(6)

0.05

0.05

0.04

0.05

0.05

0.04

(9.14)***

(9.23)***

(6.63)***

(9.13)***

(9.17)***

(6.95)***

0.07

0.07

0.07

0.07

0.07

0.07

(4.36)***

(4.56)***

(4.48)***

(4.17)***

(4.23)***

(4.33)***

0.03

0.03

0.04

0.03

0.03

0.04

(3.71)***

(3.98)***

(4.83)***

(3.57)***

(3.65)***

(4.71)***

-0.05

-0.04

-0.03

-0.05

-0.05

-0.03

(-7.21)***

(-6.42)***

(-3.79)***

(-7.35)***

(-7.32)***

(-3.93)***

Log(Return Variability)

0.03

0.02

(11.74)***

(5.85)***

Log(Home Trading Volume (%))

0.04

0.04

(18.12)***

(15.53)***

Log(Up-/Downgrades)

0.02

0.01

(10.87)***

(6.77)***

Google Ratio

0.07

0.06

(6.79)***

(6.34)***

Fixed Effects

Firm

Firm

Firm

Firm

Firm

Firm

R2

0.89

0.89

0.90

0.89

0.89

0.90

Observations 23,366 23,366 23,366 23,366 23,366 23,366 Notes: This table presents results from regressions that relate Open Market trading volume at Frankfurt Stock Exchange (FSE) to IFRS adoption. The Open Market sample comprises a total of 23,366 firm-years from 31 countries between fiscal-year 2001 and 2007. Panel A (Panel B) reports results from multi-period differences-in-differences analyses with FSE trading volume in % (the proportion of non-zero FSE trading volume days) as dependent variable. Post-IFRS equals one (zero) for fiscal-years 2005-2007 (2001-2004). Voluntary (Mandatory ) indicates companies that switched from local GAAP to IFRS before (in) fiscal-year 2005. USA equals one for US firms and zero otherwise. For a description of the remaining variables see tables 3.5.2 to 3.5.5. The table reports OLS coefficient estimates and (in parentheses) t-statistics. The t-statistics are based on standard errors that are clustered by firm. We use the natural logarithm of the raw values (plus a small constant in case of FSE Trading Volume (%) and Up-/Downgrades ) where indicated in the panels. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed) respectively.

77

3.5.7 Cross-sectional analysis of the Open Market reaction Dependent Variable: Log(FSE Trading Volume (%)) Independent Variables Post-IFRS Post-IFRS*Voluntary Post-IFRS*Voluntary*Conditional Post-IFRS*Mandatory Post-IFRS*Mandatory*Conditional Post-IFRS*USA Controls Fixed Effects 2

R

Observations Conditional Variable Threshold Conditional Variable = 0

Conditional Variable Rule of Law

MCAP/GDP

Distance to IFRS

LNW EM Aggregate

EM Measure (country level)

0.30

0.30

0.30

0.30

0.30

(5.90)***

(5.89)***

(5.87)***

(5.90)***

(5.88)***

0.22

0.22

-0.06

-0.08

0.20

1.08

1.19

-0.17

-0.32

1.39

-0.07

-0.09

0.27

0.36

-0.07

-0.28

-0.41

0.79

-1.29

-0.33

0.34

0.33

0.39

0.29

0.37

(4.14)***

(3.41)***

(4.90)***

(3.70)***

(4.75)***

-0.03

-0.01

-0.17

0.09

-0.13

-0.37

-0.05

(-1.77)*

-0.95

-1.36

-0.45

-0.45

-0.45

-0.45

-0.45

(-7.61)***

(-7.61)***

(-7.59)***

(7.62)***

(-7.59)***

Included

Included

Included

Included

Included

Firm

Firm

Firm

Firm

Firm

0.84

0.84

0.84

0.84

0.84

23,366

23,366

23,366

23,265

23,366

1.59

83

12

16.25

17.33

3,449

2,176

4,524

4,197

4,981

Conditional Variable = 1 3,978 5,251 2,903 3,129 2,446 Notes: This table presents results from regressions that relate Open Market trading volume at Frankfurt Stock Exchange (FSE) to IFRS adoption conditional on several country-level variables. The Open Market sample comprises a total of 23,366 firm-years from 31 countries between fiscal-year 2001 and 2007. The table reports results from multi-period differences-in-differences analyses with FSE trading volume (in %) as dependent variable. We include all control variables in the regression models equivalent to model (6) in table 3.5.6, panel A. Post-IFRS equals one (zero) for fiscal-years 2005-2007 (2001-2004). Voluntary (Mandatory ) indicates companies that switched from local GAAP to IFRS before (in) fiscal-year 2005. USA equals one for US firms and zero otherwise. The conditional variables are described in table 3.5.3. If these variables are continuous we transform them into binary variables splitting by the median computed over the treatment sample countries (conditional variable threshold). The table reports OLS coefficient estimates and (in parentheses) t-statistics. The t-statistics are based on standard errors clustered by firm. We use the natural logarithm of the raw values (plus a small constant in case of FSE Trading Volume (%) and Up-/Downgrades ) where indicated in the panels. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed) respectively .

78

3.6 Appendices 3.6.1 Comparison of transaction costs: Open Market versus home markets German investors have two options when trading foreign stocks that are included in the Open Market. They can either trade at FSE or they can trade abroad, i.e. in the respective home market. Both options involve transaction costs that differ considerably in nature. This appendix seeks to illustrate these differences by means of an example. Table A1 shows concurrent quotes for Fiat stocks (ISIN IT0001976403) at the home market in Milan and at the Open Market in Frankfurt. We collect this information through the web-portal of comdirect bank, the leading online broker for German individual investors. Table A2 presents order fees at comdirect bank for trading stocks at Milan Stock Exchange (MSE) and FSE, respectively. While MSE offers better prices (i.e. the MSE bid-ask spread is inside the FSE quotes), comdirect bank clients incur lower order fees (both fixed and variable) for trading at FSE. Lower order fees at FSE outweigh the price advantage at MSE if and only if the size of the trade is sufficiently small.62 For example, a buy of 100 Fiat stocks would cost 785.06 Euro at MSE, but only 772.01 Euro at FSE.63 In contrast, buying 1,000 Fiat stocks is cheaper at MSE (7,609.42 Euro versus 7,617.48 Euro at FSE).64 The break-even where MSE turns into the more cost-efficient trading alternative is about 500 units or 3,800 Euro in this particular case. This is also reflected in the different quote sizes at MSE (around 10,000 stocks) and FSE (200 stocks). Taken together, this example confirms our depiction of the Open Market as a platform for German individual investors to trade small sizes of foreign stocks.

62

The difference in transaction costs between FSE and MSE is a monotonic function, because the price advantage of MSE (about 0.5%) is larger than the advantage of FSE in variable order fees (about 0.12%). The advantage of FSE in fixed order fees therefore has less impact on the total transaction costs the higher the trade size.

63

The order value (order processing costs) is (are) 756.00 (29.06) Euro at MSE and 760.00 (12.01) Euro at FSE.

64

The order value (order processing costs) is (are) 7,560.00 (49.42) Euro at MSE and 7,600.00 (17.48) Euro at FSE. 79

Table A1: Concurrent Quotes for Fiat Stocks (ISIN IT0001976403)

Notes: This table presents concurrent quotes for Fiat stocks (ISIN IT0001976403) at the home market in Milan (German: Mailand) and at the Open Market in Frankfurt. The information was retrieved from the website of comdirect bank (www.comdirect.de) on 22 April 2009. The upper part of the table contains information on the relevant exchange (Börse), the last price (Aktuell), the time the last price was set (Zeit), the percentage difference between the last price and the price of the previous day (Diff. Vortrag) as well as the trading volume in Euro (Tages-Vol.) and in units (Gehandelte Stück). The lower part of the table provides details on the current bid (Geld) and ask quote (Brief), the time these quotes were set (Zeit), the percentage spread (Spread) as well as the size of the current bid (Geld Stk.) and ask quote (Brief Stk.).

Table A2: Order Fees at comdirect bank Type of Fee

Milan Stock Exchange (MSE)

Frankfurt Stock Exchange (FSE)

Order Provision

7.90 Euro + 0.25% of order value (min. 12.90 Euro, max. 62.90 Euro)

4.90 Euro + 0.25% of order value (min. 9.90 Euro, max. 59.90 Euro)

Brokerage Fee

-

0.08% of order value

Exchange Fee

0.20% of order value (min. 8.66 Euro)

0.0015% of order value (min. 1.50 Euro)

Delivery Fee

7.50 Euro

-

Notes: This table presents order fees at comdirect bank for trading stocks at Milan Stock Exchange (MSE) and Frankfurt Stock Exchange (FSE), respectively. The information was provided by comdirect bank customer support. Order Provision is the fee comdirect bank charges for its services. All other fees are charges by third parties that comdirect bank passes on. Brokerage Fee (Exchange Fee) is a charge for the FSE lead broker (respective exchange). Delivery Fee is a charge for stock clearing.

80

3.6.2 Google search results and trading volume Our proxy for media coverage is the number of Google News archive search results which we extract through the website http://news.google.com/archivesearch/ advanced_search. Google provides the following information with regard to the number and type of sources of its News archive search function:65 ³1HZV DUFKLYH VHDUFK VHDUFKHV DFURVV D ODUJH FROOHFWLRQ RI KLVWRULFDO DUFKLYHV LQFOXGLQJ major newspapers/magazines, news archives and legal archives. Search results include both content that accessible to all users (such as BBC News, Time Magazine and Guardian) and content that requires a fee (such as Washington Post Archives, Newspaper Archive, and New York Times Archives). In addition to crawling content online, we've also worked with newspapers to digitize materials via our News Archive Partner Program. Through partnerships with newspapers around the world, the News Archive Partner Program makes unique and previously-unavailable newspaper content searchable DQGEURZVDEOHRQOLQH´(http://news.google.com/archivesearch/help.html) We count the number of search results in the Google News archive for each firm-year in the DS Universe by imposing the following requirements:66 1) Find results with all of the words: Company name (as provided by Worldscope item WC06001) or the firm-VSHFLILF,6,1&RGHHJ³),$7´25IT0001976403) 2) Results published between: - WR  UHODWLYH WR WKH ILUP¶V ILVFDO-year end date (same measurement period as for capital market variables) 3) Results written in: a) German (Google Hits ± Germany); b) any language (Google Hits ± Worldwide) Table B1 presents correlations between the number of Google search results and trading volume variables for the Open Market sample. For Google Hits and Home/FSE Trading Volume we use the natural logarithm of the raw values, because these variables

65

$FFRUGLQJWRWKH*RRJOH1HZV$UFKLYHVXSSRUW³DOLVWRIFRQWHQWSURYLGHUVDQGVRXUFHVIRUWKLVQHZ IHDWXUHLVFXUUHQWO\QRWDYDLODEOH´

66

Note that Google is continuously extending its News archive by establishing new partnerships with sources around the world. We retrieved the number of Google News archive search results for the whole DS Universe within three weeks at the end of March / beginning of April 2009. Subsequent consistency checks confirmed that the coverage of the News archive did not increase substantially during this period. 81

are highly skewed (see table 3.5.2 for descriptive statistics).67 The correlation matrix shows that trading volume at FSE has a stronger association with media coverage in Germany (measured by Google Hits ± Germany and Google Ratio) than with worldwide media coverage (measured by Google Hits ± Worldwide). For home market trading volume it is the other way round. These results are intuitive and confirm that the number of Google search results is a useful proxy for media coverage. Table B1: Correlation Matrix Variables

[1]

[1] Log(Google Hits (Germany ))

[2]

[3]

[4]

[5]

[6]

0.78

0.24

0.19

0.14

-0.44

[2] Log(Google Hits (Worldwide))

0.80

[3] Google Ratio

0.76

0.33

-0.11

0.29

0.04

-0.28

-0.08

0.12

-0.15

[4] Log(Home Trading Volume (%))

0.25

0.32

0.14

[5] Log(FSE Trading Volume (%))

0.15

0.03

0.22

0.14

[6] FSE Zero Trading Volume (%)

-0.43

-0.30

-0.40

-0.20

0.13

-0.15 -0.68

-0.80

Notes: This table presents correlations between the number of Google News archive search results and trading volume variables for the Open Market sample. The Open Market sample is a subset of the DS Universe and consists of 23,366 firm-years. Spearman (Pearson) correlations are above (below) the diagonal. For a description of all variables see table 3.5.2. We use the natural logarithm of the raw values (plus a small constant if appropriate) where indicated in the panels.

67

82

We add 1 hit (0.000001%) to Google Hits (FSE Trading Volume) to preclude computing the natural logarithm of zero.

4 The economic consequences of fair value reclassifications under IFRS 4.1 Introduction At the peak of the financial crisis in October 2008, the IASB forwent any regular due process to issue an emergency amendment to IAS 39 in order to relax fair value accounting. These amendments leave banks reporting under IFRS with the choice to retroactively reclassify financial assets that were previously measured at fair value into categories which require measurement at amortized cost, i.e. to effectively abandon fair value accounting for these assets. This decision sharply contrDVWHG WKH ,$6%¶V JHQHUDO strategy in reporting for financial instruments (IASB, 2008a) and its strong initial position against reclassifications.68 However, the board eventually surrendered to severe political pressure by the EU Commission and EU leaders, most prominently the French president Nicolas Sarkozy. They repeatedly voiced their concerns about the procyclicality that fair value accounting may introduce and requested to align accounting rules for European banks with those that apply to its US competitors, for which a similar reclassification option already existed under SFAS 65 and 115.69 In contrast, opponents of easing fair value rules, such as the British Prime Minister Gordon Brown, who stressed that fair values simply reflect current economic reality, went unheeded.70 This conflict at the top political level in the heat of the financial crisis has been the culmination of a longtime controversial debate in standard setting (e.g. Joint Working Group of Standard Setters, 1999; IASC, 1997) and academics (e.g. Barth, 2006; Landsman, 2007; Ronen, 2008) about the pros and cons of fair value accounting. Banks made ample use of the opportunity to forgo substantial write-downs of financial assets that had become illiquid during the course of the crisis. For e[DPSOH*HUPDQ\¶V Deutsche Bank was able to boost its net income for 2008 by 3.2bn Euros by reclassifying troubled assets with a book value of 23.6bn Euros. In effect, since compliance with EDQNV¶ UHJXODWRU\ FDSLWDO UHTXLUHPHQWV LV FORVHO\ WLHG WR DFFRXQWLng numbers, the amendment to IAS 39 was politically perceived as a cheap way to put pressure off the troubled banking sector (Bushman and Landsman, 2010), especially because it was no longer politically opportune to force banks into bankruptcy after the Lehman experience 68

See, e.g., Reuters News, Philippe Danjou: Reporting rules should not favour bank cushions, September 11, 2008; Reuters News, David Tweedie: Big rewrite of accounting fair value ruled out, July 10, 2008.

69

See, e.g., 3DODLVGHO¶(O\VHH, Statement, Summit of European G8 Members, October 4, 2008.

70

See, e.g., Financial Times, Debate to ease fair value accounting intensifies, October 15, 2008. 83

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1_4, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

in September 2008. Yet, the perceived consequential benefits and, perhaps even more importantly, the potential costs of the politically forced decision to allow the suspension of fair value measurement are largely unknown.71 The purpose of this essay is to empirically investigate the economic consequences of WKHUHOD[DWLRQRIIDLUYDOXHUXOHV7RWKLVHQGDQDQDO\VLVRIWKH,$6%¶VDGRSWLRQRIWKH reclassification amendment to IAS 39 has at least two advantages compared with the reclassification option offered by SFAS 65 and 115 to US banks. First, the reclassification option under US-GAAP is hardly used by US banks (Laux and Leuz, 2010). This is due to major differences between the two accounting regimes. The IAS 39 reclassification option is applicable retroactively (at least for the short but highly relevant time window July to October 2008), i.e. banks can reverse already recognized fair value write-downs with the benefit of hindsight. Moreover, impairment rules after reclassification are more restrictive under US-GAAP, i.e. the extent of write-downs that a bank can potentially forego by reclassifying its fair value assets is, ceteris paribus, larger under IFRS. Second, the international setting allows us to exploit cross-country differences in the link between financial accounting and capital regulation to provide insight into how banking regulation has affected accounting choices related to fair value accounting during the financial crisis. Given the complexities of the short- and long-term consequences the IASB was facing in its decision to allow (or deny) the request for reclassifications of fair value assets, it is beyond the scope of this paper to make any judgment on whether the amendment decision was socially optimal. However, we are able to study and quantify the effects of UHFODVVLILFDWLRQV RQ EDQNV¶ ILQDQFLDO VWDWHPHQWV WKH UHVXOWLQJ VKRUW-term economic consequences during the extraordinary circumstances of the crisis, and ultimately the long-run effects on information asymmetry and adverse selection when fair value information about financial assets is eliminated from financial statements of banks. Our analysis proceeds in three steps. First, we provide descriptive evidence on (1) the magnitude and impact of fair value reclassifications on accounting numbers and (2) 71

84

In fact, standard setters (e.g., House of Commons, 2008) and investor advocates (e.g., CFA Institute, 2008) fear that the actions taken have adverse effects on the information the reporting standards SURYLGH WR FDSLWDO PDUNHWV SDUWLFLSDQWV WKH NH\ WDUJHW JURXS RI WKH ,$6%¶V VWDQGDUG VHWWLQJ SURFHVV (IASB Framework, para. 10). In addition, there are significant worries that the interference of the European Commission into the formally independent standard setting process of the IASB poses a major obstacle to further convergence of accounting standards, and in particular the adoption of IFRS in the US, see e.g., Financial Times, Put the brakes on convergence before it is too late, October 30, 2008.

disclosure practices related to the reclassifications. Using a comprehensive global sample of 302 publically-traded and IFRS reporting banks, we show that more than one third (124 banks) chose to reclassify some assets, thus increasing aggregate profits by a total of 22.7 billion Euros and firm-specific profits by 44% on average (see also CESR, 2009). 7KHVH VWDWLVWLFV FRQILUP WKDW EDQNV¶ ILQDQFLDO VWDWHPHQWV DURXQG WKH ZRUOG DUH substantially affected by the amendments to IAS 39. Further, we document that within the subset of reclassifying banks, only 42 banks fully comply with the newly introduced IFRS 7 disclosure requirements designed to mitigate the informational effects of reclassifications. In the second step of our analysis, we empirically analyze the factual underlying GHWHUPLQDQWVRIDQLQGLYLGXDOEDQN¶VUHFODVVLILFDWLRQFKRLFHDQGEHQFKPDUNWKHPDJDLQVW the official arguments put forward in the political discussions surrounding the amendment. We document and exploit the cross-country differences in the link between fair value accounting under IFRS and the determination of regulatory capital to provide HYLGHQFH RQ EDQN PDQDJHUV¶ LQFHQWLYHV WR IRUJR IDLU YDOXH ZULWH-downs. Major differences stem not only from the level of the minimum capital adequacy ratio, but also from the prudential filters applied on unrealized fair value gains of AFS assets (positive revaluation reserves). We find that banks are more likely to reclassify if (1) their total capital ratio before reclassifications is close to the local minimum capital adequacy ratio and (2) the country-specific discount (haircut) on unrealized fair value gains is small, i.e. more regulatory capital is at risk in times of falling market prices. Moreover, a number of system-relevant banks would have faced the risk of immediate supervisory interventions had the amendment or, alternatively, changes to regulatory capital requirements not been adopted. Thus, our findings are consistent with reclassifications granting regulatory forbearance to economically weak banks (see also Skinner, 2008), rather than reducing disadvantages relative to US banks (Laux and Leuz, 2010). In the third part of our empirical analysis, we investigate both short-term and the longterm capital market reactions to the abandonment of fair value measurement. While in times of high market volatility and uncertainty we cannot detect any robustly positive or QHJDWLYHUHDFWLRQWRHLWKHUWKHDQQRXQFHPHQWRIWKH,$6%¶VUHJXODWRU\FKDQJHRQ October 2008 or the bank-specific reclassification announcements, but we do find statistically and economically significant positive abnormal returns for banks that are close to violating regulatory capital restrictions. We complement this analysis by showing that market-adjusted buy and hold returns for reclassifying banks are 85

significantly higher than for other banks in the following year. Overall, this evidence suggest that the amendment did provide initial relief for banks with financial difficulties, and arguably helped them short-term (along with other policy mechanisms) to recover from the effects of the crisis. Where the former analyses document crisis-specific short-term benefits due to the link between fair value measurement and banking regulation, it is the question of how a decrease in fair value information affects reporting transparency, i.e. whether the shortterm relief comes at long-term costs. We therefore study the long-term changes in information asymmetry by comparing the bid-ask spreads of reclassifying and nonreclassifying banks in a difference-in-differences design. We distinguish between banks that fully comply with the accompanying disclosure requirements and banks that withhold information about the reclassified assets. Our analysis shows that the reclassification of financial assets coincides with an increase in bid-ask spreads if the impact on net income is strong and the corresponding disclosure requirements are not fully complied with. This effect is both statistically significant and economically LPSRUWDQW 7KLV ILQGLQJ FRQILUPV WKH ,$6%¶V DQG LQYHVWRU DGYRFDWHV¶ IHDUV WKDW WKH relaxation of fair value accounting of financial assets provides bank managers with leeway to reduce the transparency of financial reporting. This decrease in transparency can lead to adverse capital market effects which is particularly critical in times of crisis (Lang and Maffett, 2010). At the same time, we cannot identify any negative effects of the switch from fair value recognition to disclosure per se. In general terms, our study contributes to the discussion among regulators and standard setters in the current debate on banking supervision and accounting for financial instruments by banks (e.g., Barth and Landsman, 2010; Bushman and Landsman, 2010), in particular on the contracting and informational role of fair value measurement in an international setting (for the US, see e.g. Laux and Leuz, 2009, 2010). We also present evidence consistent with the political cost hypothesis (Watts and Zimmerman, 1978). In this crisis-specific setting, it is, however, upward management of earnings (rather than downward management) which minimizes the costs of future lobbying activities and the risk of adverse regulatory actions. By analyzing what happens when the IASB is forced into a decision by a powerful economic group among its adopting countries (in our setting, the EU), we also contribute to the literature on the politics of standard setting (e.g. Watts and Zimmermann, 1986), particularly adding the supranational dimension a global standard setter is facing (Sunder, 2009). In this respect, the reclassification 86

DPHQGPHQW FDQ EH FRQVLGHUHG DV D SURPLQHQW H[DPSOH IRU ³UHJXODWRU\ DUELWUDJH´ E\ banking lobby groups which take advantage of deviations between IFRS and US-GAAP. Further, our study contributes to the literature by examining the consequences of a decrease in fair value measurement. Due to the regulatory environment that has increasingly extended the scope of fair value notes disclosures and/or measurement, prior research has exclusively studied potential benefits of increases in fair value notes disclosures (e.g., Nelson, 1996; Barth, Beaver, and Landsman, 1996) or recognition requirements for certain financial instruments (e.g. Ahmed, Kilic, and Lobo, 2006; Hodder, Hopkins, and Wahlen, 2006). In contrast, this study exploits the reclassification amendments to IAS 39 and IFRS 7 and the resulting reclassifications as a unique setting which allows us to analyze the potential costs of decreases in fair value recognition or even notes disclosures.72 Finally, we contribute by exploiting hand-collected and comprehensive data on fair value measurement, notes disclosures and linkages of fair values to national regulatory capital requirements of IFRS banks around the world. Such data is not available in standard databases, which is presumably the reason why there is still little international evidence on the topic (Chang, Liu, and Ryan, 2009), despite the fact that the reliability concerns with fair values should be more serious in countries outside the US where enforcement mechanisms are weak (e.g., Armstrong, Barth, Jagolinzer, and Riedl, 2010; Daske, Hail, Leuz, and Verdi, 2008). We introduce this dataset into the literature on the lessons learned and the regulatory consequences to be drawn from the global financial crisis which, so far, is almost exclusively based on evidence from the US environment (e.g. Bowen, Khan, and Rajgopal, 2009; Huizinga and Laeven, 2009; Song, Thomas, and Yi, 2010). The remainder of the essay is organized as follows. In section 4.2, we describe the institutional background. In section 4.3, we review related literature and develop testable hypotheses. Section 4.4 describes our empirical strategy, data and results. We provide concluding remarks in section 4.5. Section 4.6 contains the tables.

