Handbook of Banking and Finance in Emerging Markets 1800880898, 9781800880894

Citation preview

HANDBOOK OF BANKING AND FINANCE IN EMERGING MARKETS

NGUYEN 9781800880894 PRINT.indd 1

21/09/2022 13:14

RESEARCH HANDBOOKS IN MONEY AND FINANCE The Research Handbooks in Money and Finance series presents a thorough analysis of recent scholarly developments in monetary and financial economics, forming an essential, authoritative and comprehensive reference guides to the field. Edited by esteemed international scholars, these Handbooks contain a wide range of specially commissioned chapters covering the latest advances and research, and aim to be prestigious, high-quality works of lasting significance. Each Handbook consists of original contributions by an international team of scholars, and contributes to both the expansion of current debates and the development of future research.

NGUYEN 9781800880894 PRINT.indd 2

21/09/2022 13:14

Handbook of Banking and Finance in Emerging Markets

Edited by

Duc Khuong Nguyen Professor of Finance, IPAG Business School, France

RESEARCH HANDBOOKS IN MONEY AND FINANCE

Cheltenham, UK • Northampton, MA, USA

NGUYEN 9781800880894 PRINT.indd 3

21/09/2022 13:14

© Duc Khuong Nguyen 2022 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA A catalogue record for this book is available from the British Library Library of Congress Control Number: 2022941088 This book is available electronically in the Economics subject collection http://dx.doi.org/10.4337/9781800880900

ISBN 978 1 80088 089 4 (cased) ISBN 978 1 80088 090 0 (eBook) Typeset by Cheshire Typesetting Ltd, Cuddington, Cheshire

EEP BoX NGUYEN 9781800880894 PRINT.indd 4

21/09/2022 13:14

Contents ix xxviii

List of contributors Preface PART I  FINANCIAL MARKETS, INSTITUTIONS AND MONEY   1 Booms, bubbles, blow-outs: exploring patterns in China’s credit expansion Ron McIver, Lei Xu and Shiao-Lan Chou

2

  2 Mutual fund investing in the Chinese A-share market Yeguang Chi and Xiao Qiao

32

  3 Liquidity and ex-dividend behavior in emerging markets Daniel Dupuis

51

  4 Asset-based valuation: a modified discounted cash flow approach Rafael Yanushevsky, Daniel Yanushevsky and Camilla Yanushevsky

70

  5 Financial integration in Asia: some empirical evidence An Thi Thuy Duong and Clemens Kool

83

  6 Application of the neural F-Score in Latin American stock markets Lidia Loban, Cristina Ortiz and Luis Vicente

104

  7 Market-liquidity risk modeling and reinforcement machine learning algorithms 115 under extreme market outlooks: applications to emerging markets Mazin A.M. Al Janabi   8 The components of bid–ask spread on the Warsaw Stock Exchange Paweł Miłobędzki and Sabina Nowak

131

PART II  BANKING PROFITABILITY, EFFICIENCY AND STABILITY   9 Bank profitability in the euro area: the asymmetric effect of common supervision153 Ioanna Avgeri, Yiannis Dendramis and Helen Louri 10 The journey of a thousand miles: a decade of the impact of foreign shareholders on the performance of Chinese commercial banks Constantin Gurdgiev and Li Jiaqi

176

11 The determinants of commercial banks’ profitability in the South-Eastern Europe region: a system GMM approach Francesco Guidi

200

v

NGUYEN 9781800880894 PRINT.indd 5

21/09/2022 13:14

vi  Handbook of banking and finance in emerging markets 12 Determinants of commercial banks’ performance in Mozambique Antonio Samagaio, Pedro Verga Matos and Isidora Manjate

218

13 Impact of information and communication technology on banking efficiency: the Vietnamese experience Thanh Ngo and Tu Le

238

14 Competition, efficiency and stability in Islamic and conventional banking systems: an emerging market perspective Md. Nurul Kabir and Andrew C. Worthington

254

15 Central bank independence, macroprudential policies and financial stability in the Mauritian context Manisha Chuttoor, Dinesh Ramdhony and Boopen Seetanah

274

PART III  TOWARDS FINANCIAL RESILIENCE AND SUSTAINABILITY 16 Sustainable investing in emerging markets Andreas Gruener

298

17 ESG issues in emerging markets and the role of banks Thankom Arun, Claudia Girardone and Stefano Piserà

321

18 Corporate social responsibility disclosure and cost efficiency of Islamic banks: evidence from GCC countries Anas Mohammad Hussein Al-Jbour, Lei Xu, Damien Wallace and Guodong Yuan

345

19 Credit risk management and practices in Islamic and conventional banks: an emerging market perspective Mahfod Aldoseri and Andrew C. Worthington

366

20 Stakes and challenges in the development of impact investing in emerging markets: the case of Asia Emmanuelle Dubocage and Evelyne Rousselet

378

21 Consumer financial spinning and market stress factors in emerging markets Olivier Mesly

394

22 Analyst coverage of emerging market IPOs and legal environment Romain Boissin

419

PART IV  INNOVATIVE MODELS IN BANKING AND FINANCE 23 The role of BigTech in emerging markets Silvio Andrae

433

24 The digital revolution in financial services: new business models and talent challenges464 Sylvie St-Onge, Michel Magnan and Catherine Vincent

NGUYEN 9781800880894 PRINT.indd 6

21/09/2022 13:14

Contents  ­vii 25 Ethics and trust on Fintech platforms from an emerging markets perspective Oliver Vasquez and Leire San-Jose 26 Challenges and opportunities for crowdfunding in emerging markets: an ethical perspective Johan Bouglet, Ghislaine Garmilis and Olivier Joffre

479

492

27 Rationalizing the Takaful organizational form with institutional theory Maryam Alhalboni, Muhammed Shahid Ebrahim and Wahyu Jatmiko

506

28 Artificial intelligence in an emerging portfolio manager: the case of Evovest Sylvie St-Onge, Catherine Vincent and Michel Magnan

523

PART V  EMERGING TRENDS 29 Emerging green finance hubs in Asia: regulatory initiatives for ESG investing and green bond development by the Four Tigers Artie W. Ng

539

30 Enhancing sustainability reporting and greening the finance system: institutionalization and practices in China’s banking sector Shidi Dong, Lei Xu and Ron McIver

553

31 Cross-border banking in EMDEs: trends, scale, and policy implications Erik Feyen, Norbert Fiess, Ata Can Bertay and Igor Zuccardi Huertas 32 The Swiss banking experience and lessons learned for emerging markets: the roles of digitalization and sustainability Carlo Raimondo and Patrick Coggi

573

597

33 Financial development of CEE markets and the evolving role of foreign-owned 619 multinational banks Victoria Geyfman 34 Key rate pass-through to deposit rates: experience from the pandemic era Henry Penikas

634

35 Inflation targeting and fiscal discipline in selected emerging economies Milojko Arsić, Zorica Mladenović and Aleksandra Nojković

651

PART VI  NEW PERSPECTIVES 36 The role of financial surveys for economic research and policy making in emerging markets676 Sofía Gallardo and Carlos Madeira 37 Emerging regions in the era of negative real interest rates: twenty years of convergence towards the US? Max Gillman

NGUYEN 9781800880894 PRINT.indd 7

687

21/09/2022 13:14

viii  Handbook of banking and finance in emerging markets 38 The Indonesian banking sector during the COVID-19 pandemic Sahminan Sahminan 39 Bank finance and alternative instruments in capital-intensive sectors: the case of global shipping Theodore Syriopoulos

710

731

40 The Russian financial sector: opportunities in an unstable environment and sanctions763 Vasily Tkachev, Vadim Grishchenko and Karl Summanen 41 Conflict, contention and cooperation in China’s new model of financial intermediation monitoring W. Travis Selmier II

793

Index

811

NGUYEN 9781800880894 PRINT.indd 8

21/09/2022 13:14

Contributors Mazin A.M. Al Janabi is a Full Research Professor of Finance and Banking and Financial Engineering, and recently retired from EGADE Business School, Tecnologico de Monterrey, Mexico. Before joining EGADE Business School, Professor Al Janabi worked for many years as a Full Research Professor of Finance and Banking and Financial Engineering at UAE University (UAEU) in Abu Dhabi, UAE and at Al Akhawayn University in Ifrane (AUI), Morocco, among other academic institutions. He holds a PhD degree in Nuclear Engineering from the University of London, UK, and has more than 30 years of real-world experience in science and technology think tanks, engineering enterprises, financial markets, and academic institutions and in many different roles. He has worked for top international financial groups (e.g., ING-Barings and BBVA), where he held several senior management positions, such as Director of Global Market Risk Management, Head of Trading Risk Management, and Head of Derivative Products. Professor Al Janabi has a strong interest in research and developments within emerging economies and has several publications in international refereed journals, books and chapters in books. Furthermore, his R&D in quantitative finance has been formally classified in the academic literature as the Al Janabi Model for Liquidity Risk Management (Liquidity Adjusted Value-at-Risk, LVaR Model). Professor Al Janabi has published in top-tiered journals such as: European Journal of Operational Research, Journal of Forecasting, International Review of Financial Analysis, Physica A: Statistical Mechanics and its Applications, European Actuarial Journal, Annals of Operations Research, Complexity, International Journal of Finance & Economics, Applied Economics, Economic Modelling, Review of Financial Economics, Journal of Asset Management, Service Industries Journal, Journal of Modelling in Management, Studies in Economics and Finance, Emerging Markets Finance and Trade, Studies in Economics and Finance, Journal of Risk Finance, Journal of Banking Regulation, Annals of Nuclear Energy. Anas Mohammad Hussein Al-Jbour is a PhD candidate in Finance at the University of South Australia (UniSA), sponsored by a competitive scholarship from his home country, Jordan, where he completed his master’s degree in Finance and Investment. Throughout his academic education, he has demonstrated commitment to ongoing research in finance and banking, particularly Islamic banking. Anas brings extensive industry experience to his research. He has worked for the Amman Stock Exchange (ASE), Jordan, prior to which he worked for two Islamic banks – Jordan Islamic Bank (JOIB) and Islamic International Arab Bank (IIAB) – gaining valuable practical experience in banking and finance that allows him to integrate his industry background with academic strength and contribute to academia. Mahfod Aldoseri completed his doctorate at Griffith University, Australia, under the supervision of Andrew C. Worthington. He holds an MPhil and a Master of Applied Finance from the University of Newcastle in Australia, and a Master of Comparative ix

NGUYEN 9781800880894 PRINT.indd 9

21/09/2022 13:14

x  Handbook of banking and finance in emerging markets Fiqh (Islamic jurisprudence) from the Higher Judicial Institute of Imam Mohammed Bin Saud Islamic University in Saudi Arabia. Maryam Alhalboni is a lecturer at York University, UK, having worked earlier at the University of Essex and the University of Durham. Maryam has a BSc in Economics, an MSc in Finance and Investment and a PhD in Finance from the University of Essex. Before that, she worked at an international bank for two years as a junior manager. Her  research interests are in the areas of Islamic finance, corporate finance, financial  markets and market microstructure. The bulk of her papers concentrate on Islamic  banking, high-frequency trading and market liquidity, but span other topics, including mergers and acquisitions and IPOs. She holds a Fellowship of the Higher Education Academy (FHEA) and UK Postgraduate Certificate in Academic Practice (PGCAP). Silvio Andrae is an economist. Most of his professional experience has been in the financial sector. Since 2000, he has worked as a senior consultant in various FinTech projects in Latin America, Africa and Asia. He is currently a senior advisor with the German Savings Banks Association, and is a specialist on various regulatory initiatives at the Basel Committee on Banking Supervision. Silvio Andrae studied and received his PhD at the Free University Berlin. Milojko Arsić is Professor of Public Finance and Fiscal Policy at the University of Belgrade Faculty of Economics, Serbia, where he completed his undergraduate and doctoral studies. His areas of research are fiscal and monetary policy, public finance and economic growth. Arsić is the co-author of a few books in the field of public finance and fiscal policy and a few papers published in international journals. Since 2000 he has participated in the creation of several stabilization and reform programmes in Serbia. Arsić was an advisor to the Serbian government for public finance reform, vice-governor and a member of the National Bank of Serbia Council. He is an active participant in academic and public debates on Serbian economic policy. Thankom Arun is a Professor of Global Development and Accountability at Essex Business School, UK, as well as Professor Extraordinaire at the Stellenbosch Business School, South Africa, and a Research Fellow at IZA, Bonn, Germany. He chairs an academic steering group on financial inclusion in the International Cooperative and Mutual Insurance Federation (ICMIF), and is a Fellow of the Royal Society of Arts, Manufactures and Commerce. Previously, he was Professor and Director of the Institute of Global Finance and Development (IGFD) at the Lancashire Business School (UCLan), Visiting Professor at the University of Rome, and has held academic positions at the Universities of Manchester and Ulster. Thankom’s research moved away from arbitrary disciplinary constraints towards an interdisciplinary learning process to understand global challenges, particularly in developing/emerging country contexts. Over the years, the research carried out aims to understand, theorize and tackle the problems created by the uneven relationships between business, society and economy in an interdisciplinary framework. Thankom’s recent research interests are Fintech, financial inclusion, climate change and sustainability.

NGUYEN 9781800880894 PRINT.indd 10

21/09/2022 13:14

Contributors  ­xi Ioanna Avgeri studied at Athens University of Economics and Business (BSc in Economics, MSc in Applied Economics and Finance, MSc in Economic Theory). She is currently a PhD student. Her field of research is finance, and especially the impact of euro-area institutional changes on bank performance. Her work has been published by the Bank of Greece (Working Paper Series), the London School of Economics (LEQS) and the Journal of International Financial Markets, Institutions and Money, and has been presented at academic conferences. Ata Can Bertay is an Assistant Professor of Finance at Sabancı University, Istanbul. Previously, he was a Research Economist at the World Bank’s Development Economics Research Group leading the Global Financial Development Report (GFDR) 2019–2020 on bank regulation and supervision and the GFDR 2017–2018: Bankers  without Borders on international banking. From 2010 to 2012, he also worked at the World Bank as a consultant in the Research Group. He received a BA in  Economics from Boğaziçi University (2008) and an MSc (2010) and PhD in Economics (2014) from Tilburg University. In 2013 he was a visiting scholar at New York University Stern School of Business. His research interests include banking, financial economics and macrofinance, and his research has appeared in publications such as Journal of Financial Stability, Journal of Banking and Finance and Journal of Financial Intermediation. Romain Boissin is an Associate Professor of Finance at the University of Montpellier, and head of the wealth management department, having previously been an Associate Professor at Paris Business School. His research focuses on IPOs, analyst coverage and securities laws in a global context, and he has presented his work internationally at seminars and conferences. Johan Bouglet is an Associate Professor at the University of Paris-Est Créteil (UPEC). He obtained his PhD in Management Science from the Paris Dauphine University after studying at the École Normale Supérieure de Cachan (ENS Cachan). A member of the Institut de Recherche en Gestion (IRG), his work focuses on strategy, governance and crowdfunding, and he has studied a variety of settings, including industrial companies, online platforms and public sector organizations. Yeguang Chi is a lecturer at the University of Auckland Business School, prior to which he worked in Shanghai Jiaotong University’s Advanced Institute of Finance. Yeguang’s research focuses on emerging markets, such as the Chinese A-share market, and his work has been awarded first prize by the Chartered Financial Analysts Society at various academic conferences. He received a BA and MS in Applied Mathematics from Harvard University and a PhD in the joint Financial Economics programme from the University of Chicago, where he was Nobel laureate Lars Hansen’s teaching assistant. Shiao-Lan Chou is a lecturer in Finance at the University of South Australia (UniSA) Business School. Shiao-Lan holds a Bachelor of Commerce and a Master of Commerce from Curtin University of Technology, having previously gained a few years’ research experience in the Department of Ophthalmology at the University of Melbourne on a

NGUYEN 9781800880894 PRINT.indd 11

21/09/2022 13:14

xii  Handbook of banking and finance in emerging markets project measuring personal costs and indirect costs associated with vision impairment. Shiao-Lan’s research and teaching interests encompass finance and health economics, including corporate and personal finance, financial markets, cost evaluation, healthcare costs and indirect and personal costs. Shiao-Lan has coordinated and taught large core business school courses for many years, and completed a term as Academic Integrity Officer at the UniSA School. Manisha Chuttoor has worked as an audit associate with KPMG Mauritius since August 2020. She has a background in financial science, and graduated with a finance major and law minor from the University of Mauritius in 2020. Her research interests include the systematic instabilities in the macro-economy and autofitted indicators that caution authorities about failings of the incumbent business-as-usual model. She has also worked as research assistant on various projects involving corporate governance, internet financial reporting and integrated reporting assurance. She aspires to continue on a similar path by contributing to under-researched areas and emerging trends in the fields of accounting and finance. Patrick Coggi is a member of the Group Executive Board at Ceresio Investors, an international banking group headquartered in Switzerland. He holds an MSc in Economics from the London School of Economics and a PhD in Finance from the University of St. Gallen. He was the CEO of a Swiss-based private bank and has experience as CIO and COO. He is also an Adjunct Professor at Franklin University Switzerland and lecturer at the Università della Svizzera italiana. Yiannis Dendramis is an Assistant Professor in the Department of Economics of Athens University of Economics and Business. Prior to this he was a Marie Skłodowska Curie European research fellow, a lecturer in the Department of Accounting and Finance of the University of Cyprus, and a Visiting Lecturer and Research Fellow at the School of Economics and Finance of Queen Mary University of London. His main research areas include a wide range of topics in econometrics and empirical finance such as the econometric modelling of large datasets, volatility modelling, econometric forecasting and applied macroeconomics. His work has been published in journals such as Econometric Theory, Journal of Economic Dynamics and Control, Journal of the Royal Statistical Society SA, International Journal of Forecasting, Journal of Banking and Finance, Journal of Empirical Finance and Journal of Forecasting and Energy Policy. Shidi Dong is a researcher in Accounting at the School of Management, Shenyang University of Technology (SUT), China, having joined SUT in 2016, prior to which she  worked for years at Zhongnan University of Economics and Law in Wuhan. Her  teaching and research expertise covers financial and management accounting and  sustainability reporting. In recent years she has published a growing number of research articles on firm CSR reporting and disclosures in reputable international journals with noticeable impacts and citations such as Sustainability Accounting Management and Policy Journal, Journal of Applied Accounting Research and Journal of Cleaner Production. Her current research focus is on green finance and banking in emerging markets.

NGUYEN 9781800880894 PRINT.indd 12

21/09/2022 13:14

Contributors  ­xiii Emmanuelle Dubocage is a full Professor of Corporate Finance at the University of Paris Est Créteil (UPEC). In 2019, she was appointed director of the Institut de Recherche en Gestion (IRG) laboratory. Her areas of focus in research are traditional venture capital, solidarity venture capital and crowdfunding, and she also publishes on the theme of impact investing. She is also head of research at UPEC’s IAE Gustave Eiffel School of Management, where she teaches corporate finance from bachelor to master level and is responsible for the finance major. An Thi Thuy Duong is a lecturer in the Faculty of Finance of the Banking University of Ho Chi Minh City (BUH), Vietnam. She holds a PhD in Economics from Utrecht University on financial integration, productivity and trade, and a master’s in public policy from the KDI School of Public Policy and Management, South Korea, following undergraduate studies in credit and finance at BUH. An teaches financial market, corporate finance, financial investment and portfolio management. In addition to her academic career, she also has experience in the financial industry as a financial analyst at Thien Viet Securities, a consultant at Vietinbank Securities and a securities broker at ACB Securities. Her research interests include capital markets, international capital flows, economic integration, international trade and total factor productivity, with an emphasis on empirical applications. Daniel Dupuis earned his PhD from the John Molson School of Business at Concordia University in Montreal, Canada. His research interests include emerging markets, cryptocurrencies, asset pricing and family firm performance. Dr Dupuis has featured on CNBC Arabia as an expert market commentator and as a keynote speaker at international conferences. Furthermore, he has extensive industry experience as a derivatives floor trader, options and futures (1994–2000), and in multiple management positions in the field of international investments. His consulting portfolio includes banks, brokerage firms and market regulating bodies in Canada. He also pioneered the implementation of academic trading floors and financial research labs in five universities. Muhammed Shahid Ebrahim is a Professor of Finance at the University of Durham, UK, where his research focuses on cultural political economy and development finance. Prior to joining Durham, he was a Professor of Islamic Banking and Finance at Bangor University Business School in Wales, and has previously held roles as a financial analyst/ planner with the United Bank of Kuwait and IDS-American Express. He was the recipient of the Harwood Memorial Real Estate Scholarship and Outstanding Faculty of Management Lecturer Award from the University of Brunei Darussalam. Shahid coauthors with renowned Nobel laureate Robert Shiller. Erik Feyen is Head of Global Macro-Financial Monitoring in the World Bank Group’s Finance, Competitiveness and Innovation Global Practice. There he leads the global monitoring and analysis of key macro-financial vulnerabilities and trends; co-leads analytical work on the future of finance; advises on Fintech-related country operations; and regularly coordinates IMF–World Bank collaborative efforts such as the Bali Fintech Agenda. He has written extensively on international macroeconomics, financial stability, financial development, Fintech and climate change, and his research has appeared in the

NGUYEN 9781800880894 PRINT.indd 13

21/09/2022 13:14

xiv  Handbook of banking and finance in emerging markets Journal of Financial Economics, Journal of Portfolio Management and World Bank Economic Review. Erik holds a PhD in Finance from the University of Amsterdam and an MS in Technology, Policy and Management/Electrical Engineering from Delft University of Technology. He also regularly lectures at Columbia University, New York. Norbert Fiess is a Lead Economist in the World Bank’s Macroeconomics, Trade and Investment Global Practice (MTI) on assignment to the French Treasury, supporting Official Development Assistance. In previous World Bank assignments, he led the MTI’s Macro-Financial Risk Initiative and headed the Credit Risk Department. Norbert holds a PhD in Economics from the University of Strathclyde, Scotland, and is an honorary research fellow of the University of Glasgow. Having also worked in academia and investment banking, he has been published in Journal of Money, Credit and Banking, the Journal of International Money and Finance and Journal of Development Economics. Sofía Gallardo is a BBA student and Teaching Assistant at Pontificia Universidad Católica de Chile. During her internship at the Central Bank of Chile, she worked with Carlos Madeira researching economic expectations and household finance. Prior to this she worked at the Hospital Militar de Santiago. Ghislaine Garmilis is an Associate Professor of Accounting and Head of the Accounting, Management Control and Finance Area at the Institut Mines-Télécom Business School (IMT-BS), France. Ghislaine received her PhD in Management Science from the Paris Dauphine University and her BBA from Siena College, New York. A member of the LITEM research laboratory, her research centres on trust and third parties, and research interests include ethics in accounting, financing for business growth and crowdfunding. Victoria Geyfman began her academic career at Bloomsburg University of Pennsylvania in 2005, where she is now a Professor of Finance. Before joining BU, she worked in the financial services research section of the Federal Reserve Bank of Philadelphia. She received her PhD from Temple University. Geyfman’s research has focused on financial services institutions, emerging markets finance and the status of women in business. She has also taught undergraduate and graduate courses in corporate finance, financial markets and institutions, and commercial bank management. She has published in the Journal of Money, Credit and Banking, Banking and Finance Review, Journal of Finance Case Research, Journal of Emerging Markets, Emerging Markets Finance & Trade, The Case Journal, Journal of Education for Business and Gender in Management. Max Gillman is F.A. Hayek Professor of Economic History at the University of Missouri-St. Louis. His research is on monetary economics, macroeconomics, public finance, asset pricing and development, and currently includes the optimal avoidance of both implicit inflation tax and explicit fiscal taxes for consideration in forming monetary and fiscal policy, with a view towards rules-based policy. This includes how competitive asset markets determine real interest rates while being regulated by the US Federal Reserve System. Gillman also serves as Associate Editor of Economic Modelling. He has previously been Professor of Economics at Cardiff Business School and at Central European University. He is scheduled to be a visiting scholar at the Bank of Finland and

NGUYEN 9781800880894 PRINT.indd 14

21/09/2022 13:14

Contributors  ­xv has been a visiting researcher/scholar/professor at New York University, the University of Chicago, Loughborough University, Monash University, the University of Melbourne and the University of New South Wales, and at the Federal Reserve Banks of St. Louis, Atlanta and Minneapolis. His research has been published in the Journal of Monetary Economics, Journal of Money, Credit and Banking, Economic Journal, Journal of International Money and Finance, Economica and Economic Inquiry; and his authored books include Inflation Theory in Economics (2009), Advanced Modern Macroeconomics: Analysis and Application (2011) and Principles of Macroeconomics: An Evolutionary Approach (2017). Claudia Girardone is Professor of Banking and Finance and Dean of Essex Business School, where she is also Director of the Essex Finance Centre (EFiC). She holds several external academic positions, including Academic Fellow at the Centre for Responsible Banking and Finance of the University of St Andrews and a Member of the Scientific Committee of the Euro-Mediterranean Network for Economic Studies. Claudia’s research areas are on banking sector financial and social performance, bank corporate governance and stability, the industrial structure of banking and access to finance. She has published over 60 articles in books and peer-reviewed international journals, including the Journal of Corporate Finance, European Journal of Operational Research, Small Business Economics and Journal of Financial Services Research. She is a co-author of the textbook Introduction to Banking (2021) and is on the editorial board of several journals, including Journal of Financial Economic Policy and European Journal of Finance. Vadim Grishchenko graduated in 2013 from the Higher School of Economics (HSE University) and Moscow Institute of International Relations (MGIMO), and is now Head of the Unit at the Research and Forecasting Department of the Central Bank of the Russian Federation and senior lecturer at the Higher School of Economics (HSE University). Since 2013 he has provided analytical support of key rate decisions and exchange rate policy, prepared policy documents, carried out research on monetary policy and macroeconomic issues, monitored monetary policy of various central banks and interacted with international organizations. Currently he is responsible for research and macroeconomic forecasting using a stock-flow consistent (SFC) approach. He is a coauthor of the ‘Digital Ruble’ Report (October 2020), and has published a number of articles in leading Russian economic journals. Andreas Gruener is Professor of Finance and Accounting at the School of Finance (SoF) and Academic Director of the Undergraduate Programme at the University of St. Gallen, Switzerland. He is also a Visiting Professor at Singapore Management University (SMU), St. Gallen Institute (SGI Singapore and Kuala Lumpur), the National University of Singapore (NUS) and the University of Liechtenstein. His research areas are corporate finance, asset management, alternative investments, private and public markets, sustainable finance in developed and emerging markets, impact investing, Fintech, machine learning and CSR disclosure and reporting. He also teaches finance and accounting topics to all levels at St. Gallen and in the executive courses of private and public organizations such as the Collège des Ingénieurs (Paris) and technical universities such as Giessen/ Friedberg, Kaiserslautern and Augsburg.

NGUYEN 9781800880894 PRINT.indd 15

21/09/2022 13:14

xvi  Handbook of banking and finance in emerging markets Francesco Guidi is a Senior Lecturer in Economics at the University of Greenwich Faculty of Business, from where he holds a Postgraduate Certificate in Higher Education (PGCertHE). He graduated from the University of Manchester (UK) with a MSc in Economics and Econometrics, and completed a MSc and a PhD in Economics at the University of Ancona (Italy). Francesco’s interests encompass applied research in stock markets, portfolio diversification strategies and banking, and it has been published in academic journals such as Applied Financial Economics, Journal of Emerging Market and Finance, Journal of International Financial Markets, Institutions & Money, Journal of Multinational Financial Management, International Review of Financial Analysis and International Review of Economics and Finance. Constantin Gurdgiev is an Associate Professor of Finance at Monfort College of Business, University of Northern Colorado (USA), and teaches Finance and Economics at Trinity Business School, Trinity College Dublin (Ireland). He was previously an Associate Professor at the Middlebury Institute of International Studies at Monterey, California, and has a long career in financial markets, including co-founding and running a start-up asset management company in Europe. He is also co-founder of two successful financial services non-profit organizations, and chairs the board of directors in one. His research, focusing primarily on macroeconomic and systemic risks and financial markets, has appeared in a number of academic journals, such as the Quarterly Review of Economics and Finance and the International Review of Financial Analysis. Wahyu Jatmiko is a lecturer in the Faculty of Economics and Business, University of Indonesia, and a senior researcher at the university’s Centre for Islamic Economics and Business. Before joining the university (from where he has a BEc), Wahyu was a senior researcher at Indonesia’ National Alms Agency (BAZNAS). He earned his MSc and PhD in Islamic finance from Durham University Business School, UK, where, during his doctorate, he also served as a postgrad lecturer in economics and finance. He is an Associate Fellow of the UK’s Higher Education Academy (AFHEA). His research interests include ethical finance, Islamic finance, development finance and development economics. Li Jiaqi works in the Global Finance division of the National Bank of Canada, having completed her studies in Finance at Trinity College, Dublin. Her work focuses on governance risks and structural reforms in global financial markets. Olivier Joffre is an Associate Professor at the University of Paris-Est Créteil (UPEC), having obtained a PhD in Management Science and accreditation to supervise research from Paris Dauphine University. He was the recipient of the FNEGE-AIMS 2006 award for the best PhD dissertation on Strategy. He is co-founder of the academic journal Recherche et Cas en Sciences de Gestion. A member of the IRG laboratory, his research addresses issues of post-integration processes in domestic and international M&As, sustainable development practices and crowdfunding. Md. Nurul Kabir completed his doctorate, ‘Credit Risk in Islamic and Conventional Banks’, at Griffith University (Australia) under the supervision of Andrew C. Worthington.

NGUYEN 9781800880894 PRINT.indd 16

21/09/2022 13:14

Contributors  ­xvii He holds an MBA from Ritsumeikan Asia Pacific University (Japan), and his published research on credit risk in Islamic banking includes a chapter in Contemporary Issues in Islamic Finance: Principles, Progress, and Prospects, edited by A.C. Worthington (2014), and articles in the Pacific-Basin Finance Journal and the International Review of Financial Analysis. His academic appointments include Charles Darwin University in Australia and North South University in Bangladesh. Clemens Kool studied Econometrics and Operations Research at Erasmus University, Rotterdam, where he obtained his PhD in 1989 on Bayesian forecasting, learning and monetary policy. Since 2017, he has held the full-time chair of Macroeconomics and International Monetary Economics at Maastricht University School of Business and  Economics. Previously, he also held chairs at Indiana University and Utrecht University, and has worked at national and international policy institutions such as De Nederlandsche Bank, the Netherlands Bureau for Economic Policy Analysis and the Federal Reserve Bank of St. Louis. He has published widely in national and international academic journals, has been an advisor to national and regional Dutch government authorities as well as to various financial institutions, and actively participates in societal (economic) debates. His research and teaching interests focus on monetary theory and policy, European integration, banking, the global financial infrastructure, and macroeconomic imbalances and financial crises. His research is typically empirical and policy oriented, and recognizes the need for a multidisciplinary approach. Tu Le is a researcher at the Institute for Development and Research in Banking Technology, University of Economics and Law, Vietnam. His works focuses on efficiency and productivity measurement in banking and finance and the industrial sector, the impact of e-commerce on economic growth and Fintech. His papers have been published in the International Journal of Managerial Finance, Managerial Finance, Pacific Accounting Review, Post-Communist Economies, Central Bank Review, Cogent Economics & Finance and Review of Economic Analysis. He is also a member of the review board of the Journal of Risk and Financial Management, Journal of Economics, Finance and Administrative Science and Journal of Asian Business and Economic Studies, among others. Lidia Loban is a PhD student at the University of Zaragoza (UZ), Spain. Her doctoral dissertation – ‘Active management in mutual funds with concentrated benchmarks: A major dilemma to fulfill the EU directive on portfolio concentration limits – was to be presented at the end of 2021 with an International Mention. Her main research field embraces mutual funds, specifically portfolio compositions, active management and risk determinants. She has published in journals such as International Review of Financial Analysis, Journal of Multinational Financial Management and Finance Research Letters, and has presented her work at international conferences, including the 28th Annual Global Finance Conference (Top Paper Award), the 30th Annual Meeting of the European Financial Management Association and the 11th Portuguese Finance Network Conference. In 2019 she was a Visiting Researcher at the Nova School of Business and Economics (Portugal), supervised by Professor Miguel Ferreira. She has chaired the organizing committee of the UZ Faculty of Economics and Business Brown Bag Seminar (BBS) research series.

NGUYEN 9781800880894 PRINT.indd 17

21/09/2022 13:14

xviii  Handbook of banking and finance in emerging markets Helen Louri is a Professor in the Department of Economics of Athens University of Economics and Business, where she gained a BSc in Economics, and was department head in 2015–2020. She also has degrees in Economics from the London School of Economics (MSc) and the University of Oxford (DPhil). She has been a Visiting Fellow at Oxford and Visiting Professor at LSE, where she remains a Research Associate. She has published extensively in academic journals. She was deputy governor of the Bank of Greece (2008–2014) responsible for monetary policy, bank resolution, the national mint and cash management; president of the Hellenic Deposit Guarantee Fund; and served as member of the International Relations Committee of the European Central Bank. She is also a member of the Appeal Panel of the Single Resolution Board of European banks in Brussels. Carlos Madeira is a Senior Economist at the Central Bank of Chile, having led its Household Finance team and coordinated several workshops for it and the InterAmerican Development Bank. Carlos has published over 20 articles on Fintech, credit regulation, portfolio risk, household finance, economic expectations, monetary policy disagreement and pension policy in top journals such as the Review of Economics and Statistics, Journal of Money Credit and Banking, Journal of Banking and Finance, Journal of Applied Econometrics and Journal of Economic Dynamics and Control. He is co-editor of the open access journal Economics and a Scientific Committee Member for conferences such as IFABS, LACEA and IFC. He holds a PhD in Economics from Northwestern University (USA) and a BA from the Nova School of Business and Economics (Portugal). Michel Magnan is Distinguished University Research Professor and S.A. Jarislowsky Chair in Corporate Governance at the John Molson School of Business (Concordia University, Montréal, Canada). He is also a Researcher and Fellow at CIRANO. He holds a PhD from the University of Washington (Seattle), and is a Fellow Chartered Professional Accountant (FCPA) and a Fellow of the Royal Society of Canada. He is also a Chartered Director and member of several boards of directors. His research and professional interests encompass corporate governance, financial statement analysis, corporate reporting and disclosure, corporate social responsibility and valuation. Isidora Manjante is a graduate of the Faculty of Economics at the Universidade Eduardo Mondale (Maputo, Mozambique), and holds an MSc in Accounting, Corporate Finance and Taxation from Maputo’s Universidae Politécnica. She works as Financial Manager of the Observatório do Meio Rural, an independent research organization focused primarily on agrarian and rural development in Mozambique from an integrated and interdisciplinary perspective through scientific research, studies and debates on policies and other relevant issues. Pedro Verga Matos is an Associate Professor at the Lisbon School of Economics and Management (ISEG) of the Universidade de Lisboa (Portugal), where he teaches investment appraisal and corporate governance courses. He is a graduate of the Faculty of Economics and Business Administration of the Portuguese Catholic University, and holds an MBA and an MSc in Business from the New University of Lisbon. He received his PhD in Sciences Management from the University of Porto. His research interests

NGUYEN 9781800880894 PRINT.indd 18

21/09/2022 13:14

Contributors  ­xix include corporate governance, retailing, multicriteria analysis, social innovation and microfinance, and he has been published in academic journals such as Economic Modelling, Applied Economics, Journal of Business Research, Journal of International Financial Markets, Institutions & Money and Group Decision and Negotiation. Ron McIver is a Lecturer in Financial Planning and Finance, and a member of the Markets, Values and Inclusion (MVI) research concentration at the University of South Australia (UniSA) Business School. Ron’s research and teaching interests encompass the broader areas of banking, finance and financial planning, including investment management, insurance, performance evaluation, valuation and corporate and personal risk management. His current foci are financial system regulation and reform; the roles of institutional and governance structures in delivering efficient outcomes in transition and market-oriented economies; the development and efficiency of emerging market economy financial systems and corporations; and identification and analysis of financial spillovers and spillover networks. Ron has produced a wide variety of scholarly works, including papers in Accounting and Finance, Applied Economics, Business and Society, Energy Economics, Journal of Cleaner Production, and Sustainability Accounting Management and Policy Journal. Olivier Mesly is an Associate Professor at the ICN School of Business in Nancy, and a Guest Professor at the University of Lorraine, both in France. He is a member of the European Research Centre for Financial Economics and Business Management (CEREFIGE), and holds the following degrees from Canadian universities: Certificate in Voice (University of Sherbrooke, 2017); Postdoctoral Fellowship in Arts Management (HEC Montréal, 2011; DBA in Marketing (Sherbrooke, 2010); MBA (Guelph University, 1999); Diploma in Public Relations (McGill University, 1990); and Bachelor with Distinction in East Asian Studies (also McGill, 1989). His research interests include behavioural finance, marketing, project feasibility and psychology. He has several articles in English, French and Spanish in peer-reviewed journals such as Economic Modelling, Journal of Macromarketing and International Journal of Project Management. He has also written ten books – including two on marketing, three on project feasibility and one on psychological research – as well as over 40 case studies. Professor Mesly pioneered the concept of financial spinning, whereby market agents disconnect from their financial needs, goals and preferences while engaging in a (potentially unsustainable) debt trap. This often has repercussions for financial markets, influenced by and then aggravating dysfunctionality through friction factors as market agent aggregates herd, swarm and stampede over each other. Ultimately, extreme spinning leads to financial/banking crashes. Paweł Miłobędzki holds an MSc in Econometrics and a PhD and PhD Hab in Economics from the University of Gdańsk (Poland), as well as a Professorship in Social Sciences from the President of Poland. He is a member of the Committee on Statistics and Econometrics of the Polish Academy of Sciences and Head of the Econometrics Department at the University of Gdańsk. Previously he served as Professor of Economics and Head of Management at Gdynia Maritime University, and as an expert for Poland’s National Science Centre. He is an Editor-in-Chief of Statistical Review, a quarterly journal of Statistics Poland. His research areas concern asset pricing, term structure of interest rates

NGUYEN 9781800880894 PRINT.indd 19

21/09/2022 13:14

xx  Handbook of banking and finance in emerging markets and market microstructure. He authored papers published in Economics in Transition, European Journal of Finance, Argumenta Oeconomica, Statistical Review, and has contributed to conference monographs in finance and financial economics published by Springer. Zorica Mladenović is Professor of Econometrics and Time Series Analysis at the University of Belgrade Faculty of Economics (Serbia), from where she received her PhD, and has over 30 years’ teaching experience. Her main areas of research include time series modelling of macroeconomic and financial data, with particularly focus recently on econometric issues of modelling emerging economies. She publishes in and referees for international journals, and is author and co-author (in Serbian) of several books. She was a member of the National Bank of Serbia Council and the Scientific Council of Science Fund of Serbia. Artie W. Ng is currently Dean of International Business University, Toronto, Canada. He was previously a Senior Research Fellow with the University of Waterloo Institute for Sustainable Energy (WISE), Canada, conducting interdisciplinary research in the areas of sustainability performance, green finance and policy for renewable and sustainable energy. He has been affiliated with Hong Kong Polytechnic University and has held Visiting Professorships at the Universities of Bergamo, Jinan (Institute of Resource, Environment and Sustainable Development) and National Cheng Kung. Dr. Ng is an International Associate of the Centre for Social and Environmental Accounting Research (CSEAR) at Scotland’s University of St Andrews – a global community of scholars who engage with students, practitioners, policy makers and other interested groups in order to generate and disseminate knowledge on social and environmental accounting and accountability. He has published in influential international journals on sustainability such as Energy Policy, Journal of Cleaner Production, Journal of Sustainable Finance and Investment, Renewable & Sustainable Energy Reviews, Sustainability Accounting and Management & Policy Journal. He is also an editorial board member of numerous journals, including Journal of Financial Regulation and Compliance, International Journal of Climate Change Strategies and Management, International Journal of Sustainability in Higher Education and Social and Environmental Accountability Journal. Thanh Ngo received his PhD in Economics from Massey University (New Zealand) in 2015 and joined its School of Aviation (SoA) in January 2017, where he is now a Senior Lecturer. His work involves efficiency and productivity analysis in the banking and finance, sustainability, agriculture and manufacturing sectors, and aviation and transportation economics. Dr Ngo’s research papers have been published in Transportation Research Part A, Annals of Operations Research, Transport Policy and International Journal of Managerial Finance, among others. He is a member of the Editorial Board of International Journal of Financial Studies; a Series Editor of Atlantis Highlights in Economics, Business and Management and Advances in Economics, Business and Management Research; and a reviewer for many journals, including Annals of Operations Research, Applied Economics, Pacific Accounting Review and Journal of Air Transport Management. Aleksandra Nojković is an econometrician and empirical economist with broad disciplinary interests that span themes for emerging markets, fiscal and monetary policy and

NGUYEN 9781800880894 PRINT.indd 20

21/09/2022 13:14

Contributors  ­xxi economic growth, as well as the labour market and econometric methods for programme and policy evaluation. She received her BA and completed her PhD at the University of Belgrade Faculty of Economics, where she was promoted to full professor in 2018. Aleksandra Nojković has over 20 years’ teaching and research experience in applied econometrics, and her primary research interests are time series analysis and econometric methods. She is the co-author of papers published in international journals, books published in Serbian and serves on a few editorial boards. Sabina Nowak holds a PhD in Economics from the University of Gdańsk (Poland), where she is an Assistant Professor in the Econometrics Department. Her research interests are asset pricing and market microstructure, as well as selected topics in corporate finance, including dividend policy and dividend smoothing. She serves as a Statistical Editor for Equilibrium: Quarterly Journal of Economics and Economic Policy, and has authored papers published in Emerging Markets Finance and Trade, Contemporary Economics and Journal of International Studies, and chapters in conference monographs on finance and financial economics published by Springer. Cristina Ortiz is Associate Professor of Finance in the Faculty of Economics and Business Administration at the University of Zaragoza (UZ), Spain, where she has taught undergraduate and postgraduate courses in Financial Economics since 2006. Her PhD dissertation, ‘Behavioural Finance: Application to Investors and Managers of Spanish Mutual Funds’ (2007), received European mention and was published by an international office. Her main research interests embrace portfolio management, social  responsible investment and the study of financial behaviour from the perspective  of individual and professional investors. She has published in international journals  such as Journal of Banking and Finance, Journal of Financial Services Research, Journal of Economic Behavior and Organization and Journal of Behavioral Finance, among others. She presents regularly at international conferences, winning the Best Paper Award at the 10th PFN Conference sponsored by Eurosistema-Banco de Portugal (Lisbon, 2018). She has been a Visiting Researcher and Erasmus lecturer in several European universities, including Caledonian Business School (Glasgow, UK), the Universities of Porto and Minho (Portugal) and the University of Groningen (Netherlands). Henry Penikas is a project manager in the Bank of Russia’s Research and Forecasting Department, having joined the bank after ten years working for Russia’s largest private and state banks (Alfa-Bank and Sberbank). He received his master’s degree from the Paris-I University, Paris School of Economics and Moscow’s Higher School of Economics, from where he obtained his PhD on quantitative methods in economics. He was initially responsible for redesigning the Internal Ratings-Based validation methodology from the regulator’s perspective, and is now engaged in central bank policy-related research, with particular focus on the banking sector. Henry was a visiting lecturer at the University of Pavia (Italy), where he delivered courses on the application of quantitative methods in banking. He also develops agent-based models in banking and economics as a senior research fellow at the complex systems’ mathematical modelling laboratory of the P.N. Lebedev Physics Institute of the Russian Academy of Sciences.

NGUYEN 9781800880894 PRINT.indd 21

21/09/2022 13:14

xxii  Handbook of banking and finance in emerging markets Stefano Piserà is a visiting researcher at Essex Business School of the University of Essex and researcher at the University of Genova (Italy). His research areas are empirical corporate finance and banking, corporate social responsibility, financial stability and asset pricing. He has published articles in peer-reviewed international journals including the European Journal of Finance and Research in International Business and Finance. Xiao Qiao is an assistant professor in the School of Data Science at the City University of Hong Kong (CityU) and a member of the Hong Kong Institute for Data Science. He is on the editorial board of the Journal of Portfolio Management, and his research has been featured by Forbes and Institutional Investor. Prior to CityU, Xiao worked in the investment management industry, where he participated in the creation of two awardwinning exchange-traded funds. He was a co-founder of Paraconic Technologies, an associate director at SummerHaven Investment Management and a quantitative researcher at Hull Investments. Xiao received a BS in Economics from the Wharton School of the University of Pennsylvania (USA) and a BS in Engineering from the university’s School of Engineering and Applied Sciences. His PhD in Finance was obtained from the University of Chicago, where he was Nobel laureate Eugene Fama’s teaching assistant. Carlo Raimondo is a Senior Analyst at Ceresio Investors, an international banking group based in Switzerland, and a Research Fellow and Lecturer in Finance and Communication at IALS-Università della Svizzera italiana in Lugano (Switzerland). He obtained a PhD in Finance and an MSc in Banking and Finance from the University of Bologna (Italy). Carlo’s main research interests focus on the role of information on the financial markets, IPOs and investor relations, and sustainability in banking and finance. Dinesh Ramdhony is a senior lecturer at the University of Mauritius. He holds a PhD from the University of Southern Queensland (Australia), and has a keen interest in researching topics related to corporate disclosure, corporate governance, banking and technology management. Evelyne Rousselet is a Professor of Strategy at the Université Gustave Eiffel (UGE), France. There she teaches strategy and management as well as banking and financial economics at master’s level, and is responsible for the Finance major. Her main research focus is banks and their relationships with their various stakeholders, with particular interest in the strategies employed by retail banks to deal with low-income customers. She has also worked on microfinance and on the democratization of finance. Sahminan Sahminan is a Director in the Department of Macroprudential Policy of Bank Indonesia (the country’s central bank). His career at the bank (stretching over 20 years) has involved various research projects and strategic initiatives, including assessing and recommending policy on macroeconomic and financial stability issues. From 2012 to 2014, he worked as an economist at the ASEAN+3 Macroeconomic Research Office (AMRO) in Singapore, and in 2008 was a Visiting Fellow at the Bank for International Settlement (BIS) in Basel, Switzerland. His articles have appeared in domestic and international publications such as the Journal of Financial Stability, Pacific Economic Review and Studies in Economics and Finance. He received his PhD in Economics from the

NGUYEN 9781800880894 PRINT.indd 22

21/09/2022 13:14

Contributors  ­xxiii University of North Carolina at Chapel Hill, an MA in Applied Economics from the University of Michigan at Ann Arbor (both USA) and a BS in Statistics from IPB University, Bogor, Indonesia. Antonio Samagaio is an assistant professor at ISEG–Lisbon School of Economics and Management – a faculty of the Universidade Técnica de Lisboa (Portugal) – from where he graduated and received a PhD, MBA and MSc in Management. He teaches financial auditing and management control undergraduate and graduate courses. His research interests are mainly in auditing, management control, taxation and sport management, published in academic journals such as the Australian Accounting Review, Journal of Business Research, Review of Business Management, Journal of Air Transportation Management and Journal of Small Business Strategy. Leire San-Jose is a Professor at the University of the Basque Country (UPV/EHU) in Bilbao (Spain), and a Visiting Research Fellow at Huddersfield Business School (UK), where she works with Professor Chris Cowton. She also leads the Ethics in Finance & Social Value Research Group (ECRI). She was previously a Visiting Scholar at Loyola University Chicago, University of Virginia Darden Business School and Fordham University, New York (all USA), Saïd Business School at the University of Oxford, Heriot-Watt University (both UK) and Bergamo in Italy. She organized the 2022 International Society of Business, Economics and Ethics World Congress in Bilbao, having already organized the European Business Ethics Network in 2010. Her most important publications are on ethics in finance, social value and stakeholder theory, but she has also published papers on cooperative banks, zombie companies and cash ­management. Boopen Seetanah is an Associate Professor at the University of Mauritius (UoM) with research interests in tourism and transport, international trade and finance, and development economics. At UoM he is currently Co-Chair of the World Trade Organization (WTO) Chairs Programme and the Director of Research at the International Centre for Sustainable Tourism and Hospitality (ICSTH). Boopen is an editorial board member and reviewer for numerous high-rated journals, and has consulted with the Mauritian government and numerous international organizations, including the UNEP, UNDP, UNCTAD, World Bank, ADB, ILO and the RMCE. W. Travis Selmier II is a Visiting Scholar in Political Science at Indiana University (IU), USA. Prior to academia he worked in international investments (listed in Barron’s Top 100 Portfolio Managers 1998), and field-researched investments and economics in more than 50 countries. He was Co-Director of the Investment Management Academy and a Visiting Clinical Professor in Finance at IU’s Kelley School of Business from 2008 to 2011. His research interests include CSR in extractive industries, political economy of the Belt & Road Initiative, East Asian banking, institutional dynamics of international financial markets and language economics. His publications appear in Business & Politics, Business Horizons, International Review of Economics & Finance, Journal of International Business Studies, Review of International Political Economy, Transnational Corporations, World Economy, among others, and in a variety of books.

NGUYEN 9781800880894 PRINT.indd 23

21/09/2022 13:14

xxiv  Handbook of banking and finance in emerging markets Sylvie St-Onge is a Professor of Management at HEC-Montréal (Canada), and a Researcher and Fellow at the CIRANO research centre in Canada. She holds a PhD from the Schulich School of Business (York University, Toronto), and is Distinction Fellow CRHA (Human Resources Management Professional). Her research and professional interests encompass corporate governance, human resources challenges within the financial services industry HR management. She has written or co-written several textbooks on performance management, HR management and compensation management. She holds a board governance designation (ASC). Karl Summanen graduated in 1982 from the Moscow Institute of Physics and Technology, Russia. He has PhD in Physics and Mathematics, and a Diploma in Finance and Banking from the government’s Financial Academy. Until 1993 he worked at the Russian Academy of Sciences and a number of overseas universities as a researcher in theoretical physics, specializing in the theory of multipulse nuclear magnetic resonance (NMR) in solids. Since 1994 he has worked in the banking industry, specializing in banking technologies and IT, and has implemented numerous projects in various fields, including omnichannel remote banking, digital archives, optimization, digitalization and robotization of banking processes. He is also an independent researcher in the field of theory of monetary, banking and payment systems, and has offered a new view on the notion of money as an ideal substance. Theodore Syriopoulos is Professor of Finance in the Department of Port Management and Shipping, School of Economics and Political Sciences, at the National and Kapodistrian University of Athens, Greece. He also holds adjunct visiting posts in international and domestic schools, including Audencia Business School Nantes (France) Newcastle University (UK) and Shanghai Maritime University (China). For the past 20 years, he has served in the Shipping, Trade and Transport Dept, University of the Aegean. Prior to academia, he spent nearly two decades in executive management posts in banking, investment, asset management, corporate finance and consulting. Prof. Syriopoulos publishes regularly in accredited international journals and books on applied finance topics such as global capital markets, portfolio management, mergers and acquisitions, financial derivatives and risk management, shipping finance and banking, and corporate governance and social responsibility. He is a member of international academic and professional bodies and associations including the European Finance Association, American Finance Association, European Financial Management Association, International Association of Maritime Economists, inter alia. Prof. Syriopoulos holds a PhD in Applied Economics (1990), an MA in Development Economics (1987), both from the University of Kent at Canterbury, UK, and a BA in Economics (1985) from the University of Piraeus, Greece. Vasily Tkachev is Associate Professor of Finance at the Moscow State Institute of International Relations (University) of the Russian Ministry of Foreign Relations (MGIMO-University), where he is also a scientific advisor on the Finance MBA programme. He holds a PhD in Economics. He serves as the Central Bank of Russia’s external expert at the Council of Eurasian Economic Union Financial Integration, and in 2017–2020 was Senior Visiting Fellow at Henley Business School (UK). He has over 20

NGUYEN 9781800880894 PRINT.indd 24

21/09/2022 13:14

Contributors  ­xxv years’ teaching experience in the field of finance; is the author or co-author of over 30 articles and books on corporate and international finance; and has attended and spoken at national and international conferences. His main areas of research include money and credit, international finance, financial institutions and markets. Oliver Vasquez is a proactive professional with knowledge and vision of the different departments making up and driving a business, from finance to business strategy; not forgetting his international business development duties in private and public organizations. In 2020, he completed a master’s in Business Management from an Innovation and Internationalization Perspective at the University of the Basque Country, Spain. After a year working in a constantly challenging atmosphere for one of the fastest growing e-commerce businesses in Hong Kong, Oliver moved to the European Parliament in Brussels, where for five months he was involved in several financial projects for the Directorate-General for Finance and collaborated with other organizations involved in HR and professional training. Luis Vicente is Associate Professor of Finance at the University of Zaragoza (UZ), Spain and Director of its PhD programme in Accounting and Finance. He has taught courses in financial economics at the Faculty of Economics and Business Administration there since 2001. His dissertation on Spanish mutual funds received the UZ’s Social Sciences PhD Award in 2004. He also won the 2015 Financial Research Award (Accésit) ‘Antonio Dionis Soler’ and the Best Paper Award at the 10th PFN Conference sponsored by the Eurosistema-Banco de Portugal (Lisbon, 2018). In 2020, Professor Vicente was nominated for the ‘Best Teacher in Spain’ award sponsored by Educa Abanca. His main research interests include portfolio management and behavioural finance, with several articles published in journals such as Journal of Banking and Finance, Omega: International Journal of Management Science, European Journal of Operational Research, International Review of Financial Analysis and Finance Research Letters. He has served as Visiting Researcher and Erasmus lecturer at various European universities: University Carlos III (Madrid, Spain), University of Maastricht (Netherlands), ESSCA (Paris, France), University of Minho (Braga, Portugal), University of Rostock (Germany) and University of Warsaw (Poland). Catherine Vincent works at Sobeys Inc., one of Canada’s largest food retailers. She holds an MSc in management from HEC Montréal, and has worked in management and human resources. Damien Wallace holds a PhD from the University of South Australia, and has research interests in the functioning of financial markets, FinTech, financial literacy and inclusion, and sustainable finance. Within these areas his interests include innovative securities, price discovery, market efficiency, US market impact on foreign markets and environmental finance. Andrew C. Worthington is Professor of Finance in the Department of Accounting, Finance and Economics at Griffith University, Australia. His published research in Islamic finance includes the edited volume Contemporary Issues in Islamic Finance:

NGUYEN 9781800880894 PRINT.indd 25

21/09/2022 13:14

xxvi  Handbook of banking and finance in emerging markets Principles, Progress and Prospects, and chapters in State of Islamic Finance: A Retrospective Assessment and Looking Forward; Routledge Handbook of Social and Sustainable Finance; Risk Management in Emerging Markets: Issues, Framework and Modelling; Banking: Services, Opportunities and Risks; Growth and Emerging Prospects of International Islamic Banking; Handbook of Research on Theory and Practice of Global Islamic Finance; and The Routledge Handbook of Financial Literacy. His articles have appeared in the International Journal of Social Economics, International Journal of Islamic and Middle Eastern Finance and Management, Pacific-Basin Finance Journal, International Review of Financial Analysis, Journal of the Asia Pacific Economy, Jurnal Ekonomi Malaysia, Journal of Islamic Monetary Economics and Finance, Society and Business Review, International Journal of Bank Marketing and Journal of Islamic Marketing. Lei Xu is a Senior Lecturer in Finance at the University of South Australia (UniSA) Business School. In addition to his academic roles, Lei is active on boards and committees across the university, including the Research Integrity Committee, the Business Academic Unit Board and Business Safety and Wellness. His broader community contributions include executive of the Association of Overseas Chinese Professionals in South Australia and Justice of the Peace for South Australia. His research interests cover banking, financial markets, financial systems and FinTech, with a recent focus on theoretical and empirical examination of emerging financial issues in China. His ongoing research has led to impactful publications in prestigious international and national journals and conference proceedings. He has taught various courses in finance, including global banking and financial management, financial risk analysis, Chinese banking and wealth management, risk management in financial institutions, principles of finance, international finance, managerial finance, business finance, international currency and banking markets, and financial institution management. His expertise in designing new courses has also contributed to the discipline. Camilla Yanushevsky received her education at the University of Maryland, College Park – Robert H. Smith School of Business (USA) and the Bocconi School of Economics and  Management, Milan (Italy). She currently works as a senior analyst at CFRA Research (USA), regularly publishing quarterly thematic research reports and delivering print and broadcast media interviews (New York Times, NBC News, CNBC, Reuters, Fox Business, Bloomberg, etc.). She has published several academic papers in Metroeconomica and International Journal of Economic Sciences and Applied Research, among others, and is co-author of the book Applied Macroeconomics for Public Policy (2018). Daniel Yanushevsky graduated from the Robert H. Smith School of Business at the University of Maryland, USA. He currently works as an associate for JP Morgan Bank and conducts economic research for Research & Technology Consulting. He has published research papers in Journal of Asset Management, Journal of Empirical Finance, Journal of Investing and International Journal of Multicriteria Decision Making. Rafael Yanushevsky was born in Kiev, Ukraine. He received an MS in Mathematics and in Electromechanical Engineering (with honours) from Kiev University and Kiev

NGUYEN 9781800880894 PRINT.indd 26

21/09/2022 13:14

Contributors  ­xxvii Polytechnic Institute, respectively, and a PhD in optimization of multivariable systems from the Institute of Control Sciences of the USSR Academy of Sciences, Moscow, Russia. In 1987 he started teaching at the University of Maryland (USA), first in the Department of Electrical Engineering then in Mechanical Engineering, and in the University of the District of Columbia’s Department of Mathematics. Since 1999 he has been involved in projects related to the aerospace industry, receiving a Letter of Appreciation from the Department of the Navy, the Navy Area Theatre Ballistic Missile Programme in 2002. His company, Research & Technology Consulting, focuses on economic problems related to the 2008 economic crisis, and prepared material for the US Congressional Budget Office on how to check the efficiency of proposals offered by a group of congressmen. He has co-authored published papers concerning government fiscal policy in periods of high unemployment and debt, as well as effective decision making on the stock market. Guodong Yuan is a lecturer in Finance within UniSA Business at the University of South Australia. His current research interests encompass the broader areas of finance, accounting and taxation, including the market valuation effects of financial behaviour and events, corporate governance and online education. Igor Zuccardi Huertas is an Economist at the Strategy, Policy, and Review Department of the International Monetary Fund (IMF), conducting analysis of public sector debt and macro-financial stability. He has also worked in international institutions such as the World Bank, the European Central Bank (ECB), the Inter-American Development Bank (IDB) and the Central Bank of Colombia. He has published in the Journal of Money, Credit and Banking. Igor holds a PhD in Economics from the University of Maryland, College Park, USA.

NGUYEN 9781800880894 PRINT.indd 27

21/09/2022 13:14

Preface The monthly bulletin of the European Central Bank in October 2010 noted the increasingly important role of emerging economies as a critical driver of global growth. Over the period 2004–2009, they already accounted for 63% of the worldwide output increases. By 2019, this rate rose to 76%, according to IMF statistics. Nevertheless, these economies’ resilience to external shocks at national and international levels is under scrutiny despite their high growth potential. Significant challenges include, among others, slowdown in productivity, rising debt and debt burdens, growth contraction due to trade tensions and wars, and adverse effects of the increased global uncertainty on the domestic financial system. Industrial revolution 4.0, with the rising trends of automation and data exchange in manufacturing and financial technologies, has been equally calling for enabling ­technology-driven innovation to spur productivity growth. Unexpectedly, the impact of digital technologies on firm productivity level is rather weak. The outbreak of COVID-19 in early 2020 and subsequent pandemic was another shock that broke the growth dynamics and inflicted severe health problems, human and economic costs for emerging economies. According to the World Economic Outlook report in January 2021, the growth of emerging markets and developing economies under lockdown and protection measures contracted sharply by −2.4%, compared to a growth rate of 3.7% in 2019. The OECD sovereign borrowing outlook in 2021 showed that the COVID-19 shock also triggered financial instability and vulnerability, and caused significant fluctuations in capital flows to emerging market economies, leading to higher external borrowing costs. Over two years after the start of the pandemic, the path to economic recovery remains uncertain because of vaccination inequality, renewed wave threats, and risks associated with new virus variants. Along with other global driving factors (i.e., geopolitical shifts, digitalization, production relocation, supply chain disruption, and climate emergency), the challenges as mentioned earlier pushed policymakers in emerging economies to rethink their growth models, to make the latter more resilient, and ultimately to find efficient financing  sources  necessary to spur sustainable and inclusive growth. It is worth noting that the one-size-fits-all solutions are far from possible as emerging markets are still a heterogeneous group in terms of economic structure, growth drivers, and development level. The Handbook of Banking and Finance in Emerging Markets, which you hold in your hand, examines a range of exciting topics about recent developments, current trends, and new perspectives in emerging markets in the banking and finance sectors. The Handbook is organized into six parts to provide scholars, investors, regulators, and policymakers with comprehensive insights on emerging markets today and allow them to identify critical issues for future investment, regulation, and policymaking strategies. Part I – “Financial Markets, Institutions and Money” – includes eight chapters that adopt a variety of econometric approaches to investigate the dynamics of financial markets and institutions for major emerging markets. They examine important and xxviii

NGUYEN 9781800880894 PRINT.indd 28

21/09/2022 13:14

Preface  ­xxix topical issues such as credit booms and bursts, liquidity behavior, asset valuation, mutual fund investment, and financial integration. Chapter 1 by Ron McIver, Lei Xu, and Shiao-Lan Chou focuses on economic expansion, financial deepening, and equilibrium levels of bank credit in China. Their specific objective is to evaluate China’s economic fundamentals and credit expansion univariately and to compare the Chinese context to the historical expansions of 56 developed, emerging, and transition economies. The obtained results indicate a lack of a systematic consistency between credit growth and economic growth, and thus call for particular attention to lending patterns and their associated impacts on economy-wide efficiency rather than just the level of credit growth. Chapter 2 by Yeguang Chi and Xiao Qiao examines the economic value that stock mutual funds in the Chinese A-share market can provide for investors. They find evidence of performance persistence. Portfolios consisting of the top-performing funds are found to outperform the aggregate market, without being exposed to increased volatility or tail risk. Chapter 3 by Daniel Dupuis investigates the explanatory effect of a new liquidity measure (the free-float illiquidity) on the ex-dividend day price premium in a tax-free emerging market (the United Arab Emirates). The author provides evidence that this new liquidity measure drives the ex-dividend day price anomaly and accounts for the reduction in the tradable stock associated with many frontier markets where influential block holders restrict tradable shares. Also, abnormal returns persist, even in the absence of the usual microstructure impediments. The findings clearly imply that trading restrictions can partially explain the ex-dividend return puzzle. Chapter 4 by Rafael Yanushevsky, Daniel Yanushevsky, and Camilla Yanushevsky argues that the standard discounted cash flow ignores the market value of a business, and propose the modified intrinsic value model to obtain some cost estimates that would create a certain negotiation set for both buyers and sellers. The proposed model is also advantageous in that its valuation procedure allows one to consider possible growth scenarios for a firm. Chapter 5 by An Thi Thuy Duong and Clemens Kool devotes attention to the degree and evolution of financial integration for 11 Asian equity markets, while considering three important sub-periods and three distinct groups of countries with respect to their level of economic development. Using various measures of financial integration, they find that the level and speed of financial integration evolves with the level of economic development over the recent period from 2002 to 2018. However, the period associated with the global financial crisis of 2008–2009 yields mixed results in terms of financial integration, which may reflect cross-border capital controls to reduce harmful effects from contagion risk. Chapter 6 by Lidia Loban, Cristina Ortiz, and Luis Vicente uses a network data envelopment analysis (DEA) model to investigate the value investing F-Score for a sample of the largest listed firms in Latin American stock markets. This proposed approach overcomes the binary valuation of listed companies with good/bad fundamentals as well as assessment of the interactions between the main financial areas in a listed company. The results indicate that the neural F-Score significantly improves the short-term returns over the well-known F-Score. Chapter 7 by Mazin A.M. Al Janabi discusses how robust risk modeling techniques address the market-liquidity risk of multi-asset portfolios in emerging markets under

NGUYEN 9781800880894 PRINT.indd 29

21/09/2022 13:14

xxx  Handbook of banking and finance in emerging markets extreme market outlooks, and devotes particular attention to the use of reinforcement machine learning algorithms. Specifically, it shows an optimization algorithm combining Liquidity-adjusted Value-at-Risk and a multivariate dependence modeling technique, which is helpful for taking into account large fluctuations in emerging stock markets within the context of increasing uncertainty and external shocks. Chapter 8 by Paweł Miłobędzki and Sabina Nowak also investigates liquidity risk, but focuses on the components of bid–ask spreads of shares traded on the Warsaw Stock Exchange. It finds that portions of these spreads can be attributed to private information, the existence of temporary buy–sell imbalances, and price clustering. Moreover, the trading costs are sensitive to market imperfections such as information asymmetry and the asynchronous arrival of traders at the market. Part II – “Banking Profitability, Efficiency and Stability” – presents seven chapters providing in-depth analysis of profitability, efficiency, and stability in the banking sectors of emerging markets. It also focuses on banking competition, information and communication technology (ICT) effects on banking operations, and the role of central banks in maintaining financial stability. Chapter 9 by Ioanna Avgeri, Yiannis Dendramis, and Helen Louri examines the effect of the Single Supervisory Mechanism (SSM) throughout the profitability distribution of 78 directly supervised banks. Using an unconditional quantile regression analysis, they find evidence of a robust positive effect of the SSM in the lower quantiles of the profitability distribution, while the impact in the upper quantiles depends on the profitability index examined. The introduction of the SSM was also found to reduce the probability of bank insolvency, the effect being stronger for weaker banks. Chapter 10 by Constantin Gurdgiev and Li Jiaqi analyzes the impact of foreign shareholdings on the performance of Chinese commercial banks. It finds that foreign shareholdings have a significant negative impact on banks’ return on equity (ROE). Furthermore, banks’ non-performing loan (NPL) ratio is strongly and negatively correlated with foreign shareholdings. In terms of liquidity performance, foreign share ownership has a significant negative influence on banks’ loan-to-deposit ratio and a significant positive influence on loan loss coverage ratio. Overall, these findings suggest strong positive effects of foreign shareholdings on Chinese banks’ risk profiles. Chapter 11 by Francesco Guidi investigates the determinants of bank profitability using a sample of 169 commercial banks in seven countries of South Eastern Europe (SEE) alongside alternative profitability measures and a dynamic panel data analysis. It shows that total assets and loan loss provisions usually affect SEE banks’ profitability more than other variables over 2003–2012. The determinants of profitability on small banks also have a more significant impact than large banks, irrespective of the profitability measures. Chapter 12 by Antonio Samagaio, Pedro Verga Matos, and Isidora Manjate analyzes 11 determinants of bank performance in Mozambique (bank-specific, banking-industry specific, and macroeconomic-specific factors). A bank’s return on assets (ROA) is found  to be positively influenced by capital adequacy, diversification, and inflation, while liquidity risk and banking efficiency had a negative effect. When ROE is used as a performance measure, the explanatory variables with a positive effect were asset quality and level of bank concentration, while the degree of bank diversification had a negative effect.

