Governance and Performance in Emerging Markets Empirical Study on the Link Between Performance and Corporate Governance of IFC Investment Clients ©Copyright 2018. All rights reserved. International Finance Corporation 2121 Pennsylvania Avenue, N.W. Washington, D.C. 20433 The conclusions and judgments contained in this report should not be attributed to, and do not necessarily represent the views of, IFC or its Board of Directors or the World Bank or its Executive Directors, or the countries they represent. IFC and the World Bank do not guarantee the accuracy of the data in this publication and accept no responsibility for any consequences of their use. The material in this work is protected by copyright. Copying and/or transmitting portions or all of this work may be a violation of applicable law. The International Finance Corporation encourages dissemination of its work and hereby grants permission to users of this work to copy portions for their personal, non-commercial use, without any right to resell, redistribute, or create derivative works there from. Any other copying or use of this work requires the express written permission of the International Finance Corporation. ii Governance and Performance in Emerging Markets Empirical Study on the Link Between Performance and Corporate Governance of IFC Investment Clients iii Table of Contents Acknowledgmentsix Foreward xi Key Terms and Acronyms xii Executive Summary xiv 1. Introduction 1 2. Methodology 3 2.1. Objectives 3 2.2. Firm-Level Performance Variables 3 2.3. Corporate Governance Score 5 2.4. Company Sample and CG Data Collection 6 2.5. Study Limitations and the Causality Issue 8 3. Findings 11 3.1. CG Score Analysis 11 3.2. Subgroup Comparisons 12 3.2.1. First Objective: Effect of CG at Disbursement on Firm Performance 13 3.2.2. Second Objective: Improvements of CG Practices and Performance 13 3.2.3. Third Objective: Average Level of CG Practices and Performance 15 3.2.4. Core CG Index and Performance 16 3.2.5. Corporate Governance and E&S Review Rating 17 3.2.6. Splitting the Sample by Key Project Characteristics 20 3.3. Prioritizing Specific CG Indicators 23 4. Conclusions and Improvement Opportunities 25 References27 v Appendixes Appendix A. Descriptive Analysis of CG Data 30 Appendix B. Multiple Regressions 41 Appendix C. Full Corporate Governance Index 50 Appendix D. Glossary 56 Appendix E. Operational Definition of All Variables 62 List of Boxes Box 1.1: Continuing Debate on the Real Impact of CG Practices on Performance 1 Box 2.1: Advantage of a Broad CG Index 8 Box 2.2: Sampling Bias 10 List of Figures Figure 2.1: Indicative CG Index Output 7 Figure 3.1: Credit Risk Ratings in CG Practices at Disbursement 15 Figure 3.2: Financial and Economic Improvement of CG Practices 16 Figure 3.3: ROE/ROIC and EROE/EROIC and the Core CG Index 19 Figure 3.4: ESRR and Average CG Scores 20 Figure A.1: Overall Level of Following Recommended CG Practices at Disbursement 30 Figure A.2: CG Practices: FIG versus Non-FIG at Disbursement 31 Figure A.3: CG Practices: by Investment Product at Disbursement 31 Figure A.4: CG Practices: Listed versus Unlisted at Disbursement 32 Figure A.5: CG Practice: by Investment Tier at Disbursement 32 Figure A.6: Overall Level CG Practices (June 2016) 33 Figure A.7: CG Practices: by Industry (June 2016) 33 Figure A.8: CG Practices: FIG versus Non-FIG (end 2016) 34 Figure A.9: CG Practices: by Region (end 2016) 34 Figure A.10: The 10 Most Common CG Practices (June 2016) 35 Figure A.11: The 10 Least Observed CG Practices (June 2016) 36 Figure A.12: Variation in CG Scores (2011 to 2016) 37 Figure A.13: Variation of CG Scores by Industry (2011 to 2016) 37 Figure A.14: Variation of CG Scores, FIG versus Non-FIG (2011 to 2016) 38 Figure A.15: Variation of CG Practices by Region (2011 to 2016) 38 Figure A.16: CG Practices, by Investment Product (2011 to 2016) 39 vi Governance and Performance in Emerging-Markets Firms Figure A.17: Variation of CG Practices, Listed versus Unlisted companies (2011 to 2016) 39 Figure A.18: Variation of CG Practices by Investment Tier (2011 to 2016) 40 Figure B.1: Hypothetical Effect of Firm Size as a Moderating Variable 42 Figure B.2: Conceptual Model 43 List of Tables Table 2.1: An Indicative Output of the CG Index 7 Table 2.2: Industry Breakdown of the Sample 9 Table 2.3: Regional Breakdown of the Sample 9 Table 3.1: Correlations Between Selected Variables 13 Table 3.2: CG Practices at the Disbursement Date 14 Table 3.3: Mean Comparison of CG Improvement from 2011 to 2016 17 Table 3.4: Mean Comparison of Level of Average CG Practices 18 Table 3.5: Core CG Index Average Score and Performance 19 Table 3.6: Split-Sample Key Characteristics for Equity Deals 21 Table 3.7: Split-Sample Key Characteristics for Debt Deals 22 Table 3.8: Top 20 CG Practices—Strongest Predictors of Performance Variables 24 Table B.1: Full CG Index and Performance Variables 45 Table B.2: Core CG Index and Performance Variables 46 Table B.3: CG Index and Performance Variables: Equity Holdings 48 All tables and figures were created by the assessment team. Table of Contents vii Acknowledgments This report was written under the supervision on an IFC team those experts and peer reviewers, who come from three led by David Karapetyan, Regional Corporate Governance broad areas of endeavor (in alphabetical order within each Lead, Latin America and the Caribbean, and supported by group). From academia, including prominent corporate Nazeem Wasti and Alexey Volynets, Corporate Governance governance researchers: Guillermo Boitano, Maria Andrea Group. The principal author of the report was Alexandre Trujillo Davila, Patricia Gabaldon, Christopher Jones, Aurelio Di Miceli da Silveira, founder of Direzione Management Gurrea Martínez, Rosendo Ramirez, Francisco Reyes, Consulting and professor of corporate governance at Eduardo Schiehll, Marc VanEssen, Alexander Guzmán Alvares Penteado School of Business, Sao Paolo, Brazil, Vasquez, and Alessandro Zattoni. From the private sector and a short-term consultant with IFC. experts, including Institutes of Directors, investment funds, stock exchanges, consulting companies, and members We thank Raphael Bruce for research assistance on of IFC’s Private Sector Advisory Group: Adriane Almeida, the econometric tests; Kalyani Santoshkumar, Anatoliy Paulina Beato, Heloisa Bedicks, Bistra Boeva, Valéria Café, Yefymenko, Yovel David Crispin Davila, Sheela Rahman, Tobias Coutinho, Richard Frederick, Ola Peter Krohn Jacqueline A. Williams, Tsion A. Beyene, Florence Hope-Wudu, and Omosalewa T. Fajobi for corporate governance data Gjessing, Mike Lubrano, Luiz Fernando da Costa Dalla collection through surveys and assistance with interviews Martha, Andres Oneto, Miguel Puga, Francis Stenning, John D. Sullivan, and Marco Antonio Zaldívar. From the of companies participating in the study; Katherine Downs, World Bank and IFC, representing expertise in economics, Carlos Mayorga, Haruko Koide, John Mantzavinatos, Stephen corporate governance, investment operations, portfolio Chow, and colleagues in Financial Institutions Group management, strategy development, environmental and (FIG), Manufacturing, Agriculture, and Services (MAS), social risk management, and results measurement: Henning Infrastructure and Natural Resources (INR), and Telecom, Amelung, Deniz Baharoglu, Peter Cashion, Stephen Chow, Media, Technology, and Venture Investing (CTT) in the Katherine Downs, Flavio B. Guimaraes, Natalia Lavrova, regions, who introduced corporate governance to clients Chaoying Liu, John Mantzavinatos, Edmond Mjekiqi, Temel and helped review corporate governance (CG) questionnaire Oktem, Oliver Orton, Nilesh Shrivastava, Elizabeth M. White, responses; and Saurabh Garg, Anna Ostrovskaya, Carlos and L. Colin Xu. Special thanks also go to Alexander S. Berg, Mayorga, Lei Liu, Joseph Matthews for providing financial Neil Gregory, Merima Zupcevic, Roman Zyla, Khawar Ansari, and other company-level data from IFC management Charles Canfield, James Christopher Razook, Vladislava information systems, and Shannon Roe for the editing. Ryabota, Ghita Alderman, and Darrin Hartzler. In preparing this paper, we asked many professionals, with The CG Group is grateful to the many investment and diverse experience and backgrounds, to peer review the portfolio colleagues who helped with the design, data research methodology as well as the interpretation and collection, and validation for this study. We also want presentation of the results. Their thoughtful comments were invaluable in strengthening the content and overall to thank Shannon Roe for editing, and Cristina Iglesias Espinosa for helping with the production of the report. composition of the report. Great appreciation goes to ix Foreword A growing body of academic literature seeks to demonstrate By creating a novel index of corporate governance the relationship between a firm’s environmental, social and performance, this important empirical study demonstrates governance performance and its financial performance. a clear correlation between the quality of corporate While many such studies show a positive correlation, most governance in IFC’s portfolio companies and the financial, are based on data of listed equities in developed countries economic and environmental and social performance of IFC’s and few consider this relationship in the emerging market investment. A more focused corporate governance index context. based on the indicators that IFC’s corporate governance team believe to be the most important, reveals that the IFC, the private sector arm of the World Bank Group, materiality of corporate governance issues considered seeks to demonstrate and promote the role of the private serves to enhance this correlation. These results offer IFC sector to support economic development in emerging food for thought on how we more effectively incorporate markets. Operating as a triple-bottom line investor, it corporate governance into our investment due diligence invests mainly in private debt and equity with the goal of and how we provide corporate governance advice to our delivering financial returns, development outcomes and investees to improve their (and our) returns. sound environmental, social and governance performance. Regarding corporate governance, IFC has now been The findings of this study go beyond IFC’s interests as a development finance provider. They show the importance consistently incorporating the consideration of investee of corporate governance to financial performance in corporate governance practices into its due diligence emerging markets investment and encourage any prudent assessments for over a decade and provides advice on investor or lending institutions to develop approaches for good corporate governance practices to clients with need. assessing corporate governance risk factors in addition to The performance of IFC’s portfolio therefore provides a the traditional credit and financial risk factors. I encourage unique data set that can provide insight on the relationship investment professionals, company owners and executives between corporate governance, financial performance, alike to review this study to recognize and reinforce the development outcomes and environmental and social win-win proposition of good corporate governance practices. performance of firms in emerging markets. Ethiopis Tafara Khawaja Aftab Ahmed Vice President and General Counsel Director Legal, Compliance Risk & Substantiality Investment and Credit Risk xi Key Terms and Acronyms CG corporate governance COE cost of equity Core CG Index Index based on 26 key CG indicators of 84 CG questions CRR credit risk rating CSO The Department of Special Operations DARP Distressed Asset Recovery Program DOTS Development Outcome Tracking System E&S environmental and social ESG environment, social, and governance ERR economic rate of return ESRR environment and social review rating EROE economic return on equity EROIC economic return on invested capital FRR financial rate of return Full CG Index Index based on 84 CG questions GMM generalized method of moments IRR internal rate of return KPI key performance indicator ODO overall development outcome rating P/BV price to book value xii Governance and Performance in Emerging-Markets Firms P/E price to earnings ratio PSD private sector development ROA return on assets ROE return on equity ROIC return on invested capital SOE state-owned enterprise SPV special purpose vehicle WACC weighted average cost of capital Key Terms and Acronyms xiii Executive Summary In 2016, IFC’s Corporate Governance Group launched an empirical study to explore the link between the quality of IFC portfolio clients’ corporate governance (CG) and their financial and economic performance over a four-to-five-year period. Using client surveys and portfolio financial, economic, and development outcome data, IFC tested the hypothesis that better corporate governance is associated with better performance over a defined period. The analysis of collected data showed that investing in Further analysis, using the average CG score throughout companies with better CG at the time of disbursement of the investment period, shows that companies from the IFC’s investment is associated with a better average credit top quartile exhibit an average ROE that is about 20 risk rating (CRR) by almost 1.50 points throughout the percent higher than that of the bottom CG quartile IFC investment period. Specifically, on a constructed index (13.05 percent versus 10.96 percent). of CG performance, the top quartile of IFC portfolio companies had an average CRR of 4.62 on an 11-point For this study, IFC created the Core CG Index, consisting scale, compared with an average CRR of 6.08 of those of the most important CG indicators selected based in the bottom 25 percent. This suggests that selecting on consensus of IFC corporate governance specialists companies with better CG policies and practices at the prior to undertaking the empirical analysis. Using this time of investment is correlated with a lower credit index (instead of the full index of over 80 indicators) risk for IFC. produced an even stronger result: companies from the top CG quartile exhibit an average ROE that is The study also found that companies that improved CG about three times higher than that of the bottom CG during the investment period achieved about 20 percent quartile (18.56 percent versus 6.91 percent). These higher performance as shown by ICF’s Development results indicate that the most materials indicators of Outcome Tracking System (DOTS) financial indicator, CG performance demonstrate a stronger correlation which includes return on equity and return on invested with financial performance. capital (ROE/ROIC), and by the DOTS economic indicator (DOTS Econ), which includes economic return on equity Analyzing the subsets of companies in financial and and economic return on invested capital (EROE/EROIC). nonfinancial sectors, companies where IFC invests in On a 1–4 DOTS scale, companies that improved their debt versus equity, and the size of the investment (as CG practices by more than 25 percent displayed 3.72 represented by the tier of the transaction) produced financial indicator points and 3.94 economic indicator results similar to those described above, demonstrating points, against 3.09 financial indicator points and 3.30 that CG is an equally important factor across sectors economic indicator points for the group that did not and irrespective of the type of investment relationship. improve CG at all. These findings demonstrate that Finally, the study analyzed the relationship between CG investing time and resources in improving a company’s and an environmental and social risk rating (ESRR). CG is positively associated with higher financial and Specifically, companies from the top CG quartile exhibited economic returns for IFC and its clients. an ESRR of 2.01 on a 1–4 scale, versus an ESRR of xiv Governance and Performance in Emerging-Markets Firms 2.19 for those in the bottom quartile. This suggests In conclusion, it is important to highlight that the study that companies practicing better CG also have better supports the following propositions: environmental and social (E&S) practices. • CG has statistically significant correlation with firm Moreover, the study identified specific CG practices that performance, especially regarding CRR and ROE/ have the highest correlations with financial performance: EROE. • IFC is likely to benefit from selecting companies • Having a dedicated internal audit function with its with better CG at the time of disbursement as well as own charter or terms of reference; from helping companies improve their CG practices • Following internationally recognized standards on during the investment holding period. internal controls; • Some practical actions may add significant value to • Having financial statements audited by a recognized IFC and its clients, such as 1) defining standard CG independent auditing firm; key performance indicators (KPIs) to focus on at • Having a written code of ethics/conduct; deal origination; 2) having CG specialists provide input for credit/equity ratings; 3) better monitoring • Having a board that has an audit committee; of CG covenant implementation and supporting the • Having a written policy for the approval of related- client in CG improvements (such as through CG party transactions. Advisory Services); and 4) prioritizing CG specialists’ involvement in projects with high CRRs and ESRRs. These indicators show that the broad spectrum of CG improvement actions available to a firm can be narrowed • To the extend IFC broad emerging market portfolio down to a key set of actions that make value creation illustrates the wider role of CG to commercial and a priority for the benefit of investors. investment success in emerging markets. This study provides other investors with an empirical basis for These practical findings support the proposition that CG increases scrutiny of CG performance. is associated with better results and is an important factor in the success of IFC’s investment and portfolio work. At the same time, it also suggests that IFC’s investment and portfolio teams and our clients can clearly benefit from simple efforts to improve a company’s CG policies and practices. The study covered 61 companies selected from different regions (from 45 countries) and industries (49 percent coming from the financial industry and the rest from nonfinancial sectors). While the sample of companies was relatively small for this type of empirical research, it represented more than 20 percent of IFC’s new investments during the study period (fiscal years 2011 and 2012) and an aggregate IFC disbursement of about $1.5 billion. Executive Summary xv Introduction Investors, including IFC and other development finance institutions, increasingly look at governance as an indicator of firm quality and a factor in investment selection. There is a strong business case for linking sound corporate governance practices to better firm performance. This proposition is based on the view that companies adopting governance best practices make better business decisions over time, better manage their risks, enjoy enhanced market reputation, and have improved access to capital. From a macro perspective, CG may also contribute to country-level social and economic development. “Sound corporate governance helps businesses attract investment on better terms. Clients are more accountable to investors and responsive to stakeholder concerns. They also operate more efficiently and are able to better manage risks.” —IFC website1 Over the past two decades, numerous academic studies the empirical link between CG and performance with worldwide have investigated whether the link between CG data from IFC’s investment portfolio, using a research quality and firm performance is supported by empirical methodology in line with other studies in the field but data.2 Most research finds a positive relationship between adapted to IFC’s investment environment. Specifically, following recommended CG practices and financial it investigates the extent to which having good CG performance indicators, especially when using market policies and practices—as defined by the CG Index value ratios such as Tobin’s Q, price-to-book value (P/ specifically developed for this study based on IFC’s BV), and price-to-earnings (P/E). (See Box 1.1.) CG Methodology—is associated with better financial and nonfinancial performance of IFC’s portfolio clients Because most research relies on publicly available during the investment period.3 It reflects IFC’s nature as a information, it is tailored to CG and performance analysis development institution, where the success of an investee of the largest companies and those that are publicly company is defined not only in financial terms but also listed. As a result, there is a possible gap between the by its contributions to a broader set of stakeholders findings of CG studies, as they may not be relevant to and the economy in general, including environment unlisted, often family-owned, businesses—the ones that and social aspects. The study’s methodology has been IFC typically invests in. Therefore, this study explores submitted to peer review by internal and external experts 1  http://www.ifc.org/wps/wcm/connect/topics_ext_content/ifc_external_corporate_site/ifc+cg/investment+services 2  One of the most frequent approaches in corporate governance research assigns a “score” to each company based on its adherence to certain practices. Usually this score is based on such governance dimensions as board of directors, shareholderrights, transparency, and so on. Researchers then assess whether the corporate governance score correlates with performance measures. 