Int. J. Corporate Governance, Vol. 8, No. 1, 2017 Income diversification, bank stability and owners identity: international evidence from emerging economies Naima Lassoued* Higher School of Business, University of Manouba, Campus Universitaire 2010, Manouba, Tunisia Email: naima.lassoued@sesame.com.tn *Corresponding author Houda Sassi Institute of High Business Studies, Carthage University, 2016, Tunisia Email: houdasassi01@yahoo.fr Abstract: This paper examines how ownership structure affects the relationship between income diversification and bank stability in emerging markets. Examining data on more than 171 banks from 13 Middle East and North Africa countries examined over the 2006–2012 period, the impact of foreign and state ownership on income diversification and bank stability is jointly studied by a 3SLS model. The main finding is that diversification, across interest and non-interest activities and non-traditional banking activities, increases bank stability. State-owned banks seem to take more risk by diversifying less their income. Inversely, foreign ownership decreases risk by diversifying their income. The findings have significant strategic implications for bank managers, regulators and investors who share a common interest in boosting bank stability. Keywords: bank stability; emerging economies; foreign ownership; income diversification; state ownership. Reference to this paper should be made as follows: Lassoued, N. and Sassi, H. (2017) ‘Income diversification, bank stability and owners identity: international evidence from emerging economies’, Int. J. Corporate Governance, Vol. 8, No. 1, pp.61–80. Biographical notes: Dr. Naima Lassoued is an Associate Professor of Accounting and Finance at Higher School of Business Tunisia. She holds her Habilitation in Finance from University of Manouba. Her main research interests include corporate governance, financial reporting, Islamic finance, risk management and earnings management and firm performance. She has published in Research in international business and finance (Elsevier); Journal of Accounting Auditing and Performance Evaluation (Inderscience), corporate governance: The international journal of business in society, Research in Accounting in Emerging Economies (Emerald Publishing). Copyright © 2017 Inderscience Enterprises Ltd. 61 62 N. Lassoued and H. Sassi Houda Sassi is a Teaching Assistant. She received her PhD in finance from the Institute of High Business Studies (University of Carthage). Her research interests are in corporate governance, banking institutions and earnings management. She has published in Research in international business and finance (Elsevier), Economics bulletins. 1 Introduction Over the last few decades, financial institutions had been subject to increased competition, intensified by information technology development, globalisation and deregulation. These changes incited banks to take more risk in order to maximise their expected profit and to adopt new lines of businesses in addition to their traditional interest-based activities, like loan granting (Keeley, 1990; Meslier et al., 2014). More precisely, banks diversify their activities by moving towards non-interest income activities like trading securities, brokerage, investment and other activities. By referring to modern portfolio theory, income diversification seems to be a valuable technique to reduce banking risk when all income components either negatively or only slightly correlate (Chiorazzo et al., 2008; Busch and Kick, 2009). However, most empirical studies predominantly in the US and Europe1 found that non-interest activities are often associated with higher risk because of their unstable nature. These findings may not apply to emerging/developing banking sectors, because the latter often have less mature financial systems and different banking market structures and institutional and regulatory backgrounds (Meslier et al., 2014). Relatively, little recent evidence addresses this issue in developing countries and often reach mixed findings (Sanya and Wolfe, 2011a; Meslier et al., 2014; Zhou, 2014). The mixed empirical results may be attributed to the omission of reliable variables like ownership structure. Indeed, according to Pennathur et al. (2012), studying income diversification and banking risk in emerging economies must take into account ownership structure since these countries are known by dominant controlling shareholders. These latter have a significant impact on corporate decisions because of their sizable capital stakes and weak law enforcement (LaPorta et al., 1998). In this regard, risk-taking is expected to be more practiced in firms with concentrated ownership than in firms with dispersed ownership. Accordingly, diversification may be an interesting channel to control risk (Jensen and Meckling, 1976). However, when majority stakeholders do not own diversified portfolios, they will not be incited to increase banking risk. Thereby, controlling shareholders do not have the same motivations, objectives, means and control efficiency. It would be appropriate to study the identity of controlling shareholders. Our purpose in this study to extend the literature on the effect of ownership structure, including state and foreign ownership, on the relationship between revenue diversification and banking risk in the developing Middle East and North Africa (henceforth MENA countries). Indeed, the banking sector in the MENA region seems to be an interesting area to test our research problem, since it has undergone financial openness and liberalisation aiming at limiting state intervention and establishing a legal and an institutional framework that would attract foreign investments (Ben Naceur et al., 2007). This policy has shifted the ownership of large shares of banks from government to private control and from domestic to foreign control. Such changes took place as Income diversification, bank stability and owners identity 63 governments privatised many of their state-owned banks. The most important international bank groups have generally set up a wide presence in the MENA region through subsidiaries and are competing with domestic banks. Foreign presence improves human capital skills and may lead to more diverse products, better use of up-to-date technologies and knowledge transfer forcing domestic banks to schedule major structural reforms in order to compete on an equal basis with their counterparts. In addition to the intense competition of foreign banks, banks in the MENA region also face increased competition from Islamic banks, likely to encourage income diversification (Almarzoqi et al., 2015). Consequently, the study of income diversification and risk in the banking sector assumes particular relevance in developing countries, since it would shed light on the eventual consequence of the liberalisation and deregulation process, contributing thus to an increase in financial risk-taking. Therefore, policy makers need an accurate understanding of the impact of