Mauritius Bank Profitability Determinants

Published Date: 02 Nov 2017

INTRODUCTION:

Major changes have been found in the operating environment of the banks across the globe during the last two decades. The structure, conduct and performance of the banks have been involved by both external and domestic components. However the augmented inclination towards disintermediation experimented in many countries, there would have been no economic activity in general without finance and banks. It has been indicated in recent years that small and medium enterprises (SMEs) are the backbone of most economies, and are a key source of economic development, dynamism and flexibility. SMEs in have added approximately 12 percent of the Gross Domestic Product (GDP) and have employed about 20 percent of the work force during 2010, Arnold et al. 2011. Having made the significant role that the SME sector could pay in generating economic growth and making employment, the government of Mauritius, made a number of institutions in an effort to increase the access of SMEs to financial services, as well as answering to the plight of the poor.

The private sector is also affected in the promotion of SMEs through the establishment of SME branches in banking institutions, as well as the setting up of private venture capital schemes. Goddard et al. 2004 point out that although the determinants of bank profit can be analyzed meaningfully, several problems are inadequately treated. For instance, the literature predominantly gives consideration to components affecting profitability at the bank level or macroeconomic level, with variables chosen occasionally unreliable. Also, there is a deficiency of analysis regarding the role of SMEs in adding to the profitability of banks although SMEs play a significant role in Mauritius. The objective is to furnish an assessment of the role of SMEs in affecting banks profitability and other components shall cover macroeconomic components influencing bank profitability.

The previous studies based on the following grounds:

First, emphasis shall be only on Mauritius bank profitability determinants including the role of SMEs. Moreover, we shall be made for more recent data to furnish a more suitable and up to the data empirical evidence. To achieve its broad objectives, it is organized in the following manner: part 2 furnishes a review of the literature and is divided into two sections. The first section shall draw some of the theoretical literature associated with determinants with bank profitability. The part shall also admit previous empirical studies carried out considering determinants of bank profitability both in developed and developing countries. Part 3 furnish an analysis of evolution of banking industry in Mauritius which shall be attained by time series analysis of various bank financial statements statistics to be used in our econometric models. Part 4 describes some of the econometric methodologies to be used, the source of data and describes the original models from which we shall gain our own models based on the Mauritian banking industry in order to measure the role of SMEs in affecting bank profitability in Mauritius part 5 comprises of an econometric analysis of the models developed in part 4 and gives interpretations of our consequences. Finally, part 6 shall then resolve by furnishing a summary of our findings and furnishes suggestions for future research as well as policy implications of our findings.

Problem statement, Gaps and Objectives:

Various commercial banks around the globe have had their financial performance affected during the recent decade on account of the recent sub-prime crisis. Also, some banks have had to shrink their assets and some have even failed and went out of business. Moreover, many studies regarding determinants of banking profitability have been conducted in Asian countries, European and American countries. However, very few studies have been carried out for Mauritius. Even if there are few studies, the sample size is too small or there is a wrong functional form for models utilized. The present study shall close a significant gap in the literature by trying to achieve the following objectives:

To investigate the determinants of Return on Equity (as a measure of profitability) in a sample of 15 banks over the period 2005 to 2011 in Mauritius.

To investigate the determinants of Return on Assets (as a measure of profitability) in a sample of 15 banks over the period 2005 to 2011 in Mauritius.

To investigate the determinants of Net Interest Margin (as a measure of profitability) in a sample of 15 banks over the period 2005 to 2011 in Mauritius.

The above objectives will provide an insight as to whether the banking sector in Mauritius is resilient to external shocks or not.

CHAPTER TWO

LITERATURE REVIEW

2.1 INTRODUCTION:

This part starts by providing an overview of the various concepts in the bank profitability literature which shall include inter alia, basic profitability concepts and ratios and then proceed with the theoretical determinants of bank specific, industry specific and macroeconomic indicators of the bank profitability. The second part of this part confines itself to the various earlier empirical studies carried out with respect to the latter which shall be segregated further in developing, developed as well as in a panel of countries.

2.2 Conceptual Framework – Basic Profitability Concepts:

David Cole (1972) founds a method for evaluating the performance of banks through ratio analysis. A variant of the method is the return on equity model, utilized to analyze bank profitability and identifies particular assesses the risk.

One of the most significant profitability metrics is return on equity (ROE). Return on equity brings out how much profit a bank earned in equivalence to the total amount of shareholders equity found on the balance sheet.

Formula for Return on Equity:

Shareholder equity is equal to total assets minus total liabilities. It’s what the shareholder ‘own’. Shareholder equity is creation of accounting that constitutes the assets created by the retained earnings of the bank and the paid-in capital of the owners. A higher ROE is seen as pertinent as it amends accumulated earnings for the bank so that the bank can more cash dividends when higher profits are made.

Return on Assets (ROA) evaluates how profitable a bank is relative to its total assets. In turn, it evaluates how efficiently a company uses its assets. A higher ROA is better, as it means that a bank is more efficient about using its assets. The connecting between ROA and ROE is by equity multiplier, which is the ratio of total assets to total equity.

Formula for Return on Assets:

Equivalence of a bank’s assets with its equity is made by the bank’s equity multiplier and generally a higher value indicates high bank debt, equity ratio or simple high financial leverage and therefore a profit as well as a risk assess.