NGUYEN 9781800880894 PRINT.indd 30

21/09/2022 13:14

Preface  ­xxxi Chapter 13 by Thanh Ngo and Tu Le uses a DEA approach to examine the impact of ICT development on the operational efficiency of the Vietnamese banking sector during the 2007–2019 period. The obtained efficiency and productivity measures are also regressed on several macro- and micro-characteristics of the Vietnamese banking sector. The authors find that ICT development positively contributes to the efficiency of Vietnamese banks. This contribution differs, however, among banks. Chapter 14 by Md. Nurul Kabir and Andrew C. Worthington investigates the banking competition–efficiency–stability relationship for 324 banks from 13 countries in the Middle East and Asia comprising both conventional and Islamic banking systems. They find that market power significantly increases efficiency in both banking sectors. However, the impact is higher for Islamic banks, leading to rejection of the ­competition– efficiency hypothesis. Efficiency positively affects the stability of conventional banks but has no significant impact on stability in Islamic banks. As to market power, it increases the stability of both banking sectors, suggesting that uniform competition policy can govern both banking systems since the impact of competition on stability is the same. Chapter 15 by Manisha Chuttoor, Dinesh Ramdhony, and Boopen Seetanah focuses on the operational autonomy of the Bank of Mauritius and the impact of macroprudential instruments on financial stability. Over the period 1976 to 2019, it mainly shows that central bank independence (CBI) contributes towards credit growth, stock market volatility, and exchange rate volatility. However, a negative link is found between CBI and economic growth, and among macroprudential policies, credit growth, and stock market volatility. Part III – “Towards Financial Resilience and Sustainability” – includes seven chapters covering some of the most relevant issues in corporate social responsibility, environmental, social and governance (ESG) issues, impact investing, and risk management practices. Responses to these issues would allow firms to cope with challenges and needs to improve their resilience to adverse shocks. Chapter 16 by Andreas Gruener sheds light on sustainable investments, focusing on emerging markets, including intrinsic, regulatory, and monetary developments. It discusses research on new frameworks (i.e., the ESG efficient frontier) and how ESG can be integrated into the fund universe, the role of ratings, and different approaches of active as well as index tracking ESG funds. Current trends such as thematic funds and impact investing are also examined, along with comparisons between developed and emerging markets. Chapter 17 by Thankom Arun, Claudia Girardone, and Stefano Piserà explores the role of banks in fostering the ESG agenda that has been pushed by the United Nations and regulators. Their evidence shows that banks’ specific adoption of international sustainability frameworks and agreements such as the Global Reporting Initiative are significant drivers of ESG engagement. Moreover, a stronger ESG regulatory approach enhances banks’ sustainability practices in BRICS countries (Brazil, Russia, India, China, South Africa), especially for those that have lower average ESG scores. Chapter 18 by Anas Mohammad Hussein Al-Jbour, Lei Xu, Damien Wallace, and Guodong Yuan examines the link between corporate social responsibility (CSR) disclosure and three proxies of cost efficiency – pure technical efficiency (PTE), overall technical efficiency (OTE), and scale efficiency (SE). The authors estimate the cost-efficiency of Islamic banks in the Gulf Cooperation Council, analyze annual reports, and develop a comprehensive CSR index. The overall findings show a lack of significant relationships

NGUYEN 9781800880894 PRINT.indd 31

21/09/2022 13:14

xxxii  Handbook of banking and finance in emerging markets between the three cost-efficiency proxies and the overall CSR disclosure index. By contrast, at the disaggregated level the CSR dimensions of mission and vision and commitment to community are significantly related to the banks’ PTE. The CSR dimension of commitment towards community is significantly related to the banks’ OTE. Chapter 19 by Mahfod Aldoseri and Andrew C. Worthington undertakes a comparative assessment of credit risk management and practices of Islamic and conventional banks in Saudi Arabia, using a questionnaire administered to more than 100 senior officers carrying out credit risk management across five primary aspects of the activity. The findings indicate that Islamic banks in Saudi Arabia appear to have better credit risk management and practices than conventional banks. Chapter 20 by Emmanuelle Dubocage and Evelyne Rousselet investigates the impact investing in Asian emerging markets, and illustrates the main stakes and challenges in their development. The examples discussed throughout the chapter show that the issues that are at stake for impact investing in developed markets – such as standardization of procedures, transparency, and impact-washing – also exist in emerging markets, but not in the same way. Chapter 21 by Olivier Mesly describes the phenomenon of consumer financial spinning, and analyzes its five hypothesized market stress factors in the context of emerging economies, namely: (1) external pressure for economic liberalization; (2) necessary spending on healthcare; (3) expectations of efficient regulations and banking standardization; (4) quest for political stability and integration; and (5) push for transparency despite lags in financial technology. It shows that, when exercising too much pressure, these factors may foster the accumulation of national debt that may become unsustainable, thus jeopardizing the economic growth potential. Chapter 22 by Romain Boissin addresses the influence of legal environment on the predictive abilities of financial analysts for emerging market initial public offerings (IPOs) over the period 1995 to 2018. It shows that, after controlling for IPO firm-specific characteristics as well as country-specific characteristics, there is a significant difference in the 12-month returns of non-orphan and orphan IPOs carried out in emerging markets with low legal environments. This result does not hold in emerging markets with high legal environments, suggesting that investors disregard analyst coverage when proper laws and regulations are not in place or their enforcement and sanctions are weak or non-existent. Consequently, investors could have the perception that analysts do not proceed as diligently as in a high legal environment. Part IV, “Innovative Models in Banking and Finance,” encompassing six chapters, covers the evolution of digital transformations and their impacts on the banking and finance sectors. Important issues such as the role of big tech, trust, digital platforms for financial services, and crowdfunding are also discussed. Chapter 23 by Silvio Andrae surveys the role of Bigtech companies in business life and in providing financial services through their outstanding digital platforms. It particularly examines the strategies they are pursuing in four typical emerging markets (Brazil, India, Kenya, and China). The findings show that financial technology (Fintech) plays a greater role than Bigtech, and that payments serve as the entry point into the financial services space of these emerging markets. Chapter 24 by Sylvie St-Onge, Michel Magnan, and Catherine Vincent provides an overview of how the digital revolution is upsetting the financial services industry. The

NGUYEN 9781800880894 PRINT.indd 32

21/09/2022 13:14

Preface  ­xxxiii authors specifically discuss its implications in bringing financial institutions to adopt new business models and revisit their talent management strategies. Chapter 25 by Oliver Vasquez and Leire San-Jose discusses the influence mechanism of the relevant factors of the Fintech phenomenon on user confidence in Fintech companies. It offers insights on Fintech’s development by providing a comprehensive analysis of the user–Fintech relationship and the incorporation of the determinants influencing trust as a critical and complementary element in the technology acceptance model. Chapter 26 by Johan Bouglet, Ghislaine Garmilis, and Olivier Joffre reviews the foundations of crowdfunding as an innovative model, the importance of ethics, and regulators’ responses to adapt legislative frameworks in developing countries. It also focuses on the ethics that characterize the crowdfunding platform sector today through a double response based on stakeholder theory and corporate governance theory. Chapter 27 by Maryam Alhalboni, Muhammed Shahid Ebrahim, and Wahyu Jatmiko examines the organizational structure of Takaful (Islamic insurance) as it constitutes an interesting case study of a relatively inefficient system that has grown in popularity. Using institutional theory to shed light on this puzzle, the authors show that the resolve of Muslim religious scholars and their political Islamic compatriots to regain legitimacy are the primary reasons this inefficient institution exists. This form of insurance leads to questioning the ideology of blending religious values in the broad financial services sector. Chapter 28 by Sylvie St-Onge, Catherine Vincent and Michel Magnan proposes a case study on Evovest – a portfolio management startup relying on artificial intelligence (AI). The case is intended to illustrate the challenges of launching a new portfolio manager with a novel business model based upon AI. As such, it does not aim to prescribe a particular course of action or a managerial approach. Part V – “Emerging Trends” – dedicates attention to emerging issues in the banking and finance sectors, including green finance, sustainability, digitalization, and financial transformations under the effect of today’s climate challenges and the COVID-19 pandemic. Cross-border banking, inflation targeting, and fiscal discipline also emerge as important topics in an interconnected world. Chapter 29 by Artie W. Ng provides reviews some of the emerging green finance hubs in Asia (the “Four Tigers – Hong Kong, Singapore, South Korea, and Taiwan) in light of their respective ESG policies and financial regulatory initiatives. It also offers a comparative analysis of complementary policies and regulatory approaches to advocating the development of ESG-driven equities and green bonds. Moreover, prospects and challenges of transforming the Four Tigers into a vibrant cluster of green finance hubs in Asia are articulated. Chapter 30 by Shidi Dong, Lei Xu, and Ron McIver examines commitment to sustainability reporting for a sample of Chinese listed banks. It particularly looks at changes in the extent and content of these disclosures, and the impact of the 2012 Green Credit Guidelines from the China Banking and Insurance Regulatory Commission. Overall, China’s banks have progressively committed to sustainable development through adopting sustainability reporting. Chapter 31 by Erik Feyen, Norbert Fiess, Ata Can Bertay, and Igor Zuccardi Huertas analyzes recent trends in bank activities of financial groups headquartered in 46 emerging markets and developing economies (EMDEs). It also studies the ownership structure of

NGUYEN 9781800880894 PRINT.indd 33

21/09/2022 13:14

xxxiv  Handbook of banking and finance in emerging markets 51 prominent financial groups from EMDEs. Cross-border groups in most regions have been found to grow in size, geographical reach, range of activities, and group complexity. The increasing relevance and complexity of cross-border banking pose challenges for policymakers in both home and host jurisdictions, as well as for the groups themselves to maximize the benefits of international financial integration while mitigating the risks. Chapter 32 by Carlo Raimondo and Patrick Coggi discusses the roles of digitalization and ESG goals from the perspective of policy agenda, regulatory challenges, banks’ business strategies, and external economic competition. The Swiss case is discussed as an example from which some lessons may be gained to address the future development of banking systems in emerging markets. Chapter 33 by Victoria Geyfman examines the complex components of the financial sector in emerging markets in Central and Eastern Europe (CEE) since the late 1990s. It also evaluates the continuing dominance of foreign-owned banks, and examines their evolving role from the early years of economic transformation through the period of the global financial crisis and beyond. The findings show that the Vienna Initiative was an effective crisis mitigation tool and evolves as an ongoing collaborative forum on financial regulation and risk management solutions. Chapter 34 by Henry Penikas investigates the passthrough of policy interest rates to deposit rates during the pandemic. It finds that the Bank of Russia’s double increase in the key rate had no impact on deposit rates at best, but was likely associated with their decline. Overshooting and further rate downgrade expectations could be a reason for this negative passthrough. Chapter 35 by Milojko Arsić, Zorica Mladenović, and Aleksandra Nojković examines the fiscal performance resulting from adopting an inflation targeting (IT) policy in emerging market economies of CEE and Central Asia over the period 1997–2019. The significant links between IT policy and fiscal variables are found to be limited. IT improved the cyclically adjusted overall and primary fiscal balance in 1997–2019, and has been improving the overall and primary fiscal balance since 2008. Moreover, the introduction of IT affects the reduction of the public debt-to-GDP ratio. It is important to note that, contrary to successful monetary performances of IT policy in many emerging economies, more coordination between fiscal and monetary policy seems to be needed for this policy to be effective beyond the monetary space. Part VI – “New Perspectives” – is composed of six chapters and covers discussions on new factors that need further reflection and adaptation for the sustainable development and growth of the banking and finance sectors in emerging markets. These include the role of financial surveys in economic policymaking, the convergence of interest rates, new models for financial intermediation, and COVID-wise effects on the banking sector. Chapter 36 by Sofía Gallardo and Carlos Madeira reviews the role of economic and financial surveys in the academic literature and policymaking, with a broad range of applications in EMDEs. Indeed, surveys are increasingly used in developed and emerging markets economies, providing information on a range of economic and financial ­phenomena – such as the updating of inflation expectations, future economic outlooks, the evolution of credit demand and supply factors, household balance sheets, behavioral biases, and personal finance. They also help inform some common research puzzles in finance, such as low financial market participation, unequal financial access, credit constraints, and inaccurate economic expectations.

NGUYEN 9781800880894 PRINT.indd 34

21/09/2022 13:14

Preface  ­xxxv Chapter 37 by Max Gillman focuses on the real short-term interest rates in the BRICS emerging markets and their potential convergence with the United States in the era of negative real interest rates. The finding that the US monetary policy of sustained negative real interest rates has spread to emerging regions after the financial crisis and lowered their output growth rates has important implications for emerging markets’ policy decisionmaking, particularly due to the context of increasing economic and political uncertainty. Chapter 38 by Sahminan Sahminan shows significant impacts of the COVID-19 pandemic on the Indonesian banking sector through the drop in bank lending. Future policy developments involving the banking sector should pay close attention to liquidity, capital adequacy ratio, and non-performing loans ratio. Chapter 39 by Theodore Syriopoulos investigates the dynamic adjustments and prospects in the global financial landscape, using the commercial shipping as a case study. The author shows that disruptive technologies such as digitalization, AI, machine learning, blockchain, and cryptocurrencies can eventually influence the shipping business model, inducing critical implications for financial decisions. Chapter 40 by Vasily Tkachev, Vadim Grishchenko, and Karl Summanen discusses long-term development trends of the Russian financial market in an environment of instability and sanctions. It documents that, while a modern legal framework is in place, several crucial issues remain unresolved, and the Bank of Russia has to undertake structural reforms in four key areas: promoting competition in the financial market; creating an environment of trust; maintaining financial stability; and ensuring access to financial services and capital. These policy changes would help Russia fall into a random path of accommodating sudden shocks. Finally, Chapter 41 by W. Travis Selmier II focuses on the modern Chinese financial system. Reluctant to empower financial institutions politically, and concerned about shadow banking, China’s national government is coopting data from large online technology firms to engage in financial and social monitoring through the new Social Credit System. A major finding is that China’s different approach might provide more stability than western models, but may also compromise the privacy of borrowers and capital providers. The author also argues that contention in China’s domestic financial system and within the international financial system may influence development of this algorithmic financial governance model. I believe that the elements discussed in these 41 high-quality chapters are essential for emerging markets to overcome the pandemic, shape the new normal, and foster economic sustainability. I would like to thank all contributors for their outstanding commitment and the invaluable knowledge they bring to the audience. My special thanks also go to Daniel Mather, Commissioning Editor at Edward Elgar Publishing, for his kind help and assistance in preparing this Handbook. Duc Khuong Nguyen

NGUYEN 9781800880894 PRINT.indd 35

21/09/2022 13:14

NGUYEN 9781800880894 PRINT.indd 36

21/09/2022 13:14

PART I FINANCIAL MARKETS, INSTITUTIONS AND MONEY

NGUYEN 9781800880894 PRINT.indd 1

21/09/2022 13:14

1. Booms, bubbles, blow-outs: exploring patterns in China’s credit expansion Ron McIver, Lei Xu and Shiao-Lan Chou

1. INTRODUCTION This chapter focuses on the level of credit expansion within China’s banking sector and whether patterns of credit growth in China, the world’s second largest economy, should  be identified as being excessive and risk future precipitation of banking crises (e.g., Chen and Kang 2018). Our motivations are as follows. First is the importance of bank credit to China’s gross domestic product (GDP) growth given its bankbased financial system (Allen et al. 2012). Second, emphasis on the use of credit-to-GDP and credit gap ratios as potential early warning indicators (EWIs) of banking crises and in the determination of countercyclical capital buffers (BIS 2010; Drehmann and Juselius  2014). Third, the greater use of panel data in assessing the relationship between  financial deepening and economic growth and the optimal level of financial development and economic growth (Christopoulos and Tsionas 2004, Graff and Karmann 2006, Apergis et al. 2007, Hassan et al. 2011). Finally, that econometric modelling for assessment of credit booms in transition economies has largely been based on developed and emerging market economy data (e.g., Cottarelli et al. 2005). However, financial repression is a key feature of former communist country financial systems, with  market fragmentation and pricing inefficiencies being associated with centralized  control (McKinnon 1973, Miurin and Sommariva 1993). Thus, this practice ignores potentially important information in data related to transition economy financial deepening in response to a progression towards market-based financial regulation and policy. We quantitatively examine China’s financial deepening and equilibrium level of credit using both Hodrick-Prescott (HP) filter and panel data methodologies. Consistent with a substantial component of the financial stability assessment literature, our focus is on bank credit to the private sector (BCPS).1 Our panel analysis compares China’s economic fundamentals and bank credit expansion to that of major developed, emerging and transition economies that have undergone historical expansions in this measure. It is use of data on the latter set of economies that provides a key difference in our study. We identify that credit growth in China has not, in general, outpaced levels expected for its growing economy. The rest of the study is structured as follows. Section 2 briefly reviews relevant theoretical and empirical literature on credit growth, and provides background information on China’s credit markets. Section 3 addresses China’s recent GDP and credit growth, and similarities between China and the Central and Eastern European (CEE) countries. Section 4 provides measurement of the BCPS ratio and its equilibrium level. Section 5 presents the results and Section 6 concludes the study. 2

NGUYEN 9781800880894 PRINT.indd 2

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­3

2.  LITERATURE REVIEW 2.1  Credit Behaviour: Trend, Cycle and Boom An abundant empirical literature has evolved to confirm a positive relationship between growth and financial intermediation (see Beck 2008). An efficient financial sector contributes to economic growth. However, the causality between financial development and economic growth can also be the opposite, with economic development increasing financial intermediation. These two possible patterns in the causal relationship between financial development and economic growth are the ‘supply-leading’ and ‘demand-following’ hypotheses. The supply-leading hypothesis presumes a causal relationship from financial development to economic growth (Patrick 1966). The deliberate creation and promotion of a financial system precedes the demand for its services, and thus leads to real economic growth. Numerous theoretical and empirical studies have shown that financial development is important and causes economic growth (McKinnon 1973, King and Levine 1993a and 1993b, Neusser and Kugler 1998, Levine et al. 2000). On the other hand, the demandfollowing hypothesis supposes a causal relationship from economic growth to financial development, indicating that the financial sector responds to economic growth passively from the financial side (Gurley and Shaw 1967, Jung 1986). As the real side of the economy grows, demand for new financial services increases. Thus, growth induces an expansion in the financial sector (i.e., financial deepening). With respect to the cyclical component, theoretical models can be categorized into two major groups: financial accelerator models and behavioural models. This depends on whether the pro-cyclical nature of credit growth derives from behaviour in the real economy or from the financial sector. It is the latter that is a concern in the case of China, given concerns about excess liquidity in China’s banking sector (Liu and Wray 2010). Financial accelerator models examine the cyclical component over the economic cycle through the lens of collateralized credit (Bernanke et al. 1999, Kiyotaki and Moore 1997). When economic conditions are depressed and collateral values decrease, even borrowers with profitable projects find it difficult to obtain funding due to information asymmetries. When economic conditions improve and collateral values rise, these promising borrowers are again able to gain access to external finance, which adds to the economic stimulus. Although the financial accelerator presumably plays a part in magnifying business cycles, it does not initiate these swings in economic activity. Behavioural models explain an additional source of financial pro-cyclicality: inappropriate responses by financial market participants to changes in risk over the term of the economic cycle. Borio, Furfine and Lowe (2001) argue that these inappropriate responses mainly result from difficulties in measuring the time dimension of risk and incentives of market participants to react in socially suboptimal ways to risk, even when correctly measured. It is this latter case that may be argued to be relevant in China. Our focus in this study is not, per se, on assessing the short-run dynamics associated with credit provision. Rather, it is with whether growth is evidencing cyclical (boom) or trend (financial deepening) behaviour. While, theoretically, credit growth can be separated into trend, cycle and boom, distinguishing between normal (trend) and excessive credit growth is difficult. This reflects the need to quantify an accepted definition of ‘excessive’ and a corresponding threshold value

NGUYEN 9781800880894 PRINT.indd 3

21/09/2022 13:14

4  Handbook of banking and finance in emerging markets (Buncic and Melecky 2013). The existing empirical literature, much of which considers equilibrium credit-to-GDP ratios, thus uses a variety of methods to identify credit booms.2 The first approach sees some authors establishing general ‘speed limits’ for credit growth. Credit expansion beyond this specified growth rate is identified as excessive, reflecting the common observation that rapid growth in loan portfolios is often present in individual bank failures and in some types of systemic crisis. For example, Honohan (1997) suggests that temporary ‘speed limits’ are worth considering in a market that has many new and inexperienced entrants who lack adequate supervisory resources. However, these ‘speed limits’ are unlikely to be an optimal tool in the long run. Thus, Kraft and Jankov (2005) suggest that lending booms should not be regulated away ex ante by speed limits or other preventive measures. Authorities should react in a timely and effective manner to problems as they arise, although they cannot be diagnosed in real time, limiting the effectiveness of such arbitrary structures. Given such limitations, we focus on the alternative, empirically based approaches in our study. A second approach tries to identify a trend in the evolution of credit through univariate time series methods. The Hodrick-Prescott (HP) filter is frequently used in the literature to identify the cyclical component of a time series from the raw data. The estimated trend is considered as equilibrium financial sector deepening. A credit boom is defined as credit growth higher than a given threshold around the trend. Gourinchas et al. (2001) consider two alternative threshold definitions: relative and absolute deviations. Relative deviations are based on the relative difference between the actual and predicted credit-to-GDP ratio, implying that financial deepening is not related to the lending boom. Absolute deviations are based on the absolute discrepancy between the actual and predicted credit-to-GDP ratios, implying financial deepening is related to the lending boom (i.e. countries with a more developed financial sector are more likely to experience lending booms). Specified use of the HP-filter approach under Basel III (Drehmann and Tsatsaronis 2013) supports our use of this method. A credit-to-GDP gap in excess of 2 per cent indicates a moderate excess in credit growth, while a gap exceeding 10 per cent indicates a high risk of financial distress. That noted, a problem with this approach is that it ignores institutional change and reductions in financial repression as potential determinants of large shifts in the ratio. In the case of a transition economy, such as China, both factors are present. A third approach tries to explain the equilibrium level of the credit-to-GDP ratio through fundamental macroeconomic variables. Here the estimated parameters of developed and other emerging countries with longer time series than those of transition countries are applied to assess their credit growth.3 A wide range of econometric techniques are employed, and in-sample and/or out-of-sample estimation is implemented. For example, some authors use one-country Vector Error Correction Models (VECM) (Hofmann 2001, Brzoza-Brzezina 2005). Some studies aggregate national-level data for the Eurozone and then carry out VECM estimations (Calza et al. 2003 and 2006, Schadler et al. 2005). Panel data methods are also broadly applied. Cottarelli et al. (2005) pool developed and non-transition developing countries together. Balazs et al. (2006) estimate the panel equilibrium level of private credit to GDP in 11 transition economies from the CEE region based on various combinations of OECD and emerging economy data, applying their estimated parameters out-of-sample to the transition countries. However,

NGUYEN 9781800880894 PRINT.indd 4

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­5 a weakness in each of these studies is that they ignore differences in institutional characteristics and the levels of economic and financial development of each country, each of which suggests that heterogeneity in the parameters that determine the credit-to-GDP ratio is to be expected (Buncic and Melecky 2013). 2.2  Transition to a Market Economy and Bank Credit Growth in CEE Countries The expansion of the European Union (EU) in the 1990s prompted the CEE countries to reform their domestic banking systems in line with EU practice. Key reforms implemented at the commencement of transition included: introduction of a two-tier banking system; lifting sectoral restrictions on special banks; permission to operate privately owned banks; allowing foreign banks to enter the market; liberalization of the licensing policy; and the implementation of a new legal framework and supervisory system. These policies encouraged foreign strategic investors to enter host financial markets on a large scale, capturing more than two-thirds of banking sector market shares. Western European banks dominated the consolidation process and controlled local markets, especially the credit business. Loan size for both state-owned banks (SOBs) and foreign  banks in the CEE transition countries increased year by year from the late 1990s.  The dominant position of the SOBs as lenders was taken by foreign banks, which  became the major source of credit funds for the public and private sectors by 2000.  Except for the Czech Republic, private credit in the CEE transition countries exceeded public credit from 1998, while, except for Slovenia, foreign banks in the CEE transition countries provided most of their loans to the private sector. The SOBs primarily lent to the public sector, with Estonia providing a typical example. Here SOBs did not lend to the private sector, with private credit provision being taken over by foreign banks. Such extensive foreign direct investment (FDI) had substantial positive impacts for the CEE countries. However, it created new problems through rapid credit expansion, raising concerns about macroeconomic stability and fuelling speculative excesses in several countries in the region. This triggered some central banks to initiate restrictive measures. This is despite data suggesting that, while overall capital adequacy ratios (CARs) experienced a slight decline, they still exceeded the regulatory minimums.4

3.  COMPARISON OF CHINA AND CEE COUNTRIES 3.1  China’s Recent Growth and Banking Reforms Many of China’s successes have been achieved since joining the World Trade Organization (WTO) in 2001. In the post-WTO era China has implemented a range of socio-economic reforms, including: an increased initial pace of privatization of small state-owned enterprises (SOEs), alongside a reduction in privatization of large SOEs (Yang 2007, Naughton 2011); the introduction of modern corporate governance systems (Shan and McIver 2011); and provision of both increased market access and reduced ownership restrictions for foreign investors (Kwon and Zhou 2009). Importantly, through the troika of exportled, fixed investment-led and domestic consumption-led stimuli to aggregate demand,

NGUYEN 9781800880894 PRINT.indd 5

21/09/2022 13:14

6  Handbook of banking and finance in emerging markets China has been able to preserve high levels of GDP growth throughout its transition towards a more market-oriented economy. Given claims that China will shortly overtake the US to become the world’s largest economy (e.g., Giles 2014), and that the 21st century will be the ‘China Century’, examining the sustainability of China’s economic miracle from a range of perspectives is required.5 China’s growth has been supported by its fast-growing banking industry, one that controls most of China’s aggregate financial resources (Allen et al. 2012), and which remains largely state controlled. State control has facilitated the central government’s development of lending policies for the banks. Such policies include the Four Keys (key region, key industry, key enterprise and key products) and the more recent Green Loans (loans given to clean energy enterprises and taken away from polluting manufacturers). Additionally, China’s government has been able to initiate a range of specific policies to support and maintain high levels of GDP growth through its banking system. For example, in 2008, following the inception of the global financial crisis (GFC), a 4 trillion renminbi (RMB) stimulus package was announced. This package was distributed through the banking system, resulting in a rapid increase in lending growth. In the first half of 2009, this saw new bank loans rise by a total of 7.37 trillion RMB (about 1.17 trillion USD) (Zhu 2012). Unlike most other major economies, China has not experienced a banking-sector crisis since its Open-Door policy began in 1978.6 Additionally, the GFC seemingly had limited impact on the fast-growing Chinese economy. China, again in contrast to other major economies such as the US and Japan, did not experience a credit crunch following the imposition of the Basel Accords (Berger et al. 2000, Gorton and Winton 2017).7 In fact, the lending practices of China’s banks seem to have been largely unconstrained by the introduction of stricter capital regulations. For example, the China Banking Regulatory Commission (CBRC) showed that China’s bank assets increased by 363 per cent between 2004 and 2012 – an average annual growth rate of 40 per cent.8 Such rapid growth in bank assets may indicate either structural benefits to the real economy through financial deepening (Buncic and Melecky 2013) or the initiation of an asset price bubble prior to a financial crisis (Coudert and Pouvelle 2010). 3.2  Similarities between China and CEE Banking Reforms Similarities exist in the banking reforms of China and the CEE countries. First, there is a transition process regarding attitudes to foreign banks: exclusion, prudence and acceptance. Second, China has experienced external pressure through WTO membership, while CEE countries experienced pressure through joining the EU. Third, China and the CEEs have a similar path for banking development and reform: cleaning up non-performing loans (NPLs) and recapitalization, privatization and, finally, introducing foreign strategic investors.9 This suggests the benefits of reviewing the CEE countries’ experiences during their banking reforms. Reflecting these similarities, bank credit and other credit measures for China have increased in recent years, in line with the development of non-SOE sector activity and private institutions. Table 1.1 summarizes alternative measures for the growth in credit that has financed China’s growth over the 1985 to 2019 period. These are bank credit to the private sector (BCPS) and domestic credit to the private sector (DCPS). When considered in terms of growth in level, both measures evidence rapid and similar increases,

NGUYEN 9781800880894 PRINT.indd 6

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­7 Table 1.1  Stylized facts: metrics on credit growth in China, 1985−2019 Measure Cumulative growth Average annual rate (level), % growth rate (level), % BCPS DCPS

188.01 187.62

5.53 5.52

Cumulative change Average annual (% of GDP) change (% of GDP) 99.20 100.56

2.94 2.92

Source:  IMF, International Financial Statistics; World Bank, Global Financial Development Database; authors’ calculations.

consistent with China having a bank-based financial system. For example, the cumulative growth rate in the level of BCPS is 188.01 per cent, with an average annual growth rate of 5.53 per cent. However, when considered in relation to GDP, growth in credit appears more muted. For example, the total change in the BCPS ratio is 100.06 per cent, with an average annual change of 2.94 per cent.

4.  METHODOLGY AND DATA 4.1  Speed Limits to Bank Credit Growth To examine more formally whether the recent rise in BCPS ratios in China appears abnormal relative to its previous trend, we initially apply the HP-filter approach. Consistent with the approach of the Bank for International Settlements (BIS 2010), we define a lending boom as an episode where the BCPS ratio deviates from a rolling, backwardlooking, country-specific stochastic trend by 10 per cent or more in relative terms. Moderately excessive credit growth is defined where the BCPS ratio deviates from trend growth by more than 2 but less than 10 per cent. The idea is that the stochastic trend represents the normal historical pace of credit growth. Table 1.2 reports the BCPS ratio and its trend component for the 1985 to 2019 period, together with absolute and relative deviations. As discussed earlier, in 2009 China went through a credit explosion that pushed its BCPS ratio up by just over 22 per cent. However, the statistical results from the HP-filter approach suggest a slightly different picture. We detect credit growth at moderately excessive levels (i.e., between 2 per cent and 10 per cent) for 1986, 1987, 1990, 1991, 1993, 1998, 1999, 2003, 2004, 2009, 2010, 2015 and 2016 but not a lending boom. This result shows one of the limits of the HP-filter approach with short time-series. Recent high rates of credit growth in China are not identified as credit booms because this rapid growth is included in the trend component, reflecting the stability in the rise in credit growth (Figure 1.1). 4.2  Estimating an ‘Equilibrium’ Level of Growth for the BCPS Ratio BCPS ratios in countries where market institutions have been and are being established, and the rate of change in these ratios in response to changes in economic fundamentals, are potentially useful benchmarks in evaluating whether increases in China’s BCPS ratio

NGUYEN 9781800880894 PRINT.indd 7

21/09/2022 13:14

8  Handbook of banking and finance in emerging markets 180.00 160.00

Bank credit to the private sector (% of GDP)

140.00 120.00 100.00 80.00 60.00 40.00

0.00

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019

20.00

Source:  IMF, International Financial Statistics.

Figure 1.1  BCPS in China, 1985–2019 (billion RMB) are consistent with the process of financial deepening. Given the lack of conclusiveness of the HP-filter approach, we utilize these benchmarks to determine whether China’s recent credit growth can be considered as excessive. For this purpose, we estimate an econometric model of the BCPS ratio based on an unbalanced panel data set comprised of advanced economies,10 emerging market economies,11 and transition ­economies.12 The sampled period for estimation is 1980 to 2012. We then use the econometric model results to estimate ‘equilibrium’ rates of change in the BCPS ratios for China. Finally, we compare these credit growth estimates to China’s actual growth in its BCPS ratio. While an assessment of causality is the focus in the credit-growth nexus literature, our objective is to identify the relationship between BCPS ratios and a set of economic and institutional fundamentals. This is to assess the impact of changes in these fundamentals on credit growth. The variables chosen to measure these fundamentals are as follows. PCGDPit is the log of GDP per capita, measured in terms of USD at purchasing power parity (PPP) exchange rates. This variable captures the level of economic development, which may be argued to have either a positive or negative relationship to BCPS. Arguments favouring a positive relationship between the equilibrium level of credit and economic growth suggest that countries with a higher income per capita should have a higher degree of financial deepening. Arguments for the contrary position

NGUYEN 9781800880894 PRINT.indd 8

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­9 Table 1.2  BCPS deviations from trend, 1985–2019 Year

Actual BCPS (% GDP) (a)

HP-filter estimated BCPS (% GDP) (b)

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019

65.33 74.74 76.17 71.97 75.12 84.05 86.32 84.41 94.99 84.39 83.10 88.54 96.51 105.07 110.27 111.01 109.87 117.33 125.52 118.55 111.81 109.13 105.76 101.99 124.40 126.57 123.09 128.92 134.32 140.24 152.58 156.22 154.88 157.81 165.39

68.02 71.06 73.66 75.99 78.60 81.43 83.82 85.57 86.84 87.64 89.27 92.50 97.12 102.29 107.08 110.99 114.02 116.21 116.90 115.64 113.32 111.34 110.82 112.56 116.51 120.98 125.49 130.49 136.03 141.93 147.70 152.62 156.71 160.61 164.62

Absolute Deviation (%) (a–b) −2.69 3.68 2.52 −4.02 −3.48 2.62 2.50 −1.16 8.15 −3.26 −6.17 −3.96 −0.61 2.78 3.19 0.03 −4.15 1.12 8.62 2.91 −1.52 −2.21 −5.06 −10.57 7.88 5.60 −2.40 −1.57 −1.72 −1.69 4.88 3.60 −1.83 −2.80 0.77

Relative Deviation (%) ((a–b)/a) −4.12 4.93 3.30 −5.58 −4.63 3.11 2.90 −1.37 8.58 −3.86 −7.43 −4.47 −0.63 2.64 2.89 0.02 −3.78 0.95 6.87 2.46 −1.36 −2.03 −4.78 −10.36 6.34 4.42 −1.95 −1.22 −1.28 −1.21 3.20 2.31 −1.18 −1.77 0.46

Note:  Data from 1985 to 2019 used in HP-filter process. Source:  IMF, International Financial Statistics; World Bank, Global Financial Development Database; and authors’ calculations.

suggest that increasing productivity, through higher profits in economic booms, makes it possible to use more internal funds rather than bank credit. Similarly, temporarily lower income  in  economic downturns makes it possible for households to increase debt  levels  to  smooth consumption. LENDit is the long-term lending interest rate,

NGUYEN 9781800880894 PRINT.indd 9

21/09/2022 13:14

10  Handbook of banking and finance in emerging markets which  also  simultaneously determines the equilibrium between the demand for and supply of loans. Some studies use both short and long interest rates to address this issue. For example, Calza et al. (2006) argue that whether either rate is important depends on the respective share of loans with fixed versus variable interest rates. However, other authors construct one aggregate time series. Since lower interest rates promote credit to the private sector, we expect a negative sign for this variable. SPRit is the difference between lending and deposit rates, and captures financial liberalization. The positive effect of liberalized financial and capital markets on financial deepening has been well established in the literature (McKinnon 1973, Abiad and Mody 2005, Cottarelli et al. 2005). This effect is approximated by a decrease in the spread, reflecting more intensive competition among banks and also between banks and other financial intermediaries. INFLit is measured as  the annual rate of change in the consumer price index (CPI). High inflation is expected to have a damaging effect on financial deepening and to cause a drop in bank credit to the private sector. For example, Khan et al. (2006) find that if inflation rates are within or less than the 3–6 per cent range, small increases may have either no effect or a small positive impact on financial activity. However, if inflation rates are above this threshold, inflation has a significant negative effect on financial deepening. CGSOEit is the bank credit to the government and SOE sectors as a percentage of GDP. This variable may play a role in determining private credit through the crowding-out effect, with an expected negative sign. Thus, any increase/decrease in bank credit to the government sector will lead to a decrease/increase in bank credit to the private sector. It should be noted that Cottarelli et al. (2005) use the ratio of the stock of public debt (domestic and external) to GDP to capture possible crowding-out effects. They find that public debt as a stock variable is more appropriate than different flow variables of the government budget, such as fiscal deficit, government expenditure, etc. Thus, we also include this variable. GOVDEBTit is the total government debt-to-GDP ratio. As noted above, this variable may play a role in determining private credit through the crowding-out effect, with an expected negative sign. Any increase/decrease in government debt stocks may crowd out/crowd in private sector debt, and so will lead to a decrease/increase in bank credit to the private sector. ERit is the local currency price of a USD for each of the sample economies. This variable captures an element of exchange rate risk to both borrowers and foreign lenders. Large increases in the exchange rate may make repayment of foreign currency debt more costly or increase the probability of default – potentially reducing both the demand for and supply of foreign debt funding, and so limiting expansion of the credit-to-GDP ratio. BBSDi is a dummy variable taking the value of 1 if country i has a bank-based financial system and 0 otherwise. BBSD will assist in identifying differences in the credit-to-GDP ratio based on the greater role that banks take in bank-based financial systems. This latter feature is particularly important in the case of China. BNKCRSit is a dummy variable taking the value of 1 if country i is identified as having a bankingsector crisis in year t and 0 otherwise. BNKCRS captures the impact of banking-sector crises on the ability to create credit, and thus on growth in the level of the credit-to-GDP ratio during such periods. Table 1.3 provides summary statistics on each of the key ­variables. Table 1.4 provides the correlation matrices between the above identified variables for all countries in our data set. Tables 1.5 to 1.7 provide separate correlation matrices for each

NGUYEN 9781800880894 PRINT.indd 10

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­11 Table 1.3  Summary statistics for key variables Variable

Mean

Std. Dev.

BCPS PCGDP LEND SPR GOVDEBT CGSOE INFL ER EMERG TRANS BBSD BNKCRS EMRLEND TRRLEND CRSRLEND EMCGSOE TRCGSOE BBCGSOE EMDEBT TRDEBT BBDEBT

72.91 9.56 25.09 5.44 53.17 13.96 5.44 273.92 0.28 0.19 0.62 0.14 2.54 1.08 0.83 3.42 1.66 9.91 13.18 6.03 34.32

46.32 0.84 144.52 5.73 30.49 12.69 6.58 1132.18 0.45 0.39 0.49 0.35 6.99 3.53 4.14 7.25 4.53 13.89 24.33 14.82 38.34

Min

Max

2.42 6.16 0.50 −6.91 3.69 0.04 −4.48 0.48 0.00 0.00 0.00 0.00 −16.56 −24.78 −16.56 0.00 0.00 0.00 0.00 0.00 0.00

232.10 11.02 4774.53 55.80 210.25 74.20 85.74 10389.90 1.00 1.00 1.00 1.00 91.61 39.72 91.61 41.72 42.07 74.20 164.99 98.98 210.25

Note:  Here and in Tables 1.4–1.7 see the Appendix for definitions of all variables.

sub-set: advanced economies, emerging market economies and transition economies, respectively. As noted above, differences in the relationships between key variables and the BCPS ratio may be expected considering specific institutional characteristics and the levels of economic and financial development of each country (Buncic and Melecky 2013). Additionally, institutional change and reductions in financial repression associated with transition economies, such as China, may also impact these relationships. A brief examination of Tables 1.4 to 1.7 confirms the existence of some potentially important differences in the size and sign of the correlations between key variables and the BCPS ratio, both between sub-samples and between the full data set and each subsample. These include: the size and signs of the correlations between PCGDP and BCPS and GOVDEBT and BCPS; the size of the correlations between INFL, RLEND and the SPR with the BCPS; and the size and signs of the correlations between ER and BCPS. Based on the distinctions identified for developed, emerging, and transition economies in Tables 1.5 to 1.7, we include the following additional dummy variables for use in our regression models: EMERGi is a dummy variable taking the value of 1 if country i is an emerging market economy and 0 otherwise; and TRANSi is a dummy variable taking the value of 1 if country i is a transition economy and 0 otherwise. EMERG and TRANS are intended to capture specific characteristics of sub-samples within our data set. Drawing on previous empirical models of the BCPS ratio (see Balazs et al. 2006, Cottarelli et al. 2005, Kiss et al. 2006, Zdzienicka 2011), and giving consideration to our

NGUYEN 9781800880894 PRINT.indd 11

21/09/2022 13:14

12

NGUYEN 9781800880894 PRINT.indd 12

21/09/2022 13:14

1.00 0.94 0.82 0.39 −0.01 0.39 −0.06 −0.07 −0.08 −0.08 −0.28 −0.30 −0.15 0.15

1.00 0.92 0.18 −0.04 0.21 −0.09 −0.10 −0.11 −0.10 −0.27 −0.35 −0.20 0.16 1.00 0.19 −0.03 0.12 −0.09 −0.10 −0.11 −0.09 −0.23 −0.36 −0.11 0.17 1.00 0.11 0.62 0.03 −0.01 −0.02 −0.01 −0.15 −0.08 0.15 0.00 1.00 0.27 0.03 0.06 0.04 −0.02 0.19 −0.08 0.08 −0.01 1.00 −0.13 0.02 −0.06 0.01 −0.36 −0.10 0.07 0.16 1.00 0.78 1.00 0.85 0.96 1.00 −0.02 −0.02 −0.01 0.03 0.04 0.04 0.09 0.06 0.08 −0.03 −0.02 −0.02 0.06 −0.01 −0.01

1.00 −0.10 0.15 0.13 0.02

LEND ER

1.00 −0.40 0.29 0.03

1.00 −0.19 0.00

1.00 0.01

1.00

TRANS EMERG BBSD BNKCRS

Source:  Here and in Tables 1.5–1.7 IMF, International Financial Statistics; World Bank, Global Financial Development Database; and authors’ calculations.

DCFS DCPS BCPS CGSOE PCGDP GOVDEBT INFL SPR LEND ER TRANS EMERG BBSD BNKCRS

DCFS DCPS BCPS CGSOE PCGDP GOVDEBT INFL SPR

Table 1.4  Correlation matrix (all 55 economies, including China)

13

NGUYEN 9781800880894 PRINT.indd 13

21/09/2022 13:14

DCFS DCPS BCPS CGSOE PCGDP GOVDEBT INFL SPR LEND ER BBSD BNKCRS

1.00 0.92 0.71 0.36 0.52 0.33 −0.39 −0.30 −0.54 −0.06 −0.11 0.34

DCFS

1.00 0.85 0.09 0.61 0.08 −0.45 −0.41 −0.65 −0.13 −0.25 0.41

DCPS

1.00 0.08 0.57 −0.07 −0.43 −0.38 −0.63 −0.07 −0.11 0.39

BCPS

1.00 −0.01 0.62 −0.07 0.25 0.03 0.04 0.35 −0.02

CGSOE

1.00 0.15 −0.64 −0.15 −0.70 −0.13 −0.21 0.33

PCGDP

Table 1.5  Correlation matrix (23 developed economies)

1.00 −0.20 0.17 −0.16 0.16 0.20 0.16

GOVDEBT

1.00 0.18 0.79 0.09 0.16 −0.17

INFL

1.00 0.49 0.04 0.24 −0.09

SPR

1.00 0.10 0.15 −0.25

LEND

1.00 0.21 −0.01

ER

1.00 0.02

BBSD

1.00

BNKCRS

14

NGUYEN 9781800880894 PRINT.indd 14

21/09/2022 13:14

DCFS DCPS BCPS CGSOE PCGDP GOVDEBT INFL SPR LEND ER BBSD BNKCRS

1.00 0.92 0.85 0.35 0.31 −0.04 0.01 −0.05 −0.04 −0.10 −0.29 0.03

DCFS

1.00 0.93 0.09 0.34 −0.15 −0.04 −0.09 −0.08 −0.11 −0.30 −0.02

DCPS

1.00 0.18 0.32 −0.08 −0.03 −0.09 −0.09 −0.10 −0.25 0.02

BCPS

1.00 0.13 0.47 0.06 −0.01 −0.01 −0.03 −0.06 0.01

CGSOE

1.00 0.00 −0.02 −0.01 −0.02 −0.09 −0.13 −0.05

PCGDP

Table 1.6  Correlation matrix (18 emerging economies)

1.00 0.12 0.20 0.24 −0.05 0.19 0.25

GOVDEBT

1.00 0.93 0.98 −0.04 −0.06 0.12

INFL

1.00 0.98 −0.03 −0.06 −0.02

SPR

1.00 −0.03 −0.04 −0.01

LEND

1.00 0.23 0.04

ER

1.00 −0.03

BBSD

1.00

BNKCRS

15

NGUYEN 9781800880894 PRINT.indd 15

21/09/2022 13:14

DCFS DCPS BCPS CGSOE PCGDP GOVDEBT INFL SPR LEND ER BBSD BNKCRS

1.00 0.89 0.86 0.24 0.12 −0.14 −0.07 0.00 −0.04 0.12 0.24 0.09

DCFS

1.00 0.99 0.08 0.00 −0.31 −0.15 −0.09 −0.13 0.04 0.22 0.08

DCPS

1.00 0.10 0.00 −0.31 −0.14 −0.09 −0.13 0.04 0.21 0.08

BCPS

1.00 0.38 0.43 0.15 0.05 0.06 0.12 0.08 0.12

CGSOE

1.00 0.23 0.06 0.22 0.18 −0.11 0.08 −0.03

PCGDP

1.00 0.13 0.25 0.22 0.37 −0.16 0.15

GOVDEBT

Table 1.7  Correlation matrix (16 transition economies, including China)

1.00 0.35 0.42 −0.07 0.00 −0.02

INFL

1.00 0.91 −0.08 −0.03 −0.01

SPR

1.00 −0.07 −0.02 −0.02

LEND

1.00 0.14 0.16

ER

1.00 0.00

BBSD

1.00

BNKCRS

16  Handbook of banking and finance in emerging markets observations on the differences between correlations for each of the sub-groups of countries represented in our sample data, our regression equation is formally written as:     

  

The interaction terms are between the EMERG, TRANS, BBSD and BNKCRS dummies and the CGSOE, GOVDEBT, LEND/RLEND, SPR and INFL variables. This is to capture potential differences with respect to correlations between these dummies and variables due to institutional differences between country sub-samples. Dropping China from the data set, we estimate both pooled data and random effects/ mixed panel data models using annual data for 56 economies over the 1980 to 2012 period. The 56 economies are comprised of 23 advanced economies, 18 emerging market economies, and 15 transition economies. This provides an unbalanced panel with a maximum of 32 years and a minimum of 12 years of data for each of the economies included in the sample. We test whether a pooled data or random-effects panel data model is appropriate by using a Breusch and Pagan Lagrangian multiplier test for random effects (Table 1.10).13 For the pooled data robustness checks comprise substitution of the real lending rate (RLEND) in place of LEND; regression using quantile methods; the quantiles being 25, 50 and 75; and Tobit regression to account for the left-censoring of the BCPS ratio. For the panel data model robustness checks involve substitution of the real lending rate (RLEND) in place of LEND, and the use of domestic credit to the private sector (DCPS) in place of BCPS. Additionally, to assess whether spurious regression is present due to the use of ratios involving GDP for both the dependent and selected explanatory variables (Kronmal 1993, Powell et al. 2009), regression on the level of credit is undertaken. In this case GDP is included as an explanatory variable. The estimated equations on the pooled data models, including robustness checks, are reported in Tables 1.8 and 1.12, while those for the panel data are in Tables 1.9 and 1.13, respectively.

5. RESULTS 5.1  Estimated Coefficients of the BCPS Models Table 1.8 presents the results of the pooled data regression for the 54 sample economies, without and with a year variable. Most of the coefficients have the expected signs. The coefficient on GDP per capita is positive and significant, consistent with the findings in the previous studies – that higher levels of GDP per capita are correlated with higher levels of financial deepening. The negative and statistically significant coefficient on the long-term nominal lending interest rate suggests that lower interest rates promote credit to the private sector. However, as indicated in the coefficients on the interaction terms between LEND and EMERG and TRANS, the impact of increases in the lending

NGUYEN 9781800880894 PRINT.indd 16

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­17 Table 1.8  Basic pooled data regression results Coefficients Pooled Data (1) PCGDP LEND/RLEND SPR GOVDEBT CGSOE INFL ER EMERG TRANS BBSD BNKCRS EMLEND/EMRLEND TRLEND/TRRLEND CRSLEND/CRSRLEND EMCGSOE TRCGSOE BBCGSOE EMDEBT TRDEBT BBDEBT CONSTANT

NGUYEN 9781800880894 PRINT.indd 17

5.8361*** (0.000) −5.2030*** (0.000) −0.6585* (0.055) −0.8183*** (0.000) 1.6309*** (0.000) −0.3690* (0.071) −0.0020* (0.021) −88.3580*** (0.000) −75.5187*** (0.000) −15.7885*** (0.001) 31.8727*** (0.000) 4.8487*** (0.000) 4.3900*** (0.000) −0.2046 (0.334) −1.0777*** (0.000) −1.3367*** (0.001) −0.4618 (0.172) 0.5215*** (0.000) 0.2126 (0.133) 0.2496** (0.013) 109.9675*** (0.000)

Pooled Data (2) 9.3328*** (0.000) −6.0700*** (0.000) −0.4827 (0.191) −0.6550*** (0.000) 1.5385*** (0.000) −1.9428** (0.000) −0.0012 (0.189) −59.1630*** (0.000) −48.4481*** (0.000) −14.0234*** (0.004) 32.2175*** (0.000) 5.6028*** (0.000) 4.2944*** (0.000) 0.1454 (0.661) −0.9930*** (0.002) −1.7496*** (0.000) −0.4590 (0.190) 0.3891*** (0.000) 0.2654* (0.066) 0.1831* (0.077) 57.2348*** (0.002)

21/09/2022 13:14

18  Handbook of banking and finance in emerging markets Table 1.8  (continued) Coefficients

Number of Obs. F(20, 1016) R-squared Adj. R-squared

Pooled Data (1)

Pooled Data (2)

1037 79.06*** (0.000) 0.6044 0.5967

1037 69.99*** (0.000) 0.5749 0.5667

Notes:  p values are in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1% levels. Results of Pooled Data (1) are based on LEND, while those of Pooled Data (2) are based on RLEND.

rate is weaker in transition and emerging market economies. This may reflect that credit expansion is more demand driven, and based on high real returns, for these economies over the sample period. The coefficient of the spread (SPR) indicates that a higher degree of financial liberalization in banking sector (i.e., a lower spread) has a significant and positive effect on the BCPS ratio. Based on the sign of the coefficient on INFL, our results imply that inflation is associated with a drop in the BCPS ratio. The negative and statistically significant coefficient on GOVDEBT suggests that public credit plays an important role in determining private credit through a crowding-out effect. This is weaker to some extent in transition and emerging market economies, as suggested by the coefficients on the interaction terms. Finally, current lending to government and state-owned enterprises (CGSOE) appears to be positively associated with private credit provision in general, but less so in the emerging market and transition economies. Here, the coefficients on the interaction terms suggest that the impact of CGSOEs on private sector credit provision by the banking sector is limited in emerging market and transition economies. Table 1.9 presents the results of the random-effects panel data regression for the 54 sample economies. Again, as per the pooled data regression, most of the coefficients have the expected signs. A notable exception is the coefficient for INFL, which is positive and statistically insignificant, in contrast to expectations. The above noted, the results in Table 1.10 suggest that the random effects model is preferred to the pooled data regression. Therefore, this will be the preferred model in estimating ‘equilibrium’ growth in China’s credit-to-GDP ratio (BCPS). The results of the pooled data regression will be used as a robustness check. With respect to the distinction between advanced, emerging market and transition economies, the coefficients on the EMERG and TRANS dummies capture that financial depth is lower, on average, in economies within these sub-samples. Finally, the coefficient on BNKCRS suggests a positive association between the credit-to-GDP ratio and banking sector crises. Regarding the fit of the model, with adjusted R-squared of 0.6044 and 0.5749, respectively, the fit of our models compares favourably with previous attempts to estimate BCPS ratio equations (Boyd et al. 2001, Cottarelli et al. 2005, Khan et al. 2006, Levine et al. 2000).

NGUYEN 9781800880894 PRINT.indd 18

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­19 Table 1.9  Random effects (RE) panel data regression results Coefficients

PCGDP LEND/RLEND SPR GOVDEBT CGSOE INFL ER EMERG TRANS BBSD BNKCRS EMLEND/EMRLEND TRLEND/TRRLEND CRSLEND/CRSRLEND EMCGSOE TRCGSOE BBCGSOE EMDEBT TRDEBT BBDEBT CONSTANT Number of Obs. Number of Groups R-squ. Overall

NGUYEN 9781800880894 PRINT.indd 19

RE Model (1)

RE Model (2)

33.8684*** (0.000) −2.8184*** (0.000) −1.0463*** (0.000) −1.1991*** (0.000) 1.7425*** (0.000) 0.0648 (0.624) −0.0037*** (0.000) −67.7824*** (0.000) −74.1991*** (0.000) −42.8091*** (0.000) 35.0400*** (0.000) 4.0496*** (0.000) 3.3172*** (0.000) −0.8036*** (0.000) −1.4183*** (0.000) −1.0119*** (0.003) −1.0351*** (0.002) 0.6887*** (0.000) 0.8408*** (0.000) 0.5533*** (0.000) −161.3680*** (0.000)

41.6607*** (0.000) −2.2691*** (0.000) −0.8004*** (0.006) −0.9547*** (0.000) 1.4771*** (0.000) 0.1168 (0.484) −0.0037*** (0.000) −34.6596*** (0.000) −43.0477*** (0.000) −39.8214*** (0.000) 31.7576*** (0.000) 3.6608*** (0.000) 2.3795*** (0.000) −0.8014*** (0.000) −1.7213*** (0.000) −1.2449*** (0.001) −0.9300*** (0.007) 0.6848*** (0.000) 0.7499*** (0.000) 0.4694*** (0.000) −262.9802*** (0.000)

1037 54 0.4082

1037 54 0.3623

21/09/2022 13:14

20  Handbook of banking and finance in emerging markets Table 1.9  (continued) Coefficients

Wald chi2(20) sigma_u sigma_e rho

RE Model (1)

RE Model (2)

2018.82*** (0.000) 26.5042 15.4744 0.7458

1688.08*** (0.000) 26.4293 16.3228 0.7239

Notes:  p values are in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1% levels. Results of RE Model (1) are based on LEND, while those of RE Model (2) are based on RLEND.

Table 1.10 Breusch and Pagan Lagrangian multiplier test for random effects: BCPS[CNTRYID,t] = Xb+u[CNTRYID]+e[CNTRYID,t] Estimated results DCPSBNK e u Test: Var(u) = 0 chibar2(01)

Var 2133.7320 239.4571 702.4733

sd = sqrt(Var) 46.1923 15.4744 26.5042

3709.29*** (0.000)

5.2  Equilibrium Growth in China’s BCPS Ratio We now apply the estimated coefficients of the random effects panel data model to the 2001 to 2017 values for the fundamentals in China.14 Estimates of expected growth in the BCPS ratio can help draw out whether growth in China’s credit aggregates over recent decades is consistent with financial deepening based on fundamentals or should be explained as a bubble-like phenomenon. An issue worth mentioning before discussing the results is that the estimates assume that the sample economies and China are similar regarding the manner in which fundamentals determine financial deepening. The importance of including the transition economies in our sample economies is that, unlike many emerging market economies, banks in transition economies typically have continued to provide substantial credit to both state-owned and state-controlled companies. However, the process of privatization for many emerging markets is very fast, and banks may provide limited credit to local government- and state-owned enterprises. Credit provision to the private sector is thus the major reason for development in the emerging markets. In contrast, the level of the BCPS ratio in China is well above that of emerging countries, as is its level of growth.

NGUYEN 9781800880894 PRINT.indd 20

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­21 The results of the predictions of growth in the BCPS ratio are reported in Table 1.11. Based on the estimated coefficients of the preferred random effects model (Table 1.9, RE Model (1)), predicted growth rates in the BCPS ratio are positive over the 2001 to 2017 period, with an expected average of 3.17 per cent per annum. This compares to the actual changes in the BCPS ratio, which average 2.58 per cent per annum. With the exceptions of 2002 and 2003, actual changes in the actual BCPS ratio were negative in the early and mid2000s, largely due to rapid growth in China’s GDP and some attempts to constrain credit expansion to limit inflationary pressures. Major differences (exceeding 10 per cent) between the forecast and actual changes in the BCPS ratio are found in 2009 and 2015. The rapid expansion of credit in 2009 in response to the GFC (discussed above) saw actual changes in the BCPS ratio of over 22 per cent. The significant increase in the BCPS ratio in 2015, at over 12 per cent, resulted from a policy stimulus in response to slowing GDP growth. On balance, and contrary to common perceptions, growth in China’s BCPS ratios generally appear to be lower than those suggested by the fundamentals when advanced, emerging market and other transition economies are employed as the benchmarks. This suggests that the pace of China’s credit expansion has, generally, not exceeded the level justified by its fundamentals or financial development given its considerable progress in terms of GDP growth. Table 1.11 Predicted values for changes in China’s BCPS ratio (est. from panel data regression) Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Average

Actual change BCPS ratio (a) −1.14 7.46 8.19 −6.97 −6.74 −2.68 −3.36 −3.77 22.41 2.18 −3.48 5.83 5.40 5.92 12.35 3.64 −1.34 2.58

Predicted change value (b) 2.75 3.09 3.97 4.08 4.62 4.78 6.96 2.86 3.71 4.12 4.24 2.50 2.62 2.06 −0.68 0.64 1.55 3.17

Absolute deviation (a-b) −3.89 4.37 4.23 −11.05 −11.36 −7.46 −10.33 −6.64 18.70 −1.94 −7.72 3.33 2.77 3.86 13.03 2.99 −2.88 −0.59

Note:  Estimates limited to 2017 due to incomplete observations on all variables. Source:  IMF, International Financial Statistics; World Bank, Global Financial Development Database; and authors’ calculations.

NGUYEN 9781800880894 PRINT.indd 21

21/09/2022 13:14

22  Handbook of banking and finance in emerging markets 5.3 Robustness The estimates of the coefficients – and, thus, estimates of the expected levels of changes in BCPS ratios for China – are also quite robust to alternative specifications. We report here a few alternative modelling and data specifications (Tables 1.12, 1.9 and 1.10, and 1.13). The first robustness checks employ alternative econometric modelling of the pooled data. This is via use of quantile and Tobit regression methods. The quantile method also allows distinctions to be drawn between economies with less and more developed financial systems, as indicated by the BCPS ratio, in terms of their responses to fundamentals. Use of the Tobit method allows truncation of the dependent variable to be ­accommodated. The quantile regression results (Table 1.12) highlight an increase in both the statistical significance and impact on economic growth on financial development/deepening, consistent with Patrick (1966). Additionally, levels of market efficiency, as represented in the impact of the spread, are more significant as financial deepening increases. The coefficients resulting from the use of the Tobit estimation method (also Table 1.12) are consistent in both sign and size with those of Table 1.8. GDP per capita has the expected positive sign. With regard to nominal long-term interest rate (LEND), our results are robust because the coefficient estimates are negative and statistically significant, while the coefficients on the TRANS and EMERG interaction terms show that the impact of the interest rate is less pronounced in these groups of economies. As with the earlier pooled regression, the coefficients for SPR, INFL and GOVDEBT are negative and statistically significant. The latter supports the crowding-out/crowding-in hypothesis in these countries as an increase (decrease) in public sector debt is found to cause a decline (rise) in bank credit to the private sector. Most of the remainder of the robustness tests are applied to the random effects panel data model (Tables 1.9 and 1.13). First, the nominal lending rate is replaced with the real lending rate (RLEND) (Table 1.9), with all coefficients in the alternative model being consistent with those using the nominal rate. Following this, the dependent variable BCPS is replaced with domestic credit to the private sector (DCPS), being a ratio to GDP (Table 1.13). All coefficients are consistent in sign and general order of magnitude to those of the original random-effects specification. Finally, use of the log levels of bank credit to the private sector (BNKCR) and GDP produces results consistent with earlier models. This is apart from differences for the coefficient on INFL, which is now statistically significant and positive, and for the interaction terms on CGSOE, which become statistically insignificant (Table 1.13). Table 1.14 provides alternative estimates of the equilibrium rates of growth in the BCPS ratio, using the pooled regression results. Although different in order of magnitude, like those of Table 1.11, these results highlight that major evidence of excessive credit expansion is primarily associated with the 2009 stimulus to lending associated with the GFC, and major credit expansion in 2015. We thus conclude that our coefficient estimates are relatively robust, except for the impact of inflation on credit growth, which warrants further investigation.

NGUYEN 9781800880894 PRINT.indd 22

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­23 Table 1.12  Regression results for quantile models on pooled data Coefficients

PCGDP LEND SPR GOVDEBT CGSOE INFL ER EMERG TRANS BBSD BNKCRS EMLEND TRLEND CRSLEND EMCGSOE TRCGSOE BBCGSOE EMDEBT TRDEBT BBDEBT

NGUYEN 9781800880894 PRINT.indd 23

0.25 Quantile

Median (0.5)

0.75 Quantile

Tobit

2.6937 (0.215) −4.8791*** (0.000) 0.1711 (0.700) −0.7550*** (0.000) 2.1479*** (0.000) −0.1212 (0.647) 0.0000 (0.972) −83.2467*** (0.000) −82.4053*** (0.000) −3.7337 (0.546) 21.1144*** (0.000) 4.4687*** (0.000) 4.1811*** (0.000) −0.3334 (0.225) −1.1465** (0.004) −0.7218 (0.146) −0.9247** (0.035) 0.2725* (0.052) 0.2747 (0.134) 0.2701** (0.037)

10.1278*** (0.000) −4.7014*** (0.000) −0.4281 (0.251) −0.5873*** (0.000) 0.7069** (0.048) −0.2707 (0.222) −0.0014 (0.134) −79.7346*** (0.000) −85.0041*** (0.000) −11.7798** (0.023) 28.9280*** (0.000) 4.2884*** (0.000) 4.1507*** (0.000) −0.2874 (0.211) −0.0920 (0.780) −1.2398*** (0.003) 0.3255 (0.375) 0.3035*** (0.010) 0.4102*** (0.008) 0.0306 (0.778)

12.3902*** (0.000) −4.3133*** (0.000) −1.4224** (0.015) −0.8215*** (0.000) 1.0630** (0.035) −0.3075 (0.328) −0.0025* (0.081) −87.8214*** (0.000) −69.2997*** (0.000) −17.3253** (0.033) 51.6943*** (0.000) 4.0799*** (0.000) 4.3013*** (0.000) −0.5096 (0.157) −0.9633* (0.062) −1.8552*** (0.004) 0.3096 (0.496) 0.7732*** (0.000) −0.0286 (0.905) 0.1352 (0.426)

5.8531*** (0.000) −5.2061*** (0.000) −0.6730** (0.048) −0.8184*** (0.000) 1.6355*** (0.000) −0.3562* (0.079) −0.0020** (0.019) −88.2817*** (0.000) −74.7412*** (0.000) −15.8027*** (0.001) 31.7603*** (0.000) 4.8479*** (0.000) 4.3116*** (0.000) −0.1899 (0.366) −1.0724*** (0.000) −1.3750*** (0.000) −0.4665 (0.164) 0.5192*** (0.000) 0.2276 (0.105) 0.2503** (0.012)

21/09/2022 13:14

24  Handbook of banking and finance in emerging markets Table 1.12  (continued) Coefficients

CONSTANT Number of Obs. Pseudo R−squ. LR chi2(20)

0.25 Quantile

Median (0.5)

0.75 Quantile

Tobit

100.3126*** (0.000)

57.5621*** (0.005)

65.3703** (0.041)

109.8.63** (0.028)

1037 0.3463

1037 0.4017

1037 0.4255

1037 0.0885 981.20*** (0.000)

Note:  p values are in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels.