3  This is the so-called “research problem.” The term “problem” does not mean that something is wrong; it just means there is a gap between the actual state (IFC investments on CG initiatives) and the desired ideal state (to assess whether IFC efforts to promote better CG among its investment clients do pay off). The “scientific method” to solve problems is a step-by-step logical, organized, and rigorous procedure to identify the problem, gather data, analyze the data, and draw conclusions from them. xvi Governance and Performance in Emerging-Markets Firms Box 1.1: Continuing Debate on the Real Impact of CG Practices on Performance Several studies have concluded that companies with superior governance practices tend to exhibit better operating profitability and are traded at premium prices over their peers. However, an important caveat regarding these findings is this: A positive relationship between CG practices and better performance does not necessarily mean that the superior performance is a consequence of higher CG quality. It is the difference between correlation and causation. Best governance practices, for example, may be positively correlated with better performance due to other factors (such as firm size, market power, and so on) or because it is a consequence instead of a cause of better performance. This causality issue is very difficult to solve from the econometric point of view due to three so-called endogeneity problems: omitted variables, simultaneity or reverse causality, and measurement error of the regressors (especially the construct validity of CG measures). Researchers are still trying to overcome these problems by resorting to increasingly complex econometric procedures, such as random and fixed effects procedures, instrumental variables based on changes in regulations, dynamic and systemic generalized method of moments (GMM), and so on. Thus the debate about the real impact of CG practices on performance is still not settled in the academic world. Interestingly, the most robust results have been found using data from emerging markets. Thus it is likely that having sound CG practices is a more important factor for the performance of companies located in countries with the worst investor protection. Examples of studies referred to above are Klapper and Love (2002), Gompers et al. (2003), Durnev and Kim (2005), Black, Jang, and Kim (2006), Brown and Caylor (2006), Silveira and Barros (2007), Leal and Carvalhal-da-Silva (2007), Chen, Chen, and Wei (2009), Renders, Gaeremynck, and Sercu (2010), Ammann, Oesch, and Schimd (2011), CLSA Group (2012), Black et al. (2013), and Hitz and Lehmann (2013). In addition, Claessens and Yurtoglu (2013) provide a recent review of research on corporate governance with a special focus on emerging markets. as well as investment professionals, and their feedback in IFC’s portfolio so that the effects of the investment was considered. relationship between IFC and the clients could materialize. The study sample comprised companies that IFC invested CG data were collected directly from IFC’s portfolio in during FY11 and FY12 (July 1, 2010, through June companies, based on the CG Index questionnaire, which 30, 2012). The study period was selected to ensure that included information as of the date of the original no external and extraordinary global events influenced investment and at the end date of the study (June 30, CG or performance indicators on the firm level (for 2016). Performance data, on the other hand, were already example, when the consequences of the 2007 financial available in IFC’s internal databases and comprised crisis have likely subsided) and to allow for an extended financial, economic, credit, E&S, and private sector period of five to six years of holding the companies development indicators. 4  A recent paper by Desai et al. (2017) may be the only academic paper in this line of research carried out with data from IFC operations. The authors investigate the relationship between ESG (environmental, social and governance) ratings and financial performance of some 1,000 IFC projects between 2005 and 2014 in nearly 100 middle- and low-income countries. By using an instrumental variables approach that in their view corrects for endogeneity issues, they find that the relationship between ESG and financial performance is not statistically significant. They also find evidence that both ESG and financial performance can jointly affect broader private sector development. It is important to note that, although authors use the term ESG to describe their main variable of interest, they only use E&S indicators coming from IFC DOTS and do not include governance indicators in their examination. This substantially differentiates the focus of their study from this one, as recognized by the authors in footnote 8 (page 9) of their paper. Introduction 1 The study did not seek to prove causality between CG including in the form of advisory services, to further and performance. Instead, it assessed the degree of the increase the value add to its clients and investment mutual relationship between the two and lessons to be portfolio. drawn by IFC investment teams as they structure and manage IFC’s engagement with investee firms. Future The publication and dissemination of the findings of improvements in standardizing CG and performance the study may have an effect that could go beyond data collection may make possible significantly more IFC’s own investment operations and IFC’s investee data observations, which in turn may allow for a more companies and have a broader development impact in robust analysis, including seeking to identify causality better articulating the relationship between CG and from CG to performance. While the study’s findings performance, both on the risk side and in value addition. demonstrate statistically significant correlation between In addition to these internal objectives, this study makes CG and performance, the relatively small size of the important contributions to the discourse on CG and sample limited the ability to conduct more robust performance in emerging markets. First, the study adds econometric analysis, such as multiple regressions.4 to a body of academic research by including unlisted Also, some other study limitations, often common in companies from emerging markets. Second, IFC can similar research studies conducted outside IFC, include further leverage its strong relationships with other 1) reliance on companies as the source of CG information development finance institutions by encouraging them (even though the relevant portfolio officers reviewed to conduct similar studies. Alternatively, future similar and verified information to the best of their knowledge); studies may include companies from other institutions 2) recall bias of remembering CG policies and practices and make market-level efforts to disseminate the findings dating back to 2011; 3) limited heterogeneity of the data of this study in their regions and priority countries, thus sample of IFC clients that have been prescreened and multiplying the effect of this report. are typically market leaders; and 4) absence of severely underperforming companies (due to financial distress or to lack of a strong relationship with IFC) from the study sample. Despite these limitations, inherent to most empirical studies worldwide, the key findings of the study can inform IFC’s investment operations and its approach to private sector development. First, the study findings based on hard data reinforce the business case for improving governance practices of IFC investees. Integrating CG risk assessment into deal origination, structuring, and portfolio supervision may better enable financial and nonfinancial value to be extracted across a project’s lifecycle and, more significantly, for poor CG performance to be corrected before it is manifested in debt default or erosion of equity. Second, by identifying distinct governance factors that are strongly correlated to better performance, the study helps establish a practical framework for prioritizing governance improvements by firms. Third, the findings support IFC’s 3.0 strategy of building sustainable markets and providing support, 2 Governance and Performance in Emerging-Markets Firms Methodology The study’s overall goal was to determine whether IFC clients’ practice of better corporate governance is associated with better performance during the investment period. To accomplish this, the study team designed a research methodology in line with other studies in the field but tailored to IFC’s portfolio. 2.1. OBJECTIVES creation. If the third objective is empirically supported, the findings would demonstrate that attention to CG First, the study aimed to find out whether better CG practices should be continuous rather than a one-time exercise. of IFC clients at the disbursement date are associated with better performance during the investment period. Second, it analyzed data to check whether companies 2.2. FIRM-LEVEL with greater improvement of their CG practices from PERFORMANCE VARIABLES IFC’s disbursement date until the end of the study period It is important to recognize that not all aspects of performed better throughout this period. And finally, performance are interrelated; there are many ways to the study tested whether the average level of CG in IFC measure firm performance. Most external research on clients is associated with better firm performance. CG measures financial indicators from the perspective of shareholders via measures such as the ROE or market This last objective was included in the study based on value ratios, such as Tobin’s Q, P/E, P/BV, and so on. consideration that CG practices tend to be relatively Firm performance can also be understood from a stable over time (as demonstrated by the fact that about broader perspective that includes outcomes to other one-quarter of the sample did not report any variation relevant stakeholders, such as employees, customers, of CG practices during the five years under analysis), communities, and society. and it allows for increasing of the number of firm-year observations of the sample. This methodological technique A challenge in this study was to include relevant is widely accepted in the academic field for similar measurement parameters, which IFC actively and reliably empirical studies. tracked for the study period from July 1, 2010, to June 30, 2016. Accordingly, the study considers four well- These objectives are related to perspectives that are gaining established indicators that IFC tracked across its portfolio: general acceptance among CG as well as environmental, social, and governance (ESG) practitioners. If the first • Return on Equity/Return on Invested Capital: ROE objective is supported by empirical data, then the results and ROIC are measures of financial performance would reinforce the need for IFC and other similar investors that reflect the returns to financiers. Annual ROE to screen and select companies with high-quality CG is calculated as net income for the most recent year, practices at the disbursement date. If the second objective divided by average equity (average between the most is supported by the analysis, then the findings would recent year and the previous year). ROIC is computed reinforce the case for IFC and the client companies to as the ratio of net operating profits, less adjusted focus on improving their CG practices to maximize value Methodology 3 taxes (NOPLAT) to total invested capital. (NOPLAT • Development Outcome Tracking System and Overall reflects only profits from core operations, less related Development Outcome (ODO)  Ratings: IFC’s income taxes.) proprietary DOTS measures and monitors projects’ development impacts by assigning ratings to four • Economic Return on Equity/Economic Return on performance areas, described below, with ratings on a Invested Capital: EROE and EROIC are measures of 1–4 scale (excellent, satisfactory, partly satisfactory, and economic performance that take into consideration unsatisfactory). The ratings rely on a baseline, target, returns not just to financiers but also to society as and actual results assessment and define qualitative and a whole. Annual EROE is calculated as net income quantitative benchmarks to assess projects and thus for the most recent year, adjusted for costs and complement analysis of raw financial or nonfinancial benefits to society (such as taxes paid, subsidies, performance data (also analyzed separately in this interest caps, and free or not fully priced advisory study). DOTS also provides an aggregate synthesis services), divided by average equity (average between rating, the ODO, on a 1–6 scale (highly unsuccessful, the most recent year and the previous year). Annual unsuccessful, mostly unsuccessful, mostly successful, EROIC is computed as net operating profits, less successful, and highly successful). DOTS considers the NOPLAT, adjusted for costs and benefits to society, following four dimensions of a project, with specific divided by invested capital. measurement metrics for Financial Markets and for • Credit Risk Ratings: IFC’s in-house CRRs used in this nonfinancial sectors—Manufacturing, Agribusiness, study range from 1 to 7, where 1 indicates the lowest and Services (MAS) and Infrastructure and Natural risk, and 7 the highest risk level (commensurate with Resources (INR) firms, as relevant: default). Additional modifiers (A and B) add further granularity in the CRR for categories 2 to 5, where • Financial performance (DOTS Fin) tracks project returns to determine whether they are sufficient A defines a higher credit quality than B within the to compensate financiers for risks taken. It category. The CRR categories are very good (1), good uses ROE for Financial Markets and ROIC or (2), average (3), watch (4), substandard (5), doubtful financial rate of return (FRR)  for nonfinancial (6), and loss (7). For the purposes of this study, the sector projects.6 team computed the CRRs converted into two different scales: from 1 to 7 (just removing the letters A or B • Economic performance (DOTS Econ)  tracks the benefits accrued not only to financiers but from the credit risk assigned) or from 1 to 11 (by also to stakeholders most directly affected converting letters A and B into numbers). The study by projects: customers, employees, suppliers, relied on CRR ratings per the methodology in 2011 taxpayers, and government. ROE and EROE and 2012, although this has since been replaced by are compared with cost of equity (COE)  for the IRP (Investment Risk Platform) methodology, Financial Markets firms, and ERR or EROIC which adheres to a different scale.5 are reviewed for nonfinancial sector firms.7 5  For equity operations, the team computed the equity risk rating instead of the credit risk rating—and used the CRRs for analyses when the client had both credit and equity ratings assigned. CRRs are based on general and (loan or equity) specific risk factors, which are rated and aggregated to provide a numeric score. Some of the risk factors are inputs based on quantitative data; others are based on qualitative data and thus require judgment from the IFC investment team. For financial institutions, general risk factors include country situation, regulatory environment, sector structure, internal organization, management quality, corporate governance, and earnings. Loan factors include capital, loan assets, investment in securities, liquidity, and foreign exchange open exposure. Equity factors include equity value and liquidity. For nonfinancial institutions, CRRs rate risks differently, where general risk factors considered are country situation, market situation, management quality, and profit margin. Loan factors are debt service and arrears record, debt service strength, security arrangements, and profit trends. Equity factors include value versus original cost, dividend record, future earnings growth, and exit mechanism. Although CG was one of the qualitative factors used in defining the CRR, the CG Index score used in the study is significantly broader and thus does not create circularity of having CG as both an element of CRR rating and an indicator compared with CRR. 6  DOTS Fin has two mandatory financial return indicators: ROE for Financial Markets projects, and annual ROIC for the MAS and INR industry groups. Ratings are assigned based on comparing the FRR to the cost of capital: FRR/ROIC compared with weighted average cost of capital (WACC) and ROE to cost of equity (COE). Also, DOTS Fin ratings are supplemented by an assessment of the main drivers of financial return, such as net income, sales, margins, capacity use, project cost, and so forth. 4 Governance and Performance in Emerging-Markets Firms • Environmental and Social (DOTS E&S) It is important to note that indices vary substantially performance assesses the benefits that accrue across studies and there is no universally accepted to the environment in which the project takes instrument in the literature or among practitioners. place. DOTS E&S also measures the extent to Some practitioners argue that measuring CG quality which IFC clients satisfy the environmental, via indices and scores may be potentially misleading. social, and health and safety standards set forth in IFC’s sustainability policy and performance One issue is the construct validity and whether CG indices guidelines.8 are indeed able to segregate genuinely well-governed companies from those with poor governance. For instance, • Private Sector Development (PSD) impact captures the benefits accrued to actors beyond the IFC client some firms, to please their external stakeholders, might that are influenced by the company project, such create the perception of adopting good practices while as its supply chain, industry, or country location.9 their day-to-day practices do not truly reflect the standards presented. Moreover, there is no ideal CG model that can • Environmental & Social Review Rating: ESSR indicates a company’s capability and/ be implemented by all companies. Because firms differ or management of E&S issues in accordance in size, sector, life-stage, ownership structure, strategy, with IFC’s Sustainability Framework: excellent country of origin, and so on, it would not be practical (1), satisfactory (2), partly unsatisfactory (3), to apply the same CG instrument to all firms.11 Indices unsatisfactory (4)—taking into account the are often unable to capture the country factor, because sector-specific and project-specific risks identified companies differ not only in their own practices but by IFC at the time of investment and implantation also in the requirements of the legal and regulatory of E&S action plans agreed with the company. frameworks of their home jurisdictions. While acknowledging these limitations, much of the These indicators were extracted from IFC internal systems academic literature considers CG indices to be at least or directly from the clients’ financial statements and reasonable proxies of the firm’s corporate governance annual reports. quality. Usually, these papers create indices composed of questions that are binary (yes or no), objective (based 2.3. CORPORATE GOVERNANCE on a clear criterion), and based on publicly available SCORE information. This standard reduces the subjectivity of the instrument and facilitates replicability by third parties. Evaluating corporate governance is inherently a subjective and complex process. Most research measures CG For this research, the study team constructed the CG through indices that focus on a company’s application Index and tailored it for use as the primary instrument for of recommended best practices.10 evaluating CG against performance indicators. The Index drew from the experience of previous research studies 7  Since not all economic costs and benefits can be quantified, the rating of economic performance also considers qualitative aspects, including to what extent a project has contributed to IFC’s mission of helping reduce poverty and improve people’s lives. 8  DOTS E&S ratings offer an assessment of the environmental and social benefits that accrue from the project’s operations. While DOTS E&S ratings incorporate ESRR ratings, the two can differ based on factors other than compliance with the ESAP (Environmental and Social Action Plan) and Performance Standards. 9  Because of its indirect link to the project, the private sector development impact is more difficult to assess, as quantitative data are less readily available. For this reason, DOTS PSD ratings review the project’s original objective and quantify or explain the benefits accrued to the private sector, such as improvement in the enabling environment, contribution to market efficiency, and so on. 10  Examples of academic papers that use corporate governance indices are Klapper and Love (2002), Gompers et al. (2003), Durnev and Kim (2005), Black et al. (2006), Brown and Caylor (2006), Silveira and Barros (2007), Leal and Carvalhal-da-Silva (2007), Chen, Chen, and Wei (2009), Renders et al. (2010), Ammann et al. (2011), CLSA Group (2012), Black et al. (2014), and Hitz and Lehmann (2013). An alternative approach to measuring the real value of corporate governance is based on event studies, which observe the stock price reactions to news related to corporate governance. Nguyen and Nielsen (2010) and Silveira and Dias Junior (2010) are examples of studies using this approach. 11  Bhagat et al. (2008) and Bozec and Bozec (2012) provide critiques on the use of corporate governance indices. Methodology 5 and G20/OECD Principles of Corporate Governance, 2.4. COMPANY SAMPLE AND and it relied heavily on IFC’s own CG Methodology CG DATA COLLECTION and its five dimensions: 1) firms’ commitment to CG, 2) structure and functioning of the board of directors, The study identified an initial sample of about 330 3) control environment and processes, 4) transparency portfolio companies with investments disbursed between and disclosure, and 5) shareholders’ rights. Incorporating July 1, 2010, and June 30, 2012, and where it was these five dimensions aligns the CG Index with IFC’s possible for IFC to review and oversee CG commitments existing approach for CG analysis in the investment of the company.15 The study period was selected to process.12 Because the number of questions for each ensure that no external and extraordinary global events CG dimension and category differs, the scores obtained influenced the CG or performance indicators on the were standardized as a percentage to ensure that all firm level (such as when the consequences of the 2007 dimensions have the same weight in the aggregate CG financial crisis had likely subsided) and to allow for score.13 Table 2.1 and Figure 2.1 illustrate the output an extended period of five to six years of holding the of the CG Index divided into its five dimensions. Box companies in IFC’s portfolio so that the effects of the 2.1 describes an advantage of this approach. investment relationship between IFC and the clients could materialize. The correlation analysis also used a reduced Core CG Index composed of 26 core questions (shortlisted All companies from the sample received invitations from the total 84 questions in the Full CG Index). IFC to complete the CG questionnaire, and 61 companies corporate governance specialists selected the most provided completed CG questionnaires with data on their important indicators from the Full CG Index to narrow CG practices at both the initial year of the disbursement the focus of CG on key indicators prior to undertaking (2011 or 2012) and the end of the study period (June the empirical analysis, to ensure that the results are 30, 2016). Tables 2.2 and 2.3 show the industry and not biased.14 The construction of the Core CG Index region distribution. is important, because it may be easier to replicate it for subsequent studies with a larger sample of companies Surveyed companies represent all regions and industries, and for it to become part of a routine CG assessment but there is an industry concentration on financial for all new IFC clients. institutions (49 percent of the sample).16 There also is a regional concentration in Latin America and the 12  Some companies may fit into more than one category. For example, financial institutions may also be classified as listed and/or family-owned. In such a case, the study applies all questions from every category that the firm fits into. 13  Most corporate governance indices constructed in the literature adopt equally weighted questions (each one adding one point to the overall score), because it is easier to reproduce. However, this study is designed to also be able to assign a greater weight to some governance practices perceived as more relevant. 