liberalisation and privatisation on income diversification and bank risk, in order to identify the proper financial measures. We examine the effect of ownership structure on the relationship between income diversification and bank risk using data on more than 171 banks from 13 MENA countries over the period 2006–2012. We specify a 3SLS model in which the impact of foreign and state ownership on income diversification and banking risk is jointly examined to solve the endogeneity problem that arises from the relationship between risk-taking and diversification. Our findings indicate that diversification, across interest and non-interest activities and non-traditional banking activities, decreases bank risk. State-owned banks seem to take more risk by diversifying less their income. Inversely, foreign ownership decreases risk through diversifying their income. Then, we came to confirm that diversification is a channel through which controlling shareholders influence banking risk, and diversification level depends on the monitoring practiced by different shareholders. Our study contributes to the literature in several ways. First, we extend the recent line of research to provide international evidence on the impact of income diversification on banking risk by studying unexplored emerging markets. Second, we extend the results of previous studies (Pennathur et al., 2012; Sanya and Wolfe, 2011b). Examining a sample of Indian banks, Pennathur et al. (2012) studied the effect of income diversification on banking risk for state-owned and foreign-owned banks. Unlike their study, we investigate the effect of the type of ownership (foreigner and state) on both diversification and risk while controlling for potential endogeneity between risk and diversification. Unlike Sanya and Wolfe (2011b) who examined the effect of ownership concentration on income diversification and banking risk in European banks, we explore the effect of shareholder identities as most banks in the MENA region already have a concentrated structure (Rocha et al., 2011; Lassoued et al., 2016). Thus, examining shareholder identity would provide more and relevant results. The rest of the paper is structured as follows. Section 2 presents banking systems in the MENA region. In Section 3, we review the relevant literature. In Section 4, we describe our variables and methodology. In Section 5, we present and discuss our findings. Section 6 concludes the paper. 64 2 N. Lassoued and H. Sassi Banking systems in MENA countries In most MENA economies (like other emerging economies), financial repression policies had been adopted for many decades. The banking sector, which is the main financing source, was traditionally a highly protected industry living on spreads achieved on regulated deposit and lending rates and pervasive restrictions on foreign entry (Almarzoqi et al., 2015). During 80 years, several important reforms have been undertaken in the region under the supervision of the international monetary fund. Most of MENA governments have undertaken a comprehensive financial reform agenda by nationally deregulating the banking industry and opening the financial markets to foreign competition. As a result of opening economies, many state banks have been privatised; banks have more freedom to expand their activities and entry barriers for foreign investors have been abandoned. As shown in Table 1, banks dominate MENA economies. Indeed, in most countries banking assets reach near 100%. Lebanese banking assets account more than three times the Gross Domestic Product (GDP). The smallest banking sectors are those of Algeria, Oman and Saudi Arabia. Table 1 Some features of banking sectors in MENA countries Bank assets to GDP Bank concentration Stateowned banks Foreignowned banks Limitations on foreign bank entry Noninterest income to total income Restrictions on banking activities Algeria 0.622 0.753 0.930 0.047 4 0.3600 5.5 Bahrain – 0.819 0.024 0.617 4 0.3515 7.33 Egypt 1.070 0.577 0.661 0.131 4 0.4401 7.33 Jordan 2.000 0.918 0.000 0.472 3.5 0.2844 8 Kuwait 0.980 0.818 – – 3 0.3186 5.67 Lebanon 3.290 0.493 0.010 0.345 3.5 0.3155 8 Morocco 1.100 0.711 0.293 0.204 4 0.3296 8 Oman 0.680 0.741 0.000 0.115 2.5 0.2941 8.33 Qatar 1.100 0.850 – 0.149 – 0.3546 3.67 Saudi Arabia 0.710 0.542 0.204 0.201 4 0.3323 8.5 Tunisia 1.090 0.424 0.420 0.157 4 0.3302 8 Turkey 0.940 0.477 0.314 0.089 4 0.3164 7.5 UAE 1.220 0.539 0.350 0.270 3 0.3566 7 Bank assets to GDP, Bank concentration and Non-interest income to total income are extracted from Global Financial Development (2014). State-owned banks, Foreign-owned banks, Limitations on foreign bank entry and Restrictions on banking activities are collected from Barth et al. (2013). State-owned banks column indicates the percentage of bank assets that are 50% or more government-owned during 2006 to 2011 period. Foreign-owned banks column displays the percent of the assets of banks where foreigners own 50% between 2006 and 2011. Income diversification, bank stability and owners identity 65 Limitations on foreign bank entry column indicates whether there are any limitations placed on the ownership of domestic banks by foreign banks and whether there are any limitations placed on the ability of foreign banks to enter the domestic banking industry. Lower values point to greater restrictions on foreign entries. Restrictions on banking activities column measures the number of restrictions imposed by national regulatory authorities on banking activities including securities, insurance and real estate activities plus restrictions on banks owning and controlling non-financial firms. MENA banking is highly concentrated2. Indeed, the Jordanian banking sector is the most concentrated. In fact, the three largest banks own, on average, more than 90% of the total banking assets. The three largest banks in Qatar, Bahrain and Kuwait control more than 80% of banking assets. The Tunisian and Turkish banking sector are the less concentrated. Indeed, the three largest banks own, on average, 42.2% of the total banking assets in Tunisia and 47.7% in Turkey. Column 3 displays percentage of the banking assets that are 50% or more government-owned during 2006 to 2011 period. It seems that the Algerian state monopolises an important part of banking assets in many countries (more than 90%) followed by the Egyptian State, holding more than two-thirds of banking assets. By contrast, Lebanon and Bahraini States own an insignificant part of banking assets while the Jordanian and Omani states do not own any banking assets. As for foreign ownership in column 4, on average, the percent of banking assets indicates banks where foreigners owned 50% between 2006 and 2011. More than 50% of Bahraini banking assets are owned by foreigners. Similarly, about the half of the banking assets in Jordan are owned by foreigners. While, in Algeria foreigners own the lowest banking assets (less than 5%) followed by Turkey where 16.45% are foreigner owned. The remaining MENA nations fall in the middle, allowing