Table 1.13  Random effects (RE) panel data regression alternative dependent variables Coefficients DCPS GDP PCGDP LEND SPR GOVDEBT CGSOE INFL ER EMERG TRANS BBSD BNKCRS EMLEND TRLEND

NGUYEN 9781800880894 PRINT.indd 24

– 39.6928*** (0.000) −2.4534** (0.000) −0.7547** (0.014) −0.6150*** (0.000) 0.1995 (0.560) 0.1714 (0.223) −0.0044*** (0.000) −47.5253*** (0.000) −62.0871*** (0.000) −39.1267*** (0.000) 33.4683*** (0.000) 3.4286** (0.000) 2.8395*** (0.000)

BNKCR 1.0601*** (0.000) 0.4386*** (0.000) −0.0280*** (0.000) −0.0353** (0.030) −0.0119*** (0.000) 0.0196 (0.000) 0.0034* (0.065) −0.0001*** (0.000) −0.6234*** (0.000) −0.4069*** (0.010) −0.4193*** (0.000) 0.2466*** (0.000) 0.0468*** (0.000) 0.0134*** (0.001)

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­25 Table 1.13  (continued) Coefficients DCPS CRSLEND EMCGSOE TRCGSOE BBCGSOE EMDEBT TRDEBT BBDEBT CONSTANT Number of Obs. Number of Groups R−squ. Overall Wald chi2(20) sigma_u sigma_e rho

−0.6888*** (0.000) −0.9684*** (0.002) −0.7849** (0.029) 0.2571 (0.460) 0.3500*** (0.001) 0.4827*** (0.000) 0.2242** (0.014) −228.4065*** (0.000) 1037 54 0.4522 1812.08*** (0.000) 27.7544 16.5180 0.7384

BNKCR 0.0026 (0.168) −0.0016 (0.691) −1.0047*** (0.216) −0.0112** (0.016) 0.0004 (0.754) 0.0042*** (0.010) 0.0045** (0.016) −5.3759*** (0.000) 1037 54 0.9636 15550.36*** (0.000) 0.4230 0.2150 0.7947

Note:  p values are in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 % levels.

6. CONCLUSIONS In this chapter, we investigated whether bank credit growth in China can be viewed as ‘excessive’ and so potentially detrimental to its financial and macroeconomic stability. Since quantifying excessive credit development is relatively difficult, and determining whether ‘excessive’ credit growth can lead to financial crises is even harder, the literature on the topic is still relatively limited. In the case of transition economies, much of this literature bases this assessment around estimates derived from samples of advanced and emerging market economies. Thus, a contribution of our chapter is to employ a data set that includes evidence from the CEE economies, which are often used to guide evaluation of the China experience (e.g., Miurin and Sommariva 1993). This is to ensure that the impact of reductions in financial repression experienced in most transition economies is captured in the data set. After utilizing the HP-filter approach and finding insufficient evidence regarding the trend versus cycle/boom in China’s bank credit-to-GDP ratio, we employed both pooled

NGUYEN 9781800880894 PRINT.indd 25

21/09/2022 13:14

26  Handbook of banking and finance in emerging markets Table 1.14 Predicted values for changes in China’s BCPS ratio (est. from pooled data regression) Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Average

Actual change BCPS ratio (a)

Predicted change value (b)

Absolute deviation (a−b)

−1.14 7.46 8.19 −6.97 −6.74 −2.68 −3.36 −3.77 22.41 2.18 −3.48 5.83 5.40 5.92 12.35 3.64 −1.34 2.58

−0.58 1.04 −0.36 −0.45 1.71 0.57 −2.54 3.07 0.12 −0.94 −0.75 1.82 −0.59 −0.15 0.05 −3.58 −1.31 −0.17

−0.56 6.42 8.55 −6.52 −8.45 −3.25 −0.83 −6.84 22.29 3.12 −2.73 4.01 5.98 6.07 12.30 7.21 −0.02 2.75

Note:  Estimates limited to 2017 due to incomplete observations on all variables. Source:  World Bank, Global Financial Development Database and authors’ calculations.

data and random-effects regressions to further address the issues in our study. A variety of robustness tests were employed to assess, and confirm, the reliability of our estimates. We used the coefficient estimates from the random effects model to develop estimates of the year-to-year growth in China’s BCPS ratio based on changes in its economic fundamentals. Based on these in-sample projections of growth in the BCPS ratio, we find mixed evidence that the increases in the BCPS ratio observed in China up to 2017 are not consistent with ‘normal’ patterns of financial deepening, and conclude that, with the exception of the 2009 and 2015 credit expansions, these are generally justified by economic fundamentals. This confirms the earlier analysis conducted with the HP-filter approach. Some observers argue that China has experienced, and continues to experience, excessive credit growth. However, debate exists as to whether this will end up with a ‘soft landing’ or a ‘credit boom’ episode. Some believe that the Chinese economy will collapse  in the future, and highlight the real estate market and bank credit as major sources for overheating the economy. Of concern here is increased uncertainty in real estate markets and the credit risk to which mortgage and construction loans may expose  the banking system. Additionally, there are concerns regarding whether local government has sufficient funds, and about the banks’ abilities to administer and manage  risk  in line with the speed of credit growth. However, these are matters not

NGUYEN 9781800880894 PRINT.indd 26

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­27 about the rate, but more about the composition of observed credit growth and its potential impacts. Our conclusions are, broadly, that too much emphasis has frequently been placed on the rate of credit expansion in China, and too little on the level and pattern of economic growth associated with it. Use of the BCPS ratio as our dependent variable overcomes  this distortion. This is not to say that concerns about the pattern of China’s new lending are not warranted. To the extent that there has been heavy investment in fixed assets, including real estate, much of this credit expansion may be inefficiently ­allocated.

NOTES  1. In China’s case this is technically bank credit to the non-state-owned enterprise (SOE) sector, which includes state-controlled joint-stock listed companies.  2. Much of this literature has focussed on a subset of Central and Eastern European (CEE) transition ­countries.   3. An exception is Boissay et al. (2006), who use data from the Central and Eastern European countries and  apply both the one-country ECM model and panel techniques for in-sample estimation of key ­parameters.   4. In theory, shrinking assets, especially reducing highly risky loans, is a useful way to attain an ‘ideal’ capital adequacy ratio (CAR) as compared to raising expensive capital for commercial banks.   5. The original concept of a ‘Pacific Century’ appears to have emerged in the 15th and 16th centuries in association with the European spice trade. Towards the end of the 20th century (late 1980s) it was built around the emergence of a powerful, economically dynamic Japan. More recently the emphasis has shifted to a new Pacific Century, this time built around China as a dominant global economy. Thus, it suggests a ‘Chinese Century’, with China as the major economic power flanked by Malaysia, Vietnam, Indonesia, Thailand and an emerging India (Wilkins 2010).   6. As, for example, identified in the World Bank’s Global Financial Development Database (GFDD).   7. The ‘credit crunch’ phenomenon (Bernanke and Lown 1991) was identified in the US and some Asian countries during the 1990s’ recession. Interestingly, there is no such ‘credit crunch’ in China being triggered by a tightening of capital requirements, although the ‘credit boom’ identified in China’s credit market is quite similar to that experienced by some new EU members from CEE. Additionally, China’s commercial banks’ capital adequacy ratios have improved considerably since 2003. Statistical data released by the CBRC suggests that China’s banks’ average CAR was −2.98 per cent and only eight small-sized banks satisfied international capital requirements as of 2003. However, by the third quarter of 2012, the entire weighted-average CAR for China’s commercial banks had increased to 13.03 per cent and the core CAR was 10.58 per cent.   8. The CBRC was established in 2003, and its banking industry statistics published as of Q1 2004. The CBRC was superseded by the China Banking and Insurance Regulatory Commission (CBIRC) in April 2018.   9. In addition to rapid economic growth and gradual financial market deepening, we observe substantial credit expansion for China’s commercial banks co-existing with increases in the stringency of capital requirements. 10. Austria, Australia, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Japan, Ireland, Italy, the Netherlands, New Zealand, Norway, Portugal, Spain, Singapore, South Korea, Sweden, Switzerland, the UK and the USA. 11. Argentina, Bangladesh, Brazil, Chile, Colombia, India, Indonesia, Israel, Malaysia, Mexico, Peru, the Philippines, South Africa, Thailand, Turkey, Uruguay and Venezuela. 12. Bulgaria, Croatia, the Czech Republic, Estonia, Georgia, Hungary, Latvia, Lithuania, Poland, Romania, the Russian Federation, the Slovak Republic, Slovenia and Ukraine. 13. The results of a Hausman test for fixed versus random effects were not reliable. This is due to the model, as fitted to the data, failing to meet the asymptotic assumptions of the test. 14. The choice of 2017 is due to limitations imposed by the availability of IMF and World Bank data for some of the fundamental variables.

NGUYEN 9781800880894 PRINT.indd 27

21/09/2022 13:14

28  Handbook of banking and finance in emerging markets

REFERENCES Abiad, A. and Mody, A. (2005), ‘Financial reform: what shakes it? What shapes it?’, American Economic Review, vol. 95, pp. 66–88. Allen, F., Qian, J., Zhang, C. and Zhao, M. (2012), China’s financial system: opportunities and challenges, NBER Working Study No. 17828, National Bureau of Economic Research, Cambridge, MA. Apergis, N., Filippidis, I. and Economidou, C. (2007), ‘Financial deepening and economic growth linkages: a panel data analysis’, Review of World Economics, vol. 143, pp. 179–198. Balazs, E., Backe, P. and Zumer, T. (2006), Credit growth in Central and Eastern Europe: new (over)shooting  stars? European Central Bank Working Study Series No. 687, European Central Bank, Frankfurt. Beck, T. (2008), The econometrics of finance and growth, Policy Research Working Study No. 4608, World Bank, Washington, DC. Berger, A., Kyle, M. and Scalise, J. (2000), ‘Did US bank supervision get tougher during the credit crunch? Did they get easier during the banking boom? Did it matter to bank lending?’, in Mishkin, F. (Ed.), Prudential supervision: what works and what doesn’t, University of Chicago Press, Chicago, pp. 301–356. Bernanke, B. and Lown, C. (1991), ‘The credit crunch’, Brookings Studies on Economic Activity, vol. 2, pp. 205–247. Bernanke, B., Gertler, M. and Gilchrist, S. (1999), ‘The financial accelerator in a quantitative business cycle framework’, in Taylor, B. and Woodford, M. (Eds), Handbook of Macroeconomics, Elsevier, London, pp. 1341–1393. BIS (2010), Guidance for national authorities operating the countercyclical capital buffer, Bank for International Settlements, Basel. Boissay F., Calvo-Gonzalez, O. and Kozluk, T. (2006), ‘Is lending in Central and Eastern Europe developing too fast?’, in Liebscher, K., Christl, J., Mooslechner, P. and Ritzberger-Grunwald, D. (Eds.), Financial development, integration and stability: evidence from Central, Eastern and South-Eastern Europe, Edward Elgar Publishing, Cheltenham, UK and Northampton, MA, USA, pp. 229–254. Borio, C., Furfine, C. and Lowe, P. (2001), ‘Procyclicality of the financial system and financial stability: issues and policy options’, in Marrying the macro- and micro-prudential dimensions of financial stability, BIS Study No. 1, Bank for International Settlements, Basel, pp. 1–57. Boyd, J., Levine, R. and Smith, B. (2001), ‘The impact of inflation on financial sector performance’, Journal of Monetary Economics, vol. 47, pp. 221–248. Brzoza-Brzezina, M. (2005), Lending booms in the new EU member states: will euro adoption matter? European Central Bank Working Study Series No. 0543, European Central Bank, Frankfurt. Buncic, D. and Melecky, M. (2013), Equilibrium credit: the reference point for macroprudential supervisors, Policy Research Working Study, No. 6358, World Bank, Washington, DC. Calza, A., Gartner, C. and Sousa, J. (2003), ‘Modelling the demand for loans to the private sector in the euro area’, Applied Economics, vol. 35, pp. 107–117. Calza, A., Manrique, M. and Sousa, J. (2006), ‘Credit in the euro area: an empirical investigation using aggregate data’, Quarterly Review of Economics and Finance, vol. 46, pp. 211–226. Chen, S. and Kang, J. (2018), Credit booms: Is China different? IMF Working Paper WP/18/2, International Monetary Fund, Washington, DC. Christopoulos, D. and Tsionas, E. (2004), ‘Financial development and economic growth: evidence from panel unit root and cointegration tests’, Journal of Development Economics, vol. 73, pp. 55–74. Cottarelli, C., Dell’Ariccia, G. and Vladkova-Hollar, I. (2005), ‘Early birds, late risers, and sleeping beauties: bank credit growth to the private sector in Central and Eastern Europe and in the Balkans’, Journal of Banking and Finance, vol. 29, pp. 83–104. Coudert, V. and Pouvelle, C. (2010), ‘Assessing the sustainability of credit growth: the case of Central and Eastern European countries’, European Journal of Comparative Economics, vol. 7, pp. 87–120. Demirguc-Kunt, A. and Levine, R. (1999), Bank-based and market-based financial systems: cross-country comparisons, Policy Research Working Paper Series 2143, World Bank, Washington, DC. Drehmann, M. and Juselius, M. (2014), ‘Evaluating early warning indicators of banking crises: satisfying policy requirements’, International Journal of Forecasting, vol. 30, pp. 759–780. Drehmann, M. and Tsatsaronis, K. (2013), ‘The credit-to-GDP gap and countercyclical capital buffers: questions and answers’, BIS Quarterly Review, Mar., pp. 55–73. Giles, C. (2014), ‘China poised to pass US as world’s leading economic power this year’, Financial Times, 30  April, viewed 15 June 2021, . Gorton, G. and Winton, A. (2017), ‘Liquidity provision, bank capital, and the macroeconomy’, Journal of Money, Credit and Banking, vol. 49, pp. 5–37.

NGUYEN 9781800880894 PRINT.indd 28

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­29 Gourinchas, P., Valdes, R. and Landerretche, O. (2001), ‘Lending booms: Latin America and the world’, Economia, vol. 1, pp. 47–99. Graff, M. and Karmann, A. (2006), ‘What determines the finance-growth nexus? Empirical evidence for threshold models’, Journal of Economics, vol. 87, pp. 127–157. Gurley, J. and Shaw, E. (1967), ‘Financial structure and economic development’, Economic Development and Cultural Change, vol. 15, pp. 257–268. Hassan, K., Sanchez, B. and Yu, J. (2011), ‘Financial development and economic growth: new evidence from panel data’, Quarterly Review of Economics and Finance, vol. 51, pp. 88–104. Hofmann, B. (2001), The determinants of private sector credit in industrialised countries: Do property prices matter? BIS Working Study No. 108, Bank for International Settlements, Basel. Honohan, P. (1997), Banking system failures in developing and transition countries: diagnosis and prediction, BIS Working Study No. 39, Bank for International Settlements, Basel. Jung, W. (1986), ‘Financial development and economic growth: international evidence’, Economic Development and Cultural Change, vol. 34, pp. 336–346. Khan, M., Senhadji, A. and Smith, B. (2006), ‘Inflation and financial depth’, Macroeconomic Dynamics, vol. 10, pp. 165–182. King, R. and Levine, R. (1993a), ‘Finance and growth: Schumpeter might be right’, Quarterly Journal of Economics, vol. 108, pp. 717–738. King, R. and Levine, R. (1993b), ‘Finance, entrepreneurship and growth: theory and evidence’, Journal of Monetary Economics, vol. 32, pp. 513–542. Kiss, G., Nagy, M. and Vonnák, B. (2006), Credit growth in Central and Eastern Europe: convergence or boom? NMB Working Study 10/2006, Magyar Nemzeti Bank, Budapest. Kiyotaki, N. and Moore, J. (1997), ‘Credit cycles’, Journal of Political Economy, vol. 105, pp. 211–248. Kraft, E. and Jankov, L. (2005), ‘Does speed kill? Lending booms and their consequences in Croatia’, Journal of Banking and Finance, vol. 29, pp. 105–21. Kronmal, R. (1993), ‘Spurious correlation and the fallacy of the ratio standard revisited’, Journal of the Royal Statistical Society, vol. 156, pp. 379–392. Kwon, Y. and Zhou, W. (2009), ‘A study on the changes of China’s FDI policies after WTO: implications for Korean firms’, International Area Studies Review, vol. 12, pp. 121–138. Levine, R., Loayza, N. and Beck, T. (2000), ‘Financial intermediation and growth: causality and causes’, Journal of Monetary Economics, vol. 46, pp. 31–77. Liu, X. and Wray, R. (2010), ‘Excessive liquidity and bank lending in China: a modern money perspective’, International Journal of Political Economy, vol. 39, pp. 45–63. McKinnon, R. (1973), Money and capital in economic development, Brookings Institution, Washington, DC. Miurin, P. and Sommariva, A. (1993), ‘The financial reforms in Central and Eastern European countries and in China’, Journal of Banking and Finance, vol. 17, pp. 883–911. Naughton, B. (2011), ‘China’s economic policy today: the new state activism’, Eurasian Geography and Economics, vol. 52, pp. 313–329. Neusser, K. and Kugler, M. (1998), ‘Manufacturing growth and financial development: evidence from OECD countries’, Review of Economics and Statistics, vol. 80, pp. 638–646. Patrick, H. (1966), ‘Financial development and economic growth in underdeveloped countries’, Economic Development and Cultural Change, vol. 14, pp. 174–189. Powell, J., Shi, J., Smith, T. and Whaley, R. (2009), ‘Common divisors, payout persistence, and return predictability’, International Review of Finance, vol. 9, pp. 335–357. Schadler, S., Murgasova, Z. and Elkan, R. (2005), ‘Credit booms, demand booms, and euro adoption’, in Breuss, F. and Hochreiter, E. (Eds.), Challenges for central banks in an enlarged EMU, Springer, Vienna, pp. 187–221. Shan, Y. and McIver, R. (2011), ‘Corporate governance mechanisms and financial performance in China: panel data evidence on listed non-financial companies’, Asia Pacific Business Review, vol. 17, pp. 301–324. Wilkins, T. (2010), ‘The new ‘Pacific Century’ and the rise of China: an international relations perspective’, Australian Journal of International Affairs, vol. 64, pp. 381–405. Yang, K. (2007), ‘State-owned enterprise reform in post-Mao China’, International Journal of Public Administration, vol. 31, pp. 24–53. Zdzienicka, A. (2011), A re-assessment of credit development in European transition economies’, Economie Internationale, vol. 128, pp. 33–52. Zhu, X. (2012), ‘The global financial crisis: how China responded to it through legislation’, The Chinese Economy, vol. 45, no. 3, pp. 42–55.

NGUYEN 9781800880894 PRINT.indd 29

21/09/2022 13:14

30  Handbook of banking and finance in emerging markets

APPENDIX: DEFINITION OF VARIABLES   1. Bank credit to the private sector (BCPS): bank credit to the private sector as a share of GDP. Source: IMF, International Financial Statistics; World Bank, Global Financial Development Database.   2. GDP per capita (PCGDP): the natural log of GDP per capita as measured in US dollars at purchasing power parity (PPP) prices. Source: IMF, International Financial Statistics; World Bank, World Economic Outlook Database.  3. Public credit (CGSOE): bank credit to the government sector and state-owned enterprises as a share of GDP. Source: IMF, International Financial Statistics; World Bank, World Economic Outlook Database; World Bank, Global Financial Development Database.  4. Government debt (GOVDEBT): total government debt as a percentage of GDP.  Source: IMF, International Financial Statistics; World Bank, World Economic  Outlook Database; World Bank, Global Financial Development Database.   5. Lending interest rate (LEND): the long-term nominal lending interest rate. Source: IMF, International Financial Statistics.   6. Spread between lending and deposit rate (SPR): the difference between lending and deposit rates. Source: IMF, International Financial Statistics; World Bank, Global Financial Development Database.  7. Inflation rate (INFL): changes in the consumer price index (CPI). Source: International Financial Statistics.  8. Bank-based financial system (BBSD): derived from Demirguc-Kunt and Levine (1999), Table 12: Country Classification of Financial Structure. For those economies not included in this table, classification was based on the relative size of both bank credit and stock market turnover, and the ratios of bank assets and capitalization of the stock market to GDP.   9. Banking sector crisis (BNKCRS): derived from the World Bank, Global Financial Development Database. 10. Emerging market economy (EMERG): the economies of Argentina, Bangladesh, Brazil, Chile, Colombia, India, Indonesia, Israel, Malaysia, Mexico, Peru, the Philippines, South Africa, Thailand, Turkey, Uruguay and Venezuela. 11. Transition economy (TRANS): the former CEE economies of Bulgaria, Croatia, the Czech Republic, Estonia, Georgia, Hungary, Latvia, Lithuania, Poland, Romania, the Russian Federation, the Slovak Republic, Slovenia, Ukraine and China. 12. Real interest rate (RLEND): the long-term nominal lending interest rate adjusted for changes in the CPI. Source: IMF, International Financial Statistics; and authors’ calculations. 13. Domestic credit to the private sector (DCPS): total credit to the private sector as a share of GDP. Source: IMF, International Financial Statistics; World Bank, Global Financial Development Database. 14. Domestic credit from the financial sector (BCPS): total credit to both the private and public sectors as a share of GDP. Source: IMF, International Financial Statistics; World Bank, Global Financial Development Database.

NGUYEN 9781800880894 PRINT.indd 30

21/09/2022 13:14

Exploring patterns in China’s credit expansion  ­31 15. Bank credit (BNKCR): log of bank credit to the private sector measured in local currency terms. Source: IMF, International Financial Statistics; World Bank, Global Financial Development Database; and authors’ calculations. 16. GDP (GDP): log of GDP measured in local currency terms. Source: IMF, International Financial Statistics; and authors’ calculations.

NGUYEN 9781800880894 PRINT.indd 31

21/09/2022 13:14

2. Mutual fund investing in the Chinese A-share market Yeguang Chi and Xiao Qiao*

1. INTRODUCTION The economic value of mutual funds for investors has been of long-standing interest to financial researchers, going back at least to Jensen (1968), who uses the Capital Asset Pricing Model (CAPM) to evaluate whether mutual funds have positive risk-adjusted returns. Since Jensen’s seminal work, hundreds of papers have investigated the question of whether mutual fund managers can add value for investors by outperforming their benchmark returns, including Carhart (1997), Daniel et al. (1997), Pastor and Stambaugh (2002), Cohen et al. (2005), Fama and French (2010), and Pastor et al. (2015). Although many papers have been written on mutual fund performance, there remains debate among researchers on whether mutual funds can reliably outperform their benchmarks. In this chapter we investigate whether actively managed stock mutual funds provide attractive investment opportunities for retail investors in a relatively young stock market with a developing mutual fund industry. Established in 1991, the Chinese A-share market has grown to become the second largest stock market in the world. Given the size and growing importance of the A-share market, researchers have been increasingly interested in understanding the empirical patterns of asset prices and investor behavior in this market. We set out to document foundational facts about actively managed stock mutual funds in the A-share market. For our analysis, we focus on three aspects. First, we review the size and growth of the stock mutual fund industry in the A-share market, making comparisons with commensurate figures in the U.S. markets. Second, we examine the performance of stock mutual funds – average returns, cross-sectional dispersion, and performance persistence – using portfolio sorts and Fama and MacBeth (1973) regressions. Third, we evaluate the economic value of portfolios of stock mutual funds via performance attribution regressions. Our findings can be summarized as follows. From 2006 to 2019, stock mutual funds in the A-share market experienced explosive growth. The number of stock mutual funds increased steadily, rising 17-fold. Growth in assets under management (AUM) also showed an upward path, increasing six-fold. An equal-weight portfolio of all stock mutual funds outperforms the CSI300 Index, the mostly commonly used benchmark among funds. There is a large cross-sectional spread in average returns among funds – the topperforming funds outperform the average funds by more than 20% per year, implying large gains for investors who can reliably identify the top-performing funds. Successful fund selection is predicated on two assumptions: that some fund managers possess significant ability to generate outperformance, and that this ability persists. We find evidence that supports both assumptions. Economically significant average return 32

NGUYEN 9781800880894 PRINT.indd 32

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­33 differences are observed across funds, and funds that perform well in one period also tend to perform well in the following period. Quintiles formed on past performance or CAPM alphas show little turnover from month to month, and performance persistence can be observed even at the annual horizon. Since the two assumptions for successful fund selection are satisfied, investors presumably can benefit from selecting the best funds. We illustrate the feasibility of improving the risk-return tradeoff by forming portfolios of top-performing stock mutual funds, ranked by their CAPM alphas. Portfolios of top funds improve the risk-return tradeoff compared to a buy-and-hold strategy in the aggregate market, without increased volatility or tail risk. There is a tradeoff between diversification and outperformance: a portfolio of the top 5% of funds contains more idiosyncratic risk, but offers greater marketadjusted outperformance. Taken together, our results paint the following picture for stock mutual funds in the A-share market. In a young and developing stock market, stock mutual funds behave as “smart money,” outperforming the aggregate market return on average. Although many funds are able to beat market returns, fund managers are not created equal. The best fund managers possess sufficient skill to provide economically large risk-adjusted returns, and such skill persists over time. Investors who can successfully identify the top fund managers can reap great rewards, potentially outperforming the market by double-digit returns per annum. Even if an investor cannot identify the top funds, she would still likely be better off investing in several mutual funds than in the aggregate stock market. Our work fits into the literature on evaluating the performance of mutual funds. There is a large literature, mostly based on the U.S. markets, which supports the idea that mutual funds do not outperform the aggregate stock market (Jensen, 1968; Elton et al., 1993; Fama and French, 2010). Our finding that stock mutual funds as a whole outperform the aggregate market stands in sharp contrast to these studies, and may be related to the differential maturity of the U.S. and Chinese stock markets and the saturation of institutional investors. Chi (2016) finds that, in the A-share market, stock mutual funds can outperform the market. However, he only focuses on the equal-weight portfolio of all funds, and does not investigate whether subsets of funds can further improve performance. Our work is also related to the literature on the persistence of fund performance. One set of studies rejects the notion that mutual fund performance exhibits persistence. Carhart (1997) finds economically large differences in average returns in portfolios ranked by past returns, but this persistence does not extend to the following year. Bollen and Busse (2005) find short-term return persistence up to three months, but no return persistence beyond that. Some papers argue that, although funds on average generate negative abnormal returns, relative performance persists (Goetzmann and Ibbotson, 1994; Brown and Goetzmann, 1995; Grinblatt et al., 1995; Grinblatt and Keloharju, 2000). Compared to the first set of studies, we document longer return persistence in a developing market, and we illustrate the economic value of such persistence for investors. Compared to the second set of studies, we find that not only does relative performance persist, but also that stock mutual funds generate abnormal returns as a whole. The remainder of the chapter is organized as follows. Section 2 covers some background and describes the data. Section 3 explores the performance of stock mutual funds in the A-share market. Section 4 illustrates the economic value of portfolios of funds for investors, and Section 5 concludes.

NGUYEN 9781800880894 PRINT.indd 33

21/09/2022 13:14

34  Handbook of banking and finance in emerging markets

2.  BACKGROUND AND DATA 2.1 Background China’s stock market has gone through rapid development since its establishment in 1991. By August 2016, the Chinese A-share market had become the second largest stock market, with a total capitalization exceeding $5 trillion (Chen and Chi, 2018). As a relatively young stock market, trading in the A-share market is dominated by retail investors. According to official statistics from the Shanghai and Shenzhen stock exchanges, more than 80% of trading volume can be attributed to retail investors. In comparison, institutional investors dominate trading in the U.S. markets, and retail investors make up less than 20% of trading volume. The first mutual fund primarily investing in the A-share stock market, a closed-end fund, launched in 1998. Three years later, the first open-end stock mutual fund was launched. As the mutual fund industry in China grew, regulation compelled funds to increase the transparency of their operations. Since 2003, mutual funds have had to disclose their top 10 holdings on a quarterly schedule, and their entire portfolio holdings on a semi-annual schedule. Prior to August 2015, in order to be classified as an actively managed stock mutual fund, a fund had to invest at least 60% of its assets in the A-share market. However, in August 2015, this minimum threshold increased from 60% to 80%. As a result, many funds classified as stock mutual funds prior to August 2015 became a class of funds known as hybrid stock mutual funds, which hold between 60% and 80% of assets in the A-share market. Hybrid funds have more flexibility to invest in asset classes beyond stocks, such as fixed-income instruments or financial derivatives. In our study, we focus on actively managed stock mutual funds, hereinafter simply called “stock mutual funds.” 2.2 Data Data is collected from Wind Information, a leading Chinese financial data provider. Founded in 1994, Wind serves more than 90% of the domestic financial enterprises. Our sample consists of actively managed stock mutual funds from July 2006 to September 2020. We start our sample in 2006 because there were only a small number of funds prior to 2006. We start from the universe of all mutual funds in the Chinese markets, and construct our sample of stock mutual funds using the following three filters. First, we exclude passive index funds. Second, we examine the quarterly fund holdings data, and select funds that have at least 55% of assets invested in the A-share market in any given quarter and on average more than 60% of assets invested in stocks. Third, we exclude funds whose benchmark allocates more than 5% to a non-A-share stock index such as the Hang Seng Index. Since the first stock mutual fund was introduced in China in 1998, the mutual fund industry has experienced steady and robust growth. Figure 2.1 plots the number of actively managed stock mutual funds in the Chinese A-share market from 2006 to 2019. In 2006, there were only 79 stock mutual funds. This figure has grown steadily, reaching 1,368 by 2019. In a short span of just 13 years, the number of funds in the A-share market has multiplied 17-fold.

NGUYEN 9781800880894 PRINT.indd 34

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­35 1,600

1,400

1,200

Number of Funds

1,000

800

600

400

200

0

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

Figure 2.1  Number of stock mutual funds in the Chinese A-share market, 2006–2019 Table 2.1 reports summary statistics for the actively managed stock mutual funds. While the number of funds increased every year, the total assets under management of these funds experienced a less clear upward trend. In 2006, the total AUM of stock mutual funds was ¥229 billion,1 which ballooned to ¥1,424 billion in 2007 but subsequently fell to just under ¥800 billion in the early 2010s, before rising again to ¥1,370 billion in 2019. Overall, the AUM of stock mutual funds showed significant growth, increasing six-fold from 2006 to 2019, but the path was very volatile. This erratic growth illustrates a rather young industry which has not yet reached maturity. The period between 2006 and 2020 was marked by significant growth in total stock market capitalization in the Chinese A-share market, from ¥2.3 trillion to over ¥49 trillion. As a fraction of total stock market capitalization, the total stock mutual fund AUM peaked in 2007, at over 15%, before falling to a stable level of 2–3% in the late 2010s. Currently, stock mutual funds make up a relatively small percentage of the A-share market. Figure 2.1 and Table 2.1 provide an informative overview of stock mutual funds in the A-share market. The growth of this industry in China does not resemble that of U.S. funds from the same time period, but is reminiscent of an earlier period in U.S. fund history. Khorana and Servaes (2012) document that, in the infancy of the U.S. mutual fund industry, AUM grew six-fold over a nine-year period from 1976 to 1984, a similar rate to the observed pattern in the A-share market from 2006 to 2019. By the 2000s, the

NGUYEN 9781800880894 PRINT.indd 35

21/09/2022 13:14

36  Handbook of banking and finance in emerging markets Table 2.1  Summary statistics of actively managed stock mutual funds Year 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019

Number of Funds 79 105 132 176 226 289 346 375 430 624 742 935 1,168 1,368

Total AUM (¥billion)

Total Stock Market Capitalization (¥billion)

AUM/Mktcap

229 1,424 606 1,031 998 791 770 777 747 1,087 896 1,056 942 1,370

2,366 9,043 4,436 14,938 19,069 16,329 17,985 19,789 31,465 41,544 39,081 44,674 35,228 49,131

9.7% 15.7% 13.7% 6.9% 5.2% 4.8% 4.3% 3.9% 2.4% 2.6% 2.3% 2.4% 2.7% 2.8%

U.S. mutual fund industry had matured, and from 2000 to 2009 the growth rate in the number of funds or assets under management had slowed. In this period, the AUM of all mutual funds increased to 83%. Table 2.1 reports the year-end summary statistics of the actively managed stock mutual funds in the A-share market. The second column shows the number of stock mutual funds; the third the total AUM of the stock mutual funds; the fourth shows the total market capitalization of the Chinese A-share stock market; and the final column the total AUM of stock mutual funds as a fraction of the total stock market capitalization. The market share of actively managed stock mutual funds is significantly larger in the U.S. compared to the Chinese A-share market. Barber et al. (2016, p. 2600) assert that “most mutual fund investors allocate their savings to actively managed mutual funds,” a statement supported by data. According to Pastor et al. (2017), as of 2013 mutual funds in the U.S. had a total AUM of $15 trillion, about $7.5 trillion of which were focused on stock investments. Of these stock funds, 82% were actively managed, giving a total of over $6 trillion in actively managed stock mutual funds. In comparison, the U.S. stock market was $24 trillion in 2013, which means stock mutual funds made up 25% of the total market capitalization. A more recent figure can be derived from Ma et al. (2019), who state that the total AUM of mutual funds in the U.S. was over $16 trillion in 2016. If the proportion of actively managed stock mutual funds remained roughly constant between 2013 and 2016, there would have been $6.5 trillion in actively managed stock mutual funds in the U.S. in 2016, when the total stock market capitalization was $27 trillion. The total AUM for actively managed stock mutual funds in the U.S. markets was then 24% of the total stock market capitalization. The market share estimates from Pastor et al. (2017) and Ma et al. are 6–10 times larger than those observed in the A-share market.

NGUYEN 9781800880894 PRINT.indd 36

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­37

3.  PERFORMANCE OF STOCK MUTUAL FUNDS From the perspective of an investor in the Chinese A-share market who wants to delegate her investment decisions to mutual fund managers, there are two obvious questions to ask: 1) How do stock mutual funds perform in China? and 2) Can an investor do better than the average stock mutual fund by selecting a sub-sample of the top-performing funds? To answer these questions, we investigate the performance of Chinese stock mutual funds. We will first compare aggregate fund performance to the market benchmark, and then examine the cross-sectional variation in fund performance. 3.1  Stock Mutual Funds in Aggregate The CSI300 Index represents the top 300 stocks listed on the Shanghai Stock Exchange and the Shenzhen Stock Exchange. It is the most commonly used performance benchmark for mutual funds as well as exchange-traded funds. To evaluate the behavior of the aggregate set of stock mutual funds, we construct an equal-weight portfolio of all funds, rebalanced monthly. Figure 2.2 compares the cumulative returns of this equal-weight EW stock mutual funds v. CSI300

3 EW stock mutual funds CSI300

Cumulative returns (initialized at 1)

2.5

2

1.5

1

0.5

0 2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

Figure 2.2  Aggregate stock mutual fund performance, 2007–2020

NGUYEN 9781800880894 PRINT.indd 37

21/09/2022 13:14

38  Handbook of banking and finance in emerging markets portfolio with the returns of the CSI300 Index. From August 2007 to September 2020, the portfolio of all stock mutual funds significantly outperformed the CSI300 Index. The equal-weight portfolio of stock mutual funds grew from 1 to 2.71, for an annualized average return of 10.8%, whereas the CSI300 Index’s grew from 1 to 1.31 in the same period, for an annualized average return of 6.2%. The equal-weight portfolio of funds has lower volatility compared to the CSI300 Index: the annualized volatility of the equalweight portfolio of stock mutual funds is 25.2%, whereas that of the CSI300 Index is 28.7%. Despite the CSI300 Index’s 6.2% annualized average returns, its cumulative returns over 13 year are just 31%. The index experienced high volatility, which drove a large wedge between the arithmetic and geometric returns. Given its higher average returns and lower volatility, the equal-weight portfolio of all stock mutual funds provides investors with a more attractive risk-return tradeoff. Its Sharpe ratio is 0.35, more than twice as high as that of the CSI300 Index. The tail risk of the equal-weight portfolio of funds is also smaller than that of the CSI300 Index. The maximum drawdown of the portfolio of funds is 58.0%, compared to 70.5% for the CSI300 Index. Our finding that aggregate stock mutual funds outperform the market is consistent with the results in Chi (2016), who also finds outperformance using data up to December 2013. Aggregate fund performance differs in the A-share market and the U.S. markets. There is a large body of literature, based on the U.S. markets, that supports the idea that mutual funds do not outperform the aggregate stock market (inter alia, Jensen, 1969; Elton et al., 1993; Fama and French, 2010). A separate set of studies argue that, although mutual funds on average underperform the market, relative performance across funds may persist (Goetzmann and Ibbotson, 1994; Brown and Goetzmann, 1995; Grinblatt et al., 1995; Grinblatt and Keloharju, 2000). We turn our attention to the relative performance across stock mutual funds in the next section. 3.2  Cross-Section of Fund Returns A-share investors who invest in an equal-weight portfolio of stock mutual funds can outperform the aggregate stock market. Could investors do better by selecting funds that are likely to outperform in the future? Successful fund selection requires two necessary conditions. First, some fund managers must possess significant ability to generate outperformance. Second, this ability must be consistent, so that the selected top managers continue to produce outperformance. If these two conditions are satisfied, then, provided the investor has a method of identifying the top funds, her selected portfolio of funds will have a higher return compared to the equal-weight portfolio of all funds. There is an economically large spread in the average returns of stock mutual funds. The average return difference between the top 5% of funds and the average of all funds is a staggering 23.3% per year. The top 20% of funds outperform the average of all funds by 14.9% per year. The bottom 20% of funds, in turn, underperform the average of all funds by 14.4% per year. These differences could be due to random chance or fund manager skill – consistent outperformance by a select set of funds would favor a skill-based ­explanation or a luck-based one. To better understand the return distribution of stock mutual funds in the A-share market, we rank funds into quintiles by their returns over the previous 12 months. The

NGUYEN 9781800880894 PRINT.indd 38

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­39 highest quintile contains 20% of funds with the highest cumulative returns, whereas the lowest contains 20% of funds with the lowest cumulative returns. Quintiles are equalweight among their constituent funds, and reconstituted monthly. To understand the behavior of these quintiles, we track the portfolio characteristics before and after portfolio formation. We examine the average returns of the quintiles, and use the Capital Asset Pricing Model to separate the returns into a systematic component and an idiosyncratic component:

(2.1)

where Rj,t is the portfolio return of quintile j; Rft is the risk-free rate in month t, a monthly equivalent of the three-month fixed-term deposit rate; RmRft is the aggregate Chinese stock market return, proxied by the CSI300 Index, in month t in excess of the risk-free rate; and αj is the intercept, the part of average returns unexplained by the CAPM. The pre- and post-formation statistics are shown in Panel A of Table 2.2. By construction, average returns prior to portfolio formation increase monotonically from the lowest to the highest quintile, ranging from −0.05% to 2.39% per month. The difference in average returns between the highest and lowest quintiles is 2.44% per month, which is economically large and statistically significant at the 1% level. Because the ranking period is 12 months and the rebalance frequency is one month, two adjacent rebalances have an 11-month overlap in the portfolio signals, and portfolio statistics from two adjacent months are not independent. We account for potential serial correlation using Newey and West (1987) standard errors; 11 lags are used to compute the Newey–West errors to allow for serial correlation up to 11 months. Each month, we ranked the stock mutual funds by their past 12-month performance, and formed equal-size quintile portfolios. We then tracked the quintile portfolio’s performance in the following month. The three columns on the left in Table 2.2 report the preformation average monthly returns, market betas, and CAPM alphas, and the final three columns report the post-formation statistics. The bottom two rows show the statistics for a portfolio that takes a long position in the highest quintile and a short position in the lowest quintile. We do not observe much variation in the market betas of the portfolios, which lie in a tight range between 0.82 and 0.86. Quintiles with higher average returns in the pre-­ formation period tend to have somewhat lower market betas. Because there are large differences in average returns but limited variation in market betas, portfolio average returns are not explained by the CAPM. The unexplained portion of average returns – the CAPM alphas – follows a similar pattern to the average returns themselves. The difference in CAPM alphas between the highest and lowest quintiles is 2.19% per month, economically and statistically large. In the month after portfolio formation, average portfolios returns are still monotonic in the quintile ranks, although the return differences are not as large as those in the preformation period. The average return spread between the highest and lowest quintiles is 2.26%, somewhat smaller compared to the pre-formation value. The market betas for the quintiles range from 0.83 to 0.86, virtually unchanged from their pre-formation values. The spread in CAPM alphas is still economically large, 2.04% per month, albeit smaller compared to that of the pre-formation period.

NGUYEN 9781800880894 PRINT.indd 39

21/09/2022 13:14

40  Handbook of banking and finance in emerging markets Another way to measure fund performance is by their abnormal returns relative to the CAPM (Jensen, 1969; Carhart, 1997). We estimate CAPM alphas for each fund based on past 12-month returns and sort all funds into quintiles: the highest quintile contains 20% of funds with the largest CAPM alphas in the pre-formation period, whereas the lowest quintile contains 20% of funds with the smallest CAPM alphas. Each portfolio is reconstituted at the beginning of every month. The portfolio statistics are shown in Panel B of Table 2.2. By forming quintiles based on CAPM alphas, we are maximizing the spread in preformation CAPM alphas by design, rather than the spread in average returns. Nevertheless, in the pre-formation period, the average portfolio returns are monotonically increasing in the quintile ranks. The lowest quintile formed on CAPM alphas shows an average monthly return of 0.11%, followed by 0.76% for the next quintile, 1.13% for the middle quintile, 1.53% for the next, and 2.23% for the highest quintile. There is a 2.12% spread in average returns between the highest and lowest quintiles. We observe slightly more variation in the market betas for these quintiles compared to Panel A, ranging from 0.81 to 0.88; but the difference in market exposure between the highest and lowest quintiles is still economically small (0.81 − 0.88 = − 0.07) and statistically indistinguishable from zero.

Table 2.2  Performance of quintiles formed on past returns Panel A: Quintiles based on 12-month returns   1 (low) 2 3 4 5 (high) 5−1

Pre-formation

Post-formation

Avg Return

β

α

Avg Return

β

α

−0.05% 0.71% 1.14% 1.58% 2.39% 2.44%*** (11.8)

0.86 0.84 0.84 0.83 0.82 −0.04 (−1.07)

−0.60% 0.06% 0.45% 0.84% 1.59% 2.19%*** (9.7)

0.00% 0.70% 1.11% 1.52% 2.26% 2.26%*** (11.5)

0.86 0.85 0.84 0.83 0.83 −0.04 (−1.11)

−0.56% 0.08% 0.44% 0.80% 1.48% 2.04%*** (9.0)

Panel B: Quintiles based on 12−month CAPM alphas Pre-formation

Post-formation

 

Avg Return

β

α

Avg Return

β

α

1 (low) 2 3 4 5 (high) 5−1

0.11% 0.76% 1.13% 1.53% 2.23% 2.12%*** (9.4)

0.88 0.85 0.83 0.82 0.81 −0.07 (−1.55)

−0.82% −0.01% 0.44% 0.92% 1.80% 2.61%*** (10.7)

0.15% 0.75% 1.10% 1.47% 2.12% 1.97%*** (9.1)

0.88 0.85 0.83 0.82 0.82 −0.06 (−1.37)

−0.73% 0.00% 0.43% 0.87% 1.67% 2.40%*** (9.7)

Note:  Newey and West t-statistics are reported in parentheses, with the number of lags set to 11; *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels.

NGUYEN 9781800880894 PRINT.indd 40

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­41 The CAPM alphas show significant variation across the quintiles, from −0.82% for the lowest to 1.80% for the highest (a difference of 2.61%). In the post-formation period, the average return difference between the highest and lowest quintiles is 1.97%, statistically significant at a 1% level. There is still minimal variation in the market betas; the range is only 0.06. The differences in CAPM alphas remain economically and statistically large: 2.4% per month between the highest and lowest ranked quintile portfolios. Figure 2.3 shows a time series plot of the cumulative returns of the five quintiles formed on CAPM alphas. All portfolios are initialized at 1 in 2007. The five portfolios are highly correlated with one another, primarily due to their market exposure; hence they are also positively correlated with the CSI300 Index returns. We observe persistent performance commensurate with the quintile rankings: the top 20% of funds consistently outperform, following by the next 20%, and so on. Ordered by quintiles, the annual average returns of the quintiles are 14.9%, 12.7%, 10.4%, 9.2%, and 6.8%. The bottom 20% of funds have a cumulative return of 64%, compared to 341% for the top 20% of funds: 1 RMB invested in the bottom 20% of funds in 2007 would have turned into 1.64 RMB by September 2020, whereas 1 RMB invested in the top 20% of funds in 2007 would have turned into 4.41 RMB by 2020. Quintile portfolio performance sorted by CAPM alpha 5

4.5

Low 2 3 4 High

Cumulative returns (initialized at 1)

4

3.5

3

2.5

2

1.5

1

0.5

0 2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

Figure 2.3  Cumulative returns of fund quintiles

NGUYEN 9781800880894 PRINT.indd 41

21/09/2022 13:14

42  Handbook of banking and finance in emerging markets Figure 2.3 plots the portfolio value (initialized at 1) for five quintile portfolios, sorted by past 12-month CAPM alpha:

Rj,t − Rft = αj + bjRmRft + εj,t

where Rj,t is the portfolio return of quintile j; Rft is the risk-free rate in month t, a monthly equivalent of the three-month fixed-term deposit rate; and RmRft is the aggregate Chinese stock market return, proxied by the CSI300 Index, in month t in excess of the risk-free rate. Five portfolios are formed based on the intercept αj, rebalanced monthly. The sample period is August 2007 to September 2020. 3.3  Persistence in Performance Portfolio turnover gives a measure of the persistence of fund performance. A high portfolio turnover for the top quintile would imply that most top-performing funds in one month are not the top-performing funds in the following month, suggesting little to no performance persistence. Alternatively, a low portfolio turnover would imply that the topperforming funds in one month tend to also be the top-performing funds in the following month, indicating a high degree of persistence in fund performance. Fund performance exhibits persistence from month to month. The top quintile of  funds, based on past 12-month returns, has a monthly turnover of 20.8%. This figure implies that approximately 4 out of 5 top-performing funds in a given month will also be top-performing funds in the following month. We observe similar portfolio turnover for the bottom 20% of funds based on past returns; the monthly turnover is 20.6%. Portfolios based on the CAPM alphas exhibit similar levels of turnover. A portfolio of the top 20% of funds based on the CAPM alphas has a monthly turnover of 21.5%, which means 78.5% of funds tend to remain in the top-quintile portfolio at each monthly portfolio rebalance. A portfolio of the bottom 20% of funds based on CAPM alphas exhibits a monthly turnover of 21.2%. The evidence surrounding monthly portfolio turnover strongly suggests there is performance persistence for stock mutual funds at the monthly horizon. At each monthly portfolio rebalance, the portfolio formation signal is based on a trailing 12-month period. In two adjacent monthly rebalances, the two sets of signals have an 11-month overlap, which mechanically makes the two sets of signals highly correlated. The low turnover values may be an artifact of a higher rebalance frequency relative to the portfolio formation signal. To overcome this potential limitation to our analysis, suppose we rebalance the portfolios once a year, such that two adjacent rebalance periods do not share overlapping portfolio formation signals. The turnover figures of these annually rebalanced portfolios do not suffer from any potential downward bias due to overlapping signals. The turnover of the top 20% of funds, based on the previous year’s returns, is 74.6%, indicating that about a quarter of the top-performing funds in a given year continue to be the top-performing funds in the following year. The bottom 20% of funds also show some persistence: the portfolio turnover is 68.5%, which means that almost a third of the worst-performing funds tend to continue to perform relatively poorly in the following

NGUYEN 9781800880894 PRINT.indd 42

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­43 year. Annual portfolio rebalances based on CAPM alphas show similar levels of persistence. The top 20% of funds have a portfolio turnover of 73.8%, whereas the bottom 20% have a turnover of 75.8%. While portfolio turnover provides suggestive evidence of performance persistence, a more formal test is needed to more rigorously quantify this finding. We used the methodology of Fama and MacBeth (1973) to test performance persistence at an annual horizon. Every year, we computed the cumulative returns and the CAPM alphas for each stock mutual fund, and evaluated whether the top-performing funds in one year were also the top-performing funds in the following year. Because the returns and CAPM alphas in one year do not overlap with those from another year, there is no mechanical reason for these measures to persist. Each year, we ran cross-sectional regressions of fund returns and their lagged values:

ri,t = atr + btrri,t–1 + ε tr(2.2)

where ri,t is return of fund i in year t. If a fund that performs well in one year also tends to perform well in the following year, we would expect a positive regression coefficient btr. After running separate cross-sectional regressions for each year, we computed the point estimates and standard errors of the regression coefficients from their time series. We also investigated the persistence of the CAPM alphas of individual stock mutual funds:

αi,t = aαt + bαt αi,t−1 + ε αt 

(2.3)

where ai,t is CAPM alpha of fund i in year t. Similar to the interpretation of Equation (2.2), if the top-performing funds, measured by their CAPM alphas, tend to continue to outperform, we should expect to observe a positive estimate for bαt . Table 2.3 presents the Fama-MacBeth regression results. We observe economically large regression coefficients associated with past performance. From the perspective of a given fund, a 1% increase in fund returns relative to other funds is associated with a 0.18% increase in relative performance in the following year. Similarly, a 1% increase in  the CAPM alpha relative to other funds is associated with a 0.19% increase in the  CAPM alpha in the following year, compared to the other funds. The point estimates are not only economically large, but also statistically significant at the 1% level. The table reports the results of the Fama-MacBeth procedure applied to fund performance. Table 2.3  Performance persistence of funds   a   b   Avg R-Squared

NGUYEN 9781800880894 PRINT.indd 43

(1)

(2)

0.08 (1.3) 0.18*** (4.2) 3.52%

0.00 (1.5) 0.19*** (4.4) 4.76%

21/09/2022 13:14

44  Handbook of banking and finance in emerging markets We conduct cross-sectional regression for each year, following one of the following two setups:

ri,t = atr + btrri,t–1 + ε tr



αi,t = aαt + bαt αi,t–1 + ε αt

where ri,t is the total return of fund i during the 12-month period ending in month t, and αi,t is the CAPM alpha of fund i during the 12-month period ending in month t. Point estimates and standard errors of regression coefficients are computed from their time series. The first column in Table 2.3 shows the regression results for returns, and the second the results for CAPM alphas; *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels. In the context of existing literature, the persistence of fund returns in the A-share market is striking. Papers focused on the U.S. markets almost uniformly find no performance persistence at the annual frequency. Carhart (1997) finds economically large differences in average returns in portfolios ranked on past returns at a monthly rebalance frequency, but this persistence does not extend to the following year. Funds ranked by one-year returns do not show performance persistence, except for the extreme losers that continue to underperform. Bollen and Busse (2005) document short-term return persistence in mutual funds up to three months, but no return persistence beyond three months. Taking a different approach to measure the skill of mutual fund managers that does not focus on fund returns, Berk and van Binsbergen (2015) find that the average mutual fund adds values for investors for up to ten years. In the A-share market, fund returns tell a compelling story: top fund managers are able to add value for investors through persistent outperformance relative to other funds and benchmark returns.

4.  THE ECONOMIC VALUE OF STOCK MUTUAL FUNDS 4.1  Portfolios of Top-Performing Funds We have demonstrated that stock mutual fund returns exhibit significant cross-sectional variation that persists over time. In this section, we take the perspective of an investor and explore the benefit of constructing portfolios of the top-performing funds. We examine the return profiles of these portfolios to gain insight into the economic value of stock mutual funds for market participants. Barber et al. (2016) find that CAPM alphas are the best predictor of fund flows, suggesting investors strongly value market-adjusted fund returns. As such, we used CAPM alphas as the selection criteria for top-performing funds. At the end of each month, we calculated the CAPM alphas of each stock mutual fund in our sample, using past 12-month returns. Next, we sorted all available funds by their CAPM alphas and selected the top 5% of funds, allocating equal portfolio weights to these funds. At the beginning of our sample in 2007, when the top 5% included fewer than 10 selected funds, we selected the top 10 funds. Figure 2.4 compares the performance of this portfolio with the CSI300 Index from August 2007 to September 2020. In that period the cumulative return of the CSI300 Index

NGUYEN 9781800880894 PRINT.indd 44

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­45 top 5% of funds v. CSI300

7 top 5% of funds CSI300

6

Cumulative returns (initialized at 1)

5

4

3

2

1

0 2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

Figure 2.4  Cumulative returns of the top 5% of funds

was 31%. In comparison, a portfolio of the top 5% of all stock mutual funds showed a cumulative return of 466%, turning 1 RMB in 2007 into 5.66 RMB by September 2020. This figure plots the portfolio value (initialized at 1) for a portfolio of the top 5% of funds, measured by the CAPM alphas, rebalanced monthly. Robust outperformance of the CSI300 Index is not limited to the top 5% of stock mutual funds. Investors can achieve better diversification by considering a larger set of top-performing funds. As the number of funds increases, the portfolio becomes more diversified; but the ability of the portfolio to generate outperformance also decreases. Table 2.4 presents the performance statistics for these portfolios. The top 5% of funds have an average return of 17.2% per year from 2007 to 2020 – compared to 16.0% for the top 10%, 14.6% for the top 25%, and 13.1% for the top 50% of funds. A portfolio that simply invests in all available stock mutual funds every month has an average return of 10.8% per year. Clearly, portfolios made up of stock mutual funds easily outperform the CSI300 Index in our sample. Investors could do even better by selectively investing in the topperforming funds. Table 2.4 reports the performance of portfolios formed using stock mutual funds. Every month, all available stock mutual funds are ranked by their trailing 12-month CAPM

NGUYEN 9781800880894 PRINT.indd 45

21/09/2022 13:14

46  Handbook of banking and finance in emerging markets Table 2.4  Portfolios of stock mutual funds  

Annualized Returns Annualized Volatility Sharpe Ratio Maximum Drawdown

Top 5% Top 10% Top 25% Top 50% All Funds CSI300 Index

17.2% 16.0% 14.6% 13.1% 10.8% 6.2%

28.2% 27.3% 26.4% 25.8% 25.2% 28.7%

0.54 0.52 0.48 0.43 0.35 0.15

58.0% 58.0% 57.3% 57.9% 58.0% 70.5%

alphas, and a subset of the best-performing funds is selected. “Top x%” is a portfolio of the x% best-performing funds, and “All Funds” is an equal-weight portfolio of all stock mutual funds. Table 2.4 also presents additional performance statistics for portfolios of top-­ performing stock mutual funds. Among all portfolios, the equal-weight portfolio of all stock mutual funds has the lowest volatility, 25.2% per year. Portfolio volatility increases as the set of top-performing funds becomes smaller. A portfolio of the top 5% of funds, with a volatility of 28.2%, is the most volatile. However, it is still less volatile than the CSI300 Index, which has an annual volatility of 28.7%. Although more concentrated portfolios have higher volatility, they still offer a more attractive risk-return tradeoff for investors due to their high average returns. A portfolio  of the top 5% of funds has the highest Sharpe ratio, 0.54, of all the portfolios. The  Sharpe ratios decrease monotonically as the portfolios become more diversified. All the portfolios of funds have higher Sharpe ratios compared to those of the CSI300 Index. Less diversified portfolios could be more susceptible to idiosyncratic effects associated with individual funds. However, more concentrated portfolios of stock mutual funds do not exhibit greater tail risk. As we move to more concentrated positions from an equalweight portfolio of all funds – to the top 50%, 25%, 10%, and 5% of funds – the maximum drawdowns of these portfolios are remarkably similar, hovering around 58%. It appears that the more concentrated portfolios of stock mutual funds do not suffer larger occasional losses compared to the more diversified portfolios. All the portfolios experienced smaller drawdowns compared to the CSI300 Index, which has a maximum drawdown of 70.5%. 4.2  Performance Attribution Portfolios of stock mutual funds can generate economically meaningful outperformance relative to the CSI300 Index. How do these different portfolios behave? To answer this question, we employ a standard performance attribution setup:

Rmf,t − Rft = αmf + bmf RmRft + εt

(2.4)

In this regression, Rmf,t is the return of the portfolio of mutual funds in month t; Rft is the risk-free rate in month t, computed as a monthly value of the three-month fixed-term

NGUYEN 9781800880894 PRINT.indd 46

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­47 deposit rate; RmRft represents the aggregate Chinese stock market return in month t in excess of the risk-free rate; and αmf is the portion of average returns unexplained by the CAPM. Table 2.5 reports the performance evaluation regression results. An equal-weight portfolio that includes every available stock mutual fund has a market beta of 0.78, and a CAPM alpha of 5.5% per year. Market returns explain a significant fraction of the return variation of the aggregate set of funds; only 21.2% of the return variation remains unexplained. The aggregate set of stock mutual funds in the A-share market behaves differently compared to the aggregate set of funds in the U.S. markets. Fama and French (2010) demonstrate that, in the U.S. markets, stock mutual funds in aggregate have a market beta of 1.01, and a performing attribution regression using market returns results in an R-squared of 96%. Evidently, stock mutual funds as a whole highly resemble the U.S. markets, whereas in the Chinese markets aggregate stock mutual funds do not track the market as closely. An equal-weight portfolio of the top half of funds, rebalanced monthly, has a market beta of 0.77 and an unexplained average return of 7.9%. For investors who are able to Table 2.5  Performance evaluation of fund portfolios Top 5% Intercept

CSI300-Rf

R-squared

11.96% (2.5)

0.78 (16.4)

63.2%

Top 10% Intercept

CSI300-Rf

10.77% (2.4)

0.77 (17.1)

R-squared 65.3%

Top 25% Intercept

CSI300-Rf

9.38% (2.4)

0.77 (19.4)

R-squared 70.7%

Top 50% Intercept

CSI300-Rf

7.89% (2.2)

0.77 (21.0)

R-squared 73.9%

All Funds Intercept

CSI300-Rf

5.53% (1.7)

0.78 (24.1)

NGUYEN 9781800880894 PRINT.indd 47

R-squared 78.8%

21/09/2022 13:14

48  Handbook of banking and finance in emerging markets identify the best-performing 50% of stock mutual funds, their annual average returns can be improved by 2.4%. If an investor were to form an equal-weight portfolio of the 25% top funds, the market beta of such a portfolio would be 0.77, and the residual average return would be 9.4%. As we move to more selective portfolios, the regression intercept – unexplained average returns by the CAPM – rises to 11.9% for a portfolio of the top 5% of funds. Selecting the best-performing funds not only achieves highest average returns, but also the largest risk-adjusted outperformance (Table 2.4). Interestingly, the market exposure of more concentrated portfolios does not appear to increase; a portfolio of the top 5% of funds still has a market beta of 0.78, equivalent to that of the aggregate portfolio of stock mutual funds. Table 2.4 reports the performance evaluation regressions for portfolios formed on stock mutual funds, using the Capital Asset Pricing Model:

Rj,t − Rft = αj + bjRmRft + εj,t

where Rj,t is the portfolio return of quintile j; Rft is the risk-free rate in month t, a monthly equivalent of the three-month fixed-term deposit rate; and RmRft is the aggregate Chinese stock market return, proxied by the CSI300 Index, in month t in excess of the risk-free rate. “Top x%” is a portfolio of the x% best-performing funds and “All Funds” is an equalweight portfolio of all stock mutual funds. T-statistics are reported in parentheses; *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels. Perhaps not surprisingly, portfolios with more concentrated positions in the top-­ performing funds have more idiosyncratic risk. The CAPM can explain almost 80% of the return variation associated with the aggregate portfolio of stock mutual funds, but  only 63% of the return variation associated with a portfolio of the top 5% of funds.  In fact, the explanatory power of the CAPM steadily decreases as the portfolios  become more concentrated: 73.8% for the top 50% of funds, 70.5% for the top 25%,  and 65.2% for the top 10%. Although the systematic risks of these portfolios are  all  the same, they contain increasing idiosyncratic risks not captured by market returns. Stable market betas across portfolios formed on past performance is consistent with previous work on stock mutual funds. Chi (2016) also finds that the aggregate set of stock mutual funds in the A-share market has a beta statistically significantly less than one (between 0.71 and 0.78), depending on whether other factors such as value or momentum are included in the performance attribution regression. Carhart (1997) asserts that market betas are not able to capture differences in average returns of stock mutual funds; deciles formed on past 12-month returns all have market betas around one. The outperformance of Chinese A-share market stock mutual funds relative to the aggregate stock market contrasts the findings for the U.S. markets. Among others, Fama and French (2010) document that U.S. stock mutual funds, whether in aggregate or a selected set, have not been able to beat the U.S. stock market in the past few decades. Investors in the A-share market would have easily beaten the market by holding portfolios of stock mutual funds.

NGUYEN 9781800880894 PRINT.indd 48

21/09/2022 13:14

Mutual fund investing in the Chinese A-share market  ­49

5.  CONCLUDING REMARKS In this chapter, we studied the economic value of stock mutual funds for investors in the Chinese A-share market. From 2006 to 2019, stock mutual funds in that market experienced rapid growth. The number of funds multiplied 17-fold, while the assets under management grew six-fold. An equal-weight portfolio of all stock mutual funds outperforms the CSI300 Index, the most common benchmark for these funds. We investigated the average performance, cross-sectional dispersion, and performance persistence of Chinese stock mutual funds, and uncovered a marketplace well-suited to fund selection. Economically large average return differences are observed across funds, and funds that perform well in one period also tend to perform well in the subsequent periods. Our results suggest that actively managed stock mutual funds provide attractive opportunities for investors. We illustrate the economic value of stock mutual funds by forming portfolios of the top-performing funds based on their past risk-adjusted returns, and find that these portfolios improve the risk-return tradeoff compared to a buy-andhold strategy in the aggregate stock market. In this research, we have explored two measures of past performance: past returns and Capital Asset Pricing Model alphas. A natural extension of our work is to explore a larger suite of performance measures, including information ratio, Sortino ratio, and multifactor model alphas. Moreover, fund holdings could provide a different perspective on  skill and persistence. Performance measures derived from holdings data such as industry concentration (Kacperczyk et al., 2005), holding similarity (Cohen et al., 2005), and return gap (Kacperczyk et al., 2008) all provide interesting research directions. Lastly, one could run a horse race between different performance measures in predicting future fund returns, or blend several performance measures into an optimal portfolio of funds.

NOTES * The authors thank Sabri Boubaker, Duc Khuong Nguyen, and Carrie Wang for helpful comments. All errors are our own. 1. The USD($)/RMB(¥) exchange rate ranged from 6.05 to 7.72 in our sample period.

REFERENCES Barber, B. M., Huang, X., and Odean, T. (2016). Which factors matter to investors? Evidence from mutual fund flows. Review of Financial Studies, 29, 2600–2642. Berk, J. B. and van Binsbergen, J. H. (2015). Measuring skill in the mutual fund industry. Journal of Financial Economics, 118, 1–20. Bollen, N. P. and Busse, J. A. (2005). Short-term persistence in mutual fund performance. Review of Financial Studies, 18, 569–597. Brown, S. J. and Goetzmann, W. N. (1995). Performance persistence. Journal of Finance, 50, 679–698. Carhart, M. M. (1997). On persistence in mutual fund performance. Journal of Finance, 52, 57–82. Chen, Q. and Chi, Y. (2018). Smart Beta, Smart Money. Journal of Empirical Finance, 49, 19–38. Chi, Y. (2016) Private information in the Chinese stock market: Evidence from mutual funds and corporate insiders, SSRN working paper. Cohen, R., Coval, J., and Pastor, L. (2005). Judging fund managers by the company they keep. Journal of Finance, 60, 1057–1096.

NGUYEN 9781800880894 PRINT.indd 49

21/09/2022 13:14

50  Handbook of banking and finance in emerging markets Daniel, K., Grinblatt, M., Titman, S. and Wermers, R. (1997). Measuring mutual fund performance with characteristic-based benchmarks. Journal of Finance, 52, 1035–1058. Elton, E. J., Gruber, M. J., Das, S., and Hlavka, M. (1993). Efficiency with costly information: A reinterpretation of evidence from managed portfolios. Review of Financial Studies, 6, 1–22. Fama, E. F. and French, K. R. (2010). Luck versus skill in the cross-section of mutual fund returns. Journal of Finance, 65, 1915–1947. Fama, E. F. and MacBeth, J. D. (1973). Risk, return and equilibrium: Empirical tests. Journal of Political Economy, 81, 607–636. Goetzmann, W. N. and Ibbotson, R. G. (1994). Do winners repeat? Journal of Portfolio Management, 20, 9–18. Grinblatt, M. and Keloharju, M. (2000). The investment behavior and performance of various investor types: a study of Finland’s unique data set. Journal of Financial Economics, 55, 43–67. Grinblatt, M., Titman, S., and Wermers, R. (1995). Momentum investment strategies, portfolio performance, and herding: A study of mutual fund behavior. American Economic Review, 85, 1088–1105. Jensen, M. C. (1968). The performance of mutual funds in the period 1945–1964. Journal of Finance, 23, 389–416. Jensen, M. C. (1969). Risk, the pricing of capital assets, and the evaluation of investment portfolios. Journal of Business, 42, 167–247. Kacperczyk, M., Sialm C., and Zheng L. (2005). On the industry concentration of actively managed equity mutual funds. Journal of Finance, 60, 1983–2011. Kacperczyk, M., Sialm, C., and Zheng, L. (2008). Unobserved actions of mutual funds. Review of Financial Studies, 21, 2379–2416. Khorana, A. and Servaes, H. (2012). What drives market share in the mutual fund industry? Review of Finance, 16, 81–113. Ma, L., Tang, Y., and Gomez, J. (2019). Portfolio manager compensation in the US mutual fund industry. Journal of Finance, 74, 587–638. Newey, W. K. and West, K. D. (1987). A simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica, 55, 703–708. Pastor, L. and Stambaugh, R. F. (2002). Mutual fund performance and seemingly unrelated assets. Journal of Financial Economics 63, 315–349. Pastor, L., Stambaugh, R. F., and Taylor, L. A. (2015). Scale and skill in active management. Journal of Financial Economics, 116, 23–45. Pastor, L., Stambaugh, R. F., and Taylor, L. A. (2017). Do funds make more when they trade more? Journal of Finance, 72, 1483–1528.