14  The CG questions that are part of the Core CG Index are the following: 2, 4, 6, 10, 14, 15, 16, 19, 21, 24, 30, 31, 33, 38, 39, 42, 45, 48, 54, 56, 59, 62, 69, 72, 75, and 81. Appendix C provides a full description of each question. 15  Of 440 equity and debt investment projects disbursed between July 1, 2010, and June 30, 2012, clients where CG is not a key factor and with whom IFC had very limited oversight or engagement on CG were excluded. These included projects with the following characteristics: 1) investments in funds, because the governance structures and processes of funds are significantly different from those of corporations; 2) rights issues, because these are repeat deals where IFC simply exercises anti-dilution right with no action taken on CG; 3) where the only investment product is risk management, hedging, or currency-swap arrangements; 4) projects in related companies done in the same country at the same time, or special purpose vehicles; 5) Distressed Asset Recovery Program projects, because these are investments in companies managing distressed assets of financial institutions on a contractual basis; 6) investments in state-owned enterprises and municipal enterprises; 7) inactive clients, recent prepayments, or equity exits in process with whom IFC had no active relationship; 8) clients excluded at the request of the IFC investment officer due to a poor relationship or a situation of financial distress, bankruptcy, or liquidation; and 9) repeated operations for a company, because one project per company/partner was selected for the study. 16  Although it is recognized that CG standards and practices often are different in financial and nonfinancial sectors, in listed and unlisted companies, and IFC’s ability to influence CG and performance of the client companies varies depending on the nature of the investment relationship (debt versus equity), the small overall sample size of this study did not allow for conducting a separate analysis for each of these subgroups. However, the study results within each subgroup were similar to those for the total sample, as shown in section 3.2.6. 6 Governance and Performance in Emerging-Markets Firms Table 2.1: An Indicative Output of the CG Index Contribution of each % Adherence Weight Dimension to Topic Dimension # Questions Client Score per Dimension Dimension Final Score 1 Commitment to Corporate 7 5 71% 20% 14% Governance 2 Structure & Functioning of 38 29 76% 20% 15% the Board 3 Control Environment & 25 13 52% 20% 10% Processes 4 Transparency & Disclosure 18 15 83% 20% 17% 5 Shareholders Rights 12 4 33% 20% 7% SUM 84 66 Overall % of CG Adherence 63% 0–100% Figure 2.1: Indicative CG Index Output Adherence of recommended corporate governance practices (%) Overall % of adherence = 59.0% Shareholders rights 38% Transparency and Disclosure 57% Control Environment and Processes 64% Structure and Functioning of the Board of Directors 74% Commitment to Corporate Governance 63% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Caribbean companies (36 percent of the sample) due institutions. Regarding classification, 5 respondents are to the higher response rate. The total 61 respondents Tier i, 42 are Tier ii, and 8 are Tier iii.17 Regarding the represent an aggregate disbursement of about $1.5 type of financing, 25 operations are only equity deals, billion in debt and equity investment from IFC. Of 33 are only debt operations, and 3 are both equity and the 61 respondents, 18 are listed entities, of which debt clients. 13 are nonfinancial sector entities and 5 are financial 17  Tiering of IFC investment transactions is defined as a combination of the credit risk rating and the size of the investment, with Tier iii transactions, on average, being the most risky and largest in size. Methodology 7 Box 2.1: Advantage of a Broad CG Index An advantage of constructing the CG Index with dozens of questions—resulting in a 0–100 percent score—is that it increases variance across the sample firms and the chances of finding statistically significant correlations. This is important for this study, given its relatively small sample of companies and short time span. Further, grouping questions into five distinct dimensions in the CG Index allows a more nuanced analysis of the impact on firm performance of the overall dimension, as well as of specific practices within each dimension (such as the presence of an audit committee within the Control Environment and Processes). As a hypothetical counterexample, a concise corporate governance indicator composed of just four levels (in which most companies would probably be placed in the same group) probably would not provide the variability of the CG quality indicator that allows for statistically significant results. 2.5. STUDY LIMITATIONS AND companies in financial distress were excluded, and firms THE CAUSALITY ISSUE with better relationships with IFC were more likely to participate). Box 2.2 provides further discussion of This study seeks to investigate the link between CG sampling bias. and the performance of IFC clients; it does not seek Second, given that there has been no collection of the to prove a causal relationship—from CG practices to CG data in the past, the assessment of CG practices of performance. Given the small number of firms in the IFC clients at the time of the disbursement in 2011 or study sample as well as data from a relatively short 2012 was done by asking clients about it at the end of time span of four to five years, the research provided 2016. The accuracy of data is potentially compromised a small number of firm-year observations. This in turn by recall bias and by the possibility that clients would limited the ability to employ multiple regressions based tend to give the same response for both periods. However, on sophisticated econometric procedures—such as the the collected data were screened by the portfolio officers GMM techniques—that are supposed to at least mitigate with knowledge of the company to verify, to the best endogeneity issues. Thus the study cannot claim any of their knowledge, the accuracy of the information causality running from CG to firm-level performance. It provided. Nevertheless, some conclusions may have should be seen, therefore, as a correlational study instead been impaired. of a causal one, from a methodological perspective.18 The third limitation is that the sample is composed of This caveat derives from the observational characteristic companies from different industries and regions around of the data19 as well as from important limitations related the world. This introduces wide variations in stages of to the sample and statistical methods used in this research. maturity in regulation, financial practices, investment First, the study analyzes a small sample of 61 companies climate, political risk, and so on. Influence of both the that responded to the CG questionnaire and may be industry and country factors on CG can be multifaceted potentially biased toward better firms (for example, and inadequately considered in the study. 18  A correlational study is conducted in the natural environment of the organization, with minimal or no interference by the researcher with the normal flow of work. In a causal study, on the other hand, the researcher tries to manipulate certain variables to study the effects of such manipulation on the dependent variables. 19  The data in this study are observational in the sense that the researchers are not able to manipulate their variable of interest (the adherence to CG practices) to analyze its effect on the dependent variable (firm-level performance). 8 Governance and Performance in Emerging-Markets Firms Table 2.2: Industry Breakdown of the Sample Number of Industry Companies % of Sample Telecommunications, Media, and Tech & Venture Investing (TMT) 3 5% Financial Institutions Group (FIG) 30 49% Infrastructure and Natural Resources (INR) 11 18% Manufacturing, Agribusiness & Services (MAS) 17 28% Total 61 100% Table 2.3: Regional Breakdown of the Sample Number of Region Companies % of Sample Latin America and the Caribbean (LAC) 22 36% Europe and Central Asia (ECA) 10 16% East Asia and the Pacific (EAP) 7 11% Sub-Saharan Africa (CAF) 9 15% South Asia (SA) 8 13% Middle East and North Africa (MENA) 5 8% Total 61 100% However, this last point is less of a concern for the for all companies. For example, DOTS applies different financial indicators, which (via CRR and DOTS rating performance indicators to firms belonging to different methodology) do incorporate country credit evaluation industries. While financial institutions are assessed by in the ranking assigned by IFC. For instance, the country indicators such as new loans to SMEs, manufacturing credit rating is a factor in the IFC CRR rating assigned companies are assessed by indicators such as the gross on an annual basis. But it is likely that data available value added. to create variables may not be uniformly constructed Methodology 9 Box 2.2: Sampling Bias A sampling bias occurs when a sample is collected in such a way that some members of the intended population are less likely to be included than others. This is the case with this study. For instance, it excluded companies in financial distress, and the exclusion of extreme bad performers results in a biased sample. The study also may have incurred into a selection bias, which is the selection of individuals, groups, or data for analysis in a way that precludes achievement of proper randomization, and the sample obtained is not representative of the population intended to be analyzed. Consequently, our results are not generalizable to the whole population of IFC operations. In addition, there are two particular issues related to selection bias that tend to impair the ability to generalize the study results to the market. The first derives from the fact that IFC does not randomly select its client companies (IFC may, for instance, already select companies that are more profitable and with low risk). The second is that IFC requires companies to adopt some corporate governance practices to be eligible to receive its investment. As a result, the sample is probably more homogeneous in CG practices than firms from the market that are not part of the IFC portfolio. 10 Governance and Performance in Emerging-Markets Firms Findings The database of CG and performance indicators constructed for the study facilitated two principal avenues for investigation: • Subgroup comparisons—to test whether there is a significant difference of means between at least two subgroups (for example, companies belonging to the top 25 percent in improvement of CG practices versus the bottom 25 percent) on a performance variable of interest, such as ROE, by running two-sample difference of means tests. • Data reduction analysis—to identify the most relevant CG questions, and to identify which specific CG questions are stronger predictors of performance indicators. 3.1. CG SCORE ANALYSIS Control Environment (26 percent) and Commitment to CG (25 percent), while in the area of Shareholder The review of CG scores of the companies from the Rights the real sector companies displayed better CG study sample at the time of IFC’s original disbursement practices by 4 percent than financial institutions. shows some interesting findings:20 The outperformance of financial institutions in two dimensions could be due to higher regulatory On average, companies in the sample followed 60 percent requirements for the financial industry and because of recommended CG practices at the time of IFC’s original such companies are generally more mature in their disbursement. This shows that companies usually adopt life stage. more than half of the corporate governance practices • Interestingly, CG scores of companies where IFC assessed by the study instrument at inception of their invested in equity or debt instruments are essentially relationship with IFC. the same. Moreover, while board practices are generally better in equity clients, debt clients outperform in In the analysis of each CG dimension, the highest score the area of Shareholder Rights. is in the area of Transparency and Disclosure (73.4 percent) followed closely by Control Environment (71.1 • Listed companies in the sample have on average 6 percent), while the area with the lowest score is Shareholder percent higher CG scores than those not listed on Rights (58.9 percent) followed by Commitment to CG stock exchanges, with Transparency and Disclosure (62.6 percent). and Control Environment displaying the largest difference between the scores of the two groups. • If reviewed through the industry lens, the CG scores However, unlisted companies reported on average of financial institutions overall are 14 percent better 16 percent better practices in the area of Shareholder than those of real sector companies (67 percent against Rights. Possibly, this is explained by simplicity of 53 percent). The largest differences are in the areas of shareholder protections in unlisted companies where 20  For more details on CG scores at disbursement and end of study, and the variation of the scores of different subgroups,see Appendix A. Findings 11 existing shareholders rely on private shareholder sector companies, especially in the areas of Commitment agreements to protect their interests. to CG and Control Environment. In other aspects, such as type of investment product, size of investment, and • Finally, the size of IFC investments (Tier iii versus companies’ listing on stock exchanges, material differences Tiers ii and i) did not appear to be associated with were not observed, thus making the results from the the CG scores of the companies. total sample applicable to each subgroup as well. The study then analyzed the CG performance of investee companies during the portfolio holding period. The analysis produced the following observations: 3.2. SUBGROUP COMPARISONS • On average, companies in the study sample improved Against the backdrop of baseline CG information of the their CG score by more than 7.0 percent from the companies in the study sample, the subgroup analysis time of IFC’s disbursement until the end of study was conducted by segregating firms into four quartiles period (June 2016). Largest improvements were based on the following: in the dimensions of Commitment to CG (11.3 percent)  and Board Practices (8.0 percent), while • The companies’ level of application of CG practices the least improvements were in Shareholder Rights in 2011—to test the first objective of the study (to (1.0 percent). find out whether better CG practices of IFC clients at the disbursement date are associated with better • Although real sector companies generally displayed performance during the investment period); lower CG scores at the time of disbursement compared with financial institutions, the improvements of the • CG improvement over time—to test the second CG performance in these two groups were essentially objective (to discover whether companies with the same overall. Financial institutions appeared to greater improvement of their CG practices from focus more on improvements in Commitment to CG IFC’s disbursement date until the end of the study by improving their CG policies as well as prioritizing period perform better throughout this period); and improvements in their Control Environment. The • The average level of following recommended CG real sector companies also improved their formal practices from 2011 to 2016—to test the third objective CG policies while also making relevant strides in (to learn whether the average level of CG in IFC Transparency and Disclosure. clients is associated with better firm performance). • Companies with both equity and debt investments After segregating companies into four groups, the study also improved their CG scores on average by the same compared the performance of the top versus the bottom 8 percent. The difference between variation of CG quartile through difference of means tests. Results of this scores of listed and unlisted companies (6 percent procedure for each study hypothesis are presented below. versus 7) is negligible. But first, note in Table 3.1 the correlation coefficients • Interestingly, although companies with large or between the two main CG variables (Full CG Index small investments from IFC had similar CG scores and Core CG Index) and the performance variables. at disbursement, companies with smaller investments Table 3.1 presents statistically significant correlations improved more (8 percent versus 2 percent for larger between the two CG indicators and performance variables. investments). The correlations are even stronger for the Core CG The above description of CG scores at disbursement Index than for the Full CG Index. Specifically, the study and CG improvements throughout the study period team observed that all correlations between the Core for different subgroups shows that the most striking CG Index and all performance variables are statistically differences are among the financial institutions and real significant at the 1 percent level. 12 Governance and Performance in Emerging-Markets Firms Table 3.1: Correlations Between Selected Variables Full CG Index Core CG Index ODO 1–6 SCORE 0.2288*** 0.3408*** DOTS FIN 0.1012* 0.2596*** DOTS ECON 0.1768*** 0.2346*** ROE/ROIC 0.1067** 0.3105*** EROE/EROIC 0.0689 0.3106*** CREDIT RISK 1–7 –0.3289*** –0.2775*** CREDIT RISK 1–11 –0.3250*** –0.2863*** Note: The table exhibits Pearson correlation coefficients between CG variables in the first row and performance variables in the first column. Full CG Index is the level of adherence of the sample firms to the 84-question CG questionnaire. Core CG Index measures the average level of adherence of the sample firms to the 26 questions from the CG questionnaire (the subset of CG questions considered by IFC’s ESG department as the most relevant). ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 3.2.1.  First Objective: Effect of CG at score of 6.08 of the group that displays weaker CG Disbursement on Firm Performance practices. Thus investing in companies with better CG at the disbursement date is associated with a For the first objective, the study looked for any re- lower average CRR by 1.50 points throughout the lationship between a company’s CG practices at the investment period. disbursement date and its performance during the investment period. Table 3.2 compares the perfor- Other performance variables point in the expected mance indicators of two groups: Top CG refers to direction. However, the team did not observe statis- the group composed of the top quartile of firms with tically significant results for most of these findings. the highest CG scores at the disbursement date, and Thus the data only indicated partial support for the Bottom CG refers to the group of companies with first objective. It is likely that the small number of the lowest CG scores at the same date. observations may be a cause for the absence of sta- tistically significant results in this case. Figure 3.1 On average, companies from the top quartile exhib- compares the credit risk ratings of the two quartiles ited an adherence of 82 percent to the CG Index at according to the last two rows of Table 3.2. the disbursement date, and firms from the bottom quartile exhibited a much lower level of adherence of only 37 percent. 3.2.2.  Second Objective: Improvements of CG Practices and Performance As shown in Table 3.2, the main result is that compa- nies with the highest CG score exhibit a lower credit Table 3.3 uses the same analysis as the previous table risk rating. For example, when the credit risk rating to test the second objective. It compares the perfor- of the project is converted to the scale from 1 to 11 mance indicators of the top and bottom quartiles, (CRR 1–11), the quartile of companies with the high- looking specifically at improvement of CG practices est CG practices exhibit a credit risk of 4.62, which from 2011 to 2016. is significantly lower (at the 1 percent level) than the Findings 13 Table 3.2: CG Practices at the Disbursement Date FULL GC INDEX Bottom CG Top CG P25 P75 p-value a difference Mean adherence to CG Mean adherence to CG means Performance Variable Legend Index in 2011 = 37% Index in 2011 = 82% (P75–P25) ODO 1–6 SCORE Mean 4.03 4.37 0.098* SD (0.60) (0.77) n n=16 n=14 DOTS FIN Mean 3.36 3.34 0.515 SD (1.17) (1.11) n n=16 n=14 DOTS ECON Mean 3.54 3.78 0.279 SD (0.97) (1.18) n n=16 n=14 ROE/ROIC Mean 11.69% 12.64% 0.384 SD (8.78%) (8.38%) n n=15 n=14 EROE/EROIC Mean 15.96% 17.16% 0.380 SD (11.20%) (9.74%) n n=15 n=14 CREDIT RISK 1–7 Mean 3.83 3.15 0.000*** SD (0.39) (0.37) n n=12 n=13 CREDIT RISK 1–11 Mean 6.08 4.62 0.000*** SD (1.08) (0.77) n n=12 n=13 Note: The table exhibits mean-comparison tests between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% of firms in level of adherence to the 84-question CG questionnaire at the disbursement date; BOTTOM CG P25 refers to the bottom 25% in the level of adherence to the CG Index on the same date. ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. a The p-value, or probability value, is the level of marginal significance within a hypothesis test representing the probability of the occurrence of a given event. For the purpose of this study, a small p-value (typically ≤ 0.05) provides strong evidence that there is a significant difference between the performance indicators of top 75% and bottom 25% companies based on their CG scores. So a small p-value allows us to reject the claim that performance from the two groups is equal. On the other hand, a large p-value (> 0.05) indicates weak evidence against the null hypothesis, in the sense that it does not allow us to reject the idea that performance indicators from the two groups are statistically different. 14 Governance and Performance in Emerging-Markets Firms Figure 3.1: Credit Risk Ratings in CG Practices at Disbursement Subgroup comparison: Top 25% vs. Bottom 25% in terms of adherence to CG practices at the disbursement date: Credit Risk in 1–7 and 1–11 scales 7.0 6.08 6.0 5.0 4.62*** 4.0 3.83 3.15*** 3.0 Adherence to Adherence to Adherence to Adherence to Full CG Full CG Full CG Full CG 2.0 Index in Index in Index in Index in 2011: 82% 2011: 37% 2011: 82% 2011: 37% 1.0 Credit Risk 1–7 Scale Credit Risk 1–11 Scale Top 25% CG Bottom 25% CG On average, companies from the top quartile improved Figure 3.2 compares the average DOTS Fin and DOTS their CG practices by 25 percent from 2011 to 2016, Econ performance of the top and bottom quartiles as while firms from the bottom quarter exhibited a null per the second and third rows of Table 3.3. variation relative to the CG Index over this period. From the statistical viewpoint, the main result shown in 3.2.3.  Third Objective: Average Level of Table 3.3 is that companies with the greatest improvement CG Practices and Performance of CG practices achieved better performance in the As specified in section 2.1, the study also endeavored to overall financial and economic assessment of the project check whether the overall level of CG practices during (DOTS FIN and DOTS ECON). In the first case, firms the investment period is positively correlated with from the top quartile of CG improvement achieved performance indicators. Considering that CG practices DOTS Fin of 3.72 on a 1–4 scale, while the bottom tend to be relatively stable over time (as demonstrated quartile performers achieved a score of only 3.09. In by the fact that about one-quarter of the sample did the second case, top quartile firms exhibited DOTS not report any variation of their CG practices during Econ of 3.94, against a score of 3.30 from the bottom the five years under analysis), an average CG score for group. In both cases, the difference of means between the each firm is calculated for the five years under analysis two groups is statistically significant at the 10 percent based on their CG scores at the disbursement date and level. Companies that improved their CG more during end of the study period. the investment period, therefore, achieved about 20 percent higher performance in the average DOTS Fin Although admittedly not perfect in providing the indicator and DOTS Econ indicator. exact CG score of each firm on an annual basis, this procedure allowed us to check if companies exhibiting The results for other performance variables were not higher CG standards throughout the period of analysis statistically significant. Again, the small number of indeed outperformed those exhibiting poorer standards. firms in the sample have led to not finding statistically An important advantage of this approach is that, by significant difference of means between the two groups. computing the annual average CG score for each firm, this indicator was correlated with performance metrics Findings 15 Figure 3.2: Financial and Economic Improvement of CG Practices Subgroup comparison: Top 25% vs. Bottom 25% in terms of Improvement on the Level of Adherence to Full CG Index from 2011 to 2016 4.0 3.94* 3.72* 3.30 3.09 3.0 CG varied CG varied CG varied 0% (null CG varied 0% (null 2.0 +25% on variation) on +25% on variation) on average from average from average from average from 2011 to 2016 2011 to 2016 2011 to 2016 2011 to 2016 1.0 Average DOTS 1–4 Financial Performance Average DOTS 1–4 Economic Performance Top 25% CG Bottom 25% CG that are also calculated on an annual basis. So the mean Because of the increase in the number of observations, level of recommended CG practices of the sample firms the study team finds much stronger support for the from 2011 to 2016 was computed, as shown in Table third hypothesis. 3.4. This in turn increases the number of firm-year observations belonging to each quartile. Then the results were compared for companies in the top and bottom 3.2.4.  Core CG Index and Performance quartiles of average CG scores. Instead of testing correlations between CG and Because of the larger number of observations, there are performance with the Full CG Index, the analysis above statistically significant results for all performance variables is repeated here using the Core CG Index, consisting of in Table 3.4 except for EROE/EROIC. Specifically, the 26 questions, as shown in Table 3.5.21 companies with greatest mean CG Index score exhibit As indicated by the correlation table presented earlier in better performance than the group of CG laggards in this section, correlations between CG and performance the areas of ODO, DOTS FIN, DOTS ECON, ROE, are much stronger when the Core CG Index, consisting and CREDIT RISK. of 26 questions, is used. In this case, statistically Looking at the ODO specifically, which is a measure significant results at the 1 percent level are observed of overall development impact, it can be observed that for all performance variables, showing that companies companies from the top quartile achieve, on average, a with greatest observance of recommended CG practices score of 4.41 against a much smaller ODO of 3.98 of achieve better performance than those from the group the bottom quartile. This result is statistically significant with the lower level of following recommended CG at the 1 percent level. Another example comes from the practices. ROE. Companies from the top CG quartile exhibit an The difference between the groups in this case is starker. average ROE that is about 20 percent higher than the For example, companies from the top CG quartile exhibit average ROE of the bottom CG quartile (13.05 percent an average ROE of 18.56 percent, which is about three versus 10.96 percent). 21  The team also analyzed Objectives 1 and 2 using the Core CG Index instead of the full version. The results were qualitatively the same. 16 Governance and Performance in Emerging-Markets Firms Table 3.3: Mean Comparison of CG Improvement from 2011 to 2016 FULL GC INDEX Bottom CG Top CG P25 P75 Mean variation of CG Mean variation of CG p-value difference means Performance Variable Legend practices = 0% practices = +25% (P75–P25) ODO 1–6 SCORE Mean 4.24 4.36 0.345 SD (0.83) (0.72) n n=19 n=10 DOTS FIN Mean 3.09 3.72 0.081* SD (0.97) (1.16) n n=19 n=10 DOTS ECON Mean 3.30 3.94 0.068* SD (1.11) (1.02) n n=19 n=10 ROE/ROIC Mean 10.05% 8.58% 0.6119 SD (7.33%) (15.84%) n n=18 n=11 EROE/EROIC Mean 13.35% 11.53% 0.6087 SD (9.78%) (19.04%) n n=18 n=10 CREDIT RISK 1–7 Mean 3.73 3.36 0.1553 SD (0.99) (0.92) n n=19 n=11 CREDIT RISK 1–11 Mean 5.89 5.36 0.1986 SD (1.82) (1.50) n n=19 n=11 Note: The table exhibits mean-comparison tests between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% firms in improvement in the adherence to the 84-question CG questionnaire from 2011 to 2016; BOTTOM CG P25 refers to the bottom 25% in the adherence to the CG Index over this same period. Because 14 companies from our sample reported null variation in their CG score from 2011 to 2016, we have a larger number of companies in the bottom quartile group compared with the top quartile group. ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. times the average ROE of 6.91 percent from the bottom 3.2.5.  Corporate Governance and E&S CG quartile. The same difference of magnitude is found Review Rating for the economic return on equity. While firms with the highest level of CG exhibit an EROE of 24.35 percent, The team also carried out an additional analysis to CG laggards exhibit an EROE of only 9.32 percent. examine the relationship between the recommended CG Figure 3.3 illustrates these results. practices of the sample firms and their Environmental Findings 17 Table 3.4: Mean Comparison of Level of Average CG Practices FULL CG INDEX Bottom CG Top CG P25 P75 Mean adherence to CG Mean adherence to CG p-value difference Performance Variable Legend Index 2011–2016 = 42% Index 2011–2016 = 84% means (P75–P25) ODO 1–6 SCORE Mean 3.98 4.41 0.000*** SD (0.64) (0.77) n n=96 n=90 DOTS FIN Mean 3.02 3.47 0.004*** SD (1.15) (1.11) n n=96 n=90 DOTS ECON Mean 3.25 3.82 0.000*** SD (0.94) (1.15) n n=96 n=90 ROE/ROIC Mean 10.96% 13.05% 0.040** SD (8.04%) (7.90%) n n=90 n=90 EROE/EROIC Mean 15.70% 16.72% 0.258 SD (11.20%) (9.55%) n n=90 n=84 CREDIT RISK 1–7 Mean 3.86 3.15 0.000*** SD (0.76) (0.75) n n=92 n=89 CREDIT RISK 1–11 Mean 6.15 4.80 0.000*** SD (1.44) (1.42) n n=92 n=89 Note: The table exhibits mean-comparison tests between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% firms in the mean level of adherence to the 84-question CG questionnaire from 2011 to 2016; BOTTOM CG P25 refers to the bottom 25% in the level of adherence to the CG Index over this same period. ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/ EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. & Social Review Rating (ESRR). ESRR is an index score indicates a lower environmental and social risk created by IFC to measure its clients’ E&S capability associated with the company. and their management of environmental and social issues in accordance with IFC’s Sustainability Framework. It is The analysis began with the observation that there is a computed on a 1–4-point scale: excellent, satisfactory, –0.2438 pairwise correlation between the two variables. partly satisfactory, and unsatisfactory. Thus a lower Although not large in absolute terms (correlations vary 18 Governance and Performance in Emerging-Markets Firms from –1 to +1), this correlation is significant at the 1 According to Figure 3.4, companies from the top quartile percent level. The next step was a procedure analogous exhibited an average ESRR of 2.01, against a higher ESRR to the previous subgroup analyses—dividing the sample of 2.19 for the companies in the bottom quartile (this companies into four quartiles based on CG practices and difference was statistically significant at the 5 percent then comparing the average ESRR of the two extreme level). Put simply, companies from the top quartile of quartiles. The results are presented in Figure 3.4. CG practices on average exhibit a 10 percent lower ESRR than those in the bottom quartile. Figure 3.3: ROE/ROIC and EROE/EROIC and the Core CG Index Subgroup comparison: Top 25% vs. Bottom 25% in the adherence to CG practices (Core Index, 26 selected questions): ROIC/ROE and EROIC/EROE 25.0 24.35*** 20.0 18.56** 15.0 10.0 9.32 6.91 5.0 Adherence Adherence to Adherence Adherence to Core CG Core CG Core CG Core CG Index: 33% Index: 33% 0.0 ROIC/ROE EROIC/EROE Top 25% CG Bottom 25% CG Table 3.5: Core CG Index Average Score and Performance CORE CG INDEX Bottom CG P25 Top CG P75 p-value difference Core CG Index (26q) Core CG Index (26q) means Performance Variable Legend 2011–2016 = 33% 2011–2016 = 82% (P75–P25) ODO 1–6 SCORE Mean 3.95 4.88 0.000*** SD (0.70) (0.67) n n=90 n=78 DOTS FIN Mean 3.12 4.03 0.000*** SD (1.09) (0.82) n n=90 n=78 DOTS ECON Mean 3.28 4.10 0.000*** SD (0.98) (0.85) n n=90 n=78 ROE/ROIC Mean 6.91% 18.56% 0.000*** SD (13.43%) (8.37%) n n=84 n=78 Findings 19 Table 3.5: Core CG Index Average Score and Performance (Continued) CORE CG INDEX Bottom CG P25 Top CG P75 p-value difference Core CG Index (26q) Core CG Index (26q) means Performance Variable Legend 2011–2016 = 33% 2011–2016 = 82% (P75–P25) EROE/EROIC Mean 9.32% 24.35% 0.000*** SD (15.84%) (11.29%) n n=84 n=78 CREDIT RISK 1–7 Mean 3.84 3.52 0.006*** SD (0.84) (0.82) n n=93 n=77 CREDIT RISK 1–11 Mean 6.17 5.45 0.001*** SD (1.57) (1.49) n n=93 n=77 Note: The table exhibits mean-comparison tests between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% firms in the mean level of Core CG Index scores of the CG questionnaire composed of 26 questions. BOTTOM CG P25 refers to the bottom 25% of the Core CG Index. ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 3.2.6.  Splitting the Sample by Key Figure 3.4: ESRR and Average CG Scores Project Characteristics Subgroup comparison: Top Quartile vs. Bottom Quartile in CG Scores vs. ESRR Scores To conclude the subgroup comparison, the study 4.0 segregated the sample along three key characteristics of the project: 1) equity versus debt deals; 2) Tier i or Tier ii operations versus Tier iii; and 3) financial 3.0 versus nonfinancial clients. This allows an investigation into whether the relationship between 2.19 2.01 CG and performance holds for each one of these 2.0 Companies categories of companies. Companies from the from the Top Bottom CG CG Quartile Quartile Tables 3.6 and 3.7 show a top versus bottom quartile 1.0 ESRR analysis of the sample subset of equity deals versus Top 25% CG Bottom 25% CG debt deals, using the Core CG Index.22 22  This section presents the results from using the reduced version of the CG Index. Results with the full index were qualitatively the same, although weaker in statistical significance. 20 Governance and Performance in Emerging-Markets Firms Tables 3.6 and 3.7 show that companies belonging as independent groups. For debt deals (Table 3.7:), all to the top quartile for CG exhibit substantially better results were significant at the 1 percent level. For equity performance indicators than those from the bottom quartile deals, however, Table 3.6 shows that DOTS economic for both equity and debt deals. Thus the statistically performance was the only variable with coefficients that positive correlation between CG and performance holds were not statistically significant (although pointing in for both IFC equity and debt investments when considered the expected direction). The tables also indicate that, for Table 3.6: Split-Sample Key Characteristics for Equity Deals CORE CG INDEX—ONLY EQUITY DEALS Bottom CG P25 Top CG P75 Adherence to Core CG Adherence to Core CG p-value difference Index (26q) Index (26q) means Performance Variable Legend 2011–2016 = 32% 2011–2016 = 81% (P75–P25) ODO 1–6 SCORE Mean 3.58 4.80 0.000*** SD (0.28) (0.68) n n=24 n=48 DOTS FIN Mean 2.92 4.03 0.000*** SD (0.81) (0.85) n n=24 n=48 DOTS ECON Mean 3.08 3.78 0.117 SD (1.00) (0.91) n n=24 n=48 ROE/ROIC Mean –2.19% 18.84% 0.000*** SD (20.04%) (9.17%) n n=24 n=48 EROE/EROIC Mean –1.80% 24.92% 0.000*** SD (22.84%) (12.77%) n n=24 n=48 CREDIT RISK 1–7 Mean 4.11 3.75 0.072* SD (0.99) (0.89) n n=28 n=48 CREDIT RISK 1–11 Mean 6.78 5.94 0.033** SD (1.91) (1.55) n n=28 n=48 Note: The table exhibits mean-comparison tests for equity deals between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% firms in the mean level of adherence to the reduced version (Core CG Index) of the CG questionnaire, composed of 26 questions. BOTTOM CG P25 refers to the bottom 25% in the level of adherence to the Core CG index. ODO 1–6 score is the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful. DOTS FIN and DOTS ECON are the average financial and economic performance assessment of the project on a 1–4 scale ranging from unsatisfactory to excellent. ROE/ROIC is the annual return on equity or return on invested capital of the project. EROE/EROIC is the annual economic return on equity or economic return on invested capital of the project. CREDIT RISK 1–7 and CREDIT RISK 1–11 represent the credit risk rating of the project on two different scales. Appendix E details the operational definitions of all variables. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Findings 21 some performance indicators such as ROE, the difference The team carried out the same analyses by splitting the between the two extreme CG quartiles is much larger sample into Tier i or Tier ii operations versus Tier iii, for the subsample of equity deals than for debt deals. and into financial versus nonfinancial clients. The results Specifically, while the ROE difference between the top (available on request) are qualitatively the same as for the and bottom CG quartiles is about 20 percent for equity analysis of debt versus equity. In all cases, the superior deals, the same difference amounts to only 8 percent performance of companies belonging to the top quartile for debt deals. of CG practices holds for all three subgroups analyzed. Table 3.7: Split-Sample Key Characteristics for Debt Deals CORE CG INDEX—ONLY DEBT DEALS Bottom CG P25 Top CG P75 Adherence to Core CG Adherence to Core CG p-value difference Index (26q) Index (26q) means Performance Variable Legend 2011–2016 = 33% 2011–2016 = 85% (P75–P25) ODO 1–6 SCORE Mean 4.09 5.00 0.000*** SD (0.76) (0.64) n n=66 n=30 DOTS FIN Mean 3.19 4.05 0.000*** SD (1.17) (0.78) n n=66 n=30 DOTS ECON Mean 3.35 4.62 0.000*** SD (0.97) (0.34) n n=66 n=30 ROE/ROIC Mean 10.55% 18.11% 0.000*** SD (7.08%) (7.01%) n n=60 n=30 EROE/EROIC Mean 13.77% 23.44% 0.000*** SD (8.92%) (8.54%) n n=60 n=30 CREDIT RISK 1–7 Mean 3.72 3.14 0.000*** SD (0.74) (0.52) n n=65 n=29 CREDIT RISK 1–11 Mean 5.91 4.65 0.000*** SD (1.33) (0.97) n n=65 n=29 Note: The table exhibits mean-comparison tests for debt deals between selected variables (two-sample t tests with unequal variances) of two groups: TOP CG P75 refers to the top 25% firms in the mean level of adherence to the reduced version (Core CG Index) of the CG questionnaire, composed of 26 questions. BOTTOM CG P25 refers to the bottom 25% in the level of adherence to the Core CG index. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Appendix E details the operational definitions of all variables. 22 Governance and Performance in Emerging-Markets Firms 3.3. PRIORITIZING SPECIFIC CG all measures of performance), an even stronger focus INDICATORS on proper control environment practices is made. The analysis of top-20 CG indicators done separately for The second method of investigation consists of identifying financial institutions and real sector companies shows which specific practices from the CG Index are the significant overlap between the two, which makes it strongest predictors of performance variables. This is possible to generalize the top-20 list above for the done through a data reduction analysis.23 study sample. The first step was to run separate OLS (ordinary The team also carried out an alternative analysis least squares) regressions for each question on each (available on request) that focused on the companies performance variable, then to store the absolute value that exhibited the best financial performance. It began of all t-statistics—the abs(t). The third step consisted of with identifying the 10 firms with the highest average ranking the abs(t) of each question for each performance return on equity throughout the investment period. 26 indicator (for n performance variables, therefore, the It was apparent that these companies exhibited an question is ranked n times). 24 The final step was to average ROE of 25.1 percent during this period, about sort all questions based on the mean t-stat ranking 2.5 times the average ROE of 10.8 percent of the full obtained for all performance variables. 25 Table 3.8 sample. The next step was to compute the average level lists the 20 highest-ranked practices of the CG Index. of compliance of these 10 top financial performers with each question of the CG Index and rank all As Table 3.8 indicates, 12 out of the 20 top-ranked questions by the average level. This procedure led questions belong to the dimension of Control Environment to identification of the CG practices commonplace and Processes. Many of these practices are aimed at among all firms with best financial performance of safeguarding companies’ assets, which has a direct the sample. impact on a company’s performance. These key CG practices include having a proper and independent To conclude, it is important to highlight an important internal audit function and compliance function, and limitation of the analyses carried out in this section. establishing board-approved risk management policies. To Because the sample is likely biased, as pointed out in successfully implement these practices, it is critical to have Box 2.2, the results are not generalizable to the whole a properly established and functioning board of directors, population of IFC investments. This means that top-20 often through its audit committee, which oversees the CG practices identified as the strongest predictors of management and ensures that proper compliance, risk performance in this section would not necessarily be management, and internal audit functions are in place. the same if the same exercise were done with a different Moreover, when the list of top-ranked CG practices sample of companies. So to identify the core CG practices is reviewed against credit risk rating (as opposed to that IFC should prioritize during its investment appraisal, 23  The two most common data reduction techniques are principal component analysis (PCA) and factor analysis (FA). PCA is a linear combination of variables; FA is a measurement model of a latent variable. The first approach to data reduction creates one or more index variables from a larger set of measured indicators through a linear combination of a set of variables. The second approach is a model of the measurement of a latent variable. This latent variable cannot be directly measured with a single variable. Instead, it is seen through the relationships it causes in a set of Y variables. 24  For example, if there are five performance variables, each question is ranked five times. The rank for a certain performance variable goes from 1—the lowest abs(t) obtained for this variable—up to the highest ranking, which equals the number of questions answered for this performance variable. 25  This investigation started with Stata’s principal-component factor (PCF) analysis. However, due to the small number of observations and that just 9 companies responded to all 84 questions, this technique did not generate meaningful results. The team then carried out this described alternative procedure to rank all questions based on their relevance as a predictive factor of firm performance. 