72

For example, it is far from clear whether (the magnitude of) the effects are simply opposite to the effects of first-time recognition of fair values, since the valuation techniques are already established and, thus, formerly recognized fair values that are still disclosed in the footnotes after reclassifications should produce information of similar reliability. 87

4.2 Institutional background 4.2.1 Events preceding the reclassification amendments The global financial crisis deteriorated continuously during 2008. Banks around the world were faced with substantial fair value write-downs on their trading and available for sale portfolios. Although a similar development could be observed in almost any developed market, the accounting consequences differed between US-GAAP and IFRS. In the US, the situation has been interpreted as one of the rare circumstances under which SFAS 115 (for securities) and SFAS 65 (for mortgage loans) offer an option to transfer financial assets from the trading or available for sale (held for sale) category into the held to maturity category (i.e. to abandon fair value measurement of the respective assets). Anecdotal evidence indicates that a few prominent US banks did indeed make use of this opportunity (Laux and Leuz, 2010).73 In contrast, banks reporting under IFRS could not avoid the reporting of fair value losses from trading assets that might have lost their PDUNHWDELOLW\ GXULQJWKHFULVLVEHFDXVH,$6 XQH[FHSWLRQDOO\UHTXLUHGWKDW³DQHQWLW\ shall not reclassify a financial instrument into or out of the fair value through profit or loss category while it is helGRULVVXHG´SDUD  Commercial banks outside the US identified this restrictive rule, which forced them to report, ceteris paribus, lower profits and lower regulatory capital than US competitors exercising their options, as a severe disadvantage in global capital markets, particularly during the financial crisis. Bank representatives intensely lobbied for the introduction of a reclassification option into IAS 39. In the public domain, the Institute of International Finance (IIF), an organization of international banks, asked the IASB in May 2008 to H[DPLQH ³VXJJHVWLRQV IRU HQKDQFHG YDOXDWLRQ PHWKRGRORJLHV LQ GLVORFDWHG PDUNHW conditions, or suggestions that would allow assets to be reclassified from trading to other categories in accordance with defineG FRQGLWLRQV LI PDQDJHPHQW¶V MXGJPHQW LV WKDW WKH WUDGLQJ FODVVLILFDWLRQ LV QR ORQJHU DSSURSULDWH´ ,,)   ,,) FKDLUPDQ -RVHI Ackermann, at the same time head of the IFRS reporting Deutsche Bank, underlined in a

73

88

According to its 10-K filing (footnote 16), Citigroup, which had rarely used the held to maturity category before, transferred trading assets at a carrying value of USD 33.3bn and available for sale securities at a carrying value of USD 27.0bn into that category during the fourth quarter of the financial year 2008. TKHFRPSDQ\DUJXHVWKDWWKHVHWUDQVIHUV³EHWWHUUHIOHFWWKHUHYLVHGLQWHQWLRQV«LQ UHVSRQVH WR WKH UHFHQW VLJQLILFDQW GHWHULRUDWLRQ LQ PDUNHW FRQGLWLRQV´ &LWLJURXS WKXV IRUZHQW additional fair value write-downs of approximately USD 4.1bn in 2008.

SUHVV FRQIHUHQFH WKDW LW ZRXOG EH ³YHU\ LPSRUWDQW WR KDYH PRUH FRQYHUJHQFH´ EHWZHHQ US GAAP and IFRS, in order to create a level playing field for international banks .74 In some European countries which have adopted IFRS, particularly in Continental European France, Germany and Italy, the resistance of bank lobbyists towards fair value accounting had gained the support of top-level politicians. Concerns over the reliability and procyclicality of fair value accounting have been the subject of discussions at regular meetings of the EU¶V(FRQRPLFVand Finance Ministers (Economic and Financial Affairs Council, ECOFIN) since November 2007 (ECOFIN, 2007; ECOFIN, 2008a). The skepticism towards fair value accounting might not least be due to the concerns which bank regulators had articulated for years (e.g., ECB, 2004). In the current crisis, regulators have tended to blame fair value accounting for having contributed to the banking turmoil. For example, both the Basel Bank for International Settlements (BIS) and the Financial Stability Forum, two organizations which represent, among others, financial supervisory institutions, later recommended that international accounting VWDQGDUG VHWWHUV JHQHUDOO\ DOWHU WKH DFFRXQWLQJ UXOHV ³VR WKDW WKH XVH RI IDLU YDOXH accounting is more limited for financial instrumeQWV´ %,6  )LQDQFLDO 6WDELOLW\ Forum, 2009). Initially, the IASB strongly opposed any reduction to the scope of fair value measurement in IAS 39. As a direct response to criticism raised at the ECOFIN meeting in July 2008, IASB Chairman Sir David Tweedie ruled out any rewrite of the then H[LVWLQJ IDLU YDOXH UXOHV DQG HPSKDVL]HG WKDW ³ZH DUH FHUWDLQO\ QRW WKLQNLQJ RI DQ\ HPHUJHQF\ PHDVXUHV WR FKDQJH ZKDW ZH GR DW SUHVHQW´ 75. In early September 2008, the )UHQFKERDUGPHPEHU3KLOLSSH'DQMRXVWDWHGWKDW³Where is no support within the banking FRPPXQLW\ WR FKDQJH WKH >IDLU YDOXH@ UXOHV´76. When the IASB issued a draft of guidelines on fair value measurement in illiquid markets on September 16, one day after the Lehman breakdown, there were still no indications for any short-term changes to IAS 39. The plan was rather to issue a completely new fair value standard in the second half of 2009 (IASB, 2008a). The Lehman breakdown, however, entirely changed the political situation of the IASB. When international banks faced the risk of reporting substantial losses and due to IFRS accounting numbers being used in the determination of regulatory capital a severe 74

Financial Times, Banks Told to Tackle Risk and Pay. July 18, 2008.

75

Reuters News, David Tweedie: Big rewrite of accounting fair value ruled out, July 10, 2008.

76

Reuters News, Philippe Danjou: Reporting rules should not favor bank cushions, September 11, 2008. 89

decrease in their capital ratios for the third quarter of the financial year 2008, politicians, particularly from the EU, substantially increased the pressure on the accounting body to alter IAS 39 in a way that allowed banks to forgo the recognition of those losses. At the summit of European G8 members (France, Germany, Italy, UK) on October 4 in Paris, the FrenFK 3UHVLGHQW 1LFRODV 6DUNR]\ WRRN XS WKH ,,)¶V QRWLRQ RI WKH FRPSHWLWLYH GLVDGYDQWDJHRI,)56DGRSWLQJEDQNVDQGDQQRXQFHGWKDW³ZHZLOOHQVXUHWKDW(XURSHDQ financial institutions are not disadvantaged vis-à-vis their international competitors in terms of accounting rules and of their interpretation. In this regard, European financial institutions should be given the same rules to reclassify financial instruments from the WUDGLQJERRNWRWKHEDQNLQJERRNLQFOXGLQJWKRVHDOUHDG\KHOGRULVVXHG´(XURSHDQ* Members, 2008). During the next week, the ECOFIN and the European Commission UHVSRQGHG WR WKLV UHTXHVW E\ WKH KHDGV RI WKH (8¶V ODUJHVW PHPEHU VWDWHV DQG WRJHWKHU EURZEHDWWKH,$6%2Q2FWREHU(&2),1PHWLQ/X[HPERXUJWRGHFLGHWKDWWKH³LVVXH of asset UHFODVVLILFDWLRQPXVWEHUHVROYHGTXLFNO\«ZHH[SHFWWKLVLVVXHWREHVROYHGE\ the end of the month, with the objective to implement as of the third quarter, in DFFRUGDQFH ZLWK WKH UHOHYDQW SURFHGXUHV´ (&2),1 E  2Q 2FWREHU  (8 Commissioner Charlie McCreevy gave a speech in the European Parliament and DQQRXQFHGWKDWWKH&RPPLVVLRQKDGSUHSDUHGWKHOHJLVODWLRQWRKDYHD³FDUYH-RXW´IURP IAS 39 and was ready to adopt its own EU version of IAS 39 which would allow the reclassification of financial assets but not introduce any disclosure requirements, if the IASB did not alter the accounting standard on its own. &KDLUPDQ7ZHHGLHODWHUIUDQNO\DGPLWWHGWKDWWKH,$6%LQWHUSUHWHGWKH(8¶VDFWLRQDV D ³EOXQW WKUHDW WR EORZ WKH RUJDQL]DWLRQ DZD\´ LI WKH ERard did not comply with the (XURSHDQ&RPPLVVLRQ¶VUHTXHVWV7KHVWDQGDUGVHWWHUZDVWDNHQE\VXUSULVHDQG³GLGQRW KDYHDQLQNOLQJRIWKLVFRPLQJXQWLOSUREDEO\DERXWDIRUWQLJKWEHIRUHKDQG´+RXVHRI Commons, 2008). The IASB therefore rapidly gave in primarily to prevent a standard that would give European firms the opportunity to abandon fair value measurement without any accompanying disclosures, but also in a desperate attempt to protect the ongoing process towards achieving a single global accounting language. On October 9, one day after the McCreevy speech, the IASCF trustees decided in a public meeting that the IASB was allowed to suspend the normal due process to avoid lengthy consultations before adopting fair value reclassifications. This decision made it possible to finalize an amendment to IAS 39 at the next regular board meeting held on October 13. When the board met that day, an amendment to IAS 39 allowing reclassifications of fair value assets was indeed adopted with the approval of eleven of the thirteen IASB members 90

(with dissent by James Leisenring and John Smith). In a press release, the Board stressed WKDWWKLVDFWLRQZDV³FRQVLVWHQWZLWKWKHUHTXHVW PDGHE\ (XURSHDQOHDGHUVDQGILQDQFH PLQLVWHUV´ WKXV KLJKOLJKWLQJ WKH SROLWLFDO SUHVVXUe under which the amendment came about, which had seemed highly unlikely just four weeks earlier. Only two days later, the EU Commission officially endorsed the amendments. Even though they were adopted in response to European lobbying activities, other jurisdictions where formal endorsement procedures for IFRS had been established also accepted the reclassification rules (October 15, Hong Kong; October 17, Taiwan; October 22, Australia; October 23, Philippines; October 24, South Africa; October 30, Singapore). 4.2.2 The reclassification amendments 4.2.2.1 Amendments to IAS 39: recognition and measurement The measurement of financial instruments is within the scope of IAS 39, which introduces three different measurement bases: fair value through profit or loss, fair value through OCI, and amortized cost (e.g., Spooner, 2007). Trading assets, among them all financial derivatives not designated for hedge accounting, are unexceptionally measured at fair value through profit or loss. Fair value measurement through profit or loss can also be applied if the fair value option is chosen for an instrument. Amortized cost, on the other hand, is the relevant base for the measurement of financial assets categorized as either loans & receivables (L&R) or, in the case of marketable debt securities, held to maturity (HTM). Financial assets, for which none of these categories is used, are designated as available for sale (AFS) and measured at fair value through OCI. The AFS category of IAS 39 is applicable for both securities (equivalent to the corresponding category of SFAS 115) and loans. The categorization is required at the initial recognition of any financial asset. After initial recognition, there are five types of possible reclassifications of assets measured at fair value. The original IAS 39 had allowed the reclassification out of the AFS category into the HTM category only (para. 54). The amendment from October 2008 introduces four additional types of reclassifications. Since then, trading assets can be reclassified into either the AFS, HTM, or the L&R category and AFS assets can be reclassified into the L&R category. It is, however, not allowed to reclassify any assets for ZKLFK ,$6 ¶V IDLU YDOXH RSWLRQ LV XVHG QRU LV LW SRVVLEOH WR UHFODVVLI\ D ILQDQFLDO derivative. All five typHVRIUHFODVVLILFDWLRQVFDQRQO\EHDSSOLHGLQ³UDUHFLUFXPVWDQFHV´ (paras. 50B, 54), but they differ in their accounting consequences. Overall, three different 91

effects on measurement can be distinguished. First, reclassifications out of the trading account into the HTM or L&R category affect both net income and equity because fair value gains and losses cease to be recognized in profit or loss and, thus, in equity. Second, reclassifications out of the trading account into the AFS category affect only net income and not equity because fair value gains and losses are still considered in the UHYDOXDWLRQ UHVHUYH DV SDUW RI DQ HQWLW\¶V HTXLW\ EXW WKH\ DUH QRZ VKRZQLQ 2&, UDWKHU than in profit or loss. Third, reclassifications out of the AFS category into the L&R or HTM category affect only OCI (equity) and not net income because fair value gains and losses have previously not been considered in the income statement, but only in the OCI DQGWKXVWKHUHYDOXDWLRQUHVHUYHDVSDUWRIDQHQWLW\¶VHTXLW\ The newly introduced rules differ from their US-GAAP counterparts in two important respects. First, all IFRS adopting banks are allowed to apply the relaxed rules retroactively from any date between 1 July and 1 November 2008, i.e. a reclassification could not only be used to avoid expected future fair value write-downs but also to reverse write-downs already been taken for this last quarter. Second, the rules for the recognition of other-than-temporary impairment losses that were to be applied after reclassification into cost categories are substantially less restrictive under IFRS (both for the HTM and the L&R category).77 Therefore, the fair value write-downs a bank could avoid by means of reclassifications are, ceteris paribus, potentially larger than under SFAS 65 and 115. These two reasons might explain why Citigroup is the only major US bank that considerable made use of its options (Laux and Leuz, 2010). 4.2.2.2 Amendments to IFRS 7: disclosures The IASB primarily responded to the political pressure by its own version of a reclassification amendment in order to avoid an EU-specific carve-out from IAS 39, thus ensuring the simultaneous introduction of disclosures about the use of the option (House of Commons, 2008). These extensive disclosure requirements have been introduced into IFRS 7. The disclosures comprise quantitative information about (1) the amount reclassified into and out of each measurement category, (2) the current fair values of each asset that was previously reclassified, (3) the fair value gain or loss in net income or OCI in the period of the reclassification, (4) the fair value gain or loss after reclassification that would have been reported if the asset had not been reclassified, (5) the effective 77

92

For this very reason, the two dissenting IASB members expressed concerns that a true level playing field between IFRS and US-GAAP was not achieved (IASB, 2008b).

interest rate and the estimated amounts of cash flows expected to be recovered as at the date of reclassification, as well as qualitative information about (6) the facts and circumstances of the rare situation that gave rise to the reclassifications (para. 12A). 4.2.2.3 Consequences for the regulatory capital of banks To the extent banking oversight and regulatory capital are linked to financial reporting, reclassifications may also affect the regulatory capital which banks report to national supervisory authorities. The effect depends on country-specific regulation, and we detect and document in table 4.6.1 substantial cross-sectional variation in the link between IFRS accounting numbers and prudential filters on the inclusion of unrealized fair value gains and losses into regulatory capital. In almost any jurisdiction, unrealized gains and losses from trading account securities are fully reflected in tier 1 capital. Reclassifications out RI WKH WUDGLQJ DFFRXQW WKXV DOZD\V DIIHFW D EDQN¶V WLHU  FDSLWDO FRUUHVSRQGLQJO\ RQH WR one. The major differences across countries arise from the treatment of unrealized gains from AFS securities which are recognized in the revaluation reserves (via OCI). While unrealized losses (after tax) are fully deducted from tier 1 capital, the Basel II Framework recommends the inclusion of unrealized fair value gains into tier 2 capital with a general discount (haircut) of 55% to reflect both the risk of market illiquidity and the future tax charge (Art. 49(v)). However, supervisory practice across the world differs substantially in the implementation of this guideline. In most jurisdictions, the discount is different for equity securities, debt securities and loans. In our 39 sample countries, an average haircut of 48.23% is applicable for unrealized gains on equity securities (including deductions for tax effects). Gains and losses on loans categorized as AFS are, on the other hand, fully neutralized in the determination of regulatory capital. Debt securities are treated like AFS equity securities in 27 countries and like AFS loans in 12 countries (see CEBS (2007) for an overview of EU member states). Unfortunately, disclosures on financial asset types or portfolio compositions which are subject to reclassifications are not required by IFRS 7. As our best proxy, we therefore report in table 4.6.1 the discount on debt securities for the estimation of the effect of the reclassifications on regulatory capital because evidence from voluntary disclosures by our sample banks suggests that the

93

majority of reclassified assets are debt securities. The discount on debt securities varies between 0% and 100% across countries.78 While most jurisdictions apply a net approach in determining the adjustments to the revaluation reserve (i.e. the prudential adjustment is based on the aggregate revaluation reserve), a few regulators (Netherlands, Portugal, Slovakia, Slovenia) require the determination on an item-by-item basis (i.e. the prudential adjustment is determined for each AFS asset individually). Again, IFRS 7 disclosures do not allow us to take this regulatory variation into account and, since regulatory reports are not publicly available in most jurisdictions outside the US, we need to approximate the regulatory effect of reclassifications from or into the AFS category by using the net approach for countries where the item-by-item approach is applied. Finally, jurisdictions differ in the required minimum capital ratio which defines a EDQN¶V SUR[LPLW\ WR FDSLWDO UHVWULFWLRQV EHIRUH DQG DIWHU UHFODVVLILFDWLRQV 7KH %DVHO ,, Framework recommends a total capital ratio (tier 1 plus tier 2) of at least 8% (Art. 40). Most countries follow this recommendation, while a few countries require a ratio of 10% or 12% (see table 4.6.1 for details). 4.2.3 Initial reactions to the amendment decision 7KH ,$6%¶V GHFLVLRQ LQ 2FWREHU  KDV EHHQ D VXEMHFW RI LQWHQVH SXEOLF GHbate. Unsurprisingly, bank representatives, e.g. the European Banking Federation (EBF), and (XURSHDQ UHJXODWRUVDUHVRPHRIWKHVWURQJHVWVXSSRUWHUVRIWKHERDUG¶VGHFLVLRQ 79 On the other hand, users of financial statements harshly criticize the new rules. The Corporate Reporting Users Forum, a pan-European group of investors and analysts, argues in a public letter to the Financial Times that the new reclassification option SRWHQWLDOO\ XQGHUPLQHV LQYHVWRUV¶ FRQILGHQFH LQ EDQN GLVFORVXUHV80 This concern finds VXSSRUWLQWKHFRPPHQWOHWWHUZKLFKWKH&)$,QVWLWXWHDGGUHVVHGWRWKH(8¶V$FFRXQWLQJ 78

For uniformity, we use the total discount (on an after-tax basis) in our analysis, i.e. the combination of the deduction for tax effects and the additional haircut under national banking laws. See table 4.6.1 for the two components.

79

7KH (%) ZHOFRPHV WKH DPHQGPHQW LQ D SUHVV UHOHDVH VWDWLQJ WKDW LW QRZ H[SHFWV ³PRUH UHOHYDQW information LQ WKH FRQWH[W RI LOOLTXLG PDUNHWV´ 7KH &RPPLWWHH RI (XURSHDQ %DQNLQJ 6XSHUYLVRUV (CEBS) refers to the convergence argument brought up by European politicians in a joint statement with the Committee of European Securities Regulators (CESR) and declares that European companies QRZKDYH³DOHYHOSOD\LQJILHOG«ERWKFURVV-VHFWRUDOO\DQGZLWKSUHSDUHUVIURPRWKHUMXULVGLFWLRQV´ (CEBS, 2008).

80

Financial Times, Accounting rule changes could dent confidence, say analysts, October 21, 2008.

94

Regulatory Committee as a response to the endorsement decision, expecting that the IAS  DPHQGPHQW ZRXOG ³SUHFLSLWDWH LQYHVWRU ORVV RI FRQILGHQFH´ &)$ ,QVWLWXWH 2008). Interestingly, comments in the business press are mostly negative even in countries where politicians strongly lobbied for the new rules. In France, for example, Les Echos¶ KHDGOLQH LV ³1H FDVVRQV SDV OH WKHUPRPqWUH´81 and La Tribune criticizes the political influence on the standard setter.82 Part of the criticism is due to concerns that global convergence of accounting standards is decreasing rather than increasing as a result of the IAS 39 amendment. The first reason for this concern is factual because impairment rules for reclassified assets measured at amortized cost are less restrictive under IFRS than under US-GAAP, i.e. it is easier for IFRS adopting banks to avoid additional write-downs once a financial asset is reclassified. The amendment therefore, does not lead to true convergence. The second reason is political, because the history of the amendment has demonstrated that the EU has, in fact, substantial power over the IASB. This experience casts serious doubt on the US accepting or even introducing IFRS for national companies in the near future. The amendment therefore, calls future convergence activities into question83 and the contradicting strategies of the FASB and IASB on fair value accounting in the aftermath of financial crises seem to support these concerns.84

81

(Do not break the thermometer), Les Echos, Ne cassons pas le thermomètre, October 10, 2008.

82

³&HWWH SROLWLVDWLRQ Q¶HVW SDV GH QDWXUH j UDVVXUHU FHX[ TXL UHGRXWHQW TXH FH FKDQJHPHQW HQWUDvQH XQ VXUFURvWG¶RSDFLWp´7KLVSROLWLFL]DWLRQGRHVQRWGLVDUPWKRVHZKRIHDUWKDW the amendment results in increasing opacity), La Tribune, La Commission rend sa copie sur les normes comptables, October 16, 2008.

83

Getting straight to the point, a financial analyst commented in the Financial Times (October 30, 2008): ³7KH DPHQGPHQW ZDs done without due process: no comments or discussion were sought from investors. Incredibly, that suspension of due process bore the blessing of the trustees who oversee the IASB. Seeking convergence on a tiny part of the rules in this regard may seem trivial, but it is a giant step backwards. It shows that when the going got tough, the IASB waived the interests of those - the users of accounts - whom it is supposed to serve. Instead of converging to what is best in US standards, the IASB is adopting some RIWKHZRUVWIHDWXUHV´

84

The IASB keeps amortized cost accounting at least for certain loans (IFRS 9), whereas the FASB proposes a full fair value model for financial instruments (ED ASU Topics 815 and 825), see also The Economist, To FASB or not to FASB, June 10, 2010. 95

4.3 Hypotheses development and related literature 4.3.1 Short-term benefits of the reclassification choice during the financial crisis 7KHDQDO\VLVRIEDQNV¶,)56GLVFORVXUHVDERXWWKHFLUFXPVWDQFHVWKDWJDYHULVHWRWKH reclassifications (para. 12A (c)) reveals that almost all banks provide cursory and XQVSHFLILF³ERLOHUSODWH´UHDVRQVLIDQ\DWDOOIRULWVDFFRXQWLQJFKRLFHDQGPHUHO\UHIHU to the general market turmoil during 2008. Therefore, our hypotheses development relies on positive accounting theory (e.g., Holthausen and Leftwich, 1983; Watts and Zimmerman, 1986) which offers a host of alternative explanations for the reclassification choice. Most importantly, fair value reclassifications of illiquid financial assets potentially inFUHDVHQHWLQFRPHDQGVKDUHKROGHUV¶HTXLW\LQFOXGLQJUHYDOXDWLRQUHVHUYHV LQWLPHVRI falling market prices. Thus, management incentives for fair value reclassifications may result from regulatory capital restrictions that are tied to accounting measures. In times of the worst financial crisis for decades, this appears to be a likely first order effect. 0DQDJHPHQW¶VLQFHQWLYHWRPDLQWDLQUHJXODWRU\FDSLWDOUHVWULFWLRQVPDLQO\VWHPVIURPWKH risk of regulatory costs that would be incurred if restrictions were violated and supervisory actions were taken. These regulatory measures can take various forms, up to a forced closure of the bank, may vary across countries, but always result in a substantial loss of shareholder value. For example, Jordan, Peek, and Rosengren (2000) provide evidence that share prices of US banks drop, on average, by about 5% around the announcement of formal supervisory actions against the bank. In banking crises, discretion offered by or introduced into accounting standards is an implicit means of regulatory arbitrage to satisfy the necessary thresholds, thus granting banks the avoidance of these costs without explicitly exposing this fact to the public (Skinner, 2008). Taken together, there is broad prior evidence that commercial banks manage their capital ratios (e.g. Beatty, Chamberlain, and Magliolo, 1995; Ramesh and Revsine, 2001). We conjecture that if a bank faces the risk of regulatory costs, i.e. is weakly capitalized, it is likely that it will use fair value reclassifications to increase its regulatory capital during the financial crisis. Even though the Basel Accords aim to harmonize global banking regulation, there are still notable institutional differences in the determination of regulatory capital across countries (see table 4.6.1 for details). The most relevant difference in our setting, aside from the level of the minimum capital adequacy ratio, is the inclusion of positive 96

revaluation reserves into tier 2 capital. If an AFS asset is reclassified, the revaluation reserve is frozen at the level of the reclassification date and the extent to which the reserve will be included in tier 2 capital is safeguarded against future decreases in fair value (at least if the fair value decreases are not causing simultaneous impairment writedowns). The lower the haircut on unrealized gains from AFS securities (see section 4.2.2 for details), the more regulatory capital is therefore at risk when a bank expects a decrease of the fair value of its AFS assets. Thus, we hypothesize that the haircut on the revaluation reserves adds to the probability that a bank will apply the reclassification option for regulatory capital arbitrage in order to satisfy necessary capital thresholds. We expect a bank to also consider expected political costs in its reclassification decision, even though this link to the reclassification choice is more tentative than any of the other links. In short, the reclassification amendment was the result of political pressure by governments (particularly from Continental Europe) put on the IASB, which can be traced back to intense lobbying activities by banks (see section 4.2.1). Although prior evidence on the political cost hypothesis (Watts and Zimmerman, 1978, 1990; Holthausen and Leftwich, 1983) suggests that earnings are managed downward to avoid costly political or regulatory interventions (e.g., Ramanna and Roychowdhury, 2010), matters were quite different for banks during the 2008 financial crisis. In this particular setting, political incentives may rather act in favor of reclassifications, i.e. of upward management of earnings, for two reasons. First, their stabilizing effect on regulatory capital helps to avoid adverse political actions that are associated with governmental bailouts. Since bail-outs are highly unpopular among the majority of voters, they are accompanied by tight restrictions on the business of a bank and, maybe most importantly, PDQDJHPHQW¶VFRPSHQVDWLRQVFKHPHV85 Second, if banks did not make sufficient use of the newly introduced option they intensely lobbied for, it would most probably increase the costs of opposing adverse political actions and decrease the likelihood of the success of future lobbying activities for certain accounting rules. This prediction is consistent with Ramanna (2008), who GRFXPHQWV WKDW WKH LQWHQVLW\ RI D ILUP¶V OREE\LQJ DFWLYLWLHV for goodwill impairment rules prior to the adoption of SFAS 142 was positively DVVRFLDWHGZLWKWKHILUPV¶SRVW-adoption use of accounting discretion offered by the new rules. The firm-specific size of the political incentives, therefore, are likely to depend on the extent to which the national government has been involved in the lobbying process and how much public attention the bank received during that process. 85

See, e.g., Financial Times, How to bypass populism and tackle banking, January 26, 2010. 97

In addition to these firm-specific considerations, governments might have their own LQFHQWLYHV WR LPSDFW D EDQN¶V UHFODVVLILFDWLRQ GHFLVLRQ )LUVW WKH LQFUHDVH RI EDQNV¶ QHW income and/or shareholders equity during the financial crisis can be viewed as a stabilization of the financial system, thus reducing costs eventually borne by taxpayers from a continued destabilization and its adverse economy-wide effects. Second, governments in several countries directly injected tax money into large and systemrelevant banks, which would have been lost if a bank had been closed after failing to PHHW FDSLWDO UHTXLUHPHQWV 7KHUHIRUH ZH K\SRWKHVL]H WKDW WKH ODUJHU D JRYHUQPHQW¶V equity stake, the higher the probability that benefits from a stabilization of the financial V\VWHPDUHFRQVLGHUHGLQDEDQN¶s reclassification decision. *RYHUQPHQW LQFHQWLYHV DUH DOVR FORVHO\ DOLJQHG ZLWK D EDQN¶V LQFHQWLYH WR UHSRUW DW least zero earnings (i.e. to avoid the disclosure of an accounting loss). The zero earnings threshold is particularly important in the banking industry where the sensitivity of FXVWRPHUV¶GHPDQGDQGWLPHGHSRVLWVWRDEDQN¶VILQDQFLDOFRQGLWLRQVLQKHULWVWKHULVNRI bank runs (e.g., Spiegel and Yamori, 2007; Goldberg and Hudgins, 2002) and where the zero earnings threshold has been shown to be of psychological importance for private depositors (Shen and Chih, 2005).86 Since bank failures are not only costly for immediate stakeholders in that bank, but also tend to cause contagion of other banks or even of other economic sectors, the political pressure for the reclassification option can also be interpreted as the intention of national governments to convince the public that banks were healthy. We therefore conjecture that a possibility to pass the zero earnings threshold by means of reclassifications is positively associated with the probability of a EDQNXVLQJWKLVRSWLRQ7KLVFKRLFHLVOLNHO\WRLQWHUDFWZLWKWKHGHYHORSPHQWRIDEDQN¶V deposits. A bank that already faces a decrease in deposits will be even less inclined to report a loss that coulGSRWHQWLDOO\IXUWKHUXQGHUPLQHLWVGHSRVLWRUV¶FRQILGHQFH The relevance of the zero earnings threshold for measures of regulatory forbearance relies on the assumption of functional fixation by depositors. Since depositors are, on average, lay investors, it is plausible that they do not distinguish between income before and after reclassifications. In contrast, it is less plausible that reclassifications are also XVHGIRUSDVVLQJFHUWDLQHDUQLQJVWKUHVKROGVLQRUGHUWRERRVWWKHEDQN¶VVWRFNSULFHVHH e.g., Degeorge, Patel, and Zeckhauser, 1999; Barth, Elliott, and Finn, 1999; Bartov, 86

98

Note that this relevance may also hold when there exists deposit insurance or a widespread belief in governmental interventions resulting from a too-big-to-fail policy exist because even these measures create transaction costs for customers and might not avoid temporary frictions during which access to deposits is restricted.