NGUYEN 9781800880894 PRINT.indd 50

21/09/2022 13:14

3.  Liquidity and ex-dividend behavior in emerging markets Daniel Dupuis

1. INTRODUCTION There are challenges in all studies targeting emerging markets – the sample size is limited, the market is relatively unknown to the academic community and its particularities may not be duplicated in some developed countries – but it provides a unique, quasi-laboratory setting to conduct research in an environment still under economic development. We derive a new version of liquidity, and investigate the factors that may hinder the theorized stock price-drop surrounding cash dividend events and the associated abnormal returns. The capital markets of the United Arab Emirates offer a near-perfect setting to conduct this research. The computerized stock exchanges of the UAE operate under modern rules; there are no taxes on dividends or capital gains; and the investor base is relatively homogeneous, with retail traders dominating the field. This is in stark contrast with mature markets like the US, where institutions run the roost and the fiscal burden is prominent. In a perfect frictionless economy, the price of a stock should fall by the exact value of the dividend on the day shareholders can no longer claim a right to the said dividend. The rationale is quite simple: since the payout amount no longer belongs to the firm, its aggregate value must be subtracted from the company’s equity and the stock price adjusted downwards accordingly. It is well documented in the literature that the reality differs; there is “money left on the table” as prices drop less than the expected dividend. Many empirical and theoretical studies attempt to explain this phenomenon with mixed results. Earlier work (e.g. Elton and Gruber, 1970; Kalay, 1982, Booth and Johnston, 1984; Hietala, 1990; Michaely, 1991) hold the differential in tax preference accountable for the price inefficiency as some investors may show a predilection for dividends over capital gains, or vice versa (the clientele effect). Meanwhile other studies (e.g. Bali and Hite, 1998; Frank and Jagannathan, 1998; Jakob and Ma, 2004) focus on the presence of structural frictions in the equity markets; price discreteness, tick size, bid-ask bounce, limit order imbalance and transaction costs all belong to this category. Results vary across the spectrum of research, but one constant remains: in a taxable market it is problematic to separate the clientele effect from other frictions, and conclusions drawn within this framework may be tainted by the joint interaction of two or more variables. We first argue that liquidity is an issue in emerging markets and may play a role in the loss of ex-day price efficiency. We transform the turnover-adjusted illiquidity ratio to create a new measure – the free-float illiquidity – that is particularly well adapted to the trading restrictions generated by family holdings and governmental ownership present in many emerging nations. We offer further evidence that the generally accepted structural frictions and tax effects are absent in the UAE markets. Second, we exploit this 51

NGUYEN 9781800880894 PRINT.indd 51

21/09/2022 13:14

52  Handbook of banking and finance in emerging markets near-Walrasian economic state to show that, even in the absence of microstructure impediments, the ex-day price-drop remains significantly less than the dividend yield and unrealized abnormal returns persist. Consistent with the literature on other tax-free markets, we find a mean adjusted price-drop ratio of 3.21% for an average dividend yield of 5.22% and an ex-day abnormal return of 1.69%. The implication is that other undocumented factors could be responsible for the price premium. Finally, using regression  analysis and the redefined measure of fluidity, we show that illiquidity, dividend-yield  magnitude, traded volume and, to a lesser extent, brokerage fees play important roles in the determination of ex-day abnormal returns in a market free of the usual frictions. Our results indicate that a 1% increase in liquidity reduces abnormal returns by 2.25%. The remainder of this chapter is organized as follows. Section 2 details the evolution of liquidity measures and the introduction of free-float illiquidity, while Section 3 describes the UAE market and the intrinsic microstructural impediments to trade. Section 4 provides information on our sample and data; Section 5 presents our hypotheses, empirical methodology and results; and Section 6 concludes the chapter.

2.  LIQUIDITY IN EMERGING MARKETS Liquidity can be loosely defined as the possibility to buy and sell an asset with little or no impact on price, minimal delay and at a reasonable cost (Gabrielsen et al., 2011). Low liquidity translates into higher cost and a lower market price, thus commanding higher returns – a phenomenon also known as the liquidity premium. Although commonly examined in a single-country setting (i.e. the US), the positive liquidity–valuation association is also corroborated on an international sample of firms from 40 nations, particularly in the context of high country-level investor protection (Huang et al., 2020). Scholars draw on various techniques to assess liquidity: the taxonomy of measures ranges from volume-based methods (Martin, 1975) and price-variability indices to transaction-cost metrics (Goyenko et al., 2009) and trading frequency measures (Pu, 2009). Since this study seeks to contribute to the topic of liquidity computations in the unique context of emerging markets, we focus on volume-based metrics that relate primarily to the depth and breadth dimensions of liquidity. We argue that market fluidity is directly related to the pool of shares available to trade as opposed to simply outstanding. The introduction of restricted equity shrinks the pool of stock in the active markets, a common occurrence in emerging economies. The recent literature remains oblivious to the effect of trading limitations on the liquidity premium (Yu-Thompson et al., 2016). Tran et al. (2018), Lee and Chou (2018), Marshall et al. (2013) and Kang and Zhang (2014) all use versions of Amihud’s ratio to estimate liquidity in emerging markets, a notably constrained setting. The present contribution lies the development of an alternative metric – the free-float liquidity ratio, which is particularly suitable for markets characterized by large blockholders and institutional ownership. In some cases, trading restrictions originate from the specificities of the national governance system and are deeply embedded in legal frameworks of a country. This is particularly true for China (Chen et al., 2016; Li et al., 2019) and other transitional economies that are faced with similar contextual realities or exhibit unique institutional characteristics (Chen et al., 2017).

NGUYEN 9781800880894 PRINT.indd 52

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­53 2.1  Evolution of Volume-Based Liquidity For reasons of parsimony, we only provide a cursory view of the measures relevant to this study. The most basic measure, the raw trading volume, offers an intuitive, readily available metric but suffers from potential double counting and ignores the influence of price. The next metric, the conventional liquidity ratio, introduces price impact and is defined as follows:

(3.1)

where Voli,t is the trading volume for stock i during period t, Pi,t is the price and |Δ  Pi,t| is the absolute percentage change in price. LRi,t is generally computed on a monthly basis, providing a view of the price-impact of large trades. Amihud (2002) further proposes a measure of share illiquidity based on a combination of returns and dollar-based volumes:

(3.2)

where $Voli,t is the dollar-denominated trading volume and |Ri,t| is the absolute value of the period return for stock i during period t. The Amihud ratio is a mainstay of the literature on returns and liquidity (Marshall et al., 2013; Kang and Zhang, 2014; Fong et al., 2017) and is considered a “price pressure” indicator (Liang and Wei, 2012); but it is not without critics. Cochrane (2005) argues that this measure is biased for size, as shares with a lower dollar-denominated volume will display intrinsically lower liquidity (higher illiquidity) for an equivalent return. Florackis et al. (2011) further highlight this drawback, and propose that Amihud’s ratio must be scaled by share turnover to eliminate the size bias. Thus, the turnover illiquidity ratio is defined as:

(3.3)

where Ni,t is the number of outstanding shares for firm i during period t with all other variables as previously defined. Florackis et al. further argue that TIlliq captures two distinct dimensions of liquidity, namely frequency and transaction cost. Thus, the size bias critique associated with Amihud’s ratio can be negated by the introduction of the “outstanding shares” component in Florackis’ measure. However, the intrinsic assumption behind the adoption of the turnover illiquidity measure is that all outstanding shares are available to trade. Amihud et al. (2005) find that the existence of restricted stock increases illiquidity and the associated price discount. Narayan et al. (2015) also find a strong relationship between ownership structure (closely held shares) and firm liquidity. More recently, El-Nader (2018) contributed a paper describing the impact of unrestricted shares on liquidity, concluding that a higher free-float is significantly associated with increased market fluidity. Unfortunately, none of these studies provides an alternative measure that would incorporate free-float into a liquidity model. Considering the overwhelming evidence found in the literature, we strongly feel that liquidity ratios should incorporate a free-float component. We argue that the turnover illiquidity ratio suffers from measurement bias, particularly for corporate entities with

NGUYEN 9781800880894 PRINT.indd 53

21/09/2022 13:14

54  Handbook of banking and finance in emerging markets restricted ownership. The denominator component of TIlliq is calculated using outstanding stock as a proxy for the number of tradable shares; naturally, treasury stock is excluded but restricted shares remain part of the equation. We surmise that the turnover illiquidity ratio, as previously defined, fails to capture the full effect of the constraint in potential trading activity, thereby artificially overstating illiquidity (or understating liquidity). We modify the turnover illiquidity ratio where the number of outstanding shares is scaled by the public free float:

(3.4)

where FFi,t is the public free-float ratio for firm i at time t. The free-float illiquidity ratio (FIlliqi,t) is interpreted as the price impact of a 1% change in net (unrestricted) turnover rate. Illiquidity ratios are further criticized for a lack of intuitive meaning, as there exists a negative relationship between stock liquidity and the chosen metric. To alleviate this issue, Amihud et al. (1997) use the Amivest measure – essentially equivalent to the reverse of the Amihud ratio – thus restoring cognitive interpretation. Following the Amivest method, we reverse FIlliq and obtain free-float liquidity. Finally, Hasbrouck (2005) argues that outliers influence liquidity measures and offers two solutions: a log or a square-root transformation. Following their lead, we use the natural log transform variant to normalize free-float liquidity as follows:

(3.5)

where all variables are as previously defined. 2.2  Limitations and Dynamic Response of Free-Float Liquidity The two metrics FIlliq and TIlliq differ only by the free-float ratio. It could thus be argued that the usefulness of the two new measures is limited and turnover illiquidity is sufficient for most statistical analyses. This simplification is correct for markets where little or no restricted stock exists; but the purpose of this study is to focus on emerging economies where shares are held by a family of institutional entities and these blockholders hold fast to their stock. Florackis et al. note that the impulse response of the turnover illiquidity ratio may not be linear, but actually concave. This non-linearity is a positive development as it captures the cross-sectional variability in trading frequency and the subsequent impact on returns. To investigate the particularities of the free-float liquidity ratio, we construct a multivariate dynamic response system with three dimensions including a wide range of returns, the free-float adjusted turnover

and the resulting values of

FFL. Figure 3.1 provides a visual representation of the model. Dynamic modeling highlights the limitations and properties of FFL under a wide range of market conditions. The returns cross-section is near normal with a slight leptokurtic

NGUYEN 9781800880894 PRINT.indd 54

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­55

Dynamic response of free-float liquidity

4

2 1 0 -1

FREE-FLOAT LIQUIDITY

3

-2

0.35 0.33 0.31 0.29 0.27 0.25 0.23 0.21 0.19 0.17 0.15 0.13 0.11 0.09 0.07 0.05 0.03 0.01

0.37

0.39

0.41

0.43

0.45

0.47

0.2 0.49

0.14

0.17

0.08

0.11

0.02

0.05

-0.05

-0.02

-0.2 -0.17 -0.14 -0.11 -0.08

-3

Figure 3.1  Dynamic response of free-float liquidity tendency and thin tails that flatten under extreme values – a desirable setting for regression analysis where fat tails distort results and increase model variance. The dynamic response is symmetrical around the null return axis for all turnover values except for very low ratios, where the FFL response flattens and becomes negative. Furthermore, for a given static load on returns, the cross-section of free-float liquidity is a near-natural logarithmic curve, reflecting the innate mathematical structure of the model, and thus facilitating forecasts and projections. There are two areas of specific concern. First, there exists a singularity when trading occurs but the overall net period contemporaneous return is zero, regardless of transitional variations. Because of the model’s construct, this null return’s presence in the denominator renders the solution indeterminate and such observations must be struck  from the dataset. This particular setting is graphically represented by the cropped top in the distribution, along the y–z plane, where the model fails to deliver an estimate of liquidity. Should the missing data be included, the graph would show a crater or sink instead of a flat apex. Fortunately, FIlliq does not suffer from this deficiency and represents an alternative method for zero-return events. Other measures also exist for this somewhat rare setting, such as Kang and Zhang’s (2014) “Zeros” i­lliquidity. Second, free-float liquidity is designed to be more intuitive than illiquidity measures, but the output for FFL turns negative for very low adjusted-turnover ratios. In Figure 3.1, the pronounced dip at the origin of the x–y plane represents this specific characteristic. This is easily explained by the model’s construct where the natural log transformation

NGUYEN 9781800880894 PRINT.indd 55

21/09/2022 13:14

56  Handbook of banking and finance in emerging markets stabilizes the curve but causes fractional values to revert to their negative logarithmic form. While this is not a major drawback, it impedes instinctive interpretation but only occurs for very low stock volumes; thus, negative free-float values can be useful to isolate outliers in the dataset.

3. UAE MARKET CHARACTERISTICS AND MICROSTRUCTURE FRICTIONS The United Arab Emirates is a fast-growing Middle Eastern sovereign nation and a member of the Gulf Cooperation Council (GCC). While it is still considered an emerging market, the UAE has adopted most of the stringent market practices that characterize its Western counterparts, such as increased transparency, market surveillance, sophisticated computerized platforms, and so on. Part of its success is due to the willingness to rapidly embrace reforms and enact change. Originally an oil-based economy, the UAE has, in the past few decades, endeavored to diversify into tourism, heavy/light industry and real estate, with relative success. The country’s stock exchanges include the Dubai Financial Market (DFM), established in March 2000, the Abu Dhabi exchange (ADX, November 2000) and Nasdaq Dubai (September 2005). All of the nation’s exchanges support an order-driven computerized trading system under the supervision of the Securities and Commodities Authority. There are no taxes on dividends or capital gains for individuals and corporations located in special free zones, making for one of the simplest tax codes in the world. During the period under study (April 2007–May 2016) the UAE did not allow short sales, and no derivatives (options or futures) were offered to trade. The first equity futures contracts were introduced on Nasdaq Dubai in September 2016, and cover seven stocks but fall outside the date range of this chapter. The DFM supports the equity listing of 60 firms, the ADX 67 companies and the Nasdaq Dubai accounts for 10. The total market capitalization for our sample period ranges from a high of AED 507 billion (2007) to a low of AED 180 billion (2011). AED is the abbreviation for the dirham, the legal currency in the UAE with a fixed exchange rate of 3.67 AED/USD. Each dirham can be subdivided into 100 fils). According to Dubai Financial Market (DFM) Investor Relations, stock ownership in the UAE is characterized by a strong domestic bias; local entities (Emiratis) own more than 80% of the available shares, with GCC nationals (ex-UAE) and residents of other Arab countries far behind at 15%, leaving other foreign holdings 5% of the capital stock (https://www.dfm.ae/dfm-investor-relations/dfm-ir-overview). Institutional owners claim 76% of the market capitalization, but these investors have little incentive to trade, accounting for a mere 14.6% of the volume (2007) to a maximum of 28.4% (first quarter of 2016). This behavioral characteristic may go a long way in explaining the complete lack of ex-day volume for almost half of the dividend events identified in this study. Professionals have a taste for risk, while retail investors shun it (Florentsen and Rydqvist, 2002). Gonzalez et al. (2017) study the link between ownership concentration and dividend yield to show that large shareholders encourage higher yields. There is documented evidence that dividend yields are positively correlated with abnormal returns, so we expect that high institutional ownership, paired with low trading interest, can have an effect on liquidity

NGUYEN 9781800880894 PRINT.indd 56

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­57 and ex-day premiums. Furthermore, the mean free float for the DFM exchange oscillates around the 60% level, indicating that 40% of the shares are not available to trade. Since there are few corporate entities to act as institutional investors and no local mutual funds or exchange traded funds (ETFs) available for purchase, the government is a substantial stockholder and acts as a passive investor, providing a potential explanation for the low or nil volume for many stocks and the lack of interest in dividend-yield capture. Market-maker presence is minimal, with only seven stocks (incidentally, the most liquid) actively managed out of a total sample of 57. Al-Hilu et al. (2017) study the behavior of investors in the UAE market and find that Emirati traders depend on familiarity and personal information channels to make investment decisions while exhibiting overconfidence and home bias – all characteristics associated with noise traders. Overall, the UAE markets support the trading activity of a relatively homogeneous group composed mostly of local retail investors; and the peculiarity of this investor base may have a material impact on ex-dividend day price behavior. Market operations on the DFM, ADX and Nasdaq Dubai are processed using the standard order-book method, with investors placing orders through brokers who complete the trade on the exchange. The processing time for completing a trade is T + 2 (three days) and the record date usually precedes the ex-date by the same number of days, discounting weekend holidays that are held on Friday and Saturday in the GCC countries. Listed firms distribute cash or stock dividends on a yearly basis following the traditional pattern: the board of directors proposes, and final approval belongs to stockholders during the annual general meeting. Investors usually agree with the board’s recommendation, with the exception of the 2008 period when eight planned dividend payout events were voted down. This outcome is not surprising since the board meeting generally took place before the market crash and the general meeting in its aftermath. 3.1  Market frictions and structural impediments The literature on ex-day price behavior identifies a number of microstructure impediments; however, due to the particular nature of the Emirati market, some or all of these frictions may be absent. In this section, we review the documented structural hindrance factors and ascertain their pertinence to firms listed on the UAE stock exchanges. 3.1.1  Clientele effect A sizable portion of the literature on ex-day abnormal returns focuses on the difference in tax treatment between capital gains and dividends. Contingent on their taxable status, investors may show a preference between selling on the cum-day and collecting the dividend. Multiple studies conclude that the tax effect is difficult to separate from other structural frictions, casting a shadow on the significance of many ex-day price-drop puzzle solutions (e.g., Hess, 1982; Kalay, 1982; Bali and Hite, 1998; Boyd and Jagannathan, 1994). Fortunately, the UAE markets provide an advantageous research setting as there are no taxes on income, capital gains or dividends for individual investors. Corporations can (and do!) take advantage of special tax-free zones (such as the Dubai International Financial Center), thereby delineating the confounding clientele effect from other microstructure impediments. While a complete review of the literature on the tax-induced price premium effect is beyond the scope of this chapter, the complete lack of such tax friction

NGUYEN 9781800880894 PRINT.indd 57

21/09/2022 13:14

58  Handbook of banking and finance in emerging markets is sufficient to determine that the clientele effect is absent in the price behavior of stocks traded on the DFM, ADX and Nasdaq Dubai exchanges. 3.1.2  Price discreteness and tick size Tick size can create a potential positive bias in abnormal ex-day returns, and numerous studies document the phenomenon. As stock prices fluctuate exclusively by tick increments, if the dividend is a fraction of the minimum tick then the price drop on ex-day will be rounded off to the nearest integer and may not reflect the full amount of the dividend; while studying the clientele effect, Boyd and Jagannathan (1994) note that this is often the case for US markets. Payout frequency tends to exacerbate the problem as quarterly dividends are smaller than the yearly equivalent. Even if the dividend amount is a large, discrete multiple of the minimum tick size, Bali and Hite (1998) show that the return anomaly may persist if the ex-effect is below one tick as traders cannot benefit from the fractional increment and the stock price remains above its expected value. However, they point out that the clientele-induced tax effect may skew their findings. Graham et al. (2003) study the 2001 NYSE reduction in tick size, and conclude that price discretion is a dubious candidate in the search for a solution to the ex-day premium puzzle, while Kadapakkam and Martinez (2008) propose that price decimalization limits the impact of tick size. Similarly, Al-Yahyaee (2013) also argues that in a market characterized by large dividend/tick ratios, price discreteness sheds its significance in assessing ex-day returns. He shows that a market-regulated negative change in tick size results in lower ex-day abnormal returns. To illustrate, the pricing grid in the UAE moves in 0.1 fil increments (AED 0.001). The mean dividend in our sample is AED 0.22, or 220 times the tick size. With an average ex-day opening price of 6.04 AED, the tick/price ratio is 0.000166. For price discreteness to retain its significance the ex-effect would need to be of similar magnitude, but our results indicate a price-drop ratio of at least 200 times the tick/price value. Based on this summary analysis, we follow in the footsteps of Asimakopoulos et al. (2015) and Al-Yahyaee et al. (2008) and deduce that tick size is not a factor in explaining ex-day premiums in the UAE markets. 3.1.3  Bid-ask bounce Frank and Jagannathan (1998) introduce a microstructure model where noise traders are indifferent to the state of the dividend and trade for other unknown reasons. As this group is mostly composed of retail investors, noise traders use market orders which are filled at the “worst” market price – that is, at the ask price for a purchase and the bid price for a sale. Alternatively, market makers and institutional investors tend to display more patience and sophistication, thus using limit orders when trading. They would therefore take the opposite position to noise traders in the market order-book and benefit from a better order fill. As retail investors may find it cumbersome to collect and reinvest the dividends, they tend to trade on the close of the cum-day. The final outcome is that noise traders sell on the cum-day at the bid price and purchase at the ask price on ex-day, introducing a distortion equivalent to half of the spread at the closing of the cum-day and the same at the opening on ex-day, thus affecting abnormal returns. To solve this dilemma, Frank and Jagannathan propose the use of the mid-point between bid and ask prices as an anchor point. This model performs well in explaining ex-day premiums if

NGUYEN 9781800880894 PRINT.indd 58

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­59 the dividend is relatively close in size to the bid-ask spread as the gap between both trading prices creates a band of uncertainty. The mean bid-ask spread for our sample is 0.083 dirhams (8.3 fils) with an average dividend of 0.22 AED. These values indicate that the dividend is much larger than the spread and the impact of the bid-ask bounce is negligible. 3.1.4  Transaction costs and short-term trading In a framework of risk neutrality, the price drop ratio from cum- to ex-date should equal 1. Traders would naturally sell the stock on ex-day until the arbitrage opportunity disappears. Transaction costs inhibit this relationship by creating a drag on the seller’s profit, leaving a portion of the dividend as a component of the ex-day price. The concepts of trading fees and abnormal ex-day volume (a measure of trading activity) are intertwined and impact market prices differently; greater fees reduce the price-drop ratio, while higher trading levels increase the same through higher efficiency. Earlier studies (Lakonishok and Vermaelen, 1986; Kalay, 1982) demonstrate that abnormal ex-dividend price levels (akin to a lower price-drop ratio) are positively associated with dividend yields and transaction costs, while unusually high trading volumes relate to increasing dividend yields but lower trading fees. Since the studies take place in taxable markets, the results may be influenced by the presence of the clientele effect. Recent research, focusing on tax-free markets, dispels the existence of a short-term trading effect. Al-Yahyaee et al. (2008) show an actual reduction in abnormal volume around ex-day in the Omani market – very similar to the UAE market – and conclude that ex-day behavior is not affected by short-term trading. Dasilas (2009) and Asimakopoulos et al. (2015) disagree, as they find a dividend-induced volume effect in the Greek market. Perhaps the conflicting conclusions are the result of the difference in the array of statistical tools employed in the studies (portfolio sorting, event studies, regressions, etc.) or, as indirectly implied by Isaksson and Islam (2013), the outcome of market-specific characteristics. Regardless, short-term trading remains the domain of institutional investors and market makers, both of which are in short supply in the Emirati stock exchanges. I thus surmise that, for the UAE market, short-term trading and transaction costs may not constitute a satisfactory explanation of the ex-day premium price puzzle. Of all the price-drop impediments identified in the literature (tax differential between dividends and capital gains, price discreteness and tick size, bid-ask bounce, limit order imbalance and transaction costs), none appear to have a material impact on the ex-day price drop in the UAE exchanges. The rest of this chapter is devoted to an investigation into the potential persistence of abnormal returns despite the favorable particularities of the Emirati market.

4.  SAMPLE AND DATA This study targets the universe of UAE stocks paying cash dividends from April 2007 to May 2016 (ten years). All of the stocks traded on the three UAE capital markets (DFM, ADX and NASDAQ Dubai) are included in the initial sample. In accordance with the literature, the final list is trimmed to include stocks that meet the following conditions:

NGUYEN 9781800880894 PRINT.indd 59

21/09/2022 13:14

60  Handbook of banking and finance in emerging markets 1. The firm paid a regular cash dividend. 2. The data for price, volume, float, outstanding shares and so on is publicly available and continuous (no trading halts) for the window [−240, −41] days around the ex-day. 3. There are no splits, stock dividends or rights issue on the ex-day. 4. There are positive trading volumes on cum- and ex-dividend days. 5. And, finally, there is no material discrepancy between the dataset, the exchange’s and/ or the company’s records. Rule (4) offers the opportunity for an interesting reflection: it concerns data points where there is no volume, and hence no price change on ex-dividend day. A careless direct data analysis would yield a null price drop since the reported closing price on cum-day is equivalent to that of the open and close on ex-day, but no trading took place so there is  no actual price, return or volume available. In fact, the data are undefined for these  observations. We thus eliminate them from the final sample; but their existence provides an insight into the nature of the UAE markets with regard to liquidity. The shares are listed to trade, but no one seems interested in doing so. The zero-volume observations are not specific to ex-dividend days and represent a material 47.8% of all data points, and appear to indicate that there is a strong liquidity issue with the UAE markets. The dividend amounts, ex-date and type (cash, stock, special, etc.), opening and closing prices, historical daily public float, number of outstanding shares and daily volumes are obtained from Thomson Reuters firm Zawya. We manually verify all dividend events against exchange records, company announcements and Bloomberg data; in case of material disagreement (ex-date, dividend amount or type, etc.) the observation is struck from the dataset. The final sample includes 199 dividend events from 57 firms. Table 3.1 provides the descriptive statistics for the dataset. For shares listed on the three stock markets of the United Arab Emirates during the 2007–2016 period (199 observations), the mean (median) ex-day opening stock price is AED 6.04 (2.88), with values ranging from AED 0.40 to 78.00. The corresponding mean (median) dividend is AED 0.224 (0.136) for a dividend yield (D/Pcum,cl) of 5.22% (4.65%). The yield rate for the UAE is significantly higher than that of the stock exchanges of New York (0.69%), Tokyo (0.84%), Shanghai (1.62%) and London (1.21%) – as detailed in Isaksson and Islam (2013). The explanation lies in the frequency of the payouts; dividends are only paid once per year in the UAE, compared to quarterly in most other markets. The standard deviation of the dividend yield is 3.93% and the range varies from Table 3.1  Sample statistics Variable Dividend (AED) Ex-day opening price (AED) Dividend yield Free float % Outstanding(mm) Ex-day volume (mm)

NGUYEN 9781800880894 PRINT.indd 60

Range

Mean

Median

Std. Dev.

0.09–4.00 0.40–78 0.003–0.42 12–100 20–16600 0.0007–78.18

0.2204 6.0474 0.0522 62.85 3074 5.946

0.1363 2.8800 0.0465 62.31 1750 0.952

0.3316 11.371 0.0393 29.16 3046 11.391

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­61 0.30% to 42% for one extreme special event resulting from a corporate spin-off. These sample statistics are similar to those obtained from other studies focusing on the stock exchanges of tax-free developing countries. For example, Al-Yahyaee et al. (2008) study the Omani market and find a mean dividend yield of 7.35%, while Asimakopoulos et al. (2015) report 3.16% for the Greek market. As expected for a market with a high concentration of group-controlled shares and limited institutional ownership, the average (median) free-float is fairly low at 62.85% (62.31%) with a range of 12% to 100%. Finally, the mean (median) ex-day volume is relatively light at 5.9 (0.9) million shares, but the range is wide with only 7000 shares traded at the lowest end of the range and over 16 billion at the top of the bracket.

5. METHODOLOGY, HYPOTHESIS TESTING AND RESULTS 5.1  Price Premium Ratios Early studies on ex-day behavior include Elton and Gruber (1970), who develop an arbitrage model and show that, ceteris paribus, the price of a dividend paying stock should  decrease by the amount of the dividend (D) from the cum-date (Pcum) to the ex-date (Pex):

Pcum = Pex + D(3.6)

Equation (3.6) can be rearranged to obtain the raw price ratio (RPR), a measure of the price differential scaled by the dividend:

(3.7)

In the absence of taxes and other market frictions, the theoretical value of the raw pricedrop ratio is 1; the stock price will decrease by the exact value of the payout amount on ex-day opening. External market influences can play a role in the calculation of specific stock ex-day returns, and the choice of measurement variable can have a considerable impact. If closing prices for both the cum- and ex-day are used to compute the various ratios, then Pex partially reflects market movements during the ex-day and must be scaled accordingly; likewise if the opening price is used on cum-day. For the theoretical relationship to hold (RPR = 1), one should use the closing value on cum-day and the opening price on ex-dividend day to measure the raw price ratio. Following Graham et al. (2003) and Cloyd et al. (2006), equation (3.7) is rewritten as follows:

(3.8)

where D is the dividend paid, Pcum,cl is the closing price on cum-day and Pex,op is the opening price on ex-day. Furthermore, the change in price from cum-day to ex-day can also be influenced by overnight market movements. Should a material event occur after markets close, the RPR will be skewed by external forces during the pre-market orderaccumulation period on the ex-dividend day. All of the UAE exchanges (DFM, ADX and

NGUYEN 9781800880894 PRINT.indd 61

21/09/2022 13:14

62  Handbook of banking and finance in emerging markets NASDAQ Dubai) offer a pre- and post-market order accumulation mechanism where bids and asks are collected but no trades are settled. Order cancellation is allowed up to five minutes before opening and closing times. We rewrite equation (3.8) to reflect this precision and add a subscript to the variables: cl for closing price, op for opening and ov for overnight. The overnight market return (Rm,ov) represents the difference between the closing and opening values of the market index. We also test the two market-adjusted alternatives (open-open and close-close) to find that the results are not statistically different from the close-open adjusted for overnight market movements. For parsimony, these results are not tabulated but are available from the author upon request. The marketadjusted price ratio (MAPR) thus removes exogenous influences and can be computed as follows: (3.9)



where Rm,ov is the overnight market index return. As with the RPR, the theoretical value of MAPR is also equal to 1, but it offers a more accurate estimation. We thus formulate our first hypothesis as follows: H1: In a market free of microstructure frictions, the ex-day raw and market-adjusted price ratios are equal to 1. Table 3.2 describes the results for both price ratios for my sample of 199 observations. The significance levels reported are the result of testing the null hypothesis that RPR and MAPR are different from the theoretical value of 1. These tests are conducted using a two-sample t-test with unequal variances where the premium ratios are paired with a unit vector. In particular, we find that the mean (median) raw and market-adjusted premiums are almost quantitatively equivalent at 0.6581 (0.7500) for the RPR and 0.6554 (0.7142) for the MAPR. Both tests are significant at the 99% level with t-statistics of –5.10 and –5.11 respectively. We thus conclude that the ex-dividend day price premium in the UAE is less that the theorized value of 1. These results conform closely to the literature on tax-free markets as Al-Yahyaee et al. (2008) report a mean price ratio of 0.6460 for the Omani market, Asimakopoulos et al. (2015) show 0.6223 for the Greek market, and Frank and Jagannathan (1998) indicate a slightly lower value at 0.4324 for Hong Kong. Table 3.2  Summary results for H1, H2 and H3 Theoretical value

Mean

1.0000 1.0000 Div. yield Div. yield 0.0000 –

0.6581*** 0.6554*** 0.0322*** 0.0321*** 0.0169*** 0.0522

RPR MAPR RPDR MAPDR AR Div. yield

NGUYEN 9781800880894 PRINT.indd 62

Median

Std. dev.

t-statistic

0.7500 0.7142 0.0304 0.0304 0.0068 0.0465

0.9436 0.9484 0.0391 0.0323 0.0491 0.0393

−5.1034 −5.1179 −5.0890 −5.1172 4.8386 –

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­63 5.2  Price-Drop Ratios Price premium methods are intuitive and somewhat attractive, but many authors submit that the RPR and MAPR suffer from heteroskedasticity since the ratios are scaled by the payout amount, thus introducing excessive weight on low-dividend observations (see Eades et al., 1984; Lakonishok and Vermaelen, 1986; Barclay, 1987; Michaely, 1991; Boyd and Jagannathan, 1994; Bell and Jenkinson, 2002). They suggest that the raw price-drop ratio (RPDR) provides a better measure of ex-day behavior. The RPDR is similar to the price differential from equation (3.7) but, instead of the dividend D, the denominator is the cum-price Pcum,cl:

(3.10)

Adjusting for exogenous market price movements, we compute the market-adjusted pricedrop ratio (MAPDR) as follows, with all variables previously described:



(3.11)

In the absence of market frictions, the expected drop on the ex-day open is equivalent to the dividend amount; therefore, both the raw and market-adjusted price-drop ratios take on a theoretical value equal to the dividend yield (D/Pcum,cl). We thus formulate the second hypothesis as follows: H2: In a market free of microstructure impediments, the ex-day raw and adjusted price-drop ratios are equal to the dividend yield. Table 3.2 presents the results for the price-drop ratios. The significance levels reported are the result of testing the null hypothesis that the vectors RPDR and MAPDR are different from the theoretical dividend-yield target. These tests are conducted using a two-sample t-test with unequal variances where the price-drop ratios are paired with a vector composed of the corresponding dividend yields. There are 199 observations in the sample. From Table 3.2, rows 3 and 4, the mean (median) raw and adjusted price-drop ratios are practically identical at 3.22% (3.21%) for RPDR and 3.04% (3.04%) for MAPDR, indicating the absence of overnight market shocks (the basis for adjustments) between cum-days and ex-days. Both results are strongly significant (99% level) with t-statistics of −5.08 and −5.11. With this summary analysis, I reject the null that the price-drop ratios are equivalent to the dividend yield, and support the existing opinion in the literature stating that there is “money left on the table” on ex-dividend days. Our results show a slightly higher raw price-drop ratio than Asimakopoulos et al. (2.09% for the Greek market) and Frank and Jagannathan (0.17% for Hong Kong); but this explained by the fact that the dividend-yield for our sample is also higher at 5.22% compared to 3.16% and 2.51% respectively.

NGUYEN 9781800880894 PRINT.indd 63

21/09/2022 13:14

64  Handbook of banking and finance in emerging markets 5.3  Ex-Dividend Day Abnormal Returns While both the adjusted-price and drop ratios are informative in assessing the ex-dividend behavior of stock values, abnormal returns provide a direct and intuitive measure of the impact of the friction on ex-day prices. If the price decreases by the full dividend amount, theory dictates that the abnormal return should be zero. We compute the adjusted ex-day abnormal return AR as follows:

(3.12)

where E(Ri,t) is the expected return of the dividend-paying stock for the one-day period t corresponding to the ex-day and the other variables as previously defined. Following the literature (Graham et al., 2003; Liano et al., 2003; Cloyd et al., 2006; Al-Yahyaee et al., 2008), we select the CAPM as the market model:

(3.13)

E(Rm,t) is proxied by the daily returns of the respective market (DFM, ADX or NASDAQ Dubai) where stock i is traded during the ex-day t. Since the UAE does not issue treasury bills, the accepted practice is to use the overnight inter-bank repo rate as the risk-free rate, and we obtain this data from the Central Bank of the United Arab Emirates. The  annual Rf,t observation is scaled to match the daily frequency of the dataset. Following Dasilas (2009), we estimate the model parameters using daily returns within the [−240, −41] window relative to the event and formulate the third hypothesis as follows: H3: In a market free of microstructure impediments, the ex-day abnormal return is equal to zero. Table 3.2 row 5 reports the results for the abnormal returns. We are testing the null hypothesis that the ex-day AR is different from zero. This analysis is completed using a twosample t-test with unequal variances where the abnormal return is paired with a null vector. The sample contains 199 observations. The mean (median) ex-day abnormal return is 1.69% (0.68%) with a t-statistic of 4.83 indicating that AR is statistically greater (at the 99% level) than zero. This finding ranks in the middle of the range as reported in other studies targeting similar markets: 1.18% for Al-Yahyaee et al. (2008), 4.45% for Asimakopoulos et al. (2015) and 0.8% for Dasilas (2009) at 0.8%. Contrary to theory, it appears that abnormal ex-day returns in the United Arab Emirates are positive and significant despite the absence of most, if not all, price-drop inhibiting factors. We thus propose that abnormal returns are negatively related to the float-adjusted turnover illiquidity measure and formulate the fourth hypothesis: H4: The ex-dividend day abnormal return is negatively related to the illiquidity ratio of a stock as proxied by FIlliqi,t.

NGUYEN 9781800880894 PRINT.indd 64

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­65 The purpose of this test is to determine the influence of lower liquidity, as proxied by the previously described free-float illiquidity ratio (FIlliqi,t), on ex-day abnormal returns. Multiple studies on ex-dividend day price behavior document the various control variables considered necessary for this analysis. Karpoff and Walking (1988) document that stocks with a higher dividend yield attract more trading interest as the arbitrage benefits are greater. Trading cost also induces a drag on ex-day unrealized returns. Kalay (1982) argues that, although stock prices should drop by the dividend amount, price arbitrage by short-term investors is subdued by the fees they must disburse to complete the trade. Following in the footsteps of Bhardwaj and Brooks (1992), Dhaliwal and Li (2006) and Asimakopoulos et al. (2015), we use the inverse of the closing price on cum-day (1/Pcum,cl) as an independent variable to measure frictions due to the brokerage cost (TRcostit). Investors that trade on the ex-dividend day opening also face the additional risk that the expected drop in price remains unrealized. Heath and Jarrow (1988) postulate that this constraint may hinder arbitrage. Michaely and Vila (1996), Al-Yahyaee et al. (2008) and Asimakopoulos et al. (2015) measure this risk as the standard deviation of the residuals from their chosen market model scaled by the market risk. We thus compute σ εi,t for the CAPM model (equation 3.13) normalized by the standard deviation of the market returns σmi,t to obtain the control variable Riski,t = (σεi,t/σmi,t). The estimation of these parameters is completed using the same [–240, –41] window as with the CAPM model. Finally, Lakonishok and Vermaelen (1986), Michaely and Vila (1996), Al-Yahyaee et al. (2008) and Asimakopoulos et al. (2015) observe that trading volumes and/or turnover around the ex-day are evidence of short-term arbitrage and may impact unrealized abnormal returns, although their results differ. Al-Yahyaee et al. find a significant drop in volume on the ex-day, while Asimakopoulos et al. report positive abnormal volumes. To account for the potential (negative or positive) friction due to trading volume on the ex-day, we add the control variable Voli,t to equation (3.13). The variables Voli,t and FIlliqi,t display positive skewness of the residuals and are subject to outliers; we thus use the log transformation to normalize the distribution. The regression model is as follows:         



(3.14)

where: ARi,t Divyldi,t TRcostit Riski,t logVoli,t logFIlliqi,t

is the abnormal return as previously estimated. is the dividend yield for stock i on cum-day t. is the proxy for transaction cost estimated as1/Pcum,cl. is the standard deviation of the residuals from CAPM model (8) scaled by the market risk and estimated as σ εi,t/ σ mi. is the log transformation of the trading volume on ex-day t. is the proxy for market liquidity previously described.

Table 3.3 details the results of regression model (3.14) estimated using a pooled OLS model with White standard error estimators for my sample of 199 observations. Since very few events in the sample occur on the same date, clustering is not considered an important factor. We test H4, namely that the ex-day abnormal return is negatively related to the

NGUYEN 9781800880894 PRINT.indd 65

21/09/2022 13:14

66  Handbook of banking and finance in emerging markets Table 3.3  Cross-sectional regression results for H4 Variable

Coefficient

Constant

0.−1102*** 0.8772*** −0.0177** 0.0020 −0.0162*** −0.0099***

Divyldi,t TRcostit Riski,t logVoli,t logFIlliqi,t Adj. R−square F−statistic Prob(F−statistic) Durbin−Watson statistic

Std. Error

t-statistic

0.0264 0.0770 0.0085 0.0055 0.0036 0.0017

4.21 11.51 −2.14 0.37 −4.54 −5.62

p-value 0.00 0.00 0.03 0.71 0.00 0.00

0.4519 33.65 0.0000 2.00

illiquidity ratio of a stock. The null hypothesis is that the coefficients for the independent variables are not statistically different from zero. The adjusted coefficient of determination for the least squares regression is 0.45 and the Durbin-Watson statistic is 2.00, implying no autocorrelation. The F-statistic of 33.65 is significant (p-value = 0.0000). As expected, the coefficient of the dividend yield regressor β1 is positive (0.8772) and highly significant (p-value = 0.0000), confirming the previous results reported in the literature – namely that ex-day abnormal returns increase with Divyldi,t. A 1% increase in the yield is associated with a 0.87% rise in AR. As surmised, the transaction cost does not appear to hinder arbitrage activity that would reduce the price-drop ratio and increase abnormal returns. The coefficient β2 for TRcostit is negative (−0.0177) and signifies that a decrease in abnormal ex-day return is associated with higher fees, but the significance level is somewhat weak with a p-value of 0.03. Risk does not seem to be a material factor in the determination of ex-dividend day price behavior as β3 is not significant (p-value = 0.71). In accordance with previous research, volume appears to play an important role: at −0.0162, β4 displays a t-statistic of −4.54. A 1% rise in logVoli,t leads to an AR decrease of 1.62%. Higher ex-day volume is a gauge of arbitrage activity, increasing the price-drop ratio and thus depressing abnormal returns. At −0.0099, the estimate for β5, the coefficient of the proxy for illiquidity logFIlliqi,t, is highly significant with a p-value of 0.00. The economic interpretation is that, for a 1% increase in illiquidity, AR decreases by 2.25% (100.0099). This outcome conforms to the expectations posited by Florakis et al. (2011) but disagrees with the conclusion reached by Asimakopoulos et al. (2015) with regard to liquidity. Our results show that liquidity has a significant effect on ex-dividend day stock price behavior, and that abnormal returns are negatively related to the float-adjusted turnover illiquidity ratio FIlliqi,t.

6. CONCLUSION This chapter investigated the ex-dividend day price premiums and abnormal returns in a tax-free setting unencumbered by the structural impediments that are commonly blamed

NGUYEN 9781800880894 PRINT.indd 66

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­67 for dividend-related price inefficiencies – namely the United Arab Emirates. It contributes to the literature by providing a new insight into the potential causes of the persistence of unrealized abnormal returns. Liquidity appears to be an issue in the Emirati markets. Government entities (e.g. sovereign funds) and family units hold a sizable portion of the available equity, and these holdings are not officially classified as restricted. Nevertheless, the mean free-float is still relatively low at 62.85%. We thus investigated the impact of liquidity on the ex-dividend day price behavior and introduced a new modification to the illiquidity turnover ratio (as devised by Florakis et al., 2011). We derive the “float-adjusted turnover illiquidity ratio”, a metric that normalizes liquidity by the actual stock available to trade instead of the traditional number of outstanding shares. As far as we know, this is the first time that the public free float has been integrated in a derivative of the Amihud ratio to quantify liquidity. Using this new variable and regression analysis, we find that illiquidity has a statistically significant impact on ex-day price efficiency: an increase of 1% in liquidity is associated with a 2.25% reduction in abnormal returns. The coefficients for the dividend yield (positive) and trading volume (negative) are also significant at the 99% level, while that of the transaction cost (negative) is weaker with a p-value of 0.03. Risk does not appear to be an important factor in determination of the ex-day price premium. We thus conclude that liquidity (or lack thereof) plays an important role in the behavior of ex-day prices. We selected the equity market of the United Arab Emirates for this study primarily because of its fiscal environment; there are no taxes on dividends or capital gains, thus completely eliminating the clientele effect as one possible explanation for the less-thanexpected price drop on ex-day. The Emirati market’s composition is also relatively homogenous: the trading activity is dominated by retail investors, with institutions holding a sizable portion of the capital stock (76%) but exhibiting a low participation rate in the daily turnover (16–28%). Market-maker involvement is very limited: only seven stocks in our sample (out of a total of 57) fall under their influence, and the market rules severely restrict their ability to move prices. This characteristic suggests that the ex-day price behavior in the UAE market is not attributable to inefficiencies originating from the bid-ask bounce, short-term trading or transaction cost since all require the active involvement of professionals and institutional traders. Price discreteness can also be ruled out: the Emirati exchanges use a decimal quoting system with a tick size of AED 0.001, and the mean dividend/tick for my sample is large at 220. Both of these properties indicate a negligible tick-induced ex-effect. In the absence of the usual frictions, we verify whether the ex-day price premium persists. We show that, in accordance with the literature, the stock price does not drop by the full value of the dividend on the ex-day. For our sample, the mean market-adjusted price (price drop) ratio is equal to 0.6554 (3.22%) and is statistically different from 1 (dividend yield of 5.22%), the expected theoretical value. The average abnormal return is significant at the 99% level and stands at 1.69%, indicating that the ex-day price inefficiency endures even in a market with limited structural impediments.

NGUYEN 9781800880894 PRINT.indd 67

21/09/2022 13:14

68  Handbook of banking and finance in emerging markets

REFERENCES Al-Hilu, K., Sohel Azad, A.S.M., Chazi, A., and Khallaf, A. (2017). Investors’ behavior in an emerging, tax-free market. Emerging Markets Finance and Trade 53(7), 1573–1588. Al-Yahyaee, K. (2013). The effect of a reduction in price discreteness on ex-day stock returns in a unique environment. Journal of International Financial Markets, Institutions and Money 23, 283–294. Al-Yahyaee, K., Pham, T., and Walter, T. (2008). Ex-dividend day behavior in the absence of taxes and price discreteness. International Review of Finance 8, 103–123. Amihud, Y. (2002). Illiquidity and stock returns: Cross-section and time-series effects. Journal of Financial Markets 5, 31–56. Amihud, Y., Mendelson, H., and Lauterbach, B. (1997). Market microstructure and securities values: Evidence from the Tel Aviv Stock Exchange. Journal of Financial Economics 45(3), 365–390. Amihud, Y., Mendelson, H., and Pedersen, L.H. (2005). Liquidity and asset prices. Foundations and Trends in Finance 1(4), 269–364. Asimakopoulos, N.P., Tsangarakis N. V., and Tsiritakis, E.D. (2015). Price adjustment method and ex-dividend day returns in a different institutional setting. International Review of Financial Analysis 41, 1–12. Bali, R., and Hite, G.L. (1998). Ex-dividend day stock price behavior: Discreteness or tax induced clienteles? Journal of Financial Economics 47, 127–159. Barclay, M. J. (1987). Dividends, taxes, and common stock prices before the income tax. Journal of Financial Economics 13, 31–44. Bell, L., and Jenkinson, T. (2002). New evidence of the impact of dividend taxation and on the identity of the marginal investors. Journal of Finance 57, 1321–1346. Bhardwaj, R. K., and Brooks, L.D. (1992). The January anomaly: Effects of low share price, transaction costs, and bid–ask bias. Journal of Finance 47, 553–575. Booth, L.D. and Johnston, D.J. (1984). The ex-dividend day behavior of Canadian stock prices: Tax changes and clientele effect. Journal of Finance 39, 457–476. Boyd, J., and Jagannathan, R. (1994). Ex-dividend price behavior of common stocks. Review of Financial Studies 7, 711–741. Chen, J., Cumming, D., Hou, W., and Lee, E. (2016). CEO accountability for corporate fraud: Evidence from the split share structure reform in China. Journal of Business Ethics 138(4), 787–806. Chen, R.-R., Yang, T.-H., and Yeh, S.-K. (2017). The liquidity impact on firm values: The evidence of Taiwan’s banking industry. Journal of Banking and Finance 82, 191–202. Cloyd, B., Li, O., and Weaver, C. (2006). Ticks and tax: The effects of price discreteness and taxation on exdividend day returns. Journal of American Taxation Association 28, 23–46. Cochrane, J. (2005). Asset pricing program review. NBER Reporter 5, 1–12. Dasilas, A. (2009). The ex-dividend day stock price anomaly: Evidence from the Greek stock market. Financial Markets and Portfolio Management 23, 59–91. Dhaliwal, D., and Li, O.Z. (2006). Investor tax heterogeneity and ex-dividend day trading volume. Journal of Finance 61, 463–490. Eades, K.M., Hess, P.J., and Kim, E.H. (1984). On interpreting security returns during the ex-dividend period. Journal of Financial Economics 13, 3–34. El-Nader, G. (2018). Stock liquidity and free float: Evidence from the UK. Managerial Finance 44, 1227–1236. Elton, E., and Gruber, M. (1970). Marginal stockholders tax rates and the clientele effect. Review of Economics and Statistics 52, 68–74. Florackis, C., Gregoriou, A., and Kostakis, A. (2011). Trading frequency and asset pricing on the London Stock Exchange: Evidence for a new price impact ratio. Journal of Banking and Finance 35, 3335–3350. Florentsen, B., and Rydqvist K. (2002). Ex-day behavior when investors and professional traders assume reverse roles: The case of Danish lottery bonds. Journal of Financial Intermediation 11, 152–157. Fong, K.Y.L., Holden, C.W., and Trzcinka, C. (2017). What are the best liquidity proxies for global research? Review of Finance 21, 1355–1401. Frank, M., and Jagannathan, R. (1998). Why do stock prices drop by less than the value of the dividend? Evidence from a country without taxes. Journal of Financial Economics 47, 161–188. Gabrielsen, A., Marzo, M., and Zagaglia, P. (2011). Measuring market liquidity: An introductory survey. MPRA Paper 35829, University Library of Munich. Gonzalez, M., Molina, C.A., Pablo, E. and Rosso, J. W. (2017). The effect of ownership concentration and composition on dividends: Evidence from Latin America. Emerging Markets Review 30C, 1–18. Goyenko, R., Holden, C., and Trzcinka, C. (2009). Do liquidity measures measure liquidity? Journal of Financial Economics 92, 153–181. Graham, J.R., Michaely, R., and Roberts, M. (2003). Do price discreteness and transactions costs affect stock returns? Comparing ex-dividend pricing before and after decimalization. Journal of Finance 58, 2611–2635.

NGUYEN 9781800880894 PRINT.indd 68

21/09/2022 13:14

Liquidity and ex-dividend behavior in emerging markets  ­69 Hasbrouck, J. (2005). Trading costs and returns for US equities: The evidence from daily data. Working Paper, New York University. Heath, D.C., and Jarrow, R. (1988). Ex-dividend stock price behavior and arbitrage opportunities. Journal of Business 61, 95–108. Hess, P. (1982). The ex-dividend day behavior of stock returns: further evidence of tax effects. Journal of Finance 37, 445–470. Hietala, P. T. (1990). Equity markets and personal taxation: The ex-dividend behavior of Finnish stock prices. Journal of Banking and Finance 14, 327–350. Huang, T., Wu, F., Yu, J., and Zhang, B. (2020). Investor protection and the value impact of stock liquidity. Journal of International Business Studies 51, 72–94. Isaksson, A., and Islam, M. (2013). The ex-dividend day price behaviour of blue-chip stocks: international evidence. Asian Journal of Finance and Accounting 5(1), 79–88. Jakob, K., and Ma, T. (2004). Tick size, NYSE rule 118, and ex-dividend day stock price behavior. Journal of Financial Economics 72, 605–625. Kadapakkam, P. R., and Martinez, V. (2008). Ex-dividend returns: The Mexican puzzle. Journal of Banking and Finance 32, 2453–2461. Kalay, A. (1982). The ex-dividend day behavior of stock prices: A re-examination of the clientele effect. Journal of Finance 37, 1059–1070. Kang, W. and Zhang, H. (2014). Measuring liquidity in emerging markets. Pacific Basin Financial Journal 27(1), 49–71. Karpoff, J.M., and Walkling, R.A. (1988). Short-term trading around ex-dividend days: Additional evidence. Journal of Financial Economics 21, 291–298. Lakonishok, J., and Vermaelen, T. (1986). Tax induced trading around ex-dividend days. Journal of Financial Economics 16, 287–319. Lee, C.-H., and Chou, P.-I. (2018). Financial openness and market liquidity in emerging markets. Finance Research Letters 25, 124–130. Li, S., Zhang, T., and Li, Y. (2019). Flight-to-liquidity: Evidence from China’s stock market. Emerging Markets Review 38, 159–181. Liang, S., and Wei, K.C. (2012). Liquidity risk and expected returns around the world. Journal of Banking and Finance 36, 3274–3288. Liano, K., Hardin, W., and Huang, G. (2003) The ex-dividend day stock price behavior and decimalization. Working Paper, Mississippi State University. Martin, P. (1975). Analysis of the impact of competitive rates on the liquidity of NYSE stocks. Economic Staff Paper, Securities and Exchange Commission 75(3). Marshall, B., Nguyen, N., and Visaltanachoti, N. (2013). Liquidity measurement in frontier markets. Journal of International Financial Markets, Institutions and Money 27, 1–12. Michaely, R. (1991). Ex-dividend day stock price behavior: The case of the 1986 Tax Reform Act. Journal of Finance 46, 845–860. Michaely, R., and Vila, J. (1996). Trading volume with private valuation: Evidence from the ex-dividend day. Review of Financial Studies 9, 471–509. Narayan, P.K., Zhang, Z., and Zheng, X. (2015). Some hypotheses on commonality in liquidity: New evidence from the Chinese Stock Market. Emerging Markets Finance and Trade 51, 915–944. Pu, X. (2009). Liquidity commonality across the bond and CDS markets. Journal of Fixed Income 19(1), 26–39. Tran, L.T.H., Hoang T.T.P., and Tran, H.X. (2018). Stock liquidity and ownership structure during and after the 2008 global financial crisis: Empirical evidence from an emerging market. Emerging Markets Review 37, 114–133. Yu-Thompson, Y., Lu-Andrews, R., and Fu, L. (2016). Liquidity and corporate governance: Evidence from family firms. Review of Accounting and Finance 15(2), 144–173.

NGUYEN 9781800880894 PRINT.indd 69

21/09/2022 13:14

4. Asset-based valuation: a modified discounted cash flow approach

Rafael Yanushevsky, Daniel Yanushevsky and Camilla Yanushevsky

1. INTRODUCTION The term ‘value’ can be defined as measure of goodness or desirability (e.g., the importance, worth, or usefulness of something), depending on the area in which it is used. In sociology, values are a culture’s standard for discerning what is good and just in society. In economics, according to Adam Smith, the word value can be used in two senses, namely: value-in-use (the satisfaction that one obtains from the use of a commodity); and value-in-exchange (the amount of goods and services we may obtain in the market in exchange of a particular thing). Economic value is the value a person places on an economic good or service based on its provided benefit. In accounting terms, value is the monetary worth of an asset, business entity, goods sold, services rendered, or liability or obligation acquired. For making resource allocation decisions based on economic values it is necessary to measure the net economic benefit from assets. Economic value is subjective and entirely dependent on the subjective intention of an economic agent valuing an asset. Market value is the estimated worth of an asset, based on supply and demand, and how much someone is willing to pay for that asset. In its most basic form, asset-based value is equivalent to a firm’s book value or shareholders’ equity representing the firm’s net value, or the amount that would be returned to shareholders if all of the firm’s assets were liquidated and all its debts repaid, and is determined by subtracting liabilities from assets. Since an asset can be thought of as an object able to generate cash flow in the future, the value of any asset is determined as the sum of the present values of all future cash flows it is expected to provide over the relevant time period. Financial theory defines the intrinsic (belonging to something’s true or fundamental nature) value of an asset as the present value of all its future expected cash flows. This term is widely used as a measure of what an asset is worth based on fundamental and technical analysis rather than subjective judgments. Asset valuation is the process of determining the current market value or present value (also known as discounted value, which measures the current value of an amount of money or a stream of cash flows that is expected in the future) of assets using book values, absolute valuation models where the value of the asset is derived only on the basis of characteristics of that asset – such as discounted cash flow (DCF) analysis. A slightly different absolute valuation model is the discounted asset model. In this case, the  valuation is based on the separately derived present values of each asset that a firm  currently  owns. The sum of the values of all the assets represents a value for the entire firm. This method does not take into account the synergy between assets, and 70

NGUYEN 9781800880894 PRINT.indd 70

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­71 can only be used for ­commodity businesses which involve oil, coal or other such natural resources. In contrast to absolute valuation models, relative valuation models are not based on the intrinsic value of an asset. They use similar comparable assets to evaluate another asset by comparing their metrics (e.g., a price to earnings ratios, price to book value ratios, price to sales ratios, price to free cash flow, enterprise value, etc.). Emerging markets present a unique investment opportunity because of their potential for high growth. Although, when identifying emerging markets, investors and economists are looking for countries where there is very little political or social unrest and consistent economic growth, it is important to take into account their markets’ possible volatility and related economic risk. This could come from such factors as insufficient labor and raw materials, high inflation or deflation, unregulated markets and unsound monetary policies resulting in extremely volatile currency values. This, in turn, creates difficulties in applying the existing investment valuation models to emerging markets. The DCF investment valuation models contain parameters determined by using existing forecasting methods (see, e.g., Damodaran, 2012). Decision makers have a wide choice of ways to forecast, ranging from purely intuitive or judgmental approaches to complex quantitative methods. The accuracy of forecasts produced by financial analysts depends on available data, financial reporting quality, accounting standards and practices, and effective financial disclosure, among other factors. It is reasonable to assume that the quality of financial forecasts for emerging market countries may be lower than for developed countries. Usually, the further out the forecast, the higher the chance that the estimate will be inaccurate; therefore, it is important to develop investment valuation models operating only with variables characterizing a certain finite forecasting period. Taking into account the volatility of emerging markets, it is desirable to be able to make investment decisions based not only on financial analyst forecasts but also on the judgment of experts. It is impossible to expect absolute certainty in valuation, so there any valuation recommendations have a reasonable margin of error. It is obvious that valuation is neither a science nor an objective search for true value (Damodaran, 2012) based on the assumptions we make about the future of the company and the economy. Since valuation is an art rather than a science, it is useful to make it possible to also incorporate expert opinion in the investment valuation model. The developed model possesses the above-mentioned features. In contrast to existing DCF models considering entities infinitely, the proposed finite-time interval model includes a parameter which enables one to generate various scenarios for decision making. This approach presents a modification of the DCF method that makes it more practical and extends its sphere of efficiency, especially for business valuation. The rest of the chapter is organized as follows. Section 2 analyses existing discounted cash flow models, with special attention paid to numerous factors in these models that are difficult to predict. Specifics of firm valuation are discussed in Section 3. The modified intrinsic value model that focuses only on a certain period of a firm’s activity  without  assumptions about its behavior outside the considered time interval is considered in Section 4. The described valuation procedure is tested in Section 5 by considering 14 companies from various industries, and the results are summarized in Section 6.

NGUYEN 9781800880894 PRINT.indd 71

21/09/2022 13:14

72  Handbook of banking and finance in emerging markets

2.  ANALYSIS OF THE DCF MODELS Since in an efficient market at any point in time the actual price of a security will be a good estimate of its intrinsic value (Fama, 1965), it is natural that the discounted cash flow analysis and related models are applied first to stockholder equities. Intrinsic valuation was first tested on equity valuation models, including the dividend discount model (DDM) and the residual income model. The DDM is a method of valuing a firm’s stock price based on the assumption that its stock is worth the sum of all of its future dividend payments, discounted back to their present value. In other words, it is used to value stocks based on the net present value of the future dividends – that is, the intrinsic value of a share of stock equals

(4.1)

where Dt and r are the dividend at period t and cost of equity, respectively. For the oldest discounted cash flow model, see Gordon (1959) assuming the constant dividend growth rate g (so that Dt = (1 + g)Dt–1, t = 0,1,2, ...):

(4.2)

The two-stage DDM model (see, Malkiel, 1963) allows for two stages of growth: an initial phase with a constant rate g1, and a subsequent steady state g2 that is expected to remain so for the long term. The intrinsic value of a share of stock using this model can be estimated as follows: (4.3)



where n the period of short-term growth and Dn+1 = (1 + g1)n (1 + g2) D0. The H-model (Fuller and Hsia, 1984) was devised to approximate the value of a firm whose dividend growth rate is expected to change over time. Unlike the classical two-stage model (4.2), the H-model is also a two-stage model, but the growth rate g1 in the initial growth phase is not constant; it declines linearly over time towards the terminal growth rate g2. For this model:



(4.4)

where H is the half-life of the high growth period. The three-stage model, initially formulated by Molodovsky et al. (1965), incorporates elements of all three models: an initial period of very aggressive or insignificant growth followed by a period of incremental increase or decrease that eventually stabilizes at a more moderate growth rate that is assumed to continue for the life of the company; and discount factors are assumed different for the various phases. The analytical expressions of the above-mentioned models follow directly from (4.1).

NGUYEN 9781800880894 PRINT.indd 72

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­73 The indicated models consider only dividends to be cash flows to equity. Although dividends can easily be estimated based on empirical evidence, they usually do not affect the fundamental value of a company’s share price. The value delivered to the equity owners of a firm is the result of its management’s ability to increase sales, earnings, and free cash flow, which leads to an increase in dividends and capital gains (the profits from sales of assets) for the shareholders. Usually, the major portion of the profits is kept within the company as retained earnings, which represent money to be used for the firm’s ongoing and future business activities, and firms with a greater proportion of earned capital are more likely to be dividend payers. For firms that do not give out dividends or that follow unpredictable patterns of dividend payments the residual income valuation model (also called the abnormal earnings valuation model) is used as a viable alternative to the dividend discount model. Additionally, it works well with companies that do not generate positive cash flows yet (see, Hand et al., 2017; Jorgensen et al., 2011; Ohlson, 1995). The residual income valuation model determines a firm’s equity value as the sum of its book value and the present value of expected future residual income. The residual income valuation formula is very similar to a multistage dividend discount model, substituting future dividend payments for future residual earnings. By using this model the intrinsic value of a share of stock V0 equals

(4.5)

where BV0 is the current per-share book value of equity; RIt = Et − rBt–1 is the residual per-share income at period t; Et is the expected per-share earnings for period t; and Bt is the expected per-share book value of equity at any time. Free cash flow is the most important metric in finance. It serves as a more reliable measure of a firm’s performance than the parameters of above-mentioned models and as a measure of its profitability. It can be thought of as cash potentially available for things such as dividends, share repurchases, debt repayment, or reinvesting in the firm. Free cash flow, as a metric, provides a much deeper insight into the workings of a firm. That is why the present value of free cash flow to equity (FCFE) based on the expression similar to (4.1) is widely used to determine stockholders’ equity (the equity value of a firm). The FCFE model uses an expansive definition of cash flow to equity, in comparison with the dividend discount models, and has modifications similar to (4.2)–(4.4) (see Damodaran, 2012). However, according to Cruise (2012), the accuracy of such models is low. Partially, it can be explained by the dependence of the FCFE on many parameters: FCFE = Earnings before Interest and Taxes (EBIT) − Interest − Taxes     + Depreciation & Amortization − Change in Working Capital             − Capital Expenditure + Net Borrowing

(4.6)

This is why FCFE models ignoring the existence of at least a reasonable terminal time period cannot bring about satisfactory results. This statement is also valid for the DCF models determining the value of a firm as the present value of free cash flow to the firm (FCFF), which differs from FCFE. The differences between FCFF and FCFE arise

NGUYEN 9781800880894 PRINT.indd 73

21/09/2022 13:14

74  Handbook of banking and finance in emerging markets primarily from cash flows associated with debt – interest payments, principal repayments, new debt issues and other non-equity claims such as preferred dividends: FCFF = FCFE + Interest Expense × (1 − Tax Rate) + Principal Repayments           − New Debt Issues + Preferred Dividends (4.7) Free cash flow in the firm’s models is influenced by numerous factors that are difficult to predict. The modified intrinsic value model considered below focuses only on a certain period of a firm’s activity without assumptions about its behavior outside that time ­interval.

