i Enterprise Risk Management, Corporate Governance, Performance and Risk-Taking Behaviour of the Insurance Industry: Empirical Evidence from Ghana and South Africa By Sylvester Senyo Horvey Doctoral thesis submitted in fulfilment of the requirements for the award of Doctor of Philosophy The Graduate School of Business Administration, Wits Business School University of the Witwatersrand Supervisor: Prof. Jones Odei-Mensah ©Sylvester Senyo Horvey, August 2022 ii ABSTRACT The growing complexities in the business environment have led to the adoption of enterprise risk management (ERM). ERM is a new approach to managing organisational risks holistically to achieve its goals. Regardless of the diversities in the business environment, ERM has become an essential factor for businesses and is believed to enhance shareholder value. Despite the growing number of studies on ERM, literature suffers some limitations regarding its proxies and inconclusive results between ERM and performance. This study adopts a more comprehensive measurement of ERM, which captures various characteristics (such as risk governance, operational mechanisms, and quality of risk oversight) within the risk ecosystem. The study uses a panel regression technique on a sample of 33 and 63 insurers from Ghana and South Africa, respectively, covering 2015-2019. This thesis is centred on four thematic papers. Each focuses on a specific subject (s) at the heart of the problems or research questions being investigated. The first paper provides a comprehensive and systematic literature review on the measurement and performance of ERM. Google Scholar was the primary search tool for ERM literature spanning 2001 to 2020, and papers listed in SCImago journal ranking were discussed. The study finds that most studies rely on secondary sources, particularly the Chief Risk Officer’s appointment, as a simple ERM proxy. This is widely adopted in the literature due to the difficulty in assessing ERM information. The study recommends that empirical measurement of ERM rely on both primary and secondary data as they complement each other and allow more insight and factors to be considered for a robust ERM measurement. In terms of performance, the ERM literature reveals mixed findings, but enough evidence supports the assertion that ERM enhances firm profitability and value. The study suggests that scholars consider examining the ERM-performance relationship in emerging economies as most of these studies centred on the US and European economies. The second paper analyses the determinants of ERM adoption in Ghana and South Africa using a panel logistic regression technique. Building on the contingency theory, the study posits that several factors contribute to ERM adoption. The study finds that firm size, ownership, leverage, industrial diversification and the type of audit firm are positively associated with ERM adoption in both countries. Findings from the quantile regression also highlight that the initial levels of size, profitability and leverage reduce ERM adoption, and an extreme increase in these factors promotes iii ERM adoption, which implies a nonlinear direct U-shape relationship. On the contrary, the study sees an inverted U-shape for return on assets and leverage for Ghana. Industrial diversification, Big4 audit companies and ownership show consistent patterns of a significant positive effect on ERM adoption at different quantiles for both samples. The findings support the fact that insurers could improve their risk management system by considering the factors that significantly affect them. The third paper first examines the impact of ERM on insurance performance (underwriting performance and Return on Assets) and second investigates how corporate governance (CG) characteristics such as the board size, board independence, and gender diversity interact with ERM in affecting insurance performance. The major findings are summarised as follows: (1) a positive relationship exists between ERM and insurance performance for both countries; and (2) board size, board independence and gender diversity interact with ERM in affecting underwriting performance and return on assets. This was mostly positive and significant in both samples. The study suggests that insurers interested in ensuring an effective ERM system should leverage these corporate governance factors to appreciate the overall impact of ERM on performance. In the final paper, the study examines the linear and non-linear effects of ERM and CG on risk- taking behaviour. The result from the linear regression elicits a significant positive relationship between ERM and risk-taking for both countries, implying that insurers with a strong ERM system are more likely to pursue higher risks. The empirical evidence also suggests that board size and board independence have a significant positive impact on risk-taking for both samples. In contrast, gender diversity shows an inverse relationship with risk-taking. Using the dynamic panel regression by Seo et al. (2019), the study confirms non-linearities between ERM, CG and risk- taking. Evidence from the South African sample indicates that ERM significantly increases insurers’ risk-taking beyond the threshold level. Again, the South African sample shows significant threshold levels for board size, gender diversity and board independence at 10.03, 0.274 and 0.547, respectively. The Ghanaian sample also documents significant threshold levels at 7, 0.286, and 0.692. The study recommends that insurers consider the significant threshold levels to determine the optimum level of risk that must be pursued. iv Keywords: Enterprise Risk Management, Determinants, Adoption, Corporate Governance, Threshold, Moderators, Performance, Insurance, Risk-Taking. JEL Classification: C2, C35, C36, G3, G22, G32 v PUBLICATIONS AND RESEARCH OUTPUT Prior to the submission of this thesis, portions of the thesis and other related areas have been published in peer-reviewed journals, while others are under review. Peer-reviewed journal publications Horvey, S. S., and Ankamah, J. (2020). Enterprise risk management and firm performance: Empirical evidence from Ghana equity market. Cogent Economics & Finance, 8(1), 1840102. Papers submitted/under peer-review Horvey, S. S. and Odei-Mensah, J. The Measurements and Performance of Enterprise Risk Management: A Comprehensive Literature Review. Journal of Risk Research (221445852) Horvey, S. S., Odei-Mensah, J. and Mushai, A. Th Determinants of Life Insurance Profitability in South Africa: New Evidence from a Dynamic Panel Threshold Estimation Technique. International Journal of Emerging Markets (IJOEM-08-2022-1225). Horvey, S. S., Mushai, A. and MacGregor, A. The Determinants of Firm-Level Enterprise Risk Management Adoption: Literature Review and Future Directions. Risk Management (RMGT-D- 21-00163). Dube, A. and Horvey, S. S. Institutional quality and renewable energy capital flows in Sub– Saharan Africa. Journal of Energy in Southern Africa (12297). Quaye, E. S., Ngcamu, L., Horvey, S. S. and Jarava, D. C. Personality traits, money attitudes, and consumer decision-making styles as predictors of investment products choice in South Africa. Journal of Consumer Behaviour (JCB-22-375) Conference(s) Horvey, S. S., and Odei-Mensah, J. (2022). Enterprise Risk Management and Insurance Performance in Ghana and South Africa: The Moderating Role of Corporate Governance. 6th International Conference on Applied Theory, Macro and Empirical Finance, Thessaloniki, Greece, 18-19 April 2022. Horvey, S. S., and Ouma, N. W. (2022). Retail Philanthropy: Point of Sale Giving in South Africa 3rd African Philanthropy Conference, Johannesburg, South Africa, 2-4, August, 2022. vi Horvey, S. S., Odei-Mensah, J. and Mushai, A. (2022). The Determinants of Life Insurance Profitability in South Africa: New Evidence from a Dynamic Panel Threshold Estimation Technique. 17th Economics & Finance Conference, Istanbul, Turkey, 5-7, September, 2022, upcoming event. Research Awards 2021 Third Prize Centre on African Philanthropy and Social Investment, October 4-9. 2021 Runners-up Award Wits PhD Seminar, University of the Witwatersrand, August 30-31. vii DECLARATION I, Sylvester Senyo Horvey, with student number 2395495, hereby declare that this research report is my own work except as shown in the references and acknowledgements. It is submitted in fulfilment of the requirements for the award of Doctor of Philosophy at the University of the Witwatersrand, Johannesburg, South Africa. It has not been submitted before for any degree or examination in this or any other university. …………………………………. Sylvester Senyo Horvey Signed at: Wits Business School On the 8th August 2022 viii DEDICATION To my late dad (Richard Kofi Horvey), mother(s) (Rosca Adwoa Mansah Jones, Leticia Horvey), sisters (Dela, Leticia and Rosca) and in-laws (Ernest and Rufai). ix ACKNOWLEDGEMENTS My ultimate gratitude goes to God the Father and our Lord Jesus Christ for His enabling ability and abundant grace bestowed on me. The PhD journey has been a long and extremely challenging one, but His perfect love towards me has brought me this far. For by strength shall no man prevail, but with God, all things are possible. I owe it a duty to extend my sincere appreciation to my supervisor, Prof. Jones Odei-Mensah, for all his support, tutelage and brotherly love throughout my doctoral journey. His profound encouragement at every stage of my study provided the yardstick that kept me going even when the going was tough. His unselfish mentorship and constructive feedback on every draft brought me this far. Your suggestions were superb, and words alone cannot show my appreciation to you, but I wish above all things that you may prosper and be in good health even as your soul prospers. I am also thankful to Prof Odongo Kodongo (PhD Director, Wits Business School) and Prof Paul Imhotep Alagidede for their unflinching support. I sincerely appreciate the selfless kindness and support from Dr George Tweneboah (Senior Lecturer, Wits Business School), Dr Arthur Linke (University of Stellenbosch), Dr Albert Mushai (Head of Insurance and Risk Management division, Wits University), Mrs Agata MacGregor (Lecturer, School of Business Sciences), Prof Robert Vivian (Professor, School of Business Sciences). I also thank Dr Eric Ofosu-Hene (De Montford University, United Kingdom), Prof Godfred A. Bokpin (Dean of Students, University of Ghana), and Dr Albert Gemegah (Dean of Graduate School, Wisconsin International University, Ghana) for the guidance and encouragement. I have also received vast support from the Wits Business School writing retreat team. I want to use this special opportunity to thank Dr Pia Lamberti (Coordinator), Prof Terry Carmichael (Wits Business School), and Mrs Jean Moore (Wits University) for their diverse support in writing this thesis. I also acknowledge the financial support from the University of the Witwatersrand through the Postgraduate Merit Award and the Bradlow Foundation PhD Scholarship. A big thanks go to Mrs Mmabatho Leeuw (PhD Programme Manager, Wits Business School), Ms Jennifer Mgolodela (Faculty Officer, Wits Business School), and all staff at the Wits Business School. I say a big thank you to Dr Emmanuel Silva Quaye (Wits Business School), Mr Emmanuel Kwame Korsah, Dr Dennis Osei Boahene, Mr Prince Akomiah Sekyi, Mr Jacob Ankamah and Mr Raynold x Techie Menson, who have been true friends, brothers, and a source of encouragement. God bless you for the support. I also thank Dr Maurice Omane Agyepong (Ghana) for his support. To my friends, Raymond Agyepong Antwi, Bismark Opoku Appiah, Dr Elvis Okoffo Dartey, Partison Dartey, Richard Asare Boateng, Edmund J. Mbreku, Dr Samuel Nunoo (University of Ghana), Emmanuel Quarshie, Dr Disraeli Ohene Darko, Gabriel Teye-Ali, Dr Baah Aye Kusi, Mr Ebenezer Emmanuel Owusu, Dr Daniel Kwayisi (University of Ghana), Mr Isaac Brobbey (PENSA Ghana Coordinator), Mrs Deborah Cudjoe, Mr Senyo Kofi Cudjoe, Mr Jakubose Sibanda, Ms Evelyn Nyarko, Dr