26  The top 10 firms in the average ROE from 2011 to 2016 are Nirdhan, BSP, AgBank, Transilvaniabank, Banco BHD, ACLEDA Bank, MTBank, Titan Danube, Bakhresa Rwanda, and Banco General. Findings 23 a specific study would be necessary with a larger and characteristics, such as family versus nonfamily businesses, randomly selected sample. As a suggestion, this future listed versus nonlisted companies, and equity versus study could also identify the core CG practices to be debt deals. prioritized for companies or projects with different Table 3.8: Top 20 CG Practices—Strongest Predictors of Performance Variables Rank Question Description CG Dimension 1 Q49 The internal audit has its own charter or specific terms of reference. Control Environment 2 Q40 The company follows internationally recognized standards on internal Control controls. Environment 3 Q50 Financial statements are audited by a recognized independent auditing firm. Control Environment 4 Q2 The company has a written code of conduct. Commitment to CG 5 Q39 There is one person formally responsible for the compliance initiatives of the Control company (e.g., compliance officer). Environment 6 Q45 There is one person formally responsible for the risk management initiatives Control of the company (e.g., CRO or risk manager). Environment 7 Q48 There is an internal audit function in place. Control Environment 8 Q26 The board conducts self-evaluations or other reviews of its effectiveness on Board of Directors an annual basis. 9 Q25 The board has a formal remuneration policy for board members, taking into Board of Directors account their membership of committees. 10 Q24 The company has a written policy establishing rules for the approval of Board of Directors related-party transactions (RPTs). 11 Q44 The board oversees the implementation of the risk management policies. Control Environment 12 Q3 Board members receive periodic training on CG issues, funded by the Commitment to company. CG 13 Q23 The company has a corporate secretary. Board of Directors 14 Q46 The internal audit unit has an audit work plan that is approved by the audit Control committee or by the board every year. Environment 15 Q47 The internal audit reports directly to the audit committee or to the board of Control directors. Environment 16 Q36 The audit committee oversees the implementation of the internal and Control external auditors’ recommendations. Environment 17 Q19 The board has an audit committee in place. Board of Directors 18 Q21 The board has regular meetings 6 to 8 times per year (8 to 12 for financial Board of Directors institutions). 19 Q43 The board approves the company’s risk management policies. Control Environment 20 Q42 The company has a specific whistleblower channel that ensures anonymity Control for informers and due treatment of the complaints. Environment 24 Governance and Performance in Emerging-Markets Firms Conclusions and Improvement Opportunities This empirical study on the link between CG and performance of IFC portfolio companies in emerging markets shows that CG not only is an important risk element but also has a positive association with client performance, as evidenced by lower investment risk and higher financial performance. Specifically, for IFC’s new business, the study’s conclusion is that better screening of potential IFC clients’ CG practices is correlated with lower credit risk during the investment period and thus associated with a lower probability of credit default. The ex ante CG analysis is critical for selecting the right clients and could be better integrated with other elements of investment analysis, including credit and valuation. It also would be beneficial, where relevant, to define criteria and analysis to incorporate CG risk factors into the assessment of credit risk. For IFC’s portfolio operations, the study shows that benefit substantially from CG improvements, including investing time and resources in monitoring the clients’ relying on IFC’s global experience and expertise in CG CG performance and helping them improve on CG is advisory services. associated with better financial and economic performance. This is where IFC’s pressure on and/or support to the Another key conclusion of this study is that, although clients may lead to a win-win outcome. IFC’s CG advisory the concept of CG is broad, it is still possible to have services therefore would be an important tool for value a practical and prioritized approach to it. CG spans a creation. Another relevant conclusion is that, due to the range of elements, including 1) the organization and positive correlation between CG and ESRRs and to IFC’s functioning of the board; 2) the daily operations of internal barriers for repeat investments in companies management to ensure that the company’s assets are with high ESRR scores, CG improvements can be a safeguarded through robust systems of internal control, possible tool to reduce the E&S risk and have more risk management, compliance, and internal audit; 3) the clients eligible for additional investments. proper, relevant, and timely disclosure of financial and nonfinancial information; and 4) mechanisms for The study’s findings also indicate that the positive protecting the rights of all shareholders. Yet there is correlation between CG and performance takes place no one size that fits all companies. This diverse nature for all types of portfolio clients, irrespective of the of CG should be properly captured in risk analysis. size of the investment, type of the investment product, And it is from such risk perspective that it is possible listing, industry, and other parameters. This shows that and recommended to focus on and prioritize certain a minimum level of CG risk assessment and continuous core CG indicators that have a higher chance of having monitoring of CG is probably an important factor for an impact on performance. Such core CG indicators, IFC’s business. On top of this, IFC clients (perhaps possibly in the form of corporate governance KPIs, specifically in equity transactions)  that understand should be analyzed and monitored in every single IFC CG, recognize its value, and are committed to it can transaction. Conclusions and Improvement Opportunities 25 The findings of the study show not only that CG can be an important risk management and value-enhancing instrument for investors, but also that companies themselves can see a more clearly articulated business case for good governance. More and more investors, equity and debt, incorporate better CG elements into their investment propositions and due diligence processes. Equally important, companies’ owners, boards of directors, and senior managers should recognize and take advantage of improvements in CG that would help them make better business decisions as well as improve their performance—financial and nonfinancial—to the benefit of shareholders, stakeholders, and the economy in general. This study should not be viewed as one-off and final. More can be done, perhaps with a focus on causality and price quantification of good CG. Certain enhancements of IFC’s internal due diligence and portfolio supervision, as well as its internal systems capturing and analyzing the relevant data, could be beneficial. Specifically, IFC could seek to do the following: • Standardize CG risk assessment at deal origination (for example, use corporate governance KPIs); • Improve the internal systems to collect high-quality data and feedback from clients; • Define better specific CG performance indicators and tailor them to specific sectors/types of entities; • Enhance the business case narrative for CG, internally and for clients; and • Explore ways to structure financial products by integrating CG aspects. Finally, the results and findings of this study are in line with external academic research on the topic, indicating that CG tends to be particularly relevant for emerging markets and closely held unlisted companies. It thus contributes to the body of research on this topic. 26 Governance and Performance in Emerging-Markets Firms References Ammann, M., D. Oesch, and M. Schmid. 2011. Corporate governance and firm value: International evidence. Journal of Empirical Finance 18: 35–55. Bebchuk, L., A. Cohen, and C. Wang. 2013. Learning and the disappearing association between governance and returns. Journal of Financial Economics 108 (2): 323–48. Bhagat, S., and B. Bolton. 2008. Corporate governance and firm performance. Journal of Corporate Finance 14 (3): 257–73. Bhagat, S., B. Bolton, and R. Romano. 2008. The promise and peril of corporate governance indices. Columbia Law Review 108 (8): 1803–82. Black, B., A. G. De Carvalho, V. Khanna, W. Kim, and B. Yurtoglu. 2014. Methods for multicountry studies of corporate governance: Evidence from the BRIKT countries. Journal of Econometrics 183 (2): 230–40. Black, B., H. Jang, and W. Kim. 2006. Does corporate governance predict firms’ market values? Evidence from Korea. Journal of Law, Economics, and Organization 22 (2): 366–413. Börsch-Supan, A., and J. Köke. 2002. An applied econometrician’s view of empirical corporate governance studies. German Economic Review 3 (3): 295–326. Bozec, R., and Y. Bozec. 2012. The use of governance indexes in the governance-performance relationship literature: International evidence. Canadian Journal of Administrative Sciences 29: 79–98. Brown, L., and M. Caylor. 2006. Corporate governance and firm performance. Journal of Accounting and Public Policy 25 (4): 409–34. Chen, K., Z. Chen, and J. Wei. 2009. Legal protection of investors, corporate governance, and the cost of equity capital. Journal of Corporate Finance 15: 273–89. Claessens, S., and B. B. Yurtoglu. 2013. Corporate governance in emerging markets: A survey. Emerging Markets Review 15: 1–33. CLSA Group. 2010. Corporate Governance Watch 2010. Available at https://www.clsa.com/pdf.cfm?link=/assets/ files/reports/CLSA-CG-Watch-2010.pdf. Desai, R. M., H. Kharas, and M. Amin. 2017. Combining good business and good development. Brookings Global Working Paper Series. Available at https://www.brookings.edu/wp-content/uploads/2017/05/global_20170509_ good-business-good-development.pdf. References 27 Diavatopoulos, D., and A. Fodor. 2016. Does corporate governance matter for equity returns? Journal of Accounting and Finance 16 (5): 39. Durnev, A., and H. Kim. 2005. To steal or not to steal: Firm attributes, legal environment, and valuation. Journal of Finance 60 (3): 1461–93. Gompers, P. A., J. L. Ishii, and A. Metrick. 2003. Corporate governance and equity prices. Quarterly Journal of Economics 118 (1): 107–55. Hitz, J., and N. Lehmann. 2013. The usefulness of corporate governance ratings—Insights from European settings. Working paper available at http://ssrn.com/abstract=2042019. Klapper, L., and I. Love. 2002. Corporate governance, investor protection, and performance in emerging markets. World Bank Policy Research Working Paper 2818 (April). Leal, R. P. C., and A. Carvalhal Da Silva. 2007. Corporate governance and value in Brazil (and in Chile). Inter- American Development Bank-Research Network Working Paper R-514; in Alberto Chong and Florencio Lopez- de-Silanes. Investor Protection and Corporate Governance—Firm Level Evidence Across Latin America. Palo Alto, California: Stanford University Press. 213–87. Available at http://www.iadb.org/res/pub_desc.cfm?pub_id=R-514. Nguyen, B. D., and K. M. Nielsen. 2010. The value of independent directors: Evidence from sudden deaths. Journal of Financial Economics 98 (3): 550–67. Sekaran, U., and R. Bougie. 2010. Research Methods for Business: A Skill Building Approach. 5th ed. Hoboken, New Jersey: Wiley (488 p. ISBN 978–0470744796). Silveira, A. M., and L. A. Barros. 2007. Corporate governance quality and firm value in Brazil. Working paper available at SSRN http://ssrn.com/abstract=923310. Silveira, A. M., and A. Dias. 2010. What is the impact of bad governance practices in a concentrated ownership environment? International Journal of Disclosure and Governance 7 (1): 70–91. Renders, A., A. Gaeremynck, and P. Sercu. 2010. Corporate-governance ratings and company performance: A cross-European study. Corporate Governance: An International Review 18 (2): 87–106. Wintoki, M., J. Linck, and J. Netter. 2012. Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics 105 (3): 581–606. 28 Governance and Performance in Emerging-Markets Firms APPENDIX A Descriptive Analysis of CG Data This appendix contains descriptive analysis of the questionnaire-based CG data collected for this study. This exercise allowed for comparisons of the level of adherence of the sample firms to recommended CG practices across industries, regions, governance dimensions, and so on. This comparison is available for the date of disbursement and as of the end of 2016, and it reviews the evolution of the CG changes during the study period. Figures A.1–A.8 show the CG scores of different subgroups of companies from the sample and at the two different time points. Figure A.1: Overall Level of Following Recommended CG Practices at Disbursement 80.0% 70.0% 66.8% 61.9% 63.5% 59.9% 60.0% 56.9% 51.2% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0 Full CG Commitm nt Structur & Control Env. & Tr nsp r nc & Sh r hold rs Ind x to CG Func. Bo rd Proc ss s Disclosur Ri hts 30 Governance and Performance in Emerging-Markets Firms Figure A.2: CG Practices: FIG versus Non-FIG at Disbursement CG Scores b Industr : FIG vs. non-FIG 90% 80% 77% 70% 71% 70% 67% 64% 63% 59% 60% 55% 53% 54% 51% 50% 39% 40% 30% 20% 10% 0 FIG (Fin nci l Institutions) Non-FIG (Oth r comp ni s) Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts Figure A.3: CG Practices: by Investment Product at Disbursement CG Scores: Equit vs. Debt t the time of Disbursement 70% 63% 64% 64% 64% 62% 62% 60% 59% 59% 58% 53% 50% 49% 50% 40% 30% 20% 10% 0 Equit D bt Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts Appendix A: Descriptive Analysis of CG Data 31 Figure A.4: CG Practices: Listed versus Unlisted at Disbursement CG Scores: Listed vs. Unlisted t the time of Disbursement 80% 73% 74% 69% 70% 64% 64% 60% 62% 57% 58% 59% 60% 50% 49% 46% 40% 30% 20% 10% 0 List d Unlist d Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts Figure A.5: CG Practice: by Investment Tier at Disbursement CG Scores: Tier iii vs. Tier i-ii oper tions t the time of Disbursement 80% 73% 70% 68% 65% 60% 60% 60% 62% 62% 60% 60% 55% 50% 50% 44% 40% 30% 20% 10% 0 Ti r iii Ti r i–ii Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts 32 Governance and Performance in Emerging-Markets Firms Figure A.6: Overall Level CG Practices (JUNE 2016) 80.0% 71.1% 73.4% 69.6% 70.0% 67.1% 62.6% 58.9% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0 Full CG Commitm nt Structur & Control Env. & Tr nsp r nc & Sh r hold rs Ind x to CG Func. Bo rd Proc ss s Disclosur Ri hts Figure A.7: CG Practices: by Industry (June 2016) Level of Followin CG Recommended Pr ctices per Industr 90% 85% 80% 79% 75% 76% 76% 76% 70% 68% 69% 64% 64% 63% 60% 61% 60% 60% 58% 58% 57% 53% 55% 54% 55% 52% 50% 48% 40% 38% 30% 20% 10% 0 CFG / FIG CMG / MAS CTT / CTT CNG / INR Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts Appendix A: Descriptive Analysis of CG Data 33 Figure A.8: CG Practices: FIG versus Non-FIG (end 2016) CG Scores b Industr : FIG vs. non-FIG 90% 85% 79% 80% 75% 76% 76% 71% 70% 60% 61% 60% 60% 58% 58% 50% 50% 40% 30% 20% 10% 0 FIG (Fin nci l Institutions) Non-FIG (Oth r comp ni s) Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts Figure A.9 compares the level of following recommended CG practices by all firms by region: Latin America and the Caribbean (LAC); Europe and Central Asia (ECA); East Asia and the Pacific (EAP); Sub-Saharan Africa (CAF); South Asia (SA); and Middle East and North Africa (MENA). Note that companies from South Asia exhibit the highest level of CG (76 percent), followed by the MENA firms (74 percent). On the other end, companies from Sub-Saharan Africa exhibit the lowest CG scores (62 percent). Figure A.9: CG Practices: by Region (end 2016) Level of Followin CG Recommended Pr ctices per Re ion 100% 90% 89% 86% 83%84% 80% 78% 80% 76% 75% 72% 74% 70% 71% 71% 70% 67% 68% 67% 65% 67% 67% 65% 67% 67% 65% 68% 62% 63% 62% 61% 62% 64% 63% 62% 60% 57% 52% 54% 50% 44% 40% 30% 20% 10% 0 LAC ECA EAP CAF SA MENA Full CG Ind x Commitm nt to CG Structur & Func. Bo rd Control Env. & Proc ss s Tr nsp r nc & Disclosur Sh r hold rs Ri hts 34 Governance and Performance in Emerging-Markets Firms Figure A.10 lists the 10 questions of the CG questionnaire with the highest level of adherence at the end of 2016 (only questions applicable to all companies were considered). Note that all companies of the sample had their financial statements audited by a recognized independent auditing firm. In addition, 94 percent were also compliant with three other recommended practices: boards receiving periodic reports from management about the implementation of the strategic plan; formal approval by the board of directors of the strategic and business plans of the company; and boards carrying out a dedicated meeting or session to discuss strategy at least once a year. Figure A.10: The 10 Most Common CG Practices (June 2016) 100.0% 100.0% 94.2% 94.2% 94.2% 93.2% 92.2% 91.2% 90.4% 89.2% 75.0% 50.0% 25.0% 0.0% Ar fin nci l Do s th Do s th Do s th Ar th Ar Do s th Is th r t Do dir ctors st t m nts bo rd r c iv bo rd bo rd c rr comp n ’s sh r hold rs comp n l st on r ul rl udit d b p riodic form ll out fin nci l provid d discuss ll m mb r of r c iv ll r co ni d r ports from pprov th d dic t d st t m nt with ll m t ri l th udit docum nt - ind p nd nt th str t ic m tin or udit d m t ri l tr ns ctions committ tion of bo rd uditin firm? m n - nd busin ss s ssion to b s d on inform tion with h s r l v nt m tin s m nt bout pl n of th discuss Int rn tion l nd d t il d ffili t s of udit nd with t l st th comp n ? str t t St nd rds nd for controll rs or ccountin 5 d s in impl m nt - l st onc on Auditin ? sh r’ m t in its fin. xp rtis ? dv nc ? tion of th r? m tin s? st t m nts? str t ic pl n? Appendix A: Descriptive Analysis of CG Data 35 Figure A.11 lists the 10 least observed CG practices at the end of 2016 (only questions applicable to all companies were considered). The least observed CG practices in the sample were having board members trained on CG issues funded by the company (40 percent) and carrying out self-evaluations by the board on an annual basis (50 percent). Figure A.11: The 10 Least Observed CG Practices (June 2016) 100.0% 75.0% 67.8% 66.7% 60.0% 59.5% 56.8% 55.8% 52.8% 50.0% 40.0% 40.5% 25.0% 0.0% Do s th bo rd Do s th Do s th If th udit Do s th Do s th bo rd Is th r Do s th Do s th bo rd includ bo rd h v n comp n h v firm provid s comp n h v includ wom n s ction on th bo rd h v ind p nd nt udit d si n t d non- udit succ ssion dir ctors? firm's w bsit conducts r mun r tion dir ctors? committ in offic r or s rvic s, pl n th d dic t d to s lf- v lu - polic pl c ? ov rn nc r non- udit CEO? corpor t tions nd ccountin for bod f s limit d ov rn nc ? oth r r vi w of dir ctors' r sponsibl for to 25% of it ff ctiv - tt nd nc nd ov rs in CG tot l udit n ss on n m mb rship of polici s nd or f s? nnu l b sis? committ s? pr ctic s? While the previous figures showed the level of following recommended CG practices at the end of 2016, Figures A.12–A.18 assess the improvement on the level of adherence between 2011 and 2016, by dimension.27 Figure A.12 shows the variation on the level of adherence to the CG Index of all firms from 2011 to 2016. The first column indicates that, on average, companies in the sample improved their level of adherence to the Full CG questionnaire by 7.1 percent over this period. The following columns show the level of improvement by dimension. Specifically, the data show that firms exhibited a stronger advancement (11.3 percent) in their conformity with the dimension “Commitment to CG” and almost no progress (1 percent) in the dimension “Shareholder Rights.” 27  This is measured by computing the level of adherence to the CG questionnaire in 2016 minus the level of adherence in 2011. 36 Governance and Performance in Emerging-Markets Firms Figure A.12: Variation in CG Scores (2011 to 2016) 12.0% 11.3% 10.0% 8.0% 7.9% 8.0% 7.1% 6.3% 6.0% 4.0% 2.0% 1.0% 0.0% V ri tion Full V ri tion V ri tion Struc. V ri tion Control V ri tion Tr nsp. V ri tion Sh r. CG ind x Commitm nt to CG & Funct. Bo rd & Proc. & Discl. Ri hts Figure A.13 shows the variation in the level of following recommended practices of all firms between 2011 and 2016, by industry. Note that MAS companies exhibited the largest overall improvement in CG (9.4 percent), while CTT firms exhibited the lowest level of improvement (3.3 percent) over this period. Figure A.14 shows variation in scores by industry, FIG versus non-FIG. Figure A.13: Variation of CG Scores by Industry (2011 to 2016) 20.0% 16.7% 15.0% 13.5% 13.0% 10.8% 10.6%10.0% 10.0% 9.5% 9.4% 9.1% 8.6% 7.0% 7.5% 6.3% 6.3% 5.0% 4.0% 3.6% 4.6% 3.3% 2.4% 1.9% 2.4% 2.4% 0.0% 0.0% –5.0% –8.3% –10.0% CFG / FIG CMG / MAS CTT / CTT CNG / INR V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts Appendix A: Descriptive Analysis of CG Data 37 Figure A.14: Variation of CG Scores, FIG versus Non-FIG (2011 to 2016) 14% 12% 12% 11% 10% 9% 9% 8% 8% 7% 7% 7% 7% 6% 4% 4% 2% 2% 0% 0% -2% FIG (Fin nci l Institutions) Non-FIG (Oth r comp ni s) V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts Figures A.15–A.18 show variations in following recommended CG practices—by region, investment product, listing status, and investment tier. Figure A.15: Variation of CG Practices by Region (2011 to 2016) 20.0% 18.8% 17.8% 18.0% 15.7% 15.7% 16.0% 14.1% 13.7% 14.0% 12.0% 10.9% 11.2% 11.5% 10.5% 10.2% 10.0% 9.4% 9.3% 8.6% 8.0% 7.7% 7.3% 7.2% 6.9% 6.3% 6.7% 6.0% 6.0% 5.3% 5.1% 4.5% 4.7% 4.0% 3.1% 3.2% 3.4% 3.1% 2.5% 2.4% 2.0% 1.7% 1.0% 0.6% 0.0% 0.0% 0.0% LAC ECA EAP CAF SA MENA V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts 38 Governance and Performance in Emerging-Markets Firms Figure A.16: CG Practices, by Investment Product (2011 to 2016) V ri tion of CG Scores (2011–2016): Equit vs. Debt 16% 14% 14% 12% 11% 10% 10% 10% 9% 8% 8% 8% 8% 7% 7% 6% 4% 2% 2% 1% 0% Equit D bt V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts Figure A.17: Variation of CG Practices, Listed versus Unlisted companies (2011 to 2016) V ri tion of CG Scores (2011–2016): Listed vs. Unlisted 14% 12% 12% 10% 9% 9% 8% 7% 8% 8% 7% 7% 6% 6% 5% 4% 2% 2% 0% 0% –2% List d Unlist d V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts Appendix A: Descriptive Analysis of CG Data 39 Figure A.18: Variation of CG Practices by Investment Tier (2011 to 2016) V ri tion of CG Scores (2011–2016): Tier iii vs. Tier i–ii 14% 13% 12% 10% 9% 9% 8% 8% 8% 6% 6% 5% 4% 3% 2% 2% 1% 0% –1% –1% –2% –4% Ti r iii Ti r i-ii V ri tion Full CG ind x V ri tion Commitm nt to CG V ri tion Struc .