2.3 Bank Specific Determinants: Size

In the literature, Civelek and AI-AIami, 1991, Evanoff and Fortier, 1988 contend that a bank’s total deposits, assets or an average assess based on total assets are usually used to catch bank size. Sizes commonly conceive disparities in economics of scale enjoyed by banks. Generally larger banks are improve able to enjoy more economies of scale diversify as opposed to smaller banks. Yet Evanoff and Fortier 1988 and Smirlock 1985 contend that economies of scale can have a convinced influence on profits but can be partly offset by bigger capacity for diversification of assets, which contributes to lower risk and expected return. Thus, the hit of bank size, a theoretical, is in determinant. Smirlock 1985 contend that there exists a convinced and significant link between size and profitability. Granting to Demirguc-Kunt and Maksimovic 1998 the degree to which profitability of bank is involved by financial, legal and other components, for example corruption is very much associated with form size. Furthermore, short 1979 contend the size is powerfully connected to the capital adequacy of a bank as a bigger bank have a tendency to have measure to cheaper capital and, therefore, be more profitable. Similarly, Haslem 1968, Short 1979, Bourke 1989, Molyneux and Thornton 1992 Bikker and Hu 2002, Goddard et al. 2004, associate bank size of capital ratios, which they maintain that are size increases chiefly for small to medium sized banks profitability also increases. Berger et al. 1987 propose that by increasing the size of banks large banks may experience diseconomies of scale.

2.3.1 Risk Levels:

The requirement for risk management is intrinsic in banks. Substandard asset quality and liquidity problems are the two key sources of banks failure. In periods of climbing uncertainty, banks may choose for diversification of their portfolios or maintain more liquid assets in order to minimize their risk. Against this setting, risk may be sorted into credit and liquidity risk. Molyneux and Thornton 1992 assert that there is negative and important connection between the levels of profitability and liquidity, such that the high level of liquidity contributes to lower profitability the other way around.

2.3.2 Asset Quality:

Miller and Noulas 1997 assert that the effect of credit risk on profitability is doubtless negative. This is because as banks concede higher rick loans, non-executing loans increase as well and therefore cut down the profitability of the banks. The loan-asset ratio and the non-executing loan ratio are generally employed to evaluate credit risk. Portfolio theory specifies that investment in risky assets concede with higher returns than with primary assets. Granting to Civelek and AI-AIami 1991 Evanoff and Fortier 1988 the loan asset ratio is commonly extremely concerned with bank profitability while the non-executing loan ratio is negatively concerned. This evaluate of bank risk has created severe consequences, carrying that there is risk reduction behavior among bank managers.

2.3.3 Bank Efficiency:

Bank efficiency is another bank-particular component. An efficient bank is capable to increase profitability by making efficient use of its lending resources. Granting to Demiruc-Kunt and Huizinga 1999 a dimension of banks overhead costs are exceeded on to depositors and lenders while the rest cut down profitability. A rough assess of bank efficiency is bank overhead cost, though X-efficiency, a more advanced assess, has recently become popular. The latter shows how a specific set of prices and quantities of input and outputs change, in accordance with the banks preferred strategy, and how it involves bank profitability. Berger 1995 finds that X-efficiency is systematically associated with higher profits for a large sample of banks.

2.4 Empirical Framework: Firm Particular Determinants of Bank Profitability:

Manly analysis have been done during the past few years on single country analysis such as Greece, Switzerland, UK, USA, China, Columbia, Brazil among others, from the several analysis during the few decades, capital, credit risk, productivity, expenses management and size, these firm particular components have importantly influence bank profitability. The firm particular components are generally classified into single country analysis and panel country analysis.

2.4.1 Single Country Analysis:

Berger 1995 Neeley and Wheelock 1997 furnish some evidences concerning the determinants of banks profitability in USA. Berger 1995 evaluates the connection between the profitability and the capital asset ratio for banks in the US for the period 1983-1992 applying the Granger relation and illustrates a convinced relationship between capital asset ratio of the commercial banks and their profitability. Neeley and Wheelock 1997 look into the profitability of banks in the US for the period 1980-1995. Applying GMM techniques, they find evidence of persistence and a convinced association between bank profitability and yearly percentage change in per capita income. Applying similar methodology, Angbazo 1997 researches the determinants of banks profitability for some US banks during 1989-2003 and find evidence of profit persistence. They also find convinced associations between bank profitability nonpayment risks, the opportunity cost of non-interest accepting reserves, leverage and management efficiency.

Applying pooled fixed consequences framework, Barajas et al 1999 studies the determinants of banks profitability for the period 1991 to 1998 and the finds the consequences of financial liberalization to be important on bank profitability for Columbia. Contempt profitability does not come down after the reform; the significant of the various components influencing profitability were involved by such assess. Applying pooled fixed consequences framework, Naceur and goaied 2001 research the determinants of banks profitability in Tunisia form the period 1980 to 1995 and show that banks which protest to labor and capital productivity progress in development strengthen their capital base and having higher deposits to asset ratios are most profitable. Applying similar framework, Naceur 2003 evaluates the determinants of banks profitability in the Tunisian banking sector for the period 1980 to 2000. He finds that within country inequalities in bank interest margins and net profitability is generally explained by bank particular characteristics. Highly profitable banks are those banks with high level of capital and large expenses. He also finds that banks come along to experience scale inefficiencies given that size has got a negative important consequence on net interest margin.

Also, no affect is found for macroeconomic indicators such as inflation and increase rates on banks profitability. Last, but not least, he finds that stock market growth as got an important impact on bank profitability, proposing complementarities between bank and stock market. Applying multivariate regression analysis, Guru et al. 2002 assess the internal determinants (liquidity, capital adequacy and expenses management) and external determinants (ownership, firm size and external economic conditions) of Malaysian banks profitability over the period 1986 to 1995 and find expenses management is the most important determinant. They also find that while inflation has a positive association with banks profitability, there is a negative association between high interest rates and profitability. Vong and Chan 2009 assess the internal and external determinants of the Macao banking sector for the period 1993-2007 suing shared fixed effects framework and show that capital strength is the most important in explaining profitability as highly capitalized banks are viewed as to be less risky. Yet, assets quality as described by loan loss provisions had a negative effect on profitability. Additionally, they find the most profitable banks to be those with a smaller network of retail deposit taking. Also, inflation is the only macroeconomic variable to have an important influence on banks profitability.

