UNIVERSITY OF GHANA COLLEGE OF HUMANITIES PROFIT EFFICIENCY AND CAPITAL STRUCTURE OF BANKS IN GHANA: A DEA APPROACH SANDRA NAANA AYIKU DEPARTMENT OF FINANCE JULY 2015 University of Ghana http://ugspace.ug.edu.gh UNIVERSITY OF GHANA COLLEGE OF HUMANITIES PROFIT EFFICIENCY AND CAPITAL STRUCTURE OF BANKS IN GHANA: A DEA APPROACH BY SANDRA NAANA AYIKU (ID. NO. 10277103) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE AWARD OF MASTER OF PHILOSOPHY IN FINANCE DEGREE DEPARTMENT OF FINANCE JULY 2015 University of Ghana http://ugspace.ug.edu.gh DECLARATION I do hereby declare that this work is the result of my own research undertaken under supervision and has not been presented by anyone for any academic award in this or any other university. All references used in the work have been fully acknowledged. I bear sole responsibility for any shortcomings. …………………………………………. ………………………… SANDRA NAANA AYIKU DATE (10277103) University of Ghana http://ugspace.ug.edu.gh ii CERTIFICATION I hereby certify that this thesis was supervised in accordance with procedures laid down by the University of Ghana. ………………………………………….. ………………………………….. DR. KWAKU OHENE-ASARE DATE (SUPERVISOR) ………………………………………….. ………………………………….. DR. GODFRED ALUFAR BOKPIN DATE (CO-SUPERVISOR) University of Ghana http://ugspace.ug.edu.gh iii DEDICATION This work is dedicated to my beloved parents, Mr. Alan Doe Ayiku and Mrs. Felicia Owusu Ayiku for their encouragement throughout the period of my studies. University of Ghana http://ugspace.ug.edu.gh iv ACKNOWLEDGEMENT I thank the Almighty God for all His wisdom, strength, and mercies throughout the period of this study. I would like to express my sincere appreciation to my supervisors Dr. Kwaku Ohene-Asare and Dr. Godfred Alufar Bokpin for their guidance and constructive criticisms which shaped and enriched the quality of this study. University of Ghana http://ugspace.ug.edu.gh v TABLE OF CONTENTS DECLARATION ............................................................................................................................. i CERTIFICATION .......................................................................................................................... ii DEDICATION ............................................................................................................................... iii ACKNOWLEDGEMENT ............................................................................................................. iv TABLE OF CONTENTS ................................................................................................................ v LIST OF TABLES .......................................................................................................................... x LIST OF FIGURES ....................................................................................................................... xi TABLE OF ABBREVIATIONS .................................................................................................. xii ABSTRACT .................................................................................................................................. xv CHAPTER ONE ............................................................................................................................. 1 1.1 Research Background ....................................................................................................... 1 1.2 Problem Statement ........................................................................................................... 5 1.3 Gaps and Contributions .................................................................................................... 9 1.4 Research Objectives ....................................................................................................... 10 1.5 Research Questions ........................................................................................................ 10 1.6 Significance of the Study ............................................................................................... 11 1.7 Limitations of the Study ................................................................................................. 11 1.8 Chapter Disposition ........................................................................................................ 12 CHAPTER TWO .......................................................................................................................... 13 LITERATURE REVIEW ............................................................................................................. 13 2.1 Introduction ......................................................................................................................... 13 University of Ghana http://ugspace.ug.edu.gh vi 2.2 Theoretical Review ............................................................................................................. 13 2.2.1 Capital Structure Theories and Firm Performance ....................................................... 13 2.2.2 Hypothesis on Reverse Causality from Firm Performance to Capital Structure .......... 17 2.2.3 Capital Structure, Competition and Firm Performance ................................................ 18 2.2.4 Conglomeration and Firm Performance ....................................................................... 19 2.3 Empirical Review ................................................................................................................ 21 2.3.1 Capital Structure Theories and Firm Performance ....................................................... 21 2.3.2 Reverse Causation from Performance to Capital Structure .......................................... 25 2.3.3 Capital Structure, Competition and Firm Performance ................................................ 26 2.3.4 Capital Structure, Conglomeration and Firm Performance .......................................... 27 2.3.5 Relevant Literature on Profit Efficiency in Banking .................................................... 29 2.3.6 Banking Efficiency Studies in Ghana ........................................................................... 32 2.4 Conceptual Framework ....................................................................................................... 33 2.5 Chapter Summary ................................................................................................................ 36 CHAPTER THREE ...................................................................................................................... 37 THE BANKING INDUSTRY IN GHANA ................................................................................. 37 3.1 Introduction ......................................................................................................................... 37 3.2 Historical Background......................................................................................................... 37 3.2.1 Overview of the Ghanaian Banking Industry ............................................................... 40 3.2.1.1 Implications of Some Key Developments on the Ghanaian Banking Industry ......... 40 3.2.1.2 Performance of the banking industry between 2000 to 2013 .................................... 43 Chapter Summary ...................................................................................................................... 45 CHAPTER FOUR ......................................................................................................................... 46 University of Ghana http://ugspace.ug.edu.gh vii METHODOLOGY ....................................................................................................................... 46 4.1 Introduction ......................................................................................................................... 46 4.2 Research Design .................................................................................................................. 46 4.3 Sampling and Sources of Data ............................................................................................ 46 4.4 The DEA Methodology ....................................................................................................... 47 4.5 Profit Efficiency Using the Non-parametric DEA Methodology ........................................ 51 4.5.1. An Illustrative Example of Profit Efficiency Using DEA ........................................... 56 4.6 Modelling Of Inputs and Output Variables ....................................................................... 58 4.6.1 Inputs ............................................................................................................................ 59 4.6.2 Outputs.......................................................................................................................... 61 4.6.3 Input Prices ................................................................................................................... 61 4.7 Bootstrapping the Second-Stage Regression with environmental variables ....................... 62 4.8 Capital Structure and Profit Efficiency ............................................................................... 65 4.8.1 Variable Measurements .................................................................................................... 66 4.8.1.1 Equity capital and Equity capital squared ................................................................. 66 4.8.1.2 Bank Size ................................................................................................................... 66 4.8.1.3 Sales Growth.............................................................................................................. 67 4.8.1.4 Ownership Structure .................................................................................................. 67 4.8.1.5 Regulation .................................................................................................................. 67 4.8.1.6 Technical Efficiency .................................................................................................. 68 4.9 Measures of Competition .................................................................................................... 68 4.9.1 Capital Structure, Profit Efficiency and Competition. ................................................. 69 4.9.2 Capital Structure, Conglomeration and Profit Efficiency ............................................ 70 University of Ghana http://ugspace.ug.edu.gh viii 4.9.3 Reverse Causation between Profit Efficiency and Capital Structure ............................... 71 4.9.4 Instruments for Data Analysis .......................................................................................... 71 4.9.5 Chapter Summary ............................................................................................................. 71 CHAPTER FIVE .......................................................................................................................... 72 DATA ANALYSIS ....................................................................................................................... 72 5.1 Introduction ......................................................................................................................... 72 5.2 Descriptive Statistics of Variables ...................................................................................... 72 5.3 Profit Efficiency of Banks in Ghana ................................................................................... 79 5.4 Effect of Capital Structure on Profit Efficiency .................................................................. 83 5.5 Capital Structure, Competition and Profit Efficiency. ........................................................ 90 5.6 Capital Structure, Conglomeration and Profit Efficiency. .................................................. 94 5.7 Reverse Causation Between Profit Efficiency and Capital Structure. ................................ 95 5.8 Chapter Summary ................................................................................................................ 96 CHAPTER SIX ............................................................................................................................. 98 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS................................................. 98 6.1 Introduction ......................................................................................................................... 98 6.2 Summary ............................................................................................................................. 98 6.3 Conclusions of the Study................................................................................................... 101 6.4 Recommendations ............................................................................................................. 104 REFERENCES ........................................................................................................................... 107 APPENDICES ............................................................................................................................ 132 APPENDIX A ......................................................................................................................... 133 APPENDIX B ......................................................................................................................... 138 University of Ghana http://ugspace.ug.edu.gh ix APPENDIX C ......................................................................................................................... 141 APPENDIX D ......................................................................................................................... 142 University of Ghana http://ugspace.ug.edu.gh x LIST OF TABLES TABLE PAGE Table 1: Key Developments in the Ghanaian Banking Industry from 2000 to 2013 39 Table 2: A hypothetical sample data 56 Table 3: Results for maximum profit, actual profit and profit efficiency scores. 57 Table 4: Inputs, Outputs, Input and Output Prices for Profit Efficiency Estimation 59 Table 5: Summary statistics of variables used-pooled data (GH¢) 73 Table 6: Summary statistics of variables for local and foreign banks 77 Table 7:Summary statistics of variables for conglomerates and focus banks 78 Table 8: Average Profit Efficiencies of Banks in Ghana 79 Table 9: Average Profit Efficiency Rankings Of banks in Ghana from 2000 to 2013 81 Table 10 : Comparison of DEA and traditional profitability ratios 82 Table 11: Correlation between DEA, ROA and ROE 83 Table 12: Pearson Correlation 84 Table 13: Spearman's rho 84 Table 14: Summary Statistics of variables used in truncated regression 86 Table 15: Regression Results: Profit Efficiency and Capital Structure 87 Table 16: Regression Results : Capital Structure, Competition and Profit Efficiency 91 Table 17: Regression Results : Capital Structure, Competition and Profit Efficiency 93 Table 18: Regression results: Capital Structure, Conglomeration and Profit Efficiency 95 Table 19: Results for reverse causation from profit efficiency to capital structure 96 Table 20: Regression Results: Profit Efficiency and Capital Structure 141 University of Ghana http://ugspace.ug.edu.gh xi LIST OF FIGURES FIGURE PAGE Figure 1: Structure of the Banking Industry 42 Figure 2: ROE from 2004-2013 44 Figure 3 : ROA from 2004-2013 44 Figure 4: A graphical representation of Profit Efficiency using one input and one output 57 Figure 5: Trend Analysis of Average Inputs and Outputs from 2000 to 2013 74 Figure 6: Trend Analysis of Average input and output prices from 2000 to 2013 75 University of Ghana http://ugspace.ug.edu.gh xii TABLE OF ABBREVIATIONS ACCESS - Access Bank Ghana Limited ADB - Agricultural Development Bank AMAL - Amalgamated Bank ATM - Automated Teller Machines BARODA - Bank of Baroda BBG - Barclays Bank Ghana BCC - Banker, Charnes and Cooper BI - Boone Indicator BOA - Bank of Africa BOG - Bank of Ghana BS - Bank size BSIC - Banque Sahelo-Saharienne CAL - Cal Bank Ghana Limited CCR - Charnes, Cooper and Rhodes CRS - Constant Returns to Scale CSFP - Capital Structure-Firm Performance DEA - Data Envelopment Analysis DGP - Data Generating Process DMU - Decision Making Units EBG - Ecobank Ghana Limited ECAP - Equity Capital ENERGY - Energy Bank ERP - Economic Recovery Program EU - European Union University of Ghana http://ugspace.ug.edu.gh xiii FABL - First Atlantic Bank FBL - Fidelity Bank FEAR - Frontier Efficiency Analysis with R FINSAP - Financial Sector Adjustment Program GCB - Ghana Commercial Bank Limited GDP - Gross Domestic Product GM - Gross Profit Margin GMM - Generalized Method of Moments GSS - Ghana Statistical Service GTB - Guaranty Trust Bank HFC - HFC Bank HHI - Herfindahl Hirschman Index IBG - Intercontinental Bank ICB - International Commercial Bank IMF - International Monetary Fund METRO - Metropolitan and Allied Bank NIB - National Investment Bank NPV - Net Present Value OLS - Ordinary Least Squares OWN - Ownership PBL - Prudential Bank REG - Regulation ROA - Return on Assets ROE - Return on Equity ROYAL - Royal Bank SCB - Standard Chartered Bank University of Ghana http://ugspace.ug.edu.gh xiv SFA - Stochastic Frontier Analysis SG-SSB - Societe Generale Ghana Limited SME - Small and Medium Scale Enterprises STANBIC - Stanbic Bank TE - Technical Efficiency TTB - The Trust Bank UBA - United Bank of Africa UMB - Universal Merchant Bank UNIBANK - Unibank Ghana Limited UT - UT Bank VIF - Variance Inflation Factor VRS - Variable Returns to scale ZENITH - Zenith Bank University of Ghana http://ugspace.ug.edu.gh xv ABSTRACT Using a nonparametric DEA approach, this study estimates the profit efficiency of 26 banks in Ghana over the period 2000 to 2013. This is compared to the estimates of two profitability ratios, ROA and ROE. The study then examines the influence of capital structure on the estimated profit efficiency and the extent to which the degree of competition measured by the Boone Indicator and the Herfindahl Hirschman Index influences the capital structure-profit efficiency nexus. The extent to which a bank’s ownership of a subsidiary (conglomeration) impacts the capital structure-profit efficiency link is also investigated. Two competing hypotheses-the efficiency risk and franchise value hypothesis are also tested to summarize the bi–causal relationship that exists between profit efficiency and capital structure. The results reveal that banks in Ghana operate close to the benchmark profit frontier and are 79% profit efficient. A comparison of the nonparametric DEA profit efficiency indicator with the profitability ratios suggest that these methods agree weakly on the performance of a bank. For the impact of capital structure on profit efficiency, the study found support for the trade-off and agency cost theories of capital structure. It was also found that competition and conglomeration do not necessarily influence the link between capital structure and profit efficiency of the banks. The study further found support for the efficiency risk hypothesis which indicates that profit efficient banks in Ghana choose more leverage relative to equity in financing their operations. University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE INTRODUCTION 1.1 Research Background One of the important decisions a firm must make towards achieving its objective of value maximization pertains to its capital structure (Fosu, 2013; Margaritis & Psillaki, 2010; Berger & Bonaccorsi di Patti, 2006; Modigliani & Miller, 1958). This is because, the ability to choose the appropriate combination of equity and debt can help the firm solve some of the challenges it faces in its quest to maximize the returns of its stakeholders. For instance, in an attempt to solve the problems of agency which originates from the separation of power between managers and owners, a firm may employ more debt as opposed to equity in financing its operations (Armen, Gayane & Hassan, 2004; Harvey, Lins & Roper, 2004; Williams, 1987; Grossman & Hart, 1982). Wruck (1995) contended that the use of greater debt (125% and 65% of the book and market values of their assets respectively) placed management under pressure to make efficient gains which kick started a remarkable improvement in Sealed Air Corporation’s performance. Capital structure decisions are also important because firms operate in open systems (Pearce & Robinson, 2009). This means, each firm is dependent on other institutions which provide essential resources for its operations. In this regard, for a firm to strategically position itself in the competitive industry, suitable decisions regarding its capital structure must be adopted (Abor, 2005). One industry where capital structure decisions are particularly important is the banking industry. This is because unlike other industries, agency costs are large in the banking industry. This stems from the fact that banks are informationally opaque (banks hold confidential information regarding their borrowers and other creditors) and this makes them vulnerable to University of Ghana http://ugspace.ug.edu.gh 2 increased agency costs. In addition, deposit insurance by government gives banks the incentives to increase risks (Margaritis & Psillaki, 2010, 2007; Pennacchi, 2006) which they can achieve by either increasing their assets or their leverage (Berger & Bonaccorsi di Patti, 2006). These risk- increasing-methods may increase the costs of agency for both equity holders and debt holders. For instance, the cost incurred in appointing board of directors to efficiently monitor and provide valuable advice for running the bank by equity holders (de Andres & Vallelado, 2008). Similarly, banks, by their nature engage in multiple and seemingly opposing roles which leaves them exposed to runs (Cebenoyan & Strahan, 2004; Diamond & Dybvig, 1983). These roles include, providing liquidity on demand to depositors and extending credit to borrowers (Berger & Bouwman, 2009; Kashyap, Rajan & Stein, 2002). Due to these fundamental roles, banks are usually apprehensive with their liquidity and solvency risks (Amidu, 2007). Traditionally, banks have reduced these risks by holding capital as a buffer for insolvency and liquid assets to meet unexpected withdrawals by depositors (Saidenberg & Strahan, 1999). Today, the risks associated with banks have come under increasing scrutiny with regulators setting minimum capitals (equity) for bank operations consistent with the provisions stipulated under the Basel Capital Accords. Thus, the success of banks depends largely on their ability to effectively and efficiently utilize the debt (leverage) they generate from their customers and other financiers. It has been argued in the banking and finance literature that, financial intermediation pursued by banks, play crucial roles in the economic development and growth of most countries (Levine, Loayza & Beck 2000; Levine, 1997). For example, banks ignited industrialization in England through the mobilization of capital and spurred technological advancement by recognizing and funding individuals capable of implementing state-of-art production processes and products University of Ghana http://ugspace.ug.edu.gh 3 (Schumpeter, 1912; Hicks, 1969). In the Ghanaian context, the banking industry is considered to be the driver of the services sector, which contributed about 49.5% of Gross Domestic Product (GDP) in 2013 (GSS, 2014). This implies that bank efficiency may have positive ripple effects on other sectors of the economy. Nonetheless, these value-enhancing effects of banks may not be fully attained without a careful consideration of banks’ capital structure. Within the capital structure-firm performance (CSFP) nexus, literature argues that, over-reliance on outside financing may impede the ability of a firm to fight competition and in turn allow its contenders to pursue predacious market strategies (Fosu, 2013; Campello, 2003). For example, in a perfectly competitive (unconcentrated) banking industry characterized by fixed prices of outputs, the more leverage a bank uses relative to its competitors, the higher interest incomes it is expected to generate. This is because in perfectly competitive markets, the success of business entities depends on their abilities to generate more outputs relative to their rivals. Conversely, in an uncompetitive industry, the use of more leverage given a constant demand may allow rivals to predate. Thus, the effect of leverage on performance/efficiency may depend on the degree of competition in the related industry (Campello, 2003, 2006; Kovenock & Phillips, 1997). The literature further argues that, conglomeration or diversification has potential benefits and costs effects on firm performance (Low & Chen, 2004; Berger & Ofek, 1995). For instance, the size of a diversified firm may enable it enjoy greater debt capacity than a firm that focuses on single product lines (Amihud & Lev, 1981; Lewellen, 1971). Further, through diversification, unsystematic risks are reduced and the interest tax shield related to debt utilization is higher. Conversely, a diversified firm has the potential of investing in projects with negative Net Present Values (NPV) due to its accessibility to discretionary cash flows. It may also engage in cross University of Ghana http://ugspace.ug.edu.gh 4 subsidization, where the firm disgorges cash from a well performing business segment to a less performing segment (van Lelyveld & Knot, 2009; Rajan, Servaes & Zingales, 2000) which may be detrimental to its value. These arguments suppose therefore that highly levered firms such as banks could reduce the chances of costly financial distress and bankruptcy (Berger, Hasan & Zhou, 2010) if benefits accrue from diversification. Thus, the performance effect of capital structure may also depend on whether or not the firm in question is a conglomerate or has a subsidiary. Studies on the CSFP nexus have particularly paid attention to listed firms (Antwi, Atta Mills & Zhao, 2012; Cheng, Liu & Chien, 2010; Chakraborty, 2010). It is important, however, to consider the CSFP link among service firms, including banks due to the level of risks associated with banks and their contribution to the overall GDP. Moreover, with the level of competition in the banking industry, which stems from globalization, liberalization (Addison, 2003) and technological advancement (Berger, 2007), it is essential to assess how competition and capital structure jointly affect bank efficiency. The increasing trend toward bank consolidation also necessitates a study into the interaction effects of capital structure and conglomeration/consolidation on the efficiency of banks (Nicoló, Bartholomew, Zaman & Zephirin, 2004). The central aim of this study therefore is to estimate the non-parametric profit efficiency of banks in Ghana, to examine the impact of capital structure on the estimated profit efficiency via a bootstrapped truncated regression (Simar & Wilson, 2007). The study also considers the potential reverse causation between capital structure and profit efficiency, investigates the joint effects of capital structure and competition, and capital structure and conglomeration on the estimated profit efficiency of banks. Policy implications and recommendations are provided. The novelty and contributions of the study are discussed in due course. University of Ghana http://ugspace.ug.edu.gh 5 1.2 Problem Statement Over the past decades, several authors have investigated the CSFP nexus. However, the evidence from empirical studies on this nexus are mixed. For example, whereas some studies report a positive relationship (Fosu, 2013; Kyereboah-Coleman, 2007; Ghosh, Nag & Sirmans, 2000; Champion, 1999; Roden & Lewellen, 1995) others report a negative relationship (Chakraborty, 2010; Booth, Aivazian, Demirguc-Kunt & Maksimovic, 2001; Rajan & Zingales, 1995; Titman & Wessel, 1988) and still others report no significant relationship (Ebaid, 2009). Other studies also report varied results for the different leverage ratios (Arbabiyan & Safari, 2009; Abor, 2007). Arbabiyan and Safari (2009), for instance, report a positive nexus between profitability and short term debt and a negative nexus between profitability and long term debt. The mixed empirical findings on the CSFP link is probably because previous studies employed financial ratios as proxies for performance and did not consider potential reverse causation from performance to capital structure (Margaritis & Psillaki, 2010; Berger & Bonaccorsi di Patti, 2006). Ohene-Asare (2011) argued that although financial ratios make it easier to assess firm performance, they are sometimes difficult to interpret when firms are from different industries (see also Sherman & Gold, 1985). Besides, the number of ratios that can be generated from financial statement data are unlimited and sometimes provide contradictory and confusing results (Paradi & Zhu, 2013). Similarly, Paradi, Yang and Zhu (2011) argued that, in computing each ratio, a single input and output are employed, which may confound the result to an aspect of the firm’s operations. This is because, firms use several inputs and outputs in their operations. Thus, assessing their efficiency necessitates the use of more than a single ratio (Smith, 1990). Furthermore, financial and University of Ghana http://ugspace.ug.edu.gh 6 accounting ratios tacitly assume constant returns to scale which imply that size does not matter (Smith, 1990). But in real market systems like that of banking, where competition is not perfect and market power differs (Ohene-Asare, 2011), financial ratios may not always be applicable and useful. Rather, a technique capable of capturing several inputs and outputs concurrently, such as the Data Envelopment Analysis (DEA) developed by Charnes, Cooper and Rhodes (1978) can be employed to assess relative efficiency/performance. To the best of the author’s knowledge, only Margaritis and Psillaki (2010) have used DEA to evaluate productive efficiency in the CSFP nexus. However, Margaritis and Psillaki (2010) failed to bootstrap the second-stage truncated regression analysis in which the efficiency scores obtained were regressed on capital structure and some other environmental factors. Bootstrapping is a computer intensive method which through replications or resampling, simulates the data generating process and applies the original efficiency estimator to every replicated sample so that the efficiency estimator can imitate the sampling distribution of the original efficiency estimator (de Borger, Kerstens & Staat, 2008; Simar & Wilson, 1998). Simar & Wilson (2007, 