Givoly, and Hayn, 2002). Professional investors and financial analysts are substantially less prone to functional fixation than depositors and are likely to use IFRS 7 reconcLOLDWLRQ GLVFORVXUHV LQ WKH IRRWQRWHV WR DGMXVW D FRPSDQ\¶V QHW LQFRPH IRU reclassification effects.87 There is plenty of evidence that market behavior is inconsistent with functional fixation, at least for substantial and less complex accounting choices (Kothari, 2001; Plumlee, 2003). Therefore, contrary to prior findings about the LPSRUWDQFH RI DQDO\VWV¶ IRUHFDVWV WDUJHWV LQ JHQHUDO HJ, Degeorge, Patel, and Zeckhauser, 1999; Bartov, Givoly, and Hain, 2002), it is questionable whether the proximity of pre-reclassification earnings to consensus EPS forecast targets causally H[SODLQVDEDQN¶VUHFODVVLILFDWLRQFKRLFH88 Our first hypothesis can be summarized as follows: (H1a) The probability of a bank using the reclassification option is positively associated with the size of the bank-specific benefits from avoiding (a) the violation of regulatory capital restrictions, (b) political costs and (c) a loss of depositors. $VVXPLQJ WKDW WKH PDUNHW DQWLFLSDWHV WKH HFRQRPLF EHQHILWV DQG FRVWV  RI D EDQN¶V reclassification choice, the stock price reaction around the event dates should be positively (negatively) associated with the size of those benefits (costs). With respect to WKHHYHQWGDWHVZHQHHGWRGLVWLQJXLVKEHWZHHQ WKH,$6%¶VUHJXODWRU\DQQRXQFement of the amendment to IAS 39 in October 2008 granting banks the reclassification option and (2) the announcement date of the bank-specific choice. While the latter date resolves the uncertainty about the accounting effects for a bank, investors will already have built XSWKHLURZQH[SHFWDWLRQVRIDEDQN¶VFKRLFHDWWKHGDWHRIWKHLQLWLDOUHJXODWRU\GHFLVLRQ Around the regulatory announcement, the abnormal returns of banks that are expected to use the reclassification option should, therefore, reflect the market expectation of any resulting net economic consequences. Recent evidence on fair value regulation in the US suggests an overall positive reaction of the stock market towards the perceived relaxation 87

Some anecdotal evidence can indeed be found in analyst reports and Q3 and Q4 2008 conference call transcripts. For example, in a JP Morgan report about the German Aareal Bank, the first key point on SDJH  VWDWHV WKDW WKH LQFUHDVH LQ QHW LQFRPH IRU WKH WKLUG TXDUWHU LV PDLQO\ ³GXH WR DVVHW UHFODVVLILFDWLRQV´ ,Q D VXEVHTXHQt conference call, the same analyst explicitly asks the Aareal Bank PDQDJHUV³&DQ\RXMXVWGHWDLOKRZ\RXUHFRQFLOHWKHK\SRWKHWLFDOHIIHFWRQWKH3 /RIWKH(85P on asset-EDFNHGVHFXULWLHVZKLFKFOHDUO\KDYHEHHQUHFODVVLILHG"´6LPLODUHYLGHQFHFDQEH provided for at least 29 reclassifying banks in our sample.

88

This latter variable could only come into play if management compensation were bound to achieving DQDO\VWV¶(36WDUJHWV+RZHYHUZHEHOLHYHWKDWWKLVFRPSHQVDWLRQLVVXHSOD\HGLIDQ\DPLQRr role in 2008 accounting choices due to the extraordinary circumstances of the financial crisis. 99

of fair value rules, e.g. the adoption of FSP FAS 157-3 in October 2008 (Huizinga and Laeven, 2009) or FSP FAS 157-4 in March/April 2009 by the FASB (Bowen, Khan, and Rajgopal, 2009), at least in times of the crisis. Returns around the bank-specific announcement of the reclassification choice provide an updated measure of the market assessment of the economic benefits and costs. In sum, we hypothesize: (H1b) The stock price reaction to the announcement of fair value reclassifications is positively associated with the size of the (potential) bank-specific benefits resulting from reclassifications. 4.3.2 Long-term costs of the reclassification choice (FRQRPLF WKHRU\ SULPDULO\ OLQNV WKH TXDOLW\ RI D ILUP¶V GLVFORVXUH DQG LQIRUPDWLRQ environment to its cost of capital via the adverse selection component of the bid-ask spread when market makers price-protect against informed traders (e.g., Stoll, 1978; Glosten and Harris, 1988; Diamond and Verrecchia, 1991). This theoretical link has been supported by a number of empirical studies (e.g., Leuz and Verrecchia, 2000; Muller and Riedl, 2002; Frankel and Li, 2004; Mohd, 2005; Verrecchia and Weber, 2006). These UHVXOWV LQGLFDWH WKDW QHJDWLYH HIIHFWV RI IDLU YDOXH UHFODVVLILFDWLRQV RQ D EDQN¶V information asymmetry and cost of capital depend, ceteris paribus, on whether the relaxation of fair value accounting is perceived as a decrease in disclosure quality. Overall, there is consensus that the value relevance of financial assets benefits from fair value disclosures (Barth, 1994; Barth, Beaver, and Landsman, 1996; Eccher, Ramesh, and Thiagarajan, 1996; see Wahlen et al., 2000, for an overview). Moreover, evidence from surveys conducted by the CFA Institute (2008) indicates that financial analysts consider fair value information to be useful. These findings suggest that the use of the reclassification option may signal a decrease in transparency, at least if a bank had previously built up some reputation for its high reporting quality and if the reclassifications affect the accounts in a material way. Stated differently, we predict a positive association between information asymmetry and the size of the effects of fair value reclassifications (e.g. on net income). Managers will anticipate those adverse effects of the reclassification choice and weigh off the corresponding increase in cost of equity capital and in trading costs against the benefits discussed in section 4.3.1. Therefore, we hypothesize: (H2a) The probability of a bank using the reclassification option is negatively associated with its commitment to transparency. 100

The disclosure requirements of IFRS 7 that are applicable in the case of fair value reclassifications ensure that fair value information is still available in the financial statement. Consequently, if disclosure requirements are fully complied with, the relaxation of fair value rules merely results in formerly recognized fair values now being disclosed in the footnotes. Since the different effects of recognition and disclosure are controversial (see Bernard and Schipper, 1994; Schipper, 2007, for a conceptual discussion), the existence of the capital market effects towards fair value reclassifications is not clear ex ante. Due to differences in the quality of enforcement and auditing institutions across countries, however, our international setting offers large heterogeneity in the degree of compliance with the IFRS 7 disclosure requirements. We exploit this variation in the quality of footnote disclosures to distinguish between banks which reclassify and fully disclose the effects and banks which reclassify but do not fully disclose. For the latter group of banks, theory suggests that we should observe an increase in information asymmetry and, thus, adverse selection. Our hypothesis can be summarized as follows: (H2b) The use of the reclassification option LV DVVRFLDWHG ZLWK DQ LQFUHDVH LQ D EDQN¶V information asymmetry component of the bid-ask spread. The size of the increase is negatively associated with the quality of accompanying footnote disclosures. 4.4 Empirical analysis 4.4.1 Sample selection and data sources BvD Bankscope is the primary source and starting point for our sample selection. In the first step, all 2,254 banks with publicly listed equity shares are selected (LISTINS). 594 banks from this Bankscope universe can be identified as IFRS adopters (ACCSTAND)89 for the financial year 2008. Next, we augment and verify the global representativeness of this sample by consulting Worldscope and Compustat Global as additional references. We identify and incorporate another 108 banks (Industry Groups 4310 and 4320) reporting under IFRS through these sources. This combined superset is 89

Consistent with the approach in Daske, Hail, Leuz, and Verdi (2009), we check and modify the Bankscope ACCSTAND coding in two respects: First, we treat banks from Taiwan as IFRS adopters even if they are classified as Local GAAP adopters, because Taiwanese SFAS 34 and 36 largely correspond to IAS 39 and IFRS 7, respectively; both standards have been effective since 2006 and the reclassification amendments were endorsed immediately on October 17, 2008. Second, we change the classification of banks from Malaysia from IFRS to Local GAAP, because the Malaysian Accounting Standards Board has decided that FRS 139 and FRS 7, which are the equivalent standards to IAS 39 and IFRS 7, were not effective before 2010. 101

matched with capital market data from Thomson Reuters Datastream; 264 banks are excluded due to missing or incomplete market data for the periods of interest between October 2008 and June 2009 (for bid-ask spreads) and between January 2008 and March 2009 (for stock returns). Thus, our initial sample comprises 438 banks. Even though IAS 39 is not an industryspecific standard and the reclassification option can be used by any company applying ,)56ZHUHVWULFWRXUDQDO\VLVWREDQNVIRUWZRUHDVRQV)LUVWPRUHWKDQRIDEDQN¶V balance sheet typically consists of financial instruments. The banking industry is, therefore, most strongly affected by IAS 39.90 Second, the effects of the reclassification amendments on accounting choices are likely to be idiosyncratic in the banking industry because its capital regulation is fundamentally different from other industries (even from the insurance industry). Therefore, the inclusion of non-banking firms in our sample would pose the severe issue of largely inhomogeneous accounting incentives. For this second reason, we exclude 112 institutions from our final sample which are not subject to an external capital oversight (hedge funds, brokerage houses, and securities firms which are listed as banks in Bankscope), for which we cannot retrieve any data on regulatory capital (neither manually from the financial report nor via our three databases) or that do not determine regulatory capital on the basis of their IFRS financial statements (due to country-specific options). The manual collection of financial statements from corporate websites reveals that 24 banks from our initial sample do not publish a financial report in English, French, German or Chinese on their websites. These banks are excluded due to practical impediments. Thus, our total sample comprises 302 banks from 39 countries. We use a variety of data sources for our empirical analyses. Capital market data is obtained from Thomson Reuters Datastream. Accounting data is obtained from Bankscope and Worldscope. While these two commercial databases provide information about the general asset and liabilities structure of each bank (e.g. cash, trading securities, investment securities, loans, non-financial assets), they do not offer any data on fair value reclassifications under IFRS. We therefore manually collect detailed information on reclassification choices and relevant disclosures from the footnotes to the annual financial statements and all previously filed quarterly financial statements for reporting periods which end between October 2008 and September 2009. In total, we evaluate 544 90

This is the prime reason why most prior studies on fair value accounting focus on the financial industry (see section 4.3). Moreover, there is recent evidence for IFRS firms that the magnitude of fair values outside the financial industry is rather small (Christensen and Nikolaev, 2009).

102

quarterly, interim and annual reports. In addition, we use Dow Jones Factiva as well as LexisNexis to identify thHH[DFWGDWHRIWKHLQLWLDOSXEOLFDQQRXQFHPHQWRIHDFKEDQN¶V reclassification choice. If a public announcement (e.g. in a separate press release or in combination with the general earnings release) is not available, we define the official filing date of the complete financial statement (containing the footnote disclosures on reclassifications) as the reclassification announcement. The resulting dataset includes (1) WKH FDOHQGDU GDWHV RI WKH ILUPV¶ DQQRXQFHPHQWV RQ ZKHWKHU WR WDNH WKH UHFODVVLILFDWLRQ option, (2) the reclassification amounts for each financial instrument category, (3) the reclassification effects on net income and other comprehensive income (revaluation reserves) for each financial instrument category and (4) information with regard to the type and quality of disclosures related to the reclassification choice. )XUWKHUWRVWXG\WKHHIIHFWRIUHFODVVLILFDWLRQVRQDEDQN¶VUHJXODWRU\FDSLWDOZHQHHG to collect data on country-specific capital regulation. We use the CEBS (2007) report on prudential filters for regulatory capital in European countries to update and broaden the information from the Barth, Caprio, and Levine (2001) World Bank dataset. Moreover, we contact bank regulators from each of the 39 countries represented in our sample (see table 4.6.1 for details) to verify our information on the supervisory rules governing the determination of regulatory capital. In particular, we document the proportion of the revaluation reserves, i.e. the net amount of unrealized fair value gains or losses on AFS DVVHWVWKDWLVLQFOXGHGLQDEDQN¶VWLHURUFDSLWDOIRUHDFKFRXQWU\VHHVHFWLRQ 4.2.2 for details).91 If a national regulator requires an additional haircut for future tax charges on unrealized gains, we refer to the OECD tax survey for the country-specific tax rate. This information allows us to determine precise country-specific capital adequacy ratios necessary to estimate firm-specific proxies for regulatory capital management incentives. 4.4.2 Descriptive evidence 4.4.2.1 Accounting effects of the reclassification amendments Table 4.6.1 presents details on the sample composition by country as well as selected country and bank variables. More than one third of our sample (124 banks altogether) chose to take the reclassification option during financial year 2008. This proportion is very similar to Fiechter (2010) but lower than the 61% as reported by CESR (2009) for 91

We also asked whether the amendments to IAS 39 had provoked any regulatory changes to the determination of regulatory capital. However, none of the responding authorities indicated that this was the case. 103

EU-banks only, with the difference being due to the different regional composition of our global sample. As described in section 4.2.2, there are three general types of reclassifications. We observe that 97 of the 124 reclassifying banks take the option for trading securities, while 72 banks reclassify AFS assets. Among the banks that reclassify trading securities, 40 institutions transfer assets into cost categories (HTM or L&R), 30 institutions transfer assets into the AFS category and 27 institutions transfer into both categories. These results are quite similar to CESR (2009), suggesting that the distribution of the reclassifications is not EU-specific once a bank takes the option. Table 4.6. VKRZV WKDW WKH HIIHFWV RI WKLV FKRLFH RQ WKH EDQNV¶ NH\ VXPPDUy accounting figures are quite substantial. The evidence indicates that the avoidance of the recognition of fair value decreases is only offset to a very limited extent by impairment write-downs on the assets now measured at amortized cost, suggesting that banks (and their auditors) treated the declines in value as temporary. On average, net income is EUR 182.96m or 43.8% higher after reclassifications. Earnings per share increase by an average of 0.57 Euros per share or 26.9%. The distribution of the effect is, however, highly skewed. Almost every bank increases its net income for the financial year. Only six banks experience a modest decline in net income as a result of the reclassifications, i.e. fair values of formerly reclassified assets exceed amortized costs at the balance sheet date.92 The majority of banks report moderate increases in their net income after reclassifications. For 47 banks this increase is less than 10% of net income before reclassifications. A few banks experience huge percentage increases of up to 1,100%, W\SLFDOO\ WKRVH VWDUWLQJ IURP D ORZ SURILW EDVH )RU H[DPSOH ,WDO\¶V %DQFD 3RSRODUH GL 6RQGULRUHSRUWVDSURILWRI(85PLQVWHDGRI(85P,QDEVROXWHWHUPVWKH8.¶V Royal Bank of Scotland reports the greatest increase in income of EUR 3,587.0m, IROORZHGE\6ZLW]HUODQG¶V8%6(85P (LJKWEDQNVLQWKHVDPSOHWXUQDORVV before reclassifications into a profit after reclassifications93 and 13 (11) banks would have

92

These negative effects are small in magnitude and relative to total net income for all six banks with 7XUNH\¶V 6HNHUEDQN (85 -2.67m) and the Union Bank of Taiwan (EUR -4.32m) reporting the highest reclassification losses. The effects are likely due to the retroactive application of the option which was limited until 1 November 2008, i.e. an (unexpected) increase in the fair value of reclassified assets between 1 November and the next reporting date (generally 31 December 2008) could not be recognized in profit or loss, thus resulting in a negative impact of the reclassifications on WKHSHULRG¶VQHWLQFRPH

93

These banks are Crédit Industriel et Commercial (France), Danske Bank, Sparekassen Faaborg (Denmark), JSC Rosbank (Russia), Pohjola Pankki (Finland), Boubyan Bank (Kuwait), Banca Generali and Mediobanca (Italy).

104

QRW H[FHHGHG RU DW OHDVW PHW WKH DQDO\VWV¶ PHGLDQ PHDQ  EPS forecast if they had not reclassified any financial assets. Since only net income is affected by the reclassification of trading securities, the effect RQ VKDUHKROGHUV¶ HTXLW\ ZKLFK DOVR LQFOXGHV WKH HIIHFWV IURP UHFODVVLILFDWLRQV RI $)6 securities, is even larger (EUR 287.07m on average). EUR 104.11m of this increase in VKDUHKROGHUV¶HTXLW\FDQEHDWWULEXWHGWRUHYDOXDWLRQUHVHUYHVWKHUHPDLQLQJSDUWUHVXOWV from increases in retained earnings, i.e. from net income. The effect on revaluation reserves is, however, more complex, because reclassifications out of the trading account into the AFS category and reclassifications out of the AFS category into cost categories (L&R or HTM) have opposite effects on those reserves. The 33% of the reclassifying banks in our sample that report a net decrease in revaluation reserves after reclassifications have simply shifted fair value losses from net income to OCI (thus, revaluation reserves) and as many as 25 banks have exclusively reclassified trading securities into the AFS category, i.e. safeguarded their net income (retained earnings) at the expense of their revaluation reserves with no effect on total equity. In our sample, )UDQFH¶V &UpGLW ,QGXVWULHO HW &RPPHUFLDO UHSRUWV WKH JUHDWHVW GHFUHDVH LQ UHYDOXDWLRQ reserves after reclassifications (EUR 829.8m). On the other hand, we also observe that several banks substantially increase their net revaluation reserves and, thus, their VKDUHKROGHUV¶HTXLW\E\UHFODVVLI\LQJRQO\$)6VHFXULWLHV This observation already provides some anecdotal evidence that the reclassification of AFS securities was primarily used to safeguard regulatory capital and to avoid regulatory interventions, dependent on the country-specific inclusion of fair value gains from AFS securities into the determination of tier 2 capital (see section 4.2.2). In terms of magnitude, the average impact of reclassifications on regulatory capital is, however, small when compared with the effects on income. The total capital ratio (tier 1 ratio) increases after reclassifications by an average of 24 (17) basis points or 2.61% (2.05%). Hence, it seems that enduring positive effects of fair value reclassifications on regulatory capital are confined to a relatively small number of banks, some of which, however, are system-UHOHYDQWLQWKHLUFRXQWULHV¶ILQDQFLDOV\VWHP94 There are at least two banks in the sample (Cyprus: Marfin Popular, Jordan: Ahli Bank) that would have had even more difficulty meeting the regulatory capital requirements at the end of the reporting period if they had not reclassified any financial assets. 94

We approximate that seven banks in our sample increase their regulatory capital (total capital ratio) by more than 100 basis points. 105

4.4.2.2 Footnote disclosures of reclassifications Although the reclassification amendments to IAS 39 affect the measurement of ILQDQFLDO DVVHWV LQIRUPDWLRQDO HIIHFWV VKRXOG EH DOWHUHG E\ WKH ,$6%¶V DFFRPSDQ\ing amendments to IFRS 7. A closer look into published financial statements of banks around the world reveals that disclosures about reclassifications are twofold. First, almost every bank makes a binary statement in the description of its accounting policies about whether it has applied the reclassification option for some financial assets. Second, reclassifying banks include the then required IFRS 7 disclosures in additional footnotes. However, compliance with these disclosure requirements varies significantly across countries and banks. Only 42 reclassifying banks (34%) in our sample fully comply with all six requirements laid out in the standard in the first annual report following the amendments (Complete Disclosure, see table 4.6.1). In several countries (Australia, Austria, Finland, Kuwait, Norway, Poland, Portugal, Saudi Arabia, Slovenia, Spain, Turkey, United Arab Emirates), all reclassifying banks violated at least one disclosure requirement. These findings are in accordance with CESR (2009), suggesting that compliance with certain disclosure requirements of IFRS 7 is far from perfect.95 Further analyses based on reclassification disclosures show that IFRS 7 disclosure practice varies by reclassification type (see table 4.6.4, panel A) and is also diverse in the location and format of the information presented.96 Table 4.6.4, panel B, relates Complete Disclosure to various country and bank variables that proxy for the quality of governance and enforcement mechanisms as well as other incentives for transparent reporting. Univariate analyses reveal that IFRS 7 compliance is significantly higher in countries with a developed capital market (Log(MCAP/GDP)) or governance system (CGI Score) and in Member States of the EU (EU Country). IFRS 7 compliance is significantly lower in developing economies (Emerging Country) and in code law countries (Code Law Country). The variable FV in Political News is positively related to Complete Disclosure. 95

The finding that required disclosures are not adequately enforced mirrors the earlier literature on compliance, particular in the case of IAS/IFRS (e.g., Cairns, 1999; Street and Gray, 2001; Ball, Robin, and Wu, 2003; Brown and Tarca 2005; Ball, 2006). Given the public awareness and political controversy of the reclassification amendment, along with its significant effects on the accounts (see sections 4.2.1, 4.4.2), it is, however, even more surprising that many banks (still) get away with substantial non-compliance with disclosure requirements.

96

26 reclassifying banks (21%) choose to report the reclassifications in a separate footnote, while the remaining banks disclose the information as a footnote to a previously existing footnote. 72 banks (58%) disclose at least part of the information in a clearly arranged table. The other banks rely exclusively on verbal explanations in text format.