Though multi-variable single equation regression method, Sayilgan and Yildirim 2009 analyze the determinants of banks profitability in Turkey for the period 2002-2007 and show that although statistically important, CPI inflation and lagged of balance sheet transactions to total assets have a negative association with bank profitability as lagged industrial production index, the ratio of budget balance to industrial production and the ratio of equity to total assets have a positive association with bank profitability. Changes in credit risk may reflect changes in the health of banks loan portfolio Copper, Jackson and Patterson 2003 which may consequence the performance of the institution. Duca and Mclaughlin 1990 amoung others, resolve that variations in bank profitability are largely assigned to variations in credit risk, since increased exposure to credit risk is normally associated with decreased firm profitability. This triggers discussions about not the volume but the quality of loans made. In this direction, Miller and Noulas 1997 propose that the more financial institutions being more exposed to high risk loans increase the collection of unpaid loans and decrease the profitability Sufian F. and Chong R.R 2008.

Kosmidou et al. 2004 measure the relationship between profits and asset liability composition for an example of 10 banks in Mauritius. All the models are figured applying fixed effects regression, where they eliminate the firm level heterogeneity through the use of mean deviation data. However, they include a linear time course to take account for the impact of firm level effects over time, as well as other time periods dummies in the regression. They find that capital strength has a positive and overriding influence on profitability, deposit, loan growth, size and expense management are also found important. Mamatzakis and Remoundos 2003 and Athanasoglou et al. 2005 employed GMM to account for profit persistence for bank profitability in Greece. Mamatzakis and Remoundos 2003 measure the determinants of the performance of the banks in the nineties. They find little evidence of persistence in profitability and state that the competitiveness of the banks amended on account of good management decisions, good credit risk management, and market demand features, which have a strong impact on the profitability of banks. Applying similar methodology, Athanasoglou et al. 2005 measure the impact of banks and industry particular as well as macroeconomic determinants of bank profitability spanning the period 1985 to 2001.

They, however, find that persistence of profits lives to a control extent. They also find that apart from size, all banks particular determinants like capital, credit risk, productivity, expenses management importantly effect banks profitability. Dietrich and Wanzenried 2009 employed unbalanced panel data with fixed effects to take out the heterogeneity between observations for 453 banks in Switzerland for the period 1999 to 2006, with a perceptive to analyze bank particular, industry specific and macroeconomic determinants of profitability of the commercial banks. They find important disparities between commercial banks profitability explained largely by the development of banks loans, the share of interest income relative to total income, the funding cost and the effective tax rate. Applying multivariate regression analysis, Sufian and Habibullah 2009 research the determinants of the profitability of Chinese banks during 2000 to 2005 and find that all the determinants have statistically important impact on profitability, albeit not uniform across bank types.

Liquidity, credit risk and capitalization are uncovered to have positive association with state owned banks profitability, while a negative impact of cost is found. However, the most private sectors banks were those with higher credit risk while the least profitable were those with higher cost. They also find the size and cost are associated with low profitability of city banks, where as highly capitalized and the more diversified city banks were highly profitable. The findings are that there is a positive association between economic developments and profitability whilst a negative association between developing in money supply and profitability. Empirical evidence from emerging economies emanate from Columbia Barajas et aI 1999, Brazil Afanasieff et aI 2008, Malaysia Guru et aI 2002, Tunisia Ben Naceur 2002, 2003, Macao Vong and Chan 2009, Turkey Savilgan and Yildirim 2009.

2.4.2 Country Analyses:

Most of the panel country analyses were accomplished by Molyneux and Thornton 1992, Saunders and Schumacher 2000, Abreu and Mendes 2002, Goddard et al 2004, Athanaso et al 2006, Beckan 2007, in Middle East countries by Bashir 2000, formulated and growing countries by Demerguc-Kunt and Huizingha 1999, 2001. Molyneux and Thornton 1992 first examine the determinants of banks profitability comprehensively in a panel of countries applying GMM framework. Applying a sample of 18 European countries for the period 1986 to 1989, important positive relation is found between profitability and interest rates levels in each country, concentration and government ownership. Contempt intensifying rival there is important persistence of abnormal profits from year to year. Applying multivariate regression analysis, Demerguc-Kunt and Huizingha 1999 measure the determinants of banks interest margins for 80countries for the period 1988 to 1995, the determinants covering bank features, macroeconomic conditions, taxation, regulations, financial structure and legal indicators. They find that higher banks with high bank assets to GDP ratio and lower concentration ratio have lower profits.

Local banks were least profitable than foreign owned banks in developing countries than in developed countries. Applying similar methodology Demerguc-Kunt and Huizingha 2001 analyze the effects of financial development and structure on bank profits for various developed and developing countries. They find that financial growth is of utmost significance in explaining bank performance. Explicitly, it is found that higher bank growth is connected to low profitability through acute rival. Yet, it is found that positive relationship between stock market growth and profitability of the banks, signifying complementarities between and bank and stock market. Applying multivariate regression analyses, Bashir 2000 measure the internal and external determinants of eight Islamic banks profitability and efficiency in the Middle East for the period 1993 to 1998, with macroeconomic environment, financial market situation and taxation as control elements. He finds that high profitability is the result of high leverage and high loans to asset ratios. Foreign owned banks are also seen as most profitable as local banks. Also, while taxation has a negative association with profitability, macroeconomic variables and stock market growth have a positive one.

Applying pooled fixed effects model, saunders and Schumacher 2000 measures the determinants of bank profitability of six European union countries and the US for the period 1988 to 1995 and determine that macroeconomic imbalance and regulations have an important impact on profitability. They also bring negative causality between bank solvency as showed by high capital to asset ratio and lower cost of intermediation as showed by low interest margins. Applying multivariate regression analysis, Abreu and Mendes 2002 study a set of European countries in the nineties and determine that majorly capital banks were more profitable outstanding to lower expected bankruptcy cost. As macroeconomic indicators, albeit negative, inflation and unemployment rates are found to be important. Goddard et al 2004 study European banks profitability for the period 1992 to 1998 applying GMM model. Their framework includes size, diversification, risk and ownership types as well as dynamic effects as determinants of profits. They determine that although rival is intense, there exists important persistence in banks supernormal profits over the years. Also, they determine that the size profitability relationship for the banks is quite inconsistent and disorganized capital assets ratio and profitability is positive and the relationship between the significance of off-balance sheet business in a bank portfolio and profitability is positive for the only for UK.