2011) proposed a double-bootstrapped truncated regression when undertaking a second-stage regression where the efficiency estimates are regressed on some environmental covariates. These are necessary because the first stage efficiency estimates are serially correlated with the inputs and outputs in an unknown (in a statistical sense) and complicated way This dependency issue, also imply that the stochastic error is correlated with the environmental variables making inferences inconsistent and bias. But, the bootstrapping helps to conduct statistical inferences and correct statistical biases associated with this analysis (Simar & Wilson, 2011, 2007, 2000; Barros & Assaf, 2009) University of Ghana http://ugspace.ug.edu.gh 7 On the issue of reverse causation, Margaritis and Psillaki (2010) and Berger and Bonaccorsi di Patti (2006) argued that, there is a two-way relationship between capital structure and performance. As such, not only does the choice of a firm’s capital structure influence its efficiency but also, the efficiency of a firm influences the choice of capital structure. For instance, when a firm is efficient (that is when it is able to use given resources to generate more outputs or is able to reduce inputs while maintaining the outputs), it is more likely to earn a higher return. This places the firm in a better position to employ more debt since the cost of bankruptcy and financial distress diminishes (Berger & Bonaccorsi di Patti, 2006). In effect, considering just one aspect of the nexus as have been examined by most authors in previous studies (Fosu, 2013; Antwi et al., 2012; Kyereboah- Coleman, 2007; Abor, 2005) violates the assumption of the ‘classical linear regression’ that the explanatory variables should be exogenous (Brooks, 2008 pp. 44) and biases the results they generated. Another limitation of earlier studies is that, they failed to incorporate conglomeration or diversification, which may interact with capital structure to impact on firm performance (Amihud & Lev, 1981; Lewellen, 1971). Theoretical arguments pertaining to diversification propose that, diversification can either enhance or reduce the value of a firm’s performance (Low & Chen, 2004; Berger & Ofek, 1995). One of these value-enhancing effects are the tax benefits associated with the greater debt capacity that the firm enjoys. This seems to suggest that, the value of a diversified firm one way or the other depends on the benefits and costs that it generates from operating as a conglomerate. Owing to these benefits and costs therefore, it is important that, in assessing the performance (specifically, in this case, the profit efficiency) of a diversified firm, the joint effects of conglomeration and leverage be considered. University of Ghana http://ugspace.ug.edu.gh 8 Besides, Chevalier and Scharfstein (1996) argued that leverage constrains a firm’s ability to compete in a highly concentrated market (uncompetitive industry). This is because with leverage, the firm may be compelled to charge a higher price than its rivals, which limits the profits it generates and hence its performance. Conversely, Brander and Lewis (1986) suggested that, leverage enables a firm to compete aggressively in a concentrated product market because leverage offset costly agency problems. These arguments presuppose that competition must be incorporated in analysing the CSFP nexus. However, to the best of the author’s knowledge, with the exception of Fosu (2013) who proxied competition with both Herfindahl Hirschman Index (HHI) and Boone Indicator (BI), none of the earlier studies considered CSFP nexus in the presence of competition. Another essential issue is that, preceding studies employed samples pertaining to different industries. Most of the studies with the exception of Fosu (2013) and Berger and Bonaccorsi di Patti (2006) included firms from different industries despite the differences in firm-specific conditions. Using samples from different industries can make it difficult to choose appropriate proxies (Griffin & Mahon, 1997). This stems from the fact that, each industry may have different regulatory systems, different stakeholders and peculiar attributes that distinguish it from other industries (Rowley & Berman, 2000). In effect, combining these industries would warrant disentangling the individual industry specific effects for accurate results to be generated. Primarily, this study reconciles the irregularities in previous studies by employing profit efficiency as a proxy for the performance of banks in Ghana to complement the standard financial and accounting ratios e.g. Return on Assets (ROA) and Return on Equity (ROE). In addition, the study bootstraps the second-stage regression analysis, which appears to have been neglected in both the banking efficiency literature and the CSFP nexus literature. The study advances prior methodology University of Ghana http://ugspace.ug.edu.gh 9 by considering the likelihood of reverse causation from profit efficiency to capital structure which has also not been examined in Ghana. Furthermore, it considers how competition and capital structure jointly influence the profit efficiency of banks and how conglomeration and capital structure jointly affect the profit efficiency of banks. 1.3 Gaps and Contributions From the problem statement, five (5) gaps have been identified in the literature on banking efficiency and CSFP link. The first, which is a methodological gap is the use of financial ratios in assessing firm performance (Fosu, 2013; Kyereboah-Coleman, 2007; Abor, 2005) as opposed to efficiency measures using DEA. The second gap is the limited use of bootstrapping in the second- stage non-parametric frontier estimations. The third gap is the failure to consider the interaction effect of capital structure and competition on profit efficiency. The fourth gap is the failure to incorporate the interaction effect of capital structure and conglomeration on profit efficiency. The final gap is the lack of consideration of the reverse causation from profit efficiency to capital structure. This study contributes to the existing literature by complementing the financial ratios employed in exploring the CSFP link with frontier efficiency techniques and bootstrapping the second-stage regression to resolve the potential serial correlation. The study also incorporates the interaction effect of conglomeration and capital structure as well as the effect of competition and capital structure on the profit efficiency of banks. Finally, it considers reverse causality of CSFP nexus and provides policy recommendations. University of Ghana http://ugspace.ug.edu.gh 10 1.4 Research Objectives The central objective of this research is to assess the profit efficiency of banks and to investigate the CSFP nexus. The specific research objectives are: a. To evaluate the performance of banks in Ghana using a non-parametric measure of profit efficiency. b. To examine the marginal effect of capital structure on bank performance using both profit efficiency estimates and profitability ratios. c. To investigate the interaction effect of capital structure and competition on the profit efficiency of banks in Ghana. d. To investigate the interaction effect of capital structure and conglomeration on the profit efficiency of banks in Ghana. e. To analyse the effect of profit efficiency on capital structure using two competing hypotheses (the efficiency-risk and franchise value hypotheses). 1.5 Research Questions To achieve the research objectives, the following questions are addressed. a. What are the profit efficiency estimates of banks in Ghana? b. Does capital structure have a significant effect on the profit efficiency and profitability of banks in Ghana? c. To what extent does the impact of capital structure on the profit efficiency of banks in Ghana depend on the degree of competition? University of Ghana http://ugspace.ug.edu.gh 11 d. To what extent does the impact of capital structure on the profit efficiency of banks in Ghana depend on conglomeration? e. What is the bi-directional relationship between profit efficiency and capital structure? 1.6 Significance of the Study There are immense benefits toward research and practice. For starters, this study contributes to the existing body of knowledge on the CSFP nexus in a non-parametric frontier efficiency framework by using profit efficiency scores to proxy firm performance instead of the currently used financial ratios. Furthermore, the study contributes to the existing literature by examining the interaction effect of capital structure and conglomeration as well as that of capital structure and competition on the profit efficiency of banks in Ghana. With respect to practice, the study provides the bank with a more holistic depiction of how performance can be improved based on its leverage decisions. Moreover, in computing the HHI and the BI, management is informed about the competitive nature of the industry and how it affects their performance. The interaction effect of conglomeration and capital structure also informs management on whether conglomeration decisions influence the bank’s performance significantly. 1.7 Limitations of the Study Although, the study contributes significantly to both practice and research, there are a few challenges. One of these challenges pertains to the fact that the study uses only universal banks without including rural banks which causes dimensionality problems. This problem can be mitigated by augmenting the sample to include banks from other African countries. With this, University of Ghana http://ugspace.ug.edu.gh 12 more policy recommendations centered on the African financial landscape can be brought to bear on the analysis. But this is not possible due to data unavailability. Further, the disintegration of profit efficiency into technical and allocative efficiencies would have helped to determine the real source of performance. The study looks at static profit efficiency, but could have been extended to profit productivity which could reveal patterns and trends of performance over time. Finally, the quality of this study to a large extent depends on the reliability and accuracy of the data sourced from the Bank of Ghana (BOG). 1.8 Chapter Disposition The study is divided into six chapters with sub-chapters. Chapter one covers the background of the study area, problem statement, gaps and contributions, the objectives of the study, research questions, the significance of the study and its limitations. Chapter two reviews the scholarly literature and considers the opinion of several authors on the CSFP nexus. In chapter three, an overview of the banking industry since the year 2000 is provided. The methodology of the research is discussed in chapter four. This chapter highlights the study area, source of data, study population and the mode of data analysis. Chapter five entails the data presentation, analysis, discussion and findings. Chapter six consists of the summary, conclusions and recommendations. University of Ghana http://ugspace.ug.edu.gh 13 CHAPTER TWO LITERATURE REVIEW 2.1 Introduction This chapter reviews scholarly articles and books published by researchers on capital structure and firm performance. For starters, it focuses on the various capital structure theories and firm performance, and the hypotheses related to the reverse causation of performance from capital structure. Subsequently, the chapter discusses the literature pertaining to the interaction effects of competition and conglomeration on the CSFP nexus and then empirically reviews studies on the various objectives of this study. Relevant literature on profit efficiency in the banking industry is also reviewed. 2.2 Theoretical Review 2.2.1 Capital Structure Theories and Firm Performance The CSFP nexus has been the focus of considerable debate, both theoretically and empirically. Since the seminal work of Modigliani and Miller (1958) who argued that, in a frictionless world (a perfect market characterized by no transaction costs and taxes and the existence of homogenous expectations), the value of a firm is independent of its capital structure, the emphasis throughout prior literature has been whether there is an optimal capital structure for a firm and whether or not the use of debt relative to equity is relevant for a firm (Hatfield, Cheng & Davidson, 1994). To understand the CSFP link, preceding studies have explored the trade-off, the pecking order and the agency cost theories of capital structure. University of Ghana http://ugspace.ug.edu.gh 14 The classical version of the trade-off theory dates back to Kraus and Litzenberger (1973), Scott (1976) and Kim (1978). This theory is based on the idea that, bankruptcy costs and taxes constitute market deficiencies which are fundamental in determining the effect of leverage (capital structure) on market value. The theory postulates that, there exists an optimal level of capital structure for the maximization of firm value (Frank & Goyal, 2009). At this optimal level, the marginal benefits and costs of debt are equal (Fama & French, 2002). Again, the use of debt is cheaper than the use of equity to finance a firm because interest charges on debt are tax deductible. This notion of tax deductibility implies that, debt-financed firms are capable of increasing their performance or efficiency than their counterparts who are highly equity financed. This is because, the amount to be paid as corporate taxes are lower for levered than for unlevered firms. Similarly, corporate interest deductibility encourages profitable and less volatile firms to use higher leverage suggesting a positive link between firm performance and leverage. However, the excessive use of debt is risky due to the possibility of bankruptcy and its accompanying costs (bankruptcy costs) (Titman, 1984). Bankruptcy costs are the costs incurred when the perceived tendency of default on debt financing is positive (Abor, 2005; Haugen & Senbet, 1978). To ensure that they do not employ debt to a point where they may default on their debt obligations, firms fix a target ratio, which is determined by the trade-off between the advantages (tax deductions) and the costs (bankruptcy) of debt. In effect, the value of a levered firm using the trade-off theory, includes the value of an unlevered firm plus the present value of the advantages related to tax minus the discounted value of bankruptcy and agency costs (Cheng et al, 2010). Another theory of capital structure is the pecking order theory, a concept attributable to Myers (1984) and Myers and Maljuf (1984). The theory contends that, there is no optimal target ratio - a University of Ghana http://ugspace.ug.edu.gh 15 ratio that balances the tax savings on debt with the costs of bankruptcy. Instead, to minimize the existence of information asymmetry (the fact that managers are better informed about the profitability and prospects of the firm than outside investors) and its associated financing costs, there is a hierarchy for firms with regards to financing their investments. For example, with private information, managers are more likely to issue risky securities when they are overpriced than when they are underpriced. Due to this information asymmetry, the issue of new equity to finance new projects by firms are underpriced by the investors (Harris & Raviv, 1991). This is to enable the investors gain more on the project’s value resulting in a loss to existing shareholders. In anticipation of these price discounts by investors, managers may forgo profitable projects where these projects are required to be financed by risky securities leading to underinvestment. In order to avoid this problem of underinvestment, the onus lies on the firm to finance with a security that may not be underpriced by the market such as retained earnings and riskless debt (Baker & Wurgler, 2002). The pecking order theory therefore stipulates that, firms, finance their needs with internal funds (retained earnings) first, then debt if extra funds are required and finally equity to cover any additional capital requirements (Myers & Maljuf, 1984). This does not mean the theory denies the significance of taxes and bankruptcy costs in the choice of capital structure. It does suggest however that, managers consider taxes and bankruptcy costs less important than their dispositions for internal over external funds and debt over new issues of stock in financing their operations (Brealey, Myers & Marcus, 2001). Thus, profitable firms use more internal funds in financing their operations, suggesting a negative link between capital structure and firm performance. University of Ghana http://ugspace.ug.edu.gh 16 Similarly, with the existence of information asymmetry between investors and managers, the relative costs of different sources of finance vary. For example, with retained earnings where the fund provider is the firm, and hence has an advantage of acquiring available information relative to outside investors, these investors would prefer higher return rates on their investments (Abor, 2005). This implies that, the costs of the various financing options to the firm depend on the hierarchy of its financing preferences. To avoid paying higher rates of returns to investors therefore, it behoves on a firm to make appropriate decisions regarding its capital structure in order to be more profitable and efficient. Based on the theory of asymmetric information, Ross (1977) introduced the signalling effect theory. The theory postulates that managers are aware of the return distribution of their firms than outside investors. In view of this managers are advantaged when security issues are highly priced by the market but are punished if the firm goes bankrupt. The theory also presumes that, participants in the market interpret higher levels of debt as signals of higher future cash flows and quality of the firm. Thus, an efficient firm which is capable of avoiding bankruptcy costs will resort to more debt than a less efficient firm (Barclay, Smith & Watts, 1995). Building on the early works of Fama and Miller (1972), Jensen and Meckling (1976) initiated the agency cost theory. The agency theory postulates that there is a conflict of interest between shareholders (principals) and managers (agents) of firms. Managers have a tendency to waste free cash flow. Thus, the greater the amount of discretionary funds available to a manager, the higher the likelihood of empire building (Jensen, 1986). This implies that, managers are more likely to increase the scale of their firms even if that means reducing the value of the firm or engaging in poor projects. To mitigate this problem of overinvestment, debt financing is used as a control University of Ghana http://ugspace.ug.edu.gh 17 mechanism and thus assumes a positive nexus between leverage (capital structure) and firm performance. The costs associated with agency problems also exist between shareholders and debt holders. This conflict primarily arises from the risks of default associated with employing higher leverage which in turn creates the underinvestment problem hypothesized by Myers (1977). Shareholders prefer higher risks since these risks translate into higher returns. On the other hand, higher risks are detrimental to debt holders who are the residual claimants of returns in bad states due to the limited liability nature of firms (Brander & Lewis, 1986). Hence, in situations where a firm is already levered, debt holders may require higher rates of return on their investments. This decision upsurges the cost of borrowing of the firm and reduces its efficiency/value. 2.2.2 Hypothesis on Reverse Causality from Firm Performance to Capital Structure The CSFP linkage literature has argued that there exists a bi-directional causal relationship between leverage and firm performance (Harvey, Lins & Roper, 2004; Demsetz & Villalonga, 2001; Rajan & Zingales, 1995). On one hand, the amount of leverage employed by a firm determines how well it would perform. On the other hand, the performance of the firm can determine the proportion of leverage that the firm would employ in financing its operations. Put simply, the degree of a firm’s efficiency may place it in a better position to replace equity with debt. This leads to the efficiency-risk and franchise value hypotheses of the reverse causation of performance from capital structure introduced by Berger and Bonaccorsi di Patti (2006). The efficiency-risk hypothesis states that, an efficient firm would employ higher leverage in its operations than a less efficient firm. This is because higher efficiency reduces the cost of financial University of Ghana http://ugspace.ug.edu.gh 18 distress and bankruptcy. The efficiency-risk hypothesis is therefore a byproduct of the trade-off theory in that, the dissimilarities in the levels of efficiency allow the optimal capital structure of a firm to be altered (Berger& Bonaccorsi di Patti, 2006). The franchise value hypothesis presumes that, in the attempt to protect the economic rents associated with higher efficiencies from the threat of insolvency, firms that anticipate high rates of efficiency into the future select lower debt ratios (Berger & Bonaccorsi di Patti, 2006; Demsetz, Saidenberg & Strahan, 1996; Demsetz, 1973). 2.2.3 Capital Structure, Competition and Firm Performance Theoretical extensions of the CSFP relationship propose that over-reliance on outside financing may prevent a firm from competing well with its contenders who will take advantage of that to pursue predacious market strategies since they may have lesser costs to repay (Dasgupta & Titman, 1998; Chevalier & Scharfstein, 1996; Bolton & Scharfstein, 1990; Telser, 1966). This suggests that, the effect of leverage on performance should be conditioned on the extent of competition in the related industry (Fosu, 2013). Brander and Lewis (1986) argue that, given the limited liability nature of firms, which makes debt holders the residual claimants of returns, the use of debt enables a firm to compete aggressively in a concentrated (uncompetitive) product market. This in turn mitigates the agency conflicts that exist between shareholders and managers and improves the performance of the firm. At the same time, the use of debt, which induces firms to produce more may improve the performance of the firm in a Cournot environment (Bolton & Scharfstein, 1990). In view of this, the effect of leverage on firm performance may not be direct. It may depend on the level of competition in the related industry. Chevalier and Scharfstein (1996) also note that leverage restricts the ability of a firm to invest in market shares due to the risk of default. In the case of a concentrated market for example, this University of Ghana http://ugspace.ug.edu.gh 19 restriction may allow rivals to predate on the leveraged firm. Consistently, firms that are highly leveraged charge higher prices than the less leveraged ones during a downturn. In this regard, leveraged firms are anticipated to be underprivileged in terms of competition in concentrated or uncompetitive industries. In another development, Opler and Titman (1994) suggest that highly leveraged firms lose market shares to their less leveraged counterparts when there is a dip in the industry and this lost is intense for firms in concentrated product markets. Further, the debt contracts associated with debt financing create opportunities for other firms to have a competitive advantage over the leveraged firms (Fosu, 2013). This is because, the use of debt obliges the firm to make periodic interest payments to the creditors. This has the tendency of reducing current period profits, which enable rivals to predate on the leveraged firm for the period in which positive benefits accrue to the rival firm (Bolton & Scharfstein, 1990). For example, in a highly concentrated industry characterized by lower competition, it is expected that the rival firms would predate more on the leveraged firm than they would if the leveraged firm were operating in an unconcentrated industry. In effect, the magnitude of losses that accrues to the leveraged firm decreases with the degree of competitiveness in the industry (Fosu, 2013). Put differently, the performance effect of leverage for a firm may depend on the interaction between leverage and the extent of competition in the firm’s industry. Thus, it is essential to consider these interactions when examining the link between capital structure and firm performance. 2.2.4 Conglomeration and Firm Performance The concept of conglomeration and its effect on firm performance has also been discussed in the banking and finance literature. Conglomeration in general can be traced to the emergence of environmental factors such as technology (which has reduced information and telecommunication University of Ghana http://ugspace.ug.edu.gh 20 costs) and deregulation in the modern era (Nicoló et al., 2004). The motives for conglomeration in firms have been predicted to include diversification and economies of scope (Chronopoulos Girardone & Nankervis, 2011). Theoretical arguments stipulate that, conglomeration or diversification have both value-maximizing and reducing effects (van Lelyveld & Knot, 2009; Low & Chen, 2004; Berger & Ofek, 1995). Among the arguments for the value maximization effect of conglomeration include revenue and cost economies of scope (Chandler, 1977), lower tax burdens (Lewellen, 1971), efficient internal markets and better control and monitoring of capital expenditures (Schmid & Walter, 2009). Lewellen (1971) arguing in support of the value maximizing capability of conglomeration suggests that, diversified firms enjoy greater debt capacity than their counterparts that focus on single product lines. This is because, a byproduct of greater debt is the interest tax shield that the firm can enjoy. In the same line, Weston (1970) and Chandler (1977) argue that, with conglomeration, firms enjoy economies of scale which make them efficient in their operations. Less optimistic theories followed by Berger and Ofek (1995) and Rajan et al. (2000) contend that conglomeration is associated with significant reductions in firm value. This originates from the fact that, conglomerates are more likely to support investments in segments with poor growth opportunities. To bolster this argument, Meyer, Milgrom and Roberts (1992) suggest that a firm that focuses on a single product line produces lower losses as compared to diversified firms because they are less likely to invest in negative cash flow projects. These debates presuppose that, diversification or conglomeration relative to focus has the tendency of influencing firm performance. In effect, when examining the relationship between capital structure and firm University of Ghana http://ugspace.ug.edu.gh 21 performance, it is imperative to consider the joint effect of conglomeration and capital structure on the performance of a firm where the firm in question engages in multiple product lines. 