Monica Cudjoe and Mrs Gifty Akutek and husband. I will also like to acknowledge Aps M. C. Asiedu, Aps Alexander Boakye Yiadom, Prophet Don Stewart, Pastor Matthew Owusu, Pastor Ebenezer Hagan, Pastor Ebenezer Korankye Agyapong, Pastor Kwadwo Obeng, Pastor Bernard Boakye Anyimadu, Elder(s) Foli, Johnson Akpabli, Enoch and Ernest Nyarko, Mr Frimpong and their wives for their prayers and encouragement throughout this journey. Finally, to all those who have contributed in diverse ways towards my education until now, whom I have not named here, I say a very big thank you to all of you. I pray the Almighty GOD to shower you with His eternal grace and blessings. The usual caveat applies. xi TABLE OF CONTENTS Contents ABSTRACT .................................................................................................................................................. ii PUBLICATIONS AND RESEARCH OUTPUT ......................................................................................... v DECLARATION ........................................................................................................................................ vii DEDICATION ........................................................................................................................................... viii ACKNOWLEDGEMENTS ......................................................................................................................... ix TABLE OF CONTENTS ............................................................................................................................. xi LIST OF TABLES ..................................................................................................................................... xvii LIST OF FIGURES .................................................................................................................................... xix LIST OF ABBREVIATIONS ...................................................................................................................... xx DEFINITION OF KEY TERMS ............................................................................................................... xxii CHAPTER ONE ........................................................................................................................................... 1 INTRODUCTION ........................................................................................................................................ 1 1.1 Background of the study ......................................................................................................................... 1 1.2 Overview of the Insurance Industry in Ghana and South Africa ............................................................ 3 1.3 Problem Statement .................................................................................................................................. 6 1.3.1 ERM determinants ........................................................................................................................... 6 1.3.2 ERM, Corporate Governance and Performance ............................................................................... 8 1.3.3 ERM, Corporate governance and firm risk-taking ........................................................................... 9 1.4 Research Objectives ........................................................................................................................ 10 1.5 Research Questions ............................................................................................................................... 11 1.6 Justification ........................................................................................................................................... 11 1.7 Thesis Structure .................................................................................................................................... 13 CHAPTER TWO ........................................................................................................................................ 15 THE MEASUREMENTS AND PERFORMANCE OF ENTERPRISE RISK MANAGEMENT: A COMPREHENSIVE LITERATURE REVIEW ......................................................................................... 15 2.1 Introduction ........................................................................................................................................... 15 2.2 Methodology ......................................................................................................................................... 18 2.3 Journal Credibility ................................................................................................................................ 18 2.4 Subject Areas ........................................................................................................................................ 19 2.5 Literature review on ERM measurement .............................................................................................. 19 2.5.1 CRO/ERM Keywords: ................................................................................................................... 20 xii 2.5.2 RIMS Risk Maturity Model: .......................................................................................................... 20 2.5.3 S&P Rating: ................................................................................................................................... 21 2.5.4 COSO Framework: ........................................................................................................................ 22 2.5.5 Survey: ........................................................................................................................................... 22 2.5.6 Hazard Ratio: ................................................................................................................................. 23 2.5.7 Risk Governance: ........................................................................................................................... 23 2.6 Literature review on the impact of ERM on firm performance ............................................................ 24 2.6.1 Trend Analysis of ERM papers ...................................................................................................... 24 2.6.2 Geographical context ..................................................................................................................... 25 2.6.3 Estimation Method ......................................................................................................................... 26 2.6.3.1 Ordinary least square estimation ............................................................................................. 26 2.6.3.2 Maximum likelihood ............................................................................................................... 27 2.6.3.3 Structural equation modelling (SEM) ..................................................................................... 27 2.6.3.4 Generalised least square estimation method (GLS) ................................................................ 27 2.6.3.5 Linear regression ..................................................................................................................... 28 2.6.4 Performance indicators .................................................................................................................. 28 2.7 Discussion of Empirical Studies ........................................................................................................... 29 2.8 Conclusion ............................................................................................................................................ 36 2.9 Limitations ............................................................................................................................................ 37 2.10 Recommendations ............................................................................................................................... 37 CHAPTER THREE .................................................................................................................................... 42 THE DETERMINANTS OF ENTERPRISE RISK MANAGEMENT ADOPTION AMONG INSURERS IN GHANA AND SOUTH AFRICA: A QUANTILE REGRESSION APPROACH ............................... 42 3.1 Introduction ........................................................................................................................................... 42 3.2 Literature review ................................................................................................................................... 46 3.2.1 Evolution of ERM .......................................................................................................................... 46 3.2.2 Hypothesis Development: The Determinants of ERM adoption ................................................... 49 3.2.2.1 Industrial Diversification ........................................................................................................ 49 3.2.2.2 Firm size .................................................................................................................................. 50 3.2.2.3 Auditor type ............................................................................................................................ 50 3.2.2.4 Institutional Ownership ........................................................................................................... 51 3.2.2.5 Profitability ............................................................................................................................. 51 3.2.2.6 Industry type ........................................................................................................................... 52 xiii 3.2.2.7 Financial leverage ................................................................................................................... 53 3.3 Conceptual Model ................................................................................................................................. 53 3.4 Method .................................................................................................................................................. 57 3.5 Data Description and Source................................................................................................................. 57 3.6 Empirical Strategy ................................................................................................................................ 57 3.6.1 Linear Regression .......................................................................................................................... 57 3.6.2 Logistic Quantile Regression Model .............................................................................................. 59 3.7 Description of the Determinants Variables ........................................................................................... 59 3.8 Empirical Results .................................................................................................................................. 61 3.9 Descriptive statistics ............................................................................................................................. 61 3.10 Univariate Analysis ............................................................................................................................. 64 3.11 Sensitivity Analysis ............................................................................................................................ 65 3.12 Presentation of Empirical findings ...................................................................................................... 67 3.12.1 Linear Regression Analysis ......................................................................................................... 