In increase, Athanasoglou, et al. 2006 also use GMM technique to measure the behavior of south eastern European banks for the period 1998 to 2002 and determine that the improvements of banks’ profits in these countries necessitates efficiency and new risk management standards, there by effecting profits. Also, whilst the important of macroeconomic determinants is mixed concentration is found to affect profits positively. Beckman 2007 studies the structural and cyclical determinants of banks profitability for 16 nations in Western Europe for the period 1979 to 2003 applying the Hausman-Taylor instruments variable estimator. He determinants the financial structure and higher diversification regarding banks income sources are important. Yet, concentration of national banking system, does affect profitability importantly, albeit business cycle effects as showed by lagged GDP growth, shows an important procyclical effect on banks’ profits.

Flamini et al 2009 study the determinants of 389 banks in 41 sub-Saharan African countries for a ten year period ending 2006 and determines that majorly profitability banks are those with large size, activity diversification and private ownership, albeit those with high credit risk incline to be less profitable. It is also found that macroeconomic policies that promote price stability and stable economic growth increase credit growth and thus profits. Applying GMM technique, it is also determine that there is average persistence in profitability. Further, Granger causality from profitability of capital happens with important lag proposing that profit are not retained for recapitalizing the banks and implying that higher capital demands are needed to improve financial stability.

CHAPTER THREE

ANALYSIS OF THE MAURITIUN BANKING SECTOR

3.1 INTRODUCTION:

In this part, it is an introduction to the nature and development of the domestic banking sector. The evaluation of profitability as evaluated as Return on Equity (ROE), Return on Assets (ROA), Net Interest Margin (NIM) and the various bank-specific, industry-specific elements for the ten banks taken as sample as well as the macroeconomic elements influencing the three profitability measure are made for the period spanning 2003 to 2012. Spearman rank correlation analysis is also made for all variables for the same time period.

The nature and development of the Mauritian Banking Sector:

The Mauritius financial sector was evidenced by financial liberalization during the 1980’s and early 1990’s. There was a need to liberalize and deregulate markets with a view to modernize and develop the domestic financial services sector. Reforms in the financial sector were viewed as pertinent for financial sector growth and the sector now is composed of a plethora of financial institutions which cover bank financial institutions as well as non-banks such as insurance companies, pension companies, stock brokers, and investment companies, non-bank deposits taking institutions, leasing companies, credit institutions, money changers and foreign exchange dealers. Contempt the concept that these institutions display distinct characteristics, overtime the differences are fading with the advent of diversification experienced by banks as a result of increased rival. Currently, the Mauritius banking sector is composed of 20 banks. The main activities of the Mauritius banks continue taking deposits and advancing loans.

However, it is found that banks are diversifying by enhancing their product base. Consistent to the statements made in banking legislations in 2008, banks are now able to extend Islamic products as well. In this respect, HSBC set up its first Islamic banking product in 2009. Thus the banks are not seen as homogeneous in terms of their activities. Owing to founded market share and customer loyalty, most domestic banks focused their activities domestically while others preferred to do business with non-residents. We note that credit development increased for the industry on account of additional input package and budgetary evaluates taken by the government to encourage employment through small and medium enterprises.

Their main sources of funds are still from deposits from public. Rupee and foreign currency deposits as a percentage of total deposits in the banking sector endured at 36% 64% respectively in the year 2008. As far as capitalization is pertained, all banks in the industry are well capitalized, detecting more than the 10% solvency ratio (capital adequacy ratio) dictated by the bank of Mauritius. The Basel II model was put into cause in the year 2009 and all banks are now expected to focus capital in line with the standard approach to credit risks. Following the Basel II model, capital adequacy ratios for credit and operational risks ranged from 10% to 98% in 2009.

3.3 Development of Mauritian Bank Profitability:

Bank profitability as evaluated by average net interest income, average return on assets and average return on equity depicted a general developing trend during the period 2003 to 2012. The trend in profitability for the ten banks taken in our sample is explained by following diagrams.

Chart 1 show the growth of average net interest income of the banks taken in the sample over the period studied. We note that same generally increased consistent manner from 2.46 in 2003 to 3.05 in 2012 reflecting lower interest provisioning built over the years and better margins registered on the advances built. However, we also comment that after reaching a high of 3.39 in 2007, the ratio starts to fall owing to introduction of new banks which drew banks to revise their margins and maintain their market shares.

Average return on equity (ROE) as showed in chart 2 also depicted a general arising trend from 11.60 in 2003 to 13.01 in 2012, albeit showing a declining trend up to the year 2006. The decline in the average ROE up to the year 2006 is due to the big losses because of big provisions for bad debts and collected unrecovered non-performing loans. Also, non-interest income was involved by slowing in trade flows, both local and international. However, the turnaround in profitability is due to the concept that more efficient banks took over the loss making banks and that banks in general entered on monolithic diversification strategies to maximize shareholders wealth.

In chart 3, on average banks in the sample had return on assets above 1 for the most of the period studied. This is explained by the concept that most banks followed stringent risk management practices for safety’s sake. They had to keep a buffer as they are facing tough times, particularly in the light of rival pressures emanating from non-bank financial institutions. However, the falling trend from year 2003 to 2012 in the mean ratio is accounted for by loss making banks such as Bank One Limited (Ex-First City Bank Limited) and Bramer Banking Corporation (Ex-South East Asian Bank for the years under consideration). However, other banks were determined to have acceptable returns on assets as they started to draw not only interest income but non-interest income as well which was in line with their diversification programme as credit development increased at decreasing rates.