3.  THE MODIFIED DCF METHOD OF BUSINESS VALUATION Business has all types of goods and may be subject to sale. But this is a special kind of good, and its specifics determine the principles, approaches, and methods of its valuation. First, this is an investment product, and investment focuses on bringing profit in the future, the precise size of which is unknown. It is probabilistic in nature, so the investor must take into account the risk of failure. Second, the needs of business as a product depend on the processes that occur within the business itself as well as outside – in the economy. Another feature of business as a product – its quality – includes the quality of its management. Discounted cash flow is one of the most popular methods of business valuation on which all other valuation approaches are built. Being a dynamic object, during lifespan a firm generates free cash flow – the net output, presented in monetary terms, that determines the firm’s value. A similar approach based on accounting methodology is used in practice to determine, for example, what a cow is worth. In this case, economists are dealing with a much simpler dynamic object, with known average net output (milk, calves, etc.) and mean lifespan. The presence of many such objects and their predictable market demand simplifies the procedure of determining value, which is why the highest and lowest cattle market values do not differ significantly. Moreover, economists dealing with values of cattle usually prefer to use the term ‘worth,’ the measure of value reflecting demand – people’s preferences – rather than value. For dynamic objects the concept of value is clearly dynamic, and the cost of a cow depends on its age (among other factors). Valuation of firms also has its specifics. For public companies, the demand and supply theory approach can be used only to determine the market value of common stock, of preferred equity, of debt, and the so-called enterprise value (EV) showing how much it would cost to buy a firm’s business free of its debts and liabilities – that is, the market cost of the firm (EV is calculated by adding market capitalization and total debt, then subtracting all cash and cash equivalents). Usually, EV does not present a firm’s value, since markets are shortsighted – with a limited ability to predict the distant future (see also Damodaran, 2013). In reality, a bidding firm or individual and a target firm use fundamental analysis to determine a fair value to offer and a reasonable value to accept or reject the offer, respectively. Since it is impossible to predict a firm’s lifespan, T, scientists developing the firm value theory assume that firms are immortal – that is, T tends to infinity. Usually, a firm’s

NGUYEN 9781800880894 PRINT.indd 74

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­75 activity, as an infinitely lived entity, is considered for two periods: forecasting and postforecasting. For infinite T it is extremely difficult to make any reasonably justifiable assumptions and predictions for the post-forecasting period, and that is why terms such as ‘steady mode’ or ‘intrinsic value’ look too scholastic and do not reflect the reality. Forecasting a firm’s future behavior is more difficult than forecasting stock price. However, even for stocks there is no single forecasting methodology: so-called fundamental and technical analyses are used. Usually, decisions based on both analyses are more reliable. In essence, the discounted cash flow approach is based on the fundamental analysis of a firm’s activity. In its pure academic interpretation it ignores the market value of a business, the enterprise values EV, which can represent the intrinsic value only in an efficient market. A firm’s market value is the result of a bargain been struck. It is the monetary expression of agreement between a buyer and a seller. The buyer and seller prices are monetary expressions of their estimate of a firm’s value. Although it is defined independently by each participating party, market value appears as a result of price negotiation. The market value of a business changes over time under the influence of many factors, so that it presents the estimate of value, as the most probable price of the struck bargain, at a concrete time. The intrinsic value determined by the equation similar to (4.1)–(4.5) (see, e.g., Damodaran, 2012) is not consistent with the term ‘value’ that is an attribute of a good or service characterizing its importance or usefulness. More precisely, the former equation is a firm’s valuation based on fundamental analysis and presented in a monetary metric. Being subjective (representing an analyst’s opinion based on his/her assumptions concerning the firm’s future dynamics) it cannot be considered as a firm’s real worth. Since the market reflects future expectations of a firm’s efficiency, market-related information about a firm should be used its valuation procedure. In contrast to public companies, for privately held companies market information can be used indirectly – by choosing public companies with similar basic characteristics. More or less plausible predictions of the free cash flow can be made only for a limited period of time Tp, usually not exceeding 5–10 years. That is why it is logical to try to determine a firm’s value V(0) based on this limited information, including also an expectation of the firm’s value at time Tp in the form V(Tp) = αTV(0) where αT > 0. It would have been catch-22 if we had tried to determine V(0) based on V(Tp), since both these values are unknown. However, it is possible to consider several scenarios for various αT and choose αT that would correspond to the most plausible scenario. For example, one possible formulation of a firm’s value could be: ‘What is the worth of the firm that would keep its value in Tp years?’ (i.e., αT = 1or should reflect the predicted inflation rate). The important part of such an approach is how to correctly select αT. Damodaran (2012) offered some scenarios and reasons when it is difficult to use discounted cash flow valuation, including: firms in trouble having negative earnings and cash flows; and cyclical firms for which Tp can include the change of cycles. By decreasing Tp and properly evaluating αT it is possible to overcome such difficulties. The above-mentioned deficiencies of the intrinsic value approach raise the question of whether the attempts to obtain an intrinsic value are justified. We propose using the modified intrinsic value model to obtain some cost estimates that would create a certain negotiation set for both buyers and sellers.

NGUYEN 9781800880894 PRINT.indd 75

21/09/2022 13:14

76  Handbook of banking and finance in emerging markets In contrast to the existing valuation procedures, the proposed approach offers a valuation procedure that operates only with variables characterizing a certain finite forecasting period Tp. The cost obtained from any valuation model is affected by firm-specifics as well as market information. That is why market information is included in the proposed model. However, as the examples show, the costs obtained based on this information may differ significantly from the market values.

4. VALUATION BASED ON A FINITE PERIOD FORECASTING DATA The firm’s value is determined by the equation

(4.8)

where V(0) is the firm’s value; FCFFt is the free cash flow to the firm at period t; and kf is the discount coefficient. In reality, the discount coefficients kft change in time, so that in t the more precise model the term ∏t=1 (1 + kft) should replace (1 + kf)t. In contrast to the cost of equity in the equity modes considered above, here kf is the weighted average cost of capital (WACC) calculated by taking into account the cost of both debt and equity capital (see e.g., Damodaran, 2012). Although expression (4.8) with FCFE instead of FCFF can be used to calculate the value of a firm’s equity if kf is the cost of equity, many analysts prefer the FCFF model over the FCFE model. The equity value of a firm can easily be obtained by subtracting the market value of debt from the total firm’s value. FCFF growth reflects fundamentals more clearly than FCFE growth, and the FCFF model is widely used for leveraged firms with a changing capital structure and for valuing a leveraged firm with negative FCFE. The difficulty in using the equation (4.8) is that all parameters of its right-hand part are unknown and should be predicted properly to obtain a plausible result. For the infinite time horizon, it is impossible. The existing valuation procedures suffer from various drawbacks and contain too many assumptions. We present equation (4.8) in the following form:

(4.9)

where V( j) is the firm’s value in j years. Usually, the forecast of FCFFt is produced for about five years. This is the basic and most reliable information to determine a firm’s value. It is logical, instead of the unreliable assumptions over the infinite period of time (the so-called steady mode, etc.), to establish a short-term goal, based on which the firm’s value will be determined. Let us assume that, based on analysis of the future FCFFt (t =1, 2, … , j), investors expect the firm value to be

NGUYEN 9781800880894 PRINT.indd 76

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­77

(4.10)

where the coefficient αj is chosen based upon the financial analysis of a company or/and the available market information. Equation (4.10) can then be transferred to

(4.11)

so that If

(4.12) then



(4.13)

t For the model with variable kft in (4.12) and (4.13), ∏t=1 (1 + kft) should be used instead of (1 + kf)t. The above inequality shows that the assumed value growth for a positive cash flow is limited by kf . This looks like a significant drawback of the intrinsic value model (4.8) since this inequality shows the model as if it ignores a firm’s fundamentals since the discount coefficient, rather than its free cash flow, determines the firm’s future potential. It is known that for the so-called steady mode with the growth rate g for the convergence of (4.8) this growth should be limited – that is, g < kf (see, e.g., Cruise, 2012). This weakness of the model (4.8) significantly reduces its practical usefulness. As mentioned earlier, the modified model (4.9) is used to obtain a certain negotiation set useful for decision making to find a rational trade-off. The valuation problem presented above is based on forecasting a firm’s future activity for a finite period of time Tj. In addition to forecasting the free cash flow, the proposed approach requires evaluation of future change in the firm’s value αj. If the free cash flow is determined based on rigorous analysis of the firm’s fundamentals, it is logical to determine αj based on the market information about the considered firm or similar ones since it reflects market demand, the perception of the future. If αj is chosen as



(4.14)

where t is the time of valuation, then

 (4.15)

NGUYEN 9781800880894 PRINT.indd 77

21/09/2022 13:14

78  Handbook of banking and finance in emerging markets If the current EVt is known or can be determined, its future value is forecast based on certain assumptions. It is possible to use EV multiples. It is known that changes in earnings and revenues are the most important factors influencing the long-term price of a company’s stock. This data is also the heart of the fundamental analysis of a firm’s worth. That is why, depending on the type of firm, EV/EBIT and EV/Sales or their weighted combination are the most appropriate multiples to use. However, forecasting and using, for example, EVt+j = (EV/EBIT)t × EBITt+j, to determine αj = EVt+j /EVt is equivalent to αj = EBITt+j /EBITt – that is, the ratio of future and current earnings. When using EV/Sales multiples, αj = Revenuet+j /Revenuet, EBIT, an indicator of a company’s profitability, and revenue changes are the most important factors influencing a firm’s value. This data is an essential part of the fundamental analysis of a firm’s worth. However, the indicated αj are the result of using the market multiples. They can be chosen cautiously, taking into account the inequality (4.13). Experts can also recommend, based on various factors, reasonable αj that can be used to determine a set of reasonable prices for buyers and sellers. Important elements of the negotiation set correspond to αj = 1, the case when a firm keeps its monetary value at the j-th year (the obtained amount can be increased taking inflation into account). For stable firms at a mature stage with expected growth FCFFt = FCFF0(1 + g)t we have (4.16) 



For a firm with a stable unchanged free cash flow FCFF0 (i.e., g = 0) we have

 (4.17)

For g = kf:

(4.18)

A firm’s valuation based on scenarios (4.16)–(4.18), corresponding to cases when it keeps its value in the j-th year – together with the available market information and related estimates (4.14) and (4.15) – should be used for an considered decision making. In addition, the proposed approach can be used to enhance asset management and choose realistic predictions in asset allocation models (see e.g., Yanushevsky and Yanushevsky, 2015).

NGUYEN 9781800880894 PRINT.indd 78

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­79

5. EXAMPLE The valuation procedure described above is tested by considering 14 companies from different industries (see first column of Tables 4.1–4.4). Using relevant information from Bloomberg’s website we estimate the intrinsic values of the chosen firms for various αj and compare them with their EV. The considered time interval is five years: 2001–2006 and 2008–2013. The free cash flow and the discount coefficients for these periods are given in Tables 4.1 and 4.2. The results of valuation based on the proposed procedure are presented in Tables 4.3 and 4.4, whose columns contain the following data: Table 4.1  Free cash flow into firms (FCFF) FCFF

2001

2002

2003

2004

2005

2006

2008

2009

2010

2011

2012

2013

VZ DIS JNJ PFE BA MMM AAPL CVX PG INTC GILD WMT ORCL AA

3387 13079 14286 8369 5712 10944 −1323 9483 14845 9588 1741 1694 1679 2537 2439 4488 3988 3755 4220 2949 5852 6075 7738 8450 5478 7765 9035 10985 11723 8971 5060 7697 8301 12369 8837 20776 17090 11266 11464 18331 4449 4025 789 7670 5569 4200 −3290 2614 123 −331 2475 2547 2883 3036 3341 3002 3152 3972 3812 3087 106 558 1383 1832 5773 7517 12525 20919 −160 −129 1217 5356 6657 9038 11458 11287 12985 −3291 9354 13870 4710 6018 5877 7523 5535 8778 7131 10754 12776 7833 −305 3064 6627 8460 5948 3863 6042 7245 9503 8440 397 653 732 1728 2390 2380 3029 −105 −43 −40 1787 2810 2753 6627 2169 2593 5538 11137 14924 8742 493 2552 2369 2711 3177 4409 6414 7337 9499 10516 2584 544 1336 1323 479 107 −1278 414 1837 1331

1503 4718 12157 17395 2588 3896 31225 5003 9374 4532 3249 11073 12088 746

23250 6822 11893 19077 4089 4515 31502 −7649 8975 8145 3342 12722 12361 1215

Table 4.2  Discount coefficients (WACC) WACC

2001

2002

2003

2004

2005

2006

2008

2009

2010

2011

2012

2013

VZ DIS JNJ PFE BA MMM AAPL CVX PG INTC GILD WMT ORCL AA

5.9 9.0 7.8 7.8 8.3 8.4 11.1 6.6 6.7 13.4 10.4 7.9 11.3 8.2

6.8 7.6 8.8 9.6 8.3 9.9 9.7 7.4 6.6 14.2 11.8 8.3 11.8 9.5

7.2 8.3 8.2 8.6 7.6 8.1 9.2 7.6 7.1 11.3 9.0 9.4 11.2 9.3

7.3 9.0 7.2 7.6 8.8 7.7 8.6 6.9 7.0 10.5 8.9 8.2 9.5 8.9

8.1 9.5 8.0 9.0 10.2 9.6 12.8 9.0 7.0 11.9 13.0 8.7 13.0 11.2

8.6 8.4 7.3 9.7 9.3 8.9 12.9 9.7 7.2 11.8 11.0 8.3 10.4 10.1

8.0 9.4 8.3 9.2 9.8 9.0 12.7 11.3 7.9 11.7 10.5 8.6 10.0 9.0

7.3 10.1 8.0 7.9 9.5 9.3 11.4 10.0 7.9 10.5 8.8 8.9 10.0 11.3

6.7 10.1 8.1 8.2 11.2 10.5 10.9 10.9 7.9 11.6 8.4 7.2 9.1 11.7

5.8 9.9 6.8 7.3 10.7 9.7 11.9 10.7 6.8 10.3 7.4 6.3 9.6 8.2

6.2 9.8 7.0 7.3 10.1 9.6 10.1 10.6 6.4 9.5 8.3 6.3 10.7 7.6

6.4 10.0 7.6 7.1 9.1 8.9 9.8 9.2 6.7 8.4 10.1 5.9 10.0 8.2

NGUYEN 9781800880894 PRINT.indd 79

21/09/2022 13:14

80  Handbook of banking and finance in emerging markets Table 4.3  Valuations for 2001–2006

VZ DIS JNJ PFE BA MMM AAPL CVX PG INTC GILD WMT ORCL AA

(2)

(3)

α less

α1 EBIT

(4)

(5)

α2 Rev α3 EV

1.44 1.016 1.3218 1.51 1.89 1.3406 1.46 1,738 1.65 1.53 1.409 1.666 1.53 0.8316 1.057 1.53 1.873 1.428 1.66 3.6 −7.366 1.48 3.034 2.0 1.4 2.05 1.738 1.76 2.41 1.333 1.66 −13.6 12.94 1.51 1.65 1.616 1.7 1.313 1.312 1.59 1.83 1.28

(6)

(7)

(8)

(9)

(10)

(11)

V1

V2

V3

EV2001

EV2006

V0

0.79 146950 520240 96030 212332.98 1.59 87914 48510 1.09 110660 177267 0.67 544940 77420 250783 1.72 37720 55805 42577 1.2 176620 53523 49152 38.9 1422 1.44 110051 2.28 93838 0.525 82463 28384 206304 4.99 5974 0.82 29262 275108 0.92 47787 47663 23656 80304 0.92 15293 7145 37668

167768 141640 77063 28934 193908 89124 167232 125510 73401 49771 59132 33216 55362 6175 158895 105920 214179 95866 108389 46193 29839 2763 225979 39607 74148 26367 34606 8105

Table 4.4  Valuations for 2008–2013 (2)

(3)

(4)

(5)

α less α1 EBIT α2 Rev α3 EV VZ DIS JNJ PFE BA MMM AAPL CVX PG INTC GILD WMT ORCL AA

1.37 1.61 1.43 1.44 1.62 1.58 1.67 1.63 1.41 1.61 1.51 1.4 1.6 1.57

1.729 1.27 1.39 1.024 1.708 1.206 5.884 0.777 0.884 1.3 1.7 1.25 1.8 0.536

1.238 1.246 1.119 1.068 1.422 1.222 4.559 0.83 1.011 1.4 2.1 1.243 1.66 0.856

1.35 1.74 1.54 1.79 2.73 2.13 3.46 1.63 1.08 1.67 2.68 1.17 1.24 0.99

(6)

(7)

(8)

(9)

(10)

(11)

V1

V2

V3

EV2008

EV2013

V0

236668 129536 249534 202942 97065 94965 308776 245460 238390 114396 119594 285749 145901 21251

177710 43778 147670 200050 17030 39873 177730 3517 139520 75022 33108 167360 102760 11427

501260 3499600 175152 78642 73443 74435 1412200 203190 161518 211600 23666 113226 53489 35503 61756 64500 44441 89115 25987 2 7706 150101 108850 143350 173180 220142 146580 214910 68662 44615 452050 431490 292710 244322 170950 117854 4479 6498 8009 21479

(2) restrictions on αj values (see equation (4.13)); (3) αj1 = EBITt+5 /EBITt; (4) αj2 = Revenuet+5 /Revenuet; (5) αj3 = EVt+5 /EVt (see equation (4.14)); (6)–(8) the firm’s value Vi calculated for αji (i = 1, 2, 3); (9) and (10) EV data for 2001 and 2006 (2008 and 2013); (11) the firm’s value V0 calculated for αj = 1.

NGUYEN 9781800880894 PRINT.indd 80

21/09/2022 13:14

Asset-based valuation: a modified discounted cash flow approach  ­81 The obtained firm values are compared with EV 2001 and 2008 values (in some cases also 2006 and 2013) to determine how close they are and to evaluate the efficiency of the modified intrinsic value model. As seen from Tables 4.3 and 4.4, the inequality (4.13) significantly restricts the use of EBIT, Revenue and EV ratios in the firm’s valuation procedure. For 2002–2006 period 4 and 2008–2013 period 2 companies were not able to be evaluated based on EBIT, Revenue and EV ratios data, and only in two cases were the values close to the EV of 2001 and 2008 (bold numbers in the relevant columns). The value V0 calculated for αj = 1 is close to the EV of 2001 and 2008 three and four times, respectively (numbers in bold); in addition, they are within or close to 2002–2006 and 2008–2013 intervals once and twice, respectively (bold italic numbers). The data in Tables 4.3 and 4.4 is presented without analysis of drops or even negative cash flow in separate years due to events that can be considered as temporarily or/and justified in order to improve a firm’s productivity. The expressions (4.12), (4.16) and (4.17) can be used to analyze the cases of substantial difference between V0 and the corresponding EV. In many of these sample cases the market reacted to only specific data (in many cases related to the latest years) with a high FCFF, and the related companies were overpriced. For example, in the period 2002–2006 the FFCF of the company VZ does not show any uptrend, and its 2005 value is significantly lower than in other years. By excluding this year data and recalculating V0 we obtain the value close to the EV of 2006. This allows us to assume that the estimate for 2006, as a part of the negotiation set, is more realistic. The obtained V0 of 2008, which is close to that year’s EV, supports this assumption. The overpriced DIS company is the result of evaluation focused only on its best year. The close EV value that reflects the best year expectations can be obtained by using (4.17). The same conclusion can be made for GILD, INTC (2001) and CVX (2008). The companies WMT, ORCL and AA look highly overpriced since their V0 estimates based on their best year – found using (4.17) – are significantly lower than the related EV. The V0 of AAPL shows that the company was underpriced.

6. CONCLUSION The modified intrinsic value model proposed here enables one to obtain cost estimates for both buyers and sellers to create a negotiation set useful for decision making to find a rational trade-off. In contrast to the existing valuation procedures, the considered approach offers a valuation procedure that operates only with variables characterizing a certain finite forecasting period. When valuing a firm, it is important to assess its stage in life. Comparable analysis using historic data can help create an overall picture of and value a company in the best way possible. A technical (market) analysis combined with the fundamental analysis based on (4.12) is the basis of various trading scenarios for both buyers and sellers. In practice, it is worth considering three possible scenarios for a firm: pessimistic, most probable and optimistic. This can be done by calculating the related discounted cash flow and considering various α-ratios. The case α = 1 presents a reasonable negotiation price that should be included in a negotiation set.

NGUYEN 9781800880894 PRINT.indd 81

21/09/2022 13:14

82  Handbook of banking and finance in emerging markets The offered approach supports the opinion that efficient models containing the human factor should be concise and operate with minimal number of uncertain parameters related to the problem under consideration.

REFERENCES Cruise, B. (2012). Economic profitability and the valuation of the firm. Journal of Portfolio Management, 38 (2), 122–135. Damodaran, A. (2012). Investment Valuation: Tools and Techniques for Determining the Value of any Asset, New York: John Wiley & Sons. Damodaran, A. (2013). A tangled web of values: Enterprise value, firm value and market cap. http://aswathda​ modaran.blogspot.com/2013/06/a-tangled-web-of-values-enterprise.html. Fama, E. (1965). Random walks in stock market prices. Financial Analysts Journal, 21 (5), 55–59. Fuller, R. and Hsia, C. (1984). A simplified common stock valuation model. Financial Analysts Journal, 40 (5), 49–56. Gordon, M.J. (1959). Dividends, earnings, and stock prices. Review of Economics and Statistics, 41 (2), 99–105. Hand, J.R., Coyne, J.G., Green, J.R., and Zhang, X.F. (2017). The use of residual income valuation methods by U.S. sell-side equity analysts. Journal of Financial Reporting, 2 (1), 1–29. Jorgensen, B.N, Lee, Y.G., and Yoo, Y.K. (2011). The valuation accuracy of equity value estimates inferred from conventional empirical implementations of the abnormal earnings growth model: US evidence. Journal of Business Finance & Accounting, 38 (3–4), 446–471. Malkiel, B.G. (1963). Equity yields, growth, and the structure of share prices. American Economic Review, 53 (5), 1004–1031. Molodovsky, N., May, C., Chottiner, S. (1965). Common stock valuation: Principles, tables and application. Financial Analysts Journal, 21 (2), 104–123. Ohlson, J.A. (1995). Earnings, book values, and dividends in equity valuation. Contemporary Accounting Research, 11 (2), 661–687. Yanushevsky, R. and Yanushevsky, D. (2015). An approach to improve mean-variance portfolio optimization model. Journal of Asset Management 16 (3), 209–219.

NGUYEN 9781800880894 PRINT.indd 82

21/09/2022 13:14

dence

5. Financial integration in Asia: some empirical evidence An Thi Thuy Duong and Clemens Kool

1. INTRODUCTION The 1997 Asian financial crisis (AFC), triggered by the collapse of the Thai Baht, led to a strong drive for regional financial cooperation in Asia. Examples are the 2000 Chiang Mai initiative, the 2003 Asian Bond Market Initiative, the 2004 Association of Southeast Asian Nations (ASEAN) Capital Market Forum, and the 2012 ASEAN Comprehensive Investment Agreement. Despite these efforts, empirical evidence shows that financial integration in Asia remains incomplete. Some studies confirm that the process of financial integration is moving forward (Yang et al., 2003; Boubakri and Guillaumin, 2015). However, others suggest that the region has weak financial market integration (Kim  et  al.,  2005; Kim et al., 2008; Park and Lee, 2011; Claus and Lucey, 2012) or that  financial  integration lags real economic integration (Jang, 2011; Lee et al., 2013).  Most existing  literature focuses on global integration (Beirne et al., 2010; Chevallier et al., 2018)  rather than regional financial integration. Exceptions are Guillaumin (2009), Yu et al. (2010), and Hinojales and Park (2011), who show that financial integration has recently intensified. Moreover, much of the previous literature does not fully account for the impact of the 2008 global financial crisis (GFC) on financial integration. This study aims to assess the degree to which selected Asian economies have become more financially integrated within the region in the recent two decades. We use daily data from 11 Asian equity markets from 31 May 2002 to 1 June 2018, which includes the GFC. We apply four commonly used methods to assess financial integration: beta-convergence, sigma-convergence, the Markov regime switching model, and the Dynamic Conditional Correlation (DCC)-Generalized Autoregressive Conditional Heteroscedasticity (GARCH) model. The first two examine price convergence to gauge the degree and speed of financial integration; the latter two methods examine the extent of asset return comovements to capture the short-run dynamics of financial integration. We make three contributions to the literature. First, our analysis fills a gap in the literature by analyzing financial integration dynamics during and after the GFC. Second,  we  explicitly compare the period before and after the GFC to allow for a correct  assessment of the evolution of integration without the potential distortions arising from the GFC. Third, we use four complementary methods to provide a robust overall picture. Our empirical evidence can be summarized as follows. First, we find the degree of financial integration to be positively correlated with the level of economic development, as measured by income levels. Second, unlike previous studies, we show that integration for most economies has followed a steady path and is ongoing, notwithstanding the 83

NGUYEN 9781800880894 PRINT.indd 83

21/09/2022 13:14

84  Handbook of banking and finance in emerging markets short-term discontinuity represented by the 2008 crisis. More particularly, integration dynamics before crisis and after crisis are qualitatively similar. Third, results for the crisis period are mixed and somewhat ambiguous, possibly due to the short sample period. The remainder of this chapter is structured as follows. Section 2 reviews the empirical literature. Section 3 presents the data and its properties. Section 4 concerns the research design. Section 5 presents and discusses the empirical results. Section 6 summarizes and concludes.

2. FINANCIAL INTEGRATION IN ASIA: EMPIRICAL EVIDENCE In this section we summarize the empirical evidence for financial integration in Asia in the last decades. There is little consensus so far on the degree of such integration. Early papers, including Chan et al. (1992) and Huang (1995), found no evidence of co-integration between Asian and American (U.S.) markets or among the Asian markets themselves. DeFusco et al. (1996) found no co-integrating vectors among the equity markets of South Korea, the Philippines, Taiwan, Malaysia, and Thailand from 1989 to 1993. By contrast, Chung and Liu (1994) identified two co-integration relationships among stock prices of the U.S. and five East Asian countries: Japan, Taiwan, Hong Kong, Singapore, and South Korea. Masih and Masih (1999) found a single co-integrating vector among Southeast Asian equity markets from 1992 to 1997. These early papers all treat convergence as a static phenomenon, not a time-varying process, which may provide misleading implications for policymakers and practitioners. Many subsequent papers have attempted to allow time-variant integration by using rolling or recursive co-integration approaches. For example, Awokuse et al. (2009) applied rolling co-integration methods to detect the time-varying co-integration relationship between some Asian emerging markets and the major global stock markets from 1998 to 2003. Their study also shows evidence of increasing financial integration in Asia after the 1997 AFC. Assidenou (2011) finds one co-integration vector among the Organisation for Economic Co-operation and Development (OECD) group, the Pacific group, and the Asia group during 2008–2009. Few empirical studies have applied Markov-switching models to Asian stock markets; and if they do, they mostly focus on detecting regimes in a single country rather than regime synchronization (Li and Lin, 2003; Wang and Theobald, 2008). In general, existing studies have shown that overall financial integration remained weak before the 1997 AFC but accelerated afterwards (Kim et al., 2005; Park and Lee, 2011; Kim and Lee, 2012). Some studies have argued that Asian financial markets are globally integrated (commonly represented by the United States).1 Meanwhile Chiang et al. (2007) and Singh et al. (2010), among others, have argued that Asian financial integration is more regionalized (commonly represented by Japan) than globalized. Chi et al. (2006) used an Intertemporal Capital Asset Pricing Model approach to show that integration among Asian national markets and with the region (Japan) is higher than their integration with the world (U.S.). On the other hand, Kim et al. (2005) and Park and Lee (2011) offered evidence to suggest that the region’s equity markets are integrated more globally than regionally. Similarly, Ng

NGUYEN 9781800880894 PRINT.indd 84

21/09/2022 13:14

Financial integration in Asia: some empirical evidence  ­85 (2000) examined the role of Japanese and U.S. markets in explaining the volatility of returns in six Pacific-Basin equity markets from 1975 to 1996, reporting that the influence of the world market tends to exceed that of the regional market – though together they still account for a small part of the variation in return. Beirne et al. (2010) and Chevallier et al. (2018) argued that Asian markets are mainly affected by global rather than regional shocks. While the existing literature has mostly focused on the financial integration of Asian economies with developed economies, few studies have assessed the degree of financial integration among Asian economies. Yu et al. (2010) used several methodologies to assess the degree of integration in ten Asia economies, finding a stronger degree of integration within the high-income group of Hong Kong, Taiwan, South Korea, and Singapore compared to other, lower-income economies in the region. Huyghebaert and Wang (2010) found that shocks in Hong Kong and Singapore largely affected other East Asian equity markets after 1997 but had no effect on Mainland China. Glick and Hutchison (2013) evaluated the financial linkages between China and its neighboring economies from 2005 to 2012, reporting weak cross-country correlations in long-term interest rates but stronger correlations in the equity market. They attribute these equity market linkages to increasing investor attentiveness to China as a source and destination of equity finance during the crisis rather than to any financial deepening or liberalization. Chien et al. (2015) applied the recursive co-integration technique to confirm increasing, albeit limited, integration from 1992 to 2013 between China and the region. Boubakri and Guillaumin (2015) examined East Asian financial integration between 1990 and 2012 using the multivariate DCC-GARCH framework. They found that emerging equity markets were partially segmented before 2008, and that regional integration followed an upward trend after 2008. By contrast, Rizavi et al. (2011) examined integration in the Asian region using beta- and sigma-convergence and found limited evidence of convergence from 2004 to 2009. To summarize, the literature offers no consensus regarding the degree of intra-regional financial integration in Asia; nor is there a clear preference for one methodology to measure financial integration. Furthermore, there is little evidence of how the degree of financial integration has changed after the recent global financial crisis.

3.  DATA AND TIME SERIES PROPERTIES Our 11-country sample consists of Japan (JPN), Singapore (SGP), Hong Kong (HKG), South Korea (KOR), Taiwan (TAW), China (CHN), Thailand (THA), Malaysia (MYS), the Philippines (PHL), Indonesia (IDN), and Viet Nam (VNM). For each economy, we employ the Morgan Stanley Capital International (MSCI) daily equity index extracted from the Thomson Financial DataStream from 31 May 2002 to 1 June 2018 (4,176 observations).2 All indices are expressed in U.S. dollars. For the regional benchmark, we use the MSCI ACFM Asia index, which covers 15 countries, including all economies considered in this study.3 Returns are calculated as the first difference of logarithmic prices multiplied by 100 (daily percentage return). To account for the potential impact of the GFC on financial integration, we divide the sample period into three sub-periods: before crisis (31 May 2002 to 31 July 2007), during

NGUYEN 9781800880894 PRINT.indd 85

21/09/2022 13:14

86  Handbook of banking and finance in emerging markets crisis (1 August 2007 to 31 March 2009), and after crisis (1 April 2009 to 1 June 2018). The starting date and length of the GFC are defined following existing literature.4 In addition, we divide our sample into three country groups based on income level, using the World Bank classification. The first group includes the high-income economies of Japan, South Korea, Singapore, Taiwan, and Hong Kong. The second group includes the uppermiddle-income economies of China, Malaysia, and Thailand. The third group includes the lower-middle-income economies of Indonesia, the Philippines, and Viet Nam. Table 5.1 contains summary statistics of the return series for individual countries, for the three groups of countries, and for the whole period as well as the three sub-periods. Mean returns and volatility are quite diverse among economies and across periods. The average return is rather low in Japan, Taiwan, Thailand, and Viet Nam, even turning negative for Viet Nam. Indonesia, Malaysia, and China experience the highest average returns. With respect to volatility, captured by the standard deviation of daily returns, Thailand is the least volatile market, while South Korea, China, and Indonesia are the most volatile markets. Furthermore, all returns have negative skewness and positive excess kurtosis, suggesting that the returns are left-tailed with higher peaks and have thicker tails than a normal distribution. Looking at the whole period, the highest average returns and lowest volatility are observed for the lower-middle-income group, while the high-income group has the lowest average returns and the highest volatility. Wald tests show that the differences in average returns between groups are insignificant, suggesting that the region has become increasingly financially integrated. Examining the average returns of the three sub-periods, average stock returns are generally higher before crisis compared to after crisis or during crisis. The dispersion of returns across countries appears to have declined after the crisis. We further test for differences in average returns between pairs of sub-periods: during crisis and before crisis, during crisis and after crisis, and before crisis and after crisis. Wald tests suggest that there are significant differences in the average returns between the crisis period and the other periods, while there is no significant difference between the before and after crisis periods. Only in Viet Nam are the differences in average returns weakly significant between the before and after crisis periods.5 Subsequently, we examine the time series properties in more depth. To test for unit roots, we employ a standard Augmented Dickey-Fuller (ADF) test, with lag selection through the Schwarz information criterion (SIC). The results indicate that all logarithmic price series are non-stationary, while all return series are stationary at the 1% level. The null hypothesis that the return distribution has a skewness of zero and kurtosis of three is rejected at the 1% level for all economies, indicating that the returns are not normally distributed. We also test for the presence of autocorrelation using the Ljung-Box test, and test for heteroscedasticity using the Autoregressive Conditional Heteroscedasticity (ARCH)-Lagrange Multiplier (LM) test. Overall, the evidence suggests the presence of autocorrelation and ARCH effects in the return series for all countries.6

4.  MEASUREMENT OF FINANCIAL MARKET INTEGRATION In this section, we outline the four methods that we use to assess the degree of financial integration across Asian equity markets.7 Consecutively, we discuss beta-convergence,

NGUYEN 9781800880894 PRINT.indd 86

21/09/2022 13:14

87

NGUYEN 9781800880894 PRINT.indd 87

21/09/2022 13:14

0.034

0.045 0.034 −0.000 0.015

4175 4175 3001 4175

0.041 0.018 0.041

4175 4175 4175

4175

0.033

0.012 0.022 0.025 0.031 0.014

4175 4175 4175 4175 4175

4175

0.021

4175

mean

15.04 9.33 5.22 9.40

6.93

14.04 5.78 10.52

8.82

11.47 8.56 10.45 24.99 8.23

11.72

max

 

−19.95 −14.49 −6.99 −8.43

−11.29

−12.84 −11.28 −18.08

−9.14

−9.51 −9.81 −12.57 −20.67 −7.17

−8.85

min

 

1.81 1.43 1.63 1.13

1.20

1.66 0.94 1.53

1.13

1.35 1.24 1.24 1.78 1.39

1.13

sd

 

−0.6 −0.5 −0.1 −0.5

−0.7

−0.1 −0.4 −0.6

−0.3

−0.2 −0.2 −0.2 −0.2 −0.2

−0.3

skw

 

12.5 9.1 4.4 10.1

9.4

10.3 11.3 13.6

9.6

7.9 8.9 11.1 21.0 6.2

10.5

kurt

 

0.108 0.072 0.229 0.039

0.093

0.101 0.051 0.079

0.077

0.034 0.064 0.045 0.078 0.025

 

−0.135 −0.143 −0.224 −0.155

−0.167

−0.109 −0.111 −0.16

−0.127

−0.141 −0.166 −0.121 −0.186 −0.137

−0.150

(2)

(1) 0.049

During GFC

Before GFC

 

0.042 0.045 0.024 0.033

0.037

0.033 0.022 0.056

0.037

0.028 0.032 0.041 0.044 0.036

0.036

(3)

After GFC

 

2.06** 2.35** 2.17** 2.61***

3.29***

1.90* 2.89*** 2.44**

2.84***

2.07** 2.89*** 2.08** 2.22** 1.78*

2.69***

(1) = (2)

 

−1.87* −2.48** −3.06*** −3.06***

−3.38***

−1.54 −2.54** −2.73***

−2.62***

−2.33** −2.87*** −2.33** −2.38** −2.43**

−3.02***

(2) = (3)

t-value

 

1.18 0.62 1.75* 0.17

1.52

1.47 1.01 0.47

1.2

0.15 0.9 0.12 0.68 −0.26

0.40

(1) = (3)

Note:  This table reports the descriptive statistics of the daily percentage return series for all economies and the regional equity index. sd is standard deviation, skw is skewness, and kurt is kurtosis. The t-test is the two-sample t-test on the equality of means. *, **, *** indicate significance levels of 10%, 5%, and 1% respectively.

(d) ACFM t-test for the mean differences −1.11 t-value (a) = (b) −0.04 t-value (b) = (c )   −0.79   t-value (a) = (c )

(c)Lower-middleincome economies IDN PHL VNM

(b)Upper-middleincome economies CHN THA MYS

(a) High-income economies JPN SGP HKG KOR TAW

N

Table 5.1  Descriptive statistics of the return series

88  Handbook of banking and finance in emerging markets sigma-convergence, the Markov regime switching model of Hamilton (1989), which allows for shifts with unknown timing, and the DCC-GARCH model, which accounts for dynamics in the degree of financial integration. 4.1  Beta-Convergence Approach The concept of beta-convergence originates from the growth literature (see Barro and Sala-i-Martin, 1992), and describes the catch-up effect: an economy with lower Gross Domestic Product (GDP) per capita tends to grow faster than one with higher GDP per capita. We use beta-convergence to evaluate if and at what speed equity prices in one country converge with equity prices in a benchmark country, according to equation (5.1):

(5.1)

Where ri,t, rB,t are, respectively, the equity returns of economy i and the benchmark’s equity index at time t; lnpi,t, lnpB,t are the logarithm of economy i’s equity price index and the logarithm of the benchmark’s equity price index at time t; βi is the convergence coefficient; φ measures the scale between the logarithm of economy i’s equity index and that of the benchmark’s equity index; L is the number of lags; and εi,t is the error term. A significantly negative estimate for βi indicates convergence, with the size of βi measuring the speed of convergence. Specifically, the half-life of a shock to the price differential is computed as:

(5.2)

We estimate equation (5.1) with Nonlinear Least Squares, in which the lag length is determined by the Schwarz Bayesian Information Criteria (SBIC) and standard errors are corrected for heteroscedasticity and serial correlation. 4.2  Cross-Market Return Dispersion Approach: Sigma-Convergence Standard economic theories suggest that financial integration corresponds to decreasing cross-sectional return dispersion (see for instance Adam et al., 2002). We apply this concept to test whether the cross-country standard deviation of stock returns has a declining trend. Dispersion of an equity return i relative to the benchmark return (B) is measured as follows:

(5.3)

Where ri,s, rB,s are, respectively, returns of economy i’s equity index and the benchmark’s equity index at time s; and σ is is the return dispersion between economy i and the benchmark market B over the window [s-T, s]. By definition, σ is is always positive. Lower values of σ is indicate higher levels of convergence. Sigma-convergence to zero indicates full ­integration.

NGUYEN 9781800880894 PRINT.indd 88

21/09/2022 13:14

Financial integration in Asia: some empirical evidence  ­89 4.3 Synchronization of Financial Market Regimes Approach: Markov Regime Switching Model The Markov regime switching method, developed by Hamilton (1989), identifies turning points in stock markets and has been widely used to measure financial integration. Two stock markets are considered to be integrated if they tend to be in the same regime. Stock return regimes are usually characterized as either low-volatility or high-volatility (Ang and Bekaert, 2002, 2004; Maheu and McCurdy, 2000). We follow the literature and distinguish two regimes, indicated by the variable st, which is non-observable and takes a value of 1 or 2. Assume that ri,t is the return of economy i’s equity index, with regimedependent means μi,s and variances σi,s : t



t

(5.4)

We assume that state transition probabilities are constant and characterized by a matrix ∏i, comprising the transition probabilities of being in state m at time t given state n at time t−1:



(5.5)

2 The sum of the probabilities must satisfy ∑n=1  pmn = 1 and 0 ≤ pmn ≤ 1 for m,n = 1,2. Note that we use the information set from the overall sample period to estimate the probability  of being in a regime at time t, Pr(st|ri,T ,ri,T−1,...); in other words, we use the smoothed probability. Conditional on being in state m, the expected duration Dm of the state m is:



(5.6)

The closer pmn is to 1, the higher is the expected duration of state m and the more persistent the regime. After estimating the coefficients of the model and the transition matrix, we apply the following logit transformation to remove the 0–1 range restriction from the probability values:

(5.7)

The degree of integration between equity market i and the benchmark equity market is determined by the Pearson pairwise correlation between logits of the regime probabilities of the two markets.8 4.4  Time-Varying Correlation Approach: DCC-GARCH Model Increasing correlations among equity returns over time indicate increasing market integration (Yu et al., 2010). To evaluate this for our sample, we employ the DCC-GARCH model developed by Engle (2002). Let ri,t, the daily percentage return series of economy i, be modeled as follows:

NGUYEN 9781800880894 PRINT.indd 89

21/09/2022 13:14

90  Handbook of banking and finance in emerging markets (5.8)

(5.9)

Where n is the lag length order determined by the SBIC; is the lagged U.S. stock return, which is often used as a proxy for global factor (Chiang et al., 2007; Dungey et al., 2003; Hemche et al., 2016; Sehgal et al., 2017); εi,t is the innovation on the information set Ωt−1; αi is a constant; ϕi and δi are ARCH and GARCH effects, respectively; and ϕi + δi < 1. We estimate equations (5.8) and (5.9) both for each country’s equity return and for the benchmark return. Let zi,t and zB,t be the standardized residuals of the equity market returns of economy i and the benchmark equity market B, respectively, resulting from estimating equations (5.8) and (5.9). zi,t and zB,t then obey the following GARCH process:

(5.10)



(5.11)

where qiB,t is the off-diagonal element of the variance–covariance matrix; ρ-iB is the unconditional expectation of the cross-product zi,t and zB,t; and ρiB,t is the conditional correlation between the equity market returns of economy i and the benchmark equity market B at time t. The non-negative scalar parameters λ1,λ2 capture the effects of previous shocks and of the previous dynamic conditional correlation on the current dynamic conditional correlation: λ1 + λ2 < 1. Finally, we estimate the average pairwise, conditional correlation coefficients between the equity returns of each economy and the regional benchmark equity return to estimate the degree of integration between these markets and the regional benchmark.

5.  EMPIRICAL RESULTS In this section, we present and discuss the results of our four approaches to measure financial integration in subsections 5.1 to 5.4.9 5.1  Beta-Convergence Approach Table 5.2 presents the results for beta-convergence based on estimation of equation (5.1). We estimate equity price convergence for each economy relative to the regional MSCI ACFM index for the whole period, as well as for the three sub-periods. Due to space constraints, we only report estimates and standard deviations for β, the main parameter of interest. The last three rows of Table 5.2 report the Wald test of the null hypothesis that there is no difference in beta estimates between pairs of sub-periods. Focusing first on the overall period, the beta coefficients vary across economies, suggesting that countries converge to the regional benchmark at different rates. The beta

NGUYEN 9781800880894 PRINT.indd 90

21/09/2022 13:14

91

NGUYEN 9781800880894 PRINT.indd 91

21/09/2022 13:14

−0.60*** (0.20) −0.63* (0.35) −2.89* (1.74) −1.76*** (0.51) 1.25 0.30 3.38*

−0.11* (0.07) −0.60** (0.30) −3.31** (1.47) −0.58*** (0.15) 3.46* 3.59* 0.01

SGP −0.13* (0.07) −0.83** (0.34) −3.58*** (1.18) −0.87*** (0.22) 6.48** 6.50** 0.01

HKG −0.16* (0.08) −1.41*** (0.47) −7.85*** (2.25) −0.50*** (0.16) 3.91** 5.17** 3.33*

KOR −0.49*** (0.15) −1.07*** (0.41) −1.72* (1.00) −0.78*** (0.23) 0.38 0.87 0.29

TAW −0.11** (0.05) −0.45 (0.29) −2.44** (1.14) −0.44** (0.19) 2.32 2.40 0.00

CHN −0.07 (0.05) −0.39 (0.29) −2.00* (1.15) −0.22** (0.10) 1.97 2.56 0.37

MYS

IDN

−0.12* −0.09* (0.06) (0.05) −0.55** −0.56** (0.24) (0.27) −2.32** −2.69*** (1.08) (0.90) −0.40*** −0.67*** (0.12) (0.15) 2.94* 4.51** 3.57* 4.19** 0.34 0.10

THA

−0.07 (0.05) −0.80** (0.34) −1.91* (1.00) −0.19* (0.12) 1.29 3.37* 3.16*

PHL

−0.16 (0.12) −3.74*** (1.28) −0.96* (0.57) −0.21 (0.21) 3.84* 2.07 6.80**

VNM

Note:  This table reports the beta estimates from the estimation of equation (5.1) and the Wald test statistics for the equality of the beta coefficients. Standard errors are in parentheses. *, **, *** indicate significance levels of 10%, 5%, and 1% respectively.

Wald test (2) = (3) Wald test (3) = (4) Wald test (2) = (4)

(4) After crisis

(3) During crisis

(2) Before crisis

(1) Whole period

JPN

Table 5.2  Beta convergence

92  Handbook of banking and finance in emerging markets coefficients for Japan are the highest and significant at the 1% level and indicate a half-life of 116 trading days. Taiwan, Singapore, Hong Kong, and South Korea have much lower beta coefficients, indicating a slower speed of convergence. In the upper-middle-income group, only China and Thailand display evidence of beta convergence. The speed is roughly similar to that of Singapore and Hong Kong. As a lower-middle-income country, Indonesia still shows significant convergence, though at a very slow speed. For Viet Nam, Malaysia, and the Philippines, the beta coefficients are insignificant, offering no evidence of price convergence. Next, we look at the beta coefficients across sub-periods. The speed of convergence displays considerable variation across countries and time. All β estimates are – at least marginally – significant across countries and sub-periods, with the exception of China and Malaysia in the before crisis period. Most economies experience faster convergence during crisis compared to the other two sub-periods, except for Viet Nam – though the difference is not always statistically significant due to high standard deviations. Estimated convergence appears to be insignificantly different after crisis compared to before crisis. Exceptions are Japan, where the speed of convergence is significantly higher after crisis, and South Korea and the Philippines, where it is significantly lower. Viet Nam is the only economy that shows no evidence of convergence after crisis. Overall, these results suggest that convergence occurred in most economies, albeit at different speeds. We observe faster convergence for higher-income economies and slower convergence for lower-income economies. Convergence sped up during the crisis. After the crisis, it returned to the level before the crisis, except that Japan showed increased convergence and South Korea and the Philippines decreased convergence. Little evidence for Vietnamese convergence is found. 5.2  Cross-Market Return Dispersion Approach: Sigma-Convergence In Table 5.3, we report the mean return dispersion σ is as defined in equation 5.3 across all daily dispersion estimates. Japan displays the lowest dispersion in returns, while Viet Nam displays the highest. More generally, the group of high-income economies has relatively low return dispersion, while the group of lower-middle-income economies has relatively high return dispersion, implying that the high-income economies are more integrated than the middle-income ones. Comparing sub-periods, dispersion of returns during the GFC is much larger than in the periods before and after crisis, indicating high return volatility during the crisis period. However, return dispersions declined from the period before to the period after crisis. This decline is smaller for the high-income economies and larger for the middle-income economies, suggesting increasing integration in the region, especially among middle-income economies. For Viet Nam, the return dispersion decreased considerably after the crisis, bringing it close to that of the other lower-middle-income economies. Figure 5.1(a–c) shows the time-varying nature of our dispersion measure: sigma-­ convergence for the five high-income economies, the three upper-middle-income economies, and the three lower-middle-income economies, respectively.10 Overall, we observe some cyclicality in the return dispersion series. Also, return dispersion trended upwards during the GFC but trended downwards after the crisis, returning roughly to the before crisis level for all income groups.

NGUYEN 9781800880894 PRINT.indd 92

21/09/2022 13:14

93

NGUYEN 9781800880894 PRINT.indd 93

21/09/2022 13:14

0.40*** (0.01) 0.33*** (0.33) 0.62*** (0.03) 0.39*** (0.01)

0.51*** (0.01) 0.49*** (0.01) 0.81*** (0.04) 0.46*** (0.01)

SGP 0.45*** (0.01) 0.46*** (0.01) 0.72*** (0.03) 0.39*** (0.01)

HKG 0.61*** (0.01) 0.63*** (0.02) 0.94*** (0.06) 0.53*** (0.01)

KOR 0.54*** (0.01) 0.61*** (0.02) 0.82*** (0.04) 0.45*** (0.01)

TAW 0.57*** (0.01) 0.60*** (0.01) 1.03*** (0.04) 0.47*** (0.01)

CHN 0.51*** (0.01) 0.52*** (0.01) 0.87*** (0.04) 0.43*** (0.01)

MYS 0.70*** (0.01) 0.77*** (0.02) 1.03*** (0.05) 0.61*** (0.01)

THA

0.77*** (0.01) 0.84*** (0.02) 1.14*** (0.05) 0.67*** (0.01)

IDN

0.69*** (0.01) 0.78*** (0.02) 1.00*** (0.04) 0.58*** (0.01)

PHL

VNM 0.92*** (0.02) 1.43*** (0.09) 1.44*** (0.06) 0.79*** (0.01)

Note:  This table reports the sigma estimates from the estimation of equation 5.3 and the two-sample t-test on the equality of the means of sigma across sub-periods. The values of standard error are in parentheses. *, **, *** indicate significance levels of 10%, 5%, and 1% respectively.

(4) After crisis

(3) During crisis

(2) Before crisis

(1) Whole period

JPN

Table 5.3  Sigma convergence

94

Handbook of banking and finance in emerging markets

Figure 5.1

HP sigma convergence

NGUYEN 9781800880894 PRINT.indd 94

21/09/2022 13:14

Financial integration in Asia: some empirical evidence  ­95

Figure 5.1  (continued) In the high-income group (Figure 5.1a), the before crisis period is already characterized by strong sigma-convergence. During the crisis, the differences in dispersion levels increase again, though all countries follow roughly the same, cyclical pattern. After 2009, dispersion levels gradually converge again. In the upper-middle-income group (Figure 5.1b), China and Malaysia show very similar patterns of dispersion, while Thailand converges to the other two after 2015. In the lower-middle-income group (Figure 5.1c), Indonesia and the Philippines display similar patterns in dispersion throughout the whole period. Viet Nam starts at much higher dispersion before crisis, converging gradually toward Indonesia and the Philippines after 2013. In summary, the high-income economies are the most converged and the lower-middleincome economies the least. The increased return dispersion during the crisis reflects high volatility; dispersion again fell after the crisis, suggesting an increasing degree of integration, especially for middle-income economies. Viet Nam has considerably lower return dispersion in the years after the GFC, bringing it closer to the level of integration of the lower-middle-income group. 5.3 Synchronization of Financial Market Regimes Approach: Markov Regime Switching Model We first examine whether there is evidence for two distinct regimes in equity returns. We start by estimating equation (5.4), which has both a switching mean return and variance.

NGUYEN 9781800880894 PRINT.indd 95

21/09/2022 13:14

96  Handbook of banking and finance in emerging markets This is the full model (MSM). Subsequently, we test restricted models, using the logarithmic likelihood ratio (LR) and the Akaike information criterion (AIC): a one-regime model (MSM1), a model with only a regime-dependent mean return (MSM2), and a model with only a regime-dependent variance (MSV2).11 Both the LR and AIC indicate that the full model (MSM) is preferred for all countries except Viet Nam, for which the MSV2 model is preferred.12 We estimate the preferred models and present the results in Table 5.4. First, the estimated volatility and mean returns clearly differentiate each regime. Regime 1 (2) is characterized by high (low) mean return and low (high) volatility. For all countries, the estimated mean return in regime 1 is positive and significant at least at the 5% level, while that in regime 2 is negative, although this finding is insignificant in most cases. Wald tests confirm that the standard deviation and the mean equity return differ significantly across regimes in all countries. Second, both regimes are very persistent, but regime 1 is always more persistent than regime 2.13 In the second step, we use the Markov regime estimates to assess the degree of comovement in equity markets across countries. To that end, we compute the bilateral correlation between the logits of the regime 1 probabilities of each economy’s equity index and the regional benchmark index. The results are shown in Table 5.5. Over the whole period, most countries exhibit considerable correlation (above 0.5) with the benchmark equity market; the exception, Viet Nam, has a correlation of 0.3. Before crisis, correlations are highest for the high-income countries, followed by those of two lower-middle-income countries (Indonesia and the Philippines), and only then the three upper-middleincome  countries. Viet Nam, the remaining lower-middle-income country, even has a negative correlation, although this is based on a small number of observations. After crisis, the correlations closely follow the level of development of each economy: highest for the high-income countries, followed by the three upper-middle-income countries and then the lower-middle-income countries, and lowest for Viet Nam. For the high-income countries, the correlations are roughly similar between the before and after crisis periods. The three upper-middle-income countries had a substantial rise in correlations after crisis, while the lower-middle-income countries of Indonesia and the Philippines had a small decline. For Viet Nam, the correlation became significantly positive, though still small, after crisis. During crisis, evidence is mixed. In summary, the findings show that the method well identifies the high- and low-­ volatility regimes for all economies. Both regimes show strong persistence, though more so for the low-volatility regime. Financial integration between each market and the benchmark equity market decreased during the crisis but bounced back after the crisis. Typically, the pattern of regime correlations corresponds to the country’s level of development (income), especially after crisis. The upper-middle-income countries – China, Malaysia, and Thailand – became more integrated with the benchmark index over time, catching up to a large extent with the high-income countries. Integration in Viet Nam, while increasing, still significantly lags. 5.4  Time-Varying Correlation Approach: DCC-GARCH Model In Table 5.6 we report the results of the estimation of equations (5.8)–(5.11), where the number of AR terms in the mean equation is determined by SBIC criteria. Panel A reports

NGUYEN 9781800880894 PRINT.indd 96

21/09/2022 13:14

97

NGUYEN 9781800880894 PRINT.indd 97

21/09/2022 13:14

7.14 0.01

Wald test mean Chi2 10.36 p −value 0.00 3.92 0.05

136.52 0.00

0.96 0.00

−0.42 (0.32) 3.90

0.26*** (0.08) 1.75

HKG

76.55 21.97

7.81 0.01

187.58 0.00

0.95 0.01

−0.82 (0.43) 5.31

0.43*** (0.09) 2.20

KOR

54.85 23.93

10.68 0.00

159.46 0.00

0.96 0.02

−0.62** (0.27) 3.90

0.36*** (0.09) 1.86

TAW

64.62 23.23

2.82 0.09

175.25 0.00

0.96 0.02

−0.26 (0.35) 4.76

0.38*** (0.10) 2.14

CHN

20.94 15.51

4.3 0.04

166.93 0.00

0.94 0.05

−0.16 (0.17) 2.66

0.27*** (0.07) 1.20

MYS

39.52 15.85

5.46 0.02

159.7 0.00

0.94 0.03

−0.39 (0.33) 4.48

0.44*** (0.11) 2.08

THA

26.39 15.47

3.34 0.07

230.36 0.00

0.94 0.04

−0.22 (0.34) 5.31

0.47*** (0.11) 2.10

IDN

21.41 20.86

6.91 0.01

118.06 0.00

0.95 0.05

−0.18 (0.21) 3.56

0.50*** (0.11) 1.72

PHL

28.34 11.36

– –

166.78 0.00

0.96 0.09

6.43

−0.06 (0.12) 2.26

VNM

12.96 11.85

28.15 0.00

185.57 0.00

0.92 0.08

−0.00* (0.00) 0.03

0.01*** (0.00) 0.01

ACFM

Note:  Duration is in weeks. Returns are the logarithmic differences of weekly equity index; sigma1, sigma2 are natural logarithms of the estimated standard deviations; mean1, mean2 are the estimated mean of return, respectively; p11 is the probability of staying in the regime 1, p21 is that of transforming from regime 2 to regime 1. Standard errors are in parentheses; *, **, *** are significance at 10%, 5%, 1% levels, respectively.

RCM 39.67 26.03 12.55 14.16 19.96 20 35.92 27.93 29.04 35.69 23.67 38.43 Log −1,760.11 −1,807.68 −1,804.07 −2,043.62 −1,928.86 −2,033.47 −1,673.92 −2,022.03 −2,144.19 −1,985.69 −1,562.52 2,128.60 likelihood SBIC 4.26 4.38 4.37 4.94 4.67 4.92 4.06 4.89 5.18 4.8 5.25 −5.05

Expected duration of the regime (weeks) Regime 1 22.19 30.51 114.15 Regime 2 15.79 15.96 26.84

220.49 0.00

Wald test sigma Chi2 84.54 p −value 0.00

−0.33 (0.23) 3.56

0.34*** (0.07) 1.45

SGP

0.94 0.33

−0.36* (0.19) 2.80

0.36*** (0.12) 1.40

JPN

Transition probabilities p11 0.95 p21 0.06

Sigma 2

Regime 2 Mean 2

Sigma 1

Regime 1 Mean 1

 

Table 5.4  Estimated coefficients for the Markov regime switching model

98

NGUYEN 9781800880894 PRINT.indd 98

21/09/2022 13:14

0.81*** 0.87*** 0.48*** 0.80***

0.77*** 0.73*** 0.58*** 0.73***

SGP 0.70*** 0.55*** 0.68*** 0.65***

HKG 0.71*** 0.605*** 0.43*** 0.65***

KOR 0.70*** 0.61*** 0.52*** 0.65***

TAW 0.63*** 0.34*** 0.38*** 0.62***

CHN 0.61*** 0.36*** 0.62*** 0.60***

MYS 0.52*** 0.27*** 0.23** 0.56***

THA

0.61*** 0.59*** 0.51*** 0.46***

IDN

0.60*** 0.54*** 0.52*** 0.45***

PHL

0.30*** −0.37** 0.11 0.25***

VNM

Note:  This table reports the Pearson pairwise correlations in the logit probability of regime 1 between each economy’s equity index and the regional benchmark index; *, **, *** indicate significance levels of 10%, 5%, and 1% respectively.

Whole period Before crisis During crisis After crisis

JPN

Table 5.5  Regime correlation between each economy and the regional benchmark index

Financial integration in Asia: some empirical evidence  ­99 the results of the GARCH specification for each Asian economy: Panel B reports the results of the conditional correlation equation (5.11); and the last row presents the Wald statistics for the specification conditions of the model. We observe that the GARCH and ARCH coefficients in the variance equation are always positive and significant at the 1% level, supporting time-variation. The Wald test for the null hypothesis that their sum is smaller than one cannot be rejected at the 1% level. Furthermore, the GARCH coefficients are large (ranging between 0.85 and 0.94) and always exceed the ARCH coefficients (ranging between 0.05 and 0.13), indicating the strong persistence of stock return volatility. Regarding the dynamic conditional correlations, we find that λ1 and λ2 are always statistically significant at the minimum level of 1% (except for Viet Nam, where λ1 is not statistically significant). This suggests that the conditional correlation also is time-variant. The Wald test for the null hypothesis that λ1 + λ2 < 1 cannot be rejected at the 1% level. In all specifications, λ2 always greatly exceeds λ1, confirming the strong persistence in conditional correlations. In Table 5.7, we compute the average pairwise, conditional correlation coefficients between the equity returns of each economy and the regional benchmark equity return. Overall, all equity markets display high correlation with the regional benchmark, except for Viet Nam. Japan correlates most strongly with the benchmark equity market, followed by the other high-income economies (Singapore, Hong Kong, Taiwan, and Korea). The lowest correlation is detected for the lower-middle-income countries, such as Indonesia and the Philippines. Viet Nam has a significant but very small correlation with the region, indicating only modest regional integration. In general, the conditional correlations are quite stable across sub-periods.14

6. CONCLUSION In this chapter, we investigated the dynamics of financial integration for a group of 11 Asian equity markets for the period May 2002–June 2018, using daily return data. In the analysis, we account for the potential impact of the global financial crisis on measures of financial integration, splitting the period into before crisis, during crisis, and after crisis sub-periods. We employed four complementary methods to measure integration: betaconvergence, sigma-convergence, Markov regime switching and DCC-GARCH. Finally, we distinguished groups of high-, upper-middle-, and lower-middle-income economies. We used the regional MSCI ACFM Asia index as the benchmark for our integration analysis. From our analysis, we draw three overarching conclusions. First, we find that the degree of integration closely corresponds with the level of economic development, regardless of the integration measure used. Second, while earlier studies have claimed that the degree of regional financial integration in Asian countries significantly increased after the 2008 GFC (Boubakri and Guillaumin, 2015; Yu et al., 2010), our results suggest instead that integration has remained largely unchanged when excluding the period during the crisis.15 That is, excluding the volatile crisis period, integration has proceeded at roughly the same pace in Singapore, Hong Kong, Taiwan, China, Thailand, and Indonesia.

NGUYEN 9781800880894 PRINT.indd 99

21/09/2022 13:14

100

NGUYEN 9781800880894 PRINT.indd 100

21/09/2022 13:14

Constant

GARCH

Constant   ARCH

L.rUSA

AR(2)

AR(1)

0.36*** (0.02) 0.00** (0.00) 0.07*** (0.01) 0.92*** (0.01) 0.00*** (0.00)

0.43*** (0.02) 0.000** (0.00) 0.06*** (0.01) 0.92*** (0.01) 0.00*** (0.00)

HKG

0.76*** 0.54*** 0.64*** (0.02) (0.03) (0.02) 0.03*** 0.02*** 0.03*** (0.01) 0.00 (0.01) 0.95*** 0.9705*** 0.94*** (0.01) (0.01) (0.02) (null: (ARCH+GARCH)−1 0 Quality β2 < 0 Liquidity β2 < 0 Liquidity and Quality β2 > 0.