106

This result suggests that full compliance with IFRS 7 is in fact higher in countries where the fair value debate was influenced by politicians and closely followed by the media, i.e. where the topic attracted more public attention. At the bank level, Complete Disclosure is positively related to the quality of the auditing process (Big 4 Auditor)97, the number of analysts (Analyst Following), the number of shares that are not closely held (Free Float), prior commitment to transparency (Earnings Quality) and bank size (Log(Total Assets)).98 These results are similar in the multivariate analyses, although some variables lose their statistical significance. In sum, our analyses provide evidence that IFRS 7 compliance is positively related to the quality of governance and enforcement mechanisms at bank- and country-level and that public awareness of this accounting issue has fostered full disclosure policies. 4.4.3 Economic and political drivers of the reclassification choice 4.4.3.1 Research design In our first set of analyses, we run the following probit regression to provide evidence on the determinants of the firm-specific reclassification choice: (AFS)Recl_Dummy =

ȕ0 ȕ1 Regulatory Costs ȕ2 Political Incentives + ȕ3 State Ownership ȕ4 1R/RVV7DUJHW ¨'HSRVLWRUV + ȕ5 Earnings Quality Ȉȕj Controlsj İ (4.1)

We estimate two different specifications with regard to the dependent variable. In the first specification, we use Recl_Dummy as a dummy variable that takes a value of one if the bank reclassifies either trading or AFS assets. In the second specification, the binary dependent variable is AFSRecl_Dummy and equals one if the bank reclassifies exclusively AFS assets. We estimate the determinants of AFS reclassifications separately because AFS assets feature some idiosyncratic characteristics with respect to capital regulation.99 7KH LQGHSHQGHQW YDULDEOHV FRPSULVH ERWK D EDQN¶V LQFHQWLYHV UHJXODWRU\ costs, political costs, earnings targets, commitment to transparency) and opportunities 97

Only eight reclassifying banks are not audited by one of the Big 4 auditors. Seven of them do not fully comply with the disclosure requirements of IFRS 7.

98

Note that seven reclassifying banks are cross-listed in the United States and therefore registered with the SEC. All of these banks provide complete IFRS 7 disclosures. Therefore, a cross-listing/SEC variable cannot be estimated in the probit analyses.

99

In contrast, all results for reclassifications out of the trading account only correspond highly with our findings for the general reclassification choice and are not reported for brevity. 107

(state ownership, percentage of trading and AFS assets, country-level enforcement) to use the reclassification option. Regulatory costs are captured by two different variables. First, Regulatory Capital Restriction is the difference between the minimum capital ratio at country level (as presented in table 4.6.  DQG WKH LQGLYLGXDO EDQN¶V WRWDO FDSLWDO UDWLR EHIRUH reclassifications, i.e., a higher value is representing a tighter restriction and, thus, a higher probability of regulatory costs. Second, Regulatory AFS Haircut is the proportion of unrealized available for sale (AFS) securities gains included in the determination of total regulatory capital (tier 1 plus tier 2). The variable is regulated by national banking supervisors (see table 4.6.1 for a presentation at country level) and, when necessary under WKH UHVSHFWLYH QDWLRQDO EDQNLQJ ODZ DGMXVWHG IRU WKH VLJQ RI WKH EDQN¶V UHYDOXDWLRQ reserves. The higher this proportion, the more regulatory capital will be safeguarded by a reclassification, i.e., the costs of not reclassifying are higher. The political costs are measured by the interaction term Political Incentives which FRPELQHV WKH EDQN¶V OREE\LQJ DFWLYLWLHV IIF Membership) and the political response to these activities (FV in Political News).100 IIF Membership indicates whether a bank is member of the IIF. Since the IIF had a leading role in the lobbying for reclassifications, we assume that member firms support its position on fair value reclassifications.101 FV in Political News reflects how often top-level politicians (precisely, the president, king, prime minister, secretary of treasury, and secretary of economic affairs) publicly comment on fair value accounting issues between January and October 2008. We collect this information from Google News for each country in our sample. Political Incentives has a value of one if both variables take values larger than zero and of zero otherwise. The direct governmental influence on the reclassification choice is separately proxied for by the governmental equity stakes in a bank (State Ownership). No Loss Target is a dummy variable which takes a value of one if net income before reclassifications is smaller than zero and the greatest in-sample difference between net income after and net income before reclassifications, i.e., it indicates whether it is technically possible for a bank to pass the zero earnings threshold by means of fair value 100

1RWHWKDWLWLVQRWSRVVLEOHWRLGHQWLI\HDFKEDQN¶VSROLWLFDOSRVLWLRQRQUHFODVVLILFDWLRQVLQRXUVHWWLQJ because the reclassification amendment was adopted without due process. Public commenting of the rules was therefore absent.

101

In fact, the US-based investment bank Goldman Sachs left the IIF in May 2008, because it did H[SOLFLWO\UHMHFWWKH,,)¶VGLVPLVVLYHSRVLWLRQon fair value accounting; Financial Times, Goldman set to sever IIF links, May 23, 2008.

108

reclassifications. Since we hypothesize that the incentive to avoid the reporting of a loss, among other incentives, stems from attempts to stabilize customer confidence, the interaction term 1R /RVV 7DUJHW  ǻ &XVWRPHU 'HSRVLWV is constructed as a binary variable which equals one if the sign of the change in customer deposits (ǻ &XVWRPHU Deposits) is negative and No Loss Target has a value of one.102 When the interaction term is included, the two basic terms (No Loss Target and ǻ&XVWRPHU'HSRVLWV) are excluded from the model specifications to facilitate the interpretation of the coefficient (e.g., Ai and Norton, 2003). Earnings Quality sHUYHV DV D SUR[\ IRU D EDQN¶V FRPPLWPHQW WR WUDQVSDUHQF\ DQG denotes the average abnormal loan loss provisions since IFRS adoption. We use an established approach (e.g., Beatty, Ke, and Petroni, 2002) to estimate the nondiscretionary portion of the loan loss provision by regressing loan loss provisions on loan loss reserves (t-1), net charge-offs, the change in non-performing loans between t-1 and t, as well as size (measured as the natural logarithm of the book value of total assets). All variables (except size) are scaled by total assets. However, BvD Bankscope does not contain the composition of loans by regions or by customers for the majority of international banks (Gebhardt and Novotny-Farkas, 2010). Therefore, we cannot control for these factors. The variable takes a value of one if a bank does not use loan loss provisioning to inflate its earnings (i.e., if the cumulative residuals are positive), and zero otherwise. Consistent with our first hypothesis, we expect the coefficient estimates on Regulatory Costs (Regulatory Capital Restriction) 103, Political Incentives, State Ownership, and No /RVV7DUJHW ǻ&XVWRPHU'HSRVLWV WREHSRVLWLYHLHȕ1 !ȕ2 !ȕ3 !DQGȕ4 > 0. Consistent with our second hypothesis, we expect the coefficient estimate on Earnings Quality WREHQHJDWLYHLHȕ5 < 0. We include three different sets of control variables. The first variable controls for the incentives to reduce income volatility by means of reclassifications. If banks attributed income volatility to fair value accounting, a bank with more volatile income should be more likely to use the reclassification option. We define Income Volatility as a dummy YDULDEOHEDVHGRQWKH VWDQGDUGGHYLDWLRQRIDEDQN¶VTXDUWHUO\RUKDOI-yearly percentage 102

,Q XQWDEXODWHG WHVWV ZH XVH DQDO\VWV¶ PHDQ FRQVHQVXV IRUHFDVW WR H[DPLQH ZKHWKHU RWKHU HDUQLQJV targets affect the reclassification decision. The coefficient estimate on the consensus forecast is insignificant in all specifications.

103

When AFSRecl_Dummy is the dependent variable, there is, however, an opposite effect of Regulatory AFS Haircut. 109

change in net income between 2004 and 2009. The presentation of IFRS income statements before IFRS 7 adoption in 2007 does not allow us to restrict this variable to net changes in fair values. The variable is irrelevant for reclassifications out of the AFS category which do not affect net income. The second set relates to the fact that IAS 39 allows reclassification only if a trading or at least selling intention for a specific asset has ceased due to a loss of market liquidity. We therefore expect that a bank is more likely to use the reclassification option if its assets became illiquid during the financial crisis. We use the sum of trading and AFS DVVHWV UHODWLYH WR WRWDO ILQDQFLDO DVVHWV DV D SUR[\ IRU D EDQN¶V H[SRVXUH WR WKH ULVN RI decreasing market liquidity (% Securities).104 In addition, we construct a summary PHDVXUHWKDWFDSWXUHVDEDQN¶VGLUHFWH[SRVXUHWRWKHILQDQFLDOFULVLVExposure to Crisis). 7KLV PHDVXUH LV HVWLPDWHG DV WKH ILUVW SULQFLSDO IDFWRU XVLQJ   D EDQN¶V VWRFN UHWXUQ between January and September 2008 (Stock Return 2008    D EDQN¶V VWRFN UHWXUQ volatility between January and September 2008 (Stock Return Volatility 2008) and (3) a binary variable that indicates whether a bank reports engagements in securitizations in its financial statement (Securitization Activity).105 The third set of control variables acts as a proxy for institutional differences across countries. We use EU membership (EU Country), an aggregate governance score from the World Bank (CGI Score), and the development of the capital market (Log(MCAP/GDP)). We do not have any priors as to how these variables affect the reclassification choices.106 4.4.3.2 Empirical findings Table 4.6.5 reports the results of multivariate probit regressions. Our analyses provide IRXUNH\LQVLJKWV)LUVWRXUUHVXOWVVXJJHVWWKDWWKHFORVHUDEDQN¶VWRWDOFDSLWDOUDWLRLVWR 104

We acknowledge that this relationship is mechanical, because reclassifications out of fair value categories are by definition impossible if a bank had not used those fair value categories previously.

105

We use securitization activity for two reasons. First, securitized loans suffered from market illiquidity during the financial crisis. Second, our analysis of the segment reports shows that securitization activity is correlated with investment banking activities (rho = 0.224, p < 0.001), which were most strongly affected by the market turmoil.

106

Since reclassifications are, as a consequence of the amendments to IAS 39 and IFRS 7, an absolutely permissible means of accounting choice, which cannot be mitigated through restrictive auditing or enforcement procedures, there is no unambiguous channel through which high quality enforcement or auditing could attenuate the magnitude of earnings management through reclassifications. Rather, these variables are of particular importance in the degree of compliance with the IFRS 7 disclosure requirements (Holthausen, 2009; see also section 4.4.4.2).

110

the country-specific minimum capital ratio, the higher the probability of a reclassification (either trading or AFS). For example, unreported univariate tests show that the distance of a non-UHFODVVLI\LQJ EDQN¶V WRWDO FDSLWDO UDWLR WR WKH PLQLPXP FDSLWDO UDWLR LV RQ average, more than 50% higher than the one of a reclassifying bank (650 versus 429 basis points, p<0.01). In the multivariate analyses, the coefficient estimate on Regulatory Capital Restriction is statistically significant at the 1%-level showing the expected positive sign throughout all model specifications. The marginal effects indicate that a decrease in the total capital ratio by 100 basis points is associated with an increase in the reclassification probability of two percentage points. Thus, the reclassification probability of a bank with a capital ratio of exactly the minimum adequacy level is, ceteris paribus, approximately 11 percentage points higher than of the average bank in our sample with a capital ratio of 555 basis points above the minimum level and even 66 percentage points higher than of the bank with a capital ratio equal to the first percentile in our sample (3300 basis points above the minimum level). These figures demonstrate that the effect of capital regulation on accounting choice is not only statistically significant but also economically substantial. In addition, the analysis of AFS reclassification in which we can exploit cross-country differences in the inclusion of unrealized fair value gains into regulatory capital underlines the importance of contracting incentives from capital regulation for the reclassification choice. The discount on unrealized fair value gains required by national banking law (Regulatory AFS Haircut) is negatively associated with the choice to reclassify AFS assets, suggesting that the incentive to reclassify AFS assets is decreasing with the extent to which the unrealized gains are excluded from the determination of regulatory capital, i.e. the incentive is particularly small when reclassifications cannot serve to safeguard regulatory capital due to country-level bank regulation. More specifically, a regulatory switch from a 0% to a 100% haircut on unrealized AFS gains reduces, ceteris paribus, the probability of reclassifications by approximately 15 percentage points Second, our results provide evidence that political incentives determine the reclassification choices of banks. The coefficient on Political Incentives is positive and statistically significant in all models specifications. Unreported univariate analyses further suggest that both components (IIF Membership and FV in Political News) contribute to this finding. Consistent with the political cost hypothesis, a bank is more likely to reclassify if it is a member of the IIF, i.e. was at least indirectly involved in 111

lobbying activities for the new rules, and is from a country with a high degree of political involvement in the fair value debate, i.e. where its accounting choice was more likely to receive a broad level of public attention. Third, the zero earnings target (No Loss Target) determines the reclassification choice, although the individual coefficient estimates is not strongly statistically significant. In contrast, the interaction term 1R/RVV7DUJHW ¨&XVWRPHU'HSRVLWV is significant at the 1%-level throughout all model specifications suggesting that the zero earnings target is of particular importance for banks which are already exposed to a loss of depositor confidence. This latter finding is also consistent with reclassifications serving the political objective to grant regulatory forbearance to economically weak banks. )RXUWK WKH SUR[\ IRU D EDQN¶V SULRU FRPPLWPHQW WR WUDQVSDUHQF\ Earnings Quality) loads in the hypothesized negative direction, suggesting that banks with more transparent reporting history are less likely to reclassify. This finding might also indicate how banks evaluate the usefulness of fair value recognition for investors (despite their position in the lobbying process). Taken together, our results provide support for hypotheses (H1a) and (H2a) according to which perceived benefits from avoiding supervisory interventions, adverse political actions and a loss of depositors as well as perceived costs from a decrease in transparency GHWHUPLQH EDQNV¶ UHFODVVLILFDWLRQ FKRLFHV 7KH FRHIILFLHQW HVWLPDWHV RQ WKH FRQWURO variables have the expected signs and are statistically significant in most specifications (except controls for the quality of country-level institutions). 4.4.4 Stock price reactions to reclassification announcements 4.4.4.1 Research design In this section, we test hypothesis (H1b). In order to H[DPLQH LQYHVWRUV¶ DQG WKH PDUNHW¶VRYHUDOOSHUFHSWLRQDQGDVVHVVPHQWRIWKHUHJXODWRU\FKDQJHDQDWXUDOILUVWVWHS would be to proceed with a formal analysis of the stock market reaction to the announcement of the amendment (e.g., Dechow, Hutton, and Sloan, 1996; Zhang, 2007) or to look at the market-wide reactions to the events which increased or decreased the likelihood in the run-up to the regulatory change (e.g., Armstrong, Barth, Jagolinzer, and Riedl, 2010).107 However, the reclassification amendment we examine was not the result 107

In a similar setting, Beatty, Chamberlain, and Magliolo (1996) and Cornett, Rezaee, and Tehranian (1996) study the market reaction of commercial bank shares to regulatory announcements increasing the extent of fair value disclosures in the US. While Beatty, Chamberlain, and Magliolo (1996) 112

of a long-term process, but rather the culmination of dramatic events within a few days at WKHYHU\SHDNRIWKHILQDQFLDOFULVLV,QIDFW,$6%&KDLUPDQ7ZHHGLHODWHUVWDWHVWKDW³ZH did not have a week, we had only a matter RI GD\V´ +RXVH RI &RPPRQV 2008). However, Armstrong, Barth, Jagolinzer, and Riedl (2010) analyze an environment without any discernible pattern of good or bad news during their sample period, whereas the confounding events during the few days in October 2008 that are relevant for our setting are tremendous. The reclassification amendment was adopted less than one week after the announcement of part-nationalizations of large UK banks such as Barclays, HBOS, Lloyds TSB and RBS. On October 13, the day of the IASB decision, the Financial Times reports that European governments (among them France, Germany, and the UK) pledge a total of USD 2,546bn in guarantees for new bank debt as part of coordinated plans to rescue their financial sectors. At the same time, the European Central Bank, the Bank of England and the Swiss National Bank announced unlimited capital injections at a fixed interest rate into the European money markets. Therefore, we perform two different sets of tests which aim to capture bank-specific economic consequences of the relaxation of fair value rules in the cross-section instead of focusing on market-wide effects. The first set of tests analyses stock price reactions to the IASB approval of the IAS 39 and IFRS 7 reclassification amendments. The second set of tests examines stock price reactions to the eventual reclassification choice at the bank level. We focus our analyses in this section on the potential reclassification benefits that stem from avoiding the violation of regulatory capital restrictions, because other benefits (particularly, avoidance of political costs and of losing depositors) presumably take more time to materialize and are less likely to be unambiguously identified by capital market participants, particularly given the uncertainties in this time of crisis. We begin the first set of tests by running panel regressions that relate raw bankspecific returns during the period 01 October 2008 ± 31 December 2008 to the DJ STOXX 1800 market index 108 and dummy variables that indicate various regulatory concentrate on the adoption of SFAS 115 which regulates recognition and measurement, Cornet, Rezaee, and Theranian (1996) also include events related to the adoption of SFAS 105 and 107 which regulate footnote disclosures of fair values. Both studies find negative (positive) abnormal returns for time windows around events that increase (decrease) the probability of the adoption of fair value standards. The authors conjecture that investors fear the potential volatility of equity when assets are measured at fair value. In contrast, Lys (1996) argues that the negative reaction is due to regulatory concerns rather than to negative effects of fair value measurement per se. 108

The DJ STOXX Global 1800 Index comprises the largest 600 firms, based on free float market capitalization, from each of Europe, North and South America, and the Asia/Pacific region (e.g., Armstrong, Barth, Jagolinzer, and Riedl, 2010). Since this index also includes banks, we cannot rule 113

events in October 2008 that are related to the reclassification amendments. The coefficient estimates on the event dummies represent abnormal returns (denoted as Abn_ReturnRegEvent below). Next, we focus on the abnormal return on 13/14 October 2008 and analyze its cross-sectional determinants.109 The basic regression specification is as follows: Abn_ReturnRegEvent = ȕ0 ȕ1 Expected Reclassification + ȕ2 Regulatory Capital Restriction İ

(4.2)

Since the reclassification choices at the bank level were unknown when the amendments were announced, we use a dummy variable that indicates whether investors expected a bank to reclassify or not (Expected Reclassification). We use two different DSSURDFKHV WR PRGHO LQYHVWRUV¶ H[SHFWDWLRQV ZLWKUHJDUG WRIXWXUHUHFODVVLILFDWLRQV 7KH first approach (Perfect Foresight) assumes that investors perfectly predict which banks will use the reclassification option. The second approach (Prediction Model) assumes that investors use the variables identified in the determinants analysis (see table 4.6.5) to assess the likelihood that a bank will use the reclassification option. Specifically, we code banks as expected reclassification (non-reclassification) banks if a probit model with Reclassification (Perfect Foresight) as dependent variable and Regulatory Capital Restriction, Political Incentives, State Ownership, Change in Customer Deposits, Earnings Quality, Income Volatility, %Securities, Exposure to Crisis (Factor), and Log(MCAP/GDP) as independent variables yields a probability of more than 0.5 (less than or equal to 0.5).110 This approach predicts 104 banks to reclassify. 72 of these banks eventually take the reclassification option. Around the regulatory events, there exists further uncertainty with respect to the future benefits of taking the reclassification option, particularly from avoiding costly regulatory interventions. This uncertainty is captured by the variable Regulatory Capital Restriction. out that part of the return effect we aim to detect is picked up by the market index control variable. However, this impact is likely to be rather small as the DJ STOXX Global 1800, as of December 2008, contains merely 64 of our sample banks, and works against detecting significant abnormal returns. 109

The IASB approval of the reclassification amendments was announced in the late afternoon (GMT) of 13 October (see table 4.6.6, panel A) when the exchanges in many sample had already closed. We therefore use the cumulative abnormal return on 13/14 October 2008 to ensure that the stock market reaction in all sample countries is captured.

110

We acknowledge that the prediction model also requires perfect foresight of the reclassification choice. However, we think that the probit model provides a useful tool to condense the information that was observable in October 2008 into one single measure.

114

This variable acts as a proxy for the potential savings of regulatory costs and it increases with the probability of regulatory interventions. We use three different specifications for Regulatory Capital Restriction. First, we use the continuous variable as described in table 4.6.2. Second and third, we use dummy variables based on the continuous variable that take a value of one if the difference between an LQGLYLGXDO EDQN¶V WRWDO FDSLWDO UDWLR before reclassifications and the minimum capital ratio at country level is less than 2% (61 banks in total; 34 of these banks eventually reclassify) or less than 0% (6 banks in total; 4 of these banks eventually reclassify).111 In order to account for cross-sectional heteroskedasticity and cross-correlation of the residuals, we estimate specification (2) using the weighted portfolio approach by Sefcik and Thompson (1986). The remaining uncertainty with respect to reclassification choice and effects on the accounts resolves at the reclassification announcement dates. In the second set of test, we analyze stock return around these bank-specific announcements. We use the first reclassification announcement for reclassifying banks and, as benchmark announcements, the first earnings announcement for non-reclassifying banks following the official announcement of the reclassification amendments in October 2008. Since these dates cannot be identified for all sample banks, the analyses are based on a reduced sample of 117 reclassifying and 161 non-reclassifying banks. 14 (67) (36) banks make the reclassification

announcement

before

(during)

(after)

the

respective

earnings

announcement. 78 (39) banks announce reclassifications in (interim reports prior to) the first annual report following the amendment to IAS 39. The basic regression specification for the short-term analysis is as follows: Abn_ReturnBankAnn = ȕ0 + ȕ1 Earnings Surprise ȕ2 Reclassification ȕ3 Regulatory Capital Restriction + İ

(4.3)

Abn_ReturnBankAnn is the prediction error from the market model using the DJ STOXX 1800 market index, with interval (-60, -11) and interval (+11, +60) relative to announcement day 0 as estimation window. We follow the trade-to-trade approach of Maynes and Rumsey (1993) to account for thin trading in some of the stocks. Earnings Surprise is an indicator variable that takes a value of one if the earnings number reported DW DQ HDUQLQJV DQQRXQFHPHQW LV KLJKHU WKDQ WKH PHDQ DQDO\VWV¶ IRUHFDVW DQG ]HUR otherwise. Reclassification equals one for banks that announce reclassifications, and zero 111

The two banks that violated the capital restriction and did not use the reclassification option are Gulf Bank (Kuwait) and Banca Italease (Italy). Both banks merely used the fair value categories in their 2008 financial statements. 115

otherwise. For the Regulatory Capital Restriction variable, we use the same three specifications as in the analysis of stock market reactions to regulatory events. We complement the short-term analysis with long-term analysis of mean market-adjusted returns for up to 120 days before and after the reclassification/earnings announcements. Market-adjusted returns are buy-and-hold bank returns minus buy-and-hold returns on the DJ STOXX 1800 market index. 4.4.4.2 Empirical findings Table 4.6.6 presents analyses of stock price reactions to the IASB' approval of the reclassification amendments. Panel A reports details on selected regulatory events around the approval in October 2008. Panel B shows mean abnormal returns on these event days IRUWKHWRWDOVDPSOHDQGVHYHUDOVXEVDPSOHVEDVHGRQLQYHVWRUV¶H[SHFWDWLRQVZLWKUHJDUG to future reclassifications. Across the five events, we find negative raw returns of -1.69%, but insignificant excess returns once we control for general market movements (not tabulated). Mean abnormal returns for the total sample range from -4.8% (t-statistic 14.53) on 6 October 2008 to +3.1% (t-statistic 8.13) on 9 October 2008. On October  WKH GDWH RI WKH ,$6%¶V DPHQGPHQW DQG WKH announcements of coordinated government plans to rescue their financial sectors, we measure significant positive abnormal market reactions. Overall, these results confirm that the reclassification amendments were approved in times of extremely volatile stock markets. Panel C presents results of the cross-sectional analysis of abnormal returns on 13/14 October 2008 following the IASB approval of the reclassification amendments. The coefficient estimate on the dummy variable Expected Reclassification is negative throughout all six specifications, but statistically significant in only one specification where the prediction model is used. Since these specifications are based on the strong assumption that the perfect foresight or the prediction model correctly rHIOHFWVLQYHVWRU¶VH[SHFWDWLRQVWKHVH findings have to be interpreted with caution. Our results are less ambiguous with regard to regulatory capital restriction as a crosssectional determinant of abnormal returns on the day the IASB announced the reclassification amendments. According to the perfect foresight model, the capital market expected 124 banks to reclassify. Four of these banks were expected to violate the regulatory capital restriction (Dummy Cutoff 0% = 1). Our regression analyses show that these banks experience positive abnormal returns on 13/14 October 2008, that are, on average, 5.9% higher than those of the other 120 banks that are expected to reclassify. 116

This effect is statistically significant and even stronger when the prediction model is used. As expected, the effect becomes weaker when the regulatory capital restriction is measured as a 2% cutoff dummy variable and disappears entirely when the continuous variable is applied. These findings suggest that the stock market reacted positively to the reclassification amendments if the reclassification option was expected to be used to safeguard regulatory capital of troubled banks. Table 4.6.7 presents results from short-term and long-term analyses of stock market reactions to bank-specific announcements of their reclassification choice. Panel A shows that the short-term stock market reactions to bank-specific announcements are generally weaker than the reactions to the IASB announcement of the reclassification amendments. The results of the specifications with Regulatory Capital Restriction (Dummy ± 0% Cutoff) show that the banks that are expected to violate the regulatory capital restriction and use the reclassification option experience abnormal announcement returns that are 2.6%-2.7% higher on average than those of the other reclassifying banks. 112 However, this effect is not statistically significant (t-statistic 1.53 and 1.55, respectively). The earnings surprise variable loads positively as expected. Taken together, these findings suggest that the capital market had largely anticipated the impact of fair value reclassifications on regulatory capital restrictions and impounded this information into EDQNV¶VWRFNSULFHVEHIRUHWKH\RIILFLDOO\DQQRXQFHGWKHLUUHFODVVLILFDWLRQGHFLVLRQV7KH resROXWLRQRIWKHUHPDLQLQJXQFHUWDLQW\KDVFDXVHGRQO\PLQRUUHYLVLRQVWREDQNV¶VWRFN prices. Panel B reports mean market-adjusted returns for up to 120 days before and after the reclassification/earnings announcements. The analysis shows stocks of reclassifying banks, on average, performed more poorly before the reclassification than nonreclassifying banks. This finding is line with the results from the determinants analysis (see the Exposure to Crisis variable). Interestingly, stock prices of reclassifying banks also recovered more strongly afterwards. The reclassification effect on regulatory capital does not map into a clear pattern of long-term stock returns. The 60-day window returns again confirm that the reclassification choices were taken in times of extreme volatility. We therefore hesitate to draw a causal link between the reclassification choice and subsequent long-term returns. 112

Untabulated statistics show that the banks that violated the regulatory capital restriction used the reclassification option to substantially increase their Total Capital. Specifically, for any of these banks, the magnitude of the reclassification effects on Total Capital is higher than the sample median, both in absolute and in relative terms. 117

In summary, our findings are consistent with hypothesis (H1b). 4.4.5 Long-term effects of reclassifications on information asymmetry 4.4.5.1 Research design In this section, we test hypothesis (H2b). Specifically, we analyze whether stocks of reclassifying IFRS banks experience an increase in information asymmetry in the long run. Following related literature (e.g., Leuz and Verrecchia, 2000; Muller and Riedl, 2002), we use the bid-ask spread as a proxy for information asymmetry. The basic regression specification is as follows: Log(Bid-Ask Spread) =

ȕ0 ȕ1 Reclassification ȕ2 Effect on Net Income +ȕ3 Complete Disclosure ȕ4 Effect on Net Income*Complete Disclosure Ȉȕj Controlsj İ

(4.4)

All variables are measured at the firm-quarter level. Bid-Ask Spread is the median quoted spread (i.e. the difference between the closing bid and the closing ask price divided by the midpoint). We use the natural logarithm of the bid-ask spread, because the raw values of this variable are highly skewed (see the descriptive statistics in table 4.6.8, panel A). Reclassification equals one for all reclassification quarters starting with the first quarter during which the respective bank announced IFRS reclassifications, and otherwise equals zero. Effect on Net Income is a dummy variable that takes a value of one if the percentage net income effect of the first reclassification is above the sample median (see table 4.6.2, panel A, for details on this variable), and is zero otherwise. We introduce this variable to account for the assumption that fair value reclassifications need to have a sufficient impact on financial statements in order to affect information asymmetry in capital markets. Complete Disclosure indicates whether a reclassifying bank discloses all six items required by IFRS 7, para. 12A, in the footnotes to its financial statements (see table 4.6.1 and table 4.6.4 for details on this variable).113 Consistent with our second hypothesis, we expect the coefficient estimates on Reclassification and Effect on Net Income WREHSRVLWLYHZLWKWKHODWWHUKDYLQJWKHKLJKHUYDOXHLHȕ2 !ȕ1 > 0. In line with 113

39 banks announce reclassifications in interim reports prior to the first annual report following the amendment to IAS 39. 34 of these banks apply the same IFRS 7 disclosure strategy in the interim reports than in the subsequent annual report. The other 5 banks switch from incomplete disclosure in the interim reports to complete disclosure in the annual report.