3.4 Review of Asset Quality of Banks:

Following Athanasoglou et al 2005, asset quality is evaluated by the ratio of non-performing advances to total advances. Although individually, out of the ten banks eight banks kept their ratio on non-performing advances under 10% exceptionally, Bank One Limited (previously First City Bank LTD/ Delphis Bank), which had its ratios skyrocketed as from the year 2002. As a result, the average ratio depicted a rising trend from 4.93% in 2003 to reach 5.48 in 2004. The following helps to explain such a state of affair.

Chart 4 depicted the trend in the average non-performing loans ratio. As explained previously, on a macro level, most banks save Bank One Limited were able to have good ratios. This is because of the following main reasons. (i) Better credit risk management following the setting up of the Mauritius Credit Information Bereau in 2005 (ii) more strength credit risk management policies practiced by the banks with the advent of the credit risk management and credit absorbed guidelines issued by the Bank of Mauritius. As an effect, banks were selective in conceding credits among prospective clients (iii) the banks were enjoying managerial economies of scale by employing well qualified and experienced recovery officers who were able to recover bad debts efficiently and effectively. As an effect there were able to have ratios below the mean.

An individual analysis, however, disclose that the non-performing generally increased for Bank One Limited (previously First City Bank LTD/ Delphis Bank). The high ratios shows by Ex First City Bank Limited/ Delphis Bank can be explained by firstly, settlement of the Delphis Bank in the year 2002 which means that many retail as well as corporate clients went bad credits. Secondly, when Delphis Bank was taken over by First City Bank in 2002, about 61% of the shares were owed by the public sector. As an effect, there were a lot of concerned party loans remained unpaid. Also, due to political interference, a quite important portion of the portfolio of loans was given to bad credit risks. Furthermore, the loan portfolio of the ten banks were mainly focused to the small medium textile firms, which were heavily obligated, had to lose their comparative advantage against rising economies like china, Pakistan, India and Srilanka and were finally went bankrupt. As an effect receiver managers were appointed. Last, but not least, the bank had to face adverse report risk and managerial diseconomies of scale and could not employ proficient in credit risk management and recovery of debts.

3.5 Trend in Cost Efficiency Ratio of the Banks:

The cost efficiency ratio as showed by the ratio of costs to income decline slightly from 1.76 in 2003 but continued above 1 at 1.42 in 2012. We add below chart showing the measure of the average ratio for the period 2003 to 2012 for the ten banks studies.

On an individual’s basis, some banks in the sample had a slight rise in the ratio on account of expansion in operating expenses consistent with rise in staff costs in addition to heavy investment expenses relating to building maintenance and esteem of their existing IT systems. However, the development was lower than budgeted owing to a decrease in charges for depreciation and better cost management strategies. Yet, the decline in the average cost efficiency ratio over the years may be explained by the banks management ability to keep overhead costs low while at the same time income from bank services increased on an average basis.

3.6 Trend in the Size of Banks:

To follow Sam et al 2003, we take the square of the logarithms of total assets to get the size of the banks. The average size of the ten banks in the sample increased quite quickly over the period studied. Conceive the following diagram which explains such a rate of affair.

Chart 7 makes it clear that the average size of banks made greater in size continuously over the period studied. Actually, following section 14(1) of the banking act 1988, commercial banks operating in Mauritius were expected to maintain certain level of capital to shield themselves against solvency risks. In result, in 2004, consistent to the banking act 2004, smaller banks were extended to recapitalize and to have a minimum capital of MUR 200m. The raise in the average size of banks may be also explained by the fact that the assets of the banks raised importantly owing to the diversification of their portfolio towards investments as well instead than just trusting on traditional loans.

3.7 Research Approach:

In terms of ascertain facts study there are two common ideas to business and social research

Deductive approach formulates theories and hypotheses complied by a research strategy to test the hypotheses.

Inductive approach finds data and formulates theories as a result of the data analysis.

The deductive approach brings in a high level of objectiveness in research through external notice insofar as the choice of questions and subsequent phrasing are not subjective. In contrast, the inductive approach furnishes a high level of subjectiveness and a number of theoretical future prospects based on the context of the individual research position. This study will analyze the previous findings in the literature, and apply the consequences in current practical settings. Therefore, deductive approach was followed by constructing an empirical model and hypothesizing its collinear relationship between target determinants and its dependent variable profitability of banks.

3.3 Data Analysis:

Due to time constraint and data accessibility, the study is based on secondary quantitative data. Which obtain from public database. Researcher said that the advantage of using secondary data admits the higher quality data equated with the primary data collected by the researcher themselves. The feasibility to carry on longitudinal studies, which is the case in this study and the permanency of data, which mean secondary data generally furnish a source of data that is both permanent and available in a form that may be assured relatively easily by others, i.e. more open in public examination. Therefore, increase the reliability of the data.

3.8 Limitations:

The consequences in the previous part should be interpreted with caution. Firstly, the number of banks has taken stands at only ten in the sample, which may prove to be insufficient. However, including all banks to derive more authentic examination of the determinants of profitability would be time and effort consuming. Besides, the methodology applied refuses the operating condition of the selected banks, proposing that our consequences could be bias, inaccurate and unreliable. For example, it refuses that some banks may be in restructure such as merger or acquisition which definitely has an impact on reported data. Moreover, the methodology applied is also limited from the nature of the ratios used. For instance, a variable might be evaluated by several ratios. For example, credit risk, non-performing loans ratio may be evaluated in different ways. The measures for credit risk and non-performing loans ratio chosen in our framework might be bias as they may misrepresent the true measurement of variables.

3.9 Conclusion:

In this part, we have concluded a statistical analysis by furnishing some trends in the variables for Mauritius that we shall be using in our model in the next parts. In essence, we have surveyed how bank profitability has developed over the years. We have also furnished an analysis of the bank specific, industry specific and macroeconomic components influencing bank profitability as well as spearman rank correlations between all the variables.