3.2 Population and Sample The proposed subject population covers all Chinese commercial banks during the period 2010 to 2019. Our sample comprises a panel dataset of 28 publicly listed Chinese commercial banks on the Shanghai Stock Exchange (SHSE), the Shenzhen Stock Exchange (SZSE) and the Hong Kong Stock Exchange (SEHK) for that period. Publicly listed commercial banks were chosen because their financial data is more complete, of higher quality and easier to access. The banks used in this study are divided into three tiers, as shown in Tables 10.1–10.3.2 Table 10.1  First tier: state-owned commercial banks Name

Headquarter Listed Date Stock Exchange

Industrial and Commercial Bank of China (ICBC) Agricultural Bank of China (ABC) Bank of China (BOC) China Construction Bank (CCB) Bank of Communications (BOCOM)

Beijing Beijing Beijing Beijing Shanghai

27/10/2006 15/07/2010 05/07/2006 25/09/2007 15/05/2007

SHSE:601398 SHSE:601288 SHSE:601988 SHSE:601939 SHSE:601328

Table 10.2  Second tier: joint-stock commercial banks Name

Headquarter

Listed Date

Stock Exchange

China Merchants Bank (CMB) Industrial Bank (IB) Shanghai Pudong Development Bank (SPDB) China Minsheng Bank (CMBC) China CITIC Bank (CITIC) China Everbright Bank (CEB) Huaxia Bank (HX) Ping An Bank (PAB) China Zheshang Bank (CZB)

Shenzhen Fuzhou Shanghai Beijing Beijing Beijing Beijing Shenzhen Hangzhou

09/04/2002 05/02/2007 10/11/1999 19/12/2000 27/04/2007 18/08/2010 12/09/2003 01/03/2007 30/03/2016

SHSE:600036 SHSE:601166 SHSE:600000 SHSE:600016 SHSE:601998 SHSE:601818 SHSE:600015 SZSE:000001 SEHK:2016

NGUYEN 9781800880894 PRINT.indd 185

21/09/2022 13:14

186  Handbook of banking and finance in emerging markets Table 10.3  Third tier: city commercial banks Name

Headquarter

Listed Date

Stock Exchange

Bank of Beijing (BOB) Bank of Shanghai (BSH) Bank of Ningbo (BONB) Bank of Jiangsu (BOJS) Bank of Tianjin (BTJ) Bank of Qingdao (BQD) Bank of Chongqing (BCQ) Bank of Hangzhou (BHZ) Bank of Harbin (BOH) Bank of Nanjing (BONJ) Bank of Chengdu (BCD) Bank of Zhengzhou (BOZZ) Shengjing Bank (SJB) Huishang Bank (HSB)

Beijing Shanghai Ningbo Nanjing Tianjin Qingdao Chongqing Hangzhou Harbin Nanjing Chengdu Zhengzhou Shenyang Hefei

19/09/2007 16/11/2016 19/07/2007 02/08/2016 30/03/2016 03/12/2015 24/10/2013 27/10/2016 31/03/2014 19/07/2007 31/01/2018 23/12/2015 29/12/2014 12/11/2013

SHSE:601169 SHSE:601229 SZSE:002142 SHSE:600919 SEHK:1578 SEHK:3866 SEHK:1963 SHSE:600926 SEHK:6138 SHSE:601009 SHSE:601838 SEHK:6196 SEHK:2066 SEHK:3698

3.3 Independent, Dependent and Control Variables The independent variable in our hypotheses is foreign shareholding in Chinese listed commercial banks, captured by the aggregate percentage of total shares in a publicly traded Chinese commercial bank owned by foreign investors. These are defined as investors from countries and regions outside mainland China. Investments from Taiwan, Hong Kong and Macau Special Administrative Region are included as foreign investments as all three jurisdictions are used extensively by foreign investors to hold assets from China. There are four macroeconomic and bank-specific variables among the control variables: GDP, real interest rate, bank size and capital adequacy ratio. Nominal GDP growth rate tends to be positively related to bank profitability, and high bank profits are found to be associated with low loan provisions (Laeven and Majnoni, 2003). Real interest rate is defined as nominal interest rate less inflation rate. Ismail and Burak (2007) find a positive correlation between real interest rate and foreign direct investment flows to developing countries and emerging economies. In accordance with Polizzi et al. (2020), bank size by renminbi (RMB) is chosen as one of the bank-specific control variables, proxied by total assets (RMB). Empirical studies find a positive influence of firm size and economic prosperity on firm performance (Ho and Hsu, 2010) and bank profitability (Park et al., 2010). Capital adequacy ratio is a measure of a bank’s ability to use its available capital to deal with unexpected losses. It is calculated by dividing the sum of tier 1 and tier 2 capital by the risk-weighted assets. Barrell et al. (2010) find that a high capital adequacy ratio has a positive effect on bank profitability, including during periods of exogenous shocks. Four dependent variables are selected to measure Chinese listed commercial banks’ performance from different perspectives, such as profitability, liquidity and loan quality. Commonly used ROE is chosen as the profitability indicator, which is defined as earnings before taxes scaled by owner equity. We aim to examine how increasing foreign investment in Chinese listed commercial banks through capital markets could impact commercial bank performance. Therefore, ROE is a more suitable indicator of bank profitability than

NGUYEN 9781800880894 PRINT.indd 186

21/09/2022 13:14

The impact of foreign shareholders on Chinese commercial banks  ­187 ROA. NPL ratio is a comprehensive indicator of bank balance sheet quality. It is defined as the amount of loans that are in default for 90 days over gross loans. In accordance with Chi and Li (2017), NPL ratio is chosen to measure the credit risks of Chinese listed commercial banks. Loan-to-deposit ratio, the ratio between net loans and total deposits, is a measure of liquidity. It depicts two core activities that reflect a bank’s liquidity transformation function – lending and deposits – and as the proxy for bank liquidity performance in this study. Finally, loan loss coverage ratio shows how protected a bank is against future loan losses. It is defined as the amount of reserve for potential loan losses over impaired loans. This ratio not only reflects a bank’s loan quality, but also captures liquidity available to the bank to cover its bad loans. Therefore, it is selected as a dependent variable to measure bank performance from both a loan quality and a liquidity perspective. Table 10.4 gives an overview of the variables. 3.4 Descriptive Statistics As mentioned earlier, we use annual data for all variables across 28 listed commercial banks in China for a period of ten years from 2010 to 2019. The panel dataset is yearly. The data points that were unavailable from Capital IQ or Reuters Eikon were individually collected from the banks’ annual reports and financial statements. However, due to the absence of foreign shareholders during commercial banks’ unlisted period (pre-IPO), only 220 observations of foreign ownership percentage are available. Accordingly, only 220 out of the total 280 observations for other variables are selected for analysis here. Since the range of the cross-section, n = 22, is greater than that of the time dimension, t = 10, this Table 10.4  Overview of variables Variable

Definition

Foreign  shareholding Return on   equity (ROE)

Total shareholdings of foreign investors

NPL ratio Loan-to-deposit  ratio Loan loss   coverage ratio Bank size  (RMB) Capital  adequacy ratio Real interest rate Nominal GDP  (RMB)

Net income/Total equity Non-performing loans/ Total loans

Number of observations 220 280 280

Net loan/Total deposits Allowance for loan loss/ Impaired loans

280

Total assets Total capital/Risk-weighted assets = Tier 1 capital ratio + Tier 2 capital ratio Nominal interest rate − Inflation rate (annual) Gross domestic product (annual)

280

NGUYEN 9781800880894 PRINT.indd 187

280

280  10  10

Sources Reuters Eikon Bank’s Annual Reports Capital IQ Bank’s Annual Reports Capital IQ Bank’s Annual Reports Capital IQ Bank’s Annual Reports Capital IQ Bank’s Annual Reports Capital IQ Bank’s Annual Reports Capital IQ Bank’s Annual Reports World Bank People’s Bank of China World Bank National Bureau of Statistics

21/09/2022 13:14

188  Handbook of banking and finance in emerging markets Table 10.5  Descriptive statistics Variable Foreign Shareholdings Return on Equity Non-Performing Loan Loan-to-deposit Ratio Capital Adequacy Ratio Loan Loss Coverage Ratio Bank Size (RMB billions) Real Interest Rate Nominal GDP (RMB billions)

Sample

Mean

Std. dev

Minimum

Maximum

220 220 220 220 220 220 220 220  10

0.135 0.158 0.012 0.611 0.129 2.483 5,394 0.023 77,706

17.438 0.041 0.004 0.106 0.015 0.938 6,753 0.019 16,849

0.000 0.073 0.004 0.378 0.099 0.476 187.00 (0.014) 43,232

0.791 0.253 0.024 0.935 0.175 5.264 30,109 0.048 100,851

panel is considered short. Descriptive statistics of all observations are summarized in Table 10.5. From the descriptive statistics, the average foreign shareholding over ten years is 13.48%, which does not exceed the upper limit of 25% set by the CBRC in 2003 (Podpiera, 2006). Meanwhile, the range of foreign shareholdings is wide. Speculative investment is the main driver of the volatile foreign ownership of Chinese listed commercial banks, as suggested by the large standard deviation. Distress investing is also gaining popularity among international investors, which further increases the volatility of foreign shareholdings. From our more in-depth analysis of individual banks’ foreign shareholding, foreign ownership in the state-owned and joint-stock commercial banks has been distinctly lower and more stable than foreign shareholders’ participation in city commercial banks.3 This indicates more active speculative investment in city commercial banks, where government controls are weaker compared to large state commercial banks. 3.5 Model Selection Tests Before starting the GMM estimation, we perform correlation analysis to test for multicollinearity between different variables. From Table 10.6, the figure of concern is a relatively strong negative relationship between GDP and ROE (−0.7858). The decelerating GDP growth in China in recent years implies a decreasing trend of Chinese-listed commercial banks’ ROE, in line with the literature that documents declining profitability of Chinese banks during this period. The correlation coefficient associated with the relation between foreign shareholding and NPL aligns with what is predicted in Hypothesis 2. Coefficients related to liquidity ratio and coverage ratio are also in line with Hypothesis 3 and 4, respectively. Only the correlation coefficient between foreign ownership and ROE contradicts Hypothesis 1. There are several possible explanations for the potentially negative relationship between foreign investment and ROE of Chinese-listed commercial banks. There is a general decreasing trend in Chinese commercial bank’s ROE over the period. However, this poor performance might actually incentivize potential investment. There are three main types of domestic and foreign investor attracted to the sector. The first seeks strategic partnerships with well-performing banks in order to strengthen their technical capabilities and upgrade their global competitiveness. The second seeks strategic

NGUYEN 9781800880894 PRINT.indd 188

21/09/2022 13:14

The impact of foreign shareholders on Chinese commercial banks  ­189 Table 10.6  Pearson correlation coefficients

FS ROE NPL LR CAR CR RIR GDP BS

FS

ROE

NPL

LR

CAR

CR

RIR

GDP

BS

1 −0.2303 −0.0103 −0.382 0.0354 0.055 0.0935 0.2046 −0.4512

1 −0.6662 −0.4288 −0.3374 0.4852 −0.0142 −0.7858 0.0104

1 0.3838 0.1816 −0.7424 0.0679 0.6016 0.2803

1 0.3036 −0.3914 −0.1471 0.3895 0.3056

1 −0.0541 −0.0576 0.3447 0.4469

1 −0.0541 −0.2947 −0.2656

1 0.2696 0.0382

1 0.08

1

expansion into new business areas and segments. This strategy to some extent reduces the risks of developing a business in a different market; but the embedded risks are still high due to a lack of international experience, managerial capabilities and cultural differences. The third type of investor is the speculator. Speculators seek to acquire large amounts of bank assets and make profits on rising share prices. The last two types of investor are the most common in the Chinese listed commercial bank sector.4 All these investors bear long lags to integrate their strategies and align them with bank strategies and operations, as well as substantial opportunity costs. Therefore, decreasing ROE in Chinese listed commercial banks in the short to medium run presents investors with lower costs of entering investment positions, and can be viewed as subject to change for the better within their investment horizons. We test for multicollinearity using the Variance Inflation Factor (VIF) parameter. Overall, average VIF estimated for all variables is 2.74 and the VIF range across all variables was 1.38–4.42.5 Thus, all estimated VIF statistics were well below the critical value of 10, leading us to conclude that multicollinearity is not an issue in our data. The Breusch–Pagan test is used to test for heteroscedasticity. This showed existence of heteroscedasticity with the chi-square BP statistic estimated at 83.67, indicating violation of the classical assumption used in OLS models and supporting the application of GMM.6 Before using GMM, we perform endogeneity tests to decide whether instrument variables are necessary for the regression. There are three common explanations as to why endogeneity problems arise. First, explanatory variables that have an influence on the selected variables in the regression might have been omitted. Second, the dependent variable can be affected by one or several independent variables, which in turn are influenced by the dependent variable. Finally, measurement error in explanatory variables can also result in the error term being correlated with the explanatory variables. The Hausman test is commonly used to test for endogeneity; however, it is based on the absence of conditional heteroscedasticity. Since heteroscedasticity has been identified here using the Breusch–Pagan test, the alternative Durbin–Wu–Hausman test suggested by Davidson and MacKinnon (1993, pp. 237–240) has to be applied.7 Our test rejects the null hypothesis of no endogeneity, implying that instrumental variables should be used in our model.

NGUYEN 9781800880894 PRINT.indd 189

21/09/2022 13:14

190  Handbook of banking and finance in emerging markets As is common in studies of banking sector performance, we introduce macroeconomic factors to correct for endogeneity, based on Granger causality tests used to identify suitable instruments. The Granger causality test is performed in the presence of endogeneity in order to examine the causal relationship between independent variables and dependent variables. If the lagged values of dependent variables have a significant influence on the variation of independent variables, dependent variables Granger-cause independent variables. In the case of this study, foreign investments are having positive impacts on Chinese listed commercial banks. At the same time, better bank performance, such as higher profits and lower NPL ratios, are attracting more foreign partners to invest in these banks. The Granger causality test is based on the assumption that all variables are stationary, which means the statistical properties of related variables such as mean, variance and autocorrelation structure are constant over time. Therefore, a stationary test is performed first to check whether the panel data satisfies this condition. A Fisher-type unit-root test is applied as the panel dataset used in this study is short. The test results are available on request, but our p-values of four statistical indicators used in Fisher-type unit-root test for all four dependent variables are zero. This strongly rejects the null hypothesis that all panels contain unit roots, allowing us to conclude that there is no unit root in panel data, and that the entire data set is stationary. Following stationarity tests, we perform Vector Autoregression (VAR) on our panel data with multivariate time series under GMM. As the Granger causality test is extremely sensitive to the lag condition, VAR is designed to examine the bi-directionality between each variable and its own lags as well as those of other variables. The purpose of performing VAR here is to select the lag length at which the relation between the lagged value of independent variable and that of dependent variable is significant. First, VAR is performed for ROE and foreign shareholding. The result is that only after lag 7 can we detect a significant relation between ROE and foreign shareholding. In line with this, the first seven lagged values of variables are used as instruments in VAR GMM. Applying this to the Granger causality test, we establish evidence for Granger causality from foreign shareholding to ROE (p-value 0.005) that is stronger than any potential evidence of Granger causality from ROE to foreign shareholding (p-value 0.071). In the case of NPL ratio and foreign shareholding, six lags are significant in VAR application. The Granger causality test for NPL ratio and foreign shareholding generate a p-value of 0.007, rejecting the first null hypothesis that foreign shareholding does not Granger-cause NPL ratio. With a p-value of 0.245 failing to reject the second null hypothesis – that NPL ratio does not Granger-cause foreign shareholding  – foreign shareholding Granger-causes NPL ratio but NPL ratio does not Granger-cause foreign shareholding. It takes seven lags to show a significant correlation between loan-to-deposit ratio and foreign shareholding. Results for loan-to-deposit ratio and foreign shareholding shows evidence for Granger causality from foreign shareholding to loan-to-deposit ratio (p-value 0.000) that is stronger than evidence of Granger causality from loan-to-deposit ratio to foreign shareholding (p-value 0.062). For the last group of variables, it takes six lags to reach a significant relation between loan loss coverage ratio and foreign shareholding. Accordingly, the first six lagged values of variables are used as instruments in VAR. Results for loan loss coverage ratio and foreign shareholding show evidence for Granger causality from foreign shareholding to loan loss coverage ratio (p-value 0.006). This

NGUYEN 9781800880894 PRINT.indd 190

21/09/2022 13:14

The impact of foreign shareholders on Chinese commercial banks  ­191 causality is stronger than evidence of Granger causality from loan loss coverage ratio to foreign shareholding (p-value 0.014). In conclusion, there is only singular causality direction observed between foreign shareholding and the four selected performance indicators – ROE, NPL ratio, loan-to-deposit ratio and loan loss coverage ratio: foreign shareholders have an impact on the performance of Chinese listed commercial banks. On the other hand, the performance of Chinese listed commercial banks does not necessarily affect foreign investors’ ownership percentage in these banks. This singular (in direction) causality between foreign holdings and bank performance also eliminates the possibility of initial selection bias: foreign investors do not tend to invest in commercial banks with better performance. With this, we are warranted to look at the impact of foreign investment on the performance of Chinese listed commercial banks, and not vice versa. 3.6 Instrumental Variables and the Sargan–Hansen Test Overidentification of instrumental variable occurs when the number of endogenous variables is less than that of instrumental variables. The Sargan–Hansen test is applied to test for overidentification, and the results are available on request. Overall, it shows that the null hypothesis of overidentifying restrictions being valid can be rejected (p-values for both test components are 0.0000). Therefore, overidentifying restrictions do not stand, which means that at least one of the instrument variables are correlated with the error term. In this case, it is important to assess whether our instrumental variables are strongly correlated with the endogenous variable. To do this, we perform first stage regression. With a p-value equal to 0 for a robust F-test (with 3,215 degrees of freedom) of 39.87, we reject the null hypothesis of weak instruments, which indicates that our instruments are strong. We are now ready to specify our econometric models and proceed to estimation using GMM.

4. MODEL SPECIFICATION With foreign shareholding being the instrumented variable, instrumental variables are determined through ten regression runs using different combinations of possible instruments among GDP, real interest rate, bank size and capital adequacy ratio. In conclusion, including all these four instruments generates the most robust results; thus, our most valid regression model specifications are defined in Section 3.1. An F-test was performed to test for the fixed effect model, while Breusch-Pagan LM and Likelihood-Ratio (LR) tests were conducted to examine the random effect model. In order to obtain a more robust result, we divide both foreign shareholding and bank size into four quartiles, denoted as x_fs and x_bs. GDP is proxied by taking a natural log. The results are shown in Appendix Table 10A.1. A fixed effect model controls time-invariant characteristics within Chinese listed commercial banks that may create bias in the relationship between predictor variables and the outcome. These characteristics might be certain business practices or internal policies of the banks that have been influencing bank performance absent foreign shareholder participation. In  this sense, intuitively, the fixed effect model is more suitable for analyzing

NGUYEN 9781800880894 PRINT.indd 191

21/09/2022 13:14

192  Handbook of banking and finance in emerging markets the impact of foreign ownership on the performance of Chinese listed commercial banks over the course of ten years. However, this study is designed to draw overall conclusions on how foreign entry is affecting Chinese listed commercial banks’ performance using a representative sample of 28 listed commercial banks from three categories. In  this respect, the random effect model better fits our purpose. From a theoretical sense, it is difficult to choose one model over the other. Therefore, we form our opinions on selecting the right regression model based on empirical results. Finally, in the presence of heteroskedasticity, White correction is conducted under both fixed effect and random effect models. Analysis of results associated with the dependent variable ROE are shown in Table 10A.1. The fixed effects model is statistically significant at 1% level in the F-test, and the random effects model is also statistically significant at 1% level in both LM and LR tests. In addition, using the random effect model significantly decreases the sum of squares error (SSE). However, the explanatory variable has an insignificant influence on the dependent variable in both models. Results presented in Table 10A.1 for the dependent variable NPL ratio show that both fixed and random effects are statistically significant at the 1% level. The relationships between the explanatory variable and the dependent variable are significant at the 5% level under both models. As for liquidity ratio, both fixed and random effects are statistically significant at the 1% level. Under the fixed effect model, the relation between the explanatory variable and the dependent variable is not statistically significant; but with the random effect model the relationship is confirmed at the 1% significance level. The SSE in the random effect model is also much lower than in fixed effect model. From the empirical results in Table 10A.1 related to loan allowance coverage ratio, we obtain statistically significant fixed and random effects at the 1% level. Noticeably, using the random effect model significantly lowers SSE. However, the relationship between the explanatory variable and the dependent variable is not significant in both models. It appears at first sight that the data fit is better in the random effect model than in the fixed effect model. However, empirical results show that the relationships between foreign shareholding and ROE as well as loan loss coverage ratio are not statistically significant. This contradicts the findings in the established literature. We believe that insignificance coefficients arise due to the complex endogeneity effects that are left unresolved by the use of instrumental variables. To control for these, we are therefore warranted to use GMM with instrumental variables.

5. RESULTS 5.1 GMM Model Estimation Results The Durbin–Wu–Hausman test for linear regression of all suspected endogenous and instrumental variables was run in the first step regression.8 The relation between foreign shareholding and ROE as well as NPL ratio was not statistically significant, potentially due to significant endogeneity problems in the original data. When using instrumental variables GMM (IV-GMM) instead, the regression results improve substantially. To make the GMM model robust to heteroscedasticity, Newey–West correction has to be performed. The results using robust standard errors (SE) are shown in Table 10.7.

NGUYEN 9781800880894 PRINT.indd 192

21/09/2022 13:14

193

NGUYEN 9781800880894 PRINT.indd 193

21/09/2022 13:14

220 37.42 0 0.0087 0.00424 Robust FS CAR, RIR, BS, GDP

82.08 0 0.0093 0.04112 Robust

FS CAR, RIR, BS, GDP

−0.0022 0.8733 0.0015

220

0.000 0.000 0.000

Coefficient

Number of  observations Wald chi2 Prob > Chi2 R-Squared Root MSE GMM weight  matrix Instrumented: Instruments

−6.04 −7.13 15.6

p-value: P > |z|

−0.1464 −1.0897 0.3209

z

FS CAR Constant

Coefficient

ROE

−0.77 5.62 0.70

z

NPL

0.442 0.000 0.485

p-value: P > |z|

FS CAR, RIR, BS, GDP

84.44 0 0.2220 0.09299 Robust

220

−0.1648 2.4786 0.3038

Coefficient −2.77 7.59 6.89

z 0.006 0.000 0.000

p-value: P > |z|

Loan to deposit ratio

Dependent Variables

Table 10.7  IV-GMM results for foreign shareholding and dependent variables

FS CAR, RIR, BS, GDP

19.68 0.0001 0.1633 0.98635 Robust

220

2.1434 −4.9183 2.8118

Coefficient

3.24 −1.19 4.92

z

0.001 0.233 0.000

p-value: P > |z|

Loan loss coverage ratio

194  Handbook of banking and finance in emerging markets ROE is now significantly related to foreign shareholding; however, NPL ratio is still insignificantly correlated with foreign shareholding. One explanation for this might be the heterogeneity problem with foreign shareholding and bank size reflected in the descriptive statistics (Section 3.4). Therefore, both foreign shareholding and bank size are divided into four quartiles, denoted as x_fs and x_bs. The results of instrumental variables (GMM) regression using quartiles are thus more robust and are shown in Table 10.8. The Wald test is applied to test the significance of the explanatory variables in a statistical model (Polit, 1996). If the Wald test for a particular explanatory variable, or group of explanatory variables, is statistically significant, we can conclude that the parameters associated with these variables are non-zero. The variables should therefore be included in the model. Using quartiles for foreign shareholding and bank size, we can see a significant drop in the Wald chi-square statistic compared with that in the original regression, which indicates a better-specified model with a valid set of instruments.

6. DISCUSSION AND CONCLUSIONS This study aimed to more accurately assess the impact of foreign shareholders’ presence in Chinese commercial banking by measuring bank performance from various perspectives. In a major addition to the established literature, traditionally reliant on using ROA alone as a metric of Chinese banks’ performance, four performance indicators – namely ROE, NPL ratio, loan-to-deposit ratio and loan loss coverage ratio – are used here to capture performance. Instrumental variables (GDP, real interest rate, bank size and capital adequacy ratio) are identified and integrated into GMM models in order to produce more reliable empirical results. From the instrumental variables GMM regression results shown in Table 10.7, foreign shareholding is negatively correlated to ROE of Chinese listed commercial banks. Distinguishing banks by size into four quartiles significantly reduces robust SE, which makes the selected sample more representative of the population. A one-unit increase in foreign shareholding leads to an expected decrease in ROE by 0.0132 units. The relation between foreign shareholding and ROE is significant at the 1% level, significant from the financial market perspective, and this result aligns with the correlation analysis and our prior intuition. A second set of results in Table 10.7 captures the estimated relationship between foreign investors’ shareholdings and loan quality performance. Foreign ownership has a positive impact on NPL ratio, and this impact is significant at the 1% level. With bank size and foreign ownership data separated into quartiles, the SE falls substantially. For a one-unit increase in foreign shareholding, the NPL ratio is expected to decrease by 0.0009 units. This supports the findings in previous research (e.g., García-Herrero and Santabárbara, 2008), extending them to the 2010–2019 period, and, thus, to a more modern regulatory regime for foreign shareholdings in China’s banking sector. In the meantime, the liquidity of Chinese listed commercial banks reflected by the loanto-deposit ratio is negatively influenced by increasing foreign shareholdings (as shown in Table 10.7). For a one-unit increase in foreign shareholding, the loan-to-deposit ratio is expected to decrease by 0.0416 units. This relation is significant at the 1% level and is

NGUYEN 9781800880894 PRINT.indd 194

21/09/2022 13:14

195

NGUYEN 9781800880894 PRINT.indd 195

21/09/2022 13:14

220 36.87 0 0.056 0.00413 Robust x_FS CAR, RIR, x_BS, GDP

220

67.89 0 0.0525 0.03975 Robust

x_FS CAR, RIR, x_BS, GDP

−0.0009 0.0723 0.0054

Number of  observations Wald chi2 Prob > Chi2 R-Squared Root MSE GMM weight  matrix Instrumented: Instruments

0.000 0.000 0.000

−4.27 −7.48 14.95

−0.0132 −1.2183 0.3495

Coefficient

FS CAR Constant

p-value: P > |z|

z

ROE

−2.65  4.52  2.21

z

NPL

0.008 0.000 0.027

p-value: P > |z|

x_FS CAR, RIR, x_BS, GDP

80.74 0 0.2139 0.09347 Robust

220

−0.0417 2.0569 0.4371

Coefficient −5.42 5.8  8.31

z 0.006 0.000 0.000

p-value: P > |z|

Loan to deposit ratio

Dependent Variables

IV-GMM (quartile) results for foreign shareholding and dependent variables

Coefficient

Table 10.8 

x_FS CAR, RIR, x_BS, GDP

19.68 0.0001 0.1633 0.98635 Robust

220

0.2797 −0.0680 1.8100

Coefficient

 3.92 −0.02  3.32

z

0.000 0.986 0.001

p-value: P > |z|

Loan loss coverage ratio

196  Handbook of banking and finance in emerging markets significantly large from the financial market perspective. Similarly, SE decreases when we use quartiles for bank size and foreign shareholding. The final set of results in Table 10.7 show that the loan loss coverage ratio is positively associated with foreign shareholdings. A one-unit increase in foreign shareholding is associated with a loan loss coverage ratio increase of 0.2797 units. This relationship is significant at the 1% level and is large from the financial market perspective. We can deduce that Chinese listed commercial banks, in the presence of foreign shareholders, hold greater loan allowances to mitigate the risks that arise from the increasing amount of loans. From the foreign investors’ side, their financial capital injections also provide a source of loan allowances for Chinese listed commercial banks. The banking sector in China has long played a critical role in financing state-owned enterprises, supporting the fast-growing stock market and driving economic growth, especially as the Chinese economy is pivoting toward a more domestic demand-based economic model. In the mid-2000s Chinese commercial banks accumulated massive amounts of NPLs. Low ROE posed the risk of bad debt accumulating on these banks’ balance sheets. In light of these issues, the Chinese authorities decided to accelerate the process of opening up the banking sector to foreign investors in order to bring in more foreign capital, sector know-how and technological innovations. Lifting limits on foreign ownership in the financial sector and encouraging more banking sector IPOs formed the cornerstone of these reforms. Following these changes, we examined a set of bank performance indicators using comprehensive and robust research methodologies to answer one main question: what impact do foreign shareholders have on Chinese commercial banks’ key performance indicators? Four indicators were used in this study to capture bank performance – ROE, NPL ratio, loan-to-deposit ratio and loan loss coverage ratio. Considering endogeneity issues, we used a broad set of instrumental variables (log GDP, capital adequacy ratio, real interest rate and bank size). The results show that foreign participation has a significantly positive impact on liquidity and loan quality performance of Chinese listed commercial banks. However, foreign entry is found to have a significantly negative relation on bank profitability, as measured by ROE. There are three possible reasons for this, as discussed in Section 3. The first is the continued trend in decreasing Chinese listed commercial banks’ ROE that persisted over the last two decades, most recently accelerated by slower growth in the economy. The second is the increasing number of foreign speculative investors entering the Chinese banking sector. The last reason is that it takes time for domestic banks to learn from their foreign strategic investors, and that tangible benefits regarding profitability of the banks are some years away due to these lags. Going forward with the Chinese financial sector fully open to foreign investors beyond 2021, more investors are likely to invest in the mainland’s banking sector. More importantly, there will be greater management involvement for foreign investors compared with the earlier period (covered by this study), when they were only allowed to hold minority shares in Chinese commercial banks. Therefore, our findings overall hold a very positive view of the prospects of listed commercial banks’ future performance as foreign investments in these banks scale up.

NGUYEN 9781800880894 PRINT.indd 196

21/09/2022 13:14

The impact of foreign shareholders on Chinese commercial banks  ­197

NOTES 1. City commercial banks have experienced a higher NPL ratio increase in recent years due to stricter regulations imposed them (CEIC and BBVA, 2019). 2. The Postal Savings Bank of China (PSBC) – a large state-owned commercial bank – is not included in the sample due to its very short history of trading on the stock exchange (IPO in 2019). 3. Results are available from the authors on request. 4. A rising share of speculative investments can be confirmed by the high standard deviation of foreign shareholdings in the descriptive statistics shown in Table 10.5. 5. Full results are available on request. 6. Results are available on request. 7. The results of the tests are available from the authors on request. 8. Results are available on request.

REFERENCES Barrell, R., Davis, E.P., Karim, D. and Liadze, I. (2010). Bank regulation, property prices and early warning systems for banking crises in OECD countries. Journal of Banking and Finance, 34(9), 2255–2264. CEIC (2019a). China CN: Commercial bank: Loan-to-deposit ratio 2010–2018. Available at: https://www. ceicdata.com/en/china/banking-capital-adequacy-liquidity-and-credit-risk-ratio/cn-commercial-bank-loant​ o​d​e​p​o​s​it-ratio. CEIC (2019b). China CN: NPL provision: Commercial bank: Provision coverage ratio: Large 2014–2018. Available at: https://www.ceicdata.com/en/china/non-performing-loan-npl/cn-npl-provision-commercialbank-provision-coverage-ratio-large. CEIC (2020). China non performing loans ratio. Available at: https://www.ceicdata.com/en/indicator/china/ non-performing-loans-ratio. CEIE and BBVA (2019). China banking monitor. Available at: https://externalcontent.blob.core.windows.net/ pdfs/China-Banking-Monitor_EDI-1-1.pdf. Chan, J., Choi, G., Leung, K. and Chia, S. (2014). Future directions for foreign banks in China. Journal of Financial Perspectives, 2(2). Available at: https://ssrn.com/abstract=3079411. Chi, Q.W. and Li, W.J. (2017). Economic policy uncertainty, credit risks and banks’ lending decisions: Evidence from Chinese commercial banks. China Journal of Accounting Research, 10(1), 33–50. Davidson, R. and MacKinnon, J.G. (1993). Estimation and Inference in Econometrics. New York: Oxford University Press. Deng, Z.L. (2013). Foreign Direct Investment in China: Spillover Effects on Domestic Enterprises. New York: Routledge. EY (2019). China further opens up financial sector: A compilation of EY POVs (I). China. García-Herrero, A. and Santabárbara, D. (2008). Is the Chinese banking system benefiting from foreign investors? Banco Bilbao Vizcaya Argentaria (BBVA) Economic Research Department Working Paper. No.0804. Available at: https://www.bbvaresearch.com/wp-content/uploads/mult/WP_0804_tcm348-212967.pdf. Haber, S. and Musacchio, A. (2005). Foreign banks and the Mexican economy, 1997–2004. Stanford Centre for International Development Working Paper No. 267. Available at: https://siepr.stanford.edu/sites/default/files/ publications/267wp.pdf. Hachem, K. (2018). Shadow banking in China. Annual Review of Financial Economics, 10, 287–308. Available at: https://www.annualreviews.org/doi/pdf/10.1146/annurev-financial-110217-023025. Ho, S. and Hsu, S.C. (2010). Leverage, performance and capital adequacy ratio in Taiwan’s banking industry. Japan and the World Economy, 22(4), 264–272. Hope, N.C. and Hu, F. (2006). Reforming China’s banking system: How much can foreign strategic investment help? Stanford Centre for International Development Working Paper No. 276. Available at: https://kingcenter. stanford.edu/sites/default/files/publications/276wp.pdf. Hope, N.C., Laurenceson, J., and Qin, F.M. (2008). The impact of direct investment by foreign banks on China’s banking industry. Stanford Centre for International Development Working Paper No. 362. Available at: https:// kingcenter.stanford.edu/publications/impact-direct-investment-foreign-banks-chinas-banking-industry. Huang, Y.P. and Wang, X. (2018). China’s 40 years of reform and development: 1978–2018. Canberra: Australian National University Press. Huizinga, H. and Laeven, L. (2019). The procyclicality of banking: Evidence from the euro area. European Central Bank Discussion Paper No. 2019–001. Available at: https://ssrn.com/abstract=3379641.

NGUYEN 9781800880894 PRINT.indd 197

21/09/2022 13:14

198  Handbook of banking and finance in emerging markets International Monetary Fund (IMF) (2019). Financial Development Index Database. Available at: https://data. imf.org/?sk=F8032E80-B36C-43B1-AC26-493C5B1CD33B. Ismail, Ç. and Burak, Ç. (2007). The economic determinants of foreign direct investment in developing countries and transition economies. Pakistan Development Review, 3, 285–299. Jiang, C. and Yao, S. (2017). Chinese Banking Reform: From the Pre-WTO Period to the Financial Crisis and Beyond. Nottingham China Policy Institute Series. Cham: Palgrave Macmillan. Khan, H. and Khan, U. (2019). Financial development and FDI inflows in China. Economics Discussion Papers, No 2019-54, Kiel Institute for the World Economy. Available at: http://www.economics-ejournal.org/ economics/discussionpapers/2019-54. Laeven, L. and Majnoni, G. (2003). Loan loss provisioning and economic slowdowns: Too much, too late? Journal of Financial Intermediation, 12(2), 178–197. Available at: https://doi.org/10.1016/S1042-9573(03)00016-0. Leigh, L.Y. and Podpiera, R. (2006). The rise of foreign investment in China’s banks: Taking stock. International Monetary Fund Working Paper WP/06/292. Available at: https://www.imf.org/en/Publications/WP/ Issues/2016/12/31/The-Rise-of-Foreign-Investment-in-Chinas-Banks-Taking-Stock-20152. Li, R., Li, X., Lei, W., and Huang, Y.P. (2015). Consequences of China’s opening to foreign banks. In Song, L.G., Garnaut, R., Fang, C., and Johnston, L. (Eds.), China’s Domestic Transformation in a Global Context. Acton: Australian National University Press. Available at: https://www.jstor.org/stable/j.ctt16wd0dw. Luo, D. (2016). The Development of the Chinese Financial System and Reform of Chinese Commercial Banks. London: Palgrave Macmillan. Luo, D., Dong, Y., Armitage, S. and Hou, W. (2017). The impact of foreign bank penetration on the domestic banking sector: New evidence from China. European Journal of Finance, 23 (7–9), 752–780. Available at: doi:​ 10.1080/1351847X.2014.1003314. Park, J.S., Suh, J.H. and Hahm, J.H. (2010). Global financial crisis and business strategies of Korean commercial banks: Evaluation and implications. Korea and the World Economy, 28(4), 251–307. Podpiera, R. (2006). Progress in China’s banking sector reform: Has bank behaviour changed? International Monetary Fund Working Paper WP/06/71. Available at: https://www.imf.org/en/Publications/WP/ Issues/2016/12/31/Progress-in-Chinas-Banking-Sector-Reform-Has-Bank-Behavior-Changed-18952. Polit, D.F. (1996). Data Analysis and Statistics for Nursing Research, Harlow: Pearson Education. Polizzi, S., Scannella, E. and Suarez, N. (2020). The role of capital and liquidity in bank lending: Are banks safer? Global Policy, 11(S1), 28–38. Available at: https://onlinelibrary.wiley.com/doi/epdf/10.1111/17585899.12750. Pradhan, R.P., Arvin, M.B., Hall, J.H., and Nair, M. (2017). Trade openness, foreign direct investment, and finance-growth nexus in the Eurozone countries. Journal of International Trade & Economic Development, 26(3), 336–360. PwC (2018). 2019: A sunnier outlook for international China NPL investors. Available at: https://www.pwccn. com/en/business-recovery-services/2019-sunnier-outlook-international-china-npl-investors.pdf. Svirydzenka, K. (2016). Introducing a new broad-based index of financial development. International Monetary Fund Working Paper WP/16/5. Available at: https://www.imf.org/external/pubs/ft/wp/2016/wp1605.pdf. Tan, H.L. (2017). Chinese shadow banking has slowed – but that’s not as good as it seems. Available at: https:// www.cnbc.com/2017/11/19/china-shadow-banking-has-slowed-chinese-debt-may-just-go-elsewhere.html. Turner, G., Tan, N. and Sadeghian, D. (2012). The Chinese banking system. Reserve Bank of Australia. Available at: https://www.rba.gov.au/publications/bulletin/2012/sep/pdf/bu-0912-7.pdf. Tzu, Lao (n.d.). BrainyQuote.com. Retrieved July 18, 2021, from https://www.brainyquote.com/quotes/lao_ tzu_137141. World Bank (2019a). Market capitalization of listed domestic companies (% of GDP) – China. Available at: https://data.worldbank.org/indicator/FM.AST.NFRG.CN?locations=CN. World Bank (2019b). Net foreign assets (current LCU) – China. Available at: https://data.worldbank.org/indicat​ o​r​/CM.MKT.LCAP.GD.ZS?locations=CN. Yao, K. (2018). China pledges to allow more foreign investment in financial sector by year-end. Available at: https://www.reuters.com/article/us-china-boao-pboc/china-pledges-to-allow-more-foreign-investment-in-fin​ a​ncial-sector-by-year-end-idUSKBN1HI074. Yao, S.J., Han, Z.W., and Feng, G.F. (2008). Ownership reform, foreign competition and efficiency of Chinese commercial banks: A non-parametric approach. World Economy, 31(10), 1310–1326. Available at: http:// dx.doi.org/10.1111/j.1467-9701.2008.01130x.

NGUYEN 9781800880894 PRINT.indd 198

21/09/2022 13:14

199

NGUYEN 9781800880894 PRINT.indd 199

21/09/2022 13:14

220

6.56**

214 0.7554 0.7497 0.0626 0.0183

0.0047 [0.0031] −0.1311 [0.0135] −0.5387** [0.1730] 0.5093** [0.0701] 0.0038 [0.0053] 1.6680** [0.1293] 108.56**

96.99** 59.51** 220

0.0088 0.0183 0.0187 0.7048

526.95** 214 0.7547

0.0042 [0.002] −0.1362** [0.0112] −0.4752** [0.1595] 0.5249** [0.0626] 0.006 [0.0037] 1.7713** [0.1094]

RE Model

220

3.53**

31.13** 23.28** 220

0.0001 0.0028 0.0018 0.5600

118.79** 214 0.5065

−0.0009* [0.0004] 0.0120** [0.0019] -0.0147 [0.0187] −0.0331** [0.0090] 0.0002 [0.0005] −0.1182** [0.0196]

−0.0013* [0.0006] 0.0113** [0.0023] −0.0141 [0.0243] −0.0301** [0.0100] 0.0005 [0.0009] −0.1096** [0.0228] 38.86** 214 0.5096 0.4981 0.0014 0.0028

RE Model

FE Model

NPL

Note:  Standard errors in square brackets. Statistical significance: * 0.05). Therefore, we cannot

NGUYEN 9781800880894 PRINT.indd 234

21/09/2022 13:14

Determinants of commercial banks’ performance in Mozambique  ­235 validate the hypothesis initially defined in this study, as in Petria et al. (2015), Yakubu (2016) and Alper and Anbar (2011), who concluded that macroeconomic effects are not relevant in explaining bank performance. Our tenth hypothesis (H10) related inflation to performance. The results indicate that the inflation variable positively influences performance as measured by ROA (β10 = 0.110; p-value < 0.05). This supports H10 and is in line with Karim et al. (2010), Kosmidou et al. (2008), Vong and Chan (2009), and Alper and Anbar (2011). If an inflation rate is anticipated, banks can adjust their interest rates to increase profits, thus improving performance. Our final hypothesis (H11) related interest rate to bank performance. According to the results, the hypothesis is not confirmed because the interest rate variable was not statistically significant for either model – ROA (β11 = 0.074; p-value > 0.05) and ROE (β11 = −0.030; p-value > 0.05). Therefore, it cannot be confirmed that commercial banks increase profitability with rising interest rates, thus contributing to a positive relationship between interest rate and profitability. This result was also confirmed by Aboagye (2012), Alper and Anbar (2011) and Yakubu (2016), who point out that interest rate is not relevant in explaining bank performance. Also, Massarongo (2013) argues that the interest rates set by Mozambican commercial banks do not proportionately follow the policy rates set by the BoM, since commercial banks are reluctant to reduce rates due to uncertainty about the country’s economic environment. In short, Mozambican banks do not react to changes in BoM policy rates.

6. CONCLUSIONS The objective of this study was to evaluate the main drivers of Mozambican banking performance in the period from 2008 to 2016. The results show that the variables with the greatest influence on return on assets (ROA) were capital adequacy, liquidity risk, efficiency, diversification of activities and inflation rate. Regarding return on equity (ROE), the variables that exhibited the highest statistical significance were asset quality measured by credit default (QACIN), diversification and concentration of bank activity. The variables size, level of deposit, interest rate and economic growth rate showed no significant impact on either ROA or ROE, so we did not obtain evidence of their effect on the performance of the banking sector in Mozambique. The results of this study offer important practical and academic implications. First, this work enhances the literature on banking performance in Mozambique, a context still relatively little studied, introducing in the analysis some variables (e.g. bank concentration, degree of diversification) that had not been considered in the works of Wanke et al. (2016a, 2016b) and Gil-Alana et al. (2017). Second, the results may serve to assist bank managers and the BoM in the formulation and implementation of effective strategies, as well as policy regulation, to improve the performance of commercial banks in Mozambique.

NGUYEN 9781800880894 PRINT.indd 235

21/09/2022 13:14

236  Handbook of banking and finance in emerging markets

REFERENCES Aboagye, A.Q.Q. (2012). Bank concentration and economic costs of deposit mobilization and credit extension in Ghana. Journal of Developing Areas, 46(2), 351–370. Abreu, M., Afonso, A., Escaria, V., & Ferreira, C. (2012). Economia Monetária e Financeira, 2nd edn. Lisbon: Escolar Editora. Acaravci, S.K. & Çalim, A.E. (2013). Turkish banking sectors profitability factors. International Journal of Economics and Financial Issues, 3(1), 27–41. Adeusi, S.O., Kolapo, F. & Aluko, A.O. (2014). Determinants of commercial banks’ profitability: Evidence from Nigeria. International Journal of Economics, Commerce and Management, 2(12). https://www.researchgate. net/publication/320518670. Alexiou, C. & Sofoklis, V. (2009). Determinants of bank profitability: Evidence from the Greek banking sector. Economic Annals, 54(182), 93–118. doi:10.2298/EKA0982093. Alkhazaleh, A.M. & Almsafir, M. (2014). Bank specific determinants of profitability in Jordan. Journal of Advanced Social Research, 4(10), 1–20. Alper, D. & Anbar, A. (2011). Bank Specific and Macroeconomic Determinants of Commercial Bank Profitability: Empirical Evidence from Turkey. Business & Economic Research Journal, 2(2), 139–152. Ameur, I.G.B. & Mhiri, S.M. (2013). Explanatory factors of bank performance: Evidence from Tunisia. International Journal of Economics, Finance and Management, 2(1), 143–152. Athanasoglou, P.P., Brissimis, S.N., & Delis, M.D. (2008). Bank-specific, industry-specific and macroeconomic determinants of bank profitability. Journal of International Financial Markets, Institutions and Money, 18(2), 121–136. Ben Naceur, S. & Goiaed, M. (2008). The determinants of commercial bank inters margin and profitability: Evidence from Tunisia. Frontiers in Finance and Economics, 5(1), 106–130. Berger, A.N. (2005). The relationship between capital and earnings in banking. Journal of Money, Credit and Banking, 27(2), 432–456. doi:10.2307/2077877. Curak, M., Poposki, K., & Pepur, S. (2012). Profitability Determinants of the Macedonian Banking Sector in Changing Environment. Procedia: Social and Behavioral Sciences, 44, 406–416. de Faro, C. (ed.) (2014). Administração bancária: Uma visão aplicada. Rio de Janeiro: Editora FGV. De Young, R. & Rice, T. (2004). Non-interest income and financial performance at U.S. commercial banks. Financial Review, 39, 101–127. Dietrich, A. & Wanzenried, G. (2011). Determinants of bank profitability before and during the crisis: Evidence from Switzerland. Journal of International Financial Markets, Institutions and Money, 21 (3), 307–327. Francis, M.E. (2013). Determinants of commercial bank profitability in Sub-Saharan Africa. International Journal of Economic and Finance, 5(9), 134–143. Gil-Alana, A.L., Barros, C. & Mandlhaze, D. (2017). A performance assessment of Mozambique banquets: A Bayesian stochastic frontier. Applied Economics, 49(45), 4579–4587. doi.org/10.1080/0036846.2017.1287857. Goddard, J., Molyneu, P. & Wilson, J. (2004). Dynamics of growth and profitability in banking. Journal of Money Credit and Banking, 36(3), 1069–1090. Guru, B.K., Staunton, J., & Shanmugam, B. (2002). Determinants of commercial bank profitability in Malaysia. Journal of Money, Credit, and Banking, 17, 69–82. Hassan, M.K. & Bashir, A.-H.M. (2005). Determinants of Islamic banking profitability. ERF paper. doi:10.3366/ edinburgh/9780748621002.003.0008. Kapaya, S.M. & Raphael, G. (2016). Bank-specific, industry-specific and macroeconomic determinants of bank profitability: Empirical evidence from Tanzania. International Finance and Banking, 3(2), 100–119. Karim, B.K., Sami, B.A.M. & Hichem, B.-K. (2010). Bank-specific, industry-specific and macroeconomic determinants of African Islamic banks profitability. International Journal of Business & Management Science, 3(1), 39–56. Kosmidou, K., Tanna, S. & Pasiouras, F. (2008). Determinants of profitability of domestic UK commercial banks: Panel evidence from the period 1995–2002. Economics, Finance and Accounting Applied Research Working Paper Series no. RP08-4, Coventry University. KPMG (2008). Pesquisa Sobre Sector Bancário Moçambicano [Banking survey financial services Mozambique]. Maputo: KMPG/Associação Moçambicana de Bancos. KPMG (2009). Pesquisa Sobre Sector Bancário Moçambicano. Maputo: KMPG/Associação Moçambicana de Bancos. KPMG (2012). Pesquisa Sobre Sector Bancário Moçambicano. Maputo: KMPG/Associação Moçambicana de Bancos. KPMG (2014). Pesquisa Sobre Sector Bancário Moçambicano. Maputo: KMPG/Associação Moçambicana de Bancos.

NGUYEN 9781800880894 PRINT.indd 236

21/09/2022 13:14

Determinants of commercial banks’ performance in Mozambique  ­237 KPMG (2015). Pesquisa Sobre Sector Bancário Moçambicano. Maputo: KMPG/Associação Moçambicana de Bancos. KPMG (2016). Pesquisa Sobre Sector Bancário Moçambicano. Maputo: KMPG/Associação Moçambicana de Bancos. Levine, R. (1997). Financial development and economic growth: Views and agenda. Journal of Economic Literature, 35(2), 688–726. Li, D.L. (2013), Os determinantes da lucratividade dos bancos. Thesis, Universidade da Brasília. Marôco, J. (2018). Análise Estatística com o SPSS Statistics, 7th edn, Pêro Pinheiro, Portugal: ReportNumber Massarongo, F. (2013). Porque é que os bancos comerciais não respondem à redução das taxas de referência: Reflexões. In L. de Brito, C.N. Castel-Branco, S. Chichava & A. Francisco (eds), Desafio para Moçambique. Maputo: IESE, 149–173. Mehta, A. & Bhavani, G. (2017). What determines banks’ profitability? Evidence from emerging markets: The case of the UAE banking sector. Accounting and finance Research, 6(1), 77–88. Ongore, V.O. & Kusa, G.B. (2013). Determinants of financial performance of commercial banks in Kenya. International Journal of Economics and Financial Issues, 3(1), 237–252. Onuonga, S.M. (2014). The analysis of profitability of Kenya’s top six commercial banks: Internal factor analysis. American International Journal of Social Science, 3(5), 94–103. Petria, N., Capraru, B., & Ihnatov, I. (2015). Determinants of banks’ profitability: Evidence from EU 27 banking systems. Procedia Economics and Finance, 20, 518–524. Sufian, F. & Chong, R.R. (2008). Determinants of bank profitability in a developing economy: Empirical evidence from the Philippines. Asian Academy of Management Journal of Accounting and Finance, 42(2), 91–112. Sufian, F. & Habibullah, M.S. (2009). Determinants of bank profitability in a development economy: Empirical evidence from Bangladesh. Journal of Business Economics and Management, 10(3), 207–217. Sufian, F. & Kamadurin, F. (2012). Bank-specific and macroeconomic determinants of profitability of Bangladesh’s commercial banks. Bangladesh Development Studies, 35(4), 1–28. Trujillo-Ponce, A. (2013). What determines the profitability of bank? Evidence from Spain. Accounting and Finance, 53(2), 561–586. Vong, P.I. & Chan, H.S. (2009). Determinants of bank profitability in Macau. Macau Monetary Research Bulletin, 12, 93–113. Wanke, P., Barros, P.C. & Emrouznejad, A. (2016a). Assessing productive efficiency of bank using integrates fuzzy-DEA and bootstrapping: A case of Mozambican banks. European Journal of Operation Research, 249, 378–389. Wanke, P., Barros, P.C., Azad, A.K. & Constantino, D. (2016b). The development of the Mozambican banking sector and strategic fit of mergers and acquisitions: A two-stage DEA approach. African Development Review, 28(4), 444–461. World Bank (1992). Mozambique: Financial sector study. Washington, DC: World Bank Yakubu, I.N. (2016). Bank-specific and macroeconomic determinants of commercial bank profitability in Ghana. International Finance and Banking, 3(2), 89–99.

NGUYEN 9781800880894 PRINT.indd 237

21/09/2022 13:14

Vietnamese experience

13. Impact of information and communication technology on banking efficiency: the Vietnamese experience Thanh Ngo and Tu Le

1. INTRODUCTION It is undeniable that, along with changes in deregulation, information and communication information (ICT) has significantly altered the competitive environment and strategies faced by financial institutions. Advanced technology has revolutionized back-office processes, front-office delivery systems to customers and payment systems (Humphrey et al., 2006). Therefore, these allow financial firms to expand into geographic and product markets and further promote the consolidation of the global financial market (DeYoung et al., 2009). The impact of ICT on improving performance in the banking industry is examined in several studies (Appiahene et al., 2019; Dinçer & Yüksel, 2020; Scott et al., 2017). As outlined in the study by Le and Ngo (2020), this literature can be divided into two strands. The first strand, which looks at the effect of investments in infrastructure or technological developments on bank profitability and productivity, provides mixed findings.1 Some suggest that technological progress could reduce bank costs (Berger & DeYoung, 2006), improve technical efficiency (Salim et al., 2010) and increase profits (Berger, 2003; Girmaye, 2018). However, several studies show that a reduction in bank profitability is associated with the adoption and diffusion of information technology (IT) investment (Arora & Arora, 2013; Ho & Mallick, 2010). These show that although the more internetbased banks may have the advantage of lower physical overhead costs (DeYoung, 2001), there is little relationship between IT investment and enhanced bank efficiency/profitability (Beccalli, 2007). The second strand focuses on IT-based methods of service delivery and bank performance. The use of electronic banking applications (apps) in service delivery is generally perceived as having a positive impact on bank performance (Ciciretti et al., 2009; Hernando & Nieto, 2007; Robert, 2001; Weigelt & Sarkar, 2012); however, there is continuing disagreement on the impact of Automated Teller Machines (ATMs). Several studies emphasize that ATMs impact bank profitability positively (Holden & El-Bannany, 2004; Le & Ngo, 2020) as, together with electronic payments, ATMs could reduce total operating costs over total assets by more than 30% (Valverde & Humphrey, 2009). Intensive investment in ATMs, however, may reduce the technical efficiency of Indian banks, suggesting that this expense may outweigh the benefits (Sathye & Sathye, 2017). Empirical studies on the impact of ICT on the banking system are primarily based on the US and other developed markets (e.g. Beccalli, 2007; Ho & Mallick, 2010; Le & Ngo, 2020), with much less insight and discussion on the sector in emerging economies. When accounting for the size and effect of emerging markets such as Vietnam on the global 238

NGUYEN 9781800880894 PRINT.indd 238

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­239 economy, there is a gap in the banking literature: no study attempts to investigate the relationship between ICT and banking efficiency in Vietnam. Since becoming a member of the World Trade Organization (WTO) in 2007, Vietnam has been recognized as one of the fastest-growing economies in the world, with an average annual economic growth rate of 6.25% from 2007 to 2019. Because its capital market is not yet developed, the Vietnamese banking system is the backbone of the economy (Le, 2019). Therefore, the efficiency of the country’s banking system has received much attention from academics, bank managers and policymakers. As an early adopter of technology, the Vietnamese banking system has witnessed an increasing trend in ICT investments to improve and maintain efficiency over time (this will be further explained in the next section). This therefore supports the view of the Vietnamese government that ICT investment is one of the main keys to enhancing the efficiency and effectiveness of all economic sectors. However, empirical findings are mixed, which raises the question of whether ICT would improve the efficiency of the Vietnamese banking system. Our study contributes to the literature in several ways. Firstly, it is noted that the literature on the relationship between ICT and bank performance is dominated by studies in developed countries where there are large markets, large numbers of banks and the availability of ICT data. There is limited evidence on the impact of ICT on bank efficiency in the emerging markets, especially in the Asia-Pacific region. This study is the first attempt to examine the impact of ICT on bank efficiency in Vietnam, an emerging economy. The findings in this study would therefore add more evidence to the literature, especially regarding emerging markets, and therefore provide a better understanding of the impact of ICT in the banking industry. Policy measures drawn from this study could also be implemented in the decision-making process about promoting ICT investments in other developing countries with banking structures similar to those in Vietnam. Secondly, in contrast to previous studies where ICT is measured by total expenditure on hardware and software, we use a more comprehensive measure of ICT as an index that accounts for IT infrastructure investment, human-related IT investment, IT applications, and strategies and policies to implement and develop ICT at each commercial bank in Vietnam (MIC, 2020). The use of the ICT index is therefore expected to provide an overall examination of the ICT–bank efficiency relationship in the Vietnamese sector. Thirdly, this study is an extension of the Euclidean common set of weights (ECSW) data envelopment analysis (DEA) approach (Hammami et al., 2020), where the efficiency and performance of Vietnamese banks are not only evaluated under the ‘law of one price’ but their determinants, especially ICT development, are also examined. The rest of the chapter is constructed as follows. Section 2 presents a brief overview of the Vietnamese banking system in the past decades. Section 3 describes the ECSW method as well as the data and variables of interest involved. The empirical results and discussions are consequently presented in Section 4, while Section 5 concludes.

2. BRIEF OVERVIEW OF THE VIETNAMESE BANKING SYSTEM The Vietnamese banking system has experienced several reforms over the years. The ‘Doi Moi’ (Renovation Reform) in the 1980s is considered one of the most important.

NGUYEN 9781800880894 PRINT.indd 239

21/09/2022 13:14

240  Handbook of banking and finance in emerging markets It transformed the banking industry from a one-tier into a two-tier system in which the State Bank of Vietnam acts as a true central bank, and state-owned commercial banks  (SOCBs) and privately owned commercial banks (POCBs) are responsible for commercial banking functions (Ngo, 2012). Thereafter, several reforms were implemented to further promote banks’ market functioning and efficiency. The primary focus of these reforms was on restructuring the SOCBs, especially in financing state-owned businesses (Nguyen et al., 2018). The structure of the remaining banks is relatively diversified. While two policy banks advance socio-political projects using funds from both domestic and international resources, POCBs provide universal banking products and services in different areas (or even nationwide), and are therefore seen as the most market-oriented. Due to their active operations, POCBs have gradually increased their market share of both deposits and lending (Ngo, 2012; Nguyen et al., 2016, 2018). More importantly, more foreign banks have penetrated the Vietnamese banking industry since it joined the WTO in 2007 (Vu & Turnell, 2010). The greater presence of foreign banks with advanced technology or know-how and better management put further pressure on domestic banks, in both the deposits and loans markets, and thus may affect their profitability. This may force local banks to operate more efficiently, perhaps via applying more technology to daily operations to reduce costs and communicate better with their customers. To further encourage sustainability, the 2011–2020 Socio-Economic Development Strategy was proposed, with its strong commitment to market-led development and modernization (Vietnamese Government, 2012b). In addition to promoting skills, enhancing market institutions and continuing infrastructure investment, innovation must be focused on to achieve higher market development. The subsequent Directive No. 16/CT-TTg further required ministries as well as other governmental agencies to focus on directing and organizing effective implementation of ICT to all economic sectors in all regions and provinces (Vietnamese Government, 2017). It is acknowledged that ICT can enable multiple aspects of firms’ operations, allowing them to interact and conduct business activities more efficiently (Chege et al., 2020; Gërguri-Rashiti et al., 2017; Scott et al., 2017). ICT also helps resolve distance issues and time constraints, and thus improves the efficiency of economic business and financial activities (Alshubiri et al., 2019). More importantly, ICT was adopted by the Vietnamese banking system as early as 2007, as shown in Figure 13.1. It is noted that training expenses for ICT per employee on average were generally high at the beginning but gradually fell later. In contrast, there are fluctuations in IT investment per employee; it reached a peak in 2008, decreased over the period 2008 to 2015, but then started increasing before dropping again in 2019. When observing the IT infrastructure security investment per employee, a relatively small fluctuation appears. As expected, ICT could help reduce the costs of banking intermediation services by making transactions faster and more accurate, enhancing the quality of services. It could also increase the flexibility of bank branch activities along with bank risk detection, and thus improve operational safety. Not least, ICT could improve the collection of data which is then used to determine the creditworthiness of borrowers. All in all, ICT could help banks increase their competitiveness (Appiahene et al., 2019; Le & Ngo, 2020; Scott et al., 2017). This study is the first to examine whether ICT could improve efficiency in the Vietnamese banking system.

NGUYEN 9781800880894 PRINT.indd 240

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­241 35

IT investment per employee on average (millon VND) IT infrastructure security investment per employee on averange (million VND)

30

Training expenses for ICT per employee on average (million VND)

25 20 15 10 5 0

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

Note:  Data on IT infrastructure security investment per employee is not available prior to 2011. Source:  The Vietnam ICT Index report series (MIC, 2020).

Figure 13.1  ICT investment by Vietnamese banks, 2007–2019

3.  METHODOLOGY AND DATA 3.1  Measuring the Performance of Vietnamese Banks Data Envelopment Analysis (DEA) is a nonparametric approach to measure the relative efficiency of a group of firms (Charnes et al., 1978), and is one of the most popular tools used to estimate efficiency and performance in the banking sector (Liu et al., 2013). One important reason is because DEA can examine the efficiency of banks using input and output quantities only, without a priori production function (Nguyen et al., 2019), which in turn is difficult to define in the banking (service) sector. Moreover, unlike other parametric approaches, such as regression or Stochastic Frontier Analysis, DEA is flexible in dealing with small sample sizes (Ngo & Le, 2019) – studies often encountered countries with fewer than 30 banks.2 The foundation of DEA is from the Production Possibility Frontier (PPF), where all the points on the PPF are recognized as optimal outputs of production, while the others are regarded as inefficient. A simple PPF can be constructed as the Figure 13.2, assuming that one can produce only one output X1, only one output X2, or a combination of the two. It is then observed from Figure 13.2 that points B, C and E are the efficient and feasible production points, while A and D are inefficient. Under the same principle, given the five production observations (A, B, C, D and E), DEA can envelop the PPF based on the three observations B, C and E  – those observations are efficient with 100% efficiency

NGUYEN 9781800880894 PRINT.indd 241

21/09/2022 13:14

242  Handbook of banking and finance in emerging markets Effciency

X2

A’

B

Production Possibility Frontier (PPF)

C

A

D’ D

E Ineffciency

0

X1

Figure 13.2  The DEA frontier scores. For the inefficient observation D, the distance 0D represents how efficient the production point D is compared to the PPF, or precisely to its projection point D’ on the PPF. Accordingly, the efficiency score of observation D can be calculated as 0D/0D’ (which is less than 100%). The purpose of DEA is to envelop the ‘best practice frontier’ from the sampled banks and then compare those banks to that frontier. In a multiple input and output setting, DEA assigns the optimal weights for those input/output quantities of a certain bank as those that can bring the bank closest to the frontier. The mathematical expression of DEA is

(13.1)

Subject to

≤ 1, ∀j, j = 1, 2, .., n ur νi ≥ ε, ∀i, r

where EFj ) is the efficiency score of the bank j0 (j=1,2, …,n); ur is the weight assigned 0 to  the output yr (r=1,2,…,m) of the bank j0; vi is the weight assigned to the input xi (i=1,2,…,k) of the bank j0; n is the total number of banks examined; and ε is a nonArchimedean value designed to enforce positivity on the weights.3 Note that the first constraint of Equation (13.1) is assurance that the set of weights (ur, vi) of bank j0 can also be a feasible set of weights for all other banks without making those banks’ efficiency score greater than unity, which contradicts the definition of efficiency score relevant to the discussion on Figure 13.1 above. Note again that, if available, one can use real input and output prices instead of ur and vi, and Equation (13.1) accordingly turns into a profitability ratio measuring the bank’s total sales divided by its total costs. Since ur and vi are estimations of real prices, they are also called ‘shadow prices’. It is noted that Equation (13.1) needs to be solved n times for n banks involved in the sample  – each time a set of weights or shadow prices is determined for bank j being

NGUYEN 9781800880894 PRINT.indd 242

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­243 examined. This means that a bank’s efficiency calculations using Equation (13.1) depend on different sets of shadow prices, which may be inappropriate from the managerial perspective (Hammami et al., 2020), especially in terms of the ‘law of one price’ regarding those shadow prices (Mankiw, 2020). For example, one may argue that efficiency of observation A in Figure 13.2 is more dependent on the output X2, while the efficiency of observation D is more dependent on X1. It is more problematic when more inputs and outputs are involved in Equation (13.1). As such, several studies suggest that it would be better to compare those banks using a common set of weights instead of n different sets of weights as in traditional DEA (see e.g. Hammami et al., 2020; Kao & Hung, 2005; Liu & Peng, 2008; Roll et al., 1991). Hammami et al. introduced the Euclidean common set of weights, which projects the banks by the shortest Euclidean distance to the frontier. Since the ECSW outperforms other common sets of weights, we therefore follow this approach in estimating Vietnamese banking efficiency. The basic procedure of ECSW is presented below.4 Step 1 Use Equation (13.1) to calculate the (different) set of weights (Uj*,Vj*) for each bank in the sample as normal, where Uj* = {u1*,u2*,..,ur*,..,um*} are the weights for the m outputs, and Vj* = {v1*,v2*,..,vr*,..,vs* } are the weights for the s inputs of bank j. Note that Hammami et al. (2020) used the constant returns to scale DEA (Charnes et al., 1978) in their analysis; however, other DEA models – such as the variable returns to scale DEA (Banker et al., 1984), the super-efficiency DEA (Andersen & Petersen, 1993) or the slack-based measure DEA (Tone, 2001) – can also be used in this step. Here, we extend Hammami et al.’s model by using the slack-based measure DEA since it can simultaneously account for both input minimization and output maximization, such things being very important to Vietnamese banks (Ngo, 2012; Ngo & Tripe, 2017; Nguyen et al., 2018). The linear programming problem for the slack-based measure DEA is



subject to





(13.2)







NGUYEN 9781800880894 PRINT.indd 243

and





21/09/2022 13:14

244  Handbook of banking and finance in emerging markets Step 2 Denote ECSW(UjECSW,VjECSW) as the Euclidean common set of weights that can project the banks closest to the frontier. ECSW(UjECSW,VjECSW) is therefore the optimal solution of



subject to

 





Note that the first condition of Equation (13.3) measures the Euclidean distance between the examined bank and its virtual position on the frontier. Accordingly, the ECSW estimated using that Euclidean distance outperforms other common set of weights measures because it is more realistic in terms of determination of weights, and it also allows for ranking both efficient and inefficient banks (Hammami et al., 2020). Consequently, the analysis of the Vietnamese banks in this study using the ECSW approach is expected to provide more insightful (and managerial) information compared to the traditional DEA or other DEA common set of weights approaches. 3.2  Estimating the Impact of ICT on Vietnamese Banks’ Performance Efficiency scores of the Vietnamese banks, after being adjusted under the ECSW approach as described in the previous section, are further examined to determine if ICT development is one of their key contributors. This type of two-stage DEA analysis – i.e. estimating DEA efficiency score in the first stage and then regressing those scores to some other variables of interest in the second stage – is popular in the DEA literature (e.g. Hoff, 2007; Ngo & Tsui, 2020; Nguyen et al., 2019; Simar & Wilson, 2007). This is because two-stage DEA can account for exogenous factors that are not the physical inputs of the examined firm or bank but can still influence the firm’s/bank’s performance  – such as macroeconomic policies or the firm’s/bank’s characteristics (e.g. ownership, size, location) (Le et al., 2020; Le & Ngo, 2020; Nguyen et al., 2018). In two-stage DEA, particularly in its second stage, Tobit regression and truncated regression are the most popular methods for estimating the relationship between DEA efficiency and other exogenous factors (Hoff, 2007; Ngo & Tsui, 2020; Simar & Wilson, 2007) because the dependent variable EFjECSW in Equation (13.4) is censored or truncated within (0; 1] – see the second constraint in Equation (13.3). This study, therefore, employs both the Tobit and truncated regression methods to improve the robustness of the analysis. We also improve the reliability of the regression results by applying a bootstrap technique (Algorithm #1, Simar & Wilson, 2007) which can account for biases created due to the relationship between those ­exogenous factors, including ICT, and other inputs and outputs of the banks.

NGUYEN 9781800880894 PRINT.indd 244

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­245 The general form of the second stage regression is expressed as follows:



where EFjECSW is the ECSW efficiency of bank j estimated from Equation (13.2); is an index that represents the development of ICT in bank j; Xi is a set of control variables including both bank j’s characteristics and the macro-environment that bank j is operating from; and ε is random errors. The descriptive statistics as well as other information regarding those variables are presented in the following section. 3.3  Selection of Data and Variables This chapter examines the impact of ICT development, among other factors, on the efficiency and performance of Vietnamese banks during the 2007–2019 period. The data on ICT development in Vietnam’s enterprises, including the banking sector, was extracted from the Vietnam ICT Index report series (MIC, 2020). The series started back in 2006 as the joint work of the Ministry of Information and Communications (MIC) and the Vietnam Association for Information Processing (VAIP) to examine the preparedness and readiness for ICT development at various organizations in Vietnam, including ministries, provincial government bodies, enterprises and corporations, and commercial banks. About 20–40 banks were included each year in the reports, totalling 298 bank-year observations for 2005–2019. Other variables of interest that can also contribute to the performance of Vietnamese banks include their ownership status (SOCB), loan loss provision ratio (LLP), off-balance sheet (OBS) activities and number of branches (BRANCH). There is evidence that stateowned banks outperform private banks in emerging markets such as Vietnam and China (Ariff & Can, 2008; Avkiran, 2011; Du & Girma, 2011; Ngo & Tripe, 2017; Vu & Turnell, 2010), and that banks with better lending portfolios perform better than their counterparts with inferior portfolios (Berger & Mester, 1997; Dong et al., 2014; Liang et al., 2008; Mester, 1996; Ngo & Tripe, 2017). Recent studies also found that diversification, which is represented by the increasing trend of OBS activities (Lozano-Vivas & Pasiouras, 2014), can help generate positive cash flows and thus improve a bank’s performance (Lieu et al., 2005; Lozano-Vivas & Pasiouras, 2010, 2014). Meanwhile, the number of branches (or network size) is a proxy for the geographic expansion of a bank to meet the needs of customers that can influence the bank’s performance (Berger & DeYoung, 2001; Le, 2021; Nguyen et al., 2018). Similar to Fujii et al. (2014) and Hammami et al. (2020), we examine Vietnamese banks as intermediaries that hire employees and pay operating expenses (the two inputs X1 and X2, respectively) in order to generate operating incomes and total assets (the two outputs Y1 and Y2, respectively). This data was extracted from the newest version of the Vietnamese banking database (Ngo & Le, 2017), which covers 44 banks with operational and financial data from 2002 to 2019. After matching this database with the ICT index data provided by the MCI and VAIP, we ended up with a sample consisting of 5–22 banks in the 2007–2019 period, totalling an unbalanced panel of 180 bank-year observations. The descriptive statistics of the dataset used in this study are presented in Table 13.1.

NGUYEN 9781800880894 PRINT.indd 245

21/09/2022 13:14

246  Handbook of banking and finance in emerging markets Table 13.1  The variables of interest in the dataset Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Whole  sample

No. of Banks 5 6 11 9 8 13 13 15 16 19 21 22 22 180

First-stage DEA/ECSW

Second-stage regression

X1

X2

Y1

Y2

7863.60 7021.83 5817.82 6599.67 8425.50 7626.85 7006.54 9487.87 10836.81 9248.74 9514.90 9990.59 10636.64 8930.69

3092.41 3887.54 2760.66 3994.29 6121.14 5433.99 5346.55 7560.48 8954.30 8395.18 9567.67 6462.30 7534.95 6773.69

9594.94 12074.16 8609.79 13171.69 24439.63 18655.72 16355.55 19191.97 21286.68 19709.38 22764.95 26549.52 31577.47 20800.77

122822.00 121290.39 114841.58 153839.57 209863.55 187455.10 196729.41 264110.24 313765.65 285641.03 315269.51 337207.55 386801.33 264560.62

ICT SOCB LLP OBS BRANCH 0.60 – 0.57 0.63 0.59 0.53 0.56 0.54 0.60 0.46 0.44 0.45 0.50 0.52

0.60 0.50 0.27 0.33 0.38 0.23 0.23 0.27 0.25 0.16 0.14 0.14 0.14 0.28

1.32 1.79 1.32 1.50 1.35 1.31 1.32 1.34 1.24 1.20 1.16 1.25 1.19 1.33

0.20 0.10 0.11 0.11 0.07 0.08 0.06 0.11 0.14 0.15 0.20 0.33 0.36 0.16

5.23 5.09 4.97 5.03 5.27 5.23 5.01 5.35 5.59 5.36 5.30 5.41 5.44 5.25

Notes:  X1 represents the number of employees (persons); X2 represents operating expenses (billion Vietnamese Dong/VND); Y1 represents operating income (billion VND); Y2 represents total assets (billion VND); ICT is the ICT Index (higher value is better); SOCB is a dummy variable with a value of 1 if the bank is a state-owned commercial bank, and 0 otherwise; LLP represents the loan loss provision ratio (%); OBS measures the bank’s off-balance sheet activities (% of total assets); and BRANCH is the logarithmic value of the bank’s number of branches.