only a partial ownership to foreign owners. The growing presence of foreigners in MENA countries resulted in part from abandoning barriers on foreign entries. Indeed, seven countries do not impose any form of foreign entry (column 5) like Algeria, Bahrain, Egypt, Morocco, Saudi Arabia, Tunisia and Turkey, according to the limitations on foreign banking index3. Oman is more stringent on foreign entry and imposes more barriers on foreign banks. However, foreign participation became considerably present in the banking sector in many countries4. The growing presence of foreign banks and the concentration of the banking sector have significantly increased competitive pressures on banks. This has led to deep changes in banking strategies inciting them to diversify their activities. Indeed, average noninterest income represents about the third of total income gained by banks. Egypt has the highest portion of non-interest income (more than 40%) while Jordan has the lowest portion. Policy makers in the MENA region do not promote similar diversification activities. In fact, many of them impose the same restrictions on the extent to which banks may engage in many banking activities5. Saudi banks are the most restricted in terms of their banking diversification. On average, there are more than 8 restrictions (out of 16). Similarly, Lebanon, Morocco and Tunisia impose on average 8 restrictions on diversification activities. However, Qatar restricts less banking activities (3.67). As discussed above, banking sectors play an important role in financing MENA economies. We notice many disparities in term of ownership structure. While many 66 N. Lassoued and H. Sassi countries allow foreign entries, others still restrict them. Similarly, because of these restrictions, income structure varies across these countries. Accordingly, there is a need to investigate the effect of these disparities in term of ownership structure and diversification activities on banking risk. 3 Literature review 3.1 Income diversification and banking risk Modern portfolio theory stipulates that diversification reduces portfolio total risk when assets are less correlated. Therefore, banks should diversify their revenue sources to reduce their risk. Furthermore, Gurbuz et al. (2013) argue that banks can increase shareholders’ value by shifting their focus from traditional income activities to noninterest income activities. Then, banks combine appropriately interest and non-interest income to reduce risk in banking operations. Revenue diversification brings many benefits for banks. First, the dramatic changes in the financial industry, brought about by technological advances and deregulation, have driven banks to build new skills so that they can capitalise on the first mover advantage in choosing diversification activities (Elsas et al., 2010). Second, diversification generates economies of scale for banks that allows improvements in cost/income ratios and efficiency in input sharing (e. g. labour, technology and information) which represent a major source of cost savings. At the same time, wider customer bases are reached reducing thus cost per unit (Stein, 2002). Third, banks have an information advantage about their customers compared to its competitors, thanks to the lending relationship. Indeed, the lending relationship and knowledge about customer needs and the additional requirements for non-interest financial products can be efficient incentives to match customer needs. Fourth, diversification compensates for lower competitive interest margins on traditional deposit/lending markets (Valverde and Fernandez, 2007). Empirically, Sanya and Wolfe (2011a) examine a sample of 226 banks in 11 emerging economies during 2000 and 2007, and they find that diversification activities decrease risk. In contrast to the traditional view, many studies found evidence indicating that diversification is a risk source. Studying a sample of 472 large US commercial banks, DeYoung and Roland (2001) found that non-interest income increases more bank income volatility than traditional interest-based activities do. Similarly, Stiroh (2006) shows that non-interest income is positively correlated with bank income volatility but not with returns. Furthermore, Demirguç-Kunt and Huizinga (2009) examine the effect of banking activity type and short-term funding strategies on banking risk and returns. Examining a sample of 1334 banks in 101 countries, they found that universal banking can be beneficial to diversification with very low risk levels and increasing returns. However, banking strategies exclusively relying on generating non-interest income or attracting non-deposit funding are very risky. Under this perspective, banks must focus on single line of business because diversification costs might outweigh the benefits. Chiorazzo et al. (2008) find that income diversification increases risk-adjusted returns for the Italian banking market between 1993 and 2003. One possible explanation to the harmful effect of income diversification advanced is given by Acharya et al. (2006). The authors argue that there are diseconomies of scale Income diversification, bank stability and owners identity 67 caused by poor monitoring incentives and a poorer loan portfolio quality when a risky bank expands into additional industries and sectors. Moreover, extensive diversification of income sources can reduce banking financial performance (Sahoo and Mishra, 2012). Income diversification requires specialised managerial expertise to manage new risks made possible by these additional activities. Finally, diversification may lead to higher firm-specific risk because of increased complexity and lack of experience in managing the newly diversified areas. These mixed results may be explained by economic, cultural and institutional dimensions of the studied markets or by omitting significant variables like ownership structure, in particular in emerging market known by concentrated ownership. 