CHAPTER FOUR

METHODOLOGY

4.1 INTRODUCTION:

The part starts by bringing in the data source and aimed methodology referring to the study of the determinants of banks profitability. We then continue by giving the baseline model adjusted from Flamini et al 2009. Thereafter a review of the applicable theories considering econometric study viz, the classic fixed effects, the random effects and the Hausman stipulation test are furnished. The reserve diagnostic tests of the like of Ramsey-RESET test, which test for stipulation test furnished. This furnishes the points about the methodology followed in assist in achieving research objectives. It include the approach followed to study the effect of main determinants on banks profitability, the type of data used and the techniques employed to collect the data, the sampling mechanism admitting sample size, the methods applied to manage and examine the data, and the process of constructing empirical model with recognition and the measurement of its factors.

The analyses are based on a panel of 15 commercial banks. Data is found from the annual reports of the banks. The sample data starts in 2003 to 2012 for the simple cause of data inaccessibility for some banks. Rather doing a cross section analysis, this study looks at a panel data stipulation. Baltagi 1995 highlights the advantages of applying the latter. First, panel data accounts for the conception those banks and the countries are heterogeneous in conditions of accounting policies and standards within certain dynamic period, a situation that cannot be found in either time series or cross section analysis. Second, in panel data we admit all the banks data without having an appeal to aggregation or averaging, which of course eliminates all biases linked with the latter.

4.1.1 Panel Data Analysis:

Actually panel dataset is defined as notices that are repeated on the same set of cross sectional units. Panel data applied in cases where the number of observations as well as the period conceived is greater than one. Baltagi 1995 contend that panel data analysis improve the efficiency of figures. To kunze 2002 conception can be used to explain the panel data approach. We conceive a two period panel data model and accept the following:

4.1.2 Fixed Effects Model:

4.1.3 Random Effects Estimator:

4.1.4 The Hausman Specification Test:

4.1.5 The Ramsey Reset Test:

Sometimes, some applicable variables might be excluded from a model. Similarly, the model might often include irrelevant variables. Both the inclusion of irrelevant variables and excluded of applicable variables are bound to cause a specification error. If a applicable variable is excluded, the coefficient of the variables of the model will biased as well as inconsistent estimates. Similarly, if we include an irrelevant variable, the error variation will be inflated. The Ramsey test (Ramsey, 1969) is a mis-specification test done to indicate that there is some form of mis-specification. However, it does not give us the right specification.

If the F critical value is less than the F test statistics, we refuse the null hypothesis which means that the true specification is non-linear. On the other hand, if the F critical value is greater than the F test statistics, we do not refuse the null which means that the true specification is linear and the equation passes the Ramsey Reset test.

4.1.6 Testing for Heteroscedasticity:

While performing the OLS regression it is accepted that the variance of the residuals is fixed, that is there is the presence of homoscedasticity. Yet, if the variance of the residual is non- fixed we say that heteroscedasticity is present. Breusch-Pagan/ Cook-Weisberg Test/Cook-Weisberg test is done and accordingely if we refuse the null hypothesis we say that heteroscedasticity is present. However, if we do not refuse the null hypothesis, we say that heteroscedasticity is present. The hypothesis for the test is as follows.

Results Breusch-Pagan/ Cook-Weisberg Test for Heteroscedasticity

H0: Fixed variance

H1: Non fixed variance.

Results Breusch-Pagan/ Cook-Weisberg tests test the null hypothesis that is error variances are all equal versus the substitute that the error variances are the multiplicative function of one or more variables. For example, the substitute hypothesis states that the higher the forecast value of ROA the higher the variance is. The chi-square should be equated to Prob>chi2 if chi square is high as opposed to Prob>chi2 heteroscedasticity is present.

4.1.7 Testing for Multicollinearity:

If the regressands are absolutely correlated, unique figures for them cannot be obtained. The term multicollinearity cites to a situation where two or more variables are near to absolutely correlated. The variance inflating component VIF can be applied to know whether there is multicollinearity or not. Normally, multicollinearity is severe when the VIF values are greater than 10 which require investigation. Tolerance which is 1/VIF is most usually used to see whether there is multicollinearity. A value of 0.1 is corresponding to VIF value of 10.

4.2 Framework Specification:

To furnish a bottom up analysis for banks profitability, we shall develop a framework similar to that of Flamini et al. 2009. However, we take off from his framework to include variables that are conceived more significant for the Mauritian banks. Let us first explain the Flamini et al 2009 framework- the baseline framework and then develop our framework.

4.2.1 Base line Framework:

4.2.2 Our Framework:

We depart from the framework suggested by Flamini et al. 2009 by including in our framework, variables which are conceived more significant for Mauritian banks. First and foremost in line with Athanasoglou et al 2005 and Flamini et al. 2009 we shall employ only Return on Assets as profitability assess. We continue below by giving a discussion of the bank particular influence on bank profitability and five their required signs.

4.2.3 Measurement of Bank Profitability:

As already explained in previous part, assess of bank profitability applied in this paper is the Return on Assets (ROA). ROA is a ratio calculated by dividing the net income over total assets. And evaluate the profit earned per USD of assets and reflect how bank management will use the banks real investment resources to generate profits. ROA has been chosen as the key proxy for bank profitability because it takes into consideration financial leverage and the risk related with it. However, ROA may be biased due to off-balance sheet activities, but such activities are negligible for Mauritian banks, while the risk related with leverage is likely to be significant despite the institutional institutions that these financial institutions incorporate in order to pay for informational asymmetries.

4.2.4 Specific Determinants of Bank Profitability:

Size

Size signals specific bank risk. As governments are less likely to allow large banks to fail, risk access to size would predict the larger banks would require less profit. However, if large banks have a bigger market share and operate in a non-rival environment, lending rates tend to be higher while deposit rates will be lower because of safety percepts from depositors. Thus, big banks may enjoy higher profits. Moreover, modern theory anticipates efficiency gains associated to bank size, owing to economic of scale. In most studies, Athanasoglou et al 2005, Sam et al 2003 among others, the authors show a positive relationship between bigger bank size and bank profitability. Yet, owing to mainly bureaucratic causes, when bank become extremely big, their profitability declines, explaining a non-linear relationship between bank size and bank profitability. To capture this non-linear relationship, we take the logarithm of banks assets and their square.