118

the mitigating effect of high-quality disclosure on information asymmetry we expect the coefficient estimates on Complete Disclosure and Effect on Net Income * Complete Disclosure to be negative without having priors as to how their values relate to each RWKHULHȕ3 ȕ4 < 0 . We estimate regression specification (5) using a firm fixed effects model that controls for time trends. The goal of this differences-in-differences approach is to identify a causal relationship between a treatment (reclassification choice) and an endogenous variable (bid-DVNVSUHDG E\FRPSDULQJWKHWUHDWPHQW¶VLPSDFWRQDIIHFWHGILUPVEDQNVWKDWWDNH the reclassification option) with its impact on unaffected firms (banks that do not take the reclassification option). To ensure that OLS estimation produces consistent standard errors we use standard errors clustered by firm as suggested by Bertrand, Duflo, and Mullainathan (2004). Since we estimate a firm fixed effects model, our set of control variables is confined to variables that capture firm-specific changes over time. Consistent with finance literature (e.g., Huang and Stoll, 1997), we predict that, ceteris paribus, changes in bid-ask spreads are negatively correlated with changes in share turnover or market capitalization and positively correlated with changes in stock return variability. For all control variables, we use the natural logarithm, because the raw values are highly skewed (see the descriptive statistics in table 4.6.8, panel A). 4.4.5.2 Empirical findings Table 4.6.8 presents results for the analysis of long-term consequences of fair value reclassifications on bid-ask spreads. Panel A reports descriptive statistics for the dependent and the independent variables. Panel B presents results from multi-period difference-in-differences analyses using the data for all quarters. In panel C, we drop the first two reclassification quarters to examine the long-term impact of the reclassification choice on bid-ask spreads. The sample comprises 124 reclassifying and 178 nonreclassifying banks, which results in a total of 3,467 firm-quarter observations over the period 2007/Q1 to 2009/Q4. The coefficient estimates on the quarter dummies illustrate that bid-ask spreads increased significantly for non-reclassifying banks in 2008 and moved towards pre-crisis levels at the end of the sample period. This spike in bid-ask spreads remains significant even after including the control variables suggesting that our measure of information asymmetry captures the increased uncertainty during the financial crisis. The coefficient estimates on the reclassification dummy show that reclassifications per se have only a 119

minor impact on bid-ask spreads. However, the results change considerably when we control for the reclassification effect on net income and the quality of disclosures related to the reclassification. In particular, our results suggest that banks that use the reclassification option to inflate their net income experience a significant increase in average bid-ask spreads compared to banks that reclassify with a modest impact on net income (see coefficient estimates on Effect on Net Income in specification (2) and (6)). Further analyses reveal that this increase in information asymmetry is offset if the reclassifications effects are fully disclosed (see coefficient estimates on the interaction term Effect on Net Income * Complete Disclosure in specification (4) and (8)). Translated into economic terms, our results suggest that banks with a high reclassification impact on net income and incomplete disclosure experience an increase in average bid-ask spreads of over 40% relative to non-reclassifying banks (see coefficient estimates on the interaction term Reclassification + Effect on Net Income in specification (4) and (8)). This effect is even stronger (about 60%) in panel C suggesting that the results are not driven by a large but temporary spike in bid-ask spreads around the reclassification announcements. These results are robust to the inclusion of various control variables. Overall, the analyses of the long-term consequences provide evidence for the hypothesis that the reclassification of fair value assets is positively associated with an increase in the information asymmetry component of the bid-ask spread. However, this effect is confined to banks that use the reclassifications to inflate their net income and that do not fully comply with the disclosure requirements of IFRS 7. 4.5 Conclusions and implications In this study, we examine the economic consequences of the amendments to IAS 39 and IFRS 7. These amendments leave banks reporting under IFRS with the choice to retroactively reclassify financial assets that were previously measured at fair value into categories which require measurement at amortized cost, i.e. to effectively abandon fair value accounting for these assets. Our results suggest that the reclassification option under IFRS produced both short-term benefits and long-term costs. In the short run, the relaxation of IAS 39 fair value rules served as a political means to provide short-term relief for the most troubled banks. To some extent, this objective of granting regulatory forbearance to a few arguably system-relevant economically weak banks has been achieved. Yet it is too early to judge whether the reclassifications were useful to overcome only temporary difficulties or whether they simply delayed a process of market exit because these banks are not economically viable. 120

On the other hand, there is evidence that some banks abuse the reclassification option to decrease transparency of their financial statements more generally by not providing accompanying disclosures required by IFRS 7. Our analyses show that this decrease in transparency coincides with a significant increase in information asymmetry between investors as measured by the bid-ask spread. This finding is important because it supports WKH,$6%¶VYLHZWKDW DUHVLVWDQFHWRWKH(8 SROLWLFal pressure could have come at even higher costs since an EU carve-out of the reclassification rules in IAS 39 had not been accompanied by extensive disclosure requirements. The reliance on footnote disclosure quality, however, can be difficult in an interQDWLRQDO VHWWLQJ ZKHUH ILUPV¶ UHSRUWLQJ incentives and the strengths of enforcement institutions differ considerably across countries and non-compliance is still an issue (e.g., Ball, Robin, and Wu, 2003; Barth, Caprio, and Levine, 2001, 2006; Daske, Hail, Leuz, and Verdi, 2009). Our findings seem to also support arguments that a direct relaxation of capital requirements may have been a more appropriate regulatory measure to address the consequences of the financial crisis rather than a relaxation of fair value accounting (e.g., Laux and Leuz, 2009, 2010). In addition, such a solution would have avoided effects on entities other than banks for which the amendment to IAS 39 introduced the same accounting discretion. At the same time, it is unclear whether a change in capital regulation would have fostered moral hazard on the part of bank managers (Bushman and Landsman, 2010). Other costs might have arisen from the necessity of a more complex and time-consuming coordination process between national legislations, since banking regulations are, unlike the international financial accounting standards, not set by one supra-national body but by local governments.

121

122

-

Belgium

6

Denmark

-

2

2

7

Saudi Arabia

1

3

7

7

South Africa

Spain

178

Total / Average

124

6

8

52

3

6

-

4

2

1

1

1

1

-

-

1

1

-

2

1

2

-

4

1

-

-

4

1

3

5

1

-

-

-

2

-

4

-

-

-

-

1

-

Only Trading

27

-

1

1

1

-

-

-

-

1

-

-

1

-

2

-

2

4

-

-

1

-

-

-

-

1

-

2

-

3

3

-

-

-

-

-

1

2

-

1

Only AFS

Yes

45

3

1

7

-

-

2

-

-

-

-

1

-

2

-

3

1

3

-

-

-

-

-

-

-

1

8

-

-

1

4

3

1

-

2

-

1

1

-

-

Both

Reclassifications

42

5

-

-

1

1

3

-

1

-

-

1

-

2

1

-

-

3

-

-

1

-

-

-

1

1

1

3

-

2

4

5

-

2

1

-

1

2

-

-

Yes

82

1

8

8

4

1

-

1

-

2

-

-

2

1

1

5

4

6

-

4

1

-

-

4

-

4

12

-

-

2

3

-

1

2

1

-

1

1

1

1

No

Complete Disclosure

Financial Services Authority

Central Bank

Banking Regulation and Supervision Agency

Financial Supervisory Commission

Financial Market Supervisory Authority FINMA

Swedish Finansinspektionen

Central Bank

Central Bank

Central Bank

Central Bank

Monetary Authority of Singapore

Saudi Arabia Monetary Agency

Central Bank

Central Bank

Central Bank

Polish Financial Supervision Authority

Central Bank

Central Bank

Finanstilsynet (FSA of Norway)

Central Bank

Central Bank

Financial Market Authority

Central Bank

Financial Supervision Agency

Central Bank

Central Bank

Irish Financial Regulator

Hungarian Financial Supervisory Authority

Hong Kong Monetary Authority

Federal Financial Supervisory Authority

Central Bank

Financial Supervisory Authority

Finanstilsynet (Danish FSA)

Central Bank

China Banking Regulatory Commission

Banking, Finance and Insurance Commission

Central Bank

Financial Market Authority

Australian Prudential Regulation Authority

Regulatory Authority

Handbook GENPRU 2.2.185

Circular 13/1993

Regulation OJ 26333/06

Capital Adequacy Regulation

Circular 2008/34

Regulation FFFS 2007:1

Circular 4/2004

Notice R3/2008

Regulation OJ 135/06 & 104/07

Decree 4/2007

Notice 637

SAMA Capital Requirements

Instruction on Bank Regulation

QCB Instructions Part 7

Notice 12/92

KNF Resolutions

Circular 538/06

Capital Guidelines II.A

Capital Adequacy Framework

Decree on Prudential Rules

Resolution No. 138

Regulation ERV

Instruction No. 2/BS/94/2002

Kazakhstan Banking Law

CBJ Instructions

Circular 263

Notice BSD S 2/00

HFSA Regulation

Banking (Capital) Rules

Regulation KonÜV

Regulation 90/02

FIN-FSA Standard 4

Financial Business Act

Directive 436/2006 & 328/2007

Capital Adequacy Regulation

CBFA Circular PPB-2007-1-CPB

Rulebook Vol. 1 Part A CA-2

Austrian Banking Act

Prudential Standard APS 111

Legal Source

Capital Regulation

0.09

0.08

0.10

0.08

0.08

0.08

0.08

0.08

0.10

0.08

0.08

0.10

0.08

0.10

0.10

0.08

0.08

0.10

0.12

0.08

0.08

0.08

0.08

0.12

0.12

0.12

0.08

0.08

0.08

0.08

0.08

0.08

0.08

0.08

0.10

0.08

0.08

0.12

0.08

0.08

Minimum Capital

61%

0%

55%

55%

55%

55%

0%

55%

50%

20%

0%

55%

55%

20%

55%

55%

40%

55%

55%

55%

0%

15%

100%

55%

20%

55%

50%

0%

0%

55%

55%

55%

0%

100%

0%

100%

10%

55%

30%

55%

AFS Haircuts (excl. Tax)

Selected Variables

Yes (28%)

No

No

No

No

Yes (26.3%)

No

No

Yes (21%)

Yes (19%)

No

No

No

No

No

No

No

No

No

Yes (25.5%)

Yes (20%)

No

No

No

No

Yes (27.5%)

Yes (12.5%)

No

No

No

No

Yes (26%)

No

Yes (10%)

No

Yes (35%)

No

Yes (25%)

No

Tax Deductions 1.1% 0.9%

6.9%

10.7%

21.7%

13.3%

9.0%

1.4%

7.3%

1.0%

5.6%

23.6%

0.0%

0.4%

13.3%

24.3%

9.6%

0.6%

5.3%

2.3%

12.2%

3.8%

0.8%

0.0%

28.7%

20.8%

0.0%

2.8%

0.7%

1.8%

0.3%

0.4%

5.0%

0.4%

0.8%

2.0%

0.0%

7.4%

15.6%

15.9%

19

206

0

4

2

0

7

3

2

2

2

0

0

32

0

1

0

36

0

2

19

0

0

0

1

0

27

12

1

9

107

219

1

0

0

19

7

0

2

30

FV in Political News 33%

31.4%

50%

47%

69%

0%

13%

57%

13%

25%

0%

0%

20%

33%

20%

100%

40%

15%

9%

60%

8%

20%

0%

0%

44%

0%

45%

19%

67%

50%

0%

21%

21%

0%

10%

0%

80%

100%

33%

100%

-26.6%

-39.4%

-20.8%

-4.2%

-40.4%

-29.9%

-40.3%

-36.5%

-30.4%

-15.2%

-32.4%

-1.5%

-17.1%

-27.9%

-54.2%

3.9%

-46.6%

-28.7%

-28.6%

-7.8%

-34.9%

-38.7%

-58.5%

-28.5%

-11.6%

-14.6%

3.2%

-32.4%

-58.3%

-38.7%

-35.9%

-41.1%

-38.6%

-19.8%

-46.7%

-52.4%

-55.1%

-45.3%

-1.6%

-39.5%

3.0% 3.5%

1.8%

3.5%

2.5%

3.4%

2.7%

3.0%

2.8%

2.4%

2.8%

1.4%

1.8%

1.5%

2.6%

4.0%

3.5%

2.6%

2.7%

2.2%

2.5%

2.5%

2.6%

2.9%

2.4%

1.8%

0.5%

1.8%

2.1%

5.3%

2.8%

2.7%

3.1%

2.2%

2.6%

2.4%

3.1%

3.5%

3.9%

1.8%

37%

70%

0%

31%

67%

13%

43%

100%

63%

0%

0%

60%

11%

20%

0%

100%

8%

36%

0%

25%

60%

0%

50%

0%

0%

0%

86%

67%

0%

40%

57%

32%

0%

20%

0%

60%

100%

22%

100%

100%

Securitization Activity

Effect of Financial Crisis

IIF MemberStock Return Stock Return ship Volatility

Political Influence State Ownership

Notes: This table presents the sample composition by country and selected country variables as well as country averages of selected bank variables. The full sample comprises 302 listed IFRS-reporting banks from 39 countries. The sample is divided in reclassifying (Yes ) and non-reclassifying banks (No ). A reclassifying bank is defined as a bank that chooses to reclassify fair value assets in accordance with the amendment to IAS 39 in the financial year 2008. The column Only Trading reports the number of banks that reclassified exclusively trading securities. The column Only AFS reports the number of banks that reclassified exclusively available for sale (AFS) securities. The column Both reports the number of banks that reclassified both trading and AFS securities. Complete Disclosure indicates whether a reclassifying bank discloses all six items required by IFRS 7, para. 12A, in the footnotes to its financial statements. Regulatory Authority denotes the institution which is responsible for the capital regulation of commercial banks at the country level. Legal Source provides the source of our information about the capital regulation variables. Minimum Capital is the total capital ratio (tier 1 plus tier 2) required for commercial banks by national regulators (source: The World Bank / own survey). AFS Haircuts is the proportion of unrealized AFS securities gains that are deducted from the revaluation reserves in the determination of total regulatory capital. Tax Deductions indicates whether future tax charges are deducted from unrealized gains of AFS securities in the determination of total regulatory capital in addition to the general haircut (source: CEBS / own survey). Country-specific tax rates are taken from the OECD tax survey. State Ownership is the proportion of a EDQN¶V shares held by a governmental institution in October 2008 (source: BvD BankScope). The table reports the arithmetic mean per country. FV in Political News is the number of times statements by a FRXQWU\¶V president, prime minister, secretary of finance, or secretary of economic affairs concerning fair value measurement, IAS 39, and reclassifications were captured in the news between June 2008 and October 2009 (source: Google News). IIF Membership equals 1 if a bank is a member of the International Institute of Finance, and 0 otherwise (source: IIF). The table reports the proportion per country. Stock Return is a EDQN¶V stock return between January and September 2008 (source: Thomson Reuters Datastream). Stock Return Volatility is a EDQN¶V stock return volatility between January and September 2008 (source: Thomson Reuters Datastream). For both variables, the table reports the arithmetic mean per country. Securitization Activity equals 1 if the bank reports engagements in securitizations in its financial statement and equals 0 otherwise (source: own data). The table reports the proportion per country.

4

United Kingdom

8

5

7

Turkey

United Arab Emirates

2

5

4

6

4

Sweden

Switzerland

Taiwan

1

2

-

-

Slovenia

1

4

1

Singapore

Slovakia

2

3

5

-

5

Qatar

Portugal

9

4

Russian Federation

Philippines

-

5

2

9

Oman

4

8

Norway

Poland

-

-

2

2

4

3

Lithuania

4

1

3

Netherlands

Liechtenstein

5

6

Jordan

-

-

8

Italy

5

13

2

Hungary

Ireland

Kazakhstan

7

Kuwait

5

7

6

4

14

Germany

Hong Kong

France

2

1

2

1

Cyprus

Finland

4

-

2

10

China

1

1

3

8

1

6

Australia

Total

Austria

Total

No

Bahrain

Country

4.6 Tables

4.6.1 Reclassification data and selected variables

4.6.2 Reclassification effects Panel A: All Reclassifications Absolute Effects N=124 Net Income Earnings per Share

Mean

Std. Dev.

182.960

599.238

P1 -2.666

Relative Effects Median 4.978

P99

Mean

Std. Dev.

3,225.600

43.74%

132.75%

-6.45%

P1

Median

P99

3.41%

790.91% 254.55%

0.570

2.866

-0.122

0.011

26.961

26.92%

51.41%

-18.14%

3.43%

287.072

976.356

-18.379

8.258

5,000.000

4.88%

15.66%

-4.01%

0.78%

88.64%

0.068

0.438

-0.007

0.004

0.938

47.36%

141.50%

-12.38%

3.71%

790.91%

Revaluation Reserve

104.113

702.303

-637.891

0.000

2,130.000

-85.12%

563.35%

-2556%

0.00%

110.99%

Tier 1 Capital Ratio

0.174

0.353

-0.124

0.063

1.539

2.05%

4.58%

-2.24%

0.65%

18.36%

Total Capital Ratio

0.332

0.889

-0.450

0.102

2.688

19.58%

191.95%

-1.96%

0.86%

28.44%

P99

Mean

Std. Dev.

P1

3225.600

67.20%

192.30%

-19.75%

Shareholder's Equity Return on Equity

Panel B: Trading Reclassifications Only Absolute Effects N=52 Net Income Earnings per Share

P1

Relative Effects

Mean

Std. Dev.

Median

152.786

534.499

-4.318

8.735

Median

P99

7.68%

1109.57% 214.98%

0.262

0.658

-0.122

0.025

3.569

25.77%

45.12%

-18.14%

7.32%

132.369

525.930

0.000

0.010

3225.600

0.99%

2.25%

0.00%

0.00%

10.65%

Return on Equity

0.016

0.025

-0.011

0.008

0.150

65.93%

188.76%

-19.75%

7.52%

1079.33%

Tier 1 Capital Ratio

0.182

0.288

-0.112

0.106

1.539

1.76%

2.79%

-2.24%

1.01%

16.27%

Total Capital Ratio

0.139

0.297

-0.224

0.067

1.603

1.08%

2.20%

-1.49%

0.47%

11.88%

Shareholder's Equity

Panel C: AFS Reclassifications Only Absolute Effects N=27

Mean

Revaluation Reserve

83.355

Std. Dev. 252.927

P1 -14.149

Relative Effects Median 13.877

P99

Mean

1,300.000

33.94%

Std. Dev. 30.95%

P1

Median

P99

-14.74%

36.92%

107.26%

Total Capital Ratio 0.331 0.416 -0.087 0.093 1.290 2.58% 3.50% -0.88% 0.90% 11.62% Notes: This table presents the effects of a EDQN¶V IAS 39 reclassification choice on its net income and equity capital. The sample of reclassifying banks comprises 124 IFRS-reporting banks from 39 countries (see table 4.6.1 for details). Absolute effects on net income, VKDUHKROGHU¶V equity and revaluation reserve are reported in millions of Euros. Absolute effects on earnings per share are reported in Euros. Absolute effects on return on equity, the tier 1 capital ratio and the total capital (tier 1 plus tier 2) ratio are reported in percentage points. Relative effects are calculated as the absolute effect scaled by the size of the respective variable before reclassifications and are reported in percent. Panel A comprises all reclassifying banks. Panel B comprises all banks that reclassified exclusively trading securities. Panel C comprises all banks that reclassified exclusively available for sale (AFS) securities, i.e. those reclassifications which did not affect net income and tier 1 capital, but only the revaluation reserve as part of VKDUHKROGHU¶V equity and, proportionately (as indicated by AFS Inclusion , see table 4.6.1 for details), tier 2 capital.

123

4.6.3 Descriptive statistics Panel A: Bank Variables Variables (N = 302)

Mean

Std. Dev.

P1

P25

Median

P75

P99

% AFS Assets

0.077

0.081

0.000

0.017

0.052

0.108

0.324

% Free Float

55.407

40.018

0.000

9.000

61.000

100.000

100.000

% Securities

0.109

0.096

0.000

0.039

0.083

0.160

0.449

¨&XVWRPHU'HSRVLWV Analyst Following

0.014

0.302

-0.493

-0.043

0.001

0.041

0.589

7.040

8.452

0.000

1.000

4.000

11.000

34.000

Big 4 Auditor

0.924

0.266

0.000

1.000

1.000

1.000

1.000

Consensus Forecast Target

0.523

0.500

0.000

0.000

1.000

1.000

1.000

Earnings Quality

0.510

0.501

0.000

0.000

1.000

1.000

1.000

Exposure to Crisis

0.000

0.646

-1.204

-0.429

-0.031

0.405

1.770

Income Volatility

0.631

0.423

0.063

0.336

0.533

0.796

2.363

No Loss Target

0.149

0.357

0.000

0.000

0.000

0.000

1.000

Political Incentives

0.185

0.389

0.000

0.000

0.000

0.000

1.000

Regulatory AFS Haircut

0.331

0.328

0.000

0.000

0.400

0.550

1.000

-5.554

6.824

-33.139

-6.312

-3.908

-2.409

0.700

0.067

0.158

0.000

0.000

0.003

0.020

0.775

127,127

371,986

276

3,130

11,444

47,600

2,105,760

Variables (N = 39)

Mean

Std. Dev.