4.  RESULTS AND DISCUSSION We first report the efficiency of Vietnamese banks during the 2007–2019 period. As presented in Figure 13.3, efficiency scores estimated from the ECSW approach are slightly lower than those derived from traditional DEA approaches – this is in line with findings from Hammami et al. (2020). However, DEA and ECSW efficiency scores are strongly correlated with each other, given that both Spearman’s test (ρ = 0.886, p-value = 0.001) and Kendall’s test (τ = 0.725, p-value = 0.001) for comparing the two are significant at 1%. At bank level, it is suggested from Figure 13.4 that the top performers include BAC A Bank, SCB and SeABank, while the worst performers include Agribank, VPB and KLB. There are two important things to note from Figure 13.4. First, some banks (such as Viet Capital, VPB and KLB) seem to perform well under the traditional DEA approach – that is, if they are allowed to have a dynamic set of weights that can maximize their strengths. For example, on average Viet Capital had the fewest employees (X1), while KLB had the lowest operating expenses (X2) during the examined period compared to other banks. Therefore, these banks can assign a very small weight/shadow price on their employees (or operating expenses) to reduce labour costs (or operational costs), and consequently improve their DEA efficiency scores (see the objective function of Equation (13.1)). However, when the law of one price is applied, their performance under ECSW becomes worse. This is because all banks are now facing the same shadow price(s) on the

NGUYEN 9781800880894 PRINT.indd 246

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­247 0.8 ECSW

DEA

0.6

0.4

0.2

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

Figure 13.3  Average efficiency of Vietnamese banks over time 1.0

DEA

ECSW

0.8

0.6

BLK

BPV

BTS

KNABIRGA

KNABDH

BBA

KNABLATIPACTEIV

BIV

BCO

KNABPT

KNABGP

BIE

BCA

BBM

AMAN

BCT

KNABATEIV

GTC

DIB

BSM

BCV

BHS

BCN

BCS

KNABAES

KNABACAB

0.2

TEIV NEIL

0.4

Figure 13.4  Average efficiency of Vietnamese banks (2007–2019)

NGUYEN 9781800880894 PRINT.indd 247

21/09/2022 13:14

248  Handbook of banking and finance in emerging markets same input/output, and therefore these banks could not get much advantage from the two inputs. It thus makes clearer the advantage of the ECSW approach over the traditional DEA approach regarding the ‘law of one price’. The second thing to note is that Agribank is always evaluated as inefficient – out of 27 banks examined, it ranked #27 in DEA and #25 in ECSW. This is because the current evaluation treats the banks as commercial entities – that is, how efficient they are in terms of minimizing their inputs (labour and operating costs) and maximizing their outputs (operating incomes and assets). In contrast, Agribank is the Bank for Agriculture and Rural Development which carries out the tasks of ‘accompanying the development of agriculture and rural areas, with magnitude of contributions to promoting the process of restructuring the economy, building new rural areas, and ensuring social security’.5 Given that its aim is not profit maximization (as a policy bank), Agribank’s low DEA/ECSW efficiency scores of are reasonable.6 Results from the second-stage regression analysis using Equation (13.4) are presented in Table 13.2 – note that 2000 bootstrap replications are used in this stage, as suggested in Simar and Wilson (2007). According to Table 13.2, while a significant contribution of ICT development to DEA efficiency scores could not be found (columns 2 and 3), positive and significant relationships between ICT and ECSW (at 1% level of significance) are reported Table 13.2  Second-stage regression results Tobit regression Coef

SE

Truncated regression Coef

SE

Tobit regression Coef

SE

Truncated regression Coef

SE

Dependent variable DEA

DEA

ECSW

ECSW

Independent variables ICT 0.176 LLP −0.090*** SOCB 0.141*** OBS −0.046 BRANCH −0.042*** YEAR 0.017*** Constant −33.476***

0.112 0.022 0.041 0.052 0.014 0.004 8.170

0.169 −0.080*** 0.148*** −0.010 −0.045*** 0.015*** −29.350***

0.108 0.019 0.040 0.047 0.013 0.004 7.555

0.260*** −0.050** 0.079*** −0.066 −0.020* 0.020*** −39.022***

0.086 0.021 0.029 0.048 0.011 0.004 7.580

0.251*** −0.038** 0.085*** −0.025 −0.023*** 0.017** −34.164***

0.083 0.015 0.029 0.043 0.009 0.003 5.974

Model statistics N B χ62 p-value

180 2000 41.03 0.001

177 2000 38.57 0.001

180 2000 47.56 0.001

177 2000 49.45 0.001

Notes:  Coef – Coefficient; SE – Standard Error; DEA – DEA efficiency scores; ECSW – ECSW efficiency scores; ICT – ICT Index (higher value is better); SOCB – a dummy variable of value 1 for a state-owned commercial bank, and 0 otherwise; LLP – loan loss provision ratio (%); OBS – off-balance sheet activities of the bank (% of total assets); BRANCH – the logarithmic value of the number of branches; YEAR – from 2007 to 2019; N – the number of observations; B – the number of bootstrap replications.

NGUYEN 9781800880894 PRINT.indd 248

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­249 in columns 4 and 5. This strengthens the argument that ECSW is superior to DEA since it provides more insightful information for efficiency analysis (Hammami et al., 2020). It further strengthens the argument that ICT development helps improve the efficiency and performance of commercial banks (Appiahene et al., 2019; Le & Ngo, 2020), as in the case of Vietnamese banks here. Table 13.2 also shows that, except for off-balance sheet activities (OBS), where relationship with banking efficiency is not significant, all other variables of interest can significantly influence the bank’s performance.7 First, it is suggested that state-owned banks outperform their counterparts where the coefficient of SOCB is significantly positive – this finding is consistent with previous studies on Vietnamese banking efficiency and performance (Ngo & Tripe, 2017; Nguyen et al., 2016, 2018; Vu & Turnell, 2010). Second, the performance of Vietnamese banks tends to improve over time, given the positive coefficient of YEAR. Since this variable can capture changes in the macro-economic developments in Vietnam (such as economic growth, financial market development and banking regulation/restructuring improvement),8 we further argue that over the examined period of 2007–2019, the macro-economic also positively contributes to the banks’ performance, in line with the arguments of Nguyen et al. (2016 and 2018). Third, it is observed that banks with higher problem loans (i.e. higher LLP) and/or banks with larger networks (i.e. higher BRANCH) underperformed. The former relationship has been well documented in the banking literature (Berger & DeYoung, 1997; Liang et al., 2008; Ngo & Tripe, 2017), while the latter suggests that Vietnamese banks may experience diseconomies of scale, similar to the findings of Fu and Heffernan (2008) on Chinese banks and Ngo and Tripe (2017) and Le et al. (2020) on Vietnamese banks.

5. CONCLUSIONS This study is the first attempt to examine the impacts of ICT, among other factors, on the efficiency and performance of 27 Vietnamese commercial banks during the 2­ 007–2019 period using a two-stage approach. In the first stage, it incorporates the slack-based measure (SBM) of DEA proposed by Tone (2001) into the ECSW model (Hammami et al., 2020) to estimate the efficiency scores of the banks. In the second stage, it then applies the bootstrapped DEA (Simar & Wilson, 2007) in examining the determinants of those efficiency scores. In addition, both Tobit and truncated regression models are used in this second stage to strengthen the robustness of the estimation results. In this sense, this study contributes to the literature in both a methodological way (as an extension of the ECSW/SBM/bootstrap) and a practical way (as an application on Vietnamese banks). The empirical results of this study suggest that ICT development in Vietnamese banks can significantly help improve efficiency and performance. This finding further justifies the recent restructuring of the Vietnamese banking sector toward innovation and digitalization (Vietnamese Government, 2012a, 2012b, 2017). It is also suggested that macroeconomic environment, lending quality, ownership and network are important factors that can influence a bank’s efficiency and performance. This study can be extended to other emerging economies in the Asia-Pacific as well as in other regions.

NGUYEN 9781800880894 PRINT.indd 249

21/09/2022 13:14

250  Handbook of banking and finance in emerging markets

NOTES 1. In previous studies, ICT could be measured by total expenditure on one or more aspects, including computer hardware (computers, storage devices, printers, and other peripherals), software (operating systems, programming tools, applications, and internal software development), computer services (web hosting, data processing, computer and network systems integration, etc.), communications (voice and data communication services), and wired and wireless communication equipment. 2. In contrast, a general rule of thumb for the central limit theorem of the parametric approach to work is to have at least 30 observations in the sample (Ross, 2014). 3. Equation (13.1) is the multiplier form of DEA described in Charnes et al. (1978). 4. Readers are encouraged to examine the original study by Hammami et al. (2020) for more details. 5. Retrieved 10/05/2021 from https://www.agribank.com.vn/en/ve-agribank/gioi-thieu-agribank. 6. Similar situations are found in Islamic banks, which operate under the Sharia-compliant principle instead of profit-maximization principle (Abdul-Majid et al., 2017; Beck et al., 2013; Miah & Uddin, 2017). 7. As discussed earlier, OBS activities in the Vietnamese banking sector are still limited (see also Table 13.1), and the role of OBS in bank efficiency is therefore still not clear. 8. It is noted that the restructuring of the Vietnamese banking sector toward a modernized and digitalized system was started in 2012 (Vietnamese Government, 2012a).

REFERENCES Abdul-Majid, M., Falahaty, M., & Jusoh, M. (2017). Performance of Islamic and conventional banks: A metafrontier approach. Research in International Business and Finance, 42, 1327–1335. https://doi.org/10.1016/j. ribaf.2017.07.069. Alshubiri, F., Jamil, S. A., & Elheddad, M. (2019). The impact of ICT on financial development: Empirical evidence from the Gulf Cooperation Council countries. International Journal of Engineering Business Management, 11, 1–14. https://doi.org/10.1177/1847979019870670. Andersen, P., & Petersen, N. C. (1993). A procedure for ranking efficient units in data envelopment analysis. Management Science, 39(10), 1261–1264. https://doi.org/10.2307/2632964. Appiahene, P., Missah, Y. M., & Najim, U. (2019). Evaluation of information technology impact on bank’s performance: The Ghanaian experience. International Journal of Engineering Business Management, 11, 1847979019835337. https://doi.org/10.1177/1847979019835337. Ariff, M., & Can, L. (2008). Cost and profit efficiency of Chinese banks: A non-parametric analysis. China Economic Review, 19(2), 260–273. http://www.sciencedirect.com/science/article/pii/S1043951X07000296. Arora, H., & Arora, P. (2013). Effect of investments in information technology on bank performance: Empirical evidence from Indian public sector banks. International Journal of Business Information Systems, 13(4), 400–417. Avkiran, N. K. (2011). Association of DEA super-efficiency estimates with financial ratios: Investigating the case for Chinese banks. OMEGA, 39(3), 323–334. Banker, R. D., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management Science, 30(9), 1078–1092. Beccalli, E. (2007). Does IT investment improve bank performance? Evidence from Europe. Journal of Banking & Finance, 31(7), 2205–2230. https://doi.org/10.1016/j.jbankfin.2006.10.022. Beck, T., Demirgüç-Kunt, A., & Merrouche, O. (2013). Islamic vs. conventional banking: Business model, efficiency and stability. Journal of Banking & Finance, 37(2), 433–447. https://doi.org/10.1016/j.jbankfin.​ 2012.09.016. Berger, A. N. (2003). The economic effects of technological progress: Evidence from the banking industry. Journal of Money, Credit, and Banking, 35(2), 141–176. Berger, A. N., & DeYoung, R. (1997). Problem loans and cost efficiency in commercial banks. Journal of Banking & Finance, 21(6), 849–870. Berger, A. N., & DeYoung, R. (2001). The effects of geographic expansion on bank efficiency. Journal of Financial Services Research, 19(2), 163–184. https://doi.org/10.1023/A:1011159405433. Berger, A. N., & DeYoung, R. (2006). Technological progress and the geographic expansion of the banking industry. Journal of Money, Credit and Banking, 38(6), 1483–1513. Berger, A. N., & Mester, L. J. (1997). Inside the black box: What explains differences in the efficiencies of financial institutions? Journal of Banking & Finance, 21, 895–947. Charnes, A., Cooper, W. W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6), 429–444.

NGUYEN 9781800880894 PRINT.indd 250

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­251 Chege, S. M., Wang, D., & Suntu, S. L. (2020). Impact of information technology innovation on firm performance in Kenya. Information Technology for Development, 26(2), 316–345. https://doi.org/10.1080/02681102 .2019.1573717. Ciciretti, R., Hasan, I., & Zazzara, C. (2009). Do internet activities add value? Evidence from the traditional banks. Journal of Financial Services Research, 35(1), 81–98. DeYoung, R. (2001). The financial performance of pure play internet banks. Economic Perspectives, 25(1), 60–75. DeYoung, R., Evanoff, D. D., & Molyneux, P. (2009). Mergers and acquisitions of financial institutions: A review of the post-2000 literature. Journal of Financial Services Research, 36(2–3), 87–110. https://doi. org/10.1007/s10693-009-0066-7. Dinçer, H., & Yüksel, S. (2020). The role of IT investment on the bank performance: A cointegration and causality analysis for Asian countries. In S. Sikdar, R. C. Das, & R. Bhattacharyya (Eds.), Role of IT-ITES in economic development of Asia: Issues of growth, sustainability and governance (pp. 13–25). Springer Singapore. https://doi.org/10.1007/978-981-15-4206-0_2. Dong, Y., Hamilton, R., & Tippett, M. (2014). Cost efficiency of the Chinese banking sector: A comparison of stochastic frontier analysis and data envelopment analysis. Economic Modelling, 36, 298–308. Du, J., & Girma, S. (2011). Cost economies, efficiency and productivity growth in the Chinese banking industry: Evidence from a quarterly panel dataset. Empirical Economics, 41(1), 199–226. http://dx.doi.org/10.1007/ s00181-011-0465-0. Fu, X., & Heffernan, S. (2008). Economies of scale and scope in China’s banking sector. Applied Financial Economics, 18(5), 345–356. http://dx.doi.org/10.1080/09603100600843924. Fujii, H., Managi, S., & Matousek, R. (2014). Indian bank efficiency and productivity changes with undesirable outputs: A disaggregated approach. Journal of Banking & Finance, 38, 41–50. Gërguri-Rashiti, S., Ramadani, V., Abazi-Alili, H., Dana, L.-P., & Ratten, V. (2017). ICT, innovation and firm performance: The transition economies context. Thunderbird International Business Review, 59(1), 93–102. https://doi.org/10.1002/tie.21772. Girmaye, H. (2018). Information communication technology and bank profitability: Evidence from Ethiopia. Eastern Africa Social Science Research Review, 34(2), 1–19. https://doi.org/10.1353/eas.2018.0011. Hammami, H., Ngo, T., Tripe, D., & Vo, D.-T. (2020). Ranking with a Euclidean common set of weights in data envelopment analysis: With application to the Eurozone banking sector. Annals of Operations Research. https://doi.org/10.1007/s10479-020-03759-6. Hernando, I., & Nieto, M. J. (2007). Is the internet delivery channel changing banks’ performance? The case of Spanish banks. Journal of Banking & Finance, 31(4), 1083–1099. Ho, S. J., & Mallick, S. K. (2010). The impact of information technology on the banking industry. Journal of the Operational Research Society, 61(2), 211–221. https://doi.org/10.1057/jors.2008.128. Hoff, A. (2007). Second stage DEA: Comparison of approaches for modelling the DEA score. European Journal of Operational Research, 181(1), 425–435. http://www.sciencedirect.com/science/article/pii/S​​0​3​7​7​2​2​1​7​0​6​0​0​ 3821. Holden, K., & El-Bannany, M. (2004). Investment in information technology systems and other determinants of bank profitability in the UK. Applied Financial Economics, 14(5), 361–365. Humphrey, D., Willesson, M., Bergendahl, G., & Lindblom, T. (2006). Benefits from a changing payment technology in European banking. Journal of Banking & Finance, 30(6), 1631–1652. http://dx.doi.org/10.1016/j. jbankfin.2005.09.009. Kao, C., & Hung, H.-T. (2005). Data envelopment analysis with common weights: The compromise solution approach. Journal of Operational Research Society, 56, 1196–1203. Le, C., Šević, A., G. Tzeremes, P., & Ngo, T. (2020). Bank efficiency in Vietnam: Do scale expansion strategies and non-performing loans matter? International Journal of Finance & Economics, 1–22. https://doi. org/10.1002/ijfe.2179. Le, T. (2019). The interrelationship between liquidity creation and bank capital in Vietnamese banking. Managerial Finance, 45(2), 331–347. https://doi.org/10.1108/MF-09-2017-0337. Le, T. D. Q. (2021). Geographic expansion, income diversification, and bank stability: Evidence from Vietnam. Cogent Business & Management, 8(1), 1885149. https://doi.org/10.1080/23311975.2021.1885149. Le, T. D. Q., & Ngo, T. (2020). The determinants of bank profitability: A cross-country analysis. Central Bank Review, 20(2), 65–73. https://doi.org/10.1016/j.cbrev.2020.04.001. Liang, C.-J., Yao, M.-L., Hwang, D.-Y., & Wu, W.-H. (2008). The impact of non-performing loans on bank’s operating efficiency for Taiwan banking industry. Review of Pacific Basin Financial Markets and Policies, 11(02), 287–304. https://doi.org/10.1142/s0219091508001350. Lieu, P.-t., Yeh, T.-l., & Chiu, Y.-h. (2005). Off-balance sheet activities and cost inefficiency in Taiwan’s banks. Service Industries Journal, 25(7), 925–944. https://doi.org/10.1080/02642060500134196. Liu, F.-H. F., & Peng, H. H. (2008). Ranking units of the DEA frontier with common weights. Computers & Operations Research, 35, 1624–1637.

NGUYEN 9781800880894 PRINT.indd 251

21/09/2022 13:14

252  Handbook of banking and finance in emerging markets Liu, J. S., Lu, L. Y. Y., Lu, W.-M., & Lin, B. J. Y. (2013). A survey of DEA applications. OMEGA, 41(5), 893–902. http://www.sciencedirect.com/science/article/pii/S0305048312002186. Lozano-Vivas, A., & Pasiouras, F. (2010). The impact of non-traditional activities on the estimation of bank efficiency: International evidence. Journal of Banking & Finance, 34(7), 1436–1449. Lozano-Vivas, A., & Pasiouras, F. (2014). Bank productivity change and off-balance-sheet activities across different levels of economic development. Journal of Financial Services Research, 46(3), 271–294. https://doi. org/10.1007/s10693-013-0181-3. Mankiw, N. G. (2020). Principles of economics (9th ed.). Cengage Learning. Mester, L. J. (1996). A study of bank efficiency taking into account risk-preferences. Journal of Banking & Finance, 20(6), 1025–1045. Miah, M. D., & Uddin, H. (2017). Efficiency and stability: A comparative study between Islamic and conventional banks in GCC countries. Future Business Journal, 3(2), 172–185. https://doi.org/10.1016/j.f​b​j​.​ 2017.11.001. MIC (2020). Vietnam ICT Index Reports. Ministry of Information and Communications. Ngo, T. (2012). Measuring performance of the banking system: Case of Vietnam (1990–2010). Journal of Applied Finance & Banking, 2(2), 289–312. Ngo, T., & Le, T. (2019). Capital market development and bank efficiency: A cross-country analysis. International Journal of Managerial Finance, 15(4), 478–491. https://doi.org/10.1108/IJMF-02-2018-0048. Ngo, T., & Le, T. D. (2017). Resource for research: A Vietnamese banking database. SSRN eLibrary, 14. https:// dx.doi.org/10.2139/ssrn.3028996. Ngo, T., & Tripe, D. (2017). Measuring efficiency of Vietnamese banks: Accounting for nonperforming loans in a single-step stochastic cost frontier analysis. Pacific Accounting Review, 29(2), 171–182. https://doi. org/10.1108/PAR-06-2016-0064. Ngo, T., & Tsui, K. W. H. (2020). A data-driven approach for estimating airport efficiency under endogeneity: An application to New Zealand airports. Research in Transportation Business & Management, 34, 100412, Article 100412. https://doi.org/10.1016/j.rtbm.2019.100412. Nguyen, H.-D., Ngo, T., Le, T., Ho, H., & Nguyen, H. T. (2019). The role of knowledge in sustainable agriculture: Evidence from rice farms’ technical efficiency in Hanoi, Vietnam. Sustainability, 11(9), 2472. Nguyen, T., Tripe, D., & Ngo, T. (2018). Operational efficiency of bank loans and deposits: A case study of Vietnamese banking system. International Journal of Financial Studies, 6(1), 14. http://www.mdpi.com/22277072/6/1/14. Nguyen, T. P. T., Nghiem, S. H., Roca, E., & Sharma, P. (2016). Bank reforms and efficiency in Vietnamese banks: Evidence based on SFA and DEA. Applied Economics, 48(30), 2822–2835. Robert, D. (2001). The financial performance of pure play internet banks. Economic Perspectives, 25(1), 60–75. Roll, Y., Cook, W. D., & Golany, B. (1991). Controlling factor weights in data envelopment analysis. IIE Transactions, 23(1), 2–9. http://dx.doi.org/10.1080/07408179108963835. Ross, S. M. (2014). Introduction to probability and statistics for engineers and scientists (5th ed.). Academic Press. Salim, R., Hoque, M., & Suyanto, S. (2010). The role of governance, ICT and bad loans in Australian bank efficiency: An empirical study. Asia Pacific Journal of Economics and Business, 14(1), 18–36. http://hdl.handle. net/20.500.11937/10828. Sathye, S., & Sathye, M. (2017). Do ATMs increase technical efficiency of banks in a developing country? Evidence from Indian banks. Australian Accounting Review, 27(1), 101–111. https://doi.org/10.1111/auar.​ 12110. Scott, S. V., Van Reenen, J., & Zachariadis, M. (2017). The long-term effect of digital innovation on bank performance: An empirical study of SWIFT adoption in financial services. Research Policy, 46(5), 984–1004. https://doi.org/10.1016/j.respol.2017.03.010. Simar, L., & Wilson, P. W. (2007). Estimation and inference in two-stage, semi-parametric models of production processes. Journal of Econometrics, 136, 31–64. Tone, K. (2001). A slacks-based measure of efficiency in data envelopment analysis. European Journal of Operational Research, 130(3), 498–509. http://www.sciencedirect.com/science/article/pii/S0377221799004075. Valverde, S. C., & Humphrey, D. B. (2009). Technological innovation in banking: The shift to ATMs and implicit pricing of network convenience. In L. Anderloni, D. T. Llewellyn, & R. H. Schmidt (Eds.), Financial innovation in retail and corporate banking. Edward Elgar Publishing. Vietnamese Government (2012a). Decision No. 254/QD-TTg on approving the scheme on ‘Restructuring the credit institutions system in the 2011–2015 period’. http://www.lawfirm.vn/?a=doc&id=2530. Vietnamese Government (2012b). Decision No. 432/QD-TTg on the Viet Nam Sustainable Development Strategy for the 2011–2020 period. Vietnamese Government (2017). Directive No. 16/CT-TTg on the Strengthening of the Ability to Access the Fourth Industrial Revolution.

NGUYEN 9781800880894 PRINT.indd 252

21/09/2022 13:14

Impact of ICT on banking efficiency: the Vietnamese experience  ­253 Vu, H. T., & Turnell, S. (2010). Cost efficiency of the banking sector in Vietnam: A Bayesian stochastic frontier approach with regularity constraints. Asian Economic Journal, 24(2), 115–139. Weigelt, C., & Sarkar, M. (2012). Performance implications of outsourcing for technological innovations: Managing the efficiency and adaptability trade-off. Strategic Management Journal, 33(2), 189–216.

NGUYEN 9781800880894 PRINT.indd 253

21/09/2022 13:14

an emerging market perspective

14. Competition, efficiency and stability in Islamic and conventional banking systems: an emerging market perspective Md. Nurul Kabir and Andrew C. Worthington

1. INTRODUCTION The relationship in banking systems between competition and stability is a core concern for regulators, policymakers, and practitioners. This is because, while competition in banking can lead to substantial gains for economies, one of the driving forces of financial instability is excessive competition (Dima et al., 2014). Consequently, the competition– stability nexus has been the subject of much attention. Within this, two dominant hypotheses appear in the literature concerning the relationship between competition and stability, often referred to as the competition–fragility hypothesis and the competition–stability hypothesis. The former argues that market power (as opposed to competition) increases stability  because banks with greater market power can reduce the problem of asymmetric information and have better-quality screening and monitoring methods to select creditworthy borrowers, as well as charging higher interest rates. However, more ­ recent  studies challenge this argument in that greater competition helps banks to be more innovative and efficient and increases their stability (Boyd and De Nicolo, 2005; Nicoló et al., 2006; Dima et al., 2014; Andrieş and Căpraru 2014). Both hypotheses enjoy theoretical and empirical support, and hence no conclusive findings are available to date. However, while there is ample evidence concerning the effects of competition on stability for different economies, studies investigating the ‘transmission mechanism’ through which competition affects stability remain scarce. This is especially important for policymakers in the design of a stability-enhancing policy for a bank. Existing competition–­ stability theories suggest that ‘efficiency’ could be one of the transmission mechanisms through which competition affects stability (Dima et al., 2014). We argue that if ­competition increases, banks will diversify to survive in the more competitive market, introducing new and innovative products and services, and thereby reducing their costs. This process should increase efficiency. Thus, efficiency can positively affect stability (Schaeck and Cihák, 2014). The alternative view is that when market power increases, banks will pay lower interest rates, and save on the cost of screening and monitoring by engaging in ‘relationship-type’ banking, which may both increase the efficiency of banking and have a positive impact on stability. Whether market power or competition has the more substantial impact on stability via efficiency as a channel is unknown to policymakers. We attempt to contribute to this literature by combining these two strands of literature by incorporating efficiency as a channel mechanism. 254

NGUYEN 9781800880894 PRINT.indd 254

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­255 To investigate whether efficiency plays a role as a channel between competition and stability, we apply the three dominant hypotheses in the banking literature. The first hypothesis describes the relationship between competition and efficiency, the second investigates the relationship between efficiency and stability, and the third looks at how competition affects stability. The first competition–efficiency hypothesis argues that in a non-competitive environment, managers can extract higher rents and not improve the quality of service, thus lowering their efficiency. According to this hypothesis, competition increases the efficiency of firms. In turn, the efficiency–stability hypothesis argues that efficient banks have better screening and monitoring mechanisms for borrowers, helping to lower their default probability. Furthermore, efficient allocation of resources also helps increase the stability of banks. Finally, the competition–stability hypothesis argues that competition makes a firm more innovative and forces banks to adopt better credit risk management, which makes them more stable. Combining the findings from these three hypotheses would help establish whether efficiency is an appropriate channel through which to transmit competition into stability. We postulate that if competition increases efficiency (accepting the first hypothesis) and efficiency increases stability (accepting the second hypothesis) and competition increases stability (accepting the third hypothesis), then competition enhances stability through the efficiency mechanism. Alternatively, if market power positively affects the level of efficiency and efficiency positively affects stability, we conclude that market power increases efficiency, and efficiency turns into stability. Existing empirical studies tend to examine each of these three hypotheses separately. For example, Berger (1995), Chortareas et al. (2011) and Färe et al. (2015) investigate the competition and efficiency hypothesis; Koetter and Porath (2007), KoutsomanoliFilippaki and Mamatzakis (2009) and Saeed and Izzeldin (2016) examine the efficiency– stability hypothesis; and Fungáčová and Weill (2013), Anginer et al. (2014) and Fiordelisi and Mare (2014) consider the competition–stability hypothesis. Other recent studies of banking efficiency and/or stability and/or competition in emerging market contexts include Du and Sim (2016), Moudud-Ul-Huq (2019), Ofori-Sasu et al. (2019), Abdelaziz et al. (2020), Ariefianto et al. (2020), Duho et al. (2020), Moudud-Ul-Huq et al. (2020) and Zhao et al. (2021). However, to the best of our knowledge, no extant research evaluates the transmission mechanism between competition and stability by combining all three hypotheses in a single study. In addition to these more general concerns, we are especially interested in how this relationship may differ between two alternative banking systems – namely, conventional and Islamic banking (Kabir and Worthington, 2014). Over the last two decades, Islamic banking globally has experienced unprecedented growth, especially in emerging markets. However, Islamic banking faces fierce competition from its conventional counterparts in most of the economies in which it operates. Therefore, the sustainability of Islamic banks in the longer run is a concern for the regulators. While several studies have undertaken comparative analysis of efficiency and stability between Islamic and conventional banks (Hassan, 2006; Čihák and Hesse, 2010; Srairi, 2010; Beck et al., 2013; Kabir et al., 2015; Kabir and Worthington, 2017; Sufian et al., 2017), there is no evidence comparing the varying impact of competition on stability between the two banking sectors, and none on the role of efficiency in stability. Thus, our study provides a timely contribution regarding the competition–efficiency–stability nexus for both banking systems.

NGUYEN 9781800880894 PRINT.indd 255

21/09/2022 13:14

256  Handbook of banking and finance in emerging markets The remainder of the chapter is structured as follows. Section 2 reviews the literature on the competition–efficiency–stability relationship. Section 3 discusses the methodology and provides the descriptive statistics of the variables in our models. Section 4 presents the empirical results, and Section 5 concludes.

2.  HYPOTHESIS DEVELOPMENT 2.1  Competition–Efficiency Hypothesis During the last three decades, the competition–efficiency hypothesis has been one of the most widely investigated hypotheses in the industrial organization literature. Many of these studies have examined this hypothesis in the banking industry, especially as competition is a vital consideration in the financial sector. Maudos and de Guevara (2007) list three hypotheses to describe the relationship between competition and efficiency. These are the ‘structure–conduct–performance’ (S–C–P), the ‘efficient–structure’ (E–S) and the ‘relative market power’ (RMP) hypotheses. Another, the ‘quiet life hypothesis’, is a special case of the RMP hypothesis widely used in the banking sector. We propose the following hypotheses: H1n: Competition increases efficiency H1a: Market power increases efficiency. 2.2  Efficiency–Stability Hypothesis According to this hypothesis, greater efficiency will translate into enhanced stability  because the bank will have a better asset quality, and this will reduce the likelihood  of  default (Schaeck and Cihák, 2014). Berger and DeYoung (1997) develop four hypotheses to describe the relationship between efficiency and default risk, namely the ‘bad luck’, ‘bad management’, ‘skimping’ and ‘moral hazard’ hypotheses. They argue that banks with more credit risk are located far from the best-practice frontier (Berger and Humphrey, 1992; Wheelock and Wilson, 1995). Furthermore, poor m ­ anagement is unable to control bank costs and fails to improve the stability of banks, thus further increasing credit risk. Moreover, inefficient management systems can incur more cost to screen and monitor the performance of borrowers, as well as to seize the collateral of borrowers in times of financial distress, which eventually increases the cost to banks and leads to even higher credit risk. We propose the following hypotheses: H2n: Efficiency increases stability. H2a: Efficiency decreases stability. 2.3  Competition–Stability Hypothesis The competition–stability nexus is a very widely investigated research area in the banking literature. This is because competition is a double-edged sword in the banking industry,

NGUYEN 9781800880894 PRINT.indd 256

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­257 as healthy competition is required for stability but excessive competition may result in fragility for both institutions and the sector. Both views enjoy theoretical and empirical support. The competition–fragility view argues that banks with greater market power have better screening and monitoring ability to identify creditworthy borrowers, helping to lower the default risk. Furthermore, excessive competition may erode the charter value of banks and force them to diversify their loan portfolios into riskier lines of business, hence increasing default risk. Conversely, a recent literature challenges this paradigm by arguing that excessive market power may actually induce banks to take on additional risk; or they may suffer from a moral hazard problem because incumbent  banks receive subsidies under the ‘too-big-to-fail’ policy, which indeed increases the default risk for banks (Boyd and De Nicolo, 2005). We propose the following hypotheses: H3n: Competition increases stability. H3a: Market power increases stability.

3.  METHODOLOGY AND DESCRIPTIVE STATISTICES 3.1 Data We begin with countries from the Organization of Islamic Cooperation (OIC) that have  both conventional and Islamic banks. An initial screening results in 21 countries  that fit the criteria. Then, because of the lack of availability and inconsistency in  the  data of the required variables, we remove another eight countries; so the final sample becomes 13 countries with both Islamic and conventional banks. From these sample countries, we select banks observed for at least three consecutive years. This results in 254 conventional and 70 Islamic banks with 1,995 and 358 observations, respectively. We winsorize all variables at the 1% and 99% levels to remove outliers. We collect the data from several sources, with bank-specific variables from Bankscope and the macroeconomic variables from the World Bank and the Heritage Foundation, and any bank-specific missing data manually from the banks’ annual reports. Table 14.1 provides the names, definitions and sources of all the variables discussed in this section. 3.2 Models To investigate whether efficiency is an appropriate channel through which competition affects stability, we test the three hypotheses using the following equations:

In the following sections, we describe the variables used in these regression models.

NGUYEN 9781800880894 PRINT.indd 257

21/09/2022 13:14

258  Handbook of banking and finance in emerging markets Table 14.1  Variables, definitions and sources Variable Competition   Total revenue   Total cost        

Total output Price of deposits Price of labor Price of capital

 Equity Efficiency   Total cost (TC)   Total loan (Q1)   Other earning assets (Q2)   Off balance sheet items (Q3)   Price of labour (W1)   Price of capital (W2)   Price of fund (W3)   Equity (E)   GDP growth rate   Inflation (INF) Stability   Nonperforming loans (NPL) Bank controls   Total asset (LTA)   Growth of total asset (GTA)   Equity to asset (ETA)   Liquidity (LIQ) Macroeconomic controls  Stock market capitalization (SMC)   Concentration (CON)  Economic Freedom Index (EFI)  Financial Freedom Index (FFI)  GDP   Governance (GOV)

Definition Interest + other operating income Interest + personnel + other operating expenses Loans + other earning asset Interest expense/ total deposit Personnel expense/total assets Other operating expense/total fixed assets Total equity Interest + personnel + other operating expenses Gross loans Other earning assets Off balance sheet items Personnel expense/total assets Non-interest expense/total fixed asset Interest expense/total deposit Total equity Growth rate of nominal GDP Change in CPI Net impaired loans/gross loans Natural logarithm of total assets Change in total assets Gross equity/total assets Bank deposit/customer deposit Stock market value as a % of GDP

Source   Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope   Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope World Bank World Bank   BankScope   Bankscope Bankscope Bankscope Bankscope   World Development

% market share of three largest banks Index of independence of individual economic decisions Index of financial freedom

World Development Heritage Foundation

Growth rate of nominal GDP Mean of governance measures in Kaufmann et.al (2010)

World Bank Worldwide Governance

Heritage Foundation

3.3  Competition Measure In this analysis, we use the Lerner index as our measure of competition as it is the only one calculable at the bank level. Furthermore, the Lerner index is a more accurate measure of market power than standard concentration measures (Jiménez et al., 2013). In brief,

NGUYEN 9781800880894 PRINT.indd 258

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­259 the Lerner index measures the ability of a firm to set a price above marginal cost. In other words, it directly measures the market power of an individual firm. Mathematically, we express this as follows: where Pit and MCit are the price and marginal cost of the output of bank i in year t, respectively. We calculate the price of output using the ratio of total revenues to total assets following Fungáčová and Weill (2013) and Fiordelisi and Mare (2014). In line with recent studies, we estimate the marginal cost using a translog cost function comprising one output, Qit (loans), and three input prices, Whit (where h is deposits, labour and capital): where E is each bank’s total equity, T is a time trend that captures technological change and e is the error term. Total equity (E) in this model accounts for the possible use of capital as a source of loan funding. This is in line with the intermediation approach to bank behaviour where deposits are an intermediate input for producing loans. To impose the symmetry condition and linear homogeneity restrictions, we divide the total cost and the prices of all inputs by the price of labour. As a result, the translog cost function becomes:

NGUYEN 9781800880894 PRINT.indd 259

21/09/2022 13:14

260  Handbook of banking and finance in emerging markets In this equation, the error term has two components, such that εit = υit + νit, where νit is a two-sided error term representing noise and υit is a one-sided disturbance term representing inefficiency. We estimate this equation using maximum likelihood techniques. We calculate MC as follows: Once we estimate the marginal cost and the price of output computed, we can calculate the Lerner index for each bank by replacing these two values in that equation. 3.4  Efficiency Measure A substantial body of literature investigates efficiency in banking. In a perfectly competitive market where cost minimization is the primary objective, banks require input quantities (X) at a given price (W) to produce outputs (Q) so that the total cost (TC) is optimal. The model assumes that the total cost deviates from the optimal cost by a random disturbance, and inefficiency forms part of the error term. Thus, the error term consists of two components, where ν is a two-sided component that represents the random disturbance in the model, and υ is a one-sided variable that captures inefficiency relative to the frontier. Both ν and υ are independently distributed. Following Mohanty et al. (2013), we specify the following equation to account for heteroscedasticity and allow the single-step estimation of the parameters of the cost function as follows: and where TCit denotes the observed total cost for bank i in year t; Wit is a vector of input prices; Qit is a vector of outputs of the firm; Eit is a vector of fixed netputs; β is a vector of the parameters estimated; Zit are the determinants of cost inefficiency (both bankspecific and macro variables common to all banks); νit are random fluctuations assumed to follow a symmetric normal distribution; and υit are the firm’s inefficiency assumed to follow an asymmetric truncated normal or a half-normal distribution. To implement the cost frontier, we use the following:

NGUYEN 9781800880894 PRINT.indd 260

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­261

There is a considerable amount of debate in the literature regarding the definition of cost, outputs and inputs in banking, for which there are several approaches – including the production, intermediation, asset, value-added and user-cost approaches to efficiency estimation. Given their main function is to channel funds from depositors to borrowers, the role of banks is intermediary. Hence, we use an intermediation approach to select the input and output of the above model. Total cost (TC) is the sum of interest and non-interest expenses. We select three output variables and three input prices. The output variables encompass total loans (Q1), other earning assets (Q2) and off-balance sheet items (Q3). The input prices are the prices of labour (W1), capital (W2) and funds (W3). We obtain the price of labour (W1) by dividing total personnel expense by the number of total assets, the price of capital (W2) as the ratio of non-interest expenses and total fixed assets, and the price of funds (W3) as the ratio of interest expenses and total deposits. To impose the homogeneity restriction, we divide the total cost, total outputs and all input prices by one of the input prices, which is the price of funds (W3). We also include equity capital as a netput to control for differences in risk preferences. Since we are measuring efficiency for multiple countries, it is important to control for country heterogeneity. Accordingly, we include country dummies, GDP growth and inflation to control for heterogeneity. Year dummies are also included to control for a time fixed effect. Finally, the translog cost function to estimate cost ­efficiency is as follows:

The maximum likelihood estimation technique estimates the stochastic frontier in the above equation. Given the specifications of the translog stochastic frontier cost function, the cost efficiency of the individual bank is:

NGUYEN 9781800880894 PRINT.indd 261

21/09/2022 13:14

262  Handbook of banking and finance in emerging markets 3.5  Financial Stability Measure Based on prior research (Berger and DeYoung, 1997; Jiménez et al., 2010; Fiordelisi and Mare, 2014), we use the non-performing loan ratio as a proxy for financial stability. The NPL ratio is the total amount of impaired loans to the net amount of loans, such that a high NPL ratio indicates the higher probability of a bank’s insolvency. One of the advantages of using the NPL ratio as a measurement of financial stability is that it is a direct measure and subject to managerial discretion. 3.6  Control Variables We also introduce a set of control variables that are bank-specific and macro. For bankspecific control variables, we include the logarithm of total assets (size), the equity to asset ratio (ETA), the growth of total assets (GTA) and liquidity as measured by ratio of bank deposits to customer deposits. For the macro control variables, we include stock market capitalization (SMC), bank concentration (CON), the economic freedom index (EFI), the financial freedom index (FFI), the real gross domestic product growth rate (GDP) and the governance (GOV) score. 3.7  Estimation Techniques After confirming all variables are stationary using unit root tests and to correct for endogeneity and autocorrelation (results not shown), we employ dynamic panel data estimation techniques – namely, the GMM estimator. A general dynamic model specification is as follows: where subscript i denotes the number of the cross-sections and t the time dimensions of the panel; y is the dependent variable (stability or efficiency); yit−1 is the lagged dependent variable; Xit is a k × 1 vector of explanatory variables including the variable of interest (competition or efficiency) other than the; yit−1 and εit is the disturbance with νi the observed bank-specific effect and uit the idiosyncratic error. Here, the assumption is that εit follows a one-way error component model where νi∼IID(0,σ 2ν) and is independent of ui∼IID(0,σ 2u); δ indicates the speed of adjustment to equilibrium. We employ Windmeijer (2005) corrected standard errors to produce variance-covariance estimates that are robust to heteroscedasticity.

4.  EMPIRICAL RESULTS 4.1  Descriptive Analysis Table 14.2 details the means and standard deviations of the three main variables of ­interest – namely, competition, efficiency and stability – as measured by the Lerner index, stochastic cost frontier efficiency and credit risk by country for all, conventional and

NGUYEN 9781800880894 PRINT.indd 262

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­263 Table 14.2  Descriptive statistics for competition, efficiency and stability by country Country

All banks

Conventional banks

Islamic banks

Means test

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

Competition Bahrain Bangladesh Egypt Indonesia Jordan Kuwait Malaysia Pakistan Qatar Saudi Turkey UAE Yemen

0.45 0.24 0.12 0.22 0.19 0.16 0.15 0.20 0.25 0.21 0.18 0.14 0.13

0.22 0.01 0.07 0.10 0.10 0.08 0.07 0.10 0.16 0.12 0.19 0.05 0.06

0.39 0.25 0.11 0.21 0.17 0.16 0.15 0.20 0.19 0.20 0.19 0.13 0.14

0.20 0.01 0.07 0.10 0.09 0.07 0.07 0.10 0.11 0.10 0.00 0.05 0.06

0.54 0.23 0.15 0.18 0.24 0.17 0.16 0.20 0.39 0.25 0.12 0.11 0.12

0.23 0.01 0.06 0.09 0.10 0.10 0.07 0.12 0.17 0.18 0.00 0.04 0.07

Efficiency

Bahrain Bangladesh Egypt Indonesia Jordan Kuwait Malaysia Pakistan Qatar Saudi Arabia Turkey UAE Yemen

0.72 0.69 0.73 0.72 0.72 0.73 0.72 0.72 0.73 0.73 0.69 0.73 0.73

0.03 0.16 0.02 0.06 0.08 0.02 0.05 0.09 0.02 0.02 0.19 0.02 0.02

0.73 0.70 0.73 0.72 0.71 0.73 0.73 0.72 0.72 0.73 0.69 0.72 0.72

0.02 0.14 0.02 0.06 0.09 0.01 0.03 0.10 0.03 0.01 0.19 0.02 0.01

0.72 0.63 0.72 0.69 0.74 0.72 0.71 0.71 0.73 0.72 0.82 0.73 0.73

0.05 0.23 0.01 0.04 0.04 0.03 0.09 0.05 0.02 0.03 0.00 0.02 0.02

1.35* 3.49*** 2.07** 2.76*** −1.81** 2.45*** 3.64*** 0.70 −2.19*** 2.43*** −2.46*** −3.00*** −1.75**

Stability

Bahrain Bangladesh Egypt Indonesia Jordan Kuwait Malaysia Pakistan Qatar Saudi Arabia Turkey UAE Yemen

9.05 8.03 15.19 6.71 14.45 8.91 6.45 12.84 5.06 3.60 7.53 6.47 24.58

15.15 10.6 14.2 12.31 17.06 9.59 7.82 12.00 8.77 4.25 12.97 6.22 19.36

6.48 7.72 14.84 6.54 12.87 5.98 6.81 13.33 6.10 3.42 7.73 6.51 34.42

6.39 8.48 13.65 12.07 15.15 5.82 8.16 12.31 10.18 3.44 13.76 5.82 17.02

16.82 9.42 18.27 9.21 24.48 14.77 5.64 10.56 2.58 4.05 6.26 6.30 7.24

27.18 17.15 19.26 15.55 24.21 12.64 6.92 10.27 2.40 5.78 5.72 7.60 7.18

−3.49*** −1.22 −0.644 −1.07 −3.04*** −4.75*** 1.33 * 1.46* 1.86** −0.81 0.69 0.20 6.90***

NGUYEN 9781800880894 PRINT.indd 263

t-stat. −4.86*** 1.38* −1.45* 1.54* −3.73*** −0.24 −1.27 −0.31 −65*** −1.86** −2.08 4.03*** 0.77

21/09/2022 13:14

264  Handbook of banking and finance in emerging markets Islamic banks. It also includes t-tests of the differences in means for conventional and Islamic banks. In terms of competition, larger values indicate greater market power; and we can see that Islamic banks have significantly more market power in all markets than conventional banks in Bahrain, Egypt, Jordan, Qatar and Saudi Arabia, while conventional banks have more market power in Bangladesh and the UAE. Of the markets overall, Bahrain, Bangladesh and Qatar have the least competitive banking systems, and Egypt, Yemen and Malaysia the most competitive. Regarding efficiency, conventional banks are significantly more efficient than Islamic banks in Bahrain, Bangladesh, Egypt, Indonesia, Kuwait, Malaysia and Saudi Arabia, and less efficient in Jordan, Qatar, Turkey, the UAE and Yemen. Lastly, Islamic banks are less stable (have more credit risk) than conventional banks in Bahrain, Jordan and Kuwait, and more stable in Malaysia, Pakistan, Qatar and Yemen. Over all banks (both conventional and Islamic), those in Egypt, Jordan, Pakistan and Yemen are less stable, while those in Saudi Arabia, Qatar, Malaysia and the UAE are more stable. 4.2  Competition–Efficiency Hypothesis Table 14.3 provides the results using the system GMM estimation method: columns (1)–(3) are for all banks, (4)–(6) for conventional banks and (7)–(9) for Islamic banks; columns (1), (4) and (7) are for the most basic model including just the variables of ­interest (competition and efficiency); (2), (5) and (8) also include the bank controls; and (3), (6) and (8) include both the bank and macroeconomic controls. We begin with the validity tests of the model. The overall fitness of the model is satisfactory, as the  Wald  test rejects the null hypothesis in all our regressions. To test the presence of dynamic characteristics in the model, we investigate the coefficient and level of significance of the lagged dependent variable. In all regressions, the lagged dependent variable  (cost efficiency) significantly correlates with the dependent variable, and the coefficient value remains stable in all regressions, which justifies the usage of the dynamic model. With our regression results in the full models, competition (the Lerner index) has a positive and significant relationship with efficiency in both Islamic and conventional banks. This implies that higher market power increases efficiency for both Islamic and conventional banks. Our findings are in line with those studies that reject the quiet life hypothesis (Koetter and Porath, 2007; Maudos and de Guevara, 2007; Casu and Girardone, 2009), and thereby counter studies by Ariss (2010) and Delis and Tsionas (2009). Interestingly, the positive relationship between efficiency and competition is consistent for both Islamic and conventional banks. However, the magnitude of the coefficients differs significantly between Islamic and conventional banks. The effect of market power on efficiency is significantly higher for Islamic banks (0.082) than for conventional banks (0.028), implying that Islamic banks with greater market power are more cost efficient than conventional banks. Several reasons could be behind this. First, as discussed earlier, Islamic banks typically have greater market power than conventional banks. An increase in market power gives Islamic banks the opportunity to pay a lower effective interest rate to depositors. This saves them some cost. Second, Islamic banks operate their businesses in close relationship with their customers, so screening and monitoring costs are then usually lower than for

NGUYEN 9781800880894 PRINT.indd 264

21/09/2022 13:14

265

NGUYEN 9781800880894 PRINT.indd 265

21/09/2022 13:14

GOV

GDP

FFI

EFI

CON

SMC

LIQ

GTA

ETA

LTA

Lerner

Efficiency(t−1)

Dependent var. Efficiency 0.052 (1.79) 0.043** (2.26) 0.023*** (2.98) 0.0019* (1.71) −0.007 (1.12) −0.001 (0.56) −0.002 (1.41) 0.001 (1.07) −0.001 (0.91) −0.001 (0.08) −0.223*** (3.64) 0.036 (1.51)

0.054 (2.24) 0.044** (2.44) 0.024*** (3.93) 0.001* (1.73) −0.001 (0.57) 0.001 (0.80)

0.039 (.08) 0.046** (2.12) *

(3) **

(2)

(1)

All banks

Table 14.3  Competition–efficiency relationship

0.040 1.04 0.054*** (3.29)

(4) *

0.035 (1.85) 0.038 (2.00) 0.028*** (4.62) 0.002** (2.22) 0.002 (0.57) 0.001 (0.02)

(5)

(6) **

0.035 (2.05) 0.028** (−2.21) 0.045 (0.00) 0.002 (0.00) −0.003 (0.00) −0.009 (0.00) −0.007 (0.00) 0.000 (0.00) 0.002 (0.00) 0.001 (0.00) −0.230** (2.54) 0.045 (0.00)

Conventional banks *

0.067 (1.76) 0.030* (1.98)

(7) **

0.030 (2.52) 0.077** (2.17) −0.008*** (3.40) −0.000 (0.16) −0.001* (1.72) 0.001 (0.70)

(8)

Islamic banks 0.02** (2.01) 0.082** (2.34) −0.020** (2.54) −0.001 (0.46) −0.001 (0.91) 0.001 (0.77) 0.001 (1.61) −0.001 (0.06) −0.001 (0.31) −0.001 (0.52) −0.194* (1.87) 0.004 (0.18)

(9)

266

NGUYEN 9781800880894 PRINT.indd 266

21/09/2022 13:14

0.717*** (210.81) 12.21 2276 −2.45*** −1.7 77.53

(1) 0.490*** (9.01) 27.87 2276 −2.5*** 2.28 75.17

(2)

All banks 0.580*** (4.81) 24.42 2276 −3.20*** 1.26 71.39

(3) 0.714*** (205.19) 13.24 1901 −2.46*** 0.93 76.32

(4) 0.436*** (7.18) 20.34 1901 −2.09*** 2.49 73.25

(5)

Conventional banks 0.543*** −3.21 35.05 1901 −2.75*** 1.18 75.85

(6)

0.564*** −6.44 4.95 375 −1.85** 1.88 73.77

(7)

0.652*** 8.6 20.4 375 −1.19** 0.45 37.05

(8)

Islamic banks 0.681*** 6.68 31.19 375 −0.98** 0.59 32.38

(9)

Notes:  Figures in parentheses are t-statistics. Asterisks *, ** and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively. Variable definitions are shown in Table 14.1.

Wald test N AR(1) AR(2) Hansen test

Constant

Dependent var. Efficiency

Table 14.3  (continued)

Competition, efficiency, stability in Islamic and conventional banking  ­267 conventional banks. This could be another cause for greater efficiency in Islamic banks when market power increases. Among the control variables, size has a significant positive relationship with efficiency  in the full sample as well as for conventional banks. However, it displays a ­significant negative relationship with efficiency in the Islamic bank sample, indicating that when Islamic banks get bigger, efficiency decreases. The equity to asset ratio also has a significant positive relationship with efficiency in conventional banks, but does not  show similar in Islamic banks. The growth of the total asset ratio has a negative relationship in both Islamic and conventional banks but does not appear to be significant. Similarly, the liquidity ratio has a significant but not positive relationship with efficiency. With respect to the macroeconomic variables, stock market capitalization to GDP has a negative relationship with cost efficiency in conventional banks but a positive relationship in Islamic banks. While bank concentration shows a positive relationship with efficiency in conventional banks but shows a negative relationship with Islamic banks, in either case it is not a significant determinant of cost efficiency. The Economic Freedom index has no significant impact on the efficiency. The GDP growth rate is negatively significant for both Islamic and conventional banks. The good governance score has a positive association with efficiency but does not appear to be significant in either of the banking systems. 4.3  Efficiency–Stability Relationship Table 14.4 present the results for the efficiency–stability relationship. The results of the Wald test indicate the overall validity of the model, the independence of the second-order correlation meets the first criteria for GMM estimation, and the rejection of the null hypothesis of Hansen’s test across the model indicates that instruments are not overidentified. The highly significant coefficient of the lagged NPL ratio variable confirms the dynamic character of the model specification. The estimated persistence of the NPL ratio is positive for both Islamic and conventional banks, and the level of persistence is higher for Islamic banks than conventional banks as the mean value of theta (persistence) is 0.631 and 0.824 for conventional banks and Islamic banks, respectively. In the full models including both the main variables of interest and the bank and macroeconomic controls, efficiency has a significant negative impact on stability (the NPL ratio) in the full sample and for conventional banks only. Our findings are thus consistent with the findings of Berger and DeYoung (1997) and Koetter and Porath (2007), who find that an increase in efficiency reduces the default probability of banks. However, they contradict the findings of Koutsomanoli-Filippaki and Mamatzakis (2009). Our results are also partially consistent with those of Saeed and Izzeldin (2016) regarding the insignificant impact of cost efficiency on the reduction of default probability in Islamic banks. However, Saeed and Izzeldin find that cost efficiency increases default risk for conventional banks, whereas our results support an inverse relationship between cost efficiency and the NPL ratio for conventional banks. A plausible explanation could be that the operational history of Islamic banks is of much shorter duration than that of conventional banks, and thus lack of experience makes them unable to turn efficiency into ­stability.

NGUYEN 9781800880894 PRINT.indd 267

21/09/2022 13:14

268

NGUYEN 9781800880894 PRINT.indd 268

21/09/2022 13:14

N Wald Test AR1 AR2 Hansen

Constant

GOV

GDP

FFI

EFI

CON

SMC

LIQ

GTA

ETA

LTA

Efficiency

NPL(t−1)

NPL

Dependent var.

18.75 (1.54) 2276 1094.07 −3.79 −0.478 227.93

46.19*** (2.90) 2276 1820.68 −3.56 −0.879 187.61

0.626*** (12.57) −20.44** (−2.18) −2.054** (−2.49) −0.280*** (−3.22) −0.053*** (−4.51) 0.098*** (3.82) −0.016 (−0.86) −0.032 (−1.54) −0.140 (−0.93) 0.018 (0.39) −31.94*** (−3.48) 2.877 (0.95)

0.559*** (9.44) −27.48** (−2.17) −0.829 (−0.81) −0.217 (−1.57) −0.025** (−2.20) 0.088*** (2.70)

0.713*** (12.44) −30.22** (−2.38)

23.61** (2.56) 2276 176.15 −3.38 −0.349 124.28

(3)

(2)

(1)

All banks

Table 14.4  Efficiency–stability relationship

29.08** (2.56) 1901 196.30 −3.22 −0.55 119.75

0.713*** (12.01) −37.82** (−2.44)

(4)

18.94* (1.96) 1901 972.28 −3.39 −1.05 198.27

0.554*** (8.55) −36.03*** (−3.56) 0.337 (0.47) −0.113 (−1.06) −0.032** (−2.37) 0.090*** (3.17)

(5)

(6)

41.34*** (2.58) 1901 2279 −3.10 −1.23 162.34

0.626*** (13.85) −29.54*** (−3.44) −0.881 (−1.23) −0.212** (−2.39) −0.055*** (−3.90) 0.086*** (3.71) −0.011 (−0.72) −0.035* (−1.88) −0.145 (−0.98) 0.025 (0.53) −28.94*** (−2.66) 1.084 (0.39)

Conventional banks

9.190** (2.18) 375 341.07 −2.85 −1.17 59.46

0.807*** (22.14) −10.56* (−1.77)

(7)

6.420 (0.85) 375 821.44 −2.89 −1.11 35.87

0.831*** (15.93) −2.202 (−0.27) −0.267 (−0.60) 0.041 (1.16) −0.030*** (−2.71) −0.045* (−1.78)

(8)

Islamic banks

9.132 (0.84) 375 477.28 −2.48 −0.86 26.97

0.840*** (9.88) −0.840 (−0.09) 0.434 (0.48) 0.030 (0.60) −0.023 (−1.46) −0.038 (−1.06) 0.005 (0.47) 0.044 (1.19) −0.186 (−0.99) −0.0554 (−0.06) −20.75** (−2.02) 0.166 (0.07)

(9)

269

NGUYEN 9781800880894 PRINT.indd 269

21/09/2022 13:14

Constant N Wald test AR(1) AR(2) Hansen test

GOV

GDP

FFI

EFI

CON

SMC

LIQ

GTA

ETA

LTA

Lerner

NPL(t−1)

NPL

Dependent var. ***

2.397 2276 (0.50) 90 −1.98 1.48

0.627 (14.11) −8.912* (−1.95) −0.234 (−1.26) −0.072 (−1.31) −0.048*** (−3.46) 0.052** (2.48) −0.018* (−1.67) −0.016 (−1.34) 0.087 (1.05) 0.040 (0.99) −17.10** (−2.14) −0.498 (−0.66)

0.619 (14.72) −6.697** (−2.01) −0.0299 (−0.21) −0.096 (−1.62) −0.042*** (−3.11) 0.065** (2.38) ***

(3) ***

(2)

3.663* 1.886*** 2276 2276 (2.68) (1.82) 90 90 −2.86 −2.16 −1.53 −1.06

0.718 (10.09) −2.822 (−0.89)

(1)

All banks

Table 14.5  Competition–stability relationship

***

***

0.628 (15.42) −5.284* (−1.79) 0.030 (0.24) −0.135** (−2.11) −0.043*** (−2.74) 0.063** (2.42)

(5)

2.010*** 3.287* 1901 1901 (2.94) (1.77) 90 90 −1.78 −1.85 −1.52 −2.17

0.690 (10.81) −2.188 (−0.70)

(4)

(6) ***

−0.528 1901 (−0.13) 78 −1.8 −1.01

0.616 (13.19) −8.640* (−1.82) −0.141 (−0.76) −0.034 (−0.65) −0.049*** (−3.02) 0.051** (2.21) −0.009 (−1.03) −0.020** (−2.02) 0.104* (1.67) 0.036 (0.87) −12.23 (−1.38) −1.244* (−1.82)

Conventional banks ***

3.783*** 375 (6.71) 90 −2.05 0.09

0.486 (8.39) −0.025 (−0.01)

(7) ***

6.522*** 375 (3.01) 90 −2.78 −0.93

0.821 (18.80) −7.092*** (−2.95) 0.366*** (−2.71) 0.086** (2.28) −0.026*** (−4.26) −0.046* (−1.85)

(8)

Islamic banks

7.536 375 (0.75) 78 −2.35 0.39

0.845*** (9.55) −7.813*** (−3.16) 0.330 (0.85) 0.064** (2.12) −0.034* (−1.77) −0.0162 (−0.86) −0.003 (−0.30) 0.031 (1.29) −0.045 (−0.41) −0.012 (−0.22) −8.180 (−1.26) −1.003 (−0.77)

(9)

270  Handbook of banking and finance in emerging markets Among the bank-specific control variables, the equity to asset ratio has a significant negative impact on the NPL ratio for the full sample and for conventional banks. The growth of total assets has a significant negative impact on the NPL ratio, indicating that an increase in total assets lowers the NPL ratio. This could be because of diversification strategy, when a bank diversifies its asset portfolio to a different sector. The liquidity ratio is significant for both banking systems, however – positively in the case of conventional banks and negatively for Islamic banks. Among the macroeconomic control variables, an increase in GDP lowers the NPL ratio for both banking systems, as per conventional theory. 4.4  Competition–Stability Relationship Table 14.5 presents the impact of competition on stability. The results of the Wald test indicate that the overall fitness of the model is appropriate. The rejection of the null hypothesis of Hansen’s test and the independence of the second-order correlation provides the validity of results obtained from the GMM estimation technique. Like the result in Table 14.4, NPL is a highly persistent variable, which justifies the use of dynamic panel data estimation. With respect to the impact of competition on stability, we find that the Lerner index has a negative and significant impact on the NPL ratio in both banking systems. This indicates that market power is a highly significant determinant of stability in both banking systems and lends support to the competition–fragility hypothesis. Our findings are thus like those in Ariss (2010) and Fungáčová and Weill (2013), but not Nicoló et al. (2006) and Schaeck et al. (2009). Regarding the negative relationship, it is reasonable to assume that as Islamic banks have more market power than conventional banks, they can charge higher profit (interest) from borrowers, which helps them be more stable. In addition, a portion of Islamic bank asset portfolios consists of profit-and-loss sharing agreements where Islamic banks can share losses with a partner; and, because of their market power, Islamic banks may exercise this right, which may increase their stability. As for the remaining control variables, the growth of total assets has a significant negative impact on the NPL ratio in both banking systems, supporting the diversification hypothesis; and the equity to asset ratio displays a significant negative relationship with NPL in conventional banks but a significant positive relationship in Islamic banks. Liquidity is significantly positive for conventional banks but negative for Islamic banks. Lastly, GDP again has a negative significant impact on the NPL ratio.

5. CONCLUSION In this chapter, we considered the impact of banking competition on bank stability, and whether bank efficiency is an appropriate transmission mechanism to translate competition into stability. To study this competition–efficiency–stability nexus, we presented three dominant hypotheses in the banking literature, namely the competition–efficiency hypothesis, the efficiency–stability hypothesis and the competition–stability hypothesis. We tested these three hypotheses using panel data from 13 economies with both

NGUYEN 9781800880894 PRINT.indd 270

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­271 conventional and Islamic banks. We used the Lerner index to proxy for competition, the SFA method for efficiency and the NPL ratio for credit risk. We applied the system GMM estimation technique to test our three hypotheses, and controlled for both bank-specific and macro-specific factors that could affect the variable of interest. To check the sensitivity of the results, we used Merton’s distance-to-default (DD) as a proxy of credit risk and IV Tobit regression as an alternative to the GMM estimation. In terms of results, we reject the competition–efficiency hypothesis, partially accept the efficiency–stability hypothesis and reject the competition–stability hypothesis for both banking systems. We first find that market power significantly increases efficiency in both banking sectors. However, it is of a higher magnitude for Islamic banks; thus we reject the competition–efficiency hypothesis. We then find that efficiency positively affects the stability of conventional banks but has no significant impact on stability in Islamic banks. Subsequently, we find market power increases the stability of both banking sectors. We postulate that because of their greater market power, Islamic banks can pay less profit (on deposits and profit-sharing investments) and thus gain efficiency. In practical terms, this would manifest itself as Islamic banks executing contracts such as mudaraba (fund providers to the bank) and musharaka (profit/loss sharing investments) on terms relatively advantageous to the bank. These banks would still appear as upholding Islamic banking principles, but more in their favour. However, because they lack experience and have some resource limitations, Islamic banks are unable to convert this efficiency into stability. From these findings, we can conclude that efficiency is an appropriate transmission mechanism for conventional but not Islamic banks. The findings from our study have important policy implications. First, our results indicate that market power increases stability for both banking systems, which provides an indication of the need for the regulation of competition in these economies. Second, Islamic banks are not as efficient as conventional banks; and, more importantly, are unable to convert efficiency into stability. Policymakers and Islamic bank practitioners should then emphasize the need to make Islamic banks more stable, not just more efficient. Third, a uniform competition policy can govern both banking systems given the impact of competition on stability is the same for both Islamic and conventional banks.

REFERENCES Abdelaziz, H., Rim, B., & Helmi, H. (2020). The Interactional Relationships between Credit Risk, Liquidity Risk and Bank Profitability in MENA Region, Global Business Review, https://doi.org/10.1177/09721 50919879304. Andrieş, A.M., & Căpraru, B. (2014). The Nexus between Competition and Efficiency: The European Banking Industries Experience, International Business Review, 23(3), 566–579. Anginer, D., Demirguc-Kunt, A. & Zhu, M. (2014). How Does Competition Affect Bank Systemic Risk?, Journal of Financial Intermediation, 23(1), 1–26. Ariefianto, M.D., Saheruddin, H., & Soedarmono, W. (2020). The Intertemporal Impacts of Market Power on Bank Risk: Evidence from the Indonesian Banking Industry, International Journal of Economics and Management, 14(2), 279–289. Ariss, R.T. (2010). On the Implications of Market Power in Banking: Evidence from Developing Countries, Journal of Banking and Finance, 34(4), 765–775. Beck, T., Demirgüç-Kunt, A., & Merrouche, O. (2013). Islamic vs. Conventional Banking: Business Model, Efficiency and Stability, Journal of Banking and Finance, 37(2), 433–447. Berger, A.N. (1995). The Profit-Structure Relationship in Banking: Tests of Market-Power and EfficientStructure Hypotheses, Journal of Money, Credit and Banking, 27(2), 404–431.