2.3 Empirical Review 2.3.1 Capital Structure Theories and Firm Performance Empirical studies on the CSFP nexus have provided mixed and murky results. There is no consensus on the direction of association as empirical evidence have pointed towards positive (trade-off theory), negative (pecking theory), and no relationship (irrelevance theory). For example, whereas Fosu (2013), Margaritis and Psillaki (2010) and Berger and Bonaccorsi di Patti (2006), report a positive relationship, Chakraborty (2010) and Chung et al., (2013) document a negative relationship and Ebaid (2009) report no significant relationship. These differences in empirical reasoning seem to be influenced by the sample selected, the methodological approach used and most especially the context of the study. As regards the sample selected as a reason for the differing results established in previous studies, emphasis would be placed on the papers by Shyam and Myers (1999), Frank and Goyal (2003) and Lemmon and Zender (2010). By employing a smaller sample (157 firms), Shyam and Myers (1999) in testing the trade-off theory against the pecking order model of capital structure in the United States, over the 1971-1989 period, found support for the pecking order theory. This means that, due to the costs associated with the various financing preferences, the firms in their sample resorted to retained earnings as the primary source of financing before considering debt and then equity. Thus, a negative relationship was established between capital structure and profitability. University of Ghana http://ugspace.ug.edu.gh 22 Contrary to this view, Fama and French (2002) and Frank and Goyal (2003) established that, the pecking order theory did not persist when a larger sample and a longer time series was employed. For example, using publicly traded U.S firms which included the sample employed by Shyam and Myers from 1971 to 1998, Frank and Goyal, documented that, high and small growth firms, which were more likely to face potential information asymmetry problems, were the primary issuers of equity whereas, large and mature firms, which were less likely to face problems of information asymmetry, issued debt which contradicts the predictions of the pecking order theory. However, Lemmon and Zender (2010) suggested that, the issue of equity by the small and high growth firms does not necessarily contradict the pecking order theory. Rather, the issue of debt or equity by a firm depends largely on its debt capacity constraints which are mostly motivated by demand and supply considerations (Holmstrom & Tirole, 1997). In view of this, they assessed the impact of debt capacity on the trade-off and pecking order theories between 1971 and 2001 and concluded that the small and high growth firms issued equity because they were constrained in their debt capacities. Following these arguments and focusing on banks in Ghana, it is expected that less equity relative to debt will be used by profitable banks. In this case, the more profitable the banks are, the more debt they are expected to use in their operations which follows the lines of the trade-off and the agency cost theories. Prior studies also seem to suggest that, the mixed empirical findings on CSFP link are attributable to the proxy for performance. Employing a different measure of performance (profit efficiency) as opposed to the frequently used financial ratios, ROE, ROA and Tobin’s q, Berger and Bonaccorsi di Patti (2006) assessed the CSFP nexus of 7,548 commercial banks in the U.S from 1990 through 1995 and found a positive relationship between leverage and profit efficiency which follows the University of Ghana http://ugspace.ug.edu.gh 23 agency cost and the trade-off theories of capital structure. Similarly, Margaritis and Psillaki (2010) employed technical efficiency to examine the CSFP link of French manufacturing firms from 2002 to 2005 and found higher leverage to be positively associated with improved efficiency which follows the agency cost and trade-off theories. These consistencies signal that, studies which employed efficiency measures as proxies for performance obtained similar results as opposed to those that used financial ratios. This may be because, different ratios can be fashioned out of a firm’s financial statement data and these ratios tend to be contradictory and undistinguishable. Also, those performance ratios examine only an aspect of a firm, contrary to the multifaceted nature of firms (Paradi et al., 2013) which require the use of a composite index to take simultaneous account of all resources (Thanassoulis, Boussofiane & Dyson, 1996) when assessing performance. Again, the use of leverage by a firm affects all aspects of the firm’s operations. Thus, examining the relationship between performance measures, like ratios, that focus on just a unit of a firm’s activities against leverage which affects the overall firm may not produce concrete results. To resolve these anomalies therefore, this study adopts a non-parametric profit efficiency (which would be discussed in subsection 2.3.5) measure to assess the relative efficiency of banks in Ghana. Thus, the study comes closest to the studies by Berger and Bonaccorsi di Patti (2006) but employs a non-parametric profit efficiency measure as opposed to the parametric one used by these authors. The studies reviewed also reveal that, the varied results established in the CSFP link may be due to the context in which the research was undertaken. It must be noted that, this research interest has not only focused on foreign countries. There has been quite a number of studies on CSFP link purposely targeting African countries, including Abor (2005, 2007), Kyereboah-Coleman (2007), Ebaid (2009), and Fosu (2013). Starting with Abor (2005), the impact of capital structure on the University of Ghana http://ugspace.ug.edu.gh 24 profitability of 22 listed firms on the Ghana Stock Exchange was assessed over a 5 year period from 1998 to 2002. He concluded that, short term and total debt were positively related to profitability whereas long term debt was negatively related to profitability. Contrary to his view, Kyereboah-Coleman (2007) documented an insignificantly positive relationship between long term debt and performance, but consistent with Abor (2005), found a positive relationship between performance and both short term debt and total debt for microfinance institutions over the years 1995 to 2004 in Ghana. The differences in their conclusions may be attributable to the samples employed. Whereas Abor concentrated on non-financial listed firms, Kyereboah-Coleman focused on some unlisted financial firms suggesting that, the reaction of long term leverage to performance may differ across sectors. Ebaid (2009), on his part, used three performance measures, ROA, ROE and Gross profit margin (GM) to evaluate the effect of capital structure on the performance of 64 non-financial listed firms in Egypt between 1995 and 2005. The author reported a negative relationship between short term debt and ROA and no relationship between long term debt and ROA. He also documented no relationship between both short term and long term debt and ROE and GM. In tandem with Kyereboah-Coleman, long term debt follows the Modigliani and Miller irrelevance theory of capital structure. Unlike the Ghanaian situation, there are clear differences in the Egyptian case, especially for the short term debt. Due to these inconsistencies, Abor’s (2007) paper provides a better conceptual link for differences in capital structure between countries. By assessing the reaction of the performance of Small and Medium Scale Enterprises (SMEs) in Ghana and South Africa to leverage, a positive nexus for short term debt in Ghana and negative nexus for short term debt in South Africa was found. However, for long term debt, both Ghana and South Africa experienced a negative elasticity to performance. University of Ghana http://ugspace.ug.edu.gh 25 Empirical findings are clearly mixed probably due to the nature of the samples used or the different proxy measures used for performance. Most of these studies used heterogeneous samples making it difficult to consistently and accurately measure the effect of leverage on performance in a particular industry. It is therefore important to consider this effect in a specific industry such as the banking industry, within which firms are homogenous in behavior, a gap this study fills in the African context. 2.3.2 Reverse Causation from Performance to Capital Structure Although a study has incorporated the reverse causation between performance and capital structure in econometric modelling (Fosu, 2013), to the best of the author’s knowledge, only three studies stand out as they directly test this theoretical relationship (Margaritis & Psillaki, 2010, 2007; Berger & Bonaccorsi di Patti, 2006). Berger and Bonaccorsi di Patti (2006) tested the efficiency- risk and franchise value hypotheses on the U.S banking industry. On the other hand, Margaritis and Psillaki (2007, 2010) tested the same hypotheses on firms in New Zealand and France respectively. Berger and Bonaccorsi di Patti found that, in the U.S banking industry, none of these hypotheses dominated themselves, implying, efficiency presents only an infinitesimal effect on leverage. This is consistent with findings from New Zealand firms which also revealed that, both the efficiency-risk and franchise value hypotheses operate (Margaritis & Psillaki, 2007). Conversely, for France, it was established that, the market is dominated by the efficiency-risk hypothesis. This suggests that, for French manufacturing companies, the higher the efficiency of the firm, the higher the leverage. As yet, only these three papers have purposely and empirically tested these two hypotheses. Other researchers only incorporate econometric assumptions that nullify the effect of this perceived reverse causality. For example, Fosu (2013) used a lag structure University of Ghana http://ugspace.ug.edu.gh 26 and Generalized Method of Moments (GMM) to deal with this bi-directional causality. To complement these studies, this research directly test both hypotheses on the Ghanaian banking industry. This is to ensure that the effect of leverage on the performance of banks in Ghana is not confounded by the effect of performance on leverage leading to simultaneous-equation bias. 2.3.3 Capital Structure, Competition and Firm Performance The interaction effects of capital structure and competition on the performance of firms have been explored in literature in both developed and developing economic context. However, this seems skewed towards the developed economies. For example, whereas Campello (2003), Dasgupta and Titman (1998), Kovenock and Phillips (1997), Chevalier (1995) and Opler and Titman (1994) focused on developed economies, only Fosu (2013) focused on a developing economy from papers reviewed. Empirical evidence on the relationship between performance and the joint effects of competition and capital structure has tended to be different for developing and developed economies. These mixed findings can be attributed to the different competition measures employed in the studies. Employing four-concentration based ratios as measures of competition, Opler and Titman (1994) established that, firms which are highly leveraged lose market shares to their less leveraged competitors in industries experiencing economic distress in the U.S. Similar analysis was undertaken by Chevalier (1995). Employing different measures of competition, HHI and BI, Fosu (2013) found that, leverage in the presence of competition positively affected 257 South African listed firms. These reviews appear to suggest that, the direction of the relationship between performance and the joint effects of leverage and capital structure may depend on the proxy employed for competition. In this regard, this study employs the HHI and the BI in examining the relationship between profit University of Ghana http://ugspace.ug.edu.gh 27 efficiency unlike profitability ratios and the joint effects of competition and capital structure in a homogeneous industry (the banking industry of Ghana) which is distinct from the studies by Fosu (2013). 2.3.4 Capital Structure, Conglomeration and Firm Performance One of the predominant concerns in the finance discussion on conglomeration as used in this study or diversification as used in prior studies has been whether diversification creates or destroys value (Schmid & Walter, 2009; Laeven & Levine, 2007; Menendez- Alonso, 2003; Berger & Ofek, 1995; Comment & Jarrell, 1995; Lang & Stulz, 1994). As yet, studies on the effects of diversification on firm value or performance have produced varied results and as well, have particularly been geared towards the developed economies. In relation to the varying results for example, whereas some authors report a negative relationship (Berger & Ofek, 1995; Lang & Stulz, 1994), others report a positive relationship (Hubbard & Palia, 1999; Lewellen, 1971). Emphasizing on the negative relationship, some authors are of the opinion that, the discount experienced by firms is not caused by diversification. It may be as a result of endogeneity problems, selection biases, measurement errors and most importantly, the fact that most of the firms that diversify are already discounted firms (Graham, Lemmon &Wolf, 2002; Lamont & Polk, 2002; Whited, 2001; Fluck & Lynch, 1999). Schmid and Walter (2009) in assessing the effect of diversification on the financial services industry in U.S recognized that, financial conglomeration destroys firm value except for investment banking. Similar negative nexus was documented by Laeven and Levine (2007) in a study of 43 countries, of which 3 were developing countries (South Africa, Kenya and Egypt). These studies on the financial sector did not explicitly address the causal factors underlying the University of Ghana http://ugspace.ug.edu.gh 28 diversification discount although their results were consistent with theories on diversification. Along the same lines, but focusing on non-financial firms in U.S., Berger and Ofek (1995) documented a diversification discount (a fall in firm value) of between 13% and 15%. Equally, Lins and Servaes (1999) found a diversification discount of 15% in U.K, 10% in Japan and no discount in Germany. The discount of 15% realized in both U.S and U.K was attributed to the common ownership structure (dispersed ownership structure) in these countries. Although Lins and Servaes (1999) did not explicitly identify the cause of the value loss, Berger and Ofek (1995) accredited the value loss in U.S to overinvestment and cross subsidization. On the positive effects of diversification on firm value, Hubbard and Palia (1999) found that financially unconstrained firms that acquired firms that were constrained generated positive returns in a study of some firms in the U.S. Simply put, firms that did not have any debt obligations at the time of acquisition ended up bearing the burden of the debts of the firms they acquired. However, the positive returns gained could be attributable to tax advantages they enjoyed due to the additional debts in their capital structure. Hadlock, Ryngaert and Thomas (2001) also found that the reaction of the market on average to equity issue announcement was less negative for diversified than it was for focused firms which confirm the value-enhancing effects of diversification. This was so probably because; diversified firms come with economies of scope, especially if their product lines are uncorrelated. The net effect is that, investors seem more interested in benefiting from such scope advantages than investing in a focused firm, especially in such volatile markets. Also, Lewellen (1971) and Amihud and Lev (1981) suggested that with diversification, firms enjoy greater debt capacity than their counterparts who focus on single product lines. With this, they University of Ghana http://ugspace.ug.edu.gh 29 benefit from lower tax burdens which may translate into higher performance. Thus the performance effect of leverage (capital structure) depends on whether the firm is diversified or focused which suggests an interaction between capital structure and diversification. From the review, it seems that the diverse findings documented originate from the opposing theoretical underpinnings of diversification. This implies that, where diversification translates into positive effects such as economies of scope, tax benefits and better monitoring for example, its effect on firm performance would be positive. On the other hand, where diversification informs such decisions as overinvestment and cross subsidization, it is expected that its effect on firm performance would be negative. It can therefore be concluded that, although a firm may experience both the costs and benefits associated with diversification, the true value of the firm will depend on the dominating effect. In this regard, it is necessary that, the positive or negative effects that emanate from conglomeration be interacted with capital structure in assessing performance where the firm under consideration has other segments performing other activities beside its core or traditional activities, a gap this study fills. 2.3.5 Relevant Literature on Profit Efficiency in Banking Substantial research has been conducted on the efficiency of financial institutions, particularly banks over the years. In the banking sector, 3 survey papers exist (Fethi & Pasiouras, 2010; Berger, 2007; Berger & Humphrey, 1997). The efficiency of these financial institutions have been analysed in terms of technical, cost, revenue and profit (Berger & Humphrey, 1997). Of these methods, profit efficiency has been thought to be the ultimate efficiency measure since it constitutes an essential source of information for bank management than the partial assessment offered by cost and technical efficiency analysis (Ariff & Can, 2008; Maudos, Pastor, Perez & Quesada, 2002). A University of Ghana http://ugspace.ug.edu.gh 30 survey by Berger and Humphrey (1997) indicated that, of the 130 studies conducted in 21 countries, only 14 were on profit and or revenue efficiencies. This implies that empirically, there are very limited profit and revenue efficiency studies than there are for cost and technical efficiency studies (Fethi & Pasiouras, 2010). One reason for the limited studies on profit efficiency is the difficulty in obtaining reliable and transparent information on output price which is required in the analysis of standard profit efficiency (Fethi & Pasiouras, 2010). Studies on profit efficiency in the banking industry have particularly focused on U.S (Fare, Grosskopf & Weber, 2004; Akhigbe & McNulty, 2003; Clark & Siems, 2002; Berger & DeYoung, 2001; Rogers, 1998; Berger & Mester, 1997), India (Ray & Das, 2010; Das & Ghosh, 2009), Spain (Maudos & Pastor, 2003; Vivas, 1997), China (Berger, Hasan & Zhou, 2010; Ariff & Can, 2008), Greece (Delis, Koutsomanoli-Fillipaki, Staikouras & Katerina, 2009), Turkey (Isik & Hassan, 2002) and Europe (Chronopoulos et al, 2011; Kasman & Yildirm, 2006; Bonin, Hasan & Wachtel, 2005; Bos & Schmiedel, 2003; Maudos et al., 2002; Vander Vennet, 2002). These studies have examined profit efficiency using parametric and non-parametric approaches. For example, whereas Berger and Mester (1997), Rogers (1998), Berger and DeYoung (2001), Clark and Siems (2002), Akhigbe and McNulty (2003) employed parametric approaches, Maudos and Pastor (2003), Fare et al. (2004), Ariff and Can (2008), Das and Ghosh (2009), Ray and Das (2010) and Chronopoulos et al. (2011) applied non-parametric approaches. Adopting a parametric technique (the thick frontier approach) and focusing on 54 Spanish savings banks, Vivas (1997) documented a decrease in profit inefficiency by 40% from an average of 28% between the years 1986 and 1991 (inefficiency fell from 32% in 1986 to 19% in 1991). This was a period in which the Spanish banking industry was experiencing considerable deregulation. University of Ghana http://ugspace.ug.edu.gh 31 Extending the work of DeYoung (1994) and employing the distribution free approach, Rogers (1998) estimated the cost, revenue and profit efficiency of over 10,000 Commercial banks in U.S from the year 1991 to 1995. He concluded that, standard models that omitted nontraditional activities as outputs in modelling bank profit efficiency understated these efficiencies. Contrary to this view, Clark and Siems (2002) concluded that profit efficiency remained unaffected with the inclusion of off-balance sheet items in their study of some publicly trading commercial banks in U.S from 1992 to1997. They analysed profit efficiency using the distribution free approach adopted by Rogers in addition to Stochastic Frontier Analysis. Employing a non-parametric approach DEA, Maudos and Pastor (2003) estimated the profit and cost efficiency of commercial and savings banks in Spain between 1985 and 1996. The results obtained showed that, profit efficiency levels (52.9% for commercial banks against 34.7% for savings banks) were below those of cost efficiency (90.9 % for commercial banks and 80.2 % for savings banks). Using the same technique (DEA) but focusing on banks in India, Ray and Das (2010) documented lower profit efficiency levels relative to cost efficiency for 71 commercial banks studied between 1997 and 2003 which is consistent with the findings by Maudos and Pastor. In a broader setup, Maudos et al. (2002) estimated the profit and cost efficiency of commercial banks in 10 countries of the European Union (EU) between 1993 and 1996. They noted that, the profit efficiency of the banks were lower than their cost efficiency. Similar results were documented by Bonin et al. (2005) in their analysis of commercial banks of 11 countries of the EU between 1996 and 2000, Kasman and Yildirim (2006) who analysed 8 countries of the EU between 1995 and 2002 and Chronopoulos et al. (2011) in a research conducted on 10 countries of the EU between 2001 and 2007. The difference between these cited papers is that two applied parametric University of Ghana http://ugspace.ug.edu.gh 32 techniques (Kasman & Yildirim, 2006; Bonin et al., 2005) and the other two (Maudos et al., 2002; Chronopoulos et al., 2011) used the non-parametric DEA technique. Previous studies have also explored the link between conglomeration / diversification and profit efficiency. A general finding of these studies is that diversified banks are more profit efficient than their counterparts who focus on single product lines (Chronopoulos et al., 2011; Casu & Girardone, 2004; Vander Vennet, 2002). Overall, although studies have measured the profit efficiency of banks in advanced countries, no study have been conducted on the profit efficiency of banks in Africa using DEA. This study fills this gap in the literature by estimating the profit efficiencies of banks in Ghana using a non-parametric approach. 2.3.6 Banking Efficiency Studies in Ghana A number of researches have also been conducted on the efficiency of banks in Ghana. Whereas some authors employed the nonparametric DEA technique (Saka, Aboagye & Gemegah, 2012; Adjei- Frimpong & Gan, 2014), others used the parametric SFA technique (Alhassan, 2015; Bokpin, 2013). Saka et al. (2012) assessed the impact of foreign bank entry and bank concentration on the technical efficiencies of 23 banks in Ghana from 2000 to 2008 and found that the entry of foreign banks positively influenced the efficiency of domestic banks. Bokpin (2013) also examined the impact of ownership structure, and corporate governance on the cost and profit efficiencies of 25 banks in Ghana over the period 1997 to 2007 and established that banks with inside ownership are cost and profit inefficient and foreign banks are more efficient in terms of cost than their local/domestic counterparts. University of Ghana http://ugspace.ug.edu.gh 33 Using an unbalanced panel data from 2001 to 2010, Adjei- Frimpong and Gan (2014) estimated the cost efficiencies of 25 banks in Ghana and examined the impact of capitalization, size, loan loss provision, inflation rate and GDP growth rate on the estimated efficiency scores. The findings of the study revealed that banks in Ghana operated farther away from the cost efficiency frontier. In a recent study however, Alhassan (2015) found banks in Ghana to exhibit high levels of efficiency in cost in a study of 26 banks from 2003 to 2011 which contradicts the study by Adjei- Frimpong et al. (2014). This is probably because the authors employed different techniques and time horizons in their estimations of cost efficiency in Ghana’s banking industry. 2.4 Conceptual Framework Based on theoretical predictions, preceding empirical evidence and the Ghanaian banking context, six (6) main testable hypotheses are formulated. Stulz (1990) and Jensen (1986) developed a model that envisages that, debt has both positive and negative effects on firm performance. Following this model, both effects of debt would be allowed Profit Efficiency Efficiency Competition Comp*Lev Cong*Lev + + - + + Leverage Conglomeration University of Ghana http://ugspace.ug.edu.gh 34 in specifying the model for capital structure and profit efficiency. It is expected, however, that, the effect of leverage on the profit efficiency of banks in Ghana is positive as represented by H1. This is probably because, the regulatory environment within which banks operate in Ghana require banks to be solvent at all times in order to meet the needs of customers. In this regard, banks in Ghana are less likely to use leverage to the extent that it will cause them to be bankrupt. H1: Leverage has a positive effect on profit efficiency Over the years, the banking industry of Ghana has been characterized by certain structural and regulatory reforms. These reforms include, the introduction of the Universal Banking License in 2003, the abolishment of the secondary reserve requirement in 2006, the redenomination of the local currency ‘the Cedi’ to ‘Ghana Cedi’ in 2007, the increase in minimum capital requirements from 60 million to 120 million Ghana cedis in 2013 (Ghana Banking Survey, 2008), and the consistent entry of foreign banks from 2005 to 2011(Akomea & Adusei, 2013). Following these reforms, it is expected that, the industry is highly competitive. This competitive ambience may result in lower charges on loans which may reduce the profit efficiency. In view of this, it is hypothesized that: H2: Competition has a negative effect on profit efficiency Extant literature has suggested that, the use of debt allows rival firms to predate on the leveraged firm (Opler & Titman, 1994; Campello, 2003). The intensity of the predation depends on the level of competition in the related industry. In a concentrated industry where competition is usually low, leverage firms are less susceptible to rivalry predation than when they operate in unconcentrated industries (Fosu, 2013). The banking industry in Ghana is expected to be competitive University of Ghana http://ugspace.ug.edu.gh 35 (unconcentrated). In view of this, banks that use more leverage relative to their counterparts are expected to be less vulnerable to rivalry predation in Ghana. This is because in competitive environments where prices are usually fixed, profits emanate from the amount of output that the business entity is able to produce. In view of this, the performance effect of the interaction between leverage and competition is expected to be positive. H3: The interaction effect of leverage and competition on profit efficiency is positive and significant Theories on conglomeration suggest that diversification has both benefits and costs on firm performance. To the extent that the regulatory environment requires banks to be solvent, banks are less likely to expand their product lines or engage in other activities besides those stipulated in the Banking License if the resultant effect of this diversification is reduction in profits. Hence a positive effect of conglomeration on profit efficiency is expected. H4: Conglomeration has a positive effect on profit efficiency Further, the expected positive effect of conglomeration on profit efficiency suggests that when conglomeration is interacted with leverage, the overall effect on profit efficiency is likely to be positive. H5: The interaction effect of conglomeration and leverage on profit efficiency is positive Berger and Bonaccorsi di Patti (2006) postulate that there is a reverse causation from performance to capital structure. They tested this reverse causality using two hypotheses (the efficiency-risk and the franchise value hypotheses). Following these authors, hypotheses six is formulated. University of Ghana http://ugspace.ug.edu.gh 36 H6a: More efficient banks employ higher leverage in their operations (efficiency-risk) H6b: More efficient banks employ lower leverage in their operations (franchise value) 2.5 Chapter Summary This chapter provided a review (both theoretical and empirical) of the theories underlying capital structure and firm performance, the hypotheses related to the reverse causation between capital structure and performance, the theories supporting the interaction between capital structure and competition on one hand and the interaction between capital structure and conglomeration on the other. Relevant literature on profit efficiency in banking studies was also reviewed. Based on these, a conceptual framework from which certain hypotheses were developed was provided to serve as a guide for the analysis. University of Ghana http://ugspace.ug.edu.gh 37 CHAPTER THREE THE BANKING INDUSTRY IN GHANA 3.1 Introduction This chapter provides an overview of the Ghanaian banking industry. It provides a historical background on the industry and highlights the key financial improvements and regulatory reforms introduced to enhance competition, financial conglomeration and efficiency. This is to ensure a better understanding of the arguments presented in this study. 3.2 Historical Background Banking activities in Ghana commenced in 1896 with the establishment of the British Bank of West Africa (Amidu, 2007). The main aim of the bank was to deliver principal services of lending and borrowing of money. In 1918, another bank, the Colonial bank commenced operation, but merged with other foreign banks under the leadership and name Barclays Group to compete with the British Bank of West Africa (Gatsi, 2012). In 1953, the Bank of Gold Coast emerged and was later divided into Ghana Commercial Bank and BOG. In 1957, the BOG assumed the role of managing the currency in existence. This led to the issuance of the first currency, the Cedi, to substitute the old West African currency notes in 1965 (BOG). The introduction of a new government in 1957 saw the establishment of the Ghana Investment Bank, the Merchant Bank, the Agricultural Development Bank and the Social Security Bank. In 1983, the economy experienced economic difficulties which led the government to seek assistance from the World Bank and the International Monetary Fund (IMF). Subsequently, the Economic Recovery Program (ERP) was launched. The program led to the removal of financial restrictions, University of Ghana http://ugspace.ug.edu.gh 38 divestiture of government interest in Public Corporations and the liberalization of trade (Sowa, 2003). As part of the ERP, the Financial Sector Adjustment Program (FINSAP) was enacted in 1988 with the aim of removing financial restrictions, introducing new capital, strengthening competition and efficiency within the banking sector (Quartey & Afful-Mensah, 2014; Aboagye, Akoena, Antwi- Asare & Gockel, 2008). In 1989, the Banking Law (PNDL 225) was enacted to enable local entities file applications for licenses to operate as banking institutions (Ghana Banking Law, 1989). Consequently, various corporate bodies were licensed to function as banks, including CAL Merchant Bank, Allied and Metropolitan, Meridien (BIAO) Trust Bank, and Ecobank. Although there had been attempts at enhancing the banking sector through the legislative and regulatory reforms made, the industry did not see much progress. This was evidenced in the liquidation of Ghana Co-operative Bank, Bank for Housing and Construction, and the Bank for Credit and Commerce in 2000 (Amidu, 2007). Since 2000, there has been major developments in the Ghanaian banking industry as depicted in Table 1 to strengthen bank operations. University of Ghana http://ugspace.ug.edu.gh 39 Table 1: Key Developments in the Ghanaian Banking Industry from 2000 to 2013 Year Key Developments 2002 The Bank of Ghana Act (Act 612) was enacted 2002 Introduction of Bank of Ghana Prime Rate as the policy rate 2002 Inauguration of the Monetary Policy Committee 2003 2003 Introduction of Universal Banking License. Banks with capital of GH¢7 million were allowed to undertake any form of banking. Abolishment of maintenance, transactions and transfer fee charges by Commercial banks. 2004 Replacement of the Banking Law 1989 (PNDCL 225) with Banking Act 2004 (Act 673) 2006 Abolishment of 15% secondary deposit reserve requirement. 2006 Enactment of Foreign Exchange Act 2006 (Act 723) and Whistle Blowers Act 2006 (Act 720). 2007 Approval of Credit Reporting Act 2007 (Act 726) and Banking (Amendment) Act 2007 (Act 738). 2007 Abolishment of National Reconstruction Levy 2007 Redenomination of the Cedi (10,000 = 1Ghana Cedi) 2008 Approval of Borrowers and Lenders Act 2008 (Act 773), Non-Bank Financial Institutions Act 2008 (Act 774), Home Mortgage Finance Act 2008 (Act 770) and Anti-Money Laundering Act 2008 (Act 749). 2008 Requirement of a minimum capital of GH¢60 million to maintain Class 1 banking status by the BOG 2008 Requirement for banks to adopt International Financial Reporting Standards 2009 Cheque code-line clearing system was introduced by the Ghana Inter-bank Payment and Settlement Systems (GhIPPs) 2011 Requirement to maintain the statutory reserve requirement of 9% in Ghana cedis only 2013 Increase in the reserve requirement and monetary policy rate to 10% and 16% respectively 2013 Requirement for a minimum stated capital of GH¢120 million for banks. Source: Ghana Banking Survey, 2008 University of Ghana http://ugspace.ug.edu.gh 40 3.2.1 Overview of the Ghanaian Banking Industry The Ghanaian banking industry comprises of the Bank of Ghana as the regulating body, 27 Universal banks and 140 rural and community banks (BOG Annual Report, 2013). With respect to ownership, 15 of the Universal banks are foreign-owned and 12 are domestically-owned. The structure of the banking industry is depicted in Figure 1. 