& Funct. Bo rd V ri tion Control & Proc. V ri tion Tr nsp. & Discl. V ri tion Sh r. Ri hts 40 Governance and Performance in Emerging-Markets Firms APPENDIX B Multiple Regressions The last quantitative exercise involves multiple regressions. To minimize the risk of omitted variables biasing the Because this procedure allows filtering out the effect of results, the conceptual model also contains other potential control variables on the relationship between CG and explanatory factors—the so-called control variables—that performance, it is usually the most robust procedure to may have an impact on both CG and firm performance. assess how CG influences performance from a causal The following set of control variables are based on viewpoint. However, relevant limitations in the database— previous studies of the CG-performance literature: which consists of only about 60 firms—prevented this analysis from yielding conclusive results. • Financial leverage: total liabilities/total assets; • Asset tangibility: fixed assets (property, plant, and Regardless, this appendix reports the attempt to carry out equipment)/operational revenue; multiple regressions to test the study’s third hypothesis, which argues that the mean level of adherence to the • Ownership concentration: ownership by the largest CG Index throughout the investment period is positively shareholder (percentage of voting shares held by the associated with firm performance. To the extent that controlling shareholder); the team manages to build a larger and more robust • Industry: IFC industry classification: Financial CG database, it will be possible to implement more Institutions Group (FIG); Manufacturing, Agriculture sophisticated procedures to test the CG-performance and Services (MAS); Telecom, Media, Technology relationship from an econometric perspective in the & Venture Investing (CTT); and Infrastructure and future. Natural Resources (INR); The first step in carrying out multiple regressions is • Listing on stock exchange: dummy variable assuming to establish a conceptual model that defines CG as a value of one, if the company is listed on the stock the key explanatory variable of interest and firm-level exchange; zero otherwise. performance as the dependent variable.28 28  The dependent variable is the one affected by the others from the conceptual model. It is the variable of primary interest, because the goal is to explain its variability by finding what variables influence it. The independent variable is the presumed causal factor that influences the dependent variable in either a positive or a negative way. Independent variables should meet four conditions to allow for correct statistical inferences: 1) the independent and the dependent variable should covary (a change in one should be associated with a change in the other); 2) the independent variable should precede the dependent variable (there must be a time sequence in which the two occur); 3) no other factor should be a possible cause of the change in the dependent variable (the effects of all other variables should be controlled for); 4) a logical explanation or theory is needed for why the independent variable affects the dependent variable. Appendix B: Multiple Regressions 41 Figure B.1: Hypothetical Effect of Firm Size as a Moderating Variable Eff ct for sm ll firms Firm P rform nc Firm P rform nc Eff ct for ll firms Eff ct for l r firms Corpor t Gov rn nc Qu lit Corpor t Gov rn nc Qu lit The influence of CG on performance may also vary • Size of IFC disbursement: a dummy variable significantly depending on additional elements. One categorizing firms into two groups based on the tier example comes from firm size: it is possible that a higher that the investment operation belongs to (Tier i and level of adherence to recommended CG practices is Tier ii versus Tier iii deals); more relevant for the performance of smaller companies • Type of IFC investment: a dummy variable categorizing than for the performance of large ones (or vice versa). firms into two groups based on the type of IFC’s investment (equity versus debt); If this is the case, then firm size should be used as a moderating variable in the relationship between • Role of IFC on the board: a dummy variable CG and performance.29 Figure B.1 provides a visual categorizing firms into two groups based on IFC’s representation of how the relationship between CG appointment of a nominee director (with versus and firm performance could be influenced by firm size. without the appointment of a nominee director); • Type of IFC involvement—advisory: a dummy variable As a result, this conceptual model also takes into account categorizing firms into two groups based on IFC’s the following set of potential moderating variables that provision of advisory services (with versus without may have a contingent effect on the relationship between the provision of advisory services). CG and performance: • Firm size: a dummy variable categorizing firms into two groups (large versus small) if the they are above the P67 or below the P33 of the sample in terms of its total assets; 29  29 A moderating variable is one that has a strong contingent effect on the relationship between the independent variable and the dependent variable. As a result, its presence modifies the original relationship between the independent and the dependent variables. Moderating variables are also known as “it depends” variables. For instance, someone who is asked if CG adds value might respond, “It depends. For smaller firms, it creates a lot of value. For big firms, not that much.” 42 Governance and Performance in Emerging-Markets Firms Figure B.2: Conceptual Model S t of control v ri bl s: Firm ; Fin nci l l v r ; Ass t t n ibilit ; CG Dim nsions Own. conc ntr tion; Id ntit controllin sh r hold r; Fin. indic tors Industr ; List d or not; Busin ss roup m mb rship. Commitm nt to ROE/ROIC Corpor t Gov rn nc DOTS Fin Control v ri bles Structur nd Functionin EROE/EROIC of th Bo rd DOTS Econ Perform nce Adherence to (Fin nci l & Control Environm nt nd CG Pr ctices Non fin. indic tors Proc ss s Non-Fin nci l) Ind p nd nt D p nd nt ODO 1–6 Scor V ri bl V ri bl Tr nsp r nc nd Moder tin Disclosur v ri bles Cr dit/Equit Risk Sh r hold rs ri hts S t of mod r tin v ri bl s:* Firm si : l r vs. sm ll Si of IFC disburs m nt: Ti r 1, 2 or 3 Rol of IFC on bo rd: with or without nomin dir ctor T p of IFC involv m nt: with or without dvisor s rvic s T p of IFC inv stm nt: quit or d bt Histor of r l tionship with IFC: n w cli nt vs. old cli nt Figure B.2 illustrates the full conceptual model, including the key variables of interest, control variables, and moderating variables. As stated before, the small number of observations required the study to restrict multiple regression analysis to the third hypothesis, which argues that “the mean level of adherence of IFC clients to recommended CG practices is associated with better firm performance.” The team tested this hypothesis in the following way: • For each firm, the study measured the mean level of adherence to recommended CG practices by computing the average of CG scores at both t=0 and t=n: CGit = (CGi,t=n + CGi,t=0)/2. • For each firm, the study computed an aggregate or average measure of performance from t=0 to t=n, such as, for instance, the average return on equity (ROE) or the overall level of development outcome (ODO) assigned to the project (PERFi). • For each firm, the study computed control and moderating variables for t=0, t=1,…, t=n (CVit and MVit). • The team then carried out multiple regressions on a cross-section database using the following general research model: n PERFit = α + β1 ×CG it ∗ MV it + β 3 × MV it + ∑ β j ×CV jit + n j + u it j =1 Appendix B: Multiple Regressions 43 Where: Table B.1 shows mixed findings in the relationship between the level of adherence to IFC’s Full CG Index PERFit = performance of the ith firm at time t; and performance variables. On the one hand, a higher level of adherence is associated with a lower credit risk CGit = average level of adherence to CG practices of rating (CREDIT RISK 1–11), which is a key measure the ith firm at time t; of the financial risk of the project. On the other hand, a higher CG score is also negatively related to both MVi = set of moderating variables of the ith firm at time return on equity (ROE) and economic return on equity t to be tested alternatively; (EROE), thus contradicting the study hypothesis. Also CVji = set of control variables of the ith firm at time t; observable is a non-statistically significant relationship between the adherence to the Full CG Index and the ni = firm-specific and time-invariant effect of the ith overall development outcome rating of the project (ODO firm (nonobservable fixed effect); and 1–6 SCORE). uit = random error term of the ith firm at time t. In addition, some statistically significant relationships between other explanatory variables and firm performance Table B.1 presents the results of this attempt to carry out are worth noting. Equity deals are positively related to multiple regressions. It exhibits the outcomes of different the EROE of the project. On the other hand, they are regression models relating the level of adherence to the full negatively related to ODO scores as well as associated version of the CG Index (ADHERENCE_FULL_CG) to with a higher credit risk rating. The provision of advisory four performance variables: ODO 1–6 (models 1 and 2), services is positively related to ODO scores and not the overall development outcome rating of the project significantly related to other performance indicators. on a 1–6-point scale ranging from highly unsuccessful But the appointment of a nominee director generates to highly successful; ROE/ROIC (models 3 and 4), the unexpected results: it is negatively related to both ROE annual return on equity or return on invested capital and EROE, and it is associated with a higher credit risk. of the project; EROE/EROIC (models 5 and 6), the Tier iii operations are associated with lower ROE and annual economic return on equity or economic return EROE. Listing on the stock exchange is positively related on invested capital of the project; and CREDIT RISK to financial indicators as well as associated with a lower 1–11 (models 7 and 8), the credit risk rating of the credit risk. Firm age and financial leverage are positively project on a scale from 1 to 11. related to ODO, ROE, and EROE, while the concentration of voting shares by the controlling shareholder has a For each performance variable, the team ran regressions detrimental effect on all performance variables. with both a basic model and a full model. The basic model controls the relationship between CG and performance only for firm size, firm age, and financial leverage; the full model controls for a full set of variables, including characteristics of the operation and industry dummies. 44 Governance and Performance in Emerging-Markets Firms Table B.1: Full CG Index and Performance Variables ODO 1–6 Score ROE/ROIC EROE/EROIC Credit risk 1–11 Basic Basic Basic Basic Dependent Variables Model Full Model Model Full Model Model Full Model Model Full Model Model (1) (2) (3) (4) (5) (6) (7) (8) Adherence_full_cg 0.377 0.117 –1.757 –6.139** –4.273 –9.501*** –2.503*** –1.677*** (1.57) (0.49) (–0.59) (–2.21) (–0.99) (–2.71) (–4.44) (–2.64) Equity operation – –0.355** – 0.328 – 2.665** – 1.136*** (–3.11) (0.39) (2.22) (4.79) Advisory services – 0.900*** – 0.208 – 1.965 – –0.201 (5.14) (0.12) (0.72) (–0.48) Nominee director – –1.757 – –2.847** – –2.915* – 0.809*** (–0.59) (–2.57) (–1.72) (2.76) Tier iii operation – –1.757 – –4.481*** – –9.881*** – 0.215 (–4.03) (–6.68) (0.71) (–0.92) Listed company – 0.168 – 2.135* – 3.725** – –0.450* (1.60) (1.78) (2.22) (–1.83) Firm size 0.024 –0.024 0.279 0.240 0.299 –0.271 –0.270*** –0.309*** (0.93) (–0.86) (0.86) (0.62) (0.60) (–0.50) (4.73) (–4.48) Firm age 0.097 0.429*** 3.754*** 5.937*** 4.419*** 8.916*** 0.087 0.111 (1.50) (4.86) (6.59) (6.20) (6.14) (6.87) (0.65) (0.73) Financial_leverage 0.578*** 0.449*** 7.935*** 6.382*** 9.429*** 6.620*** 0.150 0.591 (4.34) (2.87) (5.27) (3.53) (4.67) (3.06) (0.41) (1.29) Conc_shares – –0.013*** – –0.075*** – –0.108*** – 0.009** (–6.28) (–3.53) (–4.06) (2.10) Industry dummies No Yes No Yes No Yes No Yes Constant 3.022*** 3.979*** –7.382*** –4.520 –5.339 –6.408 9.914*** 8.735*** (10.44) (8.84) (–3.36) (–0.94) (–1.19) (–1.27) (15.16) (9.62) Number of 264 213 259 208 248 197 259 208 observations F statistic 8.70 22.60 24.47 31.76 20.29 32.00 15.26 16.25 R-squared 0.078 0.456 0.273 0.549 0.223 0.563 0.196 0.444 Note: This table exhibits the outcomes of different regression models aiming at analyzing the relationship between the level of adherence to IFC’s CG Index and firm-level performance. The dependent variables are ODO 1–6 (models 1 and 2), the overall development outcome rating of the project on a 1–6-point scale, ranging from highly unsuccessful to highly successful; ROE/ROIC (models 3 and 4), the annual return on equity or return on invested capital of the project; EROE/EROIC (models 5 and 6), the annual economic return on equity or economic return on invested capital of the project; and, CREDIT RISK 1–11 (models 7 and 8), the credit risk rating of the project on a scale from 1 to 11. The independent variable of interest is ADHERENCE_FULL_CG, the level of adherence to the full version of the CG Index (see Appendix C for the full list of questions of the index). Control variables are EQUITY OPERATION, dummy variable if it the operation refers to an equity deal; ADVISORY SERVICES, dummy if IFC provided advisory services to the client; NOMINEE DIRECTOR, dummy indicating if IFC appointed a nominee director to the board of the client; TIER iii OPERATION, dummy if the disbursement has been classified as a Tier iii investment by IFC; LISTED COMPANY, dummy if the client is a listed company; FIRM SIZE, natural logarithm of firm’s total assets; FIRM AGE, natural logarithm of years since company founding; FINANCIAL_LEVERAGE, total liabilities over total assets; CONC_SHARES, percentage of voting shares held by the controlling shareholder. Appendix E presents detailed operational definitions of all variables. Models are estimated through cross-sectional multiple linear regressions estimated with robust White-corrected standard errors. Regressions 2, 4, 6, and 8 are controlled for four industry dummy variables as per IFC’s industry classification. Data refer to the period from 2011 to 2016. Robust t-statistics are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Appendix B: Multiple Regressions 45 Table B.2: Core CG Index and Performance Variables ODO 1–6 Score ROE/ROIC EROE/EROIC Credit risk 1–11 Basic Basic Basic Basic Dependent Variables Model Full Model Model Full Model Model Full Model Model Full Model Model (1) (2) (3) (4) (5) (6) (7) (8) Core_cg_index 0.717*** 0.494 5.924** –4.000 9.099*** –5.134 –1.396** –0.834 (2.66) (1.41) (2.28) (–1.20) (2.60) (–1.44) (–2.41) (–0.94) Equity operation – –0.385** – 0.810 – 3.205*** – 1.267*** (–3.40) (0.92) (2.64) (5.65) Advisory services – 0.912*** – 0.067 – 1.976 – –0.207 (5.21) (0.04) (0.69) (–0.50) Nominee director – –0.167 – –2.950** – –3.183* – 0.803*** (–0.88) (–2.46) (–1.73) (2.64) Tier iii operation – –0.508*** – –4.064*** – –9.470*** – 0.313 (–3.47) (–6.04) (1.00) (–4.23) Listed company – 0.118 – 2.417* – 4.248** – –0.397 (1.05) (1.91) (2.37) (–1.54) Firm size 0.010 –0.035 0.092 0.294 –0.033 –0.241 –0.272*** –0.308*** (0.33) (–1.21) (0.28) (0.78) (–0.07) (–0.47) (4.45) (–4.10) Firm age 0.070 0.436*** 3.558*** 6.096*** 4.179*** 9.331*** 0.148 0.163 (1.05) (4.91) (6.08) (6.49) (5.74) (7.32) (1.11) (1.02) Financial_leverage 0.477*** 0.387** 6.718*** 7.181*** 7.405*** 7.458*** 0.207 0.793* (3.47) (2.40) (4.49) (3.79) (3.77) (3.25) (0.54) (1.69) Conc_shares – –0.012*** – –0.069*** – –0.098*** – 0.011*** (–5.88) (–3.17) (–3.56) (2.62) Industry dummies No Yes No Yes No Yes No Yes Constant 3.169*** 3.964*** –8.510** –8.645* –7.613* –11.295** 8.931*** 7.853*** (10.96) (9.64) (–2.59) (–1.95) (–1.66) (–2.26) (14.23) (9.79) Number of 264 213 259 208 248 197 259 208 observations F statistic 10.38 22.40 24.72 28.61 20.76 29.35 10.73 15.34 R-squared 0.095 0.599 0.286 0.540 0.240 0.550 0.149 0.421 Note: This table exhibits the outcomes of different regression models aiming at analyzing the relationship between the level of adherence to IFC’s CG Index and firm-level performance. The dependent variables are ODO 1–6 (models 1 and 2), the overall development outcome rating of the project on a 1–6-point scale ranging from highly unsuccessful to highly successful; ROE/ROIC (models 3 and 4), the annual return on equity or return on invested capital of the project; EROE/EROIC (models 5 and 6), the annual economic return on equity or economic return on invested capital of the project; and, CREDIT RISK 1–11 (models 7 and 8), the credit risk rating of the project on a scale from 1 to 11. The independent variable of interest is CORE_CG_INDEX, the level of adherence to the reduced version (Core CG Index) of the CG Index (see Appendix C for the full list of questions of the index). Control variables are: EQUITY OPERATION, dummy variable if it the operation refers to an equity deal; ADVISORY SERVICES, dummy if IFC provided advisory services to the client; NOMINEE DIRECTOR, dummy indicating if IFC appointed a nominee director to the board of the client; TIER iii OPERATION, dummy if the disbursement has been classified as a Tier iii investment by the IFC; LISTED COMPANY, dummy if the client is a listed company; FIRM SIZE, natural logarithm of firm’s total assets; FIRM AGE, natural logarithm of years since company founding; FINANCIAL_ LEVERAGE, total liabilities over total assets; CONC_SHARES, percentage of voting shares held by the controlling shareholder. Appendix E presents detailed operational definitions of all variables. Models are estimated through cross-sectional multiple linear regressions estimated with robust White-corrected standard errors. Regressions 2, 4, 6, and 8 are controlled for four industry dummy variables as per IFC’s industry classification. Data refer to the period from 2011 to 2016. Robust t-statistics are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 46 Governance and Performance in Emerging-Markets Firms The team ran the same regressions (shown in Table B.2), The process begins with the analysis of adherence to the this time using the level of adherence to the reduced Full CG Index (ADHERENCE_FULL_CG). Its relationship version (Core CG Index) of the CG Index (composed with firm performance is significantly affected only when of 26 questions) as the main explanatory variable of an equity deal takes place in model 1 (using ODO 1–6 interest. score as performance variable). This can be seen by the interaction term that multiplies the two explanatory Table B.2 also provides inconclusive results in the variables of interest in the second row of the table relationship between CG and performance. Although (ADHERENCE_FULL_CG * EQUITY OPERATION). the basic specifications of the regressions (models 1, Specifically, the negative coefficient of this interaction 3, 5, and 7) show a beneficial influence of the level of shows that an equity deal reduces the direct impact of CG adherence to the Core CG Index on all performance on performance (which, in this case, is not significant). variables, the statistical significance of the coefficients Still in model 1, a significant negative relationship can disappears in the full models employing the complete set be seen between the presence of an equity operation of controlling variables. Also observable is that the main and the overall development outcome rating of the results related to other explanatory variables remain project (ODO 1–6 SCORE). Thus an equity deal has the same as of those obtained in Table B.1. an aggregated negative impact on the ODO score of the project, which can be calculated by summing its direct The multiple regressions also make it possible to see coefficient on the fifth row (–1.106) and its indirect the contingent effect of potential moderators on the coefficient on the second row (–1.128 multiplied by CG-performance relationship. Specifically, it allows the 0–100% level of adherence to the Full CG Index). analysis of whether the relationship between CG and Since all other interaction terms of models 3, 5, 7 are performance varies depending on 1) the type of financing not statistically significant, there is very weak evidence (equity versus loan); 2) the magnitude of the investment that the relationship between CG and performance is (Tier iii versus Tier i or Tier ii); 3) the status of the significantly altered by the presence of equity deals. company as listed or not; 4) the presence of a nominee director; and 5) the provision of advisory services. To Next comes analysis to find the relationship between save space, Table B.3 presents the set of regressions using adherence to the Core CG Index (CORE_CG_INDEX) and the presence of an equity deal as a moderating variable. performance. This relationship is significantly affected (The results for the other moderating variables will be by the presence of an equity deal in models 2 and 6 described in the next paragraphs, and their full set of (using ODO 1–6 score and EROE as performance regressions can be provided on request.) variables, respectively). In both cases, the coefficients of the interaction terms are positive, indicating that the Table B.3 presents the results of the moderating variable impact of CG on both ODO and EROE is stronger when EQUITY OPERATION, which assumes a value of one if an equity deal takes place.30 Nevertheless, because the the project involves equity financing and zero otherwise. interaction terms of models 4 and 8 are not statistically Models 1, 3, 5, and 7 use the adherence to the Full CG significant, it cannot be concluded that the relationship Index (ADHERENCE_FULL_CG) as the explanatory between CG and performance is significantly altered variable of interest, while models 2, 4, 6, and 8 use for equity deals when using the reduced version of the adherence to the Core CG Index (CORE_CG_INDEX). Index as a CG indicator. 