67 3.12.2 Quantile Regression Results ........................................................................................................ 69 3.13 Discussion of Empirical Results ......................................................................................................... 73 3.13.1 Linear Regression ........................................................................................................................ 73 3.13.2 Quantile Regression ..................................................................................................................... 76 3.14 Summary and Conclusion ................................................................................................................... 77 3.15 Managerial Recommendations ............................................................................................................ 78 3.16 Limitations and Future Research ........................................................................................................ 79 Appendix ..................................................................................................................................................... 80 CHAPTER FOUR ....................................................................................................................................... 82 ENTERPRISE RISK MANAGEMENT AND INSURANCE PERFORMANCE IN GHANA AND SOUTH AFRICA: THE MODERATING ROLE OF CORPORATE GOVERNANCE ........................... 82 4.1 Introduction ........................................................................................................................................... 82 4.2 Literature Review .................................................................................................................................. 85 4.3 Theoretical Review ............................................................................................................................... 85 4.3.1 Agency theory ................................................................................................................................ 85 4.3.2 Portfolio theory .............................................................................................................................. 86 4.3.3 Contingency Theory ....................................................................................................................... 88 4.4 Traditional Risk Management System versus Enterprise Risk Management System ........................... 89 4.4.1 Traditional Risk Management (TRM) system ............................................................................... 89 xiv 4.4.2 Enterprise Risk Management (ERM) system ................................................................................. 90 4.5 Empirical Literature Review on ERM and Performance ...................................................................... 93 4.5.1 ERM in Insurance .......................................................................................................................... 95 4.6 Empirical Review on the Moderating role of Corporate Governance on the ERM and Performance relationship .................................................................................................................................................. 98 4.6.1 Board Size ...................................................................................................................................... 99 4.6.2 Board Independence ..................................................................................................................... 100 4.6.3 Gender Diversity .......................................................................................................................... 101 4.7 Conceptual Framework ....................................................................................................................... 102 4.8 Methodology ....................................................................................................................................... 103 4.9 Data and Sample ................................................................................................................................. 103 4.10 Description of Variables ................................................................................................................... 103 4.10.1 Dependent Variables .................................................................................................................. 103 4.10.2 Independent Variable: Enterprise Risk Management Index (ERMI) ......................................... 104 4.10.3 Moderating Variable .................................................................................................................. 105 4.10.4 Control Variables ....................................................................................................................... 105 4.11 Empirical Model ............................................................................................................................... 106 4.12 Analysis and Discussion ................................................................................................................... 108 4.12.1 Descriptive Statistics .................................................................................................................. 108 4.12.2 ERM Features ............................................................................................................................ 110 4.13 Empirical Results .............................................................................................................................. 114 4.13.1 ERM and Insurance Performance .............................................................................................. 114 4.13.2 The Moderating role of CG on the ERM-Performance Relationship ........................................ 119 4.13.3 Diagnostic Checks ..................................................................................................................... 125 4.14 Conclusion and Recommendations ................................................................................................... 126 4.14.1 Conclusion ................................................................................................................................. 126 4.14.2 Managerial Implications ............................................................................................................ 127 4.14.3 Study limitations ........................................................................................................................ 128 4.14.4 Recommendations for Future Studies ........................................................................................ 128 Appendix ................................................................................................................................................... 129 CHAPTER FIVE ...................................................................................................................................... 131 ENTERPRISE RISK MANAGEMENT, CORPORATE GOVERNANCE AND INSURANCE RISK- TAKING IN GHANA AND SOUTH AFRICA: EVIDENCE FROM A DYNAMIC PANEL THRESHOLD ........................................................................................................................................... 131 xv 5.1 Introduction ......................................................................................................................................... 131 5.2 Literature Review ................................................................................................................................ 134 5.3 Agency theory ..................................................................................................................................... 134 5.4 Enterprise Risk Management and Risk-Taking .................................................................................. 136 5.5 Corporate Governance and Risk-Taking ............................................................................................. 137 5.5.1 Board Size and Risk-Taking ........................................................................................................ 139 5.5.2 Gender and Risk-Taking .............................................................................................................. 140 5.5.3 Board Independence and Risk-Taking ......................................................................................... 141 5.6 Method ................................................................................................................................................ 143 5.7 Sample and Data Considerations ........................................................................................................ 143 5.8 Dependent Variable ............................................................................................................................ 143 5.9 Independent Variable .......................................................................................................................... 144 5.9.1 Enterprise Risk Management Index (ERMI) ............................................................................... 144 5.9.2 Corporate Governance ................................................................................................................. 146 5.9.3 Control Variables ......................................................................................................................... 146 5.10 Empirical Strategies .......................................................................................................................... 146 5.10.1 Linear relationship ..................................................................................................................... 147 5.10.2 Non-linear relationship .............................................................................................................. 148 5.11 Findings and Discussions .................................................................................................................. 148 5.12 Descriptive Statistics ......................................................................................................................... 148 5.13 Dynamic Panel Regression Analysis ................................................................................................ 154 5.13.1 The relationship between ERM and risk-taking behaviour ........................................................ 154 5.13.2 The relationship between Corporate Governance and risk-taking behaviour ............................ 155 5.14 Dynamic Panel Threshold Regression Analysis ............................................................................... 158 5.14.1 ERM and risk-taking behaviour ................................................................................................. 158 5.14.2 Corporate Governance and risk-taking behaviour ..................................................................... 161 5.15 Robustness Check ............................................................................................................................. 163 5.16 Conclusion ........................................................................................................................................ 164 5.17 Recommendations ............................................................................................................................. 166 5.18 Limitations and Future Studies ......................................................................................................... 166 CHAPTER SIX ......................................................................................................................................... 167 SUMMARY, CONCLUSION AND POLICY RECOMMENDATIONS ................................................ 167 6.1 Introduction ......................................................................................................................................... 