3.2 Does ownership structure affect the income diversification and banking risk relationship? Controlling shareholders with a diversified portfolio seem to encourage managers to take more risk for a higher expected profit (Shleifer and Vishny, 1986), and they can compensate for losses by diversifying their portfolios. However, when controlling shareholders own undiversified portfolios they will take less risk. Thereby, controlling shareholders do not have the same motivations, objectives, means and control incentives. In this regard, it will be more relevant to investigate the attitude of different types of controlling shareholders. In the present study, we examine state and foreign ownership that dominates the banking sector of the MENA region. Regarding state-controlled banks, government ownership is expected to preserve banks’ financial soundness and enhance good governance. Through its participation in banks, the government achieves its social and political objectives. In the case of stateowned banks, the government finances projects create more jobs especially when its projects could not obtain private financing (La Porta et al., 2002). State ownership is considered as a source of inefficiency because of government bureaucracy and lack of capital market monitoring (Lang and So, 2002). Therefore, managers may pursue their own interests rather than acting in the best interest, which may result in negative effects on performance and high risk level. Many arguments have been advanced to explain excessive risk in state-owned banks. For instance, Demirgüç-Kunt and Detragiache (2002) pointed out that state-owned banks enjoy government protection because bankers may take more risk because losses and excess costs are invariably covered by the government. Faccio et al. (2006) add that the government can protect banks by either implicit or explicit financial and regulatory support. Moreover, Dong et al. (2014) argue that the lending policy of state-owned banks may target social than financial objectives. For example, they finance unprofitable and risky projects, because it has social objectives like those undertaken by state-owned firms. Finally, state-owned banks are essentially controlled by politicians, who may transfer resources to their supporters (Shleifer and Vishny, 1986; Iannota et al., 2013). The last two arguments suggest that state-owned banks might be seen as vehicles for raising capital to finance projects with high social and political returns, but possibly with high-risk and low financial returns. Income diversification in public banks has not been sufficiently investigated, except in the study of Pennathur et al. (2012). The authors conclude that public banks earn significantly less fee-based income, and that government ownership negatively correlates 68 N. Lassoued and H. Sassi with non-interest income sources. They also find that fee-based income significantly reduces risk. Two opposite hypothesis may explain the relationship between risk and income diversification in public banks. According to the first hypothesis, lack of monitoring in these banks give managers the opportunities to pursue their own interests rather than acting in the best interest of owners. In this regard, income diversification is not an important concern for state-controlled banks and managers deploy less effort than their private counterparts to improve efficiency and attract customers to these banks. The undiversified activities of state-owned banks may lead them to be more aggressive risk-takers. However, under the second hypothesis, state-owned banks present high level of risk generated by lending inefficient projects, thus managers may diversify income to hedge risk. As for foreign ownership in emerging markets, it brings about many benefits. Indeed, entry of foreign banks improves human capital, skills and may lead to more diversified products, better use of up-to-date technologies and knowledge transfer (Berger et al., 2005). Foreign participation improves financial services and allows for an easier access to international financial markets (Levine, 1996). Then, these comparative advantages could favour foreign banks in managing operational and financial leverage, when shifting towards non-interest activities. Conversely, foreign banks may also suffer from distance problems, and large banks are disadvantaged in terms of the quality of ‘soft’ information (Stein, 2002). Thus, foreign-owned banks may avoid diversifying their activities, because they have fewer local connections and ineffective incentive schemes for managers to maximise shareholder wealth (Berger et al., 2010). Meslier et al. (2014) find that foreignowned banks in developing countries generally suffer from lack of knowledge of the local market and seem to specialise in non-interest income-activities rather than traditional banking activities. As for the impact of foreign ownership on risk-taking, it is perceived as a stimulator of risk-taking for many reasons. First, foreign owners have the opportunity to diversify better their risk internationally. Second, foreign banks are perceived more efficient and take more risk compared to their domestic counterparts, because they have better access to the capital market and are better able to serve the international clientele that is not easily served by domestic banks (Berger et al., 2005). Third, in emerging countries, they may have more expertise in collecting, evaluating and analysing quantitative information on financial statements and be less exposed to political pressure. Pennathur et al. (2012) study the impact of ownership structure on income diversification-risk relationship in Indian banking sector. The authors add that foreign banks report higher fee income that increases insolvency risk. 4 Empirical design 4.1 Data Our original sample included 202 banks (listed and unlisted bank) operating in 13 MENA countries (Algeria, Saudi Arabia, Bahrain, Egypt, Jordan, Kuwait, Lebanon, Morocco, Oman, Qatar, Tunisia, Turkey and the United Arab Emirates) observed from 2002 to 2012. We apply some filtering rules to ensure data availability and sample homogeneity. Income diversification, bank stability and owners identity 69 First, we exclude Central and cooperative banks. Second, we measure banking risk by standard deviations of ROA (return on assets) and ROE (return on equity) over a moving window of 4 years. By using data from 2002 to 2012, we are able to compute standard deviations of returns for a seven-year period from 2006 to 2012. Thus, the study period is reduced to 2006–2012. Third, we exclude banks with missing data of more than 4 years. In order to accurately calculate the standard deviations of some variables, we defined risk indicators. We ended up with an unbalanced panel of 171 banks distributed over the 2006–2012 period. Bank data are collected from the Bankscope database. Regulation data are obtained from Barth et al (2013) database and macroeconomic data are extracted from the World Bank’s website. 4.2 Measurement of key variables 4.2.1 Diversification proxies Following Mercieca et al. (2007), we construct two Herfindahl-Hirschman Indices. The first (HHIREV) is based on a breakdown of net operating revenue into net interest income (NET) and non-interest income (NON) 2 ⎛ 2 ⎛ ⎞ ⎞ HIIREV = ⎜⎝ NONNETOP ⎟⎠ + ⎜⎝ NET NETOP ⎟⎠ (1) The second index (HHINON) measures diversification into non-interest income-generating activities: Fee and commissions (COM), trading income (TRAD) and other operating income (OTHOP) for each bank in each year. We note that NON = COM + TRAD + OTHOP. ⎛ HII NON = ⎜ COM ⎝ ⎞ NON ⎟⎠ 2 ⎛ + ⎜ TRAD ⎝ ⎞ NON ⎟⎠ 2 ⎛ + ⎜ OTHO ⎝ ⎞ 2 NON ⎟⎠ (2) When HHIREV and HHINON rise, bank income becomes more concentrated and less diversified. 