P1

P25

Median

P75

P99

CGI Score Code Law Country

0.929

0.731

-0.678

0.453

0.915

1.670

1.970

0.846

0.366

0.000

1.000

1.000

1.000

1.000

Emerging Country

0.410

0.498

0.000

0.000

0.000

1.000

1.000

EU Country

0.487

0.506

0.000

0.000

0.000

1.000

1.000

Regulatory Capital Restriction State Ownership Total Assets Panel B: Country Variables

MCAP/GDP 1.206 0.959 0.084 0.540 1.022 1.414 5.005 Notes: This table provides descriptive statistics for all variables used in the analyses of the determinants of reclassification choices (table 4.6.5) and disclosure strategies (table 4.6.4, panel B). The descriptive statistics are based on the full sample of 302 banks (panel A) or 39 countries (panel B), respectively. % Securities (% AFS Assets ) is the proportion of trading assets and AFS assets (AFS assets only) relative to total financial assets. % Free Float is the proportion of total shares issued by a bank that was not closely held at the end of calendar year 2008 (source: Thomson Reuters Datastream). ¨ Customer Deposits is a dummy variable for banks that experienced a decrease in customer deposits, scaled by total liabilities, between financial years 2007 and 2008 that is greater than the median decrease (source: BvD Bankscope). Analyst Following is the number of analysts that follow the respective bank (source: IBES). Big 4 Auditor is a dummy variable that takes a value of one if a bank is audited by one of the big 4 auditors, and zero otherwise (source: BvD BankScope). Consensus Forecast Target (Dummy) indicates whether a bank is technically able to use fair value reclassifications to beat DQDO\VWV¶ consensus forecast. Earnings Quality denotes a EDQN¶V average abnormal loan loss provisions since IFRS adoption. We use the following regression model to estimate the nondiscretionary portion of the loan loss provision: Loan Loss Provisions / Total Assets = ȕ0 + ȕ1 Loan Loss Reserves t-1 / Total Assets + ȕ2 Net Charge-Offs / Total Assets + ȕ3 ǻ Non-Performing Loans / Total Assets + ȕ4 Log(Total Assets) + İ Exposure to Crisis is the extent to which a bank has been affected by the global financial crisis in 2008. It is the first principal factor of the following three variables (see table 4.6.1 for details): (1) Stock Return 2008 , (2) Stock Return Volatility 2008 , (3) Securitization Activity . Income Volatility is the standard deviation of a EDQN¶V quarterly or half-yearly change in net income, measured in percent, between 2004 and 2009 (Source: Worldscope). No Loss Target (Dummy) takes a value of one if net income before reclassifications is smaller than zero and larger than the greatest in-sample difference between net income before and net income after reclassifications, i.e. it indicates whether it is technically possible for a bank to pass the zero earnings threshold by means of fair value reclassifications. Political Incentives equals one if a bank is a member of the International Institute of Finance and if the variable Fair Value in Political News is greater than zero, and zero otherwise (see table 4.6.1 for details on the underlying variables). Regulatory AFS Haircut is the proportion of unrealized for sale (AFS) securities gain that are not included in the determination of regulatory capital (tier 1 plus tier 2). The variable is regulated by national banking supervisors (see table 4.6.1 for a presentation at country level) and, when necessary, set at 0% if a bank reports negative revaluation reserves. Regulatory Capital Restriction represents capital management incentives to reclassify fair value assets, which are defined as the difference between the minimum capital ratio at country level (as presented in table 4.6.1) and the individual EDQN¶V total capital ratio before reclassifications. State Ownership is defined in table 4.6.1. Total Assets is the book value of total assets in million Euros as of financial year 2008 (source: BvD BankScope). CGI Score is a country-specific governance score from Kaufmann, Kraay, and Mastruzzi (2009). MCAP/GDP is the ratio of a FRXQWU\¶V stock market capitalization to its Gross Domestic Product (source: The World Bank).

124

4.6.4 Reclassification disclosures Panel A: Reclassification Disclosures Type of Reclassification

Categories

HFT2Cost

HFT2AFS

AFS2Cost

Amount Reclassified

IFRS 7.12A (a)

65

(97%)

54

(95%)

69

(96%)

New Category

IFRS 7.12A (a)

66

(99%)

57

(100%)

72

(100%)

49

(86%)

66

(92%)

63

(88%)

FV of Reclassified Assets at BS Date

IFRS 7.12A (b)

62

(93%)

BV of Reclassified Assets at BS Date

IFRS 7.12A (b)

62

(93%)

Reason for Reclassification

IFRS 7.12A (c)

49

(73%)

38

(67%)

59

(82%)

Income / OCI Before Reclassification

IFRS 7.12A (d)

44

(66%)

35

(61%)

47

(65%)

Effect of Reclassification on Income / OCI

IFRS 7.12A (e)

59

(88%)

50

(88%)

59

(82%)

Effective Interest Rate

IFRS 7.12A (f)

38

(57%)

23

(40%)

46

(64%)

Estimated Cash Flow Recovery

IFRS 7.12A (f)

33

(49%)

18

(32%)

37

(51%)

not applicable

Panel B: Determinants of Complete Disclosure Indepedent Variables Log(MCAP/GDP) CGI Score Code Law Country EU Country Emerging Country FV in Political News Big 4 Auditor Analyst Following Free Float Earnings Quality Log(Total Assets) Intercept Number of observations Correct predictions (scaled)

Expected Sign + + + + + + + + +

Dependent Variable: Complete Disclosure Univariate Analyses

Multivariate Analyses

0.168

0.344

0.328

0.212

(2.51)**

(2.99)***

(2.85)***

(2.57)**

0.160

0.089

0.059

0.166

0.140

-0.054

(2.35)**

(0.82)

(0.55)

(1.53)

(1.35)

(-0.67)

-0.472

-0.088

-0.104

-0.251

-0.265

-0.120

(-3.46)***

(-0.46)

(-0.54)

(-1.35)

(-1.47)

(-0.65)

0.172

0.285

0.301

-0.024

-0.003

(2.02)**

(1.64)

(1.77)*

(-0.19)

(-0.02)

-0.199

0.569

0.414

0.350

0.217

(-2.29)**

(2.65)***

(1.97)**

(1.86)*

(1.16)

0.003

0.002

0.002

0.002

0.003

0.002

(4.22)***

(1.42)

(1.99)**

(1.87)*

(2.52)**

(1.79)*

0.245

0.222

0.233

0.191

0.204

0.222

(1.50)

(1.44)

(1.49)

(1.13)

(1.22)

(1.31)

0.012

-0.016

-0.002

-0.015

-0.000

-0.012

(2.72)***

(-2.04)**

(-0.24)

(-2.00)**

(-0.01)

(-1.61)

0.002

0.005

0.004

0.003

0.002

0.002

(1.90)*

(2.50)**

(2.23)**

(1.67)*

(1.30)

(1.43)

0.227

0.152

0.185

0.092

0.128

0.104

(2.65)***

(1.29)

(1.74)*

(0.78)

(1.23)

(0.89)

Yes

0.098

0.115

0.113

0.098

(4.04)***

(2.56)** Yes

Yes

(2.57)** Yes

(2.20)** Yes

124

124

124

124

124

0.76 (0.29)

0.77 (0.31)

0.78 (0.36)

0.77 (0.31)

0.80 (0.41)

0.31 (0.16) 0.25 (0.11) 0.26 (0.12) 0.20 (0.08) 0.27 (0.15) McFadden's (adjusted) R-squared Notes: This table presents the results from univariate and multivariate regressions that relate the reclassification choice to various country and bank variables. The dependent variable Reclass_Dummmy is a dummy variable that takes a value of one if a bank reclassifies trading or AFS assets in accordance with IAS 39 in financial year 2008, and zero otherwise (see table 4.6.1). The dependent variable AFSReclass_Dummy is a dummy variable that takes a value of one if a bank reclassifies AFS assets in accordance with IAS 39 in financial year 2008, and zero otherwise (see table 4.6.1). All independent variables are described in table 4.6.1 and table 4.6.3, respectively. The table reports marginal effects at the mean (median) of all continuous (binary) independent variables and z-statistics (in parentheses). The z-statistics are based on robust standard errors. The proportion of correct SUHGLFWLRQVLVVFDOHGDFFRUGLQJWR9HDOODQG=LPPHUPDQQ 0F)DGGHQ¶VDGMXVWHG 5VTXDUHGLVFDOFXODWHGDFFRUGLQJWR0F)DGGHQ 

 **, * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed), respectively.

125

126 + + + + + + + +

CGI Score Code Law Country EU Country Emerging Country FV in Political News Big 4 Auditor Analyst Following Free Float Earnings Quality Log(Total Assets)

124 0.77 (0.31)

0.76 (0.29)

Yes

(1.74)*

0.185

(2.23)**

0.004

(-0.24)

-0.002

(1.49)

0.233

(1.99)**

0.002

(1.97)**

0.414

(1.77)*

0.301

(-0.54)

-0.104

(0.55)

0.059

(2.85)***

0.328

124

0.115 (2.56)** Yes

0.098 (4.04)***

0.152 (1.29)

0.227 (2.65)***

0.005 (2.50)**

0.002 (1.90)*

-0.016 (-2.04)**

0.012 (2.72)***

0.222 (1.44)

0.245 (1.50)

0.002 (1.42)

0.003 (4.22)***

0.569 (2.65)***

-0.199 (-2.29)**

0.285 (1.64)

0.172 (2.02)**

-0.088 (-0.46)

-0.472 (-3.46)***

0.089 (0.82)

0.160

(2.99)***

(2.51)** (2.35)**

0.344

0.78 (0.36)

124

(2.57)** Yes

0.113

(0.78)

0.092

(1.67)*

0.003

(-2.00)**

-0.015

(1.13)

0.191

(1.87)*

0.002

(1.86)*

0.350

(-0.19)

-0.024

(-1.35)

-0.251

(1.53)

0.166

Multivariate Analyses

0.140

0.003

(1.16)

0.217

(-0.02)

-0.003

(-1.47)

-0.265

(1.35)

0.77 (0.31)

124

Yes

(1.23)

0.128

(1.30)

0.002

(-0.01)

-0.000

(1.22)

0.204

(2.52)**

Dependent Variable: Complete Disclosure

0.168

Univariate Analyses 0.212

0.80 (0.41)

124

(2.20)** Yes

0.098

(0.89)

0.104

(1.43)

0.002

(-1.61)

-0.012

(1.31)

0.222

(1.79)*

0.002

(-0.65)

-0.120

(-0.67)

-0.054

(2.57)**

0.31 (0.16) 0.25 (0.11) 0.26 (0.12) 0.20 (0.08) 0.27 (0.15) McFadden's (adjusted) R-squared Notes: This table presents the results from univariate and multivariate regressions that relate the reclassification choice to various country and bank variables. The dependent variable Reclass_Dummmy is a dummy variable that takes a value of one if a bank reclassifies trading or AFS assets in accordance with IAS 39 in financial year 2008, and zero otherwise (see table 4.6.1). The dependent variable AFSReclass_Dummy is a dummy variable that takes a value of one if a bank reclassifies AFS assets in accordance with IAS 39 in financial year 2008, and zero otherwise (see table 4.6.1). All independent variables are described in table 4.6.1 and table 4.6.3, respectively. The table reports marginal effects at the mean (median) of all continuous (binary) independent variables and z-statistics (in parentheses). The z-statistics are based on robust standard errors. The proportion of correct SUHGLFWLRQVLVVFDOHGDFFRUGLQJWR9HDOODQG=LPPHUPDQQ 0F)DGGHQ¶VDGMXVWHG 5VTXDUHGLVFDOFXODWHGDFFRUGLQJWR0F)DGGHQ 

 **, * indicate statistical significance at the 1%, 5% and 10% levels (two-tailed), respectively.

Correct predictions (scaled)

Number of observations

Intercept

+

Expected Sign

Log(MCAP/GDP)

Indepedent Variables

Panel B: Determinants of Complete Disclosure

4.6.5 Determinants of reclassification choice

4.6.6 Stock market reactions to regulatory events Panel A: Events Description Event

Source

No.

Date

Description

1

06 Oct 2008 (Monday)

European G8 Members meet and explicitly call for a reclassification option (Sarkozy statement in press conference on 04/10/2008)

2

08 Oct 2008 (Wednesday)

Following an ECOFIN decision, Commissioner Charlie McCreevy announces in the European Parliament that the EU is prepared to adopt its own EU version of IAS 39 which would include a reclassification option

3

09 Oct 2008 (Thursday)

IASCF Trustees allow the suspension of the due process for a potential reclassification amendment to IAS 39

4

13 Oct 2008 (Monday)

IASB adopts reclassification amendments to IAS 39 and IFRS 7

IASB Press Release, 3:42pm GMT

5

15 Oct 2008 (Wednesday)

EU Commission officially endorses the revised IAS 39 / IFRS 7

EU Press Release (IP/08/1513)

Statement of the Palais de l'Elysee

ECOFIN Press Release (13784/08), European Parliament Speech/08/513 Reuters News, 10:59am GMT

Panel B: Univariate Analysis Independent Variables

Event Window

Intercept Market Index (DJ STOXX) Event 1 Event 2 Event 3 Event 4

06 Oct 2008 08 Oct 2008 09 Oct 2008 13/14 Oct 2008

Total Sample

Expected Reclassification (Perfect Foresight) Yes

No

Diff.

Expected Reclassification (Prediction Model) Yes

No

Diff.

-0.004

-0.004

-0.004

-0.004

-0.005

(-2.17)**

(-1.76)*

(-2.49)**

(-1.35)

(-2.76)***

0.416

0.513

0.348

0.611

0.313

(5.59)***

(5.38)***

(5.42)***

(6.10)***

(4.92)***

-0.048

-0.061

-0.040

-0.021

-0.062

-0.041

-0.022

(-14.53)***

(-15.68)***

(-12.74)***

(-11.54)***

(-14.62)***

(-13.83)***

(-10.45)***

-0.030

-0.033

-0.029

-0.004

-0.027

-0.032

0.005

(-9.13)***

(-8.41)***

(-9.27)***

(-2.04)**

(-6.29)***

(-10.89)***

(2.62)**

0.031

0.036

0.027

0.009

0.037

0.028

0.009

(8.13)***

(8.12)***

(7.67)***

(4.25)***

(7.47)***

(8.27)***

(3.67)***

0.029

0.026

0.031

-0.005

0.019

0.035

-0.016

(4.68)***

(2.51)**

(8.11)***

(-0.57)

(2.25)**

(6.55)***

(-4.28)***

-0.000

-0.003

0.002

-0.006

0.004

-0.002

0.006

(-0.04)

(-0.53)

(0.42)

(-1.70)*

(0.55)

(-0.49)

(1.81)*

R-squared

0.13

0.15

0.12

0.16

0.12

# Banks

302

124

178

104

198

Event 5

15 Oct 2008

127

Panel C: Cross-Sectional Analysis Independent Variables Intercept Expected Reclassification (Perfect Foresight)

Dependent Variable: Cumulative Abnormal Return on 13/14 October 2008 0.030

0.030

0.030

0.031

0.032

0.032

(2.57)**

(2.63)**

(2.71)***

(2.67)***

(2.92)***

(2.93)***

-0.002

-0.009

-0.007

(-0.22)

(-1.21)

(-0.99)

Expected Reclassification (Prediction Model) Regulatory Capital Restriction (Continuous)

-0.017

-0.014

(-1.85)*

(-1.61)

-0.000

-0.000

(-0.54)

(-0.76)

Regulatory Capital Restriction (Dummy - Cut-off 2%)

0.003

0.002

(0.46)

(0.26)

Regulatory Capital Restriction (Dummy - Cut-off 0%) Expected Reclassification x Regulatory Capital Restriction

-0.000 (-0.00)

0.026

0.007

(1.59)

(0.60)

0.001

0.011

0.033

0.003

0.019

0.082

(0.54)

(1.35)

(1.23)

(1.65)

(2.08)**

(2.54)**

0.001

0.014

0.059

0.003

0.020

0.090

Additional Tests Regulatory Capital Restriction + Interaction Term

(0.45) (2.41)** (2.85)*** (1.58) (3.08)*** (3.19)*** Notes: This table presents analyses of stock market reactions to selected events around the official announcement of the amendment to IAS 39. Panel A reports details on the regulatory events. Panel B shows results from panel regressions that relate raw bank-specific stock returns during the period 01 October 2008 ± 31 December 2008 to the DJ STOXX 1800 market index and event dummy variables. The event window for Event 4 covers two days, because the amendment was announced in the late afternoon of 13 October 2008 (GMT) when the exchanges in many sample countries had already closed. The coefficient estimates on the event dummies represent abnormal returns. We report abnormal returns for the total sample and subsamples based on two different approaches that model LQYHVWRU¶V expectations with regard to future reclassifications. The first approach (Perfect Foresight ) assumes that investors perfectly predict which banks will use the reclassification option. The second approach (Prediction Model ) assumes that investors use the variables identified in the determinants analysis (see table 4.6.5) to assess the likelihood that a bank will use the reclassification option. Specifically, we code banks as expected reclassification (non-reclassification) banks if a probit model with Reclassification (Perfect Foresight) as dependent variable and Regulatory Capital Restriction, Political Incentives, State Ownership, Change in Customer Deposits, Income Volatility, Earnings Quality, %Securities, Exposure to Crisis (Factor), and Log(MCAP/GDP) as independent variables yields a probability of more than 0.5 (less than or equal to 0.5). This approach predicts 104 banks to reclassify. 72 of these banks eventually take the reclassification option. Panel C examines the cross-sectional determinants of bank-specific abnormal returns following the ,$6%¶V official announcement of the amendment to IAS 39 (Event 4: 13/14 October 2008). Expected Reclassification (Perfect Foresight) and Expected Reclassification (Prediction Model) are dummy variables that indicate whether a bank is predicted to take the reclassification option. Regulatory Capital Restriction (Continuous) is described in table 4.6.3. Regulatory Capital Restriction (Dummy ± Cut-off 2%) and Regulatory Capital Restriction (Dummy ± Cut-off 0%) are dummy variables that take a value of one if the difference between an individual EDQN¶V total capital ratio before reclassifications and the minimum capital ratio at country level is less than 2% (61 banks in total; 34 of these banks eventually reclassify) or less than 0% (6 banks in total; 4 of these banks eventually reclassify). Panel B and C report OLS coefficient estimates and tstatistics (in parentheses). The t-statistics in panel B are based on standard errors that are clustered by date. The coefficient estimates and t-statistics in panel C are based on the weighted portfolio approach by Sefcik and Thompson (1986). ***, **, * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

128

4.6.7 Stock market reactions to reclassification announcements Panel A: Short-Term Analysis Dependent Variable: Abnormal Return (DJ STOXX) cumulated over Event Days 0 and 1

Independent Variables

Reclassification and Benchmark Announcements

Intercept Earnings Surprise Reclassification Regulatory Capital Restriction (Continuous)

Reclassification Announcements only

0.002

0.003

0.004

-0.011

-0.008

-0.010

(0.37)

(0.50)

(0.78)

(-1.35)

(-0.86)

(-1.46)

0.022

0.021

0.022

0.027

0.027

0.028

(1.78)*

(1.74)*

(1.79)*

(1.43)

(1.40)

(1.53)

-0.014

-0.010

-0.014

(-1.29)

(-1.02)

(-1.66)*

-0.000

-0.001

(-0.77)

(-0.46)

Regulatory Capital Restriction (Dummy - Cutoff 2%)

0.010

-0.005

(0.65)

(-0.37)

Regulatory Capital Restriction (Dummy - Cutoff 0%)

-0.015

0.027

(-1.83)*

(1.55)

-0.000

-0.015

0.041

(-0.24)

(-0.78)

(2.17)**

R-squared

0.02

0.02

0.02

0.02

0.02

0.02

Observations

278

278

278

117

117

117

Reclassification x Regulatory Capital Restriction

Additional Tests Regulatory Capital Restriction + Interaction Term

-0.001

-0.005

0.026

(-0.50)

(-0.41)

(1.53)

Panel B: Long-Term Analysis Event Window (-120, -1) (+1, +120) (-120, -61) (-60, -1) (+1, +60) (+61, +120) # Banks

Dependent variable: Mean market-adjusted return

Total Sample

Reclassification

Effect on Regulatory Capital

Yes

No

Diff.

> Median

< Median

Diff.

-0.087

-0.131

-0.054

-0.077

-0.145

-0.118

-0.027

(-6.01)***

(-5.34)***

(-3.20)***

(-2.57)**

(-5.33)***

(-2.88)***

(-0.55)

0.133

0.263

0.036

0.227

0.212

0.314

-0.102

(4.63)***

(4.71)***

(1.43)

(3.70)***

(2.47)**

(4.37)***

(-0.91)

-0.024

-0.071

0.011

-0.082

-0.091

-0.052

-0.039

(-2.39)**

(-4.42)***

(0.90)

(-4.09)***

(-4.21)***

(-2.18)**

(-1.23)

-0.062

-0.049

-0.071

0.023

-0.080

-0.018

-0.063

(-2.16)**

(-0.75)

(-4.56)***

(0.34)

(-2.47)**

(-0.14)

(-0.49)

0.029

0.099

-0.022

0.122

0.027

0.170

-0.144

(1.48)

(2.73)***

(-1.12)

(2.93)***

(0.51)

(3.48)***

(-2.00)**

0.105

0.152

0.071

0.082

0.187

0.118

0.069

(7.13)***

(5.79)***

(4.33)***

(2.64)***

(4.71)***

(3.44)***

(1.30)

278

117

161

58

59

Notes: This table presents results from short- and long-term analyses of stock market reactions to bank-specific reclassification announcements. We use the first reclassification announcement for reclassifying banks and, as benchmark announcements, the first earnings announcement for non-reclassifying banks following the official announcement of the amendment to IAS 39 in October 2008. Since these dates cannot be identified for all sample banks, the analyses in this table are based on a reduced sample of 117 reclassifying and 161 non-reclassifying banks. 14 (67) (36) banks make the reclassification announcement before (during) (after) the respective earnings announcement. 78 (39) banks announce reclassifications in (interim reports prior to) the first annual report following the amendment to IAS 39. Panel A reports regressions that relate the abnormal announcement return to various cross-sectional determinants. Abnormal return (DJ STOXX) is the prediction error from the market model using the DJ STOXX 1800 market index, with interval (-60, 11) and interval (+11, +60) relative to announcement day 0 as estimation window. We follow the trade-to-trade approach by Maynes and Rumsey (1993) to account for thin trading in some of the stocks. Reclassification equals one for banks that announce reclassifications, and zero otherwise. Earnings Surprise is an indicator variable takes a value of one if the earnings number reported at an earnings announcement is higher than the mean DQDO\VWV¶ forecast, and zero otherwise. Regulatory Capital Restriction (Continuous) , Regulatory Capital Restriction (Dummy ± Cut-off 2%), and Regulatory Capital Restriction (Dummy ± Cut-off 0%) are described in table 4.6.3 and table 4.6.6, respectively. Panel B reports mean market-adjusted returns for up to 120 days before and after the reclassification/earnings announcements. Market-adjusted returns are buy-and-hold bank returns minus buy-and-hold returns on the DJ STOXX 1800 market index. Mean market-adjusted returns are presented for the total sample and subsamples based on the reclassification choice and the impact of this choice on the tier 1 capital ratio. Both panels report OLS coefficient estimates and t-statistics (in parentheses). The t-statistics are based on robust standard errors. ***, **, * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

129

4.6.8 Long-term effects of reclassifications on information asymmetry Panel A: Descriptive Statistics Variables

# Quarters

Mean

P1

P25

3.34%

0.04%

0.28%

0.25%

0.39%

0.00%

0.03%

0.10%

0.32%

1.71%

7,073.28

16,532.68

24.12

327.21

1,354.02

6,117.74

86,607.88

1.57%

0.64%

1.60%

2.26%

3.09%

8.31%

Bid-Ask Spread

3,467

1.60%

Share Turnover

3,467

Market Value (m Euro)

3,467

Return Variability

3,467

2.59%

Std. Dev.