167 xvi 6.2 Summary and Conclusions.................................................................................................................. 167 6.2.1 The measurement and performance of ERM ............................................................................... 167 6.2.2 Determinants of ERM adoption ................................................................................................... 168 6.2.3 The moderating role of corporate governance on the ERM-Performance nexus ......................... 169 6.2.4 The threshold dynamics between ERM, corporate governance and risk-taking .......................... 169 6.3 Policy Implications and Recommendations ........................................................................................ 170 6.4 Recommendations for further research ............................................................................................... 172 References ................................................................................................................................................. 174 Appendix: QUESTIONNAIRE ................................................................................................................ 191 ETHICS CERTIFICATE .......................................................................................................................... 195 xvii LIST OF TABLES Table 1. 1 Stylised facts about the life and non-life insurance market in Ghana and South Africa ........................................................................................................................................................ .5 Table 2. 1 Empirical Literature on the Performance of Enterprise Risk Management ..................33 Table 2. 2: List of Journals, Subject Areas, and H index………………………………………...39 Table 3. 1: Summary of empirical findings on the determinants of Enterprise Risk Management ....................................................................................................................................................... 55 Table 3. 2: Definitions of Variables and their predicted signs ..................................................... 60 Table 3. 3: Descriptive Statistics .................................................................................................. 62 Table 3. 4: Correlation Coefficients.............................................................................................. 63 Table 3. 5: Univariate Difference Across ERM Status ................................................................. 65 Table 3. 6: Sensitivity Analysis Regarding the Determinants of ERM Adoption ........................ 66 Table 3. 7: Panel Logistic Regression Results .............................................................................. 68 Table 3. 8: Quantile Regression Results (South Africa) ............................................................... 71 Table 3. 9: Quantile Regression Results (Ghana) ......................................................................... 72 Table 3. 10: ERM Adoption by years ........................................................................................... 80 Table 3. 11: Summary Statistics ................................................................................................... 81 Table 4. 1: Difference between TRM and ERM ........................................................................... 92 Table 4. 2: Descriptive Statistics ................................................................................................ 109 Table 4. 3: Distribution of ERM Features .................................................................................. 111 Table 4. 4: Mean Distribution of ERM Index ............................................................................. 112 Table 4. 5: Correlation Table ...................................................................................................... 113 Table 4. 6: Enterprise Risk Management and Insurance Performance in South Africa ............. 117 Table 4. 7: Enterprise Risk Management and Insurance Performance in Ghana ....................... 118 Table 4. 8: Interaction Effects for South Africa ......................................................................... 123 Table 4. 9: Interaction Effects for Ghana .................................................................................... 124 Table 4. 10: Diagnostic Tests...................................................................................................... 129 Table 4. 11: Variance Inflation Factor ........................................................................................ 129 xviii Table 5. 1: Descriptive statistics ................................................................................................. 150 Table 5. 2: Distribution of ERM Features .................................................................................. 151 Table 5. 3: Descriptive Statistics for ERM Categories ............................................................... 152 Table 5. 4: Correlation Matrix .................................................................................................... 153 Table 5. 5: ERM, CG and Risk-Taking ...................................................................................... 157 Table 5. 6: Threshold Model on the Effect of Enterprise Risk Management on Risk-Taking ... 160 Table 5. 7: Threshold Model on the Effect of Corporate Governance on Risk-Taking ............. 162 xix LIST OF FIGURES Figure 2. 1: Subject Areas for SCImago ....................................................................................... 19 Figure 2. 2: Empirical studies from 2008 to 2020 ........................................................................ 25 Figure 2. 3: Geographical Areas ................................................................................................... 26 Figure 3. 1: Conceptual Framework ........................................................................................................... 54 Figure 4. 1: Conceptual Model ................................................................................................... 102 Figure 4. 2: Marginal Plots ......................................................................................................... 130 xx LIST OF ABBREVIATIONS Abbreviations Meanings CAS Casualty Actuarial Society CEO Chief Executive Officer CG Corporate Governance COSO Committee of Sponsoring Organisations of the Treadway Commission CRO Chief Risk Officer ERM Enterprise Risk Management ERMI Enterprise Risk Management Index ESG Environmental, Social and Governance FSCA Financial Sector Conduct Authority GLS Generalised Least Square GMM Generalised Method of Moments GRC Governance, Risk and Compliance IM Impact Factor ISO International Organisation for Standardisation ME Marginal Effect NIC National Insurance Commission NPV Net Present Value OP Operating Mechanism OPI Operating Mechanism Index xxi POLS Pooled Ordinary Least Square QRO Risk Oversight RBS Risk-Based approach to Supervision RE Random Effect RG Risk Governance RGI Risk Governance Index RIMS Risk Management Society RO Risk Oversight ROA Return on Assets ROE Return on Equity ROI Risk Oversight Index S&P Standard and Poor’s SEM Structural Equation Modelling SJR SCImago Journal Ranking SSA Sub-Saharan Africa TRM Traditional Risk Management UK United Kingdom UNEP United Nations Environment Programme US United States VIF Variance Inflation Factor VUCA Volatile, Uncertain, Complex, Ambiguous xxii DEFINITION OF KEY TERMS Corporate Governance: Corporate governance is a set of rules, policies, and practices that govern how a company’s board of directors manages and supervises its operations. Enterprise Risk Management: A process, effected by an entity’s board of directors, management and other personnel, applied in a strategy setting and across the enterprise, designed to identify potential events that may affect the entity, and manage risk to be within its risk appetite, to provide reasonable assurance regarding the achievement of entity objective Risk Governance: Establishes the risk management system’s structure, defining roles, responsibilities and accountability while setting the guidelines for making risk management decisions Risk Oversight: Reveals the quality of risk oversight provided by the committee Risk-Taking: These are actions or decisions that involves danger or risks in order to achieve a goal Operating Mechanism: Explains the tools, techniques and procedures required for an ERM implementation 1 CHAPTER ONE INTRODUCTION 1.1 Background of the study Globalisation, an increasing number and magnitude of natural disasters, conflicts, human and social issues have contributed to the economic, business, and risk environments becoming more Volatile, Uncertain, Complex, and Ambiguous (VUCA) around the world (Gatzert & Martin, 2015). Organisations, therefore, are faced with an increasingly VUCA portfolio of risks, difficult to quantify and mitigate, which must be managed holistically whilst keeping the strategic objectives of the organisation in mind (COSO, 2017). The dynamic changing environment has led many companies to assume certain risks, such as technological changes, cyber risks, new laws and regulations, which have increased their risk-taking behaviour and risk appetite levels, leading to poor performance (Arena et al., 2010). Enterprise Risk Management (ERM) is a system that helps an organisation measure, assess and manage risk holistically, thereby protecting the firm from loss and generating value (Grace et al., 2015). Linke and Florio (2019) highlight that ERM increases the overall risk awareness of firms, leading to better results in operational and strategic decision- making processes. Unlike historical silo-based risk management, ERM is promulgated to increase firm value and performance. It enables managers to assess their risk portfolio well and determine emerging risks and opportunities for better decision-making (Gatzert & Martin, 2015; Liebenberg & Hoyt, 2003). However, this also comes with an associated cost through the employment of human resources, financial resources, and IT systems (McShane et al., 2011). ERM is closely linked to the need for Governance, Risk and Compliance (GRC) as set out by global governance standards, as well as Environmental, Social and Governance (ESG) guidelines, for example, as put forward by United Nations Environment Programme (2019). Experts have called for the integration of these factors within the risk ecosystem (Bromiley et al., 2015). ESG presents new challenges to industries and the world at large. They influence risk factors and can have a major impact on an organisation’s performance. To build a resilient industry, risks must be managed in a holistic and far-sighted manner while considering GRC issues. According to the Committee of Sponsoring Organisations of the Treadway Commission (COSO) (2017, pp. 6), “Governance sets the organisation’s tone, reinforcing the importance of, and establishing oversight responsibilities for, enterprise risk 2 management.” COSO further posits that as risk management evolves, it is very prudent to properly coordinate risk management activities and corporate governance for firm value creation. So far, the context of empirical investigations on ERM has been limited to the US and European countries without any empirical evidence in developing economies, especially African economies. Compared to developed countries, the African market is faced with higher volatilities due to political risk, interest rate and exchange rate risk, and external risk. Insurers’ activities are of high- risk nature, and the strength of risk management and its implementation thereof determines their growth and survival (Akotey & Abor, 2013). Insurance plays a key role in the economic growth and development of a country and helps protect organisations and society by mitigating risk (Liedtke, 2007). However, the insurance industry in