4.2.2 Ownership structure proxies We create two ownership variables, which represent each bank type in our sample: STAT_OWN: the capital stake held by the state (in percentage) FOR_OWN: the capital stake held by foreigners (in percentage) 4.2.3 Banking risk proxies We use two proxies of risk SD_ROA and SD_ROE that consist of the standard deviation of ROA and the standard deviation of ROE computed over a moving window of 4 years. Using data from 2002 to 2012, we were able to compute returns volatility for a sevenyear period, from 2006 to 2012. We also compute a Z-score for bank i at year t indicating bank stability (Boyd et al.,1993) as follows:: 70 N. Lassoued and H. Sassi ZSCORE it = ROA i,( t − 4,t ) + CAR it σ ( ROA )i,( t − 4,t ) (3) With capital-to-asset ratio being the capital asset ratio. 4.3 Model specification Portfolio theory suggests that diversification may influence risk. However, diversification per se might share the same determinants of banking risk, since it is a decision taken by the bank. Ownership structure may also be an endogenous outcome. Indeed, controlling shareholders may affect diversification according to their portfolio diversification degree and the risk aversion level. To address this endogeneity between banking risk, ownership structure and the decision to diversify, we construct a 3SLS simultaneous equation model6 described by two equations shown in Equations 4 and 5. Diversification, risk and ownership are treated as endogenous variables. Risk it = β 0 + β1Ownnershipit + β 2 Diversification it + β 3 Control variablesit + Year Dummies + µit Diversification it = α 0 + α1Ownnershipit + α 2 Risk it + α 3 Control variables it + Year Dummies + ϑit (4) (5) Equation (4) has either one of the banking risk measures (SD_ROA, SD_ROE and ZSCORE) as the dependent variable. The model is estimated separately for the two ownership structure variables STAT_OWN and FOREIGN_OWN. Two revenue diversification measures are used (H_REV and H_NON). The model examines the impact of ownership and diversification on banking risk. We include a set of bank-and country-specific variables in our analysis to control the sources of risk taking differences. Consistent with previous studies (Mercieca et al., 2007; Deng et al., 2013; Lassoued et al., 2016), we include the control variables of bank profitability, measured by ROA, capital-to-asset ratio (CAR) computed by the ratio of equity book value to total assets and SIZE (natural logarithm of total assets). We expect that ROA is negatively related to bank risk, since performance reflects management quality. Then, poor management practices, in unprofitable banks, lead to an increase of non-performing loans and then of bank risk. Moreover, low capital-to-assets ratio indicates that banking risk is high, because capital protects banks when asset values decline, on the other hand, higher capital ratio prevents bank failure (Deng et al., 2013). The effect of bank size is expected to be positive under the “too big to fail” hypothesis suggesting that larger banks tend to take more excessive risks. As country-level control variables, we include GDP Growth (GDP_GR) and we expect that the estimated coefficient is negative. Indeed, during periods of economic expansion, borrowers need sufficient funds to service their debts, but during a recession, the ability to service debt declines. Therefore banking risk will increase in recession periods. Finally, we include deposit insurance (DEP_INS) as a dummy variable that takes 1 or 0 indicating whether the country has explicit DEP_INS or not (yes = 1; no = 0). Deposit insurance can limit banking risk. Indeed, a DEP_INS system reduces bank risktaking incentives (Gropp and Vesala, 2001). However, bank managers may be encouraged to take more risks in order to generate higher profits, and insurance will Income diversification, bank stability and owners identity 71 cover a large part of the bank’s debts in case of non-payment (Angkinand and Wihlborg, 2010). Equation (5) has either one of the income diversification measures (H_REV and H_NON) as its dependent variable. We test the effect of ownership and banking risk on income diversification. We include a set of control variable as factors that may affect banking risk. Following Campa and Kedia (2002) and Deng et al. (2013), we include profitability, measured by ROA, bank-size (SIZE) and CAR as control variables in the diversification model (Equation 5). We expect that profitability is inversely related to diversification, because poorly performing banks seek to diversify to improve their income (Campa and Kedia, 2002). Furthermore, CAR is expected to be negatively related because, wellcapitalised banks will take less risk and may therefore be less diversified (Stiroh and Rumble, 2006). For SIZE indicators, we expect that the estimated coefficients will be positive, because large banks have a greater capacity to diversify their income since they have more opportunities to pursue a wider variety of loans and other activities (Sullivan and Spong, 2007). Finally, we include GDP Growth (GDP_GR) to control for the effect of economic growth on diversification. Indeed, Sanya and Wolfe (2011b) argue that it is more profitable to diversify during periods of rapid economic growth. We include three instruments in our model. The first instrument is the restrictions on banking activities index, which indicates whether policy makers impose the same restrictions on banking activities. The index, taken from Barth et al. (2013) influences directly income diversification activities, since it describes the extent to which banks may engage in many activities7 The second and third instruments are related to ownership variables and refer to regulatory quality since a country’s institutional environment is exogenous and closely related to ownership structure (La Porta et al., 1999) and might be considered as an external governance influence that acts at the banking industry level (Ciancanelli and Reyes, 2001; Lassoued et al., 2016). In this regard, we use the regulatory part of The Worldwide Governance Indicator developed by Kaufmann et al. (2010), which refers to the capacity of the government to effectively formulate and implement sound policies. The first index is Government Effectiveness that captures perceptions of public services quality, civil services quality and their degree of independence from political pressure, policy formulation and implementation quality and the credibility of the government’s commitment to such policies. The second is Regulatory Quality, which captures perceptions of the ability of the government to conceive and implement sound policies and regulations that encourage private sector development. 5 Results and discussion 5.1 Descriptive statistics Table 2 reports the descriptive statistics for the whole sample of 171 banks in the studied MENA countries. Returns volatility measured by SD_ROAT is on average 0.006 (SD_ROET 0.075) with a great concentration below the mean (since the median does not exceed the mean). In contrast to insolvency risk, the median of the ZSCORE is located below the mean (59.02) indicating an important risk level. 