NGUYEN 9781800880894 PRINT.indd 271

21/09/2022 13:14

272  Handbook of banking and finance in emerging markets Berger, A.N., & DeYoung, R. (1997). Problem Loans and Cost Efficiency in Commercial Banks, Journal of Banking and Finance, 21(6), 849–870. Berger, A.N., & Humphrey, D.B. (1992). Measurement and Efficiency Issues in Commercial Banking, in Z.  Griliches (ed.), Output Measurement in the Service Sectors. Chicago: University of Chicago Press, 245–300. Boyd, J.H., & De Nicolo, G. (2005). The Theory of Bank Risk Taking and Competition Revisited, Journal of Finance, 60(3), 1329–1343. Casu, B., & Girardone, C. (2009). Testing the Relationship between Competition and Efficiency in Banking: A Panel Data Analysis, Economics Letters, 105(1), 134–137. Chortareas, G.E., Garza-Garcia, J.G., & Girardone, C. (2011). Banking Sector Performance in Latin America: Market Power versus Efficiency, Review of Development Economics, 15(2), 307–325. Čihák, M., & Hesse, H. (2010). Islamic Banks and Financial Stability: An Empirical Analysis, Journal of Financial Services Research, 38(2), 95–113. Delis, M.D., & Tsionas, E.G. (2009). The Joint Estimation of Bank-Level Market Power and Efficiency, Journal of Banking and Finance, 33(10), 1842–1850. Dima, B., Dincă, M.S., & Spulbăr, C. (2014). Financial Nexus: Efficiency and Soundness in Banking and Capital Markets, Journal of International Money and Finance, 47(1), 100–124. Du, K., & Sim, N. (2016). Mergers, Acquisitions, and Bank Efficiency: Cross-Country Evidence from Emerging Markets. Research in International Business and Finance, 36, 499–510. Duho, K.C.T., Onumah, J.M., & Owodo, R.A. (2020). Bank Diversification and Performance in an Emerging Market, International Journal of Managerial Finance, 16(1), 120–138. Färe, R., Grosskopf, S., Maudos, J., & Tortosa-Ausina, E. (2015) Revisiting the Quiet Life Hypothesis in Banking Using Nonparametric Techniques, Journal of Business Economics and Management, 16(1), 159–187. Fiordelisi, F., & Mare, D.S. (2014). Competition and Financial Stability in European Cooperative Banks, Journal of International Money and Finance, 45(1), 1–16. Fungáčová, Z., & Weill, L. (2013). Does Competition Influence Bank Failures? Economics of Transition, 21(2), 301–322. Hassan, M.K. (2006). The X-Efficiency in Islamic Banks, Islamic Economic Studies, 13(2), 49–78. Jiménez, G., Lopez, J.A., & Saurina, J. (2013). How Does Competition Affect Bank Risk-Taking?, Journal of Financial Stability, 9(2), 185–195. Kabir, M.N., & Worthington, A.C. (2014). Credit Risk in Islamic Financial Products, in A.C. Worthington (ed.), Contemporary Issues in Islamic Finance: Principles, Progress, and Prospects, New York: Nova, 73–84. Kabir, M.N., & Worthington, A.C. (2017). The ‘Competition–Stability/Fragility’ Nexus: A Comparative Analysis of Islamic and Conventional Banks, International Review of Financial Analysis, 50, 111–128. Kabir, M.N., Worthington, A.C., & Gupta, R. (2015). Comparative Credit Risk in Islamic and Conventional Banks, Pacific-Basin Finance Journal, 34, 327–353. Koetter, M., & Porath, D. (2007). Efficient, Profitable and Safe Banking an Oxymoron? Evidence from a Panel VAR Approach, SSRN Electronic Journal, doi:10.2139/ssrn.958509. Koutsomanoli-Filippaki, A., & Mamatzakis, E. (2009). Performance and Merton-Type Default Risk of Listed Banks in the EU: A Panel VAR Approach, Journal of Banking and Finance, 33(11), 2050–2061. Maudos, J., & de Guevara, J.F. (2007), The Cost of Market Power in Banking: Social Welfare Loss vs. Cost Inefficiency, Journal of Banking and Finance, 31(7), 2103–2125. Mohanty, S.K., Lin, W.T., & Lin, H.-J. (2013). Measuring Cost Efficiency in Presence of Heteroskedasticity: The Case of the Banking Industry in Taiwan, Journal of International Financial Markets, Institutions and Money, 26(1), 77–90. Moudud-Ul-Huq, S. (2019). Can BRICS and ASEAN-5 Emerging Economies Benefit from Bank Diversification?, Journal of Financial Regulation and Compliance, 27(1), 43–69. Moudud-Ul-Huq, S., Zheng, C., Gupta, A.D., Hossain, S.K.A., & Biswas, T. (2020). Risk and Performance in Emerging Economies: Do Bank Diversification and Financial Crisis Matter?, Global Business Review, doi:10.1177/0972150920915301. Nicoló, D., De Nicol, G., Jalal, A.M., & Boyd, J.H. (2006). Bank Risk-Taking and Competition Revisited: New Theory and New Evidence, International Monetary Fund Working Paper No. 06/297. Ofori-Sasu, D., Mensah, L., Akuma, J.K., & Doku, I. (2019). Banking Efficiency in Emerging Economies: Does Foreign Banks Entry Matter in the Ghanaian Context?, International Journal of Finance and Economics, 24, 1091–1108. Saeed, M., & Izzeldin, M. (2016), Examining the Relationship between Default Risk and Efficiency in Islamic and Conventional Banks, Journal of Economic Behavior and Organization, 32(1), 127–154. Schaeck, K., & Cihák, M. (2014). Competition, Efficiency, and Stability in Banking, Financial Management, 43(1), 215–241. Schaeck, K., Čihák, M., & Wolfe, S. (2009). Are Competitive Banking Systems More Stable?, Journal of Money, Credit and Banking, 41(4), 711–734.

NGUYEN 9781800880894 PRINT.indd 272

21/09/2022 13:14

Competition, efficiency, stability in Islamic and conventional banking  ­273 Srairi, S.A. (2010). Cost and Profit Efficiency of Conventional and Islamic Banks in GCC Countries, Journal of Productivity Analysis, 34(1), 45–62. Sufian, F. (2011), Profitability of the Korean Banking Sector: Panel Evidence on Bank-Specific and Macroeconomic Determinants, Journal of Economics and Management, 7(1), 43–72. Sufian, F., Kamarudin, F., & Md. Nassir, A. (2017). Globalization and Bank Efficiency Nexus: Empirical Evidence from the Malaysian Banking Sector, Benchmarking: An International Journal, 24(5), 1269–1290. Wheelock, D.C., & Wilson, P.W. (1995). Explaining Bank Failures: Deposit Insurance, Regulation, and Efficiency, Review of Economics and Statistics, 77(4), 689–700. Windmeijer, F. (2005). A Finite Sample Correction for the Variance of Linear Efficient Two-Step GMM Estimators, Journal of Econometrics, 126(1), 25–51. Zhao, Y., Chupradit, S., Hassan, M., Soudagar, S., Shoukry, A.M., & Khader, J. (2021). The Role of Technical Efficiency, Market Competition and Risk in the Banking Performance in G20 Countries, Business Process Management Journal, 27(7), 2144–2160.

NGUYEN 9781800880894 PRINT.indd 273

21/09/2022 13:14

Mauritian context

15.  Central bank independence, macroprudential policies and financial stability in the Mauritian context Manisha Chuttoor, Dinesh Ramdhony and Boopen Seetanah

1. INTRODUCTION Financial stability is not a novel concept. Nonetheless, the past century has redefined the scope and magnitude of financial stability. Economic globalization has morphed financial stability into a global objective to be pursued by all world economies. The existing irregularities between markets, intermediaries and investors have exacerbated financial fragility, essentially driven by imprudent financial deregulation. The last three decades have witnessed a shift in the burden of duties assigned to central banks. Besides their traditional surveillance of price stability, central banks now undertake the responsibility of ensuring financial stability. A caveat obstructing its successful operation is the interference of incumbent government or lobbying parties that attempt to coerce policymakers into gearing policies to their advantage. This compromises the good running of financial and economic affairs at the expense of central banks’ professionalism. Since the 1970s, central bank independence (CBI) has been a stabilizing force for countries seeking politics-free monetary and macroprudential policy decisions inclined towards long-term prosperity. However, the financial crisis of 2007–08 has contributed towards negative publicity of central banks, scapegoating them for allowing the crisis to rampage and thus failing to honour their duties (Masciandaro and Romelli, 2015). Nevertheless, in 2019 the World Economic Forum re-established CBI’s essence at the heart of an economy. Despite their overburdened agenda, the centralization of both information and surveillance is believed to optimize the central bank’s effectiveness in ensuring financial stability. A central bank comes with an arsenal of tools at its disposal for tackling each of its assigned responsibilities, the most widely acknowledged and least contentious being monetary policy (MP) (Seetanah et al., 2014). How the central bank is mandated to use these tools should ideally fall within the ambit of its autonomy. Thus, an autonomous central bank will utilize both monetary policy and macroprudential limits to boost the growth of its financial sector and economy and use them as shields against instances of financial distress. Against this backdrop, the gist of this chapter is to examine whether the above is applicable in the Mauritian context, and, if so, how fruitful it has been in helping the impetus of the local financial sector. With recent financial scandals tainting the Mauritian financial sector’s reputation and policymakers questioning the efficacy of overburdening the central bank (Cochrane and Taylor, 2016), the latter’s operational autonomy is being questioned. Critics claim that the 274

NGUYEN 9781800880894 PRINT.indd 274

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­275 greed for power is causing institutions and parastatals to compromise the good of the nation by surrendering vis-à-vis powerful moguls with incontestable power, leaving Mauritius with a distorted system not geared towards attaining the best for the country and society. Berman and McNamara (1999) and Buiter (2014) assert that unfettered liberty ascribed to the central bank may well cause it to go haywire, especially in a hierarchical organizational structure where power and decision-making is highly concentrated at the top. The upper management might be driven by their vested interests rather than considering the betterment of the economy. Thus, this chapter tries to establish how fundamental CBI and MP are to the health of the Mauritian financial sector, one which aspires to become the financial hub of the African region. It is noteworthy to point at the dire lack of research linking CBI and financial stability on the global front. With CBI gaining steam as a contributor to financial stability in the literature, its compromised autonomy has left the financial sector exposed to vulnerabilities and needs to be imminently addressed. Through this study, we also attempt to give a broader definition to financial (in)stability. As noted throughout the literature, scholars such as Garcia Herrero and Del Rio (2003) and Čihák (2007) have made use of a rather narrow measure of financial (in)stability, which we have attempted to resolve by highlighting the expanse of financial stability as not being exclusive to a single sector. The rest of the chapter is organized as follows. Section 2 reviews existing theoretical explanations and empirical linkages between CBI, macroprudential instruments and financial stability. Section 3 provides the methodology used, data collection and the description of variables and other preliminary tests. Section 4 is the study’s interpretative stage, providing thorough and corroborated analysis, while Section 5 offers conclusions and recommendations.

2.  RELATED LITERATURE 2.1  Underlying Theories 2.1.1  Financial instability hypothesis Minsky’s theory (1977) postulates that financial crisis is ingrained in capitalism as periods of economic prosperity encourage borrowers and lenders to be progressively reckless. The circle of debt-borrowing and investment is recursive for prospects of anticipated profits (Kalecki 1965), supported by realized profits. In this way, debt becomes a crucial investment element that initially promotes ‘periods of tranquillity’ and growth (Papadimitriou, 2008). In the modern context, the bank will act as an intermediary between the two, seeking a purely profitable participation. As entrepreneurs in a capitalist economy, bankers are aware that innovation assures profits but perturbs stability (Schumpeter, 1961). Thus bankers, as merchants of debt, acquire assets through complex financial engineering. Minsky (1986) further proposes that deviating from this norm through hedging, speculation and Ponzi schemes would meddle with the stability component. Support from the authoritative body and dominance of stringent regulations to curb the above are the prime causes for the absence of debt-deflationary pressures (Fisher,

NGUYEN 9781800880894 PRINT.indd 275

21/09/2022 13:14

276  Handbook of banking and finance in emerging markets 1933). The ‘too big to fail’ measures, inciting moral hazard, spike inflation levels beyond inflation targets (Bernanke, 1983). Hence, balancing the scale in favour of one policy will tip the scale against another, in contrast to Smith and Walras’s classic theory (Pokorny, 1978). This creates a credit crunch situation, where money is trapped and the market is highly illiquid. This is termed a ‘balance sheet recession’ (Papadimitriou, 2008). 2.1.2  Disaster myopia hypothesis Disaster myopia has been utilized in explaining the occurrence of financial crises and some commercial property crises (Haldane, 2009). Guttentag and Herring (1986) propound that tumultuous business trends and momentums cause the recurrence of financial crises. This drives baseless investor optimism, stirring irrational behaviours and sheer disregard for risk-aversity or risk-consideration. The persistence of such conditions culminates in forming a speculative asset bubble, overshadowing all reason for doubt ex ante. The disaster myopia hypothesis aligned with Murphy’s law brings the market to a deadlock, where pessimism predominates over rationality (Allegret and Cornand, 2014). However, the market may return to stable levels if automatic stabilizers are aptly activated, a phenomenon known as reversion to the mean. Speculative bubbles are essentially characterized by high price volatility of financial assets, and eventually bust out due to over-speculation (Issing, 2003). Foot (2003) asserts that speculative bubbles are a cause of financial instability if left unattended. Asset prices are inherently a derivation of future prospects and expectations, factored in asset market valuation. But prices prevailing on the market are not always accurate, due to irrational investor behaviour. Authorities might be wary of intervening in markets to avoid foiling prospective economic growth. 2.2  Hypothesis Development 2.2.1  Central bank independence An autonomous central bank is an entity that should, above all, be free of any political entanglement; neither should it be used as an instrument for driving government mandates (Barth et al., 2003). This autonomy in operation helps the central bank deal with matters at hand according to its judgement and at its own pace, thereby preventing potential imbalance in the system (Čihák et al., 2013). Quintyn and Taylor (2003) postulate that the intrusion of political veto players contributes to the worsening of financial distress, adding to economic costs arising due to the unprofessionalism of elected ministers (Alesina and Drazen, 1991). In support of this view, Hutchison and McDill (1999) propounded the moral hazard conundrum, which became more conspicuous during the financial crisis of 2007–08. The rescue scheme was a financial strain on government budgets; and, while supporting unbuoyant firms, a weak central bank might no longer follow up as the chief of the financial arena (Mishkin, 2014). CBI was nonetheless the scapegoat of the 2007–08 financial crisis, and hence faces major backlash for its structuring (Fischer, 2015). Despite monitoring an ever-evolving banking sector, the use of unconventional policy instruments by central banks is still scrutinized and misunderstood (Cochrane and Taylor, 2016). Neo-Marxist scholars like

NGUYEN 9781800880894 PRINT.indd 276

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­277 Reichlin and Baldwin (2013), in turn, argue that CBI was never adequately attained to bear any consequential effect. Central banks are thus accused of operating under the shadow of incumbent governments, and this perception is indicative of the level of mistrust of national institutions. Nevertheless, their independence remains largely crucial to financial and macroeconomic wellbeing. The seminal paper by Klomp and Haan (2009) draws links between CBI and an authorcomputed index measuring financial instability using factor analysis. The authors criticize the use of banking crisis as an indicator of financial distress. Using a dynamic panel model over 21 years (1985 to 2005), they deduce a negative relationship, attributing this link to political independence. Greater autonomy in operations implies that the central bank is not pressured to act according to the requirements of the vetoing party, and instead acts in the principal interest of the economy by immunizing the sector and deploying early preventive action, thus ensuring financial stability. Opposing this trend, Berger and Kißmer (2013) focus on CBI and the choice of monetary policy during asset price bubbles by utilizing a social loss function with output gap, inflation and crisis probability as primary variables. They find conclusive evidence (much to the dismay of others) that the less a central bank is held on a government leash, the more it refrains from using precautionary monetary tightening policies to prevent the possibility of a crisis. They thus conclude that less autonomous central banks could contribute to inflationary pressure. In light of the above, we hypothesize: H1a: CBI has a significant effect on credit growth. H1b: CBI has a significant effect on stock market volatility. H1c: CBI has a significant effect on economic growth. H1d: CBI has a significant effect on exchange rate. 2.2.2  Monetary policy Conventionally, the only use of monetary policy was in times of regulating upending inflationary pressure (De Gregorio, 2009). Mishkin (1996) argues that interest rates can be one factor that helps precipitate a financial crisis. Along the same lines, Cukierman (1992) explains that commercial banks are faced with a time lag when adjusting constantly changing interest rates, and this might seriously impact their loan books. Issing (1993) highlights the conflicting role of the central bank in maintaining de facto independence and fending off its monetary stability mandate against a consumerist society. Bernanke and Gertler (2000) assert that interest rates should be cautiously implemented due to their trickle-down effects on credit growth and capital markets. Blinder (2006) initially pointed to irregularity in the use of monetary policy as a stabilizer of financial conditions, as outlined in many incidents preceding unstable equity markets and the ex post use of monetary policy. A central bank’s influence over monetary policy needs to be independent of interference from political bodies (Crowe and Meade, 2008). Its use should depend on the degree of inflationary pressure and inflation targeting, and should thus be implemented discretionarily (Svensson, 2010). Nevertheless, Benati and Goodhart (2010) propose that the notion of CBI will be overlooked in view of monetary policy as more pressing  issues pertaining to regulations and public debt management are prioritized. However, the spillover effect of monetary policy on the economy through the

NGUYEN 9781800880894 PRINT.indd 277

21/09/2022 13:14

278  Handbook of banking and finance in emerging markets consumption–saving nexus (wealth effect channel) and risk-return trade-off (balance sheet channel) is triggered when monetary policy is inexpertly applied (Ioannidis and Kontonikas, 2006). Borio and Lowe (2002) studied 34 countries, including the G10, for the period ­1960–1999. The authors assert that financial turbulence is characterized by low inflation shocks and disproportionate credit growth, with hikes in asset prices, which increases the probability of financial distress. They further claim that demand shocks are more likely to manifest in credit figures and asset prices rather than in price indices for consumer goods. Thus, monetary policy is an effective instrument for mitigating the adverse effects of both financial and price instability. Ueda and Valencia (2013) also made use of econometric modelling of a social loss function by incorporating variables such as monetary policy and credit growth. They profess that expansionary monetary policy can reduce the real debt burden. Moreover, they claim that the responsibility for ensuring price and financial stability falls on the central banks, whereby they face a time inconsistency problem. Before a credit crisis materializes, a central bank tends to choose the socially optimum level of inflation, while, during the crisis, it sets higher inflation targets. Lamers et al. (2016) use a structural vector autoregression (VAR) model and a panel regression analysis to assess the effect of monetary policy interventions by the European Central Bank (ECB) and the Federal Reserve on capital markets’ perception of market risk. Lamers et al. use cross-sectional data for European and US listed banks, ranging from the three last months of 2008 to the end of 2015. Their results show that banks with poor liquidity management and a time-lapse for announcing non-performing loans (NPLs) may heighten banking sector fragility. Nonetheless, continuous interest rate cuts will only repress banks’ profitability margins. We therefore propose the following hypotheses: H2a: Monetary policy has a significant effect on credit growth. H2b: Monetary policy has a significant effect on stock market volatility. H2c: Monetary policy has a significant effect on economic growth. H2d: Monetary policy has a significant effect on exchange rate. 2.2.3  Macroprudential policies Prudential supervision came into the limelight during the Great Depression (Mishkin, 2009). If an economy is to pursue financial stability, financial regulation and macroprudential policies should be implemented in an articulated way and well communicated (Claessens and Valencia 2013). Preventive macroprudential limits and supervisory power help weaken the crisis’s domino effects (Doumpos et al., 2015). Adrian and Liang (2018) claim that it is wise to keep the implementation of monetary policy and macroprudential policies at a distance. Macroprudential measures may, however, receive contradictory views (Kim and Mehrota, 2017). 2.2.4  Cash reserve ratio The cash reserve ratio (CRR) as a macroprudential tool is seconded by Monnet and Vari (2019). CRR is a conventional tool, but not archaic or anyhow ineffective in maintaining healthy limits of financial stability (Kashyap and Stein, 2012). Instead, it is more

NGUYEN 9781800880894 PRINT.indd 278

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­279 frequently used internally by commercial banks. Hence, its effect is shielded from the public gaze. Reserve requirements can also be used for monetary control purposes and adjusted, similarly to monetary policy, along the business cycle – for example to offset below trend output growth (Federico et al., 2014). CRR essentially leads to more optimistic results when geared towards financial stability (Glocker and Towbin, 2012). Indeed, Borio and White (2004) assert that procyclicality of the financial sector is abated through the imposition of reserve requirements. The aim of CRR is usually equivocal, but is principally towards stabilizing the financial sector (Central Bank of Malaysia, 2011). Its subordinate aim may be settled on ensuring price stability (Basci, 2010). In a recent paper, Cantú et al. (2019) quantify the economic trade-offs of reserve requirements with a financial stability objective. They estimate the costs of a tightening in CRR by calculating the fall of credit and industrial production growth in a panel VAR and the merits of an early warning system model. Their findings show that CRR is an effective financial stability tool. Additionally, they find that the effects of CRR are incidental in emerging market economies compared to advanced economies. We hypothesize the following: H3a: CRR has a significant effect on credit growth. H3b: CRR has a significant effect on stock market volatility. H3c: CRR has a significant effect on economic growth. H3d: CRR has a significant effect on exchange rate. 2.2.5  Liquidity coverage ratio The liquidity coverage ratio (LCR) is a desirable macroprudential instrument acting as a liquidity buffer to mitigate the effects of financial imbalances (Willem Van den End and Kruidhof, 2013). LCR gained ground after the financial crisis of 2007–08 and is fully supported by Basel III. Ex ante the crisis, many banks did not have enough liquidity reserves to effectuate payments or remit deposited money, so they began disposing of assets at large discounts, heralding a deflationary spiral (Acharya and Richardson, 2009). Contagion effects have been shown to be potentially costly (Aharony and Swary, 1983; Herring and Vankudre, 1987). The implementation of LCR can potentially limit panic during a crisis, thereby granting a concession of time for central banks to take appropriate action without rushing into decisions (Diamond and Kashyap, 2016). Runs are also less likely to occur. Banks are not incentivized to hold enough liquid assets to survive runs as this comes at the expense of securing fewer loans as they, ironically, belong to a risk-prone industry. Stein (2013) deduces that unregulated debt to the private sector will only cause excess shortterm debt, not conducive to liquidity requirement and thus a threat to financial stability. At extreme stress levels, the instrument becomes ineffective and the lender of the last resort of the central bank must underpin the system’s stability (Willem Van den End and Kruidhof, 2013). Research papers on the latter remain rare, mainly because of a lack of relevant data. However, Valla, Saes-Escorbiac and Tiesset (2006) ran an econometric model that compounded their own measure of liquidity using individual bank data from French commercial banks between 1993 and 2005. They propose that liquidity reserves remain

NGUYEN 9781800880894 PRINT.indd 279

21/09/2022 13:14

280  Handbook of banking and finance in emerging markets volatile and lethal to the survival of individual banks if not implemented as compulsory legislation. Adrian and Shin (2008), in their paper, insist on the predatory nature of the contagion effect, referring analogically to the size of the mortgage market as compared to the global financial sector; and yet we were not immunized against ripples in this sector. Based on the above discussion, we hypothesize that: H4a: LCR has a significant effect on credit growth. H4b: LCR has a significant effect on stock market volatility. H4c: LCR has a significant effect on economic growth. H4d: LCR has a significant effect on exchange rate. 2.2.6  Leverage ratio BASEL III introduced the leverage ratio (LEV), providing a cushion against credit and default risk while simultaneously contributing towards financial stability (Kiema and Jokivuolle, 2014). A leverage ratio effectively serves as a safety valve against unexpected losses and underestimation of risk, but should nonetheless be used with caution (Hilderbrand, 2008). Bair (2015) highlights resiliency, loss absorbency and shortcomings of the leverage ratio in the risk-based capital framework. Grill et al. (2018) claim that enforcing a compulsory leverage ratio may lower the tendency of taking on further risk, thereby respecting risk appetites and being sound in the process. Capital requirements reduce banks’ moral hazard and the probability of a ­system-wide crisis by approximately 25% (Gauthier et al., 2012). However, monetary policy’s impotence is caused by leverage constraints, whether adopted voluntarily or compulsorily (Barnea et al., 2015). Like LCR, leverage ratio is also a novel measure introduced by Basel III and lacks empirical precision. In this endeavour, a study by Barth and Miller (2018) established three probable scenarios using leverage ratio upgraded from 4% to 15%. Its main benefit outlined was a sharp dwindling of the probability of banking crisis, while costs included a fall in credit growth. Dermine (2015) mentions controlled bank runs and credit risk  diversification as merits of the leverage ratio. We therefore propose the following hypotheses: H5a: Leverage ratio has a significant effect on credit growth. H5b: Leverage ratio has a significant effect on stock market volatility. H5c: Leverage ratio has a significant effect on economic growth. H5d: Leverage ratio has a significant effect on exchange rate.

3.  RESEARCH METHODOLOGY 3.1  Research Design For the purpose of our study, we made use of an annual time series data stretching from 1976 to 2019, totalling 44 observations. Secondary data was the most popular dataset, with some exceptions requiring statistical computation. Macroeconomic data is derived from the World Bank Group and the International Financial Statistics databases, while

NGUYEN 9781800880894 PRINT.indd 280

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­281 banking sector-related data was collected manually from the annual reports and ­statistical bulletins of the Bank of Mauritius (BOM). 3.2  Model Specification In our endeavour to quantify the degree to which CBI and the macroprudential limits set by a central bank affect financial stability, we utilize an econometric model, largely inspired by the work of Lim et al. (2011) and Shrestha and Bhatta (2020). The choice of variables is largely inspired by the International Monetary Fund’s financial stability indicators compilation guide (IMF, 2019), and studies by Gadanecz and Jayaram (2008), Galati et al. (2011) and Demirgüč-Kunt and Detragiache (1998). We therefore present the following generic model:

(15.1) where



FS stands for financial stability indicators CBI stands for central bank independence as proxied through the turnover rate of governors at the Bank of Mauritius MP stands for monetary policy of the Bank of Mauritius CRR stands for cash reserve ratio LIQ stands for liquid assets to bank deposits, as an alternative term for LCR LEV stands for leverage ratio INFL stands for inflation rate FD stands for financial development, and MAC stands for the vector of macroeconomic variables.

This study takes different metrics of financial stability under consideration to practically reflect nearly all vital sectors of the economy that might be at risk from the destabilizing effect of financial imbalances. We therefore propose the following four regression models: A.  Financial sector



(15.2)

B.  Financial markets



NGUYEN 9781800880894 PRINT.indd 281

(15.3)

21/09/2022 13:14

282  Handbook of banking and finance in emerging markets C. Economy



(15.4)

D.  External sector



(15.5)

3.3  Description of Variables 3.3.1  Financial stability indicators Credit growth (credit to private sector) Credit growth (CreditG) was measured as the year-on-year increments in commercial bank loans given to the private sector, as they have a higher loan book weighting. Stock market volatility Stock market volatility (SEMDEX) is measured as the closing value of the main tracker index of the Stock Exchange of Mauritius collected at the end of each calendar year. It focuses on the intensity and frequency of financial asset price fluctuations in capital markets. Economic growth Economic growth (EG) measures the strength of the macroeconomy. It is the year-onyear percentage increases in GDP, denoting the current health of a country’s state of affairs. Exchange rate Exchange rate (ER) is the rate of exchange of a basket of currencies to the Mauritian rupee (MUR). Overvaluation or undervaluation of a country’s exchange rate is a sign of a currency crisis. 3.3.2  Independent variables Central bank independence CBI has been proxied through the BOM governors’ turnover rate (TOR). Simply put, CBI relates to operational autonomy and freedom from political interference in the proper execution of the central bank’s duties. Over the years, a new metric emerged in the form of TOR (Dvorsky, 2000), which focused more on de facto independence than de jure

NGUYEN 9781800880894 PRINT.indd 282

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­283 independence of central banks. TOR data for Mauritius will indicate the frequency of change in the governors of the BOM (see Appendix Table 15A.1). Given that Mauritius has an average electoral cycle of 4–5 years, the TOR should normally lie in the range of 0.2–0.25. A high TOR (above 0.2–0.25) will imply that governors have a short tenure, which is ideal in maintaining the bank’s independence. Monetary policy Reference is made to the key repo rate as monetary policy, previously known as the bank rate set by the Bank of Mauritius. This rate is now determined by the BOM’s monetary policy committee (MPC). Cash reserve ratio The CRR refers to the minimum level of cash a bank should withhold from its gross deposits, to act as cash cushion against any withdrawal or bank run. Liquid asset to depositor liquidity coverage ratio The LIQ or LCR measures liquid assets available to the banks to immunize themselves against any instance of insolvency. Leverage ratio LEV has been used as an alternative to capital requirement, as provided for under the Basel accord. 3.3.3  Control variables Papadomou et al. (2017) claim that the inclusion of control variables increases the explanatory power of any model. We therefore use the following in this study. Inflation rate In line with the literature review, inflation rate (INFL) has been computed as the percentage increase in the consumer price level, and is used to verify whether a trade-off exists between INFL and financial stability (FS). Financial development (FD)

NGUYEN 9781800880894 PRINT.indd 283

21/09/2022 13:14

284  Handbook of banking and finance in emerging markets This ratio is used to measure financial openness. FD relates to the growth of the financial sector attained through the liberalization of key policies that allow for the inflow of foreign direct investment (FDI) and portfolio investment. Vector of macroeconomic variables (MAC) The MAC used in the study is elaborated on in the Appendix. A principal component analysis (PCA) is run not only to avoid erosion of the accuracy of the tests but also to include a representative element of macroeconomic variables, for these variables also have an impact on our selected dependent variables (see Table 15A.2). Our MAC vector is thus computed using PCA on the number of selected macroeconomic variables to have their variances blended in a single root variable. PCA is a means of reducing data dimensionality while preserving the originality and reliability of the data set, and has received much consideration from scholars. PCA is distinguishable in the sense that it collates all given variables into a single principal component. Based on the Kaiser criterion (Kaiser, 1960) for its wide usage, only components with eigenvalues exceeding one would be taken. In this regard, each observed component contributes to one unit of variance of the total variance; therefore, observed components with higher eigenvalues bear higher contributions. 3.4  Multivariate Regression Analysis It should however be noted that the outcome of the bounds test indicates whether we should proceed with an autoregressive distributed lag (ARDL) or error correction model (ECM) model. Our ARDL models (assuming co-integration) will therefore be as follows: A.  Financial sector

(15.6)

B.  Financial markets             (15.7)

NGUYEN 9781800880894 PRINT.indd 284

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­285 C.  Real economy

(15.8) D.  External sector

(15.9) where α, β, γ and δ are the respective coefficients of the variables; p, q, r, s is their optimum lag; ECT the error correction term; and εt the error term.

4. ANALYSIS 4.1  Unit Root Tests Unit root test using the Augmented Dickey-Fuller test (ADF) and the Phillips-Perron unit root test indicates a mixture of I(0) and I(1) variables, so we accordingly proceed with the ARDL regression model. A bounds test was undertaken and validated the existence of cointegration – that is, the presence of a long relationship – in all specifications. An ARDL ECM is thus employed to take this into account, and such a framework interestingly allows the study of both long- and short-run relationships. 4.2  Long-Run Estimation The respective ARDL long-run estimates are summarized in Table 15.1. For the long run, most of the coefficients are in line with both the theoretical and empirical literature, except for ER, due to its moderated composition through the exclusion of inflation. Practically all coefficients of the SEMDEX regression bear significant results above the threshold of 10%. However, the leverage ratio seems ineffective in its objective of modulating the dependent variables. In the end, the coefficient of determination is considerable for all regressions, and thus shows a good overall fit. The results are further analysed below.

NGUYEN 9781800880894 PRINT.indd 285

21/09/2022 13:14

286  Handbook of banking and finance in emerging markets Table 15.1  Long-run coefficients VARIABLE

EXPECTED

CRR

Negative

FD

Positive

LIQ

Negative

LEV

Negative

INFL

Positive

MP

Negative

MAC

N/A

TO

Positive

C

N/A

R-SQUARED STD ERROR PROB (FSTAT)

CREDITG −1.026* (0.027) 0.7* (0.007) −0.097 (0.8254) 0.015* (0.0071) −0.338 (0.5968) −1.123 (0.1287) −0.045* (0.0226) 0.104 (0.3488) 0.062 (0.601) 0.537 0.096 0.0003

SEMDEX −1914.39* (0.0096) 714.935 (0.381) −2118.7* (0.100) −3.541 (0.912) 711.67 (0.718) −5621.1* (0.047) 190.58* (0.0012) 468.1* (0.07) 1296.9* (0.0092) 0.962 181.33 0.00

EG 7.396 (0.584) 0.0996 (0.986) −17.72 (0.218) 0.1247 (0.452) 42.87* (0.06) 23.803 (0.281) −0.232 (0.699) −2.614 (0.43) −4.079 (0.267) 0.426 2.884 0.328

ER 0.588 (0.947) 1.955 (0.611) 7.291 (0.347) 0.063 (0.618) −7.267 (0.541) −6.157 (0.726) 0.707 (0.310) 0.315 (0.878) 1.22 (0.782) 0.976 1.66 0.00

Notes:  * means significant at the 10% level. ** The CUSUM of squared test, which is an indicator of stability of the residuals if the latter lie within the 5% bound (Brown et al., 1975), has indeed proven the stability of our data.

4.3  Short-Run Estimates In the presence of cointegration, and with the use of an error correction model, we subsequently estimate the short-run relationships (Table 15.2). In the short run, all trends are maintained. Regarding the error correction terms, all are connoted in the negative, signalling that the short-run equation converges to a longrun equilibrium at a pace of 480% (Credit), 58% (SEMDEX), 51% (EG) and 101% for ER. 4.4  Analysis of Long-Run Relationship 4.4.1  Central bank independence CBI is positively linked to credit growth, though not statistically significant (Reject H1a). In short, a 1% change in CBI heralds a 10.4% increase in credit growth. Ordinarily, this depends entirely on the magnitude of credit growth and whether credit growth is unsustainable depending on the creditworthiness of borrowers (Hilbers et al., 2005). Borio and Shim (2008) assert that central banks, acting in their independent capacity, slow the side effects of credit growth. Nonetheless, as posited by Foos et al. (2010), excess

NGUYEN 9781800880894 PRINT.indd 286

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­287 Table 15.2  Short-run coefficients VARIABLE

SIGN

ΔCRR (-1)

Negative

ΔFD (-1)

Positive

ΔLIQ (-1)

Negative

ΔLEV (-1)

Negative

ΔINFL (-1)

Positive

ΔMP (-1)

Negative

ΔMAC (-1)

N/A

ΔTOR (-1)

Positive

C

N/A

ECT

N/A

CREDITG

R-SQUARED STD ERROR PROB (FSTAT)

−0.257 (0.676) 0.838* (0.087) −0.36 (0.598) −0.0026 (0.864) 0.63 (0.548) −1.114 (0.486) 0.017 (0.703) −0.18 (0.25) −0.011 (0.619) −4.791* (0.0069) 0.451 0.133 0.095

SEMDEX −1497.4 (0.602) 1660.3 (0.18) −1432.7 (0.184) −13.71 (0.285) 485.69 (0.819) −3257.1 (0.801) 178.75 (0.352) 584* (0.047) 21* (0.059) −0.582* (0.038) 0.841 215.9 0.099

EG −7.61 (0.684) 21.98 (0.139) −10.97 (0.581) −0.286 (0.538) −4.72 (0.879) 10.3 (0.839) 1.06 (0.395) 2.95 (0.493) −0.405 (0.547) −0.506* (0.0098) 0.314 3.84 0.183

ER 0.451 (0.857) 0.685 (0.793) 0.0165 (0.752) −1.77 (0.327) 0.731 (0.853) −3.45 (0.578) −0.045 (0.801) 0.0667 (0.91) −0.054 (0.586) −1.014* (0.00) 0.922 0.51 0.00

Note:  * means significant at the 10% level.

credit has adverse implications for a commercial bank’s profitability. These should be carefully avoided by the central bank, essentially when too much operational independence of both central and commercial banks drives credit growth exponentially, and, by extension, financial distress. CBI also has a positive impact on stock market volatility (Accept H1b). This is consistent with the findings of Papadamou et al. (2017) and Förch and Sunde (2012). In short, a 1% change in CBI results in a SEMDEX rise of 468.1 points. Haan et al. (2017) argues that financial markets would be the first to react to any reduction in independence. As per our descriptive statistics, the stock market is neither excessively volatile nor on a downward trend, and therefore not indicative of financial distress (Hutchison and McDill (1999). This mechanism works efficiently when the distancing between the role of the central bank and government influence are maximized (Umutlu et al, 2010; Esqueda et al, 2012). Opposing the trend, CBI is seen to have a negative impact on economic growth, as supported by Gharleghi (2019), albeit at a low significance level (Reject H1c). Briefly, a 1% change in CBI leads to a fall in economic growth of 2.614. The negative link established may be due to central bank priorities and is thus consistent with the findings of Barth et  al. (2013). Economic activities might not be the bank’s main concern and might be

NGUYEN 9781800880894 PRINT.indd 287

21/09/2022 13:14

288  Handbook of banking and finance in emerging markets affected by its choice of monetary or macroprudential policies, designed to meet inflation targets or restrict money supply. Additionally, there exists an affirmatory relationship between CBI and exchange rate, despite a low significance (Reject H1d). In short, a 1% change in CBI leads to a 31.5% increase in exchange rate. Fluctuations in exchange rate relayed through economic and translation risk may seriously impair the business capability of banks (Eichengreen and Hausman, 1999). With enhanced autonomy, central banks might be wary of macroeconomic indicators within their purview falling out of range, especially if they are held accountable to the public. If a bank’s autonomy allows it to independently curb galloping inflation, such that a speculative wealth effect bubble is not created or distortion of purchasing power is not palatable, any domestic currency appreciation in real terms will be sustainable. A two standard deviation improvement in the central bank independence index is estimated to reduce the exchange rate shocks associated with monetary policy movements by 50% (Ha et al., 2019). 4.4.2  Monetary policy A negative relationship is established between monetary policy and credit growth, though at low significance (Reject H2a). By extension, a 1% change in MP corresponds to a fall of 1.123 in credit growth. A tightening of monetary policy has negative implications on credit supply and loan volume (Abuka et al., 2019), although monetary policy changes do not manifest instantly (Bernanke and Blinder, 1992). Rationally speaking, a cut in the key repo rate will be translated into cuts in both prime lending rate and savings rate. Market optimism spirals borrowing upwards, blinding both bankers and authorities to any speculative bubble. Thus, credit growth becomes unsustainable, heralding multiple rounds of defaulting, especially if thorough background checks have not been carried out on the borrower. News such as this spreads like wildfire, and in less than no time depositors start reclaiming their money, causing bank runs. Commercial banks should therefore be mindful when granting loans amid market optimism. Proper customer due diligence should always be carried out. Our results highlight a significant negative link between monetary policy and stock market volatility (Accept H2b), as supported by Zhang et al. (2011). As a result, a 1% change in MP leads to a fall of 5621.1 points in SEMDEX. Cassola and Morana (2004) conclude that monetary policy aimed at long-term price stability can contribute to stock market stability. In this view, monetary policy might be used as a market corrector to lessen the prevailing precariousness and stabilize the market; but this state is only attainable when monetary policy is not frequently altered. As for monetary policy and exchange rate, a negative relationship is seen at low significance (Reject H2d). In short, a 1% change in MP leads to a 6.157% fall in exchange rate but remains insignificant. A positive relationship exists between monetary policy and economic growth, though not significant (Reject H2c). By extension, a 1% change in MP corresponds to an increase of 23.803 in economic growth. De Gregorio and Guidotti (1995) and Oosterloo and Haan (2004) rightly portrayed this relationship through an inverted parabola. Loazya and Ranciere (2004) support our findings and claim that the effects are indirect but nonetheless present. Flexible interest rates are a must for investment and economic growth (Struthers, 2003). Jappelli and Pagano (1994) postulate that, although a high interest rate might signal a fall in present consumption, it makes investment in both government bonds

NGUYEN 9781800880894 PRINT.indd 288

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­289 and capital markets attractive due to a higher return on investment. While a low interest rate incentivizes dissaving and borrowing, it also boosts economic activity as consumer purchasing power increases and, in turn, employment and production increase to meet consumer demand. 4.4.3  Cash reserve ratio CRR follows the same rationale as monetary policy. Though credit growth and SEMDEX show negative link with CRR below the significance level of 10% (Accept H3a and H3b), the macroeconomic indicators indicate positive relationships at low significance (Reject H3c and H3d). CRRs are set at benchmarks that significantly lower the probability of crisis both in the financial and the real economy. Primarily, a high CRR is symptomatic of a financially repressed economy (McKinnon, 1973) but is contemporarily symbolic to ensuring financial stability. Low reserve requirements automatically lead to higher lending rates, lowering credit growth (Cantú et al., 2019) but potentially attracting capital inflows by increasing trade opportunities, thus contributing towards economic growth (Federico et al., 2014) and less volatile exchange rates (Brei and Moreno, 2019). 4.4.4  Liquidity coverage ratio Our findings for LCR (or LIQ) conform to the literature review. Liquidity reserves serve as a cushion against any outbreaks of financial imbalance, such that the sector remains secure against any moral panic situation (Willem Van den End and Kruidhof, 2013). Liquidity holdings decline when alternative investment opportunities become more attractive, and are certainly cause for concern (Valla et al., 2006). Open market operations and liquidity reserves in hand or deposited with a central bank is callable money but helps curb the scourge of excess liquidity and the emergence of financial slump. Unregulated debt to the private sector for investment in capital markets or economic activities is not conducive to liquidity reserves, and thus a potential threat to financial stability (Stein, 2013). 4.4.5  Leverage ratio Theoretically, the leverage ratio should bear a negative link with our financial indicators (Barth and Miller, 2018), as supported by our findings, but its effectiveness remains viable exclusively in the short run. Leverage ratio controls the spiralling of credit growth, therefore keeping non-performing loans at low levels (Dermine, 2015). LEV also contributes to a robust and sound economy (Grill et al., 2018) as the financial sector and the economy are mutually inclusive (Bair, 2015). Commenting on leverage ratio’s long-run efficacy, Hildebrand (2008) proclaimed that LEV does not address issues arising predominantly in the long run. Credit concentration, excessive maturity mismatch or reliance on asset market liquidity, and thus financial cycles, remain perpetually present. Leverage ratio thus, as explained by our long-run coefficients, remains a poor financial stability management tool. 4.4.6 Inflation Our findings support a theoretically correct relationship between inflation and our dependent variables in the long run, except for credit growth. As for the short run, a time lag delays the effects of inflation on economic growth and exchange rate. In more recent

NGUYEN 9781800880894 PRINT.indd 289

21/09/2022 13:14

290  Handbook of banking and finance in emerging markets years, a wave of stability in inflation has been noticed, attributed to the accruing experience and professionalism of central banks. Central banks, at their discretion, apply contractionary monetary policy to control the trickle-down effect of inflation frictions on the macroeconomy. A stable inflation level is a sign of a stable economy, which entices producers and investors, as denoted by the positive link. 4.4.7  Financial development Financial development positively influenced all financial indicators in both the short run and the long run, despite not always being significant (Accept only H7a). In the case of Mauritius, financial development is still within healthy bounds, and contributes immensely towards economic growth and stock market development through interest rate and exchange rate liberalization. Financial development through free-floating interest rates often prompts investors to make multi-fold returns (Obstfeld and Taylor, 2004). Financial development also entails free-floating exchange rates, which are not subject to central bank intervention. Thus, any appreciation is a result of an increase in productivity, and hence exports.

5. CONCLUSIONS The main objective of this study was to assess the impact of central bank independence on financial stability. In the presence of a mix of I(0) and I(1) variables and cointegration, the ARDL-ECM model was used to estimate both long-run and short-run coefficients, using a time series over the period 1976–2019. The results showed that CBI has a positive relationship with ensuring financial stability in the financial sector, markets and the external economy; however the same cannot be concluded for the real economy. As for MP, a positive relationship was deduced with economic growth, while the link is negative for the rest. CRR, LIQ and financial development demonstrate theoretically correct relationships. Leverage ratio was found to reduce instability in the short run but became ineffective in the long run. As for inflation, a positive link is found with economic growth and stock market volatility, while the reverse is true for credit growth and exchange rate. These pertain to long-run derivations. Credit growth should ideally be considered safe until it does not cause non-performing loan figures to skyrocket. In the context of this study, Mauritius Credit Information Bureau (MCIB) background checks should always be conducted, indiscriminately and adequately for all borrowers. Leniency should not, at any cost, be exercised. The central authority should take further preventive or punitive measures against derogators, as even the slightest relaxation of policies could snowball into a financial slowdown if default rates start rising. As for stock market volatility, interest rates on a falling trend will increase demand for shares due to higher return prospects, and vice versa. Therefore, monetary policy should be weighed inside out as it affects many sectors of the economy simultaneously, from banking to open market operations, insurance and pensions funds. A central bank indirectly influences economic growth through its use of macroprudential policies. In our case, CBI has proven to keep economic growth within sustainable bounds and not as propaganda for electoral purposes. CBI and MP, for overall

NGUYEN 9781800880894 PRINT.indd 290

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­291 effectiveness, should be present interdependently. Keeping MP at a socially desirable level would keep cash hoarding, consumption and production at levels considered optimal for the economy. Given that the eroding effect of inflation on exchange rate and the impact of MP are as dictated by economic theories, CBI has a contradictory effect on exchange rate. Thus, central banks must be vigilant, as maintaining stable external currency values is one of their cardinal raisons d’être. Emphasis should be placed on other factors to maintain stable levels, such as the excess liquidity of domestic currency on the market or intervening in the market to rectify any anomalies instead of taking steps towards having a free market-determined rate of exchange. This will enable the maintenance of stable purchasing power parity. With the COVID-19 pandemic disrupting the social and economic spheres, CBI might be more prone to risks than before. There is an undeniable need for closer collaboration between governments and central banks to work on a viable plan for economic recovery and public debt restructuring (Marmefelt, 2020). However, central banks must be on their guard to operate within their mandate to avoid any usurping of duties. While avoiding further unsustainable public debt might be a fair consideration for governments, baiting central banks for short-term recovery by depleting their reserves will only compromise the good of the economy in the long run (Khushiram, 2020). As proven by our study, any adverse change in CBI or its control over macroprudential policies will trigger the four identified pillars of financial stability, thereby jeopardising the economy. Nonetheless, there is no statistically significant link between sovereign debts and monetary policy decisions (Schnabel, 2020). Instead, in this time of crisis, fiscal policy is the optimal and chosen policy for dealing with economic problems induced by COVID19 (Yashiv, 2020). Although helicopter money from tax rate cuts or liquidity injection does not necessarily undermine CBI, it might aggravate colossal amounts of public debt (Kapoor and Buiter, 2020). This study is not without limitations. It would have benefited from a higher number of observations for more robust estimates. Future research could also delve into a comparative analysis between developed and developing countries to establish more novel outlooks and reverse causality issues in the hypothesized links.

REFERENCES Abuka C., Alinda R.K., Minoiu C., Peydro J.L., and Presbitero A.F. (2019). ‘Monetary policy and bank lending in developing countries: Loan applications, rates, and real effects.’ Journal of Development Economics, 139, 185–202. Acharya V.V., and Richardson M.P. (2009). ‘Causes of the financial crisis.’ Critical Review, 21(2–3), 195–210. Adrian, T., and Liang N. (2018). ‘Monetary policy, financial conditions, and financial stability.’ International Journal of Central Banking, 14, 73–131. Adrian T., and Shin H.S. (2008). ‘Liquidity and financial contagion.’ Financial Stability Review, Banque de France, 11 (Special issue). Aharony J., and Swary I. (1983). ‘Contagion effects of bank failures: Evidence from capital markets.’ Journal of Business, 56(3), 305–322. Alesina A., and Drazen A. (1991). ‘Why are stabilizations delayed?’ American Economic Review, 81(5), 1170–1188. Allegret J-P., and Cornand C. (2014). ‘Une analyse informationnelle de la crise financière récente.’ Revue Française d’Economie, 28, 213–264.

NGUYEN 9781800880894 PRINT.indd 291

21/09/2022 13:14

292  Handbook of banking and finance in emerging markets Bair S.C. (2015). ‘How a supplemental leverage ratio can improve financial stability, traditional lending and economic growth.’ Financial Stability Review, Banque de France, 19, 75–80. Barnea E., Landskroner Y., and Sokoler M. (2015). ‘Monetary policy and financial stability in a banking economy: Transmission mechanism and policy trade-offs.’ Journal of Financial Stability, 18, 78–90. Barth J.R., and Miller S.M. (2018). ‘Benefits and costs of a higher bank “leverage ratio”.’ Journal of Financial Stability, 38, 37–52. Barth J.R., Caprio G., and Levine R. (2013). ‘Bank regulation and supervision in 180 countries from 1999 to 2011.’ Journal of Financial Economic Policy, 5(2), 111–219. Barth J.R., Nolle D.E., Phumiwasana T., and Yago G. (2003). A Cross-country Analysis of the Bank Supervisory Framework and Bank Performance. Malden. Basci, E. (2010). ‘Central banks and financial stability.’ Presentation at ECB High-Level Policy Workshop on Macroprudential Policy, Frankfurt. Beck T., and Demirguc-Kunt A. (2006). ‘Small and medium-size enterprises: Access to finance as a growth constraint.’ Journal of Banking and Finance, 30(11), 2931–2943. Benati L., and Goodhart C. (2010). ‘Monetary policy regimes and economic performance: The historical record 1979–2008.’ In B.M. Friedman and M. Woodford (eds), Handbook of Monetary Economics, Vol. 3. Elsevier, 1159–1236. Berger W., and Kißmer F. (2013). ‘Central bank independence and financial stability: A tale perfect harmony?’ European Journal of Political Economy, 31, 109–118. Berman S., and McNamara K.R. (1999). ‘Bank on Democracy: Why Central Banks Need Public Oversight.’ Foreign Affairs, 78(2), 2–8. Bernanke B.S. (1983). ‘Non-monetary effects of the financial crisis in the propagation of the Great Depression.’ American Economic Review, 73(3), 257–276. Bernanke, B.S., and Blinder A.S. (1992). ‘The Federal Funds rate and the channels of monetary transmission.’ American Economic Review, 82, 901–921. Bernanke B., and Gertler M. (2000). ‘Monetary policy and asset price volatility.’ NBER Working Paper No. 7559. Blinder A.S. (2006). ‘Monetary policy today: Sixteen questions and about twelve answers.’ Paper presented at the Banco de España Conference ‘Central Banks in the 21st Century’, Madrid, June 8–9. Borio C.E.V., and Lowe P.W. (2002). ‘Asset prices, financial and monetary stability: exploring the nexus.’ BIS Working Paper No. 114. Borio C.E.V., and Shim I. (2008). ‘What can (macro-) prudential policy do to support monetary policy?’ BIS Working Paper No. 242. Borio C.E.V., and White W.R. (2004). ‘Whither monetary and financial stability? The implications of evolving political regimes.’ BIS Working Paper No. 147. Brei M., and Moreno R. (2019). ‘Reserve requirements and capital flows in Latin America.’ Journal of International Money and Finance, 99, 102079. Brown, R.L., Durbin J. and Evans J.M. (1975). ‘Techniques for testing the constancy of regression relationships over time.’ Journal of the Royal Statistical Society: Series B (Methodological), 37(2),149–163. Buiter W.H. (2014). ‘Central banks: Powerful, political and unaccountable?’ CEPR Discussion Paper No. DP10223. Cantú C., Gondo R.M., and Martinez B. (2019). ‘Reserve requirements as a financial stability instrument.’ Banco Central de Reserva del Perú, No. 2019-014. Cassola N., and Morana C. (2004). ‘Monetary policy and the stock market in the euro area.’ Journal of Policy Modeling, 26(3), 387–399. Central Bank of Malaysia (2011). ‘Increase in the Statutory Reserve Requirement (SRR) Ratio.’ Press statement, March 11. Čihák M. (2007). Central Bank Independence and Financial Stability. IMF, Mimeo. Čihák M., Demirgüč-Kunt A., Feyen E., and Levine R. (2013). ‘Financial development in 205 economies, 1960 to 2010.’ Journal of Financial Perspectives, 1(2), 7–36. Claessens S., and Valencia F. (2013). ‘The interaction between monetary and macro prudential policies.’ VoxEU. org, 14 March. Cochrane J.H., and Taylor J.B. (2016). Central Bank Governance and Oversight Reform. Hoover Institution Press. Crowe C., and Meade E.E. (2008). ‘Central bank independence and transparency: Evolution and effectiveness.’ European Journal of Political Economy, 24(4), 763–777. Cukierman A. (1992). Central Bank Strategy, Credibility and Independence: Theory and Evidence. MIT Press. De Gregorio J. (2009). ‘Implementation of inflation targets in emerging markets.’ In G. Hammond, R. Kanbur, and E. Prasad (eds), Monetary Policy Frameworks for Emerging Markets. Edward Elgar Publishing, 40–58. De Gregorio J., and Guidotti P.E. (1995). ‘Financial development and economic growth.’ World Development, 23(3), 433–448.

NGUYEN 9781800880894 PRINT.indd 292

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­293 Demirgüč-Kunt A., and Detragiache E. (1998). ‘The determinants of the banking crisis in developing and developed countries.’ IMF Economic Review, 45(1), 81–109. Dermine J. (2015). ‘Basel III leverage ratio requirement and the probability of bank runs.’ Journal of Banking and Finance, 53, 266–277. Diamond D.W., and Kashyap A.K. (2016). ‘Liquidity requirements, liquidity choice and financial stability.’ In J.B. Taylor and H. Uhlig (eds), Handbook of Macroeconomics, Vol. 2. Elsevier, 2263–2303. Doumpos M., Gaganis C., and Pasiouras F. (2015). ‘Central bank independence, financial supervisory structure and bank soundness: An empirical analysis around the crisis.’ Journal of Banking and Finance, 61(1), S69–S83. Dvorsky S. (2000). ‘Measuring central bank independence in selected transition countries and the disinflation process.’ BOFIT Discussion Paper No. 13/2000. Eichengreen B., and Hausmann R. (1999). ‘Exchange rates and financial fragility.’ NBER Working Paper No. 7418. Esqueda O.A., Assefa T.A., and Mollick A.V. (2012). ‘Financial globalization and stock market risk.’ Journal of International Financial Markets, Institutions and Money, 22(1), 87–102. Federico P., Vegh C.A., and Vuletin G. (2014). ‘Reserve requirement policy over the business cycle.’ NBER Working Paper No. 20612. Fischer, S. (2015). ‘The Federal Reserve and the global economy.’ IMF Economic Review, 63, 8–21. Fisher I. (1933). ‘The debt-deflation theory of great depressions.’ Econometrica, 1(4), 337–357. Foos D., Norden L., and Weber M. (2010). ‘Loan growth and riskiness of banks.’ Journal of Banking and Finance, 34(12), 2929–2940. Foot, M.D.K.W. (2003). ‘Protecting financial stability: How good are we at it?’ Speech given at University of Birmingham, June 6, available at http://www.fsa.gov.uk/Pages/Library/Communication/Speeches/2003/sp​13​ 3.shtml. Förch, T. and Sunde, U. (2012). ‘Central bank independence and stock market returns in emerging economies.’ Economics Letters, 115(1), 77–80. Gadanecz B., and Jayaram K. (2008). ‘Measures of financial stability: A review.’ Irving Fisher Committee Bulletin, 31, 365–383. Galati G., Heemeijer P., and Moessner R. (2011). ‘How do inflation expectations form? New insights from a high-frequency survey.’ BIS Working Paper No. 349. Garcia Herrero, A., and Del Rio, P. (2003). ‘Financial stability and the design of monetary policy.’ Banco de España Working Paper 0315. Gauthier C., Lehar A., and Souissi M. (2012). ‘Macroprudential capital requirements and systemic risk.’ Journal of Financial Intermediation, 21(4), 594–618. Gharleghi B. (2019). ‘Economic integration and macroeconomic shocks in Eurasia.’ Munich University Library. Glocker C., and Towbin P. (2012). ‘Reserve requirements for price and financial stability: When are they effective?’ Banque de France Working Paper No. 363. Goldstein M., Kaminsky G.L., and Reinhart C.M. (2000). Assessing Financial Vulnerability: An Early Warning System for Emerging Markets. Institute for International Economics. Grill M., Lang J.H., and Smith J. (2018). ‘The leverage ratio, risk-taking and bank stability.’ Bank of England Working Paper No. 766. Guttentag J.M., and Herring R.J. (1986). Disaster Myopia in International Banking. Princeton University. Ha J., Stocker M., and Yilmazkuday H. (2019). ‘Inflation and exchange rate pass-through.’ World Bank Group, Policy Research Papers, 8 April. Haan W., Ellison M., Ilzetzki E., McMahon M., and Reis R. (2017). ‘The future of central bank independence: Results of the CFM–CEPR survey.’ VoxEU.org. Haldane A.G. (2009). ‘Why banks failed the stress test.’ Bank of England, BIS Review. Herring R.J., and Vankudre P. (1987). ‘Growth opportunities and risk-taking by financial intermediaries.’ Journal of Finance, 42(3), 583–599. Hilbers P., Ötker-Robe I., Pazarbasioglu C., and Johnsen G. (2005). ‘Assessing and managing rapid credit growth and the role of supervisory and prudential policies.’ IMF Working Paper No. 05/151. Hilderbrand P.M. (2008). ‘Is the Basel 2 enough? The benefits of the leverage ratio.’ Markets Group Lecture, London School of Economics. Hutchison M., and McDill K. (1999). ‘Are all banking crises alike? The Japanese experience in international comparison.’ Journal of the Japanese and International Economies, 13(3), 155–180. IMF (2019). Financial Soundness Indicators Compilation Guide (2019 FSI Guide). Available at https://www.imf. org/en/Data/Statistics/FSI-guide. Ioannidis C., and Kontonikas A. (2006). Monetary Policy and the Stock Market: Some International Evidence. Glasgow University. Issing O. (1993). Central Bank Independence and Monetary Stability. Institute of Economic Affairs. Issing O. (2003). ‘Monetary and financial stability: Is there a trade-off ?’ Conference on ‘Monetary Stability, Financial Stability and the Business Cycle,’ March 28–29, Bank for International Settlements, Basle.

NGUYEN 9781800880894 PRINT.indd 293

21/09/2022 13:14

294  Handbook of banking and finance in emerging markets Jappelli T., and Pagano M. (1994). ‘Saving, growth and liquidity constraints.’ Quarterly Journal of Economics, 109(1), 83–109. Kaiser H.F. (1960). ‘The application of electronic computers to factor analysis.’ Educational and Psychological Measurement, 20(1), 141–151. Kalecki M. (1965). ‘Econometric model and historical materialism.’ In On Political Economy and Econometrics: Essays in Honour of Oskar Lange, Pergamon Press/Polish Scientific Publishers, 233–238. Kaminsky G.L., and Reinhart C.M. (1999). ‘The twin crises: The causes of banking and balance-of-payments problems.’ American Economic Review, 89(3), 473–500. Kapoor S., and Buiter W. (2020). ‘To fight the COVID pandemic, policymakers must move fast and break taboos.’ In Europe in the Time of COVID-19, Vol. 1. Centre for Economic Policy Research, 250–254. Kashyap A.K., and Stein J.C. (2012). ‘The optimal conduct of monetary policy with interest on reserves.’ American Economic Journal: Macroeconomics, 4(1), 266–282. Khushiram S. (2020). ‘COVID-19 bill: No holds barred.’ L’express.mu (May), Article 376908. Kiema I., and Jokivuolle E. (2014). ‘Does a leverage ratio requirement increase bank stability?’ Journal of Banking and Finance, 39, 240–254. Kim S. and Mehrota A. (2017). ‘Managing price and financial stability objectives in inflation targeting economies in Asia and the Pacific.’ Journal of Financial Stability, 29, 106–116. Klomp J. and Haan J. (2009). Central bank independence and financial instability. Journal of Financial Stability, 5(4), 321–338. Lamers M., Mergaerts F., Meulemar E., and Vennet R.V. (2016). ‘The trade-off between monetary policy and bank stability.’ National Bank of Belgium Working Paper No. 308. Lim CH., Costa A., Columba F., Kongsamut P., Otani A., Saiyid M., Wezel T., and Wu X. (2011). ‘Macroprudential policy: What instruments and how to use them? Lessons from country experiences.’ IMF Working Paper No. 11/238. Loazya N., and Ranciere R. (2004). ‘Financial development, financial fragility, and growth.’ World Bank Group Policy Research Working Paper No. 3431. Marmefelt T. (2020). ‘COVID-19 and economic policy toward the new normal: A monetary-fiscal nexus after the crisis?’ Policy Department for Economic, Scientific and Quality of Life Policies, European Parliament (November). Masciandaro D., and Romelli D. (2015). ‘Ups and downs of Central bank independence from the Great Inflation to the Great Recession: Theory, institutions and empirics.’ Financial History Review, 22(3), 259–289. McKinnon, R.I. (1973). Money and Capital in Economic Development. Brookings Institution. Minsky H.P. (1977). ‘The financial instability hypothesis: An interpretation of Keynes and an alternative to the “standard” theory.’ Nebraska Journal of Economics and Business, 20(1), 20–27. Minsky H.P. (1986). Stabilizing an Unstable Economy. Yale University Press. Mishkin F.S. (1996). ‘The channels of monetary transmission: Lessons for monetary policy.’ NBER, Economic Fluctuation and Monetary Economics Working Paper 5464. Mishkin F.S. (2009). ‘Is monetary policy effective during financial crises?’ American Economic Review, 99(2), 573–577. Mishkin F.S. (2014). ‘Central banking after the crisis.’ In S. Bauducco, L. Christiano, and C. Raddatz (eds), Macroeconomic and Financial Stability: Challenges for Monetary Policy. Central Bank of Chile, 23–59. Monnet E., and Vari M. (2019). ‘Liquidity ratios as monetary policy tools: Some historical lessons for macroprudential policy.’ IMF Working Paper WP/19/176. Obstfeld M., and Taylor A.M. (2004). Global Capital Markets: Integration, Crisis and Growth. Cambridge University Press. Oosterloo S. and Haan J. (2004). ‘Central banks and financial stability: A survey.’ Journal of Financial Stability, 1(2), 257–273. Papadamou S., Sidiropoulos M., and Spyromitros E. (2017). ‘Does central bank independence affect stock market volatility?’ Research in International Business and Finance, 42, 885–864. Papadimitriou D.B. (2008). ‘Promoting equality through an employment of last resort policy.’ Levy Economics Institute Working Paper No. 545. Pokorny D. (1978). ‘Smith and Walras: Two theories of science.’ Canadian Journal of Economics, 11(3), 387–403. Quintyn M., and Taylor M.W. (2003). ‘Regulatory and supervisory independence and financial stability.’ CESifo Economic Studies, 49(2), 59–294. Reichlin L., and Baldwin R. (2013). ‘Is inflation targeting dead? Central banking after the crisis.’ VoxEU CEPR 14 April. Schinasi G.J. (2004). ‘Defining financial stability.’ IMF Working Paper No. 04/187. Schnabel I. (2020). ‘The ECB’s independence in times of mounting public debt.’ Contribution by Isabel Schnabel, Member of the Executive Board of the ECB. Frankfurter Allgemeine Zeitung, 10 October. Schumpeter J.A. (1961). The Theory of Economic Development. Harvard University Press.

NGUYEN 9781800880894 PRINT.indd 294

21/09/2022 13:14

Central bank independence and financial stability in Mauritius  ­295 Seetanah B., Sannassee R.V., and Dilmohamed S. (2014). ‘Central bank independence and inflation: Evidence from Mauritius.’ African Development Review, 26(4), 561–571. Shrestha M. and Bhatta G. (2020). ‘Selecting appropriate methodological framework for time series data analysis.’ Journal of Banking and Data Science, 4(2), 71–89. Stein J.C. (2013). ‘Overheating in credit markets: origins, measurement, and policy responses.’ Remarks at ‘Restoring household financial stability after the great recession: Why household balance sheets matter’, Federal Reserve Bank of St. Louis research symposium, 7 February. Struthers J. (2003). ‘The Mckinnon-Shaw hypothesis: thirty years on: a review of recent developments in financial liberalization theory.’ Semanticscholar.org, Corpus ID: 15238911. Svensson L. (2010). ‘Inflation targeting.’ In B.M. Friedman and M. Woodford (eds), Handbook of Monetary Economics, Vol. 3. Elsevier, 1237–1302. Ueda K., and Valencia F. (2013). ‘Central bank independence and macro-prudential regulation.’ Economic Letters, 125(2), 327–330. Umutlu M., Akdeniz L., and Altay-Salih A. (2010). ‘The degree of financial liberalization and aggregated stockreturn volatility in emerging markets.’ Journal of Banking and Finance, 34(3), 509–521. Valla N., Saes-Escorbiac B., and Tiesset M. (2006). ‘Bank liquidity and financial stability.’ Financial Stability Review, Banque de France, 9(December), 89–104. Willem Van Den End J., and Kruidhof M. (2013). ‘Modelling the liquidity ratio as macroprudential instrument.’ Journal of Banking Regulation, 14, 91–106. Yashiv E. (2020). ‘Breaking the taboo: The political economy of COVID-motivated helicopter drops.’ In Europe in the Time of COVID-19, Vol. 1. Centre for Economic Policy Research, 240–244. Zhang C., Zhang D., and Breece J. (2011). ‘Financial crisis, monetary policy and stock market volatility in China.’ Annals of Economics and Finance, 12(2), 371–388.

NGUYEN 9781800880894 PRINT.indd 295

21/09/2022 13:14

296  Handbook of banking and finance in emerging markets

APPENDIX Table 15A.1  Determining TOR for the Bank of Mauritius Complete autonomy

Complete dependence

BoM Act 2004 (Present context)

Governor appointed  by Governor holds any   other office

Board of Directors No

Members of executive branch No rule against

Governor can be   removed from office by Governor’s term of office

Board of Directors

Executive branch

>8 years