Africa faces several challenges that are preventing its growth and development. PwC’s (2018) report narrates some of these challenges as underinsurance, high level of business concentration, and risk management issues. They further reveal that urbanization, innovations, technological development and strengthening of the risk management system present new opportunities to broaden the insurance market. According to a survey conducted by Deloitte (2019), the African insurance industry faces several challenges that make it lag behind. Their report revealed risk management as one of the key challenges. Deloitte (2019) further explained that Africa has the lowest insurance penetration rate in the world (2.7%), while sub-Saharan Africa (SSA) has the lowest literacy rate (59%) and the highest poverty level of about 42% (World Bank, 2018). In 2020, Africa’s population was about 1.395 billion, which is 16.72% of the world’s population. However, the penetration rate of insurance in Africa revolves around 2% on average, which is lower than the average emerging markets rate of 3.4% and the global rate of 7.3% (Swiss Re, 2020). This reveals that Africa is underinsured by showing low penetration and maturity of the insurance sector. Product development and innovation, resilient investment, and better risk management practices such as ERM will help broaden the scope of insurance penetration. Based on the foregoing, it can be deduced that the ability of insurance companies to build resilience against the complexities and uncertainties, increase penetration, raise investor confidence, and enhance their performance and risk-taking behaviour is anchored on a robust risk management 3 technique. Failure to empirically establish the relationship between ERM, corporate governance, performance and risk-taking may have dire implications. First, it will impair risk governance and lead to poor risk management practices, which may affect their performance. Second, it may affect regulatory and board decisions on the policies that need to be made to enhance insurance performance and risk-taking decisions. Third, the survival of insurance companies may be in doubt amid uncertainties and economic crises. Fourth, failure to understand the relationship between ERM and risk-taking may lead to a wild west within an entity. Without a strong leadership (CG) available to establish and monitor the risk appetite and tolerance levels, managers will go for broke, which might result in goal incongruence, a major issue revealed in the agency theory. One of the key players in the insurance industry in SSA is South Africa. The South African insurance industry is very competitive and has a well-developed insurance market compared to other African markets. It includes a variety of multi-channel distribution alternatives, including technology and digitalisation of insurance markets, which strong institutions and considerable legislative changes support. Nonetheless, there are issues of overregulation and the diversion of management attention from strategic growth opportunities (PwC, 2018). It has been recommended that significant regulatory changes and an effective risk management system will enrich the insurance market. Also, the Ghanaian insurance industry is seen as one of the fastest-growing industries in SSA. The industry has undergone several regulatory transformations, which presents growth opportunities. Some of these regulatory changes include the privatisation of state-owned insurance companies, separation of life from non-life companies, and the development of the risk- based supervision framework, which emphasises implementing a strong risk management system. On this note, this study concentrates on the insurance industry in Ghana and South Africa by examining the linkages between ERM, corporate governance, risk-taking and performance. 1.2 Overview of the Insurance Industry in Ghana and South Africa This study aligns with prior empirical studies by examining two countries within the SSA context: Ghana and South Africa. The study chose the insurance industry because the concept of risk emanated from insurers, and they are in the business of managing the risks of others. Also, insurers operate in a highly regulated and competitive environment in which elements of ERM are largely required. The motivation for these two countries is because they have experienced an improvement 4 in their corporate governance practices and market capitalisation in recent years, and several regulatory policies on risk management have been conducted. First, the Ghanaian insurance industry represents one of the growing insurance industries in SSA, with a penetration rate of about 1.1%. The insurance industry in Ghana has undergone several regulatory developments over the past years as a form of risk management. Some of these include the ‘No Premium No Cover’ policy, compulsory fire insurance for private commercial buildings, reinsurance premium transfers, and others. The choice of this country is further motivated by the fact that ERM has been adapted into insurers’ risk framework in recent years. The National Insurance Commission (NIC) in 2014 issued a new solvency framework and, as part of measures put in place, adopted the Risk-Based approach to Supervision (RBS) of Insurers. A key issue in the RBS is for firms to adopt robust risk management practices to make regulated entities more efficient. These developments were made due to the recent financial crises in the banking and insurance industry, which led to the collapse and merging of companies such as Star Assurance, Starlife and Microinsurance into Star Assurance Group Limited and Regency Alliance and NEM Insurance. The NIC developed a corporate governance and risk management framework to guide the operational activities of insurers and enhance the effective supervision of the insurance companies. The new solvency framework guideline uses a risk-sensitive approach to determine the solvency of insurance companies. The risk-based approach provides a proactive assessment of risk facing an organisation. It serves as a tool for identifying and managing risks and ensures the effective management of organisational resources. Regulated entities are to ensure that robust risk management measures are embedded into their operations. Second, South Africa has the biggest insurance industry in SSA, with a penetration rate of about 16.99%. The market is competitive compared to other insurance markets on the continent, as seen by many companies and diversified product offerings. In 2019, the insurance industry was made up of 74 life companies and 91 non-life companies. Attention to risk management and corporate governance has greatly increased following various financial scandals like Guptas’ Oakbay and KPMG, which occurred in 2016 due to compliance and audit failures (Holtzblatt et al., 2020). One issue in the insurance industry was the “premium misappropriation” scandal by Insure group managers and the ghoulish fraudulent and dishonest claims from the life insurance industry in 5 2018. In response, several amendments were made, including revising risk management procedures and internal control processes, corporate governance changes and compliance controls. The South African insurance sector (both life and non-life) is regulated by the Financial Sector Conduct Authority (FSCA). The Registrar of life and non-life insurance in South Africa has developed guidelines on corporate governance and risk management for insurers which took effect in the second decade of the 21st century. These guidelines largely encapsulate the provisions of the King III code, which takes key interest in corporate governance, risk management and compliance. The King III code presents significant opportunities to firms that embrace its principles. The introduction of the new solvency framework and corporate governance guidelines in Ghana and South Africa was to ensure that the regulation of insurers is in line with international best practices and that insurers are more efficient in their activities. As the Ghanaian and South African insurance industries have different characteristics compared to the rest of the world, the findings of this study advance knowledge of ERM within a new context. In light of the above premise, this study seeks to contribute to the body of knowledge by examining the insurance industry in Ghana and South Africa. Table 1.1 presents an overview of the insurance industry in Ghana and South Africa. Table 1. 1 Stylised facts about the life and non-life insurance market in Ghana and South Africa Ghana (GHS M) Industry 2014 2015 2016 2017 2018 2019 Balance sheet Total Assets Life 1,384 1,744 2,241 2,889 3,123 3,852 Non-Life 895 1,320 1,506 1,863 2,375 2,859 Total Liabilities Life 6.4 3.8 1.4 2,003 2,280 2,711 Non-Life 17.6 13.5 14.3 890 1,179 1,467 Income Statement Gross Premiums Life 581 706 859 1082 1337 1,651 Non-Life 659 854 1070 1347 1600 1,834 Investment Income Life 177 284 378 2479 2619 3,188 Non-Life 89 118 886 142 1417 1,684 Claims Total Claims Life 243 256 159 224 301 284 Non-Life 237 555 329 272 208 241 The Industry Life 19 24 22 20 26 24 6 Non-Life 26 27 29 28 28 29 South Africa (R M) Industry 2014 2015 2016 2017 2018 2019 Balance sheet Total Assets Life 2,414,726 2,567,147 2,652,277 2,914,534 2,993,436 3,054,223 Non-Life 116,352 121,875 131,831 152,498 149,882 201,453 Total Liabilities Life 2,273,451 2,416,090 2,498,443 2,757,373 2,630,845 2,786,406 Non-Life 72,202 76,742 82,350 86,394 85,097 99,0245 Income Statement Gross Premiums Life 447,442 468,595 485,610 475,935 134,362 504,567 Non-Life 98,962 109,901 116,722 128,557 29,407 46,772 Investment Income Life 283,834 198,593 164,206 16,343 -76,999 673,003 Non-Life 5,665 6,281 7,759 8,283 366 9,456 Claims Total Claims Life 14,998 16,073 16,521 415,388 106,426 306,576 Non-Life 43,679 44,581 46,840 49,067 12,083 37,034 The Industry Life 73 74 72 71 65 74 Non-Life 92 90 90 89 70 91 Source: Extracted from the annual reports of the NIC and FSCA 1.3 Problem Statement 1.3.1 ERM determinants Following the big financial scandals and 2008 global economic crises, there have been several regulations on how firms should operate. Ghana and South Africa are no exception. The collapse and merging of financial institutions (for example, the acquisition of Alexander Forbes by Momentum Metropolitan, merging of Star Assurance, Starlife and Microinsurance into Star Assurance Group Limited, and Regency Alliance and NEM Insurance), corporate crises, and organisational scandals (example, Guptas’ Oakbay) in these countries have introduced some regulations on risk management and corporate governance. Adopting a robust and dynamic approach to risk management has become critical to every business; hence, businesses are beginning to adopt an ERM system within their risk ecosystem. Despite the regulatory policies, there are firm-specific factors that drive their ERM adoption. To establish the influence of ERM on performance, one must first understand the factors that led to their ERM adoption (Bromiley et al., 2015). This is in line with the contingency theory, which says that the successful implementation of a system within an organisation is dependent on certain firm-specific factors (Woods, 2009). These are the causative factors that affect the behaviour of risk management. These factors can change over time and across events. Knowing these drivers will enhance management 7 decisions on how to manage them. This ensures the proper management and growth of these factors. This study focuses on the insurance market in South Africa and Ghana, where ERM is still gaining prominence. Besides the growing regulatory demand on the need for an integrated risk management system, this study aims to examine firm-specific factors that might have led to ERM adoption. The question about what factors drive ERM adoption is scarce in the literature; hence there is not enough evidence as to why organisations adopt an ERM system. Identifying the determinants of ERM is very important so