72 N. Lassoued and H. Sassi Table 2 Descriptive statistics Variable Mean Median Std. Dev. Min Max SD_ROAT 0.006 0.003 0.011 0.0000 0.14 SD_ROET 0.075 0.033 0.191 0.0005 2.029 ZSCORE 59.02 38 67.28 0.02 591.25 H_REV 0.596 0.569 0.107 0 1 H_NON 0.483 0.5 0.207 0 1 STAT_OWN 0.144 0.012 0.289 0 1 FOR_OWN 0.333 0.041 0.383 0 1 ROA 0.015 0.013 0.015 –0.114 0.130 CAR 0.107 0.128 0.072 0 0.997 GDP_GR 5.257 4.802 3.902 –0.070 0.261 SIZE 9.665 7.834 0.770 5.760 11.04 REST_DIV 7 7.500 1.586 3 9 REG_REQ 0.681 0.707 0.1521 0.454 0.954 GOV_EFF 0.5 0.521 0.0701 0.5 0.75 The mean of the diversification index H_REV is 0.596 indicating a relative concentration of bank income into interest-generating activities. Moreover, banks seem to diversify within the range of non-interest income activities, since the mean and the median of H_NON are, respectively, 0.483 and 0.500. For ownership structure, states own on average 14.4% of bank capital and foreigners about 33.3%. It seems that this type of owners is significantly present since ownership shares exceed 0.012 for state ownership and 0.041 for foreign owners. 5.2 Basic model Table 3 reports the results of our model using state ownership with two panels. In panel A, diversification is measured by H_REV and in panel B by H_NON. The results of risk estimation (SD_ROA, SD_ROE and ZSCORE) are displayed in columns (1, 2 and 3) and those of diversification in columns (4, 5 and 6). Across all model specifications, the null hypothesis (suggesting that Ordinary least square (OLS) is an appropriate estimation technique) is rejected through Hausman test of endogeneity. We validate that the relevance of estimating our instrumental variables. Furthermore, the over-identifying Hansen-Sargan test does not reject the null hypothesis and supports the validity of instruments. 73 Income diversification, bank stability and owners identity Table 3 Three-stage least squares regression (3SLS) regression results of Bank risk, diversification and state ownership Panel A H_REV STAT_OWN SD_ROAT SD_ROET ZSCORE 0.010 (2.50)** 0.031 (2.03)** 0.545 (3.49)*** 0.998 (2.16)** –9.175 (–2.19)** –200.764 (–2.72)*** SIZE CAR GDP_GR DEP_INS H_REV H_REV 0.246 (3.71)*** 0.620 (22.99)*** –0.382 0.236 (3.10)*** 0.637 (23.70)*** –0.481 –0.126 –5.478 618.126 0.245 (3.71)*** 0.618 (22.99)*** –0.383 (–3.64)*** (–4.78)*** (2.36)** (–1.47) (–1.47) (–1.82)* –0.0001 –.0022 –1.556 0.0005 0.0005 0.0006 (–0.83) (–0.25) (–0.47) (0.33) (0.33) (0.40) –0.001 –0.631 135.759 0.050 0.050 0.043 (–0.12) (–2.01)** (1.88)* (2.15)** (2.21)** (2.94)*** L.H_REV ROA H_REV –0.0004 –0.007 3.109 –0.001 –0.001 –0.001 (–2.72)*** (–1.47) (2.74)*** (1.06) (–1.06) (–0.94) (546.84)*** (546.84)*** (581.6)*** 0.003 0.094 –17.775 (2.29)** (2.09)** (–1.87)* Chi2 (23.67)*** (33.71)*** (17.62)*** Hausman test (16.21)** (13.47)** (14.3)** (2.788) (2.271) (3.703) (2.788) (2.271) (3.703) SD_ROAT SD_ROE ZSCORE H_NON H_NON H_NON Hansen-Sargan statistic Panel B H_NON STAT_OWN 0.005 0.261 –34.255 (2.04)** (3.24)*** (–2.39)** 0.006 0.347 –90.655 0.281 0.280 0.279 (1.80)* (1.94)* (–2.18)** (2.87)*** (2.85)*** (2.96)*** L.H_NON ROA SIZE CAR GDP_GR DEP_INS 0.563 0.562 (13.93)*** (13.47)*** –0.106 –3.513 490.683 –1.226 –1.228 –0.942 (–5.35)*** (–5.15)*** (2.26)** (–1.83)* (–1.83)* (–2.11)** –0.0001 –0.0027 3.580 –0.010 –0.010 –0.0107 (–0.85) (–0.47) (1.07) (–2.48)** (–2.45)** (–2.49)** –0.015 –0.421 –33.837 0.254 0.255 0.274 (–3.25)*** (–2.64)*** (–0.72) (2.07)** (2.07)** (2.05)** –0.0001 –0.002 1.744 –0.003 –0.003 –0.003 (–2.77)*** (–0.64) (2.45)** (–1.19) (–1.19) (–1.24) (194.08)*** (195.32)*** (183.11)*** 0.0001 0.060 1.609 (0.21) (2.16)** (0.21) Chi2 (64.66)*** (56.03)*** (226.76)*** Hausman test (17.01)*** (12.31)* (13.74)** (8.024) (5.797) (7.096) Hansen-Sargan statistic 0.561 (13.89)*** 74 N. Lassoued and H. Sassi In panel A, the coefficients for H_REV and H_NON are positive in SD_ROA and SD_ROE regression (negative for the ZSCORE regression) and statistically significant across all specifications. This implies that diversification into and within non-interest income-generating activities correlates with lower income volatility and lower insolvency risk8. More specifically, we found that risk declines as the bank diversifies across interest and non-interest activities and diversifies within non-traditional banking activities (like trading, brokerage…). Our finding is in line with the modern portfolio theory suggesting that diversification decreases risk. This result for the MENA banks is consistent with those of many studies on different samples .For instance, Sanya and Wolfe (2011a) and Baele et al. (2007) show that income diversification reduces risk in a sample of European banks. Deng et al. (2013) found the same for US banks. However, Stiroh (2006) find the opposite relationship in a sample of US banks. In the same regression (Equation 4), the coefficient of state ownership on risk is positive (negative) and significant at the 1% level in the two first (third) columns in panel A. Panel B presents the same result but less significant. These findings imply that governments in the MENA countries encourage banks to take more risk. To see whether income diversification and banking risk can be determined through the impact of state ownership on the decision to diversify income, we report the results of Equation (5) in columns (5–6) of Table 3. Panel A reports the results of H_REV as diversification proxy. The coefficients of state ownership are positive and significant (at the 5% and 1% levels) and imply that governments seem to discourage income diversification, thus generating high risk level. The finding is very interesting and explains previous studies showing that state ownership increases banking risk like Farazi et al. (2013) and Lassoued et al. (2016) for MENA countries, Berger et al. (2005) for Argentinean banks and Angkinand and Wihlborg (2010) for a sample of banks selected from 32 countries. In fact, reluctance to diversify income (either across interest and noninterest activities or within non-traditional banking activities) for state-owned banks generates high risk level. Among the control variables in the risk equation, we found that banking performance ROA is negatively associated with SD_ROA and SD_ROE risk and positively associated to ZSCORE (in panels A and B) suggesting that performing banks are less risky and more stable, consistent with the findings of Sanya and Wolfe (2011b). The coefficient of CAR is negative and significant in column 2 and positive in column 3 for Panel A. Similarly, the coefficient of CAR is negative and significant in column 1 and positive and significant in column 3 for Panel B. In line with the findings of Mercieca et al. (2007) and Meslier et al., 2014), our results indicate that when equity levels are low, banking risk is high because capital protects banks when asset values decline. GDP_GR has a negative and a significant effect on SD_ROA and a positive and a significant effect on ZSCORE (Panels A and B) indicating that banks operating more growing economies are less risky than other banks. For DEP_INS, we found that it significantly increases risk in all specifications of Panel A. The effect is significant only