Asset Quality:

Asset quality is proxied by the ratio of non-performing loans. Theoretically, at least, increased credit risk decreases bank profitability. In perspective thereof, a negative relationship is required between bank profitability, ROA. Thus, banks can increase profitability levels by efficient screening and assure of credit risk which covers forecasting of future risk levels. Also, the Basel committee has set standards for the level of non-performing loans to be detected by banks. Therefore, credit risk is conceived as a predetermined variable in the framework.

Capital:

Capital should be a significant variable in determining bank profitability, although in the presence of capital demands, it may proxy, risk and also regulatory costs. In imperfect capital markets, well capitalized banks required to borrow less in order to support a given level of assets, and incline to face lower cost of financing due to lower position bankruptcy cost. Many analyses mention that in the presence of asymmetric information, as well as capitalized banks could furnish a signal to the market that a better than average performance should be anticipated Athanasoglou et al 2005 and Berger 1995. We proxy for capital with the ratio of equity to total assets and based on the above considerations, we model it as a predetermined instead strictly exogenous variable. Most of the analyses determine a positive and important effect of capital on bank profitability.

Credit Risk:

The main source of bank-specific risk is credit risk. Lack of information on borrowers, poor legal environment and insufficient enforcement of creditor rights disclose banks to high credit risk. On a macroeconomic view, weak economic growth increase credit risk due to deterioration of credit quality and raises the profitability of default. The proxy that we applied for credit risk is the ratio of loans to deposit and short term financing as this will furnish a forward looking evaluate of bank exposure to default and asset quality deterioration. Based on standard asset pricing arguments, we expect a negative relationship between profitability and credit risk.

Cost Efficiency:

Segregation can be made as regards the total cost of banks in terms of operating cost and other expenses, albeit only operating expenses are considered to be the outcome of bank management. Cost efficiency is evaluated by the log of operating expenses and we anticipate a negative relationship between cost efficiency and bank profitability as an advance in managing these expenses increases efficiency levels and hence profitability. Cost efficiency enforces particularly to Mauritian banks as we have seen in the recent decades that excess personal expenses in some banks resulted in lower profitability for some banks.

Activity Mix:

Bank activity mix is very significant determinant of banks profitability. It is a significant proxy for the overall level of risk undertaken by banks to the magnitude that different sources of income have different credit risk and volatility. A evaluate for activity mix is the ratio of net interest revenues over other operating income. Interest earning activity are generally considered as riskier than fee based activities, therefore they have higher returns. Dimirguckunt and Huizinga 1998 determine that banks with relatively high non-interest earning assets are in general low profitability.

Loans to SMEs:

SMEs play significant role in the development of Mauritius. The higher the loan amount to SMEs, the higher will be the profitability of the banks. A positive sign is anticipated

Bank Specific Variables

Macroeconomic Variables

Indicators

Expected Sign

Indicators

Expected Sign

Size

Positive

Inflation Expectations

Negative

Asset Quality

Positive

Central Bank Interest Rates

Positive

Capital

Positive

Cyclical Output

Positive

Credit Risk

Negative

Cost Efficiency

Negative

Loans given to SMEs

Positive

Table: Parameters to be figured and their expected signs for our framework

CHAPTER FIVE

ECONOMETRIC RESULTS AND ANALYSIS

5.1 INTRODUCTION:

This part comprises of an econometric analysis of the framework developed in part 4. Applying stata 11.1, we carry out OLS regression for equation Returns on Assets (ROA). Based on our Hausman specification test, we will choose random effects for the equation given there are important differences across banks. The part also includes post OLS estimation tests for heteroskedasticity and framework specification.

5.2 Ordinary Least Squares Results for Banks Profitability Determinants:

The results based on plain OLS for the equation Return on Assets (ROA) are given in table as follows:

Coefficient

t

P Value

Lagged ROA

0.4023

3.80

0.020

Credit Risk

-0.020

-2.61

0.0081

Capital

0.327

-2.24

0.010

Asset Quality

0.0599

0.86

0.459

Loans to SMEs

-0.1112

-0.57

0.425

Cost Efficiency

-0.0018

2.31

0.030

Net Interest Revenue to Operating Expenses

-0.0212

0.84

0.403

Size

0.0017

0.72

0.650

Fixed

0.0335

0.90

0.396

Prob>F

0.0000

R-Squared

0.7923

Adjusted R-Squared

0.77565

Source: Stata 11 output

The following has been detected: for the variables in the equation, the P-Value of the framework (Prob>F) stand at 0.000 which is less than 0.05. The later shows a statistically important relationship between the profitability indicators and the explanatory variables of the equation and show the reliability of the variables in explaining profitability. The R-Square figures indicate that the framework explains 79.23% of the variance in ROA. Adjusted R-Square, however, are a better indicator of variance than the R-Square and stood at 0.7756 which means that the framework explains 77.56% of the variance in ROA.

The P-Value is the two tail values for each parameter figures and tests the hypothesis that the figures are different from zero. To refuse the hypothesis P-Values have to be less than 0.05. Capital, credit risk and lagged ROA are determined to be statistically important for the equation. Except for these three variables, all the other variables applied in the framework statistically insignificant for the ROA equation. The statistical importance of some of the parameters is confirmed by their absolute t values which endured above 1.96.

Further, the coefficient for loans given to SMEs is negative and insignificant in explaining bank profitability in Mauritius. SMEs do not lead to bank profitability in Mauritius. Banks are more pertained about large corporates as SME funding is seem to be highly risky as the latter do not have sufficient collateral to furnish and are therefore constrained in terms of bank funding. The ideal source of finance is through leasing companies. However, the most significant determinant of bank profitability is credit risk. in line with Athanasoglou et al 2005, Flamini et al 2009, capital of banks is important in explaining profitability under OLS regression of ROA. Our result shows that a 1% raise in capital of banks contributes to 3.72% in Return on Assets, proposing that well capitalized banks experience higher returns. Also, banks diversified their portfolio towards investments instead just relying on traditional loans which increased their assets and thus profitability.