3.2.1.1 Implications of Some Key Developments on the Ghanaian Banking Industry Universal Banking The Universal banking system in Ghana was introduced in 2003. This system allowed banks to undertake development, commercial, investment and/ merchant banking activities without the need for separate licenses. The purpose of this policy was to ensure that banks became versatile in providing services to their customers (Quartey & Afful-Mensah, 2014). To perform these roles of universal banking, the BOG increased the minimum capital requirement to GH¢7 million to ensure that the banks were adequately capitalized. This was to encourage a more competitive and dynamic banking industry capable of intermediating effectively to support growth in the expanding economy (BOG Annual Report, 2004). Since the introduction of this policy, competition in the industry has heightened as is evidenced in the increase in the number of banks from 18 in 2003 to 26 in 2013. Abolishment of the Secondary Reserves Requirement In order to make funds available to the private sector, the BOG abolished the secondary reserve requirement in 2006. The secondary reserve requirement mandated banks to hold a percentage of their deposits in the form of medium term securities (BOG, 2006). With this policy, the ability of University of Ghana http://ugspace.ug.edu.gh 41 the banks to create more loans was limited. The abolishment of this requirement freed up significant liquidity for lending to businesses which has had diverse implications in terms of, for example, profits in the banking industry (Ghana Banking Survey, 2007). Currency Redenomination In an attempt to make Ghana’s economy attractive to both local and foreign investors as well as achieve the convergence criteria set by the West African Monetary Zone, the Bank of Ghana undertook the redenomination (BOG Annual Report, 2007) exercise. The redenomination exercise was also envisaged to increase efficiency in payment systems, banknote processing and improve the accounting records of banks in Ghana. A survey conducted by the research department of the BOG in 2009 (two years after the redenomination) suggested that, the objectives of the redenomination had been achieved. Minimum Capital Requirement The minimum capital requirement was increased from GH¢7 million in 2003 to GH¢60 million in 2008 and GH¢120 million in 2013 (KPMG, 2012; Ghana Banking Survey, 2014). Although, these recapitalization directives had received mixed feelings in the financial circle, the Ghana Banking Survey (2011) report indicate that, the move towards higher minimum capitals improved the liquidity and the operating assets of the industry. University of Ghana http://ugspace.ug.edu.gh 42 Figure 1: Structure of the Banking Industry Figure 1: Structure of the Banking Industry (Source: Bank of Ghana) 1. Access Bank Limited 2. Agricultural Development Bank Limited 3. Bank of Africa Limited 4. Bank of Baroda Limited 5. Barclays Bank Limited 6. BSIC Limited 7. CAL Bank Limited 8. Ecobank Ghana Limited 9. Energy Bank Limited 10. Fidelity Bank Limited 11. First Atlantic Merchant Bank Limited 12. First Capital Plus Bank 13. GCB Bank Limited 14. Guaranty Trust Bank Limited 15. HFC Bank Limited 16. International Commercial Bank Limited 17. National Investment Bank Limited 18. Prudential Bank Limited 19. Royal Bank Limited 20. SG-SSB Bank Limited 21. Stanbic Bank Limited 22. Standard Chartered Bank Limited 23. Unibank Limited 24. United Bank for Africa Limited 25. UT Bank Limited 26. Universal Merchant Bank Limited 27. Zenith Bank Limited Rural and Community Banks University of Ghana http://ugspace.ug.edu.gh 43 3.2.1.2 Performance of the banking industry between 2000 to 2013 Despite increasing competition in the banking industry, it remained a sector with the brightest opportunities for the year 2000 to 2013. Net interest income increased from ¢1.09 trillion in 2000 to ¢2.67 trillion in 2004. Similarly, industry net profit increased from ¢694 billion to ¢1170 billion. As regards capital structure, shareholders’ funds increased by 177% from ¢1.3 trillion to ¢3.6 trillion with total deposits increasing from ¢7.6 trillion to ¢21 trillion (176% increase) (Ghana Banking Survey, 2005). ROE on the other hand, decreased from 53.2% to 32.3%, while ROA decreased from 7.3% to 4.2%. This suggests that although equity funds increased, the utilization of these funds by management in the industry on average was ineffective. In essence, management’s decision on asset composition, liquidity position and cost management was poor. The industry continued to experience reductions in both ROE and ROA in the subsequent years except from 2010 to 2013 where it experienced upturns in its performance. ROE increased from 12.1% in 2009 to 16.7% in 2010, 17.97% (2011), 23.8% (2012), 27.5% (2013). Likewise, ROA increased from 1.6% in 2009 to 2.3% in 2010, 2.4% (2011), 3.5% (2012) and 4.2% (2013). This gain was partly due to the injection of capital by the local banks in their quest to meet the increase in the stated minimum capital by the BOG in 2008. In view of this, funds were made available to finance the operations of the banks in terms of investment and loan portfolios (Ghana Banking Survey, 2011, 2012, 2013, 2014) which increased the profitability of the banks. Figures1 and 2 are representations of the key performance indicators, ROE and ROA of the Ghanaian banking industry from the year 2004 to 2013. University of Ghana http://ugspace.ug.edu.gh 44 This paper argues however that, the ratios (ROA and ROE) do not provide a holistic view of the performance of banks in Ghana since they do not consider the multiple inputs and outputs used by the banks in their operations. It is to this effect that, this study employs profit efficiency, which takes a simultaneous account of the inputs and outputs used by the banks to estimate performance. Figure 2: ROE from 2004-2013 Source: Author (2015) Figure 3 : ROA from 2004-2013 Source: Author (2015) 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 ROE 0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% 4.00% 4.50% 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 ROA University of Ghana http://ugspace.ug.edu.gh 45 Chapter Summary The chapter provided an overview of the Ghanaian banking industry. Overall, the Ghanaian banking industry has undergone major transformations which have had diverse implications on its performance. It experienced downturns in performance in the early and mid-parts of the 2000’s and upturns in the later parts. The number of banks also increased tremendously, especially with the introduction of the universal banking system. This has intensified competition in the industry and has led some banks to diversify into other activities aside their core banking activities. University of Ghana http://ugspace.ug.edu.gh 46 CHAPTER FOUR METHODOLOGY 4.1 Introduction This chapter describes the procedures used to collect and analyse the data in order to answer the research questions. Specifically, it considers the design of the research, the sampling technique, the sources of data and the methods used in analysing the data. 4.2 Research Design The quantitative approach to research is adopted in this study. This is because, it helps to understand the best predictors of outcomes (Creswell, 2012). It also makes it possible to deduce, since the inferences from test of statistical hypotheses lead to general inferences about the features of the population (Harwell, 2011). A panel data methodology is employed in this study. This is because, the study comprises repeated observations on the same cross section of units over time (Wooldridge, 2012). Further, the panel data methodology is more informative, provides more degrees of freedom and less collinearity among the independent variables (Baltagi, 2008). A panel dataset also controls for cross-sectional unobserved heterogeneity (Brooks, 2008) which, when ignored can lead to biased results. 4.3 Sampling and Sources of Data The population for the study constitutes all the 26 banks operating in Ghana. Data covering the years 2000 to 2013 is extracted from the annual reports of the banks under consideration. Since not all these banks were in existence for the entire study period, an unbalanced panel method is University of Ghana http://ugspace.ug.edu.gh 47 used. Primarily, data is sourced from the Banking Supervision Department of the Bank of Ghana and cross-validated with similar data from banks’ annual reports. 4.4 The DEA Methodology Based on the inspiration of Farrell’s (1957) estimation of relative technical efficiency of Decision Making Units (DMUs), DEA was first propounded by Charnes, Cooper and Rhodes (1978) and extended by Banker, Charnes and Cooper (1984). DEA is a non-parametric linear programming optimization-based technique for assessing the relative efficiency of homogeneous DMUs that use multiple inputs to produce multiple outputs (Cook & Seiford, 2009; Fried, Lovell & Schmidt, 2008). The DMUs can be airlines, oil firms, insurance firms, football clubs, universities, hospitals and banks. The technique creates a production or cost or profit frontier which is a “best practice” frontier from observed units. The units forming the frontier become efficient because no other firm dominates them. Then, the generated efficiency or inefficiency scores of those units enveloped by the frontier or interior to it are determined relative to the boundary of the constructed frontier. Due to its powerful ability to optimize, DEA enables management to objectively recognize the best-practice units and the areas of improvement within the firms’ multidimensional operating situations. For a comprehensive discussion on the evolution and details of frontier methods and DEA, the reader is referred to Cooper, Seiford and Zhu (2011), Daraio and Simar (2007), Coelli, Rao, O’Donnell and Battese (2005), Kumbhakar and Lovell (2003) and Førsund and Sarafoglou (2002). For DEA applications in banking see Tzeremes (2015), Staub, da Silva e Souza, and Tabak (2010), Fethi and Pasiouras (2010), Ray and Das (2010), Berger (2007), Berger and Humphrey (1997). University of Ghana http://ugspace.ug.edu.gh 48 A number of reasons account for the use of DEA in this study. For starters, DEA is able to handle multiple inputs and outputs of homogenous DMUs (Charnes et al, 1978). Banks use various inputs such as workers, deposits, fixed assets, etc. and produce various outputs, including investments, loans and advances, securities, etc. The use of DEA will therefore be an appropriate method in such multifaceted operations of a business activity. The technique can also disintegrate efficiency scores into several components, including profit, revenue, cost, technical, pure technical, scale, allocative and mix efficiencies, etc. which help to identify the key sources of inefficiencies of firms. Moreover, with DEA, restrictive functional forms for the production, cost or profit technology or distributional assumptions underlying the observations need not be specified, unlike in some parametric approaches such as Stochastic Frontier Analysis (Fried et al., 2008; Coelli et al., 2005). In other words, DEA allows the ‘data to speak for themselves’ instead of imposing a structure on the data to avoid errors related to specification (Cummins, Weiss, Xie & Zi, 2010). Further, DEA is unit invariant (Lovell & Pastor, 1995) which means that the inputs and outputs that can be used for the DEA analysis do not necessarily have to be in the same metric units. For instance, although labour is measured in man-hours, whilst size is measured by space (square meters) and operating expenses is measured in a currency unit, these can be used as inputs for a particular analysis. DEA can also identify reference sets or peers for each inefficient unit which is imperative for managerial policy making. Despite these merits of DEA, the method is not a panacea for the problems that may be encountered by other performance assessment tools. DEA undoubtedly evaluates the relative efficiency of a DMU, not necessarily absolute efficiency. Hence, the extreme point of expectation to be reached University of Ghana http://ugspace.ug.edu.gh 49 in order to be efficient is determined by the best in the sample, not the ideal best point, implying, that the exclusion and/ inclusion of some DMUs can affect the efficiency estimates from the analysis (Avkiran, 1999). The method is deterministic which makes it prone to identification problems. This emanates from the fact that all deviations from the efficient frontier are attributed to inefficiency without the consideration of statistical noise. These problems have however been addressed through the emergence of bootstrapping (Simar & Wilson, 1998, 2000, 2007) which helps to solve the problem of sampling variations and serial correlations in DEA and second-stage efficiency analysis. Again, DEA is subtle to outliers and random noise from missing explanatory variables or measurement errors which can influence the efficiency estimates (Ohene-Asare, 2011). Finally, some inputs or outputs, though, heterogeneous in nature are sometimes treated as homogenous. This causes biases in the efficiency scores (Coelli et al, 2005). An example is the case where skilled and unskilled labour is not distinguished in an input orientation analysis, but is generally considered as labour, in the assessment, meanwhile, they can have different impacts on the efficiency of the DMUs. To formalize DEA method, suppose there are n observed DMUs to be evaluated, each using varying amounts of m inputs to generate different amounts of s outputs, a specific DMUj consumes xij amount of input i and produces yrj amount of output r. The technology set, T, can be defined as:  xyyxT from produced becan :),( (1) Charnes et al (1978) defined the input-oriented efficiency of a target DMUo as maximum of the ratio of weighted sum of outputs to weighted sum of inputs subject to the condition that similar ratios representing the efficiency measures for each DMU be equal or less than one. University of Ghana http://ugspace.ug.edu.gh 50 Mathematically, the Charnes, Cooper and Rhodes (CCR) fractional programming model (2) can be formalized as:   miisrrvu nj xv yu ts xv yu vuh ir m i iji s r rjr m i ioi s r ror o ,...,1,;,...,1,;0, ,...,1,1 . ,max 1 1 1 1            (2) It should be noted that 0, ijrj xy are the observed outputs and inputs of DMUj whereas 0, ir vu are the weights allocated to the outputs and inputs respectively, and are to be determined by solving the optimization problem. These shadow prices or weights are the relative value system for each firm that makes that firm as efficient as possible. This depends on the idea that the resulting value system is feasible for all other firms and that none achieves an efficiency score below one (Ohene- Asare, 2011; Berger & Mester, 1997). The CCR fractional programming model (2) can be transformed into a corresponding linear model (3) using the Charnes-Cooper transformation (Charnes & Cooper, 1962), under the assumptions of free disposability and convexity: University of Ghana http://ugspace.ug.edu.gh 51 freenj vrs sryy mixx tosubject Min j n j j n j rrjj n j iijj j       ,...2,1 ;0 )(1 ,...,2,1 ; ,...,2,1 ; : 1 1 0 1 0 , *            (3) Where, * denotes the radial input-oriented efficiency score of the DMU under evaluation, a value between 0 and 1; j is the optimal weight of referenced sets for unit j; ijx and rjy are the amount of the ith input used and rth output generated by the jth DMU. The formula of the BCC (or variable returns to scale) model is different from that of the CCR (or constant return to scale) with the activation of the convexity constraint,    n j j 1 1 . The introduction of this constraint changes the reference set from a conical hull, as will be in the case of CRS model, to convex hull (Luo, 2003). By solving this linear programming problem, the efficiency estimates of each DMU can be obtained, inefficient units are identified, targets identified, and bank policy guidelines provided. Based on the assumptions of free disposability and convexity as well as the technology set, T, other measures of efficiency can be estimated. 4.5 Profit Efficiency Using the Non-parametric DEA Methodology Profit efficiency measures how close the profit of a bank is to the profit of a “best-practice” bank producing similar outputs (y) using similar inputs (x) given particular levels of input prices (w) and University of Ghana http://ugspace.ug.edu.gh 52 output prices (p) (Ray & Das, 2010; Das & Ghosh, 2009; Cooper et al., 2006; Maudos & Pastor, 2003). To formalize, given that we have n banks (j=1,…., n) ∈ ℝ+ 𝑛 that use a vector of xi inputs (i=1,…..,m) ∈ ℝ+ 𝑚 to produce a vector of yr outputs (r =1,….,s) ∈ ℝ+ 𝑠 for which they pay a price wi= (i=1,….,m) ∈ ℝ+ 𝑚 for each input and a price pr (r =1,….s) ∈ ℝ+ 𝑠 for each output , the maximum profit can be expressed in the linear programming form as follows. nj vrs sryy mixx ts xwyp wxpy j n j j n j rorjj io n j ijj m i ioio s r roro ,...2,10 )( 1 ,...,2,1 ,...,2,1 . max 1 1 1 11 *                   (4) To obtain the profit efficiency (𝑃𝐸𝑜), the maximum profit ( * ) is estimated by solving the linear programming problem. The profit efficiency of the 0th bank then becomes the ratio of the actual profit, 0 to the maximum profit, * .               s r m i ioiororo s r m i ioiororo O xwyp xwyp PE 1 1 ** 1 1 * 0 (5) Although, the model in equation (4) is commonly used in the literature, it has been established that, it makes assumptions which are not practical in real market systems (Cooper et al, 2006). For example, it assumes that the prices of inputs and outputs are constant, prices are known with University of Ghana http://ugspace.ug.edu.gh 53 certainty and inputs and outputs are homogenous. To eliminate these shortcomings, Cooper et al. (2006) based on some theorems postulated by Tone (2002) introduced a new profit efficiency model. This is given by: nj vrs sryy mixx ts xwyp xwyp j n j j n j rorjj io n j ijj m i ioio s r roro ,...2,10 )( 1 )6(,...,2,1 ,...,2,1 . max max 1 1 1 11 *                   where x and y are cost-based inputs and price-based outputs respectively and account for the heterogeneous nature of inputs and differences in the unit prices of outputs. The new profit efficiency then becomes: )7( 1 1 ** 1 1 * 0              s r m i ioiororo s r m i ioiororo O xwyp xwyp NPE Cooper et al. (2006) and Tone (2002) further argue that, in certain circumstances, equation (7) gives negative values (refer to Table 3 column 4 for an illustration) which are difficult to deal with in terms of its interpretation and the adjustments in inputs and outputs required to achieve the maximum profit target (Fried et al., 2008). For example, although it is feasible to say that, reducing the inputs of a technically efficient bank by 40% implies reducing its cost by 40%, in a profit scenario changing the bank’s outputs and inputs by a certain percentage does not guarantee the University of Ghana http://ugspace.ug.edu.gh 54 same percentage change in profit. To solve this negativity problems, Cooper et al. (2006) and Tone (2002) propose the profit ratio model (revenue /cost efficiency ratio) as indicated in equation 8 to 10 below. Using the MaxDEA Pro 6 software which solves the negativity issues (see Table 3 column 5 for an illustration) in line with Cooper et al (2006), the revenue/cost efficiency ratio is adopted in this study. nj vrs sryy mixx t xw yp j n j j n j rorjj io n j ijj ioio roro yx ,...2,10 )( 1 ,...,2,1 ,...,2,1 . 1 1 1 ,, max                (8) Equation (8) is transformed into a linear programming form by introducing a variable t ∈ ℛ and using the Charnes-Cooper transformation of fractional programming which sets roro tyy ˆ , ioio txx ˆ ,  tˆ .Multiplying all the terms by t> 0, equation (8) changes to: University of Ghana http://ugspace.ug.edu.gh 55 nj vrs srtyy mitxx xw ts yp j n j j n j rorjj io n j ijj ioio roro tyx ,...2,10 ˆ )( 1 ˆ )9(,...,2,1 ˆ ,...,2,1 ˆ 1ˆ . ˆ 1 1 1 , ˆ ,ˆ,ˆ max                 Let an optimal solution for equation (9) be *t , *xˆ , *yˆ , *ˆ .Given that t*>0, an optimal solution for equation 8 can be derived by reversing the transformation from */*ˆ* txx ioio  , */*ˆ* tyy roro  , */*ˆ* t  . The revenue /cost efficiency, ERC of the 0 th bank then becomes )10( * * 1 1 1 1          m i ioio s r roro m i ioio s r roro RC xw yp xw yp E Profit efficiency is bounded between 0 and 1 except in the case where the maximum profit is positive while the actual profit is negative (Ray & Das, 2010). The profit efficiency estimate indicates that, there is the potential of increasing the profits of the bank by (1-ERC)*100 given input and output prices (Maudos & Pastor, 2003). University of Ghana http://ugspace.ug.edu.gh 56 4.5.1. An Illustrative Example of Profit Efficiency Using DEA Assume that there are 11 banks, producing one output (y), using two inputs, x1 and x2 for which w1 and w2 are paid as prices of the inputs and p for the output. To illustrate the calculation of profit efficiency using DEA, the sample data shown in Table 2 is used. Table 2: A hypothetical sample data Bank x1 x2 y w1 w2 p 1 5 17 12 6 8.5 25 2 42 25 26.5 5.5 9.5 22 3 37 14 31 4 10.5 23 4 27 22 30 8 10 21 5 25 10 27 7.5 9 18 6 23 6 9 6.5 5 19 7 18 14 8 5.5 8 23 8 17 15 26 4 6 25 9 16 26 25 5 7.5 24 10 16 12 14 7 8 22 11 5 13 12 6 9 20 Source: Portela and Thanassoulis, 2007 The linear programming model in equation 4 is used in solving the maximum profit for each bank. For example, for bank 1, we have: The actual profit ( 0 ) of bank 1 is computed as {(25*12)-[(6*5) + (17*8.5)]} =125. 0,,,,,,,,,,:int 1:int 1212142526892730315.2612:int 17131226151461022142517:int 551616171823252737425:int .. max 1110987654321 1110987654321 1110987654321 1110987654321 1110987664321 *            ConstranegativityNon ConstraVRS ConstraLoan ConstraLabour ConstraDeposit ts wxpy University of Ghana http://ugspace.ug.edu.gh 57 Figure 4: A graphical representation of Profit Efficiency using one input and one output Using the MaxDEA Pro 6 software, the maximum profits generated by each bank is depicted in the second column of Table 3 Table 3: Results for maximum profit, actual profit and profit efficiency scores. Bank Maximum Profit * Actual Profit 0 Profit Efficiency = * 0   Profit Efficiency = Cost venueRe 1 159.5 125.5 0.78683 1 2 361.5 114.5 0.31674 0.3025 3 418 418 1 1 4 194 194 1 0.3512 5 208.5 208.5 1 0.4257 6 -8.5 -8.5 1 1 7 378.526 -27 -0.07133 0.2120 8 492 492 1 1 9 325 325 1 0.5304 10 256.316 100 0.39014 0.3600 11 93 93 1 1 1 2 3 4 5 6 7 8 9 10 1 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 40 45 O u tp u t (y ) Input (x1) Profit Efficiency Π0 = 𝑝𝑦 − 𝑤x 𝛱∗=max py-wx University of Ghana http://ugspace.ug.edu.gh 58 The profit efficiency estimates indicate that banks 3, 4, 5, 6, 8 and 9 are efficient in the industry. To be efficient, banks 1, 2, 7 and 10 would be required to increase their profits by approximately 21%, 68%, 107% and 61% respectively. 4.6 Modelling Of Inputs and Output Variables In modelling the inputs and outputs of the banks, the intermediation approach by Sealey and Lindley (1977), unlike the production approach (Berger & Humphrey, 1991; Benston, 1965), is employed. This is because the intermediation approach does not only consider the minimization of production costs but the entire total cost of a bank which is necessary in profit maximization (Berger & Humphrey, 1997). The approach views banks as agents mediating funds between investors and supply sources. In this regard, labour and physical capital (and sometimes equity) are considered as inputs used to convert deposits into loans and investment (Halkos & Tzeremes, 2013; Ataullah & Le, 2006). Following Das and Ghosh (2009) and Moffat and Valadkhani (2011), this study considers deposits, physical capital and labour as inputs and loans and investments as outputs as indicated in Table 4 University of Ghana http://ugspace.ug.edu.gh 59 Table 4: Inputs, Outputs, Input and Output Prices for Profit Efficiency Estimation Variables Description Inputs: Deposits Customer deposits Labour Personnel Expenses Physical Capital Fixed assets or property, plant and equipment Outputs: Loans and Advances Total customer’s loans and advances Other Earning Assets Bank’s investment in different types of securities (bonds etc.) Input Prices: Price of deposits Interest expense divided by total deposits Price of Labour Personnel expenses divided by total assets Price of capital other operating expenses divided by total fixed assets Output Prices Price of loans and advances Interest received on loans and advances divided by total loans and advances Price of Other Earning Assets Investment income divided by total investments 4.6.1 Inputs 4.6.1.1 Deposits Deposits are essential resources for banks in Ghana. They aid in credit creation and investment by these banks. However, there has been a long standing debate on whether deposits are inputs or outputs of banks. This is because, deposits have both input and output features (Fethi & Pasiouras, 2010). Some studies consider deposits as inputs because they are partly paid for using interest expenses. Also, banks ‘buy’ instead of ‘sell’ deposits and deposits are used next to other funds to generate loans and investments (Nahm &Vu, 2013; Berger & Humphrey, 1997; Hughes & Mester, 1993; Elyasiani & Mehdian, 1990). Conversely, they are considered as outputs because they are linked to safekeeping, payment services and liquidity provided to depositors (Luo, 2003). To resolve this controversy, some studies employed both approaches in their models (Tortosa-Ausina, University of Ghana http://ugspace.ug.edu.gh 60 2002). Following several others, this study, however, considers deposits as inputs and represents it as the summation of all monies in customer savings accounts, current accounts and time deposits. 4.6.1.2 Labour The theory of production in microeconomics considers labour as an important resource for the generation of outputs. Some banking efficiency studies define labour as the expenses incurred by banks on their staff (Kenjegalieva, Simper, Weyman-Jones & Zelenyuk 2009; Drake, Hall & Simper, 2006). These expenses include wages and salaries, pension costs, staff provident fund contributions, staff loans and other training expenses. Other studies define labour as the average number of employees documented in the bank’s annual report during the year (Halkos & Tzeremes, 2013; Fukuyama & Matousek, 2011). The former definition of labour is applied in this study due to the data unavailability on the average number of employees for each bank in Ghana for the entire study period. Besides, using labour cost instead of the number of employees indirectly captures labour quality by virtue of the value/cost attached to the labour. 4.6.1.3 Physical Capital Physical capital, proxied by the cost of fixed assets is also an essential factor of production. In the banking literature, it refers to the book value of all property, plant and equipment, machinery, fixtures and premises acquired by the bank either through an outright purchase or by means of a lease. It is valued at cost less accumulated depreciation and impairment losses. Empirical application of this input includes studies conducted by Tortosa-Ausina et al. (2012), Assaf et al. (2011), Kenjegalieva et al. (2009) and Havrylchyk (2006). University of Ghana http://ugspace.ug.edu.gh 61 4.6.2 Outputs 4.6.2.1 Loans and Advances Loans and advances refer to debt provided by banks to households and other business entities. In this study, loans and advances is a summation of the monetary values of customer loans, corporate loans, staff loans, mortgage loans and other loans (Das & Ghosh, 2009; Havrylchyk, 2006). To ensure loan portfolio quality, provisions for bad and doubtful debt is subtracted from the loans and advances (Grigorian & Manole, 2002). 4.6.2.2 Other Earning Assets Other Earning Assets refer to the bank’s investments in different types of securities. In this study, it is an aggregate of investment in government securities (treasury bills and notes, government bonds), investment securities-available for sale, investment in other securities, investment in associate companies and subsidiaries and equity investments. 4.6.3 Input Prices The price of deposits is obtained by dividing the interest expenses incurred by the bank on any demand deposits, call deposits, time or fixed deposit, savings account and current accounts by its total deposits. The price of labour is estimated by dividing the expenses incurred by the bank on its personnel or staff by total assets (Delis et al., 2009). The price of physical capital is estimated as other operating expenses divided by total fixed assets (Ray & Das, 2010). Other operating expenses in this case are the operating expenses less the staff costs of the bank University of Ghana http://ugspace.ug.edu.gh 62 4.6.4 Output Prices The price of loans and advances is estimated by dividing the interest income on loans and advances by the total loans and advances of the bank. The price of investment is estimated by dividing the interest income on investment by the sum of all the investments made by the bank in the year under consideration. 