in column 2 of Panel B. Against the conventional wisdom that DEP_INS reduces banking risk Gropp and Vesala, 2001), it seems to give bank managers the incentive to take more risks, since insurance will cover a large part of the bank’s debts in case of non-payment (Angkinand and Wihlborg, 2010). As for the control variables in the diversification equation presented in columns (4, 5 and 6), lagged diversification exhibits positive and significant coefficients in all Income diversification, bank stability and owners identity 75 specifications for Panels A and B. This is consistent with the findings of Deng et al. (2013). ROA has a negative significant coefficient in column 6 (panel A) and columns 4, 5 and 6 (Panel B). This finding is consistent with that of Campa and Kedia (2002) who argue that poorly performing (less profitable) industrial firms tend to diversify. Finally, CAR coefficients are positive and significant in all specifications, indicating that wellcapitalised banks are less diversified. Stiroh and Rumble (2006) found that banks with high capitalisation will take less risk and may therefore be less diversified (Stiroh and Rumble, 2006). Table 4 presents results our model for foreign ownership with two panels. In panel A, diversification is measured by H_REV and in panel B by H_NON. The results of risk estimation (SD_ROA, SD_ROE and ZSCORE) are displayed in columns (1, 2 and 3) and those of diversification in columns (4, 5 and 6). Across all model specifications, we report similar statistics for the Hausman test of endogeneity and the over-identifying Hansen-Sargan test. According to these tests, we support the relevance of estimating the instrumental variables and the validity of the instruments. The coefficients for H_REV in panel A and for H_NON are positive in SD_ROA and SD_ROE regression (negative for the ZSCORE regression). They are statistically significant across all specifications. Once again, we found that diversification into and within non-interest income-generating activities is associated with lower income volatility and lower insolvency risk. In the same regression (Equation 4), the coefficient of foreign ownership on risk is negative (positive) and significant at the 1% level (at 5 % level) in the two first (third) columns in panel A. Panel B reports the same results, indicating that foreign ownership decreases banking risk in MENA countries. Like above, we investigate whether income diversification and banking risk can be determined through the impact of the foreign ownership on the decision to diversify income. We present the result of Equation (5) in columns (5–6) of Table 3. Panel A displays the results of using H_REV as a diversification proxy. The coefficients of foreign ownership are negative and significant at the 1% level in all specifications. In light of this finding, foreign participants seem to encourage income diversification, which contributes to reducing banking risk. Ours results are consistent with those of Crystal et al. (2002), Mian (2003) and Lassoued et al. (2016) which found that foreign banks decrease banking risk in emerging markets. This finding can be explained by the importance of income diversification (either across interest and non-interest activities or within non-traditional banking activities) for foreign-owned banks. Finally, we found the same results for the control variables in the risk and diversification equations. Indeed, ROA is negatively associated with SD_ROA and SD_ROE risk and positively associated to ZSCORE (in panels A and B). CAR has a negative effect on banking risk. GDP_GR seems to reduce risk while DEP_INS significantly increases risk. Regarding the control variables in the diversification equation presented in columns (4, 5 and 6), lagged diversification displays positive and significant coefficients in all specifications and CAR coefficients are positive and significant in all specifications indicating that well-capitalised banks are less diversified. 76 N. Lassoued and H. Sassi Table 4 Three-stage least squares regression (3SLS) regression results of bank risk, diversification and foreign ownership SD_ROA SD_ROE ZSCORE H_REV H_REV H_REV PANEL A H_REV FOR_OWN 0.002 0.297 –52.848 (1.97)** (3.16)*** (–2.27)** –0.005 –0.143 17.879 0.235 0.236 0.227 (–3.65)*** (–2.92)*** (2.45)** (3.43)*** (3.44)*** (2.83)*** L.H_REV ROA SIZE CAR GDP_GR DEP_INS 0.613 0.633 (23.65)*** (24.34)*** –0.091 –4.424 577.943 –0.257 –0.257 –0.401 (–5.47)*** (–7.76)*** (3.73)*** (–1.09) (–1.09) (–1.63)* –0.003 –0.021 1.651 0.0007 0.0007 0.0025 (–0.95) (–1.38) (0.46) (0.12) (0.12) (0.44) –0.014 –0.156 33.465 0.010 0.010 0.009 (–5.57)*** (–1.89)** (1.60) (2.51)** (2.51)** (2.42)** –0.0001 –0.001 1.492 –0.0004 –0.0004 –0.0002 (–3.39)*** (–0.51) (3.76)*** (–0.57) (–0.57) (–0.52) 0.001 0.028 –5.462 (3.04)*** (2.26)** (–1.74)* Chi2 (94.40)*** (119.64)*** (45.64)*** Hausman test (15.67)** (12.98)** (13.25)** (7.949) (10.095) (13.771) SD_ROA SD_ROE ZSCORE Hansen-Sargan statistic 0.613 (23.66)*** (565.1)*** H_NON (564.81)*** (600.68)*** H_NON H_NON PANEL B H_NON FOR_OWN –0.006 –0.180 –46.243 (–2.60)*** (–3.00)*** (–2.26)** –0.001 –0.069 6.761 –0.173 –0.173 –0.175 (–2.66)*** (–2.32)** (2.39)** (–3.14)*** (–3.14)*** (–3.16)*** L.H_NON ROA SIZE CAR GDP_GR DEP_INS 0.532 0.527 (14.00)*** (13.72)*** –0.116 –3.092 471.653 –0.234 –0.235 –0.143 (–5.67)** (–5.36)** (2.13)** (–0.37) (–0.37) (–0.19) –0.0005 –0.0087 3.709 –0.0128 –0.0128 –0.0127 (–1.08) (–0.71) 0.84 (–1.02) (–1.02) (–1.07) –0.013 –0.283 3.009 0.183 0.188 0.191 (–4.57)** (–3.45)** (0.11) (2.73)*** (2.71)*** (2.48)** –0.0002 0.0002 2.293 –0.0002 –0.0004 –0.0003 (–4.11)*** (0.26) (4.27)*** (–0.14) (–0.14) (–0.17) 0.001 0.031 –4.415 (1.21) (2.43)** (–1.01) Chi2 (68.24)*** (84.79)*** (29.4)*** Hausman test (14.93)** (11.28)** (12.51)** (6.927) (2.873) (8.572) Hansen-Sargan statistic 0.531 (13.97)*** (226.14)*** (226.82)*** (218.73)*** Income diversification, bank stability and owners identity 6 77 Conclusion The aim of this paper is to examine how bank ownership structure influences the relationship between income diversification and banking risk in the MENA countries. We used data on171 banks from 13 MENA countries observed over the 2006–2012 period. We specified a 3SLS model in which the impact of foreign and state ownership on income diversification and banking risk is jointly examined to solve the endogeneity problem that arises from the relationship between risk-taking and diversification. We found that diversification, across interest and non-interest activities and within non-traditional banking activities, decreases banking risk. Owner identities affect risk level through diversification. Indeed, state-owned banks seem to take more risks by less diversifying their income. Inversely, foreign ownership decreases risk by diversifying their income. Therefore, we confirm that diversification is a channel through which controlling shareholders influence banking risk and diversification level depends on the monitoring exerted by the different