that in future crises, these factors can be properly managed to sustain organisations’ viability. In recent years, maintaining organisational performance has been a major problem for most insurers in South Africa and Ghana. As indicated earlier, most insurers are experiencing underwriting losses while others are unable to sustain their business, resulting in the collapse and merging of some insurance companies. Therefore, it is important to identify some of these critical areas to ensure that the implementation of ERM is successful in these areas (Zhao et al., 2013). Moreover, previous studies have revealed mixed results on the various factors that affect ERM adoption. This may be because of differences in environmental characteristics (Lechner & Gatzert, 2018) and regulatory issues, as some organisations are more exposed to risks than others. So, empirical findings are not expected to be homogenous. For instance, Baxter et al. (2013) explained that insurers and banks are more interested in ERM implementation because they face a higher risk than others. In addition, empirical studies conducted in this area have focused on Europe (Bohnert et al., 2019; Lechner & Gatzert, 2018); Malaysia (Golshan & Rashid, 2012; Razali & Tahir, 2011); and the US (Hoyt & Liebenberg 2008, 2011; Beasley et al., 2005; Pagach & Warr, 2011) without any empirical evidence from developing countries such as Ghana and South Africa. According to Lechner and Gatzert (2018), generalisations of empirical findings may impose a lot of limitations and challenges due to differences in dataset and environmental characteristics. Therefore, results that are valid for one country may not necessarily be applied in another context. There is, therefore, the need to further investigate the determinants of ERM in an emerging market context. Based on these arguments, this study seeks to examine the factors influencing the adoption of ERM among insurers in Ghana and South Africa. Another potential factor that has been ignored in the literature 8 is the level at which the determining factors influence ERM adoption. While previous studies considered a linear regression approach to examine the determinants of ERM adoption, the impact of the determining factors on ERM adoption at different levels is still lacking in the literature. This paper contributes to the literature by addressing such issues using the quantile regression technique. 1.3.2 ERM, Corporate Governance and Performance Despite the various literature on ERM, empirical studies on the ERM-performance relationship remain limited and diverse (Florio & Leoni, 2017; Farrell & Gallagher, 2014). In addition, most of these studies are centred on the developed economies (Florio & Leoni, 2017; Lechner & Gatzert, 2018), with few (Silva et al., 2019; Horvey & Ankamah, 2020) empirical evidence from emerging markets. Hence the understanding of ERM among emerging economies is unclear and lacks empirical support. Silva et al. (2019) assert that organisational characteristics in developing countries are different from developed economies. He explained that emerging economies face several challenges and uncertainties that demand a strong risk management system. A report by PwC (2018) reveals that the insurance industry in Africa faces issues such as low penetration, underinsurance, high level of literacy and technological issues. They disclosed that technological innovations, good corporate governance and a robust risk management system present new potential to extend the insurance business. Therefore, this study contributes to knowledge by investigating the impact of ERM within the SSA context, particularly on insurance companies in Ghana and South Africa. In addition, there have been inconclusive results in the literature concerning the relationship between ERM and performance. According to the literature, some studies reveal that ERM improves performance (Hoyt & Liebenberg, 2008, 2011; McShane et al., 2011) and enhances the value of the firm (Grace et al., 2015; Eckles et al., 2014). However, Lin et al. (2012) and Li et al. (2014) argue that a negative relationship exists between ERM and performance, while Beasley, Pagach and Warr (2008) and Pagach and Warr (2010) found no evidence of the ERM-performance relationship. The inconclusive results may be because of certain moderating factors affecting the relationship. However, little academic research has been done on the possible moderating variables between ERM and performance. While there have been studies on how intellectual capital 9 dimensions (Saeidi et al., 2020), the role of the risk committee (Malik et al., 2020), financial literacy (Yang et al., 2018), and ERM maturity (Farrell & Gallagher, 2019) moderate the ERM- performance relationship, we are still not aware of how corporate governance moderate this relationship. Literature has shown that ERM is more effective when integrated with other factors such as corporate governance (Gordon et al., 2009). This supports the contingency theory, which states that the relationship between two factors is mostly influenced by another factor; hence the impact of ERM on performance is subject to a number of influences (Woods, 2009). The 2017 COSO ERM framework is greatly concerned with the alignment of risk management with corporate governance. This is because ERM is a framework that is incorporated into the corporate governance system of a firm. Caldwell (2012) affirms that one of the major factors leading to a strong ERM system is the existence of proper corporate governance initiatives, of which board oversight is an essential attribute. However, the linkages among these variables have not been empirically tested, as studies have mostly focused on the direct relationship between ERM and performance. Regardless of the growing volume of academic literature on ERM, the relationship between ERM and firm performance will not be complete without examining its linkages with corporate governance. In line with these arguments, this study contributes to empirical knowledge by addressing the combined effect of ERM and corporate governance on performance. 1.3.3 ERM, Corporate governance and firm risk-taking Risk-taking is very critical to the success of every organisation. Although organisations seek to minimise their risks, insurers cannot succeed if they do not take risks that are ex-ante profitable. Over the past two decades, the extant ERM literature has focused on the value of ERM, the driving force for ERM, and its implementation within an organisation. Despite the growing literature on ERM, empirical evidence regarding the impact of ERM on insurers’ risk-taking decisions is inchoate. Therefore, there is a need to deepen the understanding of how ERM affects insurers’ risk-taking behaviour because the essence of implementing ERM is to ensure that firms are proactive and have adequate resources to manage and control their risk. However, it remains in doubt whether ERM influences firms to take more, less or no change in their risk. Even though ERM supports firms in managing their activities within an acceptable risk appetite level, there are 10 certain situations where risk management fails to ensure that insurers take the right amount of risk. However, this is yet to be empirically examined. Furthermore, Adams and Jiang (2016) highlighted that insurers were affected by the 2008 global financial crisis due to poor corporate governance systems and risk-taking decisions. Elamer et al. (2018) state that enhancing corporate governance among insurers will affect their risk decisions as a good corporate governance system is connected to good risk-taking of insurance companies. There have been studies on corporate governance and risk-taking (Elamer et al., 2018; Matthew et al., 2016; Su & Lee, 2013; Laeven & Levine, 2009; John et al., 2008). However, empirical evidence on how ERM and corporate governance affect insurers’ risk-taking, especially in the Ghanaian and South African contexts, has been less explored. Given the significance of corporate governance and ERM, many assume that this may reduce firm risk-taking. Nevertheless, the complexity of insurance firms, attitudes of risk managers, and shareholders’ interest might influence firms’ risk-taking behaviour. In addition, empirical studies have failed to examine the threshold dynamics of these relationships. It is argued that when ERM and corporate governance factors fall below or above a certain threshold level, their impact changes (Aebi et al., 2012; Florio & Leoni, 2017; Horvey & Ankamah, 2020). However, the precise impact of these factors below or above the threshold level remains unclear in the literature. This study provides a detailed understanding of how these factors should be developed to enhance insurers’ risk-taking from the policy perspective. In light of the above premise, this study investigates the linearity and nonlinearity between ERM, corporate governance and insurers’ risk-taking. 1.4 Research Objectives The central objective of this thesis is to examine the nexus between ERM, corporate governance, performance and risk-taking among insurers in Ghana and South Africa. Specifically, the study seeks to: i. Examine the determinants of ERM adoption among insurers in Ghana and South Africa. ii. Investigate the impact of ERM on insurers’ performance. This objective also unearths the moderating effect of corporate governance on the ERM-performance relationship. iii. Examine the linearity and nonlinearity between ERM, corporate governance and insurers’ risk-taking. 11 1.5 Research Questions The above objectives yield to the following research questions: i. What are the determinants of ERM adoption among insurers in Ghana and South Africa? ii. What is the impact of ERM on insurers’ performance? Does corporate governance moderate the ERM-performance relationship? iii. Does a linear and non-linear relationship exist between ERM, corporate governance, and insurers’ risk-taking behaviour in Ghana and South Africa? 1.6 Justification ERM has received great attention over the past two decades. Many businesses have started implementing ERM because it enhances the value and performance of an organisation. Over the years, studies on ERM have been centred on its implementation, value, and determinants. While some studies adopt a qualitative and a survey approach, others adopt a quantitative approach to investigating the value and influencing factors of ERM adoption. Despite the studies that have been conducted, empirical literature remains limited, especially within the African context. Also, empirical studies face measurement problems, with many relying on a binary approach to ERM measurement. This study introduces a comprehensive measurement for ERM, which seeks to capture its governance, oversight and operating mechanisms. The study follows the work by Florio and Leoni (2017) on ERM measurement while considering other variables which have not been factored in their model to make it more robust. The first contribution of this study is to provide a review of the various measurements adopted in the literature to measure ERM and how it influences firm performance. It explores the empirical literature and introduces a missing link that serves as recommendations for further studies. The study argues that getting a comprehensive measurement for ERM will significantly contribute to the literature and be good for policymaking. In addition, the ERM-performance relationship could further be investigated by considering indirect factors that are likely to affect this relationship. According to the research objectives, the second contribution is to examine the factors influencing ERM adoption in the insurance industry. The growing complexities of organisational risks have influenced firms to adopt a holistic approach to risk management. However, there are firm factors 12 that influence their ERM adoption; hence, this study employs both the linear and quantile regression approach to examine this phenomenon. The empirical literature on the