4.7 Bootstrapping the Second-Stage Regression with environmental variables Many studies have regressed DEA efficiency estimates on certain covariates or environmental variables in the so-called two-stage processes to determine how the exogenous variables can affect the efficiency levels of firms (Simar & Wilson, 2007; 2011). These variables are usually not controlled by management, but may influence the efficiency estimates generated in the first stage such as the profit efficiency estimate of equation (10). Simar and Wilson (2007) argued that DEA efficiency estimates from the first stage are serially correlated in an unknown (in a statistical sense) and complicated way. Thus, regressing these estimates on certain exogenous covariates without recognizing this deficiency could lead to invalid inferences. To resolve the deficiency of serial correlation, preceding studies adapted Tobit regression models (in the two-stage, DEA plus regression approach) due to the censored nature of DEA scores (which are constrained from above on the right at 1). Simar and Wilson (2007) criticized these studies because of their failure to expound the underlying Data Generating Process (DGP) that allow uncontrollable covariates to affect firm’s efficiencies. They also contended that the first stage dependency issue suggests that the stochastic error term of the Tobit regression is correlated with the environmental variables making Tobit estimation inappropriate. The outcome is that inferences University of Ghana http://ugspace.ug.edu.gh 63 on the second-stage parameters will be biased and inconsistent. Also, although, employing maximum likelihood in the stage-two analysis implies that this correlation vanishes asymptotically, it occurs at a very slow pace and may produce invalid inference. Hirschberg and Lloyd (2002) and Xue and Harker (1999) had proposed a single bootstrap approach to handle the issue of correlation, which was used by Casu and Molyneux (2003), but, Simar and Wilson (2007) critiqued this “naive” bootstrap technique for resampling without considering the peculiar distributions of efficiency scores derived via non-parametric DEA approach. To address these, Simar and Wilson (2007) have proposed a double-bootstrapped truncated regression when undertaking a second-stage regression whereby the efficiency estimates are regressed on some environmental covariates instead of OLS or Tobit estimates. This is to allow valid inferences and improve statistical efficiency of the second-stage estimates Denoting 𝜃𝑗 as the true unobserved profit efficiency score of bank j and 𝑇𝑗 as the row vector of specific covariates or environmental factors for bank j (that is, those factors anticipated to affect the bank’s profit efficiency score), the second-stage truncated regression model can be specified as: 𝜃𝑗 = 𝛼 + 𝛽𝑇𝑗 + 𝜀𝑗 , 𝑗 = 1,… . , 𝑛 (11) Using Algorithm 2 of Simar and Wilson (2007) double-bootstrapped truncated regression procedure, the distribution of 𝜀𝑗 is assumed to be constrained by the condition  jj T1 . The distribution of 𝜀𝑗 is also assumed to be truncated normal with a mean of zero (before truncation), an unknown variance and a left truncation point determined by constraint imposed on it. The true unobserved 𝜃𝑗 is substituted by the profit efficiency RCE estimated in the first stage University of Ghana http://ugspace.ug.edu.gh 64 (equation 10). Since Simar and Wilson (2007) argued against Tobit estimation, a truncated econometric regression model is followed which is given by. njTERC jjo ,...,1,   (12) where  ,,0~ 2 Nj such that njT jj ,...,1,1   The estimated parameter ˆ is obtained via maximizing the correspondent likelihood function in relation to  2,  given the obtained data set. Simar and Wilson’s (2007) algorithm 2 parametric bootstrap for regression which includes information on the parametric structure and distributional assumption to obtain the bootstrap confidence intervals for the estimates of parameters  and 2 is followed. They argued using a Monte Carlo simulation that this procedure ensured feasible, unbiased and consistent estimates of the second-stage truncated regression. For a comprehensive explanation of the estimation algorithm, readers are referred to Simar and Wilson (2007, 2011). McDonald (2009) and Banker and Natarajan (2008) have in recent times argued that Ordinary Least Squares (OLS) produces consistent estimates in the second-stage regression. Similarly, Banker and Natarajan (2008) and Ramalho, Ramalho and Henriques (2010) have noted the computational burden of bootstrapping and the fact that more bootstrap replications and larger sample size are required for convergence to be achieved. McDonald (2009) indicated that the estimates are created from fractional data and not a censoring process and thus Tobit regression is inconsistent with the data generating process. Saxonhouse (1976) noted that heteroscedasticity can occur if estimated parameters are used as explained variables in regression analysis. McDonald (2009) showed that if White’s (1980) heteroskedastic-consistent-standard errors are calculated, large sample tests can be performed which are robust to heteroscedasticity and the distribution of University of Ghana http://ugspace.ug.edu.gh 65 the disturbances. Hoff (2007) however, argued, that the Tobit (censored) and OLS regressions adequately represented DEA second-stage analysis using the case study of Danish fishery. By comparing and contrasting the assumptions underlying their truncated regression model and the OLS suggested by Banker and Natarajan (2008), Simar and Wilson (2011) showed that the second-stage OLS estimation by Banker and Natarajan (2008) is consistent only under unusual and peculiar assumptions on the DGP that limit its applicability. In view of this, despite attempts to criticize Simar and Wilson’s (2007, 2011) second-stage regression, their double-bootstrapped- truncated regression is adopted in this study since it provides the only possible means for inference in the second-stage regression. 4.8 Capital Structure and Profit Efficiency The regression model for the capital structure-profit efficiency nexus is given by: 𝐸𝑅𝐶𝑖,𝑡 = 𝛽1𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝛽2𝐸𝐶𝐴𝑃 2 𝑖,𝑡 + 𝜑𝑍𝑖,𝑡 + 𝜇𝑖 + 𝜆𝑡 + 𝜀𝑖,𝑡 (13) where 𝐸𝑅𝐶𝑖,𝑡 is the profit efficiency of bank i at time t; 𝜇𝑖 represents firm-specific fixed effects; 𝜆𝑡 captures time effects and controls for macroeconomic events; 𝐸𝐶𝐴𝑃𝑖,𝑡 is equity capital; 𝐸𝐶𝐴𝑃𝑖,𝑡 2 is the squared term of equity capital, 𝑍𝑖,𝑡 is a set of control variables; 𝜀𝑖,𝑡 is the error term (Fosu, 2013; Magaritis & Psillaki, 2010; Berger & Bonaccorsi di Patti, 2006). The control variables include size, sales growth, ownership structure, bank regulations and technical efficiency. University of Ghana http://ugspace.ug.edu.gh 66 4.8.1 Variable Measurements 4.8.1.1 Equity capital and Equity capital squared In banking research, equity capital, an inverse measure of leverage captures the capital structure of banks. It is computed as the ratio of total equity to total assets. Equity capital is used as an inverse measure of leverage due to the regulations surrounding the choice of capital ratios in the banking industry. A study by Berger and Bonaccorsi di Patti (2006) established a positive nexus between leverage (lower equity) and profit efficiency which follows the agency cost hypothesis of capital structure. Equity capital squared is introduced to account for possible nonlinear effects of leverage on bank performance (Berger & Bonaccorsi di Patti, 2006). 4.8.1.2 Bank Size Bank size is measured by taking the natural logarithm of the bank’s total assets. Size is introduced to determine whether possible economies and diseconomies of scale exist in the banking industry of Ghana. The findings of preceding studies on the relationship between bank efficiency and size are inconsistent. Whereas Tecles and Tabak (2010) and Ataullah and Le (2006) report a significantly positive relationship, Pasiouras and Kosmidou (2007), Altunbas, Carbo, Gardener and Molyneux (2007) report a significantly negative relationship. Other authors, including Staub, de Silva e Souza and Tabak (2010) and Ariff and Can (2008) report insignificant influence of bank size on efficiency. University of Ghana http://ugspace.ug.edu.gh 67 4.8.1.3 Sales Growth Following Gatsi (2012), percentage change in net interest income is used as a proxy for sales growth. A positive relationship implies operational efficiency in the bank. A negative relationship would demonstrate that banks do not gain from their core business activities. 4.8.1.4 Ownership Structure As regards ownership structure, a dummy variable is used. A dummy of 1 is allocated to a foreign bank and 0 to a domestic bank. A domestic bank is one in which not less 60% of equity capital is owned by Ghanaians whereas a foreign bank is one in which at least 60% of equity capital is owned by foreigners (Banking Act, 2004). Some studies espouse that, banks that are foreign-owned perform better than their domestic counterparts (Bokpin, 2013; Berger et al., 2010; Bonin, Hassan & Wachtel, 2005; Fries & Taci, 2005) because they are well diversified and possess better technologies while others report the opposite (Berger, DeYoung, Genay & Udell, 2000). 4.8.1.5 Regulation Consistent with existing literature (Pasiouras, Tanna & Zopounidis, 2009; Pasiouras, 2008; VanHoose, 2007), capital is introduced to determine the impact of regulation on the profit efficiency of banks in Ghana. Specifically, the natural logarithm of the minimum capital requirement or stated capital for the banks in Ghana is used as the proxy for capital. Pasiouras et al. (2009) for example established a positive and statistically significant relationship between capital requirement and profit inefficiency suggesting that higher requirements lower the profits obtained by the banks. University of Ghana http://ugspace.ug.edu.gh 68 4.8.1.6 Technical Efficiency Technical efficiency which measures the ability of a bank to use given resources to generate more output or reduce inputs while keeping the same level of output (Sathye, 2003) is also used as a control variable. Miller and Noulas (1996) established that, profitable banks have higher levels of technical efficiency than their less profitable counterparts. 4.9 Measures of Competition The HHI which computes the degree of competition from the product market concentration viewpoint (Chong, Lu & Ongena, 2013) in conjunction with the BI which measures competition from the behavior of the market (Boone, 2008; Griffith, Boone & Harrison, 2005; Boone, 2000) are employed to evaluate competition in the Ghanaian banking industry. HHI is computed as the sum of the squared market shares of each bank.   21  Ni iMSHHI (14) MSi is the market share of bank i. The market shares of the banks are calculated based on three criteria .These are, industry total assets, industry deposits and industry net advances. The BI is premised on the idea that, given two firms in an industry, where one of the firms is efficient (has a lower marginal cost) and the other is inefficient, with an increase in competition in this industry, the profits of the more efficient firm will increase relative to that of the inefficient firm (Boone, 2000). Following Ohene-Asare and Latif (2014) and Van Leuvensteijn, Bikker and Rixtel (2011), the BI which measures competition is estimated using the equation below 𝑙𝑛𝑅𝑂𝐴𝑖 = 𝛼 + 𝛽𝑙𝑛𝑀𝑐𝑖 + 𝜀𝑖 (15) University of Ghana http://ugspace.ug.edu.gh 69 Where 𝑅𝑂𝐴𝑖 is the return on asset for bank i, 𝑀𝑐𝑖 is the marginal cost of bank i, 𝛽 is the BI and 𝜀𝑖 is the unobserved error term. Following the leads of Schaek and Cihak (2014), marginal cost is estimated as the ratio of average cost (interest expenses, staff cost and other operating expenses) to total income since marginal costs are not directly observable. When β < 0, it implies a competitive banking industry and β > 0 indicates an uncompetitive industry. 4.9.1 Capital Structure, Profit Efficiency and Competition. In estimating the effect of capital structure and competition on the profit efficiency of a bank, the following model is used. 𝐸𝑅𝐶𝑖,𝑡 = 𝛼1𝐸𝐶𝐴𝑃𝑖,𝑡, + 𝛼2𝐸𝐶𝐴𝑃 2 𝑖,𝑡 + 𝛼3𝐶𝑜𝑚𝑡 + 𝛼4𝐸𝐶𝐴𝑃𝑖,𝑡 ∗ 𝐶𝑜𝑚𝑡 + 𝜑𝑍𝑖,𝑡 + 𝜇𝑖 + 𝜆𝑡 + 𝜀𝑖,𝑡 (16) Where ERC i, t is the profit efficiency of bank i at time t; 𝐸𝐶𝐴𝑃𝑖,𝑡 is the equity capital of bank i at time t:𝐶𝑜𝑚𝑡 is the level of competition in the industry at time t proxied by the HHI and the BI alternatively; 𝐸𝐶𝐴𝑃𝑖,𝑡 ∗ 𝐶𝑜𝑚𝑡 is the interaction of equity capital and competition; 𝑍𝑖,𝑡 is a set of control variables (see section 4.8); 𝜇𝑖 represents firm-specific fixed effects; 𝜆𝑡 captures time varying effects; 𝜀𝑖,𝑡 is the error term. To obtain the effect of capital structure on profit efficiency amidst competition, the first derivative of equation (16) with respect to 𝐸𝐶𝐴𝑃𝑖,𝑡 is taken. This is shown in equation 17. In summarizing the effect of capital structure on profit efficiency, interesting values such as the mean value and lower and upper quartiles of equity capital and competition must be used to evaluate equation 17. To test whether the estimate is statistically different from zero, the model must be rerun with the sample mean values of equity capital and competition (Wooldridge, 2012). University of Ghana http://ugspace.ug.edu.gh 70 𝑑𝐸𝑅𝐶𝑖,𝑡 𝑑𝐸𝐶𝐴𝑃𝑖,𝑡 = 𝛼1 + 2𝛼2𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝛼4𝐶𝑜𝑚𝑡 (17) If HHI is used as a proxy for competition in equation (16), then 𝛼1 + 2𝛼2𝐸𝐶𝐴𝑃𝑖,𝑡 captures the effect of capital structure on the efficiency of banks in an unconcentrated (perfectly competitive) industry whiles 𝛼1 + 2𝛼2𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝛼4𝐶𝑜𝑚𝑡 captures the effect of capital structure at specified levels of concentration (competition). If BI is used then 𝛼1 + 2𝛼2𝐸𝐶𝐴𝑃𝑖,𝑡 captures the effect of capital structure in a concentrated (uncompetitive) industry whiles 𝛼1 + 2𝛼2𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝛼4𝐶𝑜𝑚𝑡 captures the effect of capital structure at specified levels of competition. 4.9.2 Capital Structure, Conglomeration and Profit Efficiency Vander Vennet (2002) and Nicoló et al., (2004) defined a financial conglomerate as a bank that engages in other activities (e.g. insurance, securities related activities) aside the traditional banking activities of deposit taking and lending. In Ghana, given the introduction of the Universal Banking License in 2003 (Ghana Banking Survey, 2008) which permits the banks to perform other activities besides the traditional banking activities of borrowing and lending, this study proxies a conglomerate as a bank that has subsidiaries performing activities other than those stipulated as permissible banking activities by the Banking Act 2007 (Act 738). A dummy of 1 is assigned to a bank that is a conglomerate and zero (0) to a non-conglomerate/focused bank. 𝐸𝑅𝐶𝑖,𝑡 = 𝛾1𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝛾2𝐸𝐶𝐴𝑃 2 𝑖,𝑡 + 𝛾3(𝐶𝑜𝑛𝑔) + 𝛾4𝐸𝐶𝐴𝑃𝑖,𝑡 ∗ 𝐶𝑜𝑛𝑔 + 𝜓𝑍𝑖,𝑡 + 𝜇𝑖 + 𝜆𝑡 + 𝜀𝑖,𝑡 (18) where Cong is a binary variable capturing conglomeration. University of Ghana http://ugspace.ug.edu.gh 71 4.9.3 Reverse Causation between Profit Efficiency and Capital Structure Consistent with Berger and Bonaccorsi di Patti (2006), the reverse causation between profit efficiency and capital structure is tested using a simultaneous equation model. In doing so, a two- equation structural model is constructed and estimated using two stage least squares (2SLS). 𝐸𝑅𝐶𝑖,𝑡 = 𝛿1𝐸𝐶𝐴𝑃𝑖,𝑡 + 𝜑𝑍1𝑖,𝑡 + 𝜇1𝑖 + 𝜆1𝑡 + 𝜀1𝑖,𝑡 (19) 𝐸𝐶𝐴𝑃𝑖,𝑡 = 𝜌1𝐸𝑅𝐶𝑖,𝑡−1 + 𝜙𝑍2𝑖,𝑡 + 𝜇2𝑖 + 𝜆2𝑡 + 𝜀2𝑖,𝑡 (20) 𝑍1𝑖,𝑡 includes bank size, sales growth, ownership structure, bank regulation, and technical efficiency. 𝑍2𝑖,𝑡 includes bank size, sales growth, bank regulation and asset structure. The simultaneous equations model is identified because the appropriate number of variables are excluded from each of the Z vectors. 4.9.4 Instruments for Data Analysis Data is analysed using MaxDEA Pro 6.3 and R codes with censReg (Henningsen, 2012), FEAR (Wilson, 2008) and Benchmarking (Bogetoft & Otto, 2011) packages. 4.9.5 Chapter Summary This chapter provided an in-depth explanation of the methods adopted in achieving the objectives of this study. This included clarity on the design of the research, the sources of data and the assumptions underlying DEA. Further, the chapter provided justification for the inputs and outputs selected in estimating profit efficiency and the reason for bootstrapping the second-stage regression. Instruments needed for analysing the data was also presented. University of Ghana http://ugspace.ug.edu.gh 72 CHAPTER FIVE DATA ANALYSIS 5.1 Introduction This chapter presents the results derived from analysing the data set for this study. First, it presents and analyses the descriptive statistics of the inputs, outputs, input prices and output prices used in estimating the profit efficiencies of banks in Ghana. It then presents and provides preliminary analyses and discussions on the objectives of the study. 5.2 Descriptive Statistics of Variables In the banking and finance efficiency literature, there are controversies surrounding the specification of inputs and outputs for frontier modelling. The literature recognizes that the choice of variables in efficiency studies can influence the results significantly. It is necessary therefore that, variables used for efficiency be described to ensure they are in tandem with the assumptions required for effective estimations. The data set used in this study was sourced from BOG and the annual reports of 26 banks. For generalization, pooled summary statistics of the variables used in the estimation of profit efficiency for the banks in Ghana are reported in Table 5. A year-by-year summary statistics of these variables have been attached in Appendix A. University of Ghana http://ugspace.ug.edu.gh 73 Table 5: Summary statistics of variables used-pooled data (GH¢) Variables Mean Std. Dev. Min Max F-stat across time Pooled Data-N-304 Inputs Deposits 347,000,000 463,000,000 532,192.1 3.221E+09 12.78*** Labour 17,530,602 25,937,965 49,094 169,996,000 8.53*** Fixed Assets 13,228,737 15,780,213 59,111 94,756,640 10.4*** Input Prices Deposits 0.08 0.05 0.01 0.4 5.29*** Labour Cost 0.04 0.08 0 1.47 0.91 Fixed Assets 1.71 1.65 0.15 17.72 1.50 Outputs Loans and Advances 217,000,000 281,000,000 154,708 2.125E+09 13.53*** Investment 129,000,000 226,000,000 61,000 1.747E+09 8.07*** Output Prices Loans and Advances 0.19 0.07 0.02 0.52 8.43*** Investment 0.23 0.59 0 8.45 0.79 Others Total Assets 503,645,096 653,752,686 884,009 4.624E+09 11.99*** Equity 67,946,016 89,061,008 -1,606,000 557,106,000 21.56*** An examination of the prices of inputs suggest that on average, the most expensive factor of production in Ghana’s banking industry is capital or fixed assets (GH¢1.71) which is typical of most developing countries. The results reported also suggest that, banks in Ghana on average price their deposits (GH¢0.08) much lower than the loans (GH¢0.19) they grant their customers. The implication of this pricing strategy is that, ceteris paribus, on average, banks in Ghana generate more from loans and incur less expenses on the deposits. The results also reveal that, banks in Ghana gave out approximately 62.54% of their deposits as loans to their customers and invested about 37.18% for the period under consideration. The minimum and maximum values of the variables, particularly that of total assets (min=GH¢884,009, max=GH¢4,624,405,000) and their relatively high standard deviations indicate that banks in Ghana have different sizes. This justifies the use of the Variable Returns to Scale (VRS) assumption of Banker et al. (1984) in the profit efficiency estimation. Further, the relatively high standard deviations of the inputs, outputs and equity indicate that the actual University of Ghana http://ugspace.ug.edu.gh 74 averages of these variables are largely dispersed from the expected suggesting volatility in the variables over the 14 year period. A test of differences in the variables across time using a one way anova reveals significant differences of each input and output across time. This is also presented in Table 5. This suggests that, pooling data across years for the purposes of generalization as have been done in Table 5 may be flawed. This is because pooling tacitly implies that all the banks in Ghana operated in the same environment or under a common condition. However, the variables may have been affected by certain yearly factors including changes in the regulatory framework which may not be visible when the data is pooled or other macroeconomic conditions peculiar to some years. To resolve this potential unobserved year-specific factors, a trend analysis of the variables is provided in Figures 5 and 6. Figure 5: Trend Analysis of Average Inputs and Outputs from 2000 to 2013 0 100000000 200000000 300000000 400000000 500000000 600000000 700000000 800000000 900000000 1E+09 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 M ea n Years Deposits Labour Fixed Assets Loans and Advances Investments University of Ghana http://ugspace.ug.edu.gh 75 Figure 5 shows that with the exception of investments which fell between 2005 and 2007 and in 2010, there has been a steady rise in the inputs and the other output. The fall in investment is probably because of the decrease in the discount rate of the 91 day Treasury bill from 25% in 2002 to 10.7% in 2006 (Ghana Banking Survey, 2007) which made investments less attractive to the banks. Figure 6: Trend Analysis of Average input and output prices from 2000 to 2013 Notable observations in Figure 6 are the prices of deposits and the prices of loans and advances. The price of deposits fell in 2002, 2004, 2010 and 2011. Similarly, the price of loans and advances fell between 2002 and 2006 and between 2010 and 2011. This is probably because, the economy experienced falls in inflation rates during these periods. For example, the inflation rate fell from 40.5% in 2000 to 15% in 2002 and 12% in 2004 (Ghana Banking Survey, 2005). It also fell from 16.9% in 2009 to 8.6% in 2010 and 2011. Likewise, the monetary policy rates fell from 18% in 2009 to 13.5% in 2010 and 13% in 2011 (Ghana Banking Survey, 2011, 2012). Thus, the fall in 0 0.5 1 1.5 2 2.5 3 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 M ea n Years Price of Deposits Labour Cost Price of Fixed Assets Price of Loans and Advances Price of Investments University of Ghana http://ugspace.ug.edu.gh 76 prices (interest rates) mirrored the fall in inflation and the changes in the monetary policy rates. The year 2006 was also characterized by the abolition of the secondary reserve requirement of 35%. This increased the amount of deposits in the industry and thus the amount of loans that could be created. Given a constant demand, it is expected that the excess supply of loans would lead to a fall in the prices of loans and advances. Table 6 shows summary statistics of the variables by accounting for bank ownership. Although, the yearly average prices of labour (see Appendix A) seem to suggest that local banks pay more salaries per worker than foreign banks, the aggregate of these observations indicate that, foreign banks pay higher salaries per employee and invest more in expensive capital than the local banks. This substantiates the perception that, foreign banks have enhanced the human resources quality in the banking industry. However, as denoted by the standard deviations, the staff costs incurred by foreign banks is volatile. This is probably because these foreign banks differ in terms of origin. University of Ghana http://ugspace.ug.edu.gh 77 Table 6: Summary statistics of variables for local and foreign banks Variable Name Ownership Pooled Data Mean Std Dev. Min Max Inputs Deposit Local 32,7481,345 430,908,027.2 2,172,653.1 263,0283,000 Foreign 365,569,503.9 491,850,232.7 532,192.1 3,220,777,000 Labour Local 18,532,582.51 28,441,355.04 153,358.8 169,996,000 Foreign 16,580,336.52 23,371,322.05 49,094 140,748,000 Fixed Assets Local 13,998,499.26 16,074,029.87 264,843.9 94,756,640 Foreign 12,508,003.74 15,516,793.37 59,111 82,726,830 Input Prices Deposit Local 0.1* 0.05 0.02 0.4 Foreign 0.06* 0.04 0.01 0.26 Labour Local 0.03 0.01 0.01 0.07 Foreign 0.04 0.11 0 1.47 Fixed Assets Local 1.38* 0.87 0.32 5.34 Foreign 1.99* 2.05 0.15 17.72 Outputs Loans and Advances Local 229,684,428.4 268,448,625.3 646,601.5 1,265,517,000 Foreign 205,127,027.5 293,460,737.3 154,708 2,124,530,000 Investments Local 109,775,452.7 226,612,072.4 61,000 1,747,087,000 Foreign 147,308,672.1 225,504,143.3 114,786 1,229,731,000 Output Prices Price of Loans and Advances Local 0.21* 0.07 0.07 0.49 Foreign 0.18* 0.08 0.02 0.52 Price of Investments Local 0.16* 0.06 0 0.4 Foreign 0.3* 0.82 0 8.45 Total Assets Local 472,346,036.3 567,446,443.6 2,712,278.6 3,391,100,000 Foreign 532,950,585.8 725,937,822.7 884,009 4624,405,000 Equity Local 56,551,757.31* 71,165,468.79 405,446.4 447,156,000 Foreign 78,614,526.54* 102,124,229.6 -1,606,000 557,106,000 Observations 304 304 304 304 * indicates that there are significant differences in the mean of the variable for local and foreign banks at 5% An independent sample t-test of the means of the variables also indicated that there are significant differences in the prices of deposits, prices of fixed assets, prices of loans and advances, prices of investments and equity between local and foreign banks. Table 7 reports the summary statistics of the variables according to the degree of conglomeration. A conglomerate in this context refers to University of Ghana http://ugspace.ug.edu.gh 78 a bank that engages in other activities aside those stipulated by the Banking Act 2004 (Act 673). In other words, it is a bank that has a subsidiary performing non-banking activities. A focused bank is one that engages solely in the activities stated in Act 673 and has no subsidiary. Weston (1970) and Chandler (1977) argue that with conglomeration, firms enjoy economies of scale which make them efficient in their operations. The total assets of conglomerates (mean=GH¢706,797,749) exceed that of the focused banks (mean= GH¢381,753,505) on average. An independent sample t-test of the means implies that in Ghana, banks that are conglomerates are significantly larger than their focused counterparts. Table 7:Summary statistics of variables for conglomerates and focus banks Variable Name Degree of conglomeration Pooled Data Mean Std Dev Min Max Inputs Deposit Conglomerate 524,038,255* 597,034,365 12,725,429 3,220,777,000 Focus 241,020,046* 316,835,670 532,192.1 1,592,324,000 Labour Conglomerate 27,894,943* 34,423,376 284,481.2 169,996,000 Focus 11,245,842* 16,176,370 49,094 92,667,010 Fixed Assets Conglomerate 18,749,607* 18,161,973 340,725.4 82,726,830 Focus 9,916,214.4* 13,137,079 59,111 94,756,640 Input Prices Deposit Conglomerate 0.08 0.05 0.02 0.28 Focus 0.08 0.05 0.01 0.4 Labour Conglomerate 0.03 0.01 0.01 0.01 Focus 0.04 0.11 0 1.47 Fixed Assets Conglomerate 1.3* 0.63 0.32 4.8 Focus 1.95* 1.99 0.15 17.72 Outputs Loans and Advances Conglomerate 331,211,089* 350,034,709 3,206,220 2,124,530,000 Focus 148,476,264* 202,855,195 154,708 975,584,000 Investments Conglomerate 186,471,918* 288,051,418 4,953,578.8 1,747,087,000 Focus 94,771,866* 171,551,720 61,000 1,229,731,000 Output Prices Price of Loans and Advances Conglomerate 0.19 0.07 0.07 0.48 Focus 0.19 0.08 0.02 0.52 Price of Investments Conglomerate 0.16* 0.06 0.05 0.35 Focus 0.28* 0.75 0 8.45 Total Assets Conglomerate 706,797,749* 792,182,406 17,228,548 462,4405,000 Focus 381,753,505* 519,761,008 884,009 293,0852,000 Equity Conglomerate 90,283,252* 107,113,842 702,313 557,106,000 Focus 54,543,675* 73,313,645 -1,606,000 445,232,000 Observations 304 304 304 304 * indicates that there are significant differences in the mean of the variable for conglomerate and focused banks at 5%. University of Ghana http://ugspace.ug.edu.gh 79 5.3 Profit Efficiency of Banks in Ghana To achieve the first objective of this study, the profit efficiency scores of each bank under consideration is estimated under VRS. This is attached in Appendix B. For generalization, the means and standard deviations of the profit efficiencies estimated are reported in Table 8. Considering the fact that there were significant differences in the variables for estimation (see Table 5), the profit efficiency estimates are estimated relative to each year-specific frontier. Table 8 presents the average year wise distribution of profit efficiency of banks in Ghana. Both the arithmetic and geometric means are computed. This is because for normalized benchmark scores, using arithmetic mean alone may lead to wrong inferences (Ohene-Asare & Asmild, 2012). The high levels of profit efficiencies accompanied by lower standard deviations suggest that most banks in Ghana lie close to the benchmark profit frontier. On average, 79% of the potential profits Table 8: Average Profit Efficiencies of Banks in Ghana Year Number of Banks PROFIT EFFICIENCY Geometric Mean Arithmetic Mean Std Dev. 2000 16 0.77 0.79 0.20 2001 17 0.80 0.84 0.24 2002 17 0.76 0.79 0.21 2003 18 0.90 0.91 0.13 2004 18 0.81 0.83 0.19 2005 19 0.94 0.94 0.11 2006 23 0.90 0.92 0.16 2007 23 0.89 0.90 0.15 2008 24 0.89 0.90 0.13 2009 27 0.80 0.83 0.20 2010 26 0.77 0.80 0.21 2011 25 0.68 0.73 0.25 2012 25 0.66 0.72 0.26 2013 26 0.60 0.66 0.28 Average 0.79 0.82 0.22 University of Ghana http://ugspace.ug.edu.gh 80 that a best practice bank could make under similar conditions are earned by most of the banks in Ghana. Compared to other banking industries in other countries, the Ghanaian banking industry has recorded higher profit efficiency results. The average profit efficiency was found to be 57.5% for Spanish banks (Maudos & Pastor, 2003), 50.5% for Chinese banks (Ariff & Can, 2008) and 52.14% for Indian banks (Ray & Das, 2010). This is probably because, unlike these countries, Ghana’s capital market is underdeveloped. Thus, the disintermediation process that has triggered declines in these banking industries has not yet threatened Ghana. To test whether there are significant differences in the profit efficiency estimates over time, both the non-parametric Kruskal Wallis test and the parametric anova as well as Tukey HSD multiple comparison tests were conducted. Both the Kruskal Wallis (𝜒2 = 35.786, 𝑝 = .001) and the anova (F=3.905, p=.000) reveal significant differences over time. Profit efficiency since 2009 has fallen monotonically over the years. This is probably because of the change in government in 2009 as well as the global financial crisis in 2008. Changes in economic management policies of the new government may have had adverse effects on banks. The global financial downturn may also have influenced banks in Ghana potentially because of their direct exposure to their partners abroad in the form of nostro placements and balances. Further, rankings of the average profit efficiencies of each bank over time indicate that, the best performing banks are GCB and BARODA. These are followed by SCB, BBG and UT. University of Ghana http://ugspace.ug.edu.gh 81 Table 9: Average Profit Efficiency Rankings Of banks in Ghana from 2000 to 2013 Bank OWN AM GM Ranka Bank OWN AM GM Ranka GCB L 1.00 1.00 1 FABL L 0.82 0.77 16 BARODA F 1.00 1.00 1 CAL L 0.81 0.79 17 SCB F 0.96 0.95 3 ZENITH F 0.81 0.79 18 BBG F 0.96 0.95 4 EBG F 0.80 0.78 19 UT L 0.92 0.90 5 PBL L 0.79 0.77 20 TTB L 0.91 0.90 6 UMB L 0.78 0.70 21 UNIBANK L 0.89 0.88 7 ADB L 0.76 0.74 22 GTB F 0.87 0.85 8 UBA F 0.72 0.68 23 ACCESS F 0.87 0.85 9 STANBIC F 0.71 0.68 24 SG-SSB F 0.87 0.85 10 IBG F 0.70 0.68 25 HFC L 0.85 0.84 11 NIB L 0.68 0.65 26 ICB F 0.85 0.81 12 BSIC F 0.46 0.41 27 AMAL/BOA F 0.85 0.82 13 ENERGY F 0.29 0.28 28 FBL L 0.84 0.82 14 ROYAL F 0.21 0.21 29 METRO/BPI F 0.83 0.80 15 - - - - a The ranking is from 1 - 29. The value 1 represents the best performing bank and 29 the least performing bank. OWN-Ownership; F-Foreign bank; L-Local bank; AM-Arithmetic mean; GM-Geometric mean In terms of ownership, local banks have been more profit efficient (0.8375) than foreign banks (0.7506) over the years on average. However, a Wilcoxon rank sum test (W=89.5, p=0.5945) shows that the difference is infinitesimal and insignificant. Also, the scores of the non-parametric DEA approach are compared with two traditional profitability ratios, ROA and ROE in Table 10. This is to determine whether DEA scores give similar rankings as the traditional ratios. Whereas the DEA scores are the geometric means of the profit efficiency scores, ROA and ROE scores are the arithmetic means of each bank for the entire study period (2000 to 2013). The geometric mean instead of arithmetic mean is used for the DEA scores because DEA scores are benchmarked scores which may be biased when arithmetic mean is used. University of Ghana http://ugspace.ug.edu.gh 82 Table 10 : Comparison of DEA and traditional profitability ratios DEA ROA ROE Bank Score Rank Score Rank Score Rank ACCESS 0.8531 8 0.0489 7 0.2207 12 ADB 0.7403 21 0.0373 11 0.2084 15 AMAL/BOA 0.8215 13 -0.0059 27 -0.1905 29 BARODA 1.0000 1 0.0680 2 0.2487 11 BBG 0.9484 4 0.0684 1 0.6113 2 BSIC 0.4139 27 -0.0492 29 -0.0969 28 CAL 0.7926 17 0.0460 9 0.3006 10 EBG 0.7824 18 0.0526 5 0.4978 5 ENERGY 0.2781 28 0.0304 12 0.1812 17 FABL 0.7693 19 0.0155 20 0.2100 14 FBL 0.8230 12 0.0123 21 0.1506 21 GCB 1.0000 2 0.0458 10 0.4372 7 GTB 0.8496 9 0.0052 24 0.0821 24 HFC 0.8368 11 0.0290 14 1.0565 1 IBG 0.6799 25 0.0099 22 0.1651 18 ICB 0.8074 14 0.0301 13 0.1609 19 METRO/BPI 0.8000 15 -0.0210 28 -0.0289 26 NIB 0.6530 26 0.0288 15 0.1209 22 PBL 0.7668 20 0.0230 17 0.3512 8 ROYAL 0.2083 29 0.0068 23 0.1201 23 SCB 0.9517 3 0.0622 3 0.5479 3 SG-SSB 0.8454 10 0.0530 4 0.3376 9 STANBIC 0.6821 23 0.0486 8 0.4507 6 TTB 0.8994 5 0.0512 6 0.5037 4 UBA 0.6811 24 0.0051 25 -0.0485 27 UMB 0.7048 22 0.0197 19 0.1956 16 UNIBANK 0.8755 7 0.0008 26 0.0180 25 UT 0.8988 6 0.0261 16 0.2144 13 ZENITH 0.7935 16 0.0227 18 0.1550 20 a The ranking is from 1 - 29. The value 1 represents the best performing bank and 29 the least performing bank. Wilcoxon signed rank test(DEA-ROA):V=212, p=0.8462 Wilcoxon signed rank test (DEA-ROE):V=211,p=0.8643 A careful comparison of DEA and the other profitability ratios reveal that the techniques generally do not present similar rankings. Except for SCB, the ranks of most of the other banks are not precisely the same. Out of the 29 banks, 14 are ranked highly for DEA than ROA and 15 for DEA University of Ghana http://ugspace.ug.edu.gh 83 than ROE. A correlation matrix of the three techniques as presented in Table 11 also shows that the techniques are positively correlated but the association is generally weak. Table 11: Correlation between DEA, ROA and ROE DEA ROA ROE DEA 1 ROA 0.4633 1 ROE 0.3372 0.6697 1 This is probably because, DEA provides a more holistic picture of a bank’s performance by taking simultaneous account of their inputs and outputs than the profitability ratios. A Wilcoxon signed rank test (see Table 10) however, reveal no significant differences in the pairwise comparison of DEA scores and the two profitability ratios. This implies that DEA scores provide similar conclusions as the profitability ratios, but incorporates other factors into its computation. The weak correlation between the DEA and the profitability ratios suggests however that using only profitability ratios to assess bank performance may not capture their true overall performance. 