shareholders. Our findings have several implications for regulators and investors in the MENA region. Policy makers should, first, encourage banks to diversify more their income by removing restrictions on banking activities. Second, regulators should accelerate the banking privatisation process. Third, it will be more appropriate to alleviate barriers to foreign investment to decrease bank risk-taking. Investors may use ownership structure and income diversification in their decisions. In conclusion, our study highlighted that ownership structure affects bank risk-taking through income diversification. Future research could provide additional insights by jointly examining the effect of other types of diversification. References Acharya, V.V, Hasan, I. and Saunders, A. (2006) ‘Should banks be diversified? evidence from individual bank loan portfolios’, Journal of Business, Vol. 79, No. 3, pp.1355–1411. Almarzoqi, R., Ben Naceur, S. and Scopelliti, A. (2015) How Does Bank Competition Affect Solvency, Liquidity and Credit Risk? Evidence from the MENA Countries, IMF Working Paper Series Nos. 15/210 Angkinand, A. and Wihlborg, C. (2010) ‘Deposit insurance coverage, ownership, and banks’ risktaking in emerging markets’, Journal of International Money and Finance, Vol. 29, No 2, pp.252–274. Baele, L., De Jonghe, O. and Vennet, R.V. (2007) ‘Does the stock market value bank diversification?’, Journal of Banking and Finance, Vol. 31, No. 7. pp.1999–2023. Barth, J.R., Caprio, J. and Levine, R. (2013) Bank Regulation and Supervision in 180 Countries from 1999 to 2011, National Bureau of Economic Research Working Paper 18733. Ben Naceur, S., Ghazouani, S. and Omran, M. (2007) ‘The performance of newly privatized firms in selected MENA countries: the role of ownership structure, governance and liberalization policies’, International Review of Financial Analysis, Vol. 16, No. 4, pp.332–353. Berger, A.N., Clarke, G.R., Cull, R., Klapper, L. and Udell, G. (2005) ‘Corporate governance and bank performance: a joint analysis of the static, selection, and dynamic effects of domestic, foreign, and state ownership’, Journal of Banking and Finance, Vol. 29, No. 8, pp.2179–2221. Berger, A.N., Hasan, I. and Zhou, M.M. (2010) ‘The effects of focus versus diversification on bank performance: evidence from Chinese banks’, Journal of Banking and Finance, Vol. 34, pp.1417–1435. 78 N. Lassoued and H. Sassi Boyd, J.H., Graham, S. and Hewitt, S. (1993) ‘Bank holding company mergers with nonbank financial firms on the risk of failure’, Journal of Banking and Finance, Vol. 17, No. 1, pp.43–63. Busch, R. and Kick, T. (2009) Income Diversification in the German banking Industry. Deutsche Bundes bank, Discussion Paper Series 2: Banking and Financial Studies No 09/2009. Campa, J.M. and Kedia S. (2002) ‘Explaining the diversification discount’, The Journal of Finance, Vol. 57, pp.1731–1762. Chiorazzo, V., Milani, C. and Salvini, F. (2008) ‘Income diversification and bank performance: evidence from Italian banks’, Journal of Financial Services Research, Vol. 33, No. 3, pp.181–203. Ciancanelli, P. and Reyes, J. (2001) Corporate Governance in Banking: A Conceptual Framework, Working Paper SSRN, Available at: http://ssrn.com/abstract=253714. Crystal, J.S., Dages, B.G. and Goldberg, L.S. (2002) ‘Has foreign bank entry led to sounder banks in Latin America?’, Current Issues in Economics and Finance, Vol. 8, No. 1, pp.1–6. Demirgüç-Kunt, A. and Detragiache, E. (2002) ‘Does deposit insurance increase banking system stability? An empirical investigation’, Journal of Monetary Economics, Vol. 49, No. 7, pp.1373–1406. Demirguç-Kunt, A. and Huizinga, H. (2009) ‘Bank activity and funding strategies: the impact on risk and returns’, Journal of Financial Economics, Vol. 98, No. 3, pp.626–650. Deng, S., Elyasiani, E. and Jia, J. (2013) ‘Institutional ownership, diversification, and riskiness of bank holding companies’, Financial Review, Vol. 48, No. 3, pp.385–415. DeYoung, R. and Roland, K.P. (2001) ‘Product mix and earnings volatility at commercial banks: evidence from a degree of leverage model’, Journal of Financial Intermediation, Vol. 10 No. 1, pp.54–84. Dong, Y., Meng, C., Firth, M. and Hou, W. (2014) ‘Ownership structure and risk-taking: comparative evidence from private and state-controlled banks in China’, International Review of Financial Analysis, Vol. 36, pp.120–130. Elsas, R., Hacketha, A. and Holzhäuser, M. (2010) ‘The anatomy of bank diversification’, Journal of Banking and Finance, Vol. 34, No. 6, pp.1274–1287. Faccio, M., Masulis, R.W. and McConnell, J.J. (2006) ‘Political connections and corporate bailouts’, Journal of Finance, Vol. 61, No. 6, pp.2597–2635. Farazi, S., Feyen, E. and Rocha, R. (2013) ‘Bank ownership and performance in the middle east and North Africa Region’, Review of Middle East Economics and Finance, Vol. 9, No. 2, pp.159–196. Global Findex Database (2014) World Bank. Gropp, R. and Vesala, J. (2001) Deposit Insurance and Moral Hazard: Does the Counterfactual Matter?, European Central bank Working Paper 41. Gurbuz, A.O., Yanik, S. and Ayturk, Y. (2013) ‘Income diversification and bank performance: evidence from Turkish banking sector’, Journal of BRSA Banking and Financial Markets, Vol. 7, No. 1, pp.9–29. Iannota, G., Giacomo, N. and Sironi, A. (2013) ‘The impact of government ownership on bank risk’, Journal of Financial Intermediation, Vol. 22, No. 2, pp.152–176. Jensen, M. and Meckling, W. (1976) ‘Theory of the firm: managerial behavior, agency costs and ownership structure’, Journal of Financial Economics, Vol. 3, No. 4, pp.305–360. Kaufmann, D., Kraay, A. and Mastruzzi, M. (2010) The Worldwide Governance Indicator: A Summary of Methodology, data, and Analytical Issues, WorldBank Policy Research Working Paper No. 5430. Keeley, M.C. (1990) ‘Deposit insurance, risk and market power in banking’, American Economic Review, Vol. 80, No. 5, pp.1183–1200. La Porta, R., Florencio L., Andrei, S. and Robert, W. (2002) ‘Government ownership of banks’, The Journal of Finance, Vol. 57, No. 1, pp.256–301 Income diversification, bank stability and owners identity 79 La Porta, R., Lopez-de-Silanes, F. and Shleifer, A. (1999) ‘Corporate ownership around the world’, Journal of Finance, Vol. 54, No. 2, pp.471–517. Lang, L. and So, R. (2002) Bank Ownership Structure and Economic Performance, SSRN Working Paper. LaPorta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R.W. (1998) ‘Law and finance’, The Journal of Political Economy, Vol. 106, No. 6, pp.1113–1155. Lassoued, N., Sassi, H. and Ben Rejeb-Attia, M. (2016) ‘The impact of state and foreign ownership on banking risk: Evidence from the MENA countries’, Research in International Business and Finance, Vol. 36, No. 1, pp.167–178. Lepetit, L., Nys, E., Rous, P. and Tarazi, A. (2008) ‘Bank income structure and risk: an empirical analysis of European banks’, Journal of Banking and Finance, Vol. 32, No. 9, pp.1452–1467. Levine, R. (1996) ‘Foreign banks, financial development and economic growth’, in Barfield, C.E. (