30  Take the results of model 6 (using EROE), for example. For a firm with a level of adherence of 80% to IFC’s Core CG Index, the impact of CG on EROE for equity deals would be computed by summing its direct impact (–12.823 * 80% = –10.2584) and its indirect impact (15.442 * 80% * 1 = 12.354), ending in a net positive impact of approximately 2.10% on EROE. Appendix B: Multiple Regressions 47 Table B.3: CG Index and Performance Variables: Equity Holdings Dependent Variables ODO 1–6 Score ROE/ROIC EROE/EROIC Credit risk 1–11 Model (1) (2) (3) (4) (5) (6) (7) (8) adherence_full_CG –0.569 – –2.685*** – –3.811 – –2.082* – (–1.08) (5.18) (–0.72) (–1.86) adherence_full_CG * –1.128* – –5.952 – –9.573 – 0.659 – equity operation (–1.87) (–0.90) (–1.21) (0.48) Core_CG_index – –0.640 – –9.561* – –12.823** – –1.800* (–1.25) (–1.93) (–2.39) (–1.73) Core__CG_index * equity – 2.030*** – 11.034 – 15.442** – 1.713 operation (3.24) (1.65) (2.07) (1.34) Equity operation –1.106*** –1.554*** 4.312 –5.651 8.969* –5.671 0.694 0.276 (–2.68) (0.99) (1.76) (0.73) (–4.21) (–1.44) (–1.31) (–3.40) Advisory services 0.929*** 1.039*** 0.123 0.622 1.905 2.667 –0.185 –0.099 (5.10) (0.07) (0.72) (–0.44) (6.01) (0.32) (0.90) (–0.23) Nominee director –0.148 –0.083 –2.988*** –2.493** –3.164* –2.539 0.825*** 0.872*** (–0.80) (–2.75) (–1.93) (2.75) (–0.49) (–2.04) (–1.34) (2.79) Tier iii operation –0.531*** –0.651*** –4.201*** –4.929*** –9.475*** –10.653*** 0.186 0.194 (–4.41) (–3.47) (–3.61) (–6.05) (–6.26) (0.60) (0.60) (–5.10) Listed company 0.212** 0.094 1.945 2.143* 3.497** 3.767** –0.425* –0.415 (1.98) (1.58) (2.06) (–1.66) (0.90) (1.75) (2.13) (–1.64) Firm size –0.019 –0.043 0.158 0.386 –0.421 –0.089 –0.306*** –0.315*** (–0.69) (–1.48) (0.42) (1.03) (–0.77) (–0.17) (–4.40) (–4.13) Firm age 0.450*** 0.454*** 5.827*** 6.181*** 8.784*** 9.427*** 0.123 0.178 (4.93) (5.37) (6.04) (6.63) (6.84) (7.43) (0.79) (1.11) Financial_leverage 0.327* 0.042 –0.075*** 5.110*** 7.600*** 4.701** 0.518 0.495 (1.92) (0.22) (–3.55) (2.78) (3.48) (2.03) (1.06) (0.96) Conc_shares –0.013*** –0.013*** –0.075*** –0.074*** –0.109*** –0.104*** 0.009** 0.010 (–6.37) (–3.55) (–4.05) (2.12) (–6.55) (–3.34) (–3.78) (2.51) Industry Dummies Yes Yes Yes Yes Yes Yes Yes Yes Constant 4.924*** 5.480*** 10.189** –6.458*** –9.741* –5.083 9.042*** 8.793*** (9.64) (10.87) (–2.16) (–1.46) (–1.75) (–0.90) (8.19) (8.52) (continued on next page) 48 Governance and Performance in Emerging-Markets Firms Table B.3: CG Index and Performance Variables: Equity Holdings (continued) Dependent Variables ODO 1–6 Score ROE/ROIC EROE/EROIC Credit risk 1–11 Model (1) (2) (3) (4) (5) (6) (7) (8) Number of observations 213 213 208 208 197 197 213 208 F statistic 21.06 18.40 31.75 25.79 32.51 26.38 22.40 14.45 R-squared 0.597 0.581 0.551 0.546 0.566 0.557 0.599 0.428 Note: This table exhibits the outcomes of different regression models using the variable EQUITY OPERATION as a moderator of the relationship between the level of adherence to the CG Index and firm-level performance. This variable assumes a value of one if the project involves equity financing and zero otherwise. Models 1, 3, 5, and 7 use adherence to the Full CG Index (ADHERENCE_FULL_ CG) as the explanatory variable of interest, while models 2, 4, 6, and 8 use the adherence to the Core CG Index (CORE_CG_INDEX). See Appendix C for the full list of questions of the index. The dependent variables are ODO 1–6 (models 1 and 2), the overall development outcome rating of the project on a 1–6-point scale, ranging from highly unsuccessful to highly successful; ROE/ROIC (models 3 and 4), the annual return on equity or return on invested capital of the project; EROE/EROIC (models 5 and 6), the annual economic return on equity or economic return on invested capital of the project; and CREDIT RISK 1–11 (models 7 and 8), the credit risk rating of the project on a scale from 1 to 11. Control variables are ADVISORY SERVICES, dummy if IFC provided advisory services to the client; NOMINEE DIRECTOR, dummy indicating if IFC appointed a nominee director to the board of the client; TIER iii OPERATION, dummy if the disbursement has been classified as a Tier iii investment by IFC; LISTED COMPANY, dummy if the client is a listed company; FIRM SIZE, natural logarithm of firm’s total assets; FIRM AGE, natural logarithm of years since company founding; FINANCIAL_LEVERAGE, total liabilities over total assets; CONC_SHARES, percentage of voting shares held by the controlling shareholder. Appendix E presents detailed operational definitions of all variables. Models are estimated through cross-sectional multiple linear regressions estimated with robust White-corrected standard errors. All regressions are controlled for four industry dummy variables as per IFC’s industry classification. Data refer to the period from 2011 to 2016. Robust t-statistics are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Similar analyses for the other moderating variables • For the other performance variables, the analysis yielded results that were qualitatively similar to those yielded mixed results, in some instances contradictory obtained for the presence of an equity deal. In most to the hypothesis. cases, the coefficients of the interaction terms were not • The analysis of potential moderating variables also statistically significant. did not allow drawing robust conclusions. In most cases, the coefficients of the variables of interest Again, note that the absence of statistically significant were not statistically significant. coefficients and conclusive results in the multiple regressions is likely a consequence of a dataset that is too small as well as relatively poor in terms of the variables’ construction (likely leading to a relevant measurement error). In summary, the following are the main results from the attempt to carry out multiple regressions with the database available so far: • The credit risk rating of the project was the only performance variable for which there were consistent results. In virtually all specifications, a higher level of adherence to IFC’s CG Index is associated with a lower credit risk rating of the project. Appendix B: Multiple Regressions 49 APPENDIX C Full Corporate Governance Index The CG Index focuses on a limited set of critical CG characteristics, selected through a survey of literature on CG indices, internationally recognized codes of best practices, and direct consultation with IFC corporate governance specialists using an online survey.31 It contains 84 questions, divided into five dimensions: commitment to CG, structure and funding of the board of directors, control environment and processes, transparency and disclosure, and shareholders rights; 51 are applicable to all firms, 5 for family-owned businesses, 7 for financial institutions, and 21 for listed companies. 31  However, the CG Index used in this research is not necessarily a complete CG indicator, and it may not have included some elements that can be relevant for measuring the true quality of the CG practices of a particular firm. 50 Governance and Performance in Emerging-Markets Firms Type of Control # Dimension # Question All 1 Commitment 1 Does the company have a Corporate Governance Code (or “Policy” or to CG “Guidelines”) in addition to the Articles of Association/By-laws addressing, at a minimum, the rights and treatment of shareholders, the role of the board of directors, transparency and disclosure, and business ethics? All 1 Commitment 2 Does the company have a written Code of Conduct? (1.0 if Code has been to CG approved by the Board; 0.5 if approved by the Management Team) All 1 Commitment 3 Do board members receive periodic training on corporate governance related to CG issues sponsored/funded by the company? All 1 Commitment 4 Does the company have a designated officer or governance body responsible to CG for overseeing corporate governance policies and practices? (1.0 if reports to the Board; 0.5 if reports to the Management Team) Family 1 Commitment 5 Does the company have a family-employment policy (e.g. establishing the rules Owned to CG for hiring and promoting members of the controlling family)? Family 1 Commitment 6 Does the company have written policies (e.g. family constitution) addressing key Owned to CG family governance element, such as – Succession planning; – Human resources and family member employment; – Family member share ownership? Listed 1 Commitment 7 Does the company periodically disclose the extent to which its corporate to CG governance code and practices conform to the code of best practice of its country? Listed 1 Commitment 8 Is the company a signatory of national and/or international initiatives to combat to CG corruption (e.g., the Principles for Countering Bribery, voluntary industry-specific codes of practice)? All 2 Structure & 9 Does the board receive periodic reports from the management about the Functioning of implementation of the strategic plan? the BoD All 2 Structure & 10 Does the board formally approve the strategic and business plan of the company? Functioning of the BoD All 2 Structure & 11 Does the board have an internal regulation (e.g. Charter, by-law or other formal Functioning of document) establishing its role, composition and functioning rules? the BoD All 2 Structure & 12 Does the board carry out a dedicated meeting or session to discuss strategy at Functioning of least once a year? the BoD All 2 Structure & 13 Does the company only appoint alternate directors if required by the local law? Functioning of (please assign zero point if there are alternate directors in place which are NOT the BoD mandatory by local law) All 2 Structure & 14 Does the board include women directors? (0.5 if the company has one female Functioning of director; 1.0 if there are two or more women on the board) the BoD All 2 Structure & 15 Does the board include directors who are independent of management and Functioning of controlling shareholders? (0.5 if the company has one independent director; 1.0 if the BoD two or more) (continued on next page) Appendix C: Full Corporate Governance Index 51 (continued) Type of Control # Dimension # Question All 2 Structure & 16 Are Chairman and CEO positions occupied by different people? Functioning of the BoD All 2 Structure & 17 Is the audit committee only composed by non-executive directors (NEDs)? (1.0 Functioning of if only composed by NEDs; 0.5 if ithere are non-executive members who are the BoD not board members) (For listed and financial institutions, consider independent directors instead of NEDs) All 2 Structure & 18 Is there at least one member of the audit committee has relevant audit and Functioning of accounting expertise? (1.0 if someone has audit and accounting expertise; 0.5 if at the BoD least one member has relevant finance expertise) All 2 Structure & 19 Does the board have an audit committee in place? Functioning of the BoD All 2 Structure & 20 Do all board committees have internal regulations specifying their role, Functioning of composition and functioning? (1.0 if all have internal regulations; 0.5 if just some the BoD of them have internal regulations) All 2 Structure & 21 Does the board have regular meetings between 6 to 8 times per year? (please use Functioning of 8 to 12 for financial institutions) the BoD All 2 Structure & 22 Do directors regularly receive all documentation of board meetings with at least 5 Functioning of days in advance? (please adopt 7 days for listed companies and 2 days for family- the BoD owned firms) All 2 Structure & 23 Does the board have a corporate secretary? (1.0 if corporate secretary is full time; Functioning of 0.5 if corporate secretary functions are combined with other functions, such as the BoD legal) All 2 Structure & 24 Does the company have a written policy establishing rules for the approval of Functioning of related parties’ transactions (RPTs)? (1.0 if policy requires conflicted directors to the BoD abstain from voting; 0.5 if policy exists but does not require conflicted directors to abstain) All 2 Structure & 25 Does the board have a formal remuneration policy for board members that takes Functioning of into account their attendance and membership of committees? the BoD All 2 Structure & 26 Does the board conduct self-evaluations or other reviews of its effectiveness on Functioning of an annual basis? (1.0 if self-evaluation is conducted by an external facilitator; 0.5 the BoD if conducted internally. Pleaso also assign 1.0 if company alternates between the two types of evaluation) All 2 Structure & 27 Does the company have a succession plan for the CEO? Functioning of the BoD All 2 Structure & 28 Does the board adopt Key Performance Indicators (KPIs) for the CEO and senior Functioning of management and periodically review their performance? the BoD All 2 Structure & 29 Does the board formally appoint the CEO? Functioning of the BoD (continued on next page) 52 Governance and Performance in Emerging-Markets Firms (continued) Type of Control # Dimension # Question Family 2 Structure & 30 Are Chairman and CEO occupied by non-relatives? Owned Functioning of the BoD Financial 2 Structure & 31 Does the company have a risk committee at the board level? (1.0 if there is a Functioning of separate risk committee; 0.5 if there is a joint audit and risk committee) the BoD Financial 2 Structure & 32 Is the compensation of the senior management team linked to individual Functioning of performance (not the overall company performance results)? the BoD Listed 2 Structure & 33 Does the company have a corporate governance committee in place? Functioning of the BoD Listed 2 Structure & 34 Is the board chaired by an independent director? (0.5 if there is a lead Functioning of independent director) the BoD Listed 2 Structure & 35 Is the compensation of the senior management team linked to individual Functioning of performance (not the overall company performance results)? the BoD All 3 Control 36 Does the audit committee oversee the implementation of the internal and Environment & external auditors’ recommendations? Processes All 3 Control 37 Is the board’s audit committee involving in the appointment of the external Environment & auditors? Processes All 3 Control 38 Does the company follow internationally recognized standards on internal audit? Environment & (e.g. the International Standards for the Professional Practice Internal Auditing Processes issued by the International Internal Audit Standards Board – IIASB) All 3 Control 39 Is there one person formally responsible for the compliance initiatives of the Environment & company (e.g. compliance officer or similar)? (1.0 for full-time compliance officer; Processes 0.5 for half-time) All 3 Control 40 Does the company follow internationally recognized standards on internal Environment & controls? Processes All 3 Control 41 Have the external auditors consistently flag repetitive shortcomings in the Environment & company’s internal control system during the last 3 years? Processes All 3 Control 42 Does the company have a specific whistleblower channel that ensures anonymity Environment & for informers and due treatment of the denounces? Processes All 3 Control 43 Does the board approve the company’s risk management policies? Environment & Processes All 3 Control 44 Does the board oversees the implementation of the risk management policies? Environment & Processes (continued on next page) Appendix C: Full Corporate Governance Index 53 (continued) Type of Control # Dimension # Question All 3 Control 45 Is there one person formally responsible for the risk management initiatives of Environment & the company? (e.g. CRO or risk manager) Processes All 3 Control 46 Does the Internal Audit unit have an audit work plan that is approved by the Environment & Audit Committee or by the board every year? Processes All 3 Control 47 Does the Internal Audit report directly to the Audit Committee or to the board of Environment & directors? Processes All 3 Control 48 Is there an internal audit function in place? Environment & Processes All 3 Control 49 Does the internal audit have its charter or specific terms of reference? (1.0 if Environment & approved by the Board; 0.5 if approved by the Management Team) Processes All 3 Control 50 Are financial statements audited by a recognized independent auditing firm? (1.0 Environment & if company uses an international audit firm; 0.5 if it hires a local audit firm) Processes All 3 Control 51 If the audit firm provides non-audit services, are non-audit fees limited to 25% of Environment & total auditor fees? Processes All 3 Control 52 Does the external auditor report on the adequacy of the company’s system of Environment & internal controls? Processes All 3 Control 53 Are the company’s financial statement audited based on International Standards Environment & on Auditing or other internationally recognized standards? Processes Financial 3 Control 54 Is there a formal compliance program in place? (1.0 if compliance program has Environment & been approved by the Board; 0.5 if approved by the Management Team) Processes Financial 3 Control 55 Does the internal audit carry out a risk-based audit? Environment & Processes Listed 3 Control 56 Is there a formal compliance program in place? (1.0 if compliance program has Environment & been approv ed by the Board; 0.5 if approved by the Management Team) Processes Listed 3 Control 57 Does the internal audit carry out a risk-based audit? Environment & Processes All 4 Transparency 58 Does the company prepare financial statements in accordance with international and Disclosure accounting standards (e.g. IFRS)? All 4 Transparency 59 Does the company publish an annual report in addition to its financial statement? and Disclosure (1.0 if annual report is both in the local language 0.5 if annual report is only in local language) All 4 Transparency 60 Is there a section on the firm’s website dedicated to corporate governance? and Disclosure (continued on next page) 54 Governance and Performance in Emerging-Markets Firms (continued) Type of Control # Dimension # Question All 4 Transparency 61 Does the company discuss all material transactions with affiliates of the and Disclosure controlling shareholders, directors or management in its statements? Financial 4 Transparency 62 Does the company disclose its main risks in the annual report or website? and Disclosure Financial 4 Transparency 63 Does the company periodically disclose in the Annual Report or website its and Disclosure political contributions and other donations? Listed 4 Transparency 64 Does the company have a designated officer to speak on behalf of the company? and Disclosure Listed 4 Transparency 65 Does the company hold periodic meetings with equity or credit analysts? and Disclosure Listed 4 Transparency 66 The company did NOT receive any regulatory sanctions for poor disclosure in the and Disclosure past 5 (five) years? Listed 4 Transparency 67 Does the company have a full time Investor Relations (IR) officer and IR unit? (1.0 and Disclosure if company has a IR unit and IR full time officer; have a IR unit) Listed 4 Transparency 68 Does the company disclose its main risks in the annual report or website? and Disclosure Listed 4 Transparency 69 Does the company disclose its sustainability policy? and Disclosure Listed 4 Transparency 70 Does the company disclose the individual compensation of board members and and Disclosure senior managers? (1.0 if company discloses individual compensation of directors and senior managers; 0.5 if discloses remuneration of only one group) Listed 4 Transparency 71 Does the company periodically disclose its political contributions and other and Disclosure donations? All 5 Shareholders 72 Is there any mechanism that permits minority shareholders to nominate Rights members of the board (cumulative voting, block voting, etc.)? All 5 Shareholders 73 Are shareholders provided with all material information and detailed agenda on Rights the date of the first notice of the shareholders’ meetings? All 5 Shareholders 74 Does the company have clearly articulated and enforceable policies with respect Rights to treatment of minority shareholders in changes of control, such as tag along, preemptive rights, etc.? All 5 Shareholders 75 Does the company only use common shares? Rights Family 5 Shareholders 76 Is there an exit provision for family shareholders? (A provision allowing Owned Rights minority shareholders to sell their shares in a manner which is equitable to all shareholders. E.g. put option, tag along clause, drag along) Family 5 Shareholders 77 Is there a family council established (if a number of family members are not Owned Rights working in the business)? Financial 5 Shareholders 78 Has the regulator, central bank, or the stock exchange conducted any Rights investigation into the company’s treatment of shareholders in the last five years? Listed 5 Shareholders 79 Does the company adopt mechanisms such as proxy voting and electronic voting? Rights Appendix C: Full Corporate Governance Index 55 APPENDIX D Glossary32 Annual General Meeting (AGM) (General Meeting of Shareholders): A shareholders’ gathering, usually held at the end of each fiscal year, at which shareholders and management discuss the previous year and the outlook for the future, directors are elected, and other shareholder concerns are addressed. The AGM usually makes decisions only on major issues affecting the company. Annual Report: A document issued annually by companies to their shareholders and stakeholders. It contains information on financial results and overall performance during the previous fiscal year and comments on the future outlook. The annual report might also include the corporate governance report and other narrative reports. Auditor’s Opinion: A certification that accompanies financial statements, provided by independent auditors who audit a company’s financial statements and records. The opinion indicates whether or not, overall, the financial statements present a fair reflection of the company’s financial condition. Audit: A review of the historical financial statements to enhance the degree of confidence in them. It consists of an examination and verification of a company’s financial and accounting records and supporting documents by a competent, qualified professional and independent external auditor in order to assure readers that they are in accordance with applicable reporting and accounting requirements, are free from material misstatement due to fraud or error and are true and fair representation of the company’s financial condition. Audit Committee: A committee constituted by the board of directors, typically charged with oversight of company reporting and disclosure of both financial and nonfinancial information to stakeholders. The committee usually is responsible for selecting and recommending the company’s audit firm to be approved by the board/shareholders. Board of Directors: The collective group of individuals elected by the shareholders of a company to direct and control the company. They define vision and mission, set the strategy and business priorities, and oversee the management of the company. The board is usually charged with issues such as selecting the chief executive officer, defining the compensation package of officers, and setting the long-term objectives of the firm and oversight of risk and compliance. There are two main board models around the world: one-tier board (unitary 32  This glossary is based primarily on previous IFC publications, especially 1) Who’s Running the Company—A Guide to Reporting on Corporate Governance (IFC Global Corporate Governance Forum, in partnership with International Center for Journalists, 2012); and 2) Practical Guide to Corporate Governance Experiences from the Latin American Companies Circle (IFC GCGF, in partnership with OECD, 2009). 