5.4 Effect of Capital Structure on Profit Efficiency To investigate the marginal effect of capital structure on the profit efficiency of banks in Ghana (the second objective of this study), a bootstrapped truncated regression is estimated. This involves regressing profit efficiency on capital structure and other control variables. To do this, the study first tests for the degree of multicollinearity among the independent variables by including a correlation matrix in Tables 12 and 13. University of Ghana http://ugspace.ug.edu.gh 84 Correlation Matrix Table 12: Pearson Correlation ERC ECAP 𝑬𝑪𝑨𝑷𝟐 BS SG REG OWN TE ERC 1 ECAP -.116* 1 𝑬𝑪𝑨𝑷𝟐 -0.078 .905** 1 BS -0.037 -0.103 -.130* 1 SG 0.061 .167** .158** -0.053 1 REG -.183** .230** .175** .641** 0.109 1 OWN -0.009 .196** .181** -0.017 .168** .172** 1 TE .362** -0.103 -.125* 0.059 -0.066 -0.102 -0.078 1 * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). Table 13: Spearman's rho ERC ECAP 𝑬𝑪𝑨𝑷𝟐 BS SG REG OWN TE ERC 1 ECAP -.157** 1 𝑬𝑪𝑨𝑷𝟐 -.176** .989** 1 BS -0.044 0.009 -0.009 1 SG 0.084 -0.055 -0.075 -.116* 1 REG -.193** .247** .237** .667** 0.062 1 OWN 0.013 .152** .164** 0.005 0.107 .217** 1 TE .349** -0.059 -0.078 0.068 -0.03 -0.099 -0.026 1 ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). The correlation matrix shows a positive and statistically significant correlation between profit efficiency (ERC) and technical efficiency (TE) but a negative and statistically significant correlation with equity capital (ECAP) and regulation (REG). Bank size (BS) and Ownership (OWN) correlate negatively with profit efficiency but these results are insignificant. Sales growth has a positive, but insignificant correlation with profit efficiency. Equity capital has a positive and statistically significant correlation with sales growth, regulation and ownership, but has a University of Ghana http://ugspace.ug.edu.gh 85 statistically insignificant negative correlation with size and technical efficiency. Size correlates positively with regulation and technical efficiency and negatively with sales growth and ownership. With the exception of regulation, the correlation between size and the other variables are insignificant. Sales growth has a positive correlation with regulation and ownership but a negative correlation with technical efficiency. The correlation between sales growth and regulation is significant but insignificant for ownership and technical efficiency. Regulation is statistically positively correlated with ownership but negatively correlated with technical efficiency. Ownership correlates negatively with technical efficiency but this is insignificant. In the specification of a model, independent variables that are correlated (have values exceeding 0.50) cannot be placed in the same model (Cohen, 1988). This is because they play similar roles and including them in the same model may make the regression sensitive to small changes in specification. With this, the confidence intervals for the parameters will be wide and the conclusions that might be made from the significance tests may be inappropriate (Brooks, 2009). To resolve this issue of multicollinearity, it is required that one of the highly correlated independent variables be dropped from the model to avoid misspecification. This can be done through the use of the Variance Inflation Factor (VIF) which quantifies the severity of multicollinearity through a stepwise procedure. However, some econometricians have argued that, dropping an independent variable that belongs to the population model can lead to biases (Wooldridge, 2012). Also, setting an arbitrary cutoff point for the VIF above which multicollinearity is a problem is questionable and not particularly useful. Thus, if the model is adequate in terms of the coefficients having the appropriate signs and being of plausible magnitude, the multicollinearity problem can be ignored. Further, University of Ghana http://ugspace.ug.edu.gh 86 multicollinearity is less a problem with the model than with the data (Brooks, 2009). In view of these, the multicollinearity between regulation and size (0.64) is ignored. Findings from Bootstrapped Truncated Regression Summary statistics of the variables used in the bootstrapped truncated regressions are presented in Table 14 and the results for the regressions in Tables 15 and 16. Table 14: Summary Statistics of variables used in truncated regression N Mean Std. Dev Min Max ERC 304 0.82 0.22 0.18 1 ECAP 304 0.15 0.12 -0.13 0.97 304 0.04 0.08 0 0.94 BS 304 19.17 1.54 13.69 22.25 SG 289 44.4 62.18 -134.94 507.11 REG 304 0.09 0.018 0.06 0.10 OWN 304 0.54 0.5 0 1 TE 304 0.9 0.15 0.31 1 HHI 304 0.09 0.03 0.06 0.15 BI 304 -2.40 1.72 -7.32 0.01 CONG 304 0.37 0.48 0 1 ROA 304 0.03 0.04 -0.21 0.18 ROE 304 0.27 0.7 -4.4 9.58 The original results of the bootstrapped truncated regression shows that, the effect of equity capital and equity capital squared on profit efficiency are both statistically insignificant at 5% (see Table 20 Appendix C) but the first derivative ti ti dECAP dERC , , is statistically significant and negative at the sample mean of 0.15 (see Table 15). This is consistent with the predictions of the trade-off and the agency cost theories that lower equity (higher leverage) is related to improved performance (profit efficiency). It is also in line with the findings of Berger and Bonaccorsi di Patti (2006) on U.S commercial banks and Fosu (2013) on South African firms. The result also satisfies hypothesis University of Ghana http://ugspace.ug.edu.gh 87 one (H1) and implies that, in Ghana, capital structure is important in determining the profit efficiency of banks. The more leverage a bank uses relative to equity, the more profit efficient it becomes. This is probably because, leverage reduces their tax burdens as indicated by the trade- off theory and helps them give out more loans from which they generate higher interest incomes. It is also possible that leverage reduces the agency costs predominant in these banks leading to higher profit efficiency. Table 15: Regression Results: Profit Efficiency and Capital Structure (At Sample Mean) ERC ROA ROE (Intercept) 0.1951 - -0.9101 (0.1767) - (0.5945) ECAP -0.3299 0.1719 *** -1.0497 (0.1626) * (0.0467) (1.3671) 0.3917 -0.0719 3.3223 (0.337) (0.045) (2.088) BS 0.0192 0.0022 0.073 (0.0107) . (0.0068) (0.0484) SG 0.0004 0.0001 0.0002 (0.0002) * (0) (0.0003) REG -0.015 -0.0027 -0.0437 (0.0054) ** (0.0025) (0.0281) OWN 0.0143 - -0.025 (0.0234) - (0.0944) TE 0.5421 0.0493 0.6177 (0.0823) *** (0.0215) * (0.2621) * Sigma 0.191 - - (0.008) *** - - Log Likelihood 68.46 R squared 0.1603 0.0562 F-statistic 8.1157*** 2.3898* Profit Efficiency estimates are truncated at zero. ‘***’ 0.001 ‘**’ 0.01 ‘*’0.05 ‘.’ 0.1 Standard errors are white robust to serial correlation and heteroscedasticity Coefficients have been bootstrapped University of Ghana http://ugspace.ug.edu.gh 88 To buttress the main argument of the study that, the proxy for performance may be a cause of the inconclusive results in the CSFP link, two financial ratios, ROA and ROE were used. The Hausman specification test, a test performed to determine whether the researcher should choose the fixed effects or random effects model estimation prior to running a panel regression was used. The rule of thumb for this test is that when p < 0.05, there is a correlation between the error terms and the explanatory variables and a fixed effects estimation is adopted else the random effects estimator is more appropriate. The test showed that, a fixed effects estimator is appropriate for using ROA as the dependent variable and a random effects estimator for ROE (see Appendix C for results on the Hausman Test). The panel regression results suggest a statistically significant and positive relationship between equity capital and ROA (see Table 20 Appendix C). A similar relationship is documented when the first derivative ti ti dECAP dROA , , is taken and the model is rerun at the sample equity capital mean of 0.15 (see Table 15). The results imply that, profitable banks in Ghana rely less on leverage than equity, which is consistent with the pecking order theory of capital structure. It is also in tandem with the results documented by Amidu (2007) in his study of 19 banks in Ghana. However, equity capital has a negative and insignificant impact on ROE. This is in line with Modigliani and Miller’s irrelevance theory of capital structure. It is evident therefore that, although ROA and ROE are created from the same financial statements, they provide differing results on the relationship between capital structure and bank performance. This is probably because each of these ratios examine only a part of the bank’s operations and thus ineffective for assessing the overall performance of banks (Paradi & Zhu, 2013). In view of this, employing financial ratios as proxies University of Ghana http://ugspace.ug.edu.gh 89 for performance in the CSFP link may not be appropriate. Instead, a measure that takes simultaneous account of the inputs and outputs used in the bank such as profit efficiency is effective. From Table 15, bank size has a positive effect on profit efficiency and significant at 10%. This means that the larger a bank is in Ghana, the higher its profit efficiency. One reason for this relationship may be the fact that, bigger banks enjoy economies of scale, are more diversified and develop better means and prospects for diversifying risk than smaller banks (Tecles & Tabak, 2010; Isik & Hassan, 2002). Sales growth also has a positive and statistically significant relationship with profit efficiency. This implies that, in Ghana, profit efficient banks gain from their core activity which is taking deposits and granting loans. In other words, they are operationally efficient. As Gatsi (2012) posit, the strong link between sales growth of banks and performance is probably as a result of improvement in the general economic climate thereby translating into the banking sector. Regulation as measured by the natural logarithm of the minimum capital requirement has a negative and statistically significant impact on profit efficiency. This suggests that higher capital requirements decreases the profit efficiency of banks in Ghana. This finding is consistent with Pasiouras et al (2009). The decrease in profit efficiency is probably because, banks substitute loans which generate interest incomes with equity which does not bear interest incomes to meet the higher requirements (VanHoose, 2007). This finding is different from a similar work by Bokpin (2013) who saw that regulation has a positive effect on profit efficiency of banks in Ghana. Perhaps, the difference is as a result of the differing proxies for measuring regulation. Whereas, Bokpin (2013) uses the capital adequacy ratio, this study uses the minimum capital requirement. University of Ghana http://ugspace.ug.edu.gh 90 Similar to Bokpin (2013), a positive but statistically insignificant relationship is found between ownership and profit efficiency. This means that, the type of ownership is irrelevant in determining the profit efficiencies of banks in Ghana. This supports the argument by Molyneux and Thornton (1992) and Bourke (1989). The results imply however that, foreign banks are more profit efficient than local banks which is consistent with studies by Berger et al. (2010) and Bonin et al. (2005). This is probably because foreign banks are well diversified and possess better technologies than local banks (Berger et al, 2000). Technical efficiency has a positive and statistically significant impact on profit efficiency. This suggests that in Ghana, banks that are capable of producing more outputs (loans and advances, investments) given fixed inputs (deposits, fixed assets and labour) or use lesser inputs to produce given outputs are more profit efficient. Miller and Noulas (1996) documented same findings in their study of technical efficiency of large bank production. 5.5 Capital Structure, Competition and Profit Efficiency. To achieve objective three, which is examining the marginal effect of capital structure on profit efficiency amidst the degree of competition, a bootstrapped truncated regression was run. In a unique contribution of this study, two measures of competition, HHI and BI were used to proxy the levels of competition in the Ghanaian banking industry (see Appendix D for yearly HHI and BI scores). The results for using HHI are presented in models 1 to 3 in Table 16 and models 4 to 6 in Table 17 for BI. Models 1 and 4 present the original results of the bootstrapped truncated regression. In models 2 and 5, the results for the first derivative of profit efficiency with respect to equity capital at a mean of 0.15 for equity capital, 0.09 for HHI and 2.40 for BI are presented. University of Ghana http://ugspace.ug.edu.gh 91 Models 3 and 6 show the first derivative of profit efficiency with respect to competition for HHI and BI respectively. The partial effect of equity capital on profit efficiency is significant at 10% and negative (see model 2). The partial effect of competition measured by the degree of concentration, HHI, is negative and insignificant (see model 3). This does not provide support for hypothesis two (H2) of this study. The result implies that, bank concentration has a negative impact on profit efficiency. This means that, the more monopolistic the banking industry becomes, the lesser its ability to convert its resources into profits. This is probably because of the inefficiencies and deadweight losses attributable to monopolistic markets (Baye, 2010). In less concentrated industries, firms are able to utilize the resources in order to stay competitive. Table 16: Regression Results : Capital Structure, Competition and Profit Efficiency Herfindahl Hirschman Index (HHI) Model 1 Model 2 Model 3 (Intercept) 0.6402 * 0.6306 . 0.6402 * (0.324) (0.3233) (0.324) ECAP -0.5521 -0.3162 . -0.5521 (0.574) (0.1627) (0.574) 0.4257 0.4257 0.4257 (0.3839) (0.3839) (0.3839) BS 0.0169 0.0169 0.0169 (0.0108) (0.0108) (0.0108) SG 0.0004 * 0.0004 * 0.0004 * (0.0002) (0.0002) (0.0002) REG -0.0304 ** -0.0304 ** -0.0304 ** (0.011) (0.011) (0.011) OWN 0.0213 0.0213 0.0213 (0.0239) (0.0239) (0.0239) TE 0.5399 *** 0.5399 *** 0.5399 *** (0.0821) (0.0821) (0.0821) HHI -1.9114 -1.9114 -1.7311 (1.2948) (1.2948) (1.0973) ECAP*HHI 1.2024 1.2024 1.2024 (5.3271) (5.3271) (5.3271) Sigma 0.1902 *** 0.1902 *** 0.1902 *** University of Ghana http://ugspace.ug.edu.gh 92 (0.0079) (0.0079) (0.0079) Log Likelihood 69.77 69.77 69.77 HHI was measured using squared market shares of total Assets. ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1. Coefficients have been bootstrapped Further, unlike in previous CSFP studies in the banking industry, this study attempts to determine the impact of the interaction between competition and capital structure on profit efficiency. This is found to be positive, but insignificant as shown by ECAP*HHI. This suggests that, the effect of capital structure on profit efficiency increases as bank concentration increases. In other words, the negative effect of equity capital on profit efficiency increases as competition in Ghana’s banking industry decreases. Thus, the benefits banks gain from using more leverage (deposits, borrowings, and accruals) in financing their activities become higher in periods when the banking industry is less competitive. Using BI as a proxy for competition, the impact of the interaction between capital structure and competition on profit efficiency is negative and insignificant (see Table 17). This finding suggests that the effect of equity capital on profit efficiency increases as the level of competition in the banking industry reduces. But this effect is insignificant. Interestingly, the partial effect of competition on profit efficiency is negative and significant (see model 6) contrary to the positive link that seems more intuitive in most industries (Casu & Girardone, 2009). The results suggest that, higher levels of competition, reduce the profit efficiencies of banks in Ghana. Ideally, it is expected that increases in competition would precipitate increases in profit efficiency. This is because with competition, managers are forced to exert more effort and allocate resources efficiently in order to reach market equilibrium. However, in this study, competition is negatively University of Ghana http://ugspace.ug.edu.gh 93 related to profit efficiency. This is probably because, heightened competition is associated with shorter relationships between customers and banks (Boot & Schmeits, 2006; Petersen & Rajan, 1995) since the tendency for customers to switch to other banks upsurges in competitive environments. Table 17: Regression Results : Capital Structure, Competition and Profit Efficiency Boone Indicator Model 4 Model 5 Model 6 (Intercept) 0.1623 0.1521 0.1623 (0.1829) (0.1828) (0.1829) ECAP -0.7619 . -0.2288 -0.7619 . (0.4065) (0.1646) (0.4065) 0.4502 0.4502 0.4502 (0.3464) (0.3464) (0.3464) BS 0.0231 * 0.0231 * 0.0231 * (0.0108) (0.0108) (0.0108) SG 0.0004 * 0.0004 * 0.0004 * (0.0002) (0.0002) (0.0002) REG -0.0166 ** -0.0166 ** -0.0166 ** (0.0054) (0.0054) (0.0054) OWN 0.0158 0.0158 0.0158 (0.0231) (0.0231) (0.0231) TE 0.5262 *** 0.5262 *** 0.5262 *** (0.0815) (0.0815) (0.0815) BI 0.0026 0.0026 -0.0223 ** (0.0164) (0.0164) (0.0082) ECAP*BI -0.1658 -0.1658 -0.1658 (0.1312) (0.1312) (0.1312) Sigma 0.1886 *** 0.1886 *** 0.1886 *** (0.0079) (0.0079) (0.0079) Log Likelihood 72.21 72.21 72.21 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 Coefficients have been bootstrapped The effect of this is that, information asymmetries between banks and customers increase, which causes an increase in the costs incurred by the banks in monitoring and screening borrowers. Further, given a higher unbanked population of over 70 % (Akosah, 2013) relative to the number University of Ghana http://ugspace.ug.edu.gh 94 of banks in Ghana, to survive in the competitive financial landscape, banks incur greater expenses in maintaining and attracting customers through greater marketing efforts such as investments in Automated Teller Machines (ATMs) and advertisements. These may cause declines in their profit efficiencies. 5.6 Capital Structure, Conglomeration and Profit Efficiency. The results for objective four (that is investigating the impact of the interaction between capital structure and conglomeration on profit efficiency) are presented in Table 18. Model 7 presents the original results of the bootstrapped truncated regression. Models 8 and 9 depicts the results of the partial effects of equity capital and conglomeration on profit efficiency respectively. Consistent with Chronopoulos et al (2011) conglomeration has a positive impact on profit efficiency. However, in this study the impact is insignificant (model 9). The relationship implies that in Ghana, banks with subsidiaries are more efficient in generating profits than their counterparts that are focused. This may be due to the economies of scale and scope benefits associated with diversification (Chronopoulos et al., 2011; Vander Vennet, 2002). The interaction between conglomeration and profit efficiency has a negative and insignificant impact on profit efficiency. This suggests that, the effect of equity capital on profit efficiency reduces if the bank is a conglomerate than if it is focused. University of Ghana http://ugspace.ug.edu.gh 95 Table 18: Regression results: Capital Structure, Conglomeration and Profit Efficiency Model 7 Model 8 Model 9 (Intercept) 0.2667 0.2635 0.2667 (0.1792) (0.1798) (0.1792) ECAP -0.1961 -0.3967 * -0.1961 (0.2702) (0.1824) (0.2702) 0.1417 0.1417 0.1417 (0.3569) (0.3569) (0.3569) BS 0.0149 0.0149 0.0149 (0.011) (0.011) (0.011) SG 0.0004 * 0.0004 * 0.0004 * (0.0002) (0.0002) (0.0002) REG -0.0156 ** -0.0156 ** -0.0156 ** (0.0053) (0.0053) (0.0053) OWN 0.0243 0.0243 0.0243 (0.0239) (0.0239) (0.0239) TE 0.5166 *** 0.5166 *** 0.5166 *** (0.0823) (0.0823) (0.0823) CONG 0.1243 * 0.1243 * 0.0257 (0.0572) (0.0572) (0.0291) ECAP*CONG -0.6569 -0.6569 -0.6569 (0.4298) (0.4298) (0.4298) Sigma 0.1892 *** 0.1892 *** 0.1892 *** (0.0079) (0.0079) (0.0079) Log Likelihood 71.21 71.21 71.21 Signif. Codes: ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 5.7 Reverse Causation Between Profit Efficiency and Capital Structure. To achieve objective five which is testing whether there is a reverse causation between profit efficiency and capital structure using two competing hypotheses, the two stage least squares (2SLS) technique is used. The technique is used to resolve the potential endogeneity problems likely to occur when a bi- causal relationship exists between a dependent and an independent variable. The technique requires the models in the simultaneous-equation to be either just identified or over identified. To test the identification of the models, the order condition by Ramanathan (1995, pp.666) is used. The order condition suggests that, the equation models University of Ghana http://ugspace.ug.edu.gh 96 (equations 22 and 23) are econometrically identified since an appropriate number of the specified variables are excluded from each of the equations. Table 19 presents results for the predictions of the franchise value and efficiency-risk hypotheses of reverse causation from profit efficiency to capital structure. Table 19: Results for reverse causation from profit efficiency to capital structure ECAP (Intercept) 0.3682 *** (0.0986) ERC -0.0475 (0.0787) BS -0.02 *** (0.0057) SG 0.0002 * (0.0001) REG 0.0136 *** (0.003) AS -0.3076 (0.2434) The results show a negative and statistically insignificant relationship between profit efficiency and equity capital, thereby providing support for the efficiency-risk hypothesis. This is similar to the finding of Margaritis and Psillaki (2010) on French manufacturing firms. This means that in Ghana, more profit efficient banks choose lesser equity capital than their counterparts that are less profit efficient. This may be because the higher profit efficiencies may translate into higher expected returns which substitute for equity capital to manage potential risks and costs (for example bankruptcy, financial distress) that the bank may be faced with. 5.8 Chapter Summary This chapter presented the findings and the discussions of each research objective. First, it provided descriptive statistics on all the variables used in the estimation of each objective. The University of Ghana http://ugspace.ug.edu.gh 97 first objective was achieved by estimating the profit efficiency of the banking industry relative to each yearly frontier. It was found that banks in Ghana over the fourteen year period were 79% profit efficient on average. The profit efficiency estimates were then compared with two profitability ratios, ROA and ROE to determine the extent to which they agreed on the performance of a bank. The findings revealed that the methods agreed weakly. The other objectives were achieved by regressing the estimated profit efficiencies on capital structure, the degree of competition in the industry, and a bank’s ownership of a subsidiary. For the reverse causation, capital structure was regressed on profit efficiency and other control variables by using a two stage least squares (2SLS) method. University of Ghana http://ugspace.ug.edu.gh 98 CHAPTER SIX SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 6.1 Introduction This chapter is categorized into three sub-headings. The first sub-heading which is the summary clearly highlights what the study sought to achieve and presents the techniques employed in achieving the research objectives and the findings. Subsequently, conclusions on the study are provided, from which essential recommendations for practice, policy and further research are made. 6.2 Summary The main objective of this study was to investigate the impact of capital structure decisions of banks in Ghana on their performance by employing a non-parametric profit efficiency technique as the indicator of bank performance. Additionally, reverse causation between capital structure and profit efficiency was tested in the Ghanaian banking industry, by using two competing hypotheses of Berger and Bonaccorsi di Patti (2006) - the efficiency-risk and the franchise value hypotheses. Previous studies that considered the CSFP nexus mostly used profitability ratios as proxies for performance and did not recognize the possible endogeneity problems that could occur as a result of the bi-causal relationship between capital structure and firm performance. This study is therefore the first to employ profit efficiency to measure performance in the Ghanaian banking industry. It is also the first to consider the possible bi-causal relationship between capital structure of banks and profit efficiency. University of Ghana http://ugspace.ug.edu.gh 99 In this study, simultaneous equations modelling was used to assess the reverse causation between capital structure and profit efficiencies of banks in Ghana. This helped to understand the reasons for the mixed empirical results seen in earlier studies on the CSFP nexus. Interaction effects of capital structure and competition in the industry, as well as between capital structure and conglomeration were also further examined. This was to determine whether the level of competition and conglomeration of banks significantly affects the relationship between capital structure and bank profit efficiency, which is also the first to be investigated in the capital structure- efficiency studies in the banking industry. The study used an unbalanced panel data of 26 banks from the year 2000 to 2013 to achieve the research objectives. Data was sourced from BOG and cross-validated with annual reports of universal banks in Ghana to ensure consistency and to handle missing data. R-based statistical packages- Frontier Efficiency Analysis with R (FEAR) by Wilson (2008) and Benchmarking by Bogetoft and Otto (2011) were the primary software used for the analysis. However, MaxDEA Pro version 6 was also used together with other default packages in the R software. The findings of the study included the following: a. On average, about 79% of the profits that can be earned by a best practice bank under very similar conditions are earned by banks in Ghana. By implication, banks in Ghana are on average, 21% profit inefficient in their operations. The production of inappropriate output and input mix that maximizes cost while minimizing revenues may be the cause of these inefficiencies. b. A comparison of the non-parametric profit efficiency with two predominantly used profitability ratios as measures of bank performance revealed that profit efficiency weakly agrees with the ROA and ROE of banks in Ghana. This is probably because whereas profit University of Ghana http://ugspace.ug.edu.gh 100 efficiency mirrors overall or total-factor performance, profitability ratios only reflect factor-specific performance. In other words, a profitable bank may not necessarily be profit efficient. c. In the assessment of the CSFP link, a significantly negative relationship was found between profit efficiency of banks and capital structure (equity capital). This is consistent with the views of the agency cost hypothesis of capital structure. However, when ROA and ROE were used as measures of bank performance, whereas a positive and significant relationship was found for ROA, a negative but insignificant relationship was found for ROE. This shows that, albeit both profitability ratios, they gave different and conflicting results, indicating that researchers need to be wary in their interpretations when using different accounting and financial ratios. d. A negative and insignificant relationship was also found between bank concentration, measured by the HHI, and profit efficiency. A negative and significant relationship was, however, found between competition, measured by the BI and profit efficiency. This could mean that researchers need to be careful with the proxy measure used for competition as these could lead to different and misleading findings. This study endorses the BI as a better measure of competition. e. For the results of the interaction between capital structure and competition, a positive relationship was found when HHI was used, whereas a negative relationship was identified when BI was adopted as measures of industry competition. It must be stressed, however, that in all two cases the interaction effects of competition and capital structure on profit efficiency was not statistically significant. University of Ghana http://ugspace.ug.edu.gh 101 f. For the interactions between capital structure and conglomeration, its effect on profit efficiencies of banks in Ghana was found to be negative and insignificant. This implies that, having a subsidiary which engages in other activities besides those permissible as banking activities in Ghana does not necessarily influence the link between capital structure and profit efficiency. g. Finally, the reverse causation between capital structure and profit efficiency, was found to support the efficiency-risk hypothesis. This was because a negative relationship was found between equity capital and profit efficiency. This means that, on average, in Ghana, profit efficient banks employ more debts than equity in their operations. 