determinants of ERM adoption remains inconclusive. Again, little is known about why insurers adopt ERM in Ghana and South Africa, which has not been done so far. Thus, an analysis of the drivers of ERM adoption is expected to advance our understanding of the under-investigated insurance market in Ghana and South Africa. This may be of support to industries and organisations that are addressing the new risk management requirements. Another justification for this study is that the linear regression is complemented by the quantile regression approach, which provides a more comprehensive picture of the determining factors at different levels. The quantile regression approach allows for variations in the coefficients for the entire distribution (Bassett & Koenker, 1978). Researchers have conducted studies on ERM from diverse perspectives with emphasis mostly on Europe, Malaysia, and the US. However, empirical evidence from the context of developing economies remains limited. As Lechner and Gatzert (2018) stated, generalisations of empirical findings across industries and geographical locations may impose a lot of limitations due to the differences in characteristics. Following the number of regulatory interventions stressing the relevance of risk management in corporate organisations, studies on ERM in Ghana and South Africa could still be undertaken to make justifiable contributions to the empirical literature. In doing so, this study contributes to the diverse argument on ERM and performance relationship in a new context. More so, the complex relationship between ERM and performance may be moderated by some other factors. It is argued in the literature that there is an indirect linkage between ERM and performance through corporate governance (Sekerci & Pagach, 2019; Gordon et al., 2009); however, empirical studies remain silent about this. Corporate governance plays a key role in ERM implementation and its relevance. The board of directors sets the tone and provides resources for an efficient ERM system (Lundqvist, 2015). Sekerci and Pagach (2019) support this argument by stating that an improved CG enhances ERM adoption because CG helps in risk management monitoring. The second objective contributes to knowledge by considering an indirect approach to the ERM-performance nexus by making a case for corporate governance as a moderating factor. This study highlights which corporate governance variable contributes to the ERM process. This supports the theoretical proposition made by COSO (2017) about the linkages 13 between ERM and corporate governance. The findings of this study are important for policy implementation. It supports organisations that need further directives on ERM implementation and firms that seek to adopt a holistic approach to implement risk management. Finally, just as literature on ERM and performance keeps growing, empirical studies on how ERM and corporate governance affects risk-taking decisions remain limited. While literature states that firms with high risk adopt a risk management system (Hoyt & Liebenberg, 2008, 2011; Pagach & Warr, 2011), no study, to the best of the researcher’s knowledge, seems to look at how the adoption of ERM affects firms risk-taking decisions. The success of insurance firms depends on how they take risks. Insurers’ ability to manage risks leads to an increase in shareholders’ value. For risk- taking to increase shareholders’ value and profitability, there must be an improved ERM system and good corporate governance structures (Sekerci & Pagach, 2019). The relationship between ERM, CG and firm risk-taking has many implications. This is an important gap in the literature that will provide insight into how ERM influences risk-taking, which will also provide insight into the necessary policy decisions that need to be made. To deal with the endogeneity problem, the study employs the generalised method of moments (GMM), which considers the lag of the dependent variable to examine how previous risk-taking decisions affect current decisions. The findings of this study, without any doubt, are very important for policymaking. Again, the study examines the nonlinearity between ERM, CG and risk-taking, which has not received much attention. This study reveals the effect of these factors under two different regimes, thus below and above the threshold level. With this, the study proffers concrete policy recommendations for managing these relationships. 1.7 Thesis Structure The thesis consists of six chapters. Chapter Two: This chapter provides a comprehensive literature review on the measurements and impacts of ERM. Google Scholar was the main search engine used in collecting empirical works between 2001 and 2020. Articles listed in the Scimago journal ranking were chosen for the review. SCImago is a multidisciplinary database that provides index ratings for all journals based on Scopus/Elsevier’s extracted data. This chapter provides fresh insight into the gaps in the literature and provides recommendations for future studies. 14 Chapter Three: This chapter foretells the various determinants of ERM adoption among insurers in Ghana and South Africa. The logistic panel and quantile regression techniques are used to examine this relationship. This is due to the binary nature of the dependent variable (ERM adoption). A sample of 96 insurance companies from both countries was examined between 2015- 2018. It addresses the similarities and differences of the determinants between these two countries. Sensitivity and univariate analysis across the ERM status first compare the univariate differences between ERM firms (ERM=1) and non-ERM firms (ERM=0). Chapter Four: This chapter first examines the impact of ERM and Corporate governance on the performance of the insurance industry. An ERM index was developed based on twelve main features categorised under risk governance, operating mechanism and the quality of risk oversight. A static model was chosen to examine the impact of the individual categories on performance. The overall ERM index was also examined to determine its relationship and significance. Also, the interactions between ERM and the corporate governance variables were assessed in relation to performance. The study followed the conditions specified in Brambor et al. (2006) to test for the joint significance of the constitutive and interaction terms to arrive at the net effect of the ERM and corporate governance factors on insurance performance. Chapter Five: This chapter focuses on the linearity and nonlinearity between ERM, corporate governance and insurers’ risk-taking behaviour. The study employed several techniques to assess these relationships. The generalised method of moments (GMM) proposed by Arellano and Bond (1991) was adopted to investigate the linear relationship. This technique was adopted due to the significance of the lagged dependent variable (Risk-Taking). Incorporating the lagged term may cause endogeneity problems that the GMM controls. There is a suspicion of a possible endogeneity problem between ERM and risk-taking. While ERM may influence risk-taking behaviour, insurers’ risk-taking decisions may also influence the implementation of ERM. The study further employs the dynamic panel threshold model developed by Seo and Shin (2016) and Seo et al. (2019) to examine the non-linear relationship. This aspect sought to determine if nonlinearity exists in the model and to know the effect of the threshold variables below and above the threshold indicator. 15 CHAPTER TWO THE MEASUREMENTS AND PERFORMANCE OF ENTERPRISE RISK MANAGEMENT: A COMPREHENSIVE LITERATURE REVIEW 2.1 Introduction Enterprise risk management (ERM) has gained popularity among business practitioners and academics over the past two decades. Uncertainties, high volatilities, and complexities in the business environments have exposed the traditional risk management (silo) system’s limitations. Global economic crises, coupled with pandemics, have triggered various risk factors affecting every aspect of an organisation. These have raised awareness, hence calling for the need for a comprehensive and robust risk management system for businesses since most of these risks cannot be treated in isolation (Hoyt & Liebenberg, 2011). ERM provides a shield to organisations by anticipating and providing resources to manage any unforeseen events. Before the development of ERM, businesses were managing risks in silos. This approach of concentrating on specific risks was not helpful due to issues associated with risk oversight, duplication of resources and its focus on internal risks. Enterprise risk management is the strategic process of planning, identifying, assessing, managing, monitoring, and communicating organisational risks in an integrated manner. ERM provides a top-down holistic strategic view of risks (Horvey & Ankamah, 2020). The Committee of Sponsoring Organization of the Treadway Commission (COSO) defines ERM as: ‘A process, effected by an entity’s board of directors, management and other personnel, applied in a strategy setting and across the enterprise, designed to identify potential events that may affect the entity, and manage risk to be within its risk appetite, to provide reasonable assurance regarding the achievement of entity objectives’ (COSO, 2004). Industries, professionals, regulators, rating agencies, international standard organisations and business consultants have been encouraged to implement ERM as its implementation is believed to enhance shareholder value and profitability (Lechner & Gatzert, 2018). Firms need to take advantage of this opportunity to overcome any financial uncertainties. Despite the relevance of ERM, there have been several arguments in the literature about the benefits and pitfalls of ERM. Theoretically, the implementation of a risk management system such as ERM could enhance firm 16 value. According to the modern portfolio theory, firms operate in an imperfect market; hence, risk management’s value creation attribute can be enhanced when firms exploit the advantages in an imperfect market (Markowitz & Todd, 2000). Markowitz (1952) posits that shareholder’s value is not enhanced by risk management because shareholders can, without much stress, spread their risk; hence, only the systematic risk is relevant. In view of that, all risk management practices have a negative net present value (NPV) and must, therefore, not be implemented. However, Beasley et al. (2008) explain that the hostile NPV project resulting from the firm’s reduction of idiosyncratic risk is based on the notion that capital markets work devoid of resistances and deficiencies. When such resistances and deficiencies are introduced, it reveals the value creation activity of ERM. Meulbroek (2002) adds that a firm should not focus on an individual risk as this may create other types of risks that may affect the firm. Hence, when ERM is adopted, it enables the firm to manage all kinds of risks in an integrated manner. Thus, ERM supports the firm in identifying the full potential of risks and ways of eliminating unfavourable risks and hedging contracts that affect the natural hedging of risks within the portfolio (Eckles et al., 2014). While some scholars argue that ERM is a problem-solving activity, others see it as part of the problem. According to Power (2009), ERM’s security is illusory, which leads to the risk management of nothing. However, Farrell and Gallagher (2019) argue that ERM serves as a facilitator, which creates a competitive advantage for firms. The foredeal and afterdeal of ERM have been extensively discussed in the literature. One of the major obstacles in the ERM literature is non-disclosure, as many firms fail to disclose their ERM activity (Lundqvist, 2014). As a result, different methodological approaches have been employed to measure ERM. This phenomenon has created many inconsistencies in the result of ERM. Against this backdrop, this paper provides a comprehensive and systematic review of ERM measurement in the existing literature. It