Moreover, in contrast with Mamatzakis and Remoundos 2003, Athanasoglou et al 2005, Vong and Chan 2009, the non-performing loans ratio (NPL) has no statistically important impact on profitability. This statistically insignificance of the asset quality is explained as follows:

The loan portfolio of banks is so high than the non-performing loans represent a very less percentage of the total assets of the banks and thus it does not affect the profitability.

More stringent credit risk management policies practiced by the banks following by the recommendations of Basel II. As an effect, banks are more selective in granting credits among potential clients.

Banks nowadays have a well-diversified portfolio of assets instead the traditional focus on loan assets. Therefore banks are investing on other financial instruments which are low risky and earning more returns. As an effect, non-performing leases do not have an important impact on banks profitability.

Following the recent financial crisis, banks are more stringent in debtors’ management. They have a very close look on the debt lists and they are using the recommendations made in Basel II for credit management.

Further, in line with Bashir 2000, Athanasoglou et al 2005, credit risk as showed by the credit/deposit ratio is important ROA. Our result proposes a negative relationship between credit risk and ROA, proposing that a 1% raise in the credit risk ratio leads to a decline of 0.20 in ROA. The statistical importance is explained by the fact that banks had been able to gather more deposits than furnishing credits, resulting in excess liquidity in the banking system. Also the relative decline in credit growth is attributed to the recent financial crisis where investors are more conservative on their investments thus preferring a less return but a safer investment.

Further, in line with Guru et al 2002, Kosmidouet al 2004, Athanasoglou et al 2005, Famini et al 2009 but in contrast to Naceur 2003, the cost efficiency ratio is insignificant for the ROA. Out result propose a positive relationship between cost efficiency and ROA, proposing that a 1% raise in the ratio of operating expenses leads to a raise of 0.18% in ROA. Since overheads costs in Africa, we would anticipate this variable to enter the regression importantly and with a negative sign. The positive and insignificant coefficient in our results, rather, propose that banks are able to pass on most of the overheads costs to customers through larger spreads in order to keep profits unaffected. To the level that banks’ ability to overcharge is a function of their market power, this outcome presents manifest of market power incidence in the banking sector.

The ratio of net interest revenues to other operating income enters a regression with a negative and insignificant coefficient. This showed that greater bank activity diversification, as implied by bigger shares of services in the bank activity mix, positively influence returns. This is probably due to the fact that fees represent a more stable source of income than loans. In contrast to Naceur and Goaied 2001, Naceur 2003, Kosmidou et al 2004, Sufian and Habibullah 2009, size of banks is insignificant in explaining profitability under OLS regression of ROA. Our results depict that a 1% raise in the size of banks leads to 0.17% decline in Return on Assets. The negative coefficient of size, important at the 5 percent level, shows that this relation might be non-linear due to potential bureaucratic bottlenecks and managerial inefficiencies endured by banks has the become too large. The marginal statistical importance of the regression coefficient, on the other hand, contributes further evidence to the hypothesis that. Thanks to some degree of market power, banks manage to pass on to depositors and borrowers possible inefficiency without affecting profit in a significant way.

5.3 The Hausman Specification Test:

For the profitability regression, the Hausman specification test gives us a tabulated value of Prob>chi2 which is less than the stata output, below which means that the Random Effects is favored

Hausman Specification Tests Results:

Chi2(10)

41.17

Prob>chi2

0.0000

Table Source: Stata 11.1 Output

The value for rho as depicted in table in the annexure mean that differences across banks have important influence on profitability which further strengthens our selection of the Random Effects. Similar to what has been determined in the OLS regression, loans given to SMEs is insignificant in explaining banks profitability in Mauritius. Banks are more pertained about large corporates as SME funding is seen to be highly risky as the latter do not have sufficient collateral to furnish and are therefore constrained in terms of bank funding. The ideal source of funding is through leasing companies. However, the most significant determinant of bank profitability is credit risk.

Table: Random Effects Estimates

Coefficient

t

P Value

Lagged ROA

0.5013

2.69

0.0012

Credit Risk

-0.004

-2.44

0.0016

Capital

0.0475

2.12

0.0015

Asset Quality

0.0945

0.84

0.321

Loans to SMEs

-0.0185

-0.49

0.421

Cost Efficiency

-0.754

-2.13

0.0152

Net Interest Revenue to Operating Expenses

-0.0475

-0.79

0.301

Size

0.0045

0.558

0.599

Fixed

0.0145

0.90

0.241

Prob>F

0.0000

R-Squared

0.7857

Adjusted R-Squared

0.7714

Source: Stata 11 Output

5.4 Data and Model Analysis:

We have also done the following tests for the data gathered to find meaningful estimates and results of the framework:

Breusch-Pagan/ Cook-Weisberg Test to test for heteroscedasticity.

Testing for Multicollinearity applying the various inflating components.

Ramset reset test to test for framework specification.

5.4.1 Testing for Heteroscedasticity:

Table: Breusch-Pagan/ Cook-Weisberg Test for Heteroscedasticity

Chi2(1)

7.55

Prob>chi2

0.0060

Table Source: Stata 11.1 Output

Breusch-Pagan/ Cook-Weisberg Tests test the null hypothesis that the error variations are all similar versus the alternative that the error variance are a multiplicative function of one or more variables. For example, the alternative hypothesis states that the larger the predicted value of ROA, the higher the error variance is. The bigger chi-square as against the Prob>chi2 showed that heteroscedasticity was present in the equation. This is what was anticipated given that banks of different sizes were taken in the sample. The consequences also mean that the standard errors are not only biased but also inefficient.