6.3 Conclusions of the Study The findings of the study reveal that average profit efficiency levels in the banking industry are quite high, indicating an industry which is quite sound and economically profitable. More room, however, exists for improvement based on effective regulatory and policy controls. It should not be seen as the prerogative of only banks to improve profit efficiency. Regulators need careful consideration of the effects of policy directions on the level of profit efficiency in the industry. This is because a more profit efficient banking industry has implications for economic growth since banks will reduce waste in their intermediation roles. The second major finding of the study was observed when the three profitability measures were compared. When ROA, ROE and DEA measures of performance were compared, these three measures provided slightly different conclusions. This is probably because whereas ROA and ROE are single-factor ratios which shows how much of a single output (income) can be generated from only one input (either assets or equity capital), DEA scores provide better and more holistic picture University of Ghana http://ugspace.ug.edu.gh 102 of performance. ROA and ROE are more advantageous when simple profitability analyses are required. DEA, on the other hand, provides a comprehensive profitability assessment that takes into consideration the multiple inputs and multiple output dimensions of bank performance. Furthermore, DEA computations, by their nature, not only provide information on bank profitability, but provides also, useful benchmarks or guidelines for policies towards achieving higher levels of profit efficiency. Coupled with this, DEA provides empirically and theoretically grounded models for both cross sectional and time series measurement of performance, whereas these profitability ratios mainly account for cross sectional performance which ignores any time dependencies. Consequently, DEA provides more holistic insights for policy-oriented decisions than ROA and ROE. The mixed empirical results that were seen by Berger and Bonaccorsi di Patti (2006) when profitability ratios are employed, was also proved in this study. This is as a result of the divergent conclusions on the effects of capital structure on profitability when ROE and ROA were used as proxies for performance. Whereas the relationship between capital structure and ROA supports the pecking order theory, ROE supported the Modigliani and Miller irrelevance theory. The differences in the conclusions when the two financial ratios are used makes it difficult for effective policy formulation. However, because DEA is capable of encapsulating both the asset and equity dimensions of firm profitability, it provides a better and relatively more consistent conclusions to aid in effective policy formulation. Consequently, the relationship between profit efficiency and capital structure (equity capital) can be concluded to be negative. This follows the agency cost hypothesis and the trade-off theory of capital structure. Similar results were found by Berger and Bonaccorsi di Patti (2006) and Margaritis and Psillaki (2010) who employed Stochastic Frontier University of Ghana http://ugspace.ug.edu.gh 103 Analysis (SFA) and DEA respectively in their study. The implication of the result is that banks in Ghana who use more debts/leverage are able to generate more loans and investments which generates more interest incomes and makes them more profit efficient. Considering the reverse causation between profit efficiency and capital structure, the evidence from the study is that not only does more leverage result in higher profit efficiencies, but more profit efficient banks tend to also use more leverage. A clear empirical grounding has therefore been provided to believe leverage is critical in sustaining banks in Ghana. Therefore, policies that have the possibility of affecting the leverage decisions of banks in Ghana need careful rethink before they are enacted. This is because such policies have the possibility of affecting (reducing) profit efficiency in the industry, which may have a multiplier effect on economic growth in the country. Also evident in this study was the contradictory results between competition and profit efficiency when HHI and BI were employed. It is possible that the relationship between competition and profit efficiency is nonlinear. As seen, when the HHI was used, it is possible that competition will positively affect profit efficiency at some levels of competition, but at other levels of competition, a negative relationship may be manifested as seen by the BI. However, from current results, the direct linear impact of leverage decisions and policies of banks in Ghana has little to do with the level of competition in the industry. Therefore, banks’ leverage decisions can provide consistent effects on profit efficiency irrespective of the level and nature of competition. Finally, there is no empirical evidence in this study to conclude that the effects of capital structure decisions on profit efficiency in Ghana depends on the existence of subsidiary business operations. Therefore, owning a subsidiary firm is not a prerequisite for a larger impact of capital structure University of Ghana http://ugspace.ug.edu.gh 104 decisions on the level of profit efficiency of banks in Ghana. This notwithstanding, it is possible that owning such conglomerates would directly lead to higher levels of profit efficiency. 6.4 Recommendations Findings and conclusions of this study provide important policy implications and recommendations for practice, and further research. These are summarized as follows: For Practice and Policy: a. First, banks were seen to be 21% profit inefficient on average. It is necessary for policy makers and management to establish policies and procedures that will cut wastages in inputs whiles also strengthening their risk management practices. These may include better screening of borrowers and effective monitoring of loan performance to reduce and avoid adverse selection and moral hazards. These banks can also cut down expenses in operations by encouraging customers to use technologies like ATMs, mobile and internet banking in most of their bank transactions. b. Also in the study were substantial disagreements in the results of the DEA and two profitability ratios as proxies for performance. Unlike these profitability ratios, the results of the DEA non-parametric profit efficiency incorporates the multiple inputs and multiple outputs employed by these banks in their operation. Consequently, banks are advised to use profitability ratios for mostly simple analysis, but for more inferential and holistic analysis, DEA provides a better assessment. These performance measures should be seen as complementary measures of profitability rather than alternatives when evaluating performance over time. University of Ghana http://ugspace.ug.edu.gh 105 c. In the CSFP link, DEA and profitability ratios provided different conclusions. Even ROA and ROE which are both profitability ratios presented contradictory results on the nature of the relationship between capital structure and bank performance. To avoid complexities that can occur with the reliance on profitability ratios, it is recommended that management and policy makers routinely monitor decisions based on profitability ratios whiles also using more advanced performance measurement techniques like DEA. d. Another key finding in this study is how essential leverage decisions are to bank profit efficiency in Ghana. Therefore, the regulator must be careful when formulating policies that can adversely affect the leverage decisions on banks. This is key because leverage is important in sustaining banks in Ghana. e. On the negative nexus found between competition and profit efficiency, policy makers are advised to be careful in setting policies intended to increase competition in the banking industry. This is because information asymmetries and lower credit evaluations of borrowers may result from heightened competition that weakens the relationship between banks and customers. As a result, monitoring and screening costs may increase, which may lead to a decrease in profit efficiency of these banks. For example, in the US, increased competition resulted in banks lowering their lending stringency to subprime mortgage markets. This instigated the global financial crisis of 2007 to 2009. For Further Research: a. This study focused on estimating the profit efficiencies of the Ghanaian banking industry. A cross-country banking efficiency assessment of African countries can be undertaken to understand profit efficiencies in the banking industries of these countries. This way, a University of Ghana http://ugspace.ug.edu.gh 106 bigger picture can be observed given that some of these countries are forming currency unions or political or economic blocks. b. A decomposition of profit efficiency into technical and allocative efficiencies can also be considered to help determine the real source of performance so that managers can improve performance via those sources. One can decompose profit efficiency into revenue and cost efficiencies or better still, along the lines of Portela and Thanassoulis (2005), use the geometric distance function to estimate profit efficiency. c. This study looked at static profit efficiency but further studies can be extended to profit productivity dynamics which can reveal patterns and trends of performance over time. For instance, a study can adopt the principles of Maniadakis and Thanassoulis (2004) cost productivity change to profit productivity change. d. It is also recommended that further studies should clarify if the mixed results in the CSFP nexus is truly due to the employment of other profitability ratios as performance measures without the consideration of possible reverse causation as stated by Berger and Bonaccorsi di Patti (2006). e. Further research should also consider other performance assessment techniques (e.g.Thick Frontier Approach, SFA, Bayesian Efficiency) besides profitability ratios in investigating the impact of capital structure on the performance of other industries in Africa. f. Further empirical analyses may also be targeted at exploring the possible nonlinear relationship between competition and profit efficiency by applying nonlinear regression estimation. This is in light of the contradictory results when two different competition methods were applied. University of Ghana http://ugspace.ug.edu.gh 107 REFERENCES Aboagye, A. Q., Akoena, S. K., Antwi‐Asare, T. O., & Gockel, A. F. (2008). Explaining Interest Rate Spreads in Ghana*. African Development Review, 20(3), 378-399. Abor, J. (2005). The effect of capital structure on profitability: an empirical analysis of listed firms in Ghana. The Journal of Risk Finance, 6(5), 438–445. Abor, J. (2007). Debt policy and performance of SMEs: Evidence from Ghanaian and South African firms. The Journal of Risk Finance, 8(4), 364–379. Addison, E. (2003). Managing The Constraints To Development Finance: Is Universal Banking The Solution. 4th ISSER-Merchant Bank Annual Economic Lectures. Adjei-frimpong, K., & Gan, C. (2014). Cost efficiency of ghana’s banking industry: a panel data analysis. The International Journal of Business and Finance Research, 8(2), 69–86 Akhigbe, A., & McNulty, J. E. (2003). The profit efficiency of small US commercial banks. Journal of Banking & Finance, 27(2), 307–325. Akomea, S. Y., & Adusei, M. (2013). Bank recapitalization and Market Concentration in Ghana’s banking Industry: a Herfindahl-Hirschman Index. Global Journal of Business Research, 7 (3) Akosah, N. K. (2013). Dynamics of Inflation and Financial Development: Empirical Evidence from Ghana. Journal of Economics and Sustainable Development, 4(15), 20-37. Alhassan, A. L. (2015). Income diversification and bank efficiency in an emerging market. Managerial Finance, 41(12), 1318-1335. University of Ghana http://ugspace.ug.edu.gh 108 Altunbas, Y., Carbo, S., Gardener, E. P., & Molyneux, P. (2007). Examining the relationships between capital, risk and efficiency in European banking. European Financial Management, 13(1), 49-70. Amidu, M. (2007). Determinants of Capital Structure of Banks in Ghana: An Empirical Approach. Baltic Journal of Management, 2 (1), 67-79 Amihud, Y., Lev, B., 1981. Risk reduction as a managerial motive for conglomerate mergers. The Rand Journal of Economics, 12, 605–617. Antwi, S., Atta,E. & Zhao,X. (2012). Capital Structure and Firm Value : Empirical Evidence from Ghana. International Journal of Business and Social Science 3(22), 103–111 Arbabiyan, A., & Safari, M. (2009). The effects of capital structure and profitability in the listed firms in Tehran Stock Exchange. Journal of Management Perspective, 33, 159-175. Ariff, M., & Can, L. (2008). Cost and profit efficiency of Chinese banks: A non-parametric analysis. China Economic Review, 19(2), 260–273. Armen, H., Gayane, H., & Hassan, T. (2004). Determinants of target capital structure: The case of dual debt and equity issues. Journal of Financial Economics, 71(3), 517–540 Assaf, A. G., Barros, C. P., & Matousek, R. (2011). Technical efficiency in Saudi banks. Expert Systems With Applications, 38(5), 5781–5786. Asmild, M., Paradi, J. C., Reese, D. N., & Tam, F. (2007). Measuring overall efficiency and effectiveness using DEA. European Journal of Operational Research, 178(1), 305-321. University of Ghana http://ugspace.ug.edu.gh 109 Ataullah, A., & Le, H. (2006). Economic reforms and bank efficiency in developing countries: the case of the Indian banking industry. Applied Financial Economics, 16(9), 653-663. Avkiran , Necmi K. (1999). An application reference for data envelopment analysis in branch banking: helping the novice researcher. International Journal of Bank Marketing, 17, 206– 220. Baker, M.P. & Wurgler, J. (2002). Market timing and capital structure. Journal of Finance 57, 1– 32 Baltagi, B. (2008). Econometric analysis of panel data (Vol. 1). John Wiley & Sons. Banker, R. D., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management science, 30(9), 1078-1092. Banker, R. D., & Natarajan, R. (2008). Evaluating contextual variables affecting productivity using data envelopment analysis. Operations research, 56(1), 48-58 Bank of Ghana: Ghana Banking Act (2004) Bank of Ghana: Ghana Banking Act (2007) Bank of Ghana: Annual Report (2013) Bank of Ghana: Annual Report (2007) Bank of Ghana: Annual Report (2004) Barclay, M.J., Smith, C.W., & Watts, R.L. (1995). The determinants of corporate leverage and dividend policies. Journal of Applied Corporate Finance, 7, 4–19 University of Ghana http://ugspace.ug.edu.gh 110 Barros, C. P., & Assaf, A. (2009). Bootstrapped efficiency measures of oil blocks in Angola. Energy Policy, 37(10), 4098–4103. Baye, R.M. (2010). Managerial Economics and Business Strategy (7th ed.). McGraw Hill Benston, G.J. (1965). Branch banking and economies of scale. The Journal of Finance, 20(2), 312- 332. Berger A.N., & Mester, L.J. (1997). Inside the Black Box: What Explains Differences in the Efficiencies of Financial Institutions? Journal of Banking and Finance, 21, 895-947. Berger, A. N., & Bonaccorsi di Patti, E. (2006). Capital structure and firm performance: A new approach to testing agency theory and an application to the banking industry. Journal of Banking & Finance, 30(4), 1065–1102. Berger, A. N., & DeYoung, R. (1997). Problem loans and cost efficiency in commercial banks. Journal of Banking & Finance, 21(6), 849–870. Berger, A. N., Hasan, I & Zhou, M. (2010). "The effects of focus versus diversification on bank performance: Evidence from Chinese banks. Journal of Banking & Finance, 34(7), 1417- 1435 Berger, A. N., & DeYoung, R. (2001). The effects of geographic expansion on bank efficiency. Journal of Financial Services Research, 19(2-3), 163-184. Berger, A.N. & Bouwman, C.H.S. (2009). Bank liquidity creation. Review of Financial Studies, 22, 3779–3837 Berger, A.N. (2007). International Comparisons of Banking Efficiency. Financial Markets, Institutions & Instruments, 16(3), 119-144. University of Ghana http://ugspace.ug.edu.gh 111 Berger, A.N., & Humphrey, D. (1991). The dominance of inefficiencies over scale and product mix economies in banking. Journal of Monetary Economics, 28(1), 117–148. Berger, A.N., & Humphrey, D.B. (1997). Efficiency of financial institutions: International survey and directions for future research. European Journal of Operational Research, 98(2), 175- 212. Berger, A. N., DeYoung, R., Genay, H., & Udell, G. F. (2000). Globalization of financial institutions: Evidence from cross-border banking performance. Brookings-Wharton papers on financial services, 2000(1), 23-120. Berger, P. G., & Ofek, E. (1995). Diversification's effect on firm value. Journal of financial economics, 37(1), 39-65 Bogetoft, A. P., Otto, L., & Otto, M. L. (2011). Package “ Benchmarking .” Bokpin, G. A. (2013). Ownership structure, corporate governance and bank efficiency: an empirical analysis of panel data from the banking industry in Ghana. Corporate Governance, 13(3), 274-287. Bolton, P., & Scharfstein, D. (1990). A theory of predation based on agency problems in financial contracting. American Economic Review, 80(1), 93–106. Bonin, J. P., Hasan, I., & Wachtel, P. (2005). Bank performance, efficiency and ownership in transition countries. Journal of Banking & Finance, 29(1), 31–53 Boone, J. (2000). Competition, CEPR Discussion Paper Series No. 2636. Boone, J. (2008). A New Way to Measure Competition*. The Economic Journal, 118(531), 1245- 1261. University of Ghana http://ugspace.ug.edu.gh 112 Booth, L., Aivazian, V., Demirguc-Kunt, A., & Maksimovic, V. (2001). Capital Structure in Developing Countries. Journal of Finance, 56, 87–130. Bos, J. W., & Schmiedel, H. (2003). Comparing efficiency in European banking: a Meta frontier approach. De Nederlandsche Bank Research Paper, (57). Boot, a, & Schmeits, A. (2006). The competitive challenge in banking. In Advances in Corporate Finance and Asset Pricing (pp. 133 – 160). Bourke, P. (1989). Concentration and other determinants of bank profitability in Europe, North America and Australia. Journal of Banking & Finance, 13(1), 65-79. Brander, J., & Lewis, T. (1986). Oligopoly and financial structure: The limited liability effect. The American Economic Review, 76(5), 956–970. Brealey, R., Myers, S., & Marcus, A. (2001). Fundamentals of Corporate Finance. McGraw-Hill Irwin. Boston, MA. Brooks, C. (2008). Introductory Econometrics for Finance (2nd ed.). New York: Cambridge University Press Campello, M. (2003). Capital structure and product markets interactions: evidence from business cycles. Journal of Financial Economics, 68(3), 353–378. Campello, M. (2006). Capital structure and product markets interactions: Evidence from business cycles. Journal of Financial Economics, 68, 353–378 Casu, B., & Molyneux, P. (2003). A comparative study of efficiency in European banking. Applied Economics, 35(17), 1865-1876. University of Ghana http://ugspace.ug.edu.gh 113 Casu, B., & Girardone, C. (2004). Financial conglomeration: efficiency, productivity and strategic drive. Applied Financial Economics, 14(10), 687-696. Casu, B., & Girardone, C. (2009). Competition issues in European banking. Journal of Financial Regulation and Compliance, 17(2), 119-133. Cebenoyan, A. S., & Strahan, P. E. (2004). Risk management, capital structure and lending at banks. Journal of Banking & Finance, 28(1), 19–43. Chakraborty, I. (2010). Capital structure in an emerging stock market: The case of India. Research in International Business and Finance, 24(3), 295–314. Champion, D. (1999). Finance: The joy of leverage. Harvard Business Review, 77(4), 19-22. Chandler, A. (1977). The visible hand. Cambridge, MA: Belknap Press. Charnes, A., Cooper, W.W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2, 429-444. Cheng, Y., Liu, Y., & Chien, C. (2010). Capital structure and firm value in China : A panel threshold regression analysis. African Journal of Business Management, 4(12), 2500–2507. Chevalier, J. A. (1995). Capital structure and product-market competition: Empirical evidence from the supermarket industry. The American Economic Review, 415-435. Chevalier, J., & Scharfstein, D. (1996). Capital-market imperfections and countercyclical markups: Theory and evidence. American Economic Review, 85, 703–725. Chong, T. T. L., Lu, L., & Ongena, S. (2013). Does banking competition alleviate or worsen credit constraints faced by small-and medium-sized enterprises? Evidence from China. Journal of Banking & Finance, 37(9), 3412-3424 University of Ghana http://ugspace.ug.edu.gh 114 Chronopoulos, D. K., Girardone, C., & Nankervis, J. C. (2011). Are there any cost and profit efficiency gains in financial conglomeration? Evidence from the accession countries. The European Journal of Finance, 17(8), 603–621. Chung, Y. P., Na, H. S., & Smith, R. (2013). How important is capital structure policy to firm survival? Journal of Corporate Finance, 22, 83–103. Clark, J. A., & Siems, T. (2002). X-Efficiency in Banking: Looking beyond the Balance Sheet. Journal of Money, Credit, and Banking, 34(4), 987–1013 Coelli, T. J., Rao, D. S. P., O'Donnell, C. J., & Battese, G. E. (2005). An introduction to efficiency and productivity analysis. Springer Science & Business Media. Cohen, J. W. (1988). Statistical power analysis for the behavioural sciences (2nd ed). Hillsdale, New Jersey: Lawrence Erlbaum Asssociates. Comment, R., & Jarrell, G. A. (1995). Corporate focus and stock returns. Journal of Financial Economics, 37(1), 67-87. Cook, W. D., & Seiford, L. M. (2009). Data envelopment analysis (DEA)–Thirty years on. European Journal of Operational Research, 192(1), 1-17. Cooper, W. W., Seiford, L. M., & Tone, K. (2006). Introduction to Data Envelopment Analysis and Its Uses. Data envelopment analysis: a comprehensive text with models, applications, references and DEA-solver software Cooper, W. W., Seiford, L. M., & Zhu, J. (2011). Handbook on data envelopment analysis (Vol. 164). Springer Science & Business Media. University of Ghana http://ugspace.ug.edu.gh 115 Creswell, J. W. (2012). Educational Research:Planning,Conducting and Evaluating Quantitative and Qualitative Research. Cummins, J. D., Weiss, M. a., Xie, X., & Zi, H. (2010). Economies of scope in financial services: A DEA efficiency analysis of the US insurance industry. Journal of Banking & Finance, 34(7), 1525–1539. Das, A., & Ghosh, S. (2009). Financial deregulation and profit efficiency: A nonparametric analysis of Indian banks. Journal of Economics and Business, 61(6), 509–528. Dasgupta, S., & Titman, S. (1998). Pricing strategy and financial policy. The Review of Financial Studies, 11, 705–737. Daraio, C., & Simar, L. (2007). Advanced robust and nonparametric methods in efficiency analysis: Methodology and applications (Vol. 4). Springer Science & Business Media De Andres, P., & Vallelado, E. (2008). Corporate governance in banking: The role of the board of directors. Journal of banking & finance, 32(12), 2570-2580. De Borger, B., Kerstens, K., & Staat, M. (2008). Transit costs and cost efficiency: Bootstrapping non-parametric frontiers. Research in Transportation Economics, 23(1), 53–64. Delis, M. D., Koutsomanoli-Fillipaki, A., Staikouras, C. K., & Katerina, G. (2009). Evaluating cost and profit efficiency: a comparison of parametric and nonparametric methodologies. Applied Financial Economics, 19(3), 191-202 Demsetz, H. & Villalonga, B. (2001). Ownership structure and corporate performance. Journal of Corporate Finance 7, 209–233 University of Ghana http://ugspace.ug.edu.gh 116 Demsetz, H. (1973). Industry structure, market rivalry and public policy. Journal of Law and Economics, 16, 1–9. Demsetz, R. S., Saidenberg, M. R., & Strahan, P. E. (1996). The disciplinary role of franchise value: Banks with Something to Lose. Federal Reserve Bank of New York Economic Policy Review, 2, 1-14. Diamond, D.W. & Dybvig, P.H. (1983). Bank runs, deposit insurance, and liquidity. Journal of Political Economy 91 (3), 401–419. Drake, L., Hall, M. J. B., & Simper, R. (2006). The impact of macroeconomic and regulatory factors on bank efficiency: A non-parametric analysis of Hong Kong’s banking system. Journal of Banking & Finance, 30(5), 1443–1466 Ebaid, I. E.-S. (2009). The impact of capital-structure choice on firm performance: empirical evidence from Egypt. The Journal of Risk Finance, 10(5), 477–487. Elyasiani, E., & Mehdian, S. M. (1990). A nonparametric approach to measurement of efficiency and technological change: The case of large US commercial banks. Journal of Financial Services Research, 4(2), 157-168. Fama, E. F., & Miller, M. H. (1972). The theory of finance (Vol. 3). Hinsdale, IL: Dryden Press. Fama, E.F. & French, K.R. (2002). Testing trade-off and pecking order predictions about dividends and debt. Review of Financial Studies, 15 (1), 1–33. Färe, R., Grosskopf, S., & Weber, W. L. (2004). The effect of risk-based capital requirements on profit efficiency in banking. Applied Economics, 36(15), 1731–1743. University of Ghana http://ugspace.ug.edu.gh 117 Farrell, M. J. (1957). The measurement of productive efficiency. Journal of the Royal Statistical Society. Series A (General), 253-290 Fethi, M. D., & Pasiouras, F. (2010). Assessing bank efficiency and performance with operational research and artificial intelligence techniques: A survey. European Journal of Operational Research, 204(2), 189–198. Fluck, Z., & Lynch, A. W. (1999). Why Do Firms Merge and Then Divest? A Theory of Financial Synergy. The Journal of Business, 72(3), 319-346. Fosu, S. (2013). Capital structure, product market competition and firm performance: Evidence from South Africa. The Quarterly Review of Economics and Finance, 53(2), 140–151. Førsund, F. R., & Sarafoglou, N. (2002). On the origins of data envelopment analysis. Journal of Productivity Analysis, 17(1-2), 23-40 Frank, M. Z., & Goyal, V. K. (2003). Testing the pecking order theory of capital structure. Journal of financial economics, 67(2), 217-248. Frank, M. Z., & Goyal, V. K. (2009). Capital structure decisions: which factors are reliably important? Financial Management, 38(1), 1-37. Fried, H. O., Lovell, C. K., & Schmidt, S. S. (Eds.). (2008). The measurement of productive efficiency and productivity growth. Oxford University Press Fries, S., & Taci, A. (2005). Cost efficiency of banks in transition: Evidence from 289 banks in 15 post-communist countries. Journal of Banking & Finance, 29(1), 55-81. University of Ghana http://ugspace.ug.edu.gh 118 Fukuyama, H., & Matousek, R. (2011). Efficiency of Turkish banking: Two-stage network system. Variable returns to scale model. Journal of International Financial Markets, Institutions and Money, 21(1), 75-91. Gatsi, J. G. (2012). Capital Structure of Ghanaian Banks: An Evaluation of its Impact on Performance. The IUP Journal of Bank Management, 11(4), 86-99. Ghana Banking Survey (2005). Is local bank aligned to macro-economic objectives? Accra. PWC Ghana Banking Survey (2007). Is the industry on the ball? Accra. PWC Ghana Banking Survey (2008). Raising the bar: Increase in the minimum Capital requirements, and Implications for the industry. Accra: PWC Ghana Banking Survey (2011).Sustaining Growth: Challenges and Opportunities. Accra. PWC Ghana Banking Survey (2012).Enhancing customer value to sustain profitable growth. Accra. PWC Ghana Banking Survey (2013). Harnessing the SME potential. Accra. PWC Ghana Banking Survey (2014). The future of banking in Ghana. What’s next? Accra. PWC Ghana Statistical Service (2014). National Account Statistics: Gross Domestic Product. Accra: GSS Ghosh, C., Nag, R., & Sirmans, C. (2000). The pricing of seasoned equity offerings: Evidence from REITs. Real Estate Economics, 28(3), 363–834. University of Ghana http://ugspace.ug.edu.gh 119 Graham, J. R., Lemmon, M. L., & Wolf, J. G. (2002). Does corporate diversification destroy value? The Journal of Finance, 57(2), 695-720. Griffin, J.J., & Mahon, J.F. (1997). The Corporate Social Performance and Corporate Financial Performance Debate: Twenty-Five Years of Incomparable Research. Business Society, 36(1), 5-31. Griffith, R., Boone, J., & Harrison, R. (2005). Measuring competition. Advanced Institute of Management Research Paper, (022). Grigorian, D., & Manole, V. (2002). Determinants of commercial bank performance in transition: An application of data envelopment analysis. World Bank Policy Research Working Paper, (2850). Grossman, S., & Hart. O. (1982). Corporate Financial Structure and Managerial Incentives. The Economics of Information and Uncertainty. Chicago: University of Chicago Press Hadlock, C. J., Ryngaert, M., & Thomas, S. (2001). Corporate Structure and Equity Offerings: Are There Benefits to Diversification? The journal of business, 74(4), 613-635. Halkos, G. E., & Tzeremes, N. G. (2013). Estimating the degree of operating efficiency gains from a potential bank merger and acquisition: A DEA bootstrapped approach. Journal of Banking and Finance, 37(5), 1658–1668. Harwell, M. R. (2011). Research design in qualitative/quantitative/mixed methods. The Sage handbook for research in education. 