identifies the procs and cons of each measurement and makes suggestions for further studies. Again, the paper examines the value implications of ERM implementation. This is a response to the increased attention given to ERM in the 21st century. Google Scholar was the main search engine used in selecting the ERM literature. The search spanned a period of two decades, that is, from 2001 to 2020. The final criteria used in selecting ERM articles was that the paper must be indexed in the SCImago Journal Ranking (SJR) indicator. SCImago reveals the importance of scholarly journals 17 and citations received, which is explained by the H index. A total of twenty-nine (29) articles were used for the discussion on ERM The findings from the systematic literature review reveal different approaches to ERM measurements. Scholars either use a survey approach or rely on secondary sources for their ERM information and measurements. According to the literature, most scholars rely on the appointment of the Chief Risk Officer (CRO) as a simple proxy for ERM adoption. This approach is widely used because of the difficulty in identifying ERM information. In the absence of the CRO, other scholars adopt the Standard and Poor’s (S&P) risk management rating, COSO (2004) objectives, RIMS, and the survey approach. A comprehensive ERM measurement approach, which considers risk governance issues and operating principles, was also considered in some studies. These inconsistencies affect the value implications of ERM as scholars are comparing “apples to oranges.” Regarding performance, our findings show that the ERM literature cuts across different industries and geographical areas. Also, different estimation techniques, time periods, and performance measures are used to examine this relationship. Quantitative approaches and secondary sources of information were mostly adopted. Due to these diversities, the literature on ERM and performance reveals mixed findings; however, the study finds enough evidence that ERM enhances firm performance and shareholder value. Besides, a higher level of ERM maturity is found to significantly impact firm performance. This chapter contributes to the ERM literature in several ways. First, it provides a comprehensive and up-to-date discussion on the various measures and value of ERM. It identifies the strengths and weaknesses of each approach in measuring ERM and provides suggestions for future research. Second, it serves as a reference point for researchers who are interested in having a detailed understanding of the ERM literature. Third, the paper provides trend analysis, geographical contexts, and different estimation techniques used to examine the relationship. This serves as an initial point for identifying research gaps for further studies. 18 The rest of the chapter is structured as follows: The second section discusses the methodological approach used in selecting empirical literature for the study. Section three presents empirical findings on the measurement of ERM. Section four presents empirical findings and discussions on the relationship between ERM and firm performance. The final section addresses the conclusion, limitations, and recommendations for further studies. 2.2 Methodology This chapter presents a systematic and comprehensive review of ERM literature on the measurement and performance of ERM. Only published papers in peer-reviewed journals were selected to examine the impact and measurement of ERM. An advanced literature search was conducted on Google Scholar for peer-reviewed articles. A keyword search on the terms “ERM and firm performance” and “Value of ERM” were used to select empirical studies. The search period for the literature was two decades. That is, from 2001 to 2020 because empirical studies on Enterprise Risk Management were first seen in literature in 2001 (Dickinson, 2001). The first search was done on 17th March 2020, and the last on 20th December 2020. A search within the database produced over 2000 articles. Each page revealed ten articles. To ensure that more recent studies were included in the study, an advanced search was conducted to include articles published from 2015 to 2020. This was done to ensure that the study reveals current issues on the measurement and the ERM-performance relationship. Articles selected were based on the titles, abstracts and full text. The final sample of papers was twenty-nine (29). 2.3 Journal Credibility The final criteria used in selecting peer-reviewed articles was that the paper must be listed in the SCImago Journal Ranking (SJR) indicator. SCImago is a multidisciplinary database that provides index ratings for all journals based on Scopus/Elsevier’s extracted data. This is to ensure that relevant articles in the subject area were included in the review. SCImago reveals the importance of scholarly journals and citations received by that journal explained by the H index. SCImago corrects many of the factors criticised in the journal impact factors (IF) because of the inclusion of more journals, longer periods of citations (3 years), and limits self-citations. “It also weighs citations according to the importance of the journal where they were published, using an algorithm similar to that of Google PageRank” (Falagas et al., 2008, pp. 2623). Table 3 in the appendix 19 provides the details of the journals and their SCImago H index. The journal H index of the papers ranged from 3 to179. Figure 2.1 describes the subject area of the journals. 2.4 Subject Areas The journals were grouped based on their subject areas in SCImago. It reveals the importance of scientific influence and the prestige of the journals. The subject areas are presented in Figure 2.1. The subject area with the highest journal publication was Business Management and Accounting, representing about 65.52% of the total papers used for the review, followed by Economics, econometrics, and finance, which had 20.69%. Decision Science, Psychology, Engineering, and Arts and humanities were the subject areas with the lowest journal publication, which had a score of 3.45%. Figure 2. 1: Subject Areas for SCImago 2.5 Literature review on ERM measurement The empirical literature on ERM measurement is multifaceted. This is as a result of the different approaches used in collecting ERM information. While some researchers preferred public data, others identified the presence of ERM through surveys. The section explains the various approach used in measuring ERM. 65.52 3.45 3.45 3.45 20.69 3.45 Business, Management, and Accounting Psychology Decision Sciences Engineering Economics, Econometrics and Finance Arts and Humanities Subject Area 20 2.5.1 CRO/ERM Keywords: Liebenberg and Hoyt (2003) were the first to adopt the CRO as a simple proxy to measure ERM. Many scholars followed the same approach by relying on the CRO or ERM keyword search on the company’s profile for their ERM measurement. That is, CRO equals one if the firm has a CRO, otherwise zero (Beasley et al., 2005; Beasley et al., 2008; Hoyt & Liebenberg, 2011; Pagach & Warr, 2011; Golshan & Rashid, 2012; Lin et al., 2012; Eckles et al., 2014; Farrell & Gallagher, 2015; Lechner & Gatzert, 2018; Silva et al., 2019; Nguyen & Vo, 2020). Most studies rely on this simple proxy because of ERM measurement complexities (Pagach & Warr, 2011), and lack of disclosure of ERM requirements (Lundqvist, 2014). More so, this proxy is used because firms that have adopted an ERM system must have a risk expert or CRO (Pagach & Warr, 2011). It is also argued that a CRO’s employment is a signal that firms have adopted an integrated approach to managing their risk (Beasley et al., 2005; Bohnert et al., 2019). However, this approach is not without limitations. The challenge with this approach is that a single variable may not be a good predictor of ERM adoption because the hiring of an individual may not necessarily reflect a well-implemented and an effective ERM system (Lundqvist, 2014), especially, in situations where CRO may not correspond to an ERM system (Sekerci & Pagach, 2019). Again, a simple proxy makes it challenging to identify the extent of an ERM implementation (Gatzert & Martin, 2015). It is also possible that firms may create this position as ‘window dressing’ to comply with regulatory or corporate governance requirement; hence, most firms create this position to please regulators and the board. This implies that studies that focus only on the CRO may fail to provide a good impression of ERM implementation and performance. It is also possible that firms may employ a different expert, such as the internal auditor or the chief financial officer to perform the CRO’s responsibilities. 2.5.2 RIMS Risk Maturity Model: Instead of relying on the CRO, other scholars relied on a more sophisticated approach that gives a broader ERM measurement perspective. For instance, Farrell and Gallagher (2015, 2019) construct an ERM index using the seven attributes of the Risk and Insurance Management Society (RIMS) Risk Maturity Model. The performance of each attribute was measured based on specific competency drivers. These are ERM adoption-based approach (the level of support for ERM by 21 the board), ERM process management (the integration of ERM process into daily practices), risk appetite management (the risk portfolio view and risk-reward tradeoffs), root cause discipline (ability to identify specific trends to minimise adverse effects and maximise value creation), uncovering risk (ability to identify threats and opportunities for effective mitigation and exploitation), performance management (the ability to align strategy with risk management to execute business plans), and business resilience and sustainability (the ability to recover from business setbacks and maintain value). Each of these attributes was classified into five maturity levels which are stated as, 1. Ad hoc: ERM depends on individuals with insufficient understanding. 2. Initial: where risk management is done in silos. 3. Repeatable: the approach to risk management is established with the help of risk framework. 4. Managed: ERM fully integrated into the management process), and 5. Leadership: risk-based discussions are embedded in the strategic process with clear understanding. Notwithstanding, Farrell and Gallagher (2015) relied on a dummy approach for their ERM measurement. ERM equals one if the firm’s RIMS Maturity model is Ad hoc or Initial and zero otherwise if its RIMS Maturity model is either Repeatable, managed or Leadership. As compared to the S&P risk management rating, the RIMS index provides a comprehensive assessment of an organisation’s risk attributes, including risk appetite and performance management. However, the RIMS index is based on survey data which is exposed to response bias. 2.5.3 S&P Rating: Another approach to measuring ERM is the Standard and Poor (S&P) risk management rating. McShane et al. (2009) first adopted this approach to construct an ERM index with the S&P rating. The S&P rating is a risk management tool developed to assist in rating the insurance and banking companies (Linke & Florio, 2019). The results from their ERM rating were classified as: very strong, strong, adequate, and weak. Other studies followed a similar approach to investigate the dynamic capability and value of ERM (Quon et al., 2012; Baxter et al., 2013; Nair et al., 2014; Bohnert et al., 2019). The challenge with this approach is that the S&P rating is limited and not applicable to every firm but can only be applied by some selected firms (Insurance and banks) hence, there is the need for a proxy that can be broadly applied. Florio and Leoni (2017) note that the S&P rating is highly dependent on their definition of ERM and must be further evaluated for 22 its appropriateness for measuring ERM and content validity. Bailey (2019) further adds that the S&P rating does not include risk experts’ expertise and ignores ERM quality variations. Farrell and Gallagher (2015) stress that the S&P rating is based on a limited number of subfactors rated as p