56 Governance and Performance in Emerging-Markets Firms board system) and two-tier board (dual system). The unitary board system is characterized by a single board that governs the company (this board might include both executive and non-executive members). The dual system is characterized by distinct supervisory and management bodies. The former is commonly referred to as the supervisory board, the latter as the executive board. (Please see “one-tier board” and “two-tier board,” below.) In countries adopting the dual system, the terminology is “supervisory board” instead of “board of directors,” and “management board” instead of “senior management team.” Board Internal Regulation (Board Charter or Statute): A document detailing the roles, responsibilities, composition, and functioning of the board of directors and its committees. Bylaws: A written document stating the rules of internal governance for a company as adopted by its board of directors or shareholders. Includes topics such as election of directors, duties of officers, and how share transfers should be conducted. Chair of the Board: The person responsible for leadership of the board and the effectiveness of the board’s functioning. The chair is in charge of activities such as the elaboration of the board agenda as well as ensuring that the board’s business is conducted in the interest of all shareholders. Chief Executive Officer (CEO): The highest-ranking management officer of the company, who reports to the board of directors. The CEO is usually responsible for short-term decisions, leadership of employees, and implementation of strategy, risk management, and oversight of management. Code of Conduct (or Code of Ethics): A document describing the company’s principles, values, standards, or rules of behavior that should guide its decisions, procedures, and systems. It defines appropriate behaviors and actions on relevant and potentially delicate subjects. Committees of the Board (Board Committees): Committees are ancillary bodies of the board, established to assist it in the analysis of specific subjects outside of regular board meetings. Common committees are the audit, remuneration, and nomination committees. Common Shares: Equity securities representing ownership in a corporation and providing the holders with voting rights and the right to a share in the company’s residual earnings through dividends, stock buybacks, and/or capital appreciation. Compliance: Agreeing to and abiding by rules and regulations. In general, compliance means conforming to a specification or policy (internal or external), standard, or law that has been clearly defined. Company Charter: An official document filed with the relevant government agency in the country where the firm is incorporated. The charter outlines the corporation’s purpose, powers under law, authorized classes of securities to be issued, and the rights and liabilities of shareholders and directors. Conflict of Interest: A situation when a company official or a staff member has a competing professional or personal interest with the company, which can impede unbiased fulfillment of duties by such person. It includes a situation that has the potential to undermine the impartiality of a person because of the possibility of a clash between the person’s self-interest and professional interest or public interest. It may also be a situation in which a party’s Appendix D: Glossary of the Corporate Governance Indicator 57 responsibility to a second party limits its ability to discharge its responsibility to a third party. Directors have a duty to avoid conflicts of interest and should always act in the best interests of the company and all shareholders. Controlling Shareholders: Shareholders who own enough of the company’s voting capital to control the composition of the board of directors. Typically, a threshold of 30 percent or more is adopted for identifying a controlling shareholder. Corporate Governance: The structures and processes for the direction and control of companies. It aims to assure protection of shareholder rights and efficiency and effectiveness of interaction among the general meeting of shareholders, the board of directors and the senior management team. Corporate Governance Code: A document, typically approved by the company’s general shareholders’ meeting or board of directors, aiming to demonstrate the company’s commitment to sound corporate governance by providing guidelines on issues such as board practices, shareholder rights, related-party transactions, relations with stakeholders, and information disclosure and transparency policies. Corporate Secretary: A key figure of a corporate governance system. The corporate secretary assures the company’s compliance with corporate governance legislation and internal documents and fosters efficient coordination of the operations of all company governance bodies, information sharing between such bodies and shareholders, and resolution of other corporate governance issues within the company. Tasks usually comprise the organization of the meetings of governance bodies, compliance of the procedures of the board of directors with the law, keeping the book of shareholders, and preparing and recording minutes of all meetings of shareholders, board of directors, and other governance bodies. Cumulative Voting: A voting system that gives minority shareholders more power, by allowing them to cast all of their board of director votes for a single candidate, as opposed to regular or statutory voting, in which shareholders must vote for a different candidate for each available seat or distribute their votes between a number of candidates. Disclosure: Refers to the obligation of a firm to provide material information in accordance with the requirements of a number of parties, including regulatory authorities and the public, or in accordance with standards, such as accounting standards, and self-regulatory contracts. Disclosure contributes to the transparency of the firm, which is one of the main corporate governance principles. Dual-Class Shares: Shares that have different rights, such as A Class and B Class shares, where one class has voting rights and the other does not. Executive Director: A member of a company’s board of directors who is also part of the company’s executive team (for example, the CEO or the chief financial officer). Executive Session: The portion of a board of directors’ meeting that excludes the chief executive or any other executive. External Auditors (Independent Auditors): Professionals from an external auditing firm charged with undertaking an audit of the financial statements. An audit may be required annually, semiannually, or quarterly. In most countries, the independent auditors undertake an annual audit. The independent auditor is required to render 58 Governance and Performance in Emerging-Markets Firms an unbiased judgment that the financial statements and accounting records of the firm are likely to be free from material misstatement and are a fair reflection of the financial position of the firm. Family Constitution: Guidelines for the rights and duties of family members who will share in the family’s resources, mainly those associated with invested companies. Family Council: Organized forum for family members to meet and discuss the current and future state of the family business. Members may or may not be directly involved in the day-to-day business operations. The family council is a way of building family unity and cohesiveness through a shared vision of the family’s guiding principles and to separate the professional management of the firm from the personal family issues. It is usually the forum to determine how the family shareholding will be voted on any matter. Family-Owned Businesses: Companies in which the controlling shareholders belong to the same family (immediate or wider family members) or group of families. Financial Statements: a complete set of financial statements comprises a balance sheet, an income statement, a statement of changes in equity, a cash flow statement, and notes. They collectively communicate an entity’s economic resources or obligations at a point in time or the changes therein for a period in accordance with a financial reporting framework. Free-Float: The portion of shares negotiated in the market, giving liquidity to shares. These shares are not held by large owners and are not shares held in the company’s treasury. Generally Accepted Accounting Principles (GAAP): Accounting rules, conventions, and standards for companies, established by reporting requirements and accounting standard setters in the country. Each country is likely to have a GAAP, which is unlikely to be identical to any other country’s GAAP. Independent Director: Someone whose only nontrivial professional, familial, personal, or financial connection to the corporation, its chairman, CEO, or any other executive officer is his or her directorship. The independent director is expected to be capable of applying objective judgment to all company decisions. Internal Audit: An independent, objective assurance and consulting activity designed to add value and improve an organization’s operations. It helps an organization accomplish its objectives by bringing a systematic, disciplined approach to evaluate and improve the effectiveness of risk management, control, and governance processes. Investor Relations: The corporate communications department of a company. This department specializes in information and disclosure management for public and private companies as they communicate with the investment community at large. Minority Shareholders: Those shareholders with minority stakes in a company controlled by a majority shareholder— usually less than a 5 percent stake. However, each country may determine various thresholds applicable to the term “minority shareholder.” Non-Executive Director (NED) (Outside Director): A member of a company’s board of directors who is not part of the company’s executive team. A non-executive director typically does not engage in the day-to-day management Appendix D: Glossary of the Corporate Governance Indicator 59 of the organization. Since the non-executive director may have relationships with the controlling shareholder, he or she might not necessarily be considered an independent director. Non-Voting Shares: Owners holding this share class do not commonly have voting rights at the AGM, except on some matters of highest importance. Usually, non-voting shareowners have preferential rights for receiving dividends. One-Tier Board: A board of directors composed of both executive and non-executive members. It delegates day-to-day business to the management team. Found in the United States, the United Kingdom, and Commonwealth countries. Ownership Structure: The way in which company shares are distributed among shareholders. Preferred Shares: Equity securities representing ownership in a corporation that have preferential rights over other share classes regarding the payment of dividends and distribution of assets upon liquidation. Preferred shares usually do not carry voting rights. Pyramidal Structure: An organizational structure common in family-dominated companies. Legally independent companies are controlled by the same family through a chain of ownership relations. Related Party: A party is related to an entity if it can directly or indirectly control the other party or exercise control through other parties; it may also be where parties are subject to a common control from the same source. Related parties tend to have influence over the financial or operating policies of a firm or have the power to influence another party’s actions. A related party may be a close family member (including partners, spouses, children, other relatives), a key manager in the entity (and their close family members), or entities such as subsidiaries of the entity, holding companies, joint ventures, and associates. Risk Management: The process of identifying, analyzing, managing, and monitoring a corporation’s exposure to risk and determining optimal approaches to handling such exposure. Senior Management Team: A company’s body carrying out the day-to-day management of its operations, including implementation of a company’s strategic development plan, introduction of a corporate culture in accordance with business ethics standards, and drafting the company’s bylaws and guidelines. A company may establish a sole executive body (for example, the CEO) or a collegial executive body (the management board). Share: A security issued by a joint-stock company and authorizing the right to participate in a company’s general meeting of shareholders, to receive dividends, and to receive a portion of a company’s assets in case of the company’s dissolution. Shareholders: Holders of shares issued by companies. Shareholders’ Agreement: A written document governing the relations among shareholders and defining how the company will be managed and controlled. The agreement helps align the objectives of controlling shareholders to safeguard common interests and to protect the interests of minority shareholders. 60 Governance and Performance in Emerging-Markets Firms Shareholders Rights: The rights resulting from ownership of shares, which may be based in legal rights or other rights contracted with the company. The basic shareholder rights include the right to information on the company, to attend the meeting of shareholders, to elect directors, and to appoint the external auditor, as well as voting rights and cash flow rights. Stakeholder: A person or organization with a legitimate interest in a project or company. In a more general sense, it refers to suppliers, creditors, clients, employees, and the local community—all affected by the actions of the company. Tag-Along Rights: If a majority shareholder sells his or her stake, minority holders have the right to participate and sell their stake under the same terms and conditions as the majority shareholder. This right protects minority shareholders and is a standard inclusion in shareholders’ agreements. Transparency: The corporate governance principle of publishing and disclosing timely information relevant to stakeholders’ interests and to shareholders on all material matters, including its financial position, performance, and ownership and governance structure. Two-Tier Board: A board of directors that divides supervisory and management duties into two separate bodies. The supervisory board, comprising non-executive directors, oversees the management board, comprising executive directors. Common in France, Germany, and Eastern Europe. Not all styles of two-tier boards are identical. Voting Rights: The right to vote at shareholders’ meetings on issues of importance for the company. Voting Shares: Shares that give the shareholder the right to vote on matters of corporate policy, including elections to the board of directors. Appendix D: Glossary of the Corporate Governance Indicator 61 APPENDIX E Operational Definition of All Variables Type of Code Variable Construct Definition Dimension Ind_Name Control Industry IFC Industry Classification NA Dummy_Ind Control Industry Dummy of Industry Classification NA Sector Control Sector IFC Secondary Sector Name NA Country Control Country Client Country NA Reg_Name Control Region Client Region NA Dummy_Reg Control Region Dummy of Client Region NA Inception Control Firm Age Year of Inception of Firm’s Operation NA Conc_Shares Control Ownership Percentage of voting shares held by the NA concentration controlling shareholder Listed Control Listed on stock 1 if listed on stock exchange; 0, otherwise. NA exchange Cross_List Control Cross-listed on 1 if cross-listed on foreign stock exchange (e.g. NA foreign stock ADR Level 2 or 3 on NYSE); 0, otherwise. exchange Bus_Group Control Business group If firm belongs to business group, then 1; if it NA membership stand-alone firm, then 0. Assets Control Firm Size Total assets (in US$) NA Revenues Control Firm Size Total revenues in US$ NA Shar_Eq Control Shareholders’ Equity (average in US$) NA Equity Net_Income Control Profitability Net Income in US$ NA Fix_Asset Control Fixed Assets Premises & Equipment in US$ NA Liabilities Control Liabilities Total liabilities in US$ NA Debt_Asset Control Financial Total liabilities / total assets NA leverage (continued on next page) 62 Governance and Performance in Emerging-Markets Firms (continued) Type of Code Variable Construct Definition Dimension Tangibility Control Asset Fixed assets (property, plant, & equipment) / NA tangibility operational revenue Tier Moderating Size of IFC Tier i, ii or iii NA disbursement Loan Moderating Type of IFC If loan deal, then 1; otherwise, 0 NA operation Equity Moderating Type of IFC If equity deal, then 1; otherwise, 0 NA operation New_Client Moderating History of If new client, then 1; if repeated client, then 0 NA relationship with IFC Nom_Right Moderating Right for If IFC has the right to indicate a Nominee Director, NA Nominee 1; Otherwise, 0 Director Nominee Moderating Nominee If IFC exercises its rights and indicates a Nominee NA Director Director, 1; Otherwise, 0 Female Moderating Female If IFC indicates a Female Director, 1; Otherwise, 0 NA Nominee Director Staff_FStaff Moderating Staff or If IFC indicates a Staff or Former Staff as Director, NA Former Staff 1; Otherwise, 0 as Nominee Director Advisory Moderating Type of IFC If IFC carried out CG-related advisory services, NA involvement then 1; otherwise, 0 SIZE Moderating Size of IFC Amount (US$) involved operation (Loan or Equity) NA disbursement DOTS_1–6_ Dependent Performance 6-point scale Multidimensional Score DOTS_Fin_ Dependent Performance 4-point scale Financial Perf DOTS_Econ_ Dependent Performance 4-point scale Economic Perf DOTS_E&S Dependent Performance 4-point scale E&S Performance DOTS_PSD Dependent Performance 4-point scale Private Sector Dev. Av_DOTS_1– Dependent Performance 6-point scale Multidimensional 6_Score (Average from 2011 to 2016) Av_DOTS_Fin_ Dependent Performance 4-point scale Financial (Average Perf from 2011 to 2016) (continued on next page) Appendix E: Operational Definition of All Variables 63 (continued) Type of Code Variable Construct Definition Dimension Av_DOTS_ Dependent Performance 4-point scale Economic Econ_Perf (Average from 2011 to 2016) Av_DOTS_E&S Dependent Performance 4-point scale E&S Performance (Average from 2011 to 2016) Av_DOTS_PSD Dependent Performance 4-point scale Private Sector Dev. (Average from 2011 to 2016) ROE_Annual Dependent Performance Return on Invested Capital / Equity Financial EROE_Annual Dependent Performance Economic Return on Equity Financial Avrg_ROE Dependent Performance Return on Invested Capital / Equity (Average from Financial 2011 to 2016) Avrg_EROE_ Dependent Performance Economic Return on Equity (Average from 2011 to Financial Annual 2016) Credit_Risk Dependent Performance Credit Risk Rating. CRRs range from 1 to 7, where Financial 1 indicates the lowest risk, and 7 the highest risk level (commensurate with default). Additional modifiers (A and B) are used to add further granularity in the CRR for categories 2 to 5, where A defines a higher credit quality than B within the category. Indicators (‘+’ improving; or ‘–‘ deteriorating; or ‘=’ neutral) are used to signify the expected future trend in the CRR. The CRR categories are Very Good (1), Good (2), Average (3), Watch (4), Substandard (5), Doubtful (6) and Loss (7). Cred_Risk_1–7 Dependent Performance Credit Risk Rating (1–7 Scale) Financial 1–1 Very Good; 2–2A Good; 2–2B Good; 3–3A Average; 3–3B Average; 4–4A Watch; 4–4B Watch; 5–5A Substandard; 5–5B Substandard; 6–6 Doubtful; 7–7 Loss. Cred_Risk_1–11 Dependent Performance Credit Risk Rating (1–11 Scale) Financial 1–1 Very Good; 2–2A Good 3–2B Good; 4–3A Average; 5–3B Average; 6–4A Watch; 7–4B Watch; 8–5A Substandard; 9–5B Substandard 10–6 Doubtful; 11–7 Loss. ESRR Dependent / Performance Environmental & Social Review Rating: E&S Risk Rating Explanatory / E&S Index indicating Client’s capability and/or management of E&S issues in accordance with IFC’s Sustainability Framework: 1 – Excellent; 2 – Satisfactory; 3 – Partly Unsatisfactory; 4 – Unsatisfactory. Adh_Full_CG Explanatory Corp. Overall percentage of adherence with all CG Multidimensional Governance dimensions (continued on next page) 64 Governance and Performance in Emerging-Markets Firms (continued) Type of Code Variable Construct Definition Dimension Adh_Commit_ Explanatory Corp. Overall percentage of adherence with Multidimensional CG Governance Commitment to CG Adh_Board Explanatory Corp. Overall percentage of adherence with Structure & Structure & Governance Functioning of the Board Functioning of the Board Adh_Control Explanatory Corp. Overall percentage of adherence with Control Control Governance Environment and Processes Environment and Processes Adh_Transp_ Explanatory Corp. Overall percentage of adherence with Transparency and Disc Governance Transparency and Disclosure Disclosure Adh_Shar_ Explanatory Corp. Overall percentage of adherence with Shareholders Rights Governance Shareholders Rights Rights Var_Full_CG Explanatory Corp. Evolution of percentage of adherence with all CG Multidimensional Governance dimensions (2011–2016) Core_ Explanatory Corp. Overall percentage of adherence with the 26 Selected Indicators CG_26Ind Governance selected questions (2, 4, 6, 10, 14, 15, 16, 19, 21, 24, 30, 31, 33, 38, 39, 42, 45, 48, 54, 56, 59, 62, 69, 72, 75, and 81) Var_Core_ Explanatory Corp. Evolution of percentage of adherence with the 26 Selected Indicators CG_26Ind Governance selected questions (2, 4, 6, 10, 14, 15, 16, 19, 21, 24, 30, 31, 33, 38, 39, 42, 45, 48, 54, 56, 59, 62, 69, 72, 75, and 81) Appendix E: Operational Definition of All Variables 65 2121 Pennsylvania Ave., NW Washington, D.C. 20433, USA www.ifc.org November/2018