Median 0.66%

P75

P99

1.55%

16.05%

Panel B: All Quarters Independent Variables Reclassification

Dependent Variable: Log(Bid-Ask Spread) (1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

0.077

-0.113

0.143

-0.144

0.042

-0.090

0.142

-0.073

(1.13)

(-1.76)*

(1.76)*

(-1.98)**

(0.65)

(-1.47)

(1.87)*

(-1.01)

Effect on Net Income

0.359

0.544

0.250

0.407

(3.43)***

(4.31)***

(2.45)**

(3.28)***

Complete Disclosure

-0.210

0.094

-0.319

-0.052

(-1.99)**

(0.83)

(-3.28)***

(-0.49)

Effect on Net Income x Complete Disclosure

-0.580

-0.501

(-2.98)***

(-2.80)***

Log(Share Turnover) Log(Market Value) Log(Return Variability)

-0.189

-0.184

-0.194

-0.187

(-10.40)***

(-10.18)***

(-10.32)***

(-9.95)***

-0.316

-0.313

-0.322

-0.315

(-6.09)***

(-6.18)***

(-6.31)***

(-6.44)***

0.293

0.281

0.305

0.292

(7.40)***

(7.08)***

(7.71)***

(7.44)***

Selected Quarter Dummies (Benchmark: 2007Q1) Dummy 2007Q2 Dummy 2007Q4 Dummy 2008Q2 Dummy 2008Q4 Dummy 2009Q2 Dummy 2009Q4

Fixed Effects R-squared # Observations

-0.035

-0.035

-0.035

-0.035

0.015

0.012

0.016

0.014

(-1.56)

(-1.59)

(-1.57)

(-1.59)

(0.63)

(0.54)

(0.70)

(0.61)

0.031

0.033

0.032

0.035

-0.004

-0.002

-0.003

0.000

(1.12)

(1.19)

(1.17)

(1.27)

(-0.17)

(-0.07)

(-0.13)

(0.01)

0.143

0.144

0.144

0.146

0.011

0.015

0.008

0.015

(4.26)***

(4.29)***

(4.28)***

(4.36)***

(0.31)

(0.42)

(0.24)

(0.42)

0.787

0.777

0.785

0.777

0.278

0.285

0.260

0.273

(17.96)***

(17.75)***

(17.94)***

(17.81)***

(4.85)***

(4.97)***

(4.44)***

(4.70)***

0.550

0.554

0.552

0.556

0.183

0.194

0.177

0.191

(10.69)***

(10.76)***

(10.73)***

(10.78)***

(3.43)***

(3.65)***

(3.33)***

(3.61)***

0.297

0.300

0.298

0.303

0.046

0.054

0.043

0.053

(5.43)***

(5.49)***

(5.45)***

(5.52)***

(0.93)

(1.08)

(0.86)

(1.08)

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

0.91

0.91

0.91

0.91

0.92

0.92

0.92

0.92

3,467

3,467

3,467

3,467

3,467

3,467

3,467

3,467

Additional Tests Reclassification + Effect on Net Income Reclassification + Complete Disclosure Reclassification + Effect on NI + Compl. Discl. + Interaction Term

130

0.246

0.400

0.160

(2.53)**

(3.50)***

(1.69)*

-0.067

-0.050

(-0.78)

(-0.48) -0.086 (-0.70)

0.334 (3.04)*** -0.177

-0.125

(-2.28)** (-1.39) -0.219 (-2.01)**

Panel C: Without first two reclassification quarters Independent Variables Reclassification

Dependent Variable: Log(Bid-Ask Spread) (1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

0.110

-0.124

0.218

-0.140

0.095

-0.076

0.231

-0.044

(1.34)

(-1.58)

(2.21)**

(-1.61)

(1.25)

(-1.07)

(2.55)**

(-0.54)

Effect on Net Income

0.432

0.675

0.316

0.517

(3.51)***

(4.57)***

(2.62)***

(3.57)***

Complete Disclosure

-0.333

0.051

-0.417

-0.097

(-2.71)***

(0.35)

(-3.64)***

(-0.73)

Effect on Net Income x Complete Disclosure

-0.727

-0.595

(-3.17)***

(-2.78)***

Log(Share Turnover) Log(Market Value) Log(Return Variability)

Fixed Effects R-squared # Observations Additional Tests Reclassification + Effect on Net Income Reclassification + Complete Disclosure Reclassification + Effect on NI + Compl. Discl. + Interaction Term

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

-0.190

-0.184

-0.195

-0.187

(-10.39)***

(-10.06)***

(-10.45)***

(-9.92)***

-0.328

-0.325

-0.328

-0.318

(-6.03)***

(-6.15)***

(-6.23)***

(-6.34)***

0.288

0.279

0.295

0.286

(7.00)***

(6.79)***

(7.27)***

(7.12)***

Firm, Quarter

Firm, Quarter

Firm, Quarter

Firm, Quarter

0.91

0.91

0.91

0.91

0.92

0.92

0.92

0.93

3,220

3,220

3,220

3,220

3,220

3,220

3,220

3,220

0.308

0.535

0.240

0.473

(2.69)***

(3.95)***

(2.15)**

(3.59)***

-0.115

-0.089

-0.186

-0.141

(-1.18)

(-0.66)

(-2.08)**

(-1.20)

-0.141

-0.219

(-1.09) (-1.85)* Notes: This table presents results from regressions that relate bid-ask spreads to the effect of IFRS reclassifications. Panel A reports descriptive statistics for the dependent as well as the control variables. All variables are measured at the firm-quarter level. The sample comprises of 124 reclassifying and 178 non-reclassifying banks, which results in a total of 3,467 firm-quarter observations over the period 2007/Q1 to 2009/Q4. Panel B present results from multi-period difference-in-differences analyses using the data for all quarters. In panel C, we drop the first two reclassification quarters to examine the long-term impact of the reclassification choice on bid-ask spreads. Bid-Ask Spread is the median quoted spread (i.e. the difference between the closing bid and the closing ask price divided by the midpoint, source: Thomson Reuters Datastream). Reclassification equals one for all reclassification quarters starting with the first quarter during which the respective bank announced IFRS reclassifications, and zero otherwise. Effect on Net Income is a dummy variable that takes a value of one if the percentage net income effect of the reclassification is above the sample median (see table 4.6.2), and zero otherwise. Complete Disclosure indicates whether a reclassifying bank discloses all six items required by IFRS 7, para. 12A, in the footnotes to its financial statements (see table 4.6.1). Share Turnover is the average percentage trading volume (i.e. trading volume in units divided by the number of outstanding shares, source: Thomson Reuters Datastream). Market Value (mEuro) is the median market value of outstanding equity in Million Euros (source: Thomson Reuters Datastream). Return Variability is the standard deviation of daily stock returns (source: Thomson Reuters Datastream). Panel B and C report OLS coefficient estimates and t-statistics (in parentheses). The t-statistics are based on standard errors that are clustered by firm. We use the natural logarithm of the raw values where indicated in the panels. ***, **, * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

131

5 Summary and Conclusions This thesis comprises three essays on the economic consequences of mandatory IFRS reporting around the world. While the first two essays focus on reactions to the adoption process, the third essay analyzes the effects of a subsequent change in IFRS accounting rules. The first essay is a literature review on the economic consequences of mandatory IFRS adoption in the EU. Based on the explicitly stated objectives of the IAS Regulation, we distinguish between intended and unintended consequences. We document that empirical research on the intended consequences of mandatory IFRS adoption generally fails to find an increase in the comparability or transparency of financial statements. In contrast, evidence of positive reactions in capital markets and at the macroeconomic level is plentiful and almost unanimous. We argue that this disagreement in research findings is likely due to the difficulty of separating a potential IFRS effect from concurrent changes that are unrelated to financial reporting. Direct empirical evidence on the unintended consequences is still scarce. However, extant research and insights from non-IFRS settings suggest that mandatory IFRS adoption substantially affects contractual outcomes and provokes opportunistic anticipatory actions. We conclude from our literature review that both the intended and the unintended consequences of mandatory IFRS adoption deserve further scrutiny and provide guidance for future research. In the second essay, we empirically test the prediction that mandatory IFRS adoption enhances cross-border equity investments of individual investors. We test this prediction by analyzing trading activity in the Open Market. The Open Market is a trading segment at Frankfurt Stock Exchange and provides German individual investors with a costefficient alternative to trade a large selection of foreign (i.e., non-German) stocks. Using a global sample of 4,869 firms, we find that stocks experience an increase in Open Market trading activity following mandatory adoption of IFRS. This increase is both economically and statistically significant, and robust to a multitude of alternative regression specifications. In summary, our results are consistent with the idea that collective IFRS adoption enhances cross-border equity investments of individual investors, although we cannot fully rule out alternative explanations. The third essay analyzes the economic consequences of fair value reclassifications following the emergency amendment to IAS 39. The IASB issued this amendment in October 2008 after severe political pressure by EU leaders to relax fair value accounting 133

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1_5, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

rules. Using a sample of 302 publicly listed IFRS banks from around the world, we find that 124 banks choose to take the reclassification option in the first annual report after the amendment was issued. These fair value reclassifications increase aggregate profits by a total of 22.7 billion Euros and firm-specific profits by 44% on average. Accounting choice tests show that banks are more likely to reclassify if they are close to violating regulatory capital restrictions. The most troubled of these banks experience positive abnormal stock returns around the reclassification announcement. We further find that two thirds of the reclassifying banks do not fully comply with the accompanying footnote disclosures that IFRS 7 requires. These non-complying banks experience a significant increase in bid-ask spreads, especially if the reclassifications substantially affect net income. In sum, we provide evidence that the emergency amendment to IAS 39 facilitates regulatory arbitrage by economically weak banks. However, the fair value reclassifications are also attended by a high degree of non-compliance with disclosure requirements and, thus, a decrease in the transparency of financial statements. Taken together, this thesis shows that mandatory IFRS reporting around the world has the potential to produce both intended and unintended consequences. On the one hand, we provide evidence that a single set of global accounting standards enhances crossborder equity investments by individual investors. On the other hand, we find that IFRS are not always complied with. Uneven implementation of IFRS and its impact on accounting-based contracts whose features vary substantially across countries are likely to dampen the benefits of uniform accounting standards. Moreover, the example of the emergency amendment to IAS 39 raises doubts whether the IASB will be able to maintain its position as an independent standard-setter. Since mandatory IFRS reporting around the world is a recent phenomenon, its longterm economic consequences have yet to emerge. The findings presented in this thesis are therefore inherently preliminary, but hopefully stimulate further research to establish a balanced view on the pros and cons of IFRS as a single set of global accounting standards.

134

References Abreu, M., V. Mendes, and J. A. C. Santos, 2009. Home country bias: does domestic experience help investors enter foreign markets? Working paper, Instituto Superior de Economia e Gestao. AcSB, 2010. Exposure draft: adoption of IFRSs by entities with rate-regulated activities. Canadian Accounting Standards Board AGFRC, 2002. Adoption of International Accounting Standards by 2005. Australian Government Financial Reporting Council, Bulletin 2002/4 of 3 July 2002. Aharony, J., R. Barniv, and H. Falk, 2010. The impact of mandatory IFRS adoption on equity valuation of accounting numbers for security investors in the EU. European Accounting Review 19 (3): 535-578. Ahmed, A. S., E. Kilic, and G. J. Lobo, 2006. Does recognition versus disclosure matter? Evidence from value-relevance of EDQNV¶Uecognized and disclosed derivative financial instruments. The Accounting Review 81 (3): 567±588. Ahmed, A. S., M. Neel, and D. Wang, 2009. The effects of mandatory adoption of International Financial Reporting Standards on smoothness, conservatism and timeliness of accounting earnings. Working paper, Texas A & M University. Ai, C. and E. C. Norton, 2003. Interaction terms in logit and probit models. Economics Letters 80 (1): 123±129. Amihud, Y. and H. Mendelson, 1986. Asset pricing and the bid-ask spread, Journal of Financial Economics 17 (2): 223-249. Amir, E., Y. Guan, and D. Oswald, 2010. The effect of pension accounting on corporate pension asset allocation. Review of Accounting Studies 15 (2): 345-366. Amiram, D., 2009. Financial information globalization and foreign investment decisions. Working paper, University of North Carolina. Armstrong, C. S., Barth, M. E., Jagolinzer, A. D., and E. J. Riedl, 2010. Market reaction to the adoption of IFRS in Europe. The Accounting Review 85 (1): 31-61. Armstrong, M. S., 1977. The politics of establishing accounting standards. Journal of Accountancy 143 (2): 76-79. Bae, K.-H., H. Tan, and M. Welker, 2008. International GAAP Differences: the impact on foreign analysts. The Accounting Review 83 (3): 593-628. Bailey, W., A. Kumar, and D. Ng, 2008. Foreign investments of US. individual investors: causes and consequences. Management Science 54 (3): 443-459. Ball, R., 2006. International Financial Reporting Standards (IFRS): pros and cons for investors. Accounting and Business Research 36, Special Issue: 5±27. Ball, R., S. P. Kothari, and A. Robin, 2000. The effect of international institutional factors of properties of accounting earnings. Journal of Accounting and Economics 29 (1): 1-51. 135

U. Brüggemann, Essays on the economic consequences of mandatory IFRS reporting around the world, DOI 10.1007/978-3-8349-6952-1, © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2011

Ball, R., A. Robin, and J. S. Wu, 2003. Incentives versus standards: properties of accounting income in four East Asian countries. Journal of Accounting and Economics 36 (1±3): 235±270. Barber, B. M. and T. Odean, 2008. All that glitters: the effect of attention and news on the buying behavior of individual and institutional Investors. The Review of Financial Studies 21 (2): 785-818. Barth, J. R., G. Caprio Jr., and R. Levine, 2001. The regulation and supervision of banks around the word ± a new database. The World Bank. Barth, J. R., G. Caprio Jr., and R. Levine, 2006. Rethinking bank regulation ± till angels govern. Cambridge: Cambridge University Press. Barth, M. E., 1994. Fair value accounting: evidence from investment securities and the market valuation of banks. The Accounting Review 69 (1): 1±25. Barth, M. E., 2006. Including estimates of the future in tRGD\¶V financial statements. Accounting Horizons 20 (3): 271±285. Barth, M. E., W. H. Beaver, and W. R. Landsman, 1996. Value-relevance of bDQNV¶fair value disclosures under SFAS No. 107. The Accounting Review 71 (4): 513±537. Barth, M. E., W. H. Beaver, and W. R. Landsman, 2001. The relevance of the value relevance literature for financial accounting standard setting: another view. Journal of Accounting and Economics 31 (1-3): 77-104. Barth, M. E., G. Clinch, and T. Shibano, 1999. International accounting harmonization and global equity markets. Journal of Accounting and Economics 26 (1-3): 201-235. Barth, M. E., J. A. Elliott, and M. W. Finn, 1999. Market rewards associated with patterns of increasing earnings. Journal of Accounting Research 37 (2): 387-413. Barth, M. E. and W. R. Landsman, 2010. How did financial reporting contribute to the financial crisis? European Accounting Review: forthcoming. Bartlett, S. A., and R. A. Chandler. The corporate report and the private shareholder: Lee and Tweedie twenty years on. British Accounting Review 29 (3): 245-261. Bartov, E., D. Givoly, and C. Hayn, 2002. The rewards for meeting-or-beating earnings expectations. Journal of Accounting and Economics 33 (2): 173±204. Beatty, A., 2007. How does changing measurement change management behavior? A review of the evidence. Accounting and Business Research 37: 63-71. Beatty, A. L., S. L. Chamberlain, and J. Magliolo, 1995. Managing financial reports of commercial banks: the influence of taxes, regulatory capital, and earnings. Journal of Accounting Research 33 (2): 231±261. Beatty, A. L., S. L. Chamberlain, and J. Magliolo, 1996. An empirical analysis of the economic implications of fair value accounting for investment securities. Journal of Accounting and Economics 22 (1±3): 43±77.

136

Beatty, A. L., Ke, B., and K. R. Petroni, 2002. Earnings management to avoid earnings declines across publicly and privately held banks. The Accounting Review 77 (3): 547-570. Beneish, M. D., B. P. Miller, and T. L. Yohn, 2010. The effect of IFRS adoption on cross-border investment in equity and debt markets. Working paper, Indiana University. Beneish, M. D. and T. L. Yohn, 2008. Information frictions and investors home bias: a perspective on the effect of global IFRS adoption on the extent of equity home bias. Journal of Accounting and Public Policy 27 (6): 433-443. Bernard, V. and K. Schipper, 1994. Recognition and disclosure in financial reporting. Working Paper, University of Michigan. Bertrand, M., E. Duflo, and S. Mullainathan, 2004. How much should we trust differences-in-differences estimates? Quarterly Journal of Economics 119 (1): 249± 275. Beuselinck, C., P. Joos, I. Khurana, and S. Van der Meulen, 2010. Mandatory IFRS reporting and stock price informativeness. Working paper, Tilburg University. Bhat, G., J. L. Callen, and D. Segal, 2010. Credit risk and IFRS: the case of credit default swaps. Working paper, Washington University. Bhattacharya, N., 2001. ,QYHVWRUV¶ WUDGH VL]H DQG Wrading responses around earnings announcements: an empirical investigation. The Accounting Review 76 (2): 221-244. Biondi, Y. and T. Suzuki, 2007. Socio-economic impacts of international accounting standards: an introduction. Socio-Economic Review 5 (4): 585-602. BIS Bank for International Settlements, 2009. The role of valuation and leverage in procyclicality. Committee on the Global Financial System Paper No. 34. Blaufus, K. and D. Lorenz, 2009. Wem droht die Zinsschranke? Eine empirische Untersuchung zur Identifikation der Einflussfaktoren. Zeitschrift für Betriebswirtschaft 79: 503-526. Bluethgen, R., A. Gintschel, A. Hackethal, and A. Müller, 2008. Financial advice and individual iQYHVWRUV¶portfolios. Working Paper, European Business School. Bova, F. and R. Pereira, 2010. The determinants and consequences of heterogeneous IFRS compliance levels following mandatory IFRS adoption: evidence from a developing country. Working paper, University of Toronto. Bowen, R. M., U. Khan, and S. Rajgopal, 2009. The economic consequences of relaxing fair value accounting and impairment rules on banks during the financial crisis of 2008-2009. Working Paper, University of Washington. Brown, P. R. and A. Tarca, 2005. A commentary on issues relating to the enforcement of International Financial Reporting Standards in the EU. European Accounting Review 14 (1): 181±212.

137

Brown, P. R., J. P. Preiato, and A. Tarca, 2009. Mandatory IFRS and properties of DQDO\VWV¶IRUHFDsts: how much does enforcement matter? Working paper, University of Western Australia. Burgstahler, D., L. Hail, and C. Leuz, 2006. The importance of reporting incentives: earnings management in European private and public firms. The Accounting Review 81 (5): 983-1016. Bushman, R. and W. R. Landsman, 2010. The pros and cons of regulating corporate reporting: a critical review of the arguments. Accounting and Business Research: forthcoming. Cairns, D., 1999. The FT International Accounting Standards survey. Financial Times. London. Cascino, S. and J. Gassen, 2010. Mandatory IFRS adoption and accounting comparability. Working paper, London School of Economics. CEBS, 2007. Analytical report on prudential filters for regulatory capital. Committee of European Banking Supervisors, London. CEBS, 2008. Joint statement from CESR, CEBS and CEIOPS regarding the latest developments in accounting. Committee of European Banking Supervisors, Paris/London/Frankfurt am Main. CESR, 2007. C(65¶VUHYLHZRIWKHLPSOHPHQWDWLRQDQd enforcement of IFRS in the EU. Committee of European Securities Regulators, Paris, Ref: CESR/07-352. CESR, 2009. Application of and disclosures related to the reclassification of financial instruments. Committee of European Securities Regulators, Paris, Ref.: CESR/09-575. CFA Institute, 2008. Re: Amendments to International Accounting Standards no 39 (IAS 39). Letter to the EC Accounting Regulatory Committee, London, October 14. Chang, Y.-L., C.-C. Liu, and S. G. 5\DQ  6)$6 1R ¶V fair value option: eventually used as intended. Working Paper, National Taiwan University. Chen, H., Q. Tang, Y. Jiang, and Z. Lin, 2010. The role of International Financial Reporting Standards in accounting quality: evidence from the European Union. Journal of International Financial Management and Accounting 21 (3): 220-278. Chen, K. C. W. and F. Tang, 2009. Do firms use the unrealized gains mandated by IFRS to increase executive cash compensation? Evidence from family-owned property companies in Hong Kong. Working paper, Hong Kong University of Science and Technology. Christensen, H. B., E. Lee, and M. Walker, 2007: Cross-sectional variation in the economic consequences of international accounting harmonization: the case of mandatory IFRS adoption in the UK. The International Journal of Accounting 42 (4): 341-379. Christensen, H. B., E. Lee, and M. Walker, 2008. Incentives or standards: what determines accounting quality changes around IFRS adoption? Working paper, University of Chicago. 138

Christensen, H. B., E. Lee, and M. Walker, 2009. Do IFRS reconciliations convey information? The effect of debt contracting. Journal of Accounting Research 47 (5): 1167-1199. Christensen, H. B. and V. Nikolaev, 2009. Does fair value accounting for non-financial assets pass the market test? Working Paper, University of Chicago. Citron, D. B., 1992. Accounting measurement rules in UK bank loan contracts. Accounting and Business Research 23 (89): 21-30. Clarkson, P., J. D. Hanna, G. D. Richardson, and R. Thompson, 2010. The impact of IFRS adoption on the value relevance of book value and earnings. Working paper, Simon Fraser University. Comprix, J. J., K. A. Muller III, and M. Stanford, 2003. Economic consequences from mandatory adoption of IASB standards in the European Union. Working paper, Arizona State University. Cornett, M. M., Z. Rezaee, and H. Tehranian, 1996. An investigation of capital market reactions to pronouncements on fair value accounting. Journal of Accounting and Economics 22 (1±3): 119±154. DAI, 2008. Factbook 2008. Deutsches Aktieninstitut. Frankfurt am Main. Daske, H., 2006. Economic benefits of adoption IFRS or US-GAAP ± have the expected cost of equity capital really decreased? Journal of Business Finance and Accounting 33 (3-4): 329-373. Daske, H., L. Hail, C. Leuz, and R. S. Verdi, 2008. Mandatory IFRS reporting around the world: early evidence on the economic consequences. Journal of Accounting Research 46 (5): 1085-1142. Daske, H., L. Hail, C. Leuz, and R. S. Verdi, 2009. Adopting a label: heterogeneity in the economic cRQVHTXHQFHVRI,)56DGRSWLRQV¶:RUNLQg paper, University of Mannheim. De Bondt, W. F. M., 1998. A portrait of the individual investor. European Economic Review 42 (3-5): 831-844. de Jong, A., M. Rosellón, and P. Verwijmeren, 2006. The economic consequences of IFRS: the impact of IAS 32 on preference shares in the Netherlands. Accounting in Europe 3 (1): 169-185. de las Heras, E., J. A. Moreira, and P. Teixeira, 2010. Are institutional factors influencing accounting policy choices under IFRS? A European perspective. Working paper, Universidad Autonoma de Madrid. Dechow, P. M., A. P. Hutton, and R. G. Sloan, 1996. Economic consequences of accounting for stock-based compensation. Journal of Accounting Research 34, Supplement: 1±20. DeFond, M., X. Hu, M. Hung, and S. Li, 2009. The impact of IFRS adoption on US mutual fund ownership: the role of comparability. Working paper, University of Southern California. 139

Degeorge, F., J. Patel, and R. J. Zeckhauser, 1999. Earnings management to exceed thresholds. Journal of Business 72 (1): 1±33. Deutsche Bundesbank, 2005. Securities dHSRVLWV¶ special statistical publication 9. Frankfurt am Main. Dewing, I., and P. O. Russell, 2008. Financial integration in the EU: the first phase of EU endorsement of International Accounting Standards. Journal of Common Market Studies 46 (2): 243-264. Diamond, D. W. and R. E. Verrecchia, 1991. Disclosure, liquidity, and the cost of capital. Journal of Finance 46 (4): 1325±1359. Ding, Y., O.-K. Hope, T. Jeanjean, and H. Stolowy, 2007. Differences between domestic accounting standards and IAS: Measurement, determinants and implications. Journal of Accounting and Public Policy 26 (1): 1-38. Dixon, A. D., and A. H. B. Monk, 2009. The power of finance: accounting KDUPRQL]DWLRQ¶V effect on pension provision. Journal of Economic Geography 9 (5): 619-639. Durtschi, C., and P. D. Easton, 2005. Earnings management? The shapes of the frequency distributions of earnings metrics are not evidence ipso facto. Journal of Accounting Research 43 (4): 557-592. EC, 2002. Regulation No 1606/2002 of the European Parliament and of the Council of 19 July 2002 on the application of international accounting standards. Official Journal of the European Communities, L 243: 1-4. ECB, 2004. Fair value accounting and financial stability. European Central Bank, Occasional Paper Series No. 13. Eccher, E. A., K. Ramesh, and S. R. Thiagarajan, 1996. Fair value disclosures by bank holding companies. Journal of Accounting and Economics 22 (1±3): 79±117. ECOFIN, 2007. 2822nd meeting of the Council of the European Union. Economic and Financial Affairs, Luxembourg, 9 October. Draft Minutes 13661/07. ECOFIN, 2008a. 2882nd meeting of the Council of the European Union. Economic and Financial Affairs, Brussels, 8 July. Press Release 11236/08. ECOFIN, 2008b. 2894th meeting of the Council of the European Union. Economic and Financial Affairs, Luxembourg, 7 October. Press Release 13784/08 Elliott, W. B., F. D. Hodge, and K. E. Jackson, 2008. The association between nonprofessional invesWRUV¶ information choices and their portfolio returns: the importance of investing experience. Contemporary Accounting Research 25 (2): 473498. Endres, D., A. Oestreicher, W. Scheffler, and C. Spengel, 2007. The determination of corporate taxable income in the EU member states. Kluwer Law International, Alphen aan den Rijn. Ernst & Young, 2007. IFRS: observations on the implementation of IFRS. London. 140