2nd ed. Los Angeles, CA: Sage, 147. Harris, M., & Raviv, A. (1991). The theory of capital structure. The Journal of Finance, 46(1), 297–355. University of Ghana http://ugspace.ug.edu.gh 120 Harvey, C. R., Lins, K. V., & Roper, A. H. (2004). The effect of capital structure when expected agency costs are extreme. Journal of Financial Economics, 74(1), 3–30. Hatfield, B. G, Cheng, T. W. & Davidson, N. W. (1994). The Determination of Optimal Capital Structure. The Effect of Firm and Industry Debt Ratio on Market Value. Journal of Financial and Strategic Decision, 7 (3), 1 – 14 Haugen, R. & Senbet, L. (1978). The insignificance of bankruptcy costs to the theory of optimal capital structure. Journal of Finance, 33 (2), 383-93. Havrylchyk, O. (2006). Efficiency of the Polish banking industry: Foreign versus domestic banks. Journal of Banking & Finance, 30(7), 1975-1996 Henningsen, A. (2010). Estimating censored regression models in R using the censReg Package. R package vignettes collection, v0. 5, 2. Hirschberg, J. G., & Lloyd, P. J. (2002). Does the technology of foreign-invested enterprises spill over to other enterprises in China? An application of post-DEA bootstrap regression analysis. Modelling the Chinese Economy. Edward Elgar Press, London. Hicks, J. (1969). A theory of economic history. Ox- ford: Clarendon Press. Holmstrom, B. & Tirole, J. (1997). Financial intermediation, loanable funds, and the real sector. Quarterly Journal of Economics, 62, 663–691 Hoff, A. (2007). Second stage DEA: Comparison of approaches for modelling the DEA score. European Journal of Operational Research, 181(1), 425-435 Hubbard, R. G., & Palia, D. (1999). A reexamination of the conglomerate merger wave in the 1960s: An internal capital markets view. The Journal of Finance, 54(3), 1131-1152. University of Ghana http://ugspace.ug.edu.gh 121 Hughes, J. P., & Mester, L. J. (1993). A quality and risk-adjusted cost function for banks: Evidence on the “too-big-to-fail” doctrine. Journal of productivity analysis, 4(3), 293-315. Isik, I., & Hassan, M. K. (2002). Cost and profit efficiency of the Turkish banking industry: An empirical investigation. Financial Review, 37(2), 257-279. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behaviour, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. Jensen, M.C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review, 76, 323–329. Kashyap, A.K., Rajan, R. & Stein, J.C. (1999). Banks as Liquidity Providers: An Explanation for the Co-existence of Lending and Deposit-Taking. NBER Working Paper Series, No. 6962. Kasman, A., & Yildirim, C. (2006). Cost and profit efficiencies in transition banking: the case of new EU members. Applied Economics, 38(9), 1079-1090. Kenjegalieva, K., Simper, R., Weyman-Jones, T., & Zelenyuk, V. (2009). Comparative analysis of banking production frameworks in Eastern European financial markets. European Journal of Operational Research, 198(1), 326-340. Kim, E.H. (1978). A Mean-Variance Theory of Optimal Capital Structure and Corporate Debt Capacity. Journal of Finance, 45-63 Kovenock, D. & Phillips, G. (1997). Capital structure and product market behavior: an examination of plant exit and investment decisions. Review of Financial Studies, 10, 767– 803. KPMG (2012). Doing Business in Ghana. Accra University of Ghana http://ugspace.ug.edu.gh 122 Kraus, A. & Litzenberger, R.H. (1973). A state preference model of optimal financial leverage. Journal of Finance, 28, 911–922 Kumbhakar, S. C., & Lovell, C. K. (2003). Stochastic frontier analysis. Cambridge University Press. Kyereboah-Coleman, A. (2007). The impact of capital structure on the performance of microfinance institutions. The Journal of Risk Finance, 8(1), 56–71 Laeven, L., & Levine, R. (2007). Is there a diversification discount in financial conglomerates?. Journal of Financial Economics, 85(2), 331-367 Lamont, O. A., & Polk, C. (2002). Does diversification destroy value? Evidence from the industry shocks. Journal of Financial Economics, 63(1), 51-77 Lang, L. & Stulz, R. (1994). Tobin's q, corporate diversification and firm performance. Journal of Political Economy, 102, 1248-1280 Lemmon, M. L., & Zender, J. F. (2010). Debt Capacity and Tests of Capital Structure Theories. Journal of Financial and Quantitative Analysis, 45(05), 1161–1187. 99 Levine, R. (1997). Financial development and economic growth: Views and agenda. Journal of Economic Literature, 35, 688–726. Levine, R., Loayza, N., & Beck, T. (2000). Financial intermediation and growth: Causality and causes. Journal of Monetary Economics, 46(1), 31-77. Lewellen, W. G. (1971). A Pure Financial Rationale for the Conglomerate Merger. The Journal of Finance, 26(2), 521–537. University of Ghana http://ugspace.ug.edu.gh 123 Lins, K., & Servaes, H. (1999). International evidence on the value of corporate diversification. The Journal of Finance, 54(6), 2215-2239. Lovell, C. K., & Pastor, J. T. (1995). Units invariant and translation invariant DEA models. Operations Research Letters, 18(3), 147-151. Low, P., & Chen, K. (2004). Diversification and capital structure: some international evidence. Review of Quantitative Finance and Accounting, 23, 55–71. Luo, X. (2003). Evaluating the profitability and marketability efficiency of large banks. Journal of Business Research, 56(8), 627–635. Maniadakis, N., & Thanassoulis, E. (2004). A cost Malmquist productivity index. European Journal of Operational Research, 154(2), 396-409. Margaritis, D., & Psillaki, M. (2007). Capital structure and firm efficiency. Journal of Business Finance & Accounting, 34(9‐10), 1447-1469. Margaritis, D., & Psillaki, M. (2010). Capital structure, equity ownership and firm performance. Journal of Banking & Finance, 34(3), 621–632. Maudos, J., & Pastor, J. M. (2003). Cost and profit efficiency in the Spanish banking sector (1985– 1996): a non-parametric approach. Applied Financial Economics, 13(1), 1–12. Maudos, J., Pastor, J., Perez, F., & Quesada, J. (2002). Cost and profit efficiency in European banks. Journal of International Financial Markets Institutions and Money, 12, 33−58 McDonald, J. (2009). Using least squares and tobit in second stage DEA efficiency analyses. European Journal of Operational Research, 197(2), 792-798. University of Ghana http://ugspace.ug.edu.gh 124 Menéndez-Alonso, E. J. (2003). Does diversification strategy matter in explaining capital structure? Some evidence from Spain. Applied Financial Economics, 13(6), 427–430. Meyer, M., Milgrom, P., & Roberts, J. (1992). Organizational prospects, influence costs, and ownership changes. Journal of Economics & Management Strategy, 1(1), 9-35 Miller, S. M., & Noulas, A. G. (1996). The technical efficiency of large bank production. Journal of Banking and Finance, 20, 495–509. Modigliani, F., & Miller, M. (1958). The cost of capital, corporation finance and the theory of investment. American Economic Review, 48, 261–297. Moffat, B. D., & Valadkhani, A. (2011). Efficiency of Botswana's financial institutions: a data envelopment analysis. Applied Economics Letters, 18(7), 697-702. Molyneux, P., & Thornton, J. (1992). Determinants of European bank profitability: A note. Journal of banking & Finance, 16(6), 1173-1178. Myers, S. C. (1977). Determinants of corporate borrowing. Journal of financial economics, 5(2), 147-175. Myers, S. C. (1984). The capital structure puzzle. The journal of finance, 39(3), 574-592. Myers, S.C. & Majluf, N.S. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 12, 187-221. Myers, S., & Majluf, N. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13, 187–222. University of Ghana http://ugspace.ug.edu.gh 125 Nahm, D., & Vu, H. (2013). Profit efficiency and productivity of Vietnamese banks: a new index approach. Journal of Applied Finance and banking, 3, 45-65. Nicoló, G. D., Bartholomew, P., Zaman, J., & Zephirin, M. (2004). Bank consolidation, internationalization, and conglomeration: Trends and implications for financial risk. Financial markets, institutions & instruments, 13(4), 173-217. Ohene-Asare, K. (2011). Nonparametric efficiency and productivity change measurement of banks with corporate social responsibilities: the case for Ghana (Doctoral dissertation, University of Warwick). Ohene-Asare, K., & Asmild, M. (2012). Banking efficiency analysis under corporate social responsibilities. International Journal of Banking, Accounting and Finance, 4(2), 146-171. Ohene-Asare, K., & Latif. (2014). Competition And Bank Efficiency in Emerging Markets: Empirical Evidence from Ghana. African Journal of Economic and Management Studies. Opler, T., & Titman, S. (1994). Financial distress and corporate performance. Journal of Corporate Finance, 49, 1015–1040 Paradi, J. C., & Zhu, H. (2013). A survey on bank branch efficiency and performance research with data envelopment analysis. Omega, 41(1), 61–79. Paradi, J.C., Yang, Z., & Zhu, H. (2011). Handbook on Data Envelopment Analysis. Boston, MA: Springer. Pasiouras, F., & Kosmidou, K. (2007). Factors influencing the profitability of domestic and foreign commercial banks in the European Union. Research in International Business and Finance, 21(2), 222-237. University of Ghana http://ugspace.ug.edu.gh 126 Pasiouras, F. (2008). International evidence on the impact of regulations and supervision on banks’ technical efficiency: An application of two-stage data envelopment analysis. Review of Quantitative Finance and Accounting, 30, 187–223 Pasiouras, F., Tanna, S., & Zopounidis, C. (2009). The impact of banking regulations on banks ’ cost and profit efficiency: Cross-country evidence, International Review of Financial Analysis,18, 294–302. Pearce, J.A. & Robinson, R.B. (2009). Strategic Management, Formulation, Implementation and Control (11th ed.). U.S.A: McGraw-Hill Pennacchi, G. (2006). Deposit insurance, bank regulation, and financial system risks. Journal of Monetary Economics, 53(1), 1–30. Petersen, M. A., & Rajan, R. G. (1995). The effect of credit market competition on lending relationships. The Quarterly Journal of Economics, 407-443. Portela, M. C. A. S., & Thanassoulis, E. (2005). Profitability of a sample of Portuguese bank branches and its decomposition into technical and allocative components. European journal of Operational Research, 162(3), 850-866. Quartey, P., & Afful-Mensah, G. (2014). Financial and monetary policies in Ghana: A review of recent trends. Review of Development Finance, 4(2), 115-125. Rajan, R., & Zingales, L. (1995). What do we know about capital structure? Some evidence from international data. The Journal of Finance, 50(5), 1421–1460. University of Ghana http://ugspace.ug.edu.gh 127 Rajan, R., Servaes, H., & Zingales, L. (2000). The cost of diversity: The diversification discount and inefficient investment. The Journal of Finance, 55(1), 35–80. Ramalho, E. A., Ramalho, J. J., & Henriques, P. D. (2010). Fractional regression models for second stage DEA efficiency analyses. Journal of Productivity Analysis, 34(3), 239-255 Ramanathan, R. (1995). Introductory Econometrics with Applications (3rd ed.). Dryden Press, Fort Worth TX Ray, S. C., & Das, A. (2010). Distribution of cost and profit efficiency: Evidence from Indian banking. European Journal of Operational Research, 201(1), 297–307 Roden, D.M., & Lewellen, W.G. (1995). Corporate capital structure decisions: Evidence from leveraged buyouts. Financial Management, 24, 76-87. Rogers, K. E. (1998). Nontraditional activities and the efficiency of US commercial banks. Journal of Banking & Finance, 22(4), 467-482. Ross, S. (1977). The determination of financial structure: the incentive-signalling approach. The Bell Journal of Economics, 8(1), 23–40. Rowley, T., & Berman, S. (2000). A Brand New Brand of Corporate Social Performance. Business & Society, 39(4), 397- 418. Saidenberg, M.R., & Strahan, P.E. (1999). Are Banks Important for Financing Large Businesses? Current Issues in Economics and Finance 5 (12). Saka, A. N. A., Aboagye, A. Q. Q., & Gemegah, A. (2012). Technical Efficiency of the Ghanaian Banking Industry and the Effects of the Entry of Foreign Banks. Journal of African Business, 13(3), 232–243. University of Ghana http://ugspace.ug.edu.gh 128 Sathye, M. (2003). Efficiency of banks in a developing economy : The case of India, 148, 662– 671. Saxonhouse, G. R. (1976). Estimated parameters as dependent variables. The American Economic Review, 178-183. Schaeck, K., & Cihák, M. (2014). Competition, efficiency, and stability in banking. Financial Management, 43(1), 215-241. Schmid, M. M., & Walter, I. (2009). Do financial conglomerates create or destroy economic value? Journal of Financial Intermediation, 18(2), 193–216. Schumpeter, J. (1912). A Theory of Economic Development. Harvard University Press, Cambridge, MA Scott, J. (1976). A Theory of Optimal Capital Structure. The Bell Journal of Economics and Management Science, 33-54 Sealey, C.W., & Lindley, J.T. (1977). Inputs, Outputs, and a Theory of Production and Cost at Depository Financial Institutions. Journal of Finance, 32(4), 1251-1266. Sherman, H.D., & Gold, F. (1985). Bank Branch Operating Efficiency - Evaluation with Data Envelopment Analysis. Journal of Banking & Finance, 9(2), 297-315. Shyam-Sunder, L., & C Myers, S. (1999). Testing static tradeoff against pecking order models of capital structure. Journal of financial economics, 51(2), 219-244. Simar, L., & Wilson, P. W. (2011). Two-stage DEA: caveat emptor. Journal of Productivity Analysis, 36(2), 205-218. University of Ghana http://ugspace.ug.edu.gh 129 Simar, L., & Wilson, P.W. (1998). Sensitivity Analysis of Efficiency Scores: How to Bootstrap in Nonparametric Frontier Models. Management Science, 44(1), 49-61. Simar, L., & Wilson, P.W. (2000). A general methodology for bootstrapping in non-parametric frontier models. Journal of Applied Statistics, 27(6), 779–802. Simar, L., & Wilson, P.W. (2007). Estimation and inference in two-stage, semi-parametric models of production processes. Journal of Econometrics, 136(1), 31-64. Smith, A. (1990). Corporate ownership structure and performance: the case of management buyouts. Journal of Financial Economics, 27, 143–164. Sowa, N. K. (2003). Financial sector reform policies and poverty reduction. Centre for Policy Analysis. Staub, R. B., da Silva e Souza, G., & Tabak, B. M. (2010). Evolution of bank efficiency in Brazil: A DEA approach. European Journal of Operational Research, 202(1), 204–213 Stulz, R. (1990). Managerial discretion and optimal financing policies. Journal of Financial Economics, 20, 3–27. Tecles, P. L., & Tabak, B. M. (2010). Determinants of bank efficiency: The case of Brazil. European Journal of Operational Research, 207(3), 1587–1598 Telser, L. G. (1966). Cutthroat competition and the long purse. Journal of Law and Economics, 259-277. Thanassoulis E., Boussofiane A. & Dyson R. A. (1996). Comparison of Data Envelopment Analysis and Ratio Analysis as tools for performance assessment. Omega, 24, 229–44. University of Ghana http://ugspace.ug.edu.gh 130 Titman, S. (1984). The effect of capital structure on a firm's liquidation decision. Journal of Financial Economics, 13, 137-151. Titman, S., & Wessels, R. (1988). The determinants of capital structure choice. The Journal of Finance, 43(1), 1–19. Tone, K. (2002). A strange case of the cost and allocative efficiencies in DEA. Journal of the Operational Research Society, 53(11), 1225–1231. Tortosa-Ausina, E. (2002). Bank cost efficiency and output specification. Journal of Productivity Analysis, 18(3), 199-222. Tortosa-Ausina, E., Armero, C., Conesa, D., & Grifell-Tatjé, E. (2012). Bootstrapping profit change: An application to Spanish banks. Computers & Operations Research, 39(8), 1857- 1871. Tzeremes, N. G. (2015). Efficiency dynamics in Indian banking: A conditional directional distance approach. European Journal of Operational Research, 240(3), 807–818. VanHoose, D. (2007). Theories of bank behavior under capital regulation. Journal of Banking and Finance, 31, 3680–3697. Van Lelyveld, I., & Knot, K. (2009). Do financial conglomerates create or destroy value? Evidence for the EU. Journal of Banking & Finance, 33(12), 2312–2321. University of Ghana http://ugspace.ug.edu.gh 131 Van Leuvensteijn, M., Bikker, J. A., van Rixtel, A. A., & Sørensen, C. K. (2011). A new approach to measuring competition in the loan markets of the euro area. Applied Economics, 43(23), 3155-3167. Vander-Vennet, R. (2002). Cost and profit efficiency of financial conglomerates and universal banks in Europe. Journal of Money Credit and Banking, 34(1), 254–282 Vivas, A. L. (1997). Profit efficiency for Spanish savings banks. European Journal of Operational Research, 98(2), 381-394. Wanzenried, G. (2003). Capital structure decisions and output market competition under demand uncertainty. International Journal of Industrial Organization, 21(2), 171-200. Weston, J. (1970). The nature and significance of conglomerate firms. St. John’s Law Review, 44, 66–80. Whited, T. M. (2001). Is it inefficient investment that causes the diversification discount? The Journal of Finance, 56(5), 1667-1691 Williams, J. (1987). Perquisites, risk, and capital structure. Journal of Finance, 42, 29–49. Wooldridge, J. (2012). Introductory econometrics: A modern approach. Cengage Learning. Wruck, K. H. (1995). Financial policy as a catalyst for organizational change: Sealed Air's leveraged special dividend. Journal of Applied Corporate Finance, 7(4), 20-35. Xue, M., & Harker, P. T. (1999). Overcoming the inherent dependency of DEA efficiency scores: a bootstrap approach. Unpublished Working Paper, Wharton Financial Institutions Center, University of Pennsylvania University of Ghana http://ugspace.ug.edu.gh 132 APPENDICES University of Ghana http://ugspace.ug.edu.gh 133 APPENDIX A Table I Definitions and summary statistics of inputs, outputs and prices (Amount in GH¢) Variables Mean Std. Dev Min Max 2000- N=16 Inputs Deposits 47525451.07 56487564.68 532192.1 169328300 Labour 2013705.86 2559970.46 49094 9881200 Fixed Assets 2228776.94 2634590.52 142846.3 9732800 Deposits 0.12 0.04 0.06 0.2 Input Prices Labour Cost 0.03 0.01 0.01 0.06 Fixed Assets 1.45 0.68 0.68 2.89 Outputs Loans and Advances 29086950.65 38677493.3 154708 128695800 Investment 22064396.41 28981648.68 149307.8 100405200 Output Prices Loans and Advances 0.25 0.11 0.11 0.52 Investment 0.25 0.09 0.08 0.4 Total Assets 76822341.07 92990986.31 884009 314956500 Equity 8202300.99 8365797.32 278719 23232156.2 2001-N-17 Inputs Deposits 52962196.98 63943805.27 2172653.1 192441900 Labour 2711770.34 3511081.52 71530 13744100 Fixed Assets 2986023.28 4011719.67 124941.4 15981700 Input Prices Deposits 0.13 0.05 0.06 0.26 Labour Cost 0.04 0.01 0.02 0.07 Fixed Assets 1.8 1.32 0.42 5.82 Outputs Loans and Advances 31438771.87 47063933.86 646601.5 182619800 Investment 23868658.24 30630042.82 729233.7 102411000 Outputs Prices Loans and Advances 0.28 0.08 0.14 0.45 Investment 0.28 0.05 0.16 0.34 Total Assets 84915639.31 106957102 2712278.6 379428800 Equity 10074851.61 11343803.6 274300 32919500 2002-N-17 Deposits 67998601.79 77989522.37 3584258.9 240860900 Inputs Labour 3597164.92 4687376.34 122659.4 17773200 Fixed Assets 3670531.5 4828090.32 181583.5 18417800 Deposits 0.08 0.03 0.04 0.13 Input Prices Labour Cost 0.03 0.01 0.01 0.05 Fixed Assets 1.58 0.76 0.6 3.05 University of Ghana http://ugspace.ug.edu.gh 134 Outputs Loans and Advances 33269547.19 35930226.14 1721614.6 102606200 Investment 34823396.95 60658492.85 1010145.2 247857600 Output Prices Loans and Advances 0.24 0.09 0.12 0.48 Investment 0.19 0.05 0.06 0.25 Total Assets 190212505.6 399255004.6 4921309.2 1667882000 Equity 12571913.56 14522440.83 -1606000 42072100 Table II: Definitions and summary statistics of inputs, outputs and prices (Amount in GH¢) Variables Mean Std. Dev. Min Max 2003- N=18 Inputs Deposits 90286711.37 103379755.5 7355846.2 318383000 Labour 4825164.49 7100041.48 287497.5 29715800 Fixed Assets 3708830.12 4301705.9 232341.3 16696200 Deposits 0.11 0.09 0.03 0.4 Input Prices Labour Cost 0.03 0.01 0.01 0.06 Fixed Assets 1.85 1.19 0.62 5.49 Outputs Loans and Advances 49400346.86 56206249.21 3005457.4 175429700 Investment 43462048.81 51488373.82 2374278.6 194454600 Output Prices Loans and Advances 0.21 0.07 0.11 0.39 Investment 0.2 0.05 0.11 0.32 Total Assets 229315211.5 456337936.1 9624320.3 1971062000 Equity 15715700.9 17051521.53 -1555775.4 47289000 2004-N-18 Inputs Deposits 118411804.1 131852493.5 11220601.8 426573300 Labour 6079249.24 7844060.83 485331.7 32348200 Fixed Assets 4915225.84 4274901.06 485032 14160300 Input Prices Deposits 0.09 0.06 0.02 0.28 Labour Cost 0.03 0.01 0.02 0.06 Fixed Assets 1.66 0.73 0.47 2.91 Outputs Loans and Advances 62331211.86 66475868.74 4361800 209506100 Investment 64366725.94 67473125.24 5716200 231319000 Outputs Prices Loans and Advances 0.2 0.05 0.12 0.3 Investment 0.15 0.06 0.08 0.27 Total Assets 285095149.7 551110556.5 16094300 2390684000 Equity 20363230.9 19723247.95 1037000 57432400 2005-N-19 Deposits 127977020.7 131558904.7 13917800 472994000 Inputs Labour 7779016.28 10185863.64 772009 43519700 Fixed Assets 5740045.27 4522069.71 466300 16155300 Deposits 0.07 0.04 0.03 0.19 Input Prices Labour Cost 0.04 0.01 0.02 0.07 Fixed Assets 1.31 0.53 0.62 2.52 Outputs Loans and Advances 83595014.71 85063853.95 6371900 274552700 University of Ghana http://ugspace.ug.edu.gh 135 Investment 51350746.41 50898858.24 3655700 184333000 Output Prices Loans and Advances 0.18 0.05 0.08 0.3 Investment 0.17 0.03 0.11 0.23 Total Assets 179099375.2 166204596.6 17805900 586471300 Equity 23295129.22 23022791.42 1401200 70094200 Table III: Definitions and summary statistics of inputs, outputs and prices (Armount in GH¢) Variables Mean Std. Dev Min Max 2006- N=23 Inputs Deposits 154009498.6 172081337.4 11860300 634572700 Labour 8604648.37 10962427.66 837639 47425300 Fixed Assets 7556785.03 6683872.26 982510 24066000 Deposits 0.06 0.04 0.02 0.18 Input Prices Labour Cost 0.03 0.01 0.02 0.07 Fixed Assets 1.13 0.59 0.15 2.61 Outputs Loans and Advances 109417439.2 125175115.5 3206220 430177000 Investment 46392645.9 53429119.76 1791100 220616248 Output Prices Loans and Advances 0.15 0.04 0.06 0.22 Investment 0.16 0.07 0 0.35 Total Assets 207051753.4 212586977.1 1351847.84 775992315 Equity 26755148.76 26955216.69 6380163 89034865 2007-N-23 Inputs Deposits 228585933 217503326.6 20089129 839382600 Labour 11617113.83 13715165.64 1466174 57884160 Fixed Assets 10884911.04 10580454.85 1024386 42913000 Input Prices Deposits 0.06 0.03 0.03 0.13 Labour Cost 0.1 0.3 0.02 1.47 Fixed Assets 1.22 0.65 0.38 3.18 Outputs Loans and Advances 170080358.1 189978094.4 12689450 750663500 Investment 61735108.57 72127676.32 5973484 255842000 Outputs Prices Loans and Advances 0.15 0.04 0.07 0.22 Investment 0.13 0.06 0 0.21 Total Assets 321363495.1 314567924.2 29066347 1191015000 Equity 36334821.96 41843549.75 5660191 173691300 2008-N-24 Deposits 302044219.8 273061382.6 5061311 1030106000 Inputs Labour 16144511.98 18049327 650752 67714010 Fixed Assets 14183716.96 13181039.41 731065 57412000 Deposits 0.07 0.04 0.02 0.21 Input Prices Labour Cost 0.03 0.01 0.01 0.06 Fixed Assets 1.35 0.74 0.5 3.19 University of Ghana http://ugspace.ug.edu.gh 136 Outputs Loans and Advances 237676639.4 245606506.6 2247845 1087119000 Investment 68052465.75 66445808.06 4362458 258367000 Output Prices Loans and Advances 0.16 0.06 0.02 0.23 Investment 0.12 0.05 0 0.22 Total Assets 429227927.8 388547628.4 13971170 1645797000 Equity 45550179.92 47393686.81 7705363 203863700 Table IV: Definitions and summary statistics of inputs, outputs and prices (Amount in GH¢) Variables Mean Std. Dev Min Max 2009- N=27 Inputs Deposits 351596015.8 309866451.2 5214116 1259470000 Labour 18472206.71 19440894.64 196280.79 84988700 Fixed Assets 16313772.96 14038533.26 140573 51918000 Deposits 0.11 0.06 0.03 0.25 Input Prices Labour Cost 0.03 0.02 0.01 0.1 Fixed Assets 1.41 0.86 0.57 4.41 Outputs Loans and Advances 234653349.6 249774224.4 4315785 1265517000 Investment 111662410.2 131655724.4 61000 505781000 Output Prices Loans and Advances 0.23 0.07 0.12 0.49 Investment 0.48 1.63 0 8.45 Total Assets 498769276.2 449445580.7 15375874 1917083000 Equity 68796622.78 59616981.85 702313 205413000 2010-N-26 Inputs Deposits 476856782 377994629.3 33464677 1584055000 Labour 24540923.54 26176055.93 237590.13 107422800 Fixed Assets 17991556.27 13586531.67 96133 54684000 Input Prices Deposits 0.08 0.04 0.02 0.2 Labour Cost 0.03 0.01 0 0.07 Fixed Assets 2.44 2.46 0.48 11.53 Outputs Loans and Advances 272253586.4 211694101 8127071 995356000 Investment 182636135 182216614.3 10548232 741297000 Outputs Prices Loans and Advances 0.2 0.08 0.07 0.47 Investment 0.14 0.04 0.01 0.2 Total Assets 635663123 478275851.4 64244649 2076361000 Equity 88535757.73 62047303.93 3697802 242265000 2011-N-25 Deposits 614144882.3 518882741 20004683 2061390000 Inputs Labour 28036940.71 30715489.1 294540.21 135912000 Fixed Assets 20357687.88 16852588.41 59111 63339490 Deposits 0.05 0.03 0.02 0.1 Input Prices Labour Cost 0.03 0.01 0 0.06 University of Ghana http://ugspace.ug.edu.gh 137 Fixed Assets 2.35 3.35 0.57 17.72 Outputs Loans and Advances 315964551.1 225911827.2 6020889 848459000 Investment 220231346.2 277584613.6 5158520 1215140000 Output Prices Loans and Advances 0.16 0.05 0.06 0.27 Investment 0.22 0.49 0.04 2.56 Total Assets 783871666.2 613010614 91405281 2454564000 Equity 114398916.2 77515784.55 40940000 316860000 Table V: Definitions and summary statistics of inputs, outputs and prices (Amount in GH¢) Variables Mean Std. Dev Min Max 2012- N=25 Inputs Deposits 790917500 634341682.2 36370986 2407615000 Labour 37894000 36798060.86 473696.34 144435000 Fixed Assets 24241680 19732991.59 176030 73404000 Deposits 0.06 0.03 0.01 0.13 Input Prices Labour Cost 0.03 0.01 0 0.07 Fixed Assets 1.98 1.5 0.54 7.42 Outputs Loans and Advances 472565000 328271555.1 16616224 1394967000 Investment 272602800 340334319.7 5158520 1586813000 Output Prices Loans and Advances 0.16 0.05 0.04 0.27 Investment 0.29 0.63 0.07 3.28 Total Assets 1045397000 797626631.9 115147396 3378843000 Equity 158107500 103269543.1 61772515 456547000 2013-N-26 Inputs Deposits 902408000 766824500 60844336 3220777000 Labour 46158110 42856690 585017.71 169996000 Fixed Assets 31562130 25605750 129841 94756640 Input Prices Deposits 0.08 0.04 0.02 0.16 Labour Cost 0.03 0.01 0 0.07 Fixed Assets 2.02 2.52 0.32 13.65 Outputs Loans and Advances 597826400 470003600 13689042 2124530000 Investment 415100000 445654700 5158520 1747087000 Outputs Prices Loans and Advances 0.17 0.05 0.06 0.27 Investment 0.41 0.89 0.06 4.34 Total Assets 1398414000 1100222000 149042412 4624405000 Equity 212729100 141366000 70351502 557106000 University of Ghana http://ugspace.ug.edu.gh 138 APPENDIX B Table VI: Profit Efficiency Scores (in %) of banks from 2000- 2008 Bank Profit Efficiency (%) 2000 2001 2002 2003 2004 2005 2006 2007 2008 GCB 100 100 100 100 100 100 100 100 100 BBG 100 100 100 100 100 100 100 100 64.80 SCB 100 100 100 100 50.01 100 100 100 100 NIB 72.21 29.34 60.90 84.64 91.24 63.92 64.91 74.46 72.29 UMB 55.31 100 63.53 100 100 100 100 100 100 ADB 68.58 71.41 62.44 100 69.52 100 100 94.02 84.80 AMAL/BOA 100 100 100 100 100 100 64.77 83.26 100 SG-SSB 100 100 71.93 74.26 73.63 100 92.65 100 100 UNIBANK 100 100 100 100 100 100 73.26 71.28 STANBIC 40.13 40.95 47.26 100 70.15 95.94 74.86 100 100 EBG 66.74 100 63.83 83.01 67.29 70.94 80.24 73.81 75.91 CAL 68.34 48.86 68.04 83.82 56.31 90.68 91.59 71.87 77.17 TTB 74.22 100 65.70 88.65 100 95.97 100 100 100 FABL 100 100 100 100 52.61 100 100 100 100 METRO/BPI 67.01 61.51 42.97 60.64 100 100 100 100 100 PBL 58.75 79.95 100 62.97 71.76 100 100 100 78.43 ICB 100 100 100 100 100 100 100 100 100 HFC 100 100 75.47 100 100 100 UBA 100 42.93 48.88 100 FBL 100 100 75.76 GTB 100 100 100 University of Ghana http://ugspace.ug.edu.gh 139 IBG 100 73.64 63.43 ZENITH 100 69.42 100 BSIC 100 BARODA ACCESS UT ENERGY ROYAL Table VII: Profit Efficiency Scores (in %) of banks from 2009- 2013 Bank Profit Efficiency (%) 2009 2010 2011 2012 2013 GCB 100 100 100 100 100 BBG 73.53 100 100 100 100 SCB 100 100 100 100 100 NIB 72.99 50.25 100 71.16 44.11 UMB 94.66 100 27.80 30.56 26.41 ADB 64.39 68.51 47.64 58.42 71.36 AMAL/BOA 92.13 70.27 78.94 59.26 39.04 SG-SSB 100 100 100 54.94 47.55 UNIBANK 58.73 73.49 78.83 100 100 STANBIC 79.88 51.85 62.82 62.08 74.67 EBG 76.20 100 55.75 100 100 CAL 78.22 100 100 100 100 TTB 100 75.06 FABL 100 63.94 45.90 38.59 42.70 METRO/BPI 100 PBL 100 80.24 68.66 55.19 48.02 ICB 100 56.33 41.45 49.46 43.32 HFC 74.64 87.11 69.58 58.79 70.21 UBA 53.82 46.89 59.58 100 100 University of Ghana http://ugspace.ug.edu.gh 140 FBL 75.19 100 100 66.03 55.94 GTB 100 86.60 58.75 100 53.35 IBG 57.89 53.73 ZENITH 78.11 78.88 60.47 60.74 100 BSIC 23.16 41.65 36.04 43.01 33.61 BARODA 100 100 100 100 100 ACCESS 100 100 100 69.56 64.96 UT 100 100 100 100 58.66 ENERGY 40.52 18.13 29.29 ROYAL 20.83 University of Ghana http://ugspace.ug.edu.gh 141 APPENDIX C Table 20: Regression Results: Profit Efficiency and Capital Structure ERC ROA ROE (Intercept) 0.2039 - -0.8353 (0.1761) - (0.6185) ECAP -0.4474 0.1935 -2.0464 (0.243) . (0.0595) ** (1.9872) 0.3917 -0.0719 3.3223 (0.337) (0.045) (2.088) BS 0.0192 0.0022 0.073 (0.0107) . (0.0068) (0.0484) SG 0.0004 0.0001 0.0002 (0.0002) * (0) (0.0003) REG -0.015 -0.0027 -0.0437 (0.0054) ** (0.0025) (0.0281) OWN 0.0143 - -0.025 (0.0234) - (0.0944) TE 0.5421 0.0493 0.6177 (0.0823) *** (0.0215) * (0.2621) * Sigma 0.191 - - (0.008) *** - - Log Likelihood 68.46 R squared 0.1603 0.0562 F-statistic 8.1157*** 2.3898* Profit Efficiency estimates are truncated at zero. ‘***’ 0.001 ‘**’ 0.01 ‘*’0.05 ‘.’ 0.1 Standard errors are white robust to serial correlation and heteroscedasticity Coefficients have been bootstrapped Hausman Specification Test for ROA and ROE Variable Chi-square p value University of Ghana http://ugspace.ug.edu.gh 142 ROA=ECAP+𝐸𝐶𝐴𝑃2+BS+SG+REG+OWN+TE 26.15 0.0002 ROE= ECAP+𝐸𝐶𝐴𝑃2+BS+SG+REG+OWN+TE 5.6 0.47 APPENDIX D Year HHIA HHID HHIL BI 2000 0.1484 0.1453 0.1661 -2.2603 2001 0.1467 0.1395 0.1829 -2.4558 2002 0.1361 0.1317 0.1234 -2.4716 2003 0.1162 0.1243 0.1235 -3.2549 2004 0.1092 0.1206 0.1152 0.0064 2005 0.0979 0.1053 0.1043 -4.0754 2006 0.0864 0.0954 0.1133 -0.2957 2007 0.0831 0.0811 0.0954 -7.3154 2008 0.0741 0.0743 0.0843 -2.281 2009 0.0670 0.0647 0.0774 -2.8148 2010 0.0607 0.0617 0.0608 -1.4571 2011 0.0644 0.0674 0.0596 -1.7947 2012 0.0630 0.0647 0.0585 -1.8489 2013 0.0614 0.0652 0.0613 -1.5464 Average 0.0939 0.0958 0.1019 -2.4190 University of Ghana http://ugspace.ug.edu.gh