Ernst, E., J. Gassen, and B. Pellens, 2009. Verhalten und Präferenzen deutscher $NWLRQlUH´6WXGLHQGHs Deutschen Aktieninstituts 42. Frankfurt am Main. European G8 Members, 2008. Statement at the summit of European G8 members. Palais GHO¶(O\VHH4 October. FESE, 2008a. Share ownership structure in Europe. Federation of European Securities Exchanges, Brussels. FESE, 2008b. Alternative markets/segments in equity analysis. Federation of European Securities Exchanges, Brussels. Fiechter, P., 2010. Reclassification of financial assets. Accounting in Europe: forthcoming. Financial Stability Forum, 2009. Report of the Financial Stability Forum on addressing procyclicality in the financial system. 2 April. FINRA, 2008. Remarks by Mary L. Shapiro, FINRA Fall Securities Conference, 23 October. Financial Industry Regulatory Authority. Florou, A. and U. Kosi, 2009. The economic consequences of mandatory IFRS adoption for debt financing. Working paper, University of Macedonia. Florou, A. and P. F. Pope, 2009. Mandatory IFRS adoption and investor asset allocation decisions. Working paper, University of Macedonia. Frankel, R. and X. Li&KDUDFWHULVWLFVRIDILUP¶VLQIRUPDWLRQHQYLURQPHQWDQGWKH information asymmetry between insiders and outsiders. Journal of Accounting and Economics 37 (2): 229±259. Freihube, T., C.-H. Kehr, J. P. Krahnen, and E. Theissen, 1999. Was leisten die Kursmakler: Eine empirische Untersuchung am Beispiel der Frankfurter Wertpapierbörse. Kredit und Kapital 32: 426-460. French, K., 2008. The cost of active investing. Journal of Finance 63 (4): 1537-1573. FSE, 2000. Factbook 2000. Deutsche Börse Group, Frankfurt am Main. FSE, 2008. Factbook 2008. Deutsche Börse Group, Frankfurt am Main. Fülbier, R. U., J.-M. Hitz, and T. Sellhorn, 2009. Relevance of academic research and researchers' role in the IASB's financial reporting standard setting. Abacus 45 (4): 455492. Gaeremynck, A., D. B. Thornton, and A. Verriest, 2009. Quality of IFRS adoption. Working paper, Catholic University of Leuven. Gao, F., J. S. Wu, and J. L. Zimmerman, 2009. Unintended consequences of granting small firms exemptions from securities regulation: evidence from the Sarbanes-Oxley Act. Journal of Accounting Research 47 (2): 459-506. Garcia Lara, J. M., B. Garcia Osma, and B. Gill de Albornoz Noguer, 2006. Effects of database choice on international accounting research. Abacus 42 (3-4): 426-454. 141

Garcia Osma, B. and P. F. Pope, 2009. Earnings quality effects of mandatory IFRS adoption. Working paper, Lancaster University. Gassen, J., 2008. Are stewardship and valuation usefulness compatible or alternative objectives of financial accounting? Working paper, Humboldt-University Berlin. Gebhardt, G. and Z. Novotny-Farkas, 2010. The effects of IFRS adoption on the financial reporting quality of European banks. Working paper, Goethe University of Frankfurt am Main. Gee, M., A. Haller, and C. Nobes, 2010. The influence of tax on IFRS consolidated statements: the convergence of Germany and the UK. Accounting in Europe 7 (1): 97122. Gerhardt, R., and A. Hackethal, 2009. The influence of financial advisors on household portfolios: a study on private investors switching to financial advice. Working paper, Goethe University of Frankfurt am Main. Giner, B. and M. Arce, 2004. Lobbying on accounting standards: the due process of IFRS 2 on share-based payments. Working paper, University of Valencia. Gkougkousi, X. and G. Mertens, 2010. Capital market effects of mandatory IFRS adoption in the financial sector. Working paper, Rotterdam School of Management. Glosten, L. R. and L. E. Harris, 1988. Estimating the components of the bid/ask spread. Journal of Financial Economics 21 (1): 123±142. Goldberg, L. G. and S. C. Hudgins, 2002. Depositor discipline and changing strategies for regulating thrift institutions. Journal of Financial Economics 63 (2): 263±274. Goncharov, I. and S. van Triest, 2009. The strange case of the world record profit and the missing dividends: unintended consequences of an accounting change. Working paper, Amsterdam Business School. Goncharov, I. and S. van Triest, 2011. Do fair value adjustments influence dividend policy? Accounting and Business Research, 42 (1): forthcoming. Graham, J. R., C. R. Harvey and H. Huang, 2009. Investors competence, trading frequency, and home bias. Management Science 55 (7): 1094-1106. Greil, R., 2005. Ab 29. Dezember werden in Frankfurt keine türkischen Aktien mehr gehandelt. Börse Online. Avaiable at: http://www.boerse-online.de/aktien/ deutschland_europa/481497.html?p=1. Hail, L. and C. Leuz, 2007. Capital market effects of mandatory IFRS reporting in the EU: empirical evidence. Report made available by the Netherlands Authority for the Financial Markets (AFM). Hail, L., C. Leuz, and P. D. Wysocki, 2010. Global accounting convergence and the potential adoption of IFRS by the US (part I): conceptual underpinnings and economic analysis. Accounting Horizons 24 (3): 355-394. Harrer, H. and R. Müller, 2006. Die Renaissance des Freiverkehrs ± Eine aktuelle Analyse mit internationalem Vergleich. Wertpapier-Mitteilungen: 653-696. 142

Healy, P. M., 1985. The impact of bonus schemes on the selection of accounting principles. Journal of Accounting and Economics 7 (1-3): 85-107. Heckman, J., 1979. Sample selection bias as a specification error. Econometrica 47 (1): 153-161. Hiller von Gaertringen, C., 2006. Ausländische Aktien sind wieder gefragt. Frankfurter Allgemeine Zeitung, 23 March. Hitz, J.-M., 2007. The decision usefulness of fair value accounting ± a theoretical perspective. European Accounting Review 16 (2): 323-362. Hodder, L. D., P. E. Hopkins, and J. M. Wahlen, 2006. Risk-relevance of fair-value income measures for commercial banks. The Accounting Review 81 (2): 337±375. Holthausen, R. W., 2009. Accounting standards, financial reporting outcomes, and enforcement. Journal of Accounting Research 47 (2): 447-458. Holthausen, R. W. and R. W. Leftwich, 1983. The economic consequences of accounting choice. Journal of Accounting and Economics 5 (1): 77±117. Holthausen, R. W. and R. L. Watts, 2001. The relevance of the value-relevance literature for financial accounting standard setting. Journal of Accounting and Economics 31 (13): 3-75. Hope, O.-K., J. Jin, and T. Kang, 2006. Empirical evidence on jurisdictions that adopt IFRS. Journal of International Accounting Research 5 (2): 1-20. Horton, J. and G. Serafeim, 2009. Market reaction to and valuation of IFRS reconciliation adjustments: first evidence from the UK. Review of Accounting Studies: forthcoming. Horton, J., G. Serafeim, and I. Serafeim, 2009. Does mandatory IFRS adoption improve the information environment? Working paper, London School of Economics. House of Commons, 2008. Banking crisis: oral evidence by Sir David Tweedie, Paul Boyle and Michael Izza taken before the Treasury Committee on Tuesday, 11 November. London, HC 144±I. Huang, R. D. and H. R. Stoll, 1997. The components of the bid-ask spread: a general approach. Review of Financial Studies 10 (4): 995±1034. Huizinga, H. and L. Laeven, 2009. Accounting discretion of banks during a financial crisis. Working Paper, Tilburg University. IASB, 2008a. Reducing complexity in reporting for financial instruments. International Accounting Standards Board, London. IASB, 2008b. Reclassification of financial assets ± amendments to IAS 39 Financial Instruments: recognition and measurement and IFRS 7 Financial Instruments: disclosures. International Accounting Standards Board, London. IASC, 1997. Accounting for financial assets and financial liabilities. International Accounting Standards Committee, London. 143

IASCF, 2001. Framework for the preparation and presentation of financial statements. International Accounting Standards Committee Foundation, London. IIF, 2008. Report of the IIF Committee on market best practices: principles of conduct and best practice recommendations. Institute of International Finance, Washington, DC. Jackson, H. E., 2008. The impact of enforcement: a reflection. University of Pennsylvania Law Review Pennumbra 156 (229): 400-411. Jeanjean, T. and H. Stolowy, 2008. Do accounting standards matter? An exploratory analysis of earnings management before and after IFRS adoption. Journal of Accounting and Public Policy 27 (6): 480-494. Jensen, M. C. and W. H. Meckling, 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4): 305-360. Johnson, D., 2009. Political systems, lobbying, and IFRS adoption. Working paper, Massachusetts Institute of Technology. Joint Working Group of Standard Setters, 1999. Financial instruments and similar items. Draft Standard and Basis for Conclusions, London. Jones, J. J., 1991. Earnings management during import relief investigations. Journal of Accounting Research 29 (2): 193-228. Jordan, J. S., J. Peek, and S. Rosengren, 2000. The market reaction to the disclosure of supervisory actions: implications for bank transparency. Journal of Financial Intermediation 9 (3): 298±319. Jorissen, A., N. Lybaert, and K. Van de Poel, 2006. Lobbying towards a global standard setter ± do national characteristics matter? An analysis of comment letters written to the IASB. In: G. Gregoriou and M. Garber (eds.): International accounting - standards, regulations, and financial reporting. Amsterdam, Elsevier: 1-40. Kaufmann, D., A. Kraay, and M. Mastruzzi, 2009. Governance matters VIII: aggregate and individual governance indicators 1996-2008. The World Bank. Kessler, G., 2008. Accounting standards wilt under pressure. Washington Post, 27 December. Kim, Y. and S. Li, 2010. Mandatory IFRS adoption and intra-industry information transfers. Working paper, Santa Clara University. Kiosse, V. and K. Peasnell, 2009. Have changes in pension accounting changed pension provision? A review of the evidence. Accounting and Business Research 39 (3): 255267. Königsgruber, R., 2009. A political economy of accounting standard setting. Journal of Management and Governance: forthcoming. Kothari, S.P., 2001. Capital markets research in accounting. Journal of Accounting and Economics 31 (1±3): 105±231. KPMG and I. von Keitz, 2006. The application of IFRS: choices in practice. London. 144

Kvaal, E. and C. Nobes, 2010. International differences in IFRS policy choice. Accounting and Business Research: forthcoming. Lagoarde-Segot, T., 2009. Financial reforms and time-varying microstructures in emerging equity markets. Journal of Banking and Finance 33 (10): 1755-1769. Lambert, R., C. Leuz, and R. E. Verrecchia, 2007. Accounting information, disclosure, and the cost of capital. Journal of Accounting Research 45 (2): 385-420. Landsman, W. R., 2007. Is fair value accounting information relevant and reliable? Evidence from capital market research. Accounting and Business Research 37, Special Issue: 19±30. Landsman, W. R., E. L. Maydew, and J. R. Thornock, 2010. The information content of annual earnings announcements and mandatory adoption of IFRS. Working paper, University of North Carolina. Lang, M. H. and M. G. Maffett, M, 2010. Transparency and liquidity uncertainty in crisis periods. Working Paper, University of North Carolina. Lang, M., M. Maffett, and E. Owens, 2010. Earnings comovement and accounting comparability: the effects of mandatory IFRS adoption. Working paper, University of North Carolina. Larson, R. K., 1997. Corporate lobbying of the International Accounting Standards Committee. Journal of International Financial Management and Accounting 8 (3): 175-203. Larson, R. K., 2007. &RQVWLWXHQW SDUWLFLSDWLRQ DQG WKH ,$6%¶V ,QWHUQDWLRQDO )LQDQFLDO Reporting Interpretations Committee. Accounting in Europe 4 (2): 207-254. Laux, C. and C. Leuz, 2009. The crisis of fair-value accounting: making sense of the recent debate. Accounting, Organizations and Society 34 (7±8): 826±834. Laux, C. and C. Leuz, 2010. Did fair-value accounting contribute to the financial crisis? Journal of Economic Perspectives 24 (1): 93±118. Lee, C. M. C., 1992. Earnings news and small traders: An intraday analysis. Journal of Accounting and Economics 15 (2-3): 265-302. Lee, T. A. and D. P. Tweedie, 1977. The private shareholder & the corporate report. The Institute of Chartered Accountants in England and Wales, London. Leftwich, R. W., 1983. Accounting information in private markets: evidence from private lending agreements. The Accounting Review 58 (1): 23-42. Leuz, C., 2003. IAS versus US-GAAP: information asymmetry-based evidence from *HUPDQ\¶V1HZ0DUNHW-RXUQDORI$FFRXQWLQJResearch 41 (3): 445-472. Leuz, C., 2006. &URVV OLVWLQJ ERQGLQJ DQG ILUPV¶ UHSRUWLQJ LQFHQWLYHV D GLVFXVVion of Lang, Raedy and Wilson. Journal of Accounting and Economics 42 (1-2): 285-299. Leuz, C., 2010. Different approaches to corporate reporting regulation: how jurisdictions differ and why. Accounting and Business Research: forthcoming. 145

Leuz, C., D. Deller, and M. Stubenrath, 1998. An international comparison of accounting-based payout restrictions in the United States, United Kingdom and Germany. Accounting and Business Research 28 (2): 111-129. Leuz, C., D. J. Nanda, and P. D. Wysocki, 2003. Earnings management and investor protection: an international comparison. Journal of Financial Economics 69 (3): 505527. Leuz, C., A. Triantis, and T. Y. Wang, 2008. Why do firms go dark? Causes and economic consequences of voluntary SEC deregistrations. Journal of Accounting and Economics 45 (2-3): 181-208. Leuz, C. and R. E. Verrecchia, 2000. The economic consequences of increased disclosure. Journal of Accounting Research 38 (3): 91±124. Leuz, C. and P. D. Wysocki, 2008. Economic consequences of financial reporting and disclosure regulation: a review and suggestions for future research. Working paper, University of Chicago. Li, N., 2010. Negotiated measurement rules in debt contracts. Journal of Accounting Research: forthcoming. Li, S., 2010. Does mandatory adoption of International Financial Reporting Standards in the European Union reduce the cost of equity capital? The Accounting Review 85 (2): 607-636. Liao, Q., T. Sellhorn, and H. A. Skaife, 2009. The cross-country comparability of IFRS earnings and book values: evidence from France and Germany. Working paper, University of Wisconsin-Madison. LSE, 2005. Stock exchange AIM notice: AIM Rules ± IAS confirmation & consultation London Stock Exchange, London. Lys, T., 1996. Abandoning the transactions-based accounting model: weighing the evidence. Journal of Accounting and Economics 22 (1±3): 155±175. Malmendier, U. and D. Shantikumar, 2007. Are small investors naïve about incentives? Journal of Financial Economics 85 (2): 457-489. Marosi, A. and N. Massoud, 2007. Why do firms go dark? Journal of Financial and Quantitative Analysis 42 (2): 421-442. Marquez-Ramos, L., 2008. The effect of IFRS adoption on trade and foreign direct investments. Working paper, Universitat Jaume I, Castellon. Maynes, E. and J. Rumsey, 1993. Conducting event studies with thinly traded stocks. Journal of Banking and Finance 17 (1): 145-157. McCreevy, C., 2005. IFRS: No pain, no gain? European Commissioner for Internal Market and Services Speech 05-621. Brussels, 18 October. McFadden, D. L., 1973. Conditional logit analysis of qualitative choice behavior. In: P. Zarembka (ed.): Frontiers in Econometrics. New York, Academic Press: 105±142. 146

Mohd, E., 2005. Accounting for software development costs and information asymmetry. The Accounting Review 80 (4): 1211±1231. Morais, A. I. and J. D. Curto, 2009. Mandatory adoption of IASB standards: value relevance and country-specific factors. Australian Accounting Review 19 (2): 128143. Morricone, S., R. Oriani, and M. Sobrero, 2009. The value relevance of intangible assets and the mandatory adoption of IFRS. Working paper, University of Bologna. Muller III, K. A. and E. J. Riedl, 2002. External monitoring of property appraisal estimates and information asymmetry. Journal of Accounting Research 40 (3): 865± 881. Muller, K. A., E. J. Riedl, and T. Sellhorn, 2010. Mandatory fair value accounting and information asymmetry: evidence from the European Real Estate Industry. Working paper, Pennsylvania State University. Müller-Michaels, O. and J. Wecker, 2005. Freiverkehr: gesetzliche Rahmenbedingungen und Börsenordnungen. Finanz Betrieb 11: 736-743. Nelson, K. K., 1996. Fair value accounting for commercial Banks: an empirical analysis of SFAS No. 107. The Accounting Review 71 (2): 161±182. Ng, J., 2009. Tax and non-tax incentives for voluntary IFRS adoption: evidence from the UK. Working paper, University of Chicago. Nini, G., D. C. Smith, and A. Sufi, 2009. Creditor control rights and firm investment policy. Journal of Financial Economics 92 (3): 400-420. Nobes, C. W., 2006. The survival of international differences under IFRS: towards a research agenda. Accounting and Business Research 36 (3): 233-245. Odean, T., 1998. Are investors reluctant to realize their losses? The Journal of Finance 53 (5): 1775-1798. Ormrod, P. and P. Taylor, 2004. The impact of the change to International Accounting Standards on debt covenants: a UK perspective. Accounting in Europe 1 (1): 71-94. Pae, J., D. B. Thornton, and M. Welker, 2008. Agency cost reduction associated with EU financial reporting reform. Journal of International Accounting Research 7 (1): 51-76. Pagratis, S. and M. Stringa, 2009. Modeling bank senior unsecured ratings: a reasoned structured approach to bank credit assessment. International Journal of Central Banking 5 (2): 1-39. Panaretou, A., M. Shackleton, and P. Taylor, 2009. Corporate risk management and hedge accounting. Working paper, Lancaster University. Pellens, B., J. Gassen, and M. Richard, 2003. Ausschüttungspolitik börsennotierter Unternehmen in Deutschland. Die Betriebswirtschaft 63 (3): 309-332. Plumlee, M., 2003. The effect of information complexity on anal\VWV¶ use of that information. The Accounting Review 78 (1): 275±296. 147

PWC, 2008. Application of IFRS to rate regulations in North American utilities companies. PricewaterhouseCoopers, Chicago. Ramanna, K., 2008. The implications of unverifiable fair-value accounting: Evidence from the political economy of goodwill accounting. Journal of Accounting and Economics 45 (2-3): 253-281. Ramanna, K. and S. Roychowdhury, 2010. Elections and discretionary accruals: evidence from 2004. Journal of Accounting Research 48 (2): 445±475. Ramanna, K. and E. Sletten, 2009. Why do countries adopt International Financial Reporting Standards? Working Paper, Harvard University. Ramesh, K. and L. Revsine, 2001. The effects of regulatory and contracting costs on EDQNV¶ FKRLFH RI DFFRXQWLQg method for other postretirement employee benefits. Journal of Accounting and Economics 30 (2): 159±186. Rixon, D. and A. Faseruk, 2009. Valuation in public sector agencies: impact on financial reporting through the implementation of international financial standards: focus on Canadian workers compensation boards. Journal of Financial Management and Analysis 22 (1): 16-27. Ronen, J., 2008. To fair value or not to fair value: A Broader Perspective. Abacus 44 (2): 181±208. Schipper, K., 2007. Required disclosures in financial reports. The Accounting Review 82 (2): 301±326. Schlitt, M. and S. Schäfer, 2006. Der neue Entry Standard der Frankfurter Wertpapierbörse. Die Aktiengesellschaft (5): 147-155. Schön, W., 2005. The odd couple: a common future for financial and tax Accounting? Tax Law Review 58: 111-148. SEC, 2008. SEC 2008 Report. US Securities and Exchange Commission. Sefcik, S. E. and R. Thompson, 1986. An approach to statistical inference in crosssectional models with security abnormal returns as dependent variable. Journal of Accounting Research 24 (2): 316±334. Sellhorn, T. and S. Gornik-Tomaszewski, 2006. ,PSOLFDWLRQVRIWKHµ,$65HJXODWLRQ¶IRU research into the international differences in accounting systems. Accounting in Europe 3 (1): 187-217. Shackelford, D. A. and T. Shevlin, 2001. Empirical tax research in accounting. Journal of Accounting and Economics 31 (1-3): 321-387. Shahzad, K., 2010. The quality of financial reporting under IFRS: evidence from credit ratings. Working paper, Erasmus University Rotterdam. Shen, C.-H. and H.-L. Chih, 2005. Investor protection, prospect theory, and earnings management: an international comparison of the banking industry. Journal of Banking and Finance 29 (10): 2675±2697. 148

Skinner, D. J., 2008. The rise of deferred tax assets in Japan: The role of deferred tax accounting in the Japanese banking crisis. Journal of Accounting and Economics 46 (2±3): 218±239. Smith Jr., C. W. and J. B. Warner, 1979. On financial contracting: an analysis of bond covenants. Journal of Financial Economics 7 (2): 117-161. Soderstrom, N. S. and K. J. Sun, 2007. IFRS adoption and accounting quality: a review. European Accounting Review 16 (4): 675-702. Song, C. J., W. Thomas, and H. Yi, 2010. Value relevance of FAS 157 fair value hierarchy information and the impact of corporate governance mechanisms. The Accounting Review: forthcoming. Soonawalla, K. and J. Ireland, 2010. The pooling of interests to end the pooling method in IFRS. International Journal of Accounting, Auditing and Performance Evaluation 6 (2-3): 129-157. Spiegel, M. M. and N. Yamori, 2007. Market price accounting and depositor discipline: The case of Japanese regional banks. Journal of Banking and Finance 31 (3): 769±786. Spooner, A., 2007. Fair Value and Financial Instruments. In: P. Walton (ed.): The Routledge Companion to Fair Value and Financial Reporting. London, Routledge: 370±384. Stadler, C. and S. Lobe, 2010. Voluntary pension funding in Germany and the effect of international accounting. Working paper, University of London (Royal Holloway). Stoll, H. R., 1978. The supply of dealer services in securities markets. Journal of Finance 33 (4): 1133±1151. Street, D. L. and S. J. Gray, 2001. Observance of International Accounting Standards: factors explaining non-compliance by companies referring to the use of IAS. ACCA Research Monograph. Sudmeyer, J., S. Rückert, and T. Kuthe, 2006. Entry Standard ± Das neue Börsensegment für den Mittelstand. Betriebs-Berater (50): 2703-2706. Sunder, S., 2009. IFRS and the accounting consensus. Accounting Horizons 23 (1): 101± 111. Sunder, S., 2010. Adverse effects of uniform written reporting standards on accounting practice, education, and research. Journal of Accounting and Public Policy 29 (2): 99114. Swinkels, L., 2006. Have pension plans changed after the introduction of IFRS? Working paper, Erasmus University Rotterdam. Tan, H., S. Wang, and M. Welker, 2009. Foreign analyst following and forecast accuracy around mandated IFRS adoptions. Working paper, University of Waterloo. Theissen, E., 2002. Floor versus screen trading: evidence from the German stock market. Journal of Institutional and Theoretical Economics 158 (2002): 32-54. 149

Tweedie, D., 2006. Prepared statement of Sir David Tweedie, Chairman of the IASB, before the Economic and Monetary Affairs Committee of the European Parliament on 31 January 2006, available at: http://www.iasplus.com/resource/0601tweedie euspeech.pdf. Veall, M. R. and K. F. Zimmermann, 1996. Pseudo-R2 measures for some common limited dependent variable models. Journal of Economic Surveys 10 (3): 241±259. Verrecchia, R. E. and J. Weber, 2006. Redacted disclosure. Journal of Accounting Research 44 (4): 791±814. Vogelheim, P., D. D. Schoenbachler, G. L. Gordon, and C. C. Gordon, 2001: The importance of courting the individual investor. Business Horizons 44 (JanuaryFebruary): 69-76. Vulcheva, M. I., 2009. International accounting standardization across countries with unequal enforcement ± questionable benefits at a high price? Dissertation proposal, Emory University. Wahlen, J. M., J. R. Boatsman, R. H. Herz, G. J. Jonas, K. G. Palepu, S. G. Ryan, K. Schipper, C. M. Schrand, and D. J. Skinner, 2000. Response to the FASB preliminary views: reporting financial instruments and certain related assets and liabilities at fair value. Accounting Horizons 14 (4): 501±508. Wang, S. and M. Welker, 2008. Timing equity issuance in response to information asymmetry arising from IFRS adoption in Australia and Europe. Working paper, Hong Kong University of Science and Technology. Wang, X., D. Young, and Z. Zhuang, 2008. The effects of mandatory adoption of International Financial Reporting Standards on information environments. Working paper, The Chinese University of Hong Kong. Watts, R. L. and J. L. Zimmerman, 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review 53 (1): 112-134. Watts, R. L. and J. L. Zimmerman, 1986. Positive accounting theory. Englewood Cliffs, NJ: Prentice Hall. Watts, R. L. and J. L. Zimmerman, 1990. Positive accounting theory: a ten year perspective. The Accounting Review 65 (1): 131±156. Wooldridge, J., 1995. Selection corrections for panel data methods under conditional mean independence assumptions. Journal of Econometrics 68 (1): 115-132. Wu, J. S. and I. X. Zhang, 2009a. The voluntary adoption of internationally recognized accounting standards and firm internal performance evaluation. The Accounting Review 84 (4): 1281-1310. Wu, J. S. and I. X. Zhang, 2009b. The adoption of internationally recognized accounting standards: implications for the credit markets. Working paper, University of Rochester.

150

Wu, J. S. and I. X. Zhang, 2010. Accounting integration and comparability: evidence from relative performance evaluation around IFRS adoption. Working paper, University of Rochester. Yu, G., 2009. Accounting standards and international portfolio holdings: analysis of cross-border holdings following mandatory adoption of IFRS. Working paper, University of Michigan at Ann Arbor. Zeff, S. A., 1978. The rLVHRIµHFRQRPLFFRQVHTXHQFHV¶ Journal of Accountancy 146 (6 December): 56-63. Zeff, S. A., 2006. Political lobbying on accounting standards ± national and international experience. In: C. Nobes and R. Parker (eds.): Comparative international accounting. Upper Saddle River, Prentice Hall: 189-218. Zhang, I. X., 2007. Economic consequences of the Sarbanes-Oxley Act of 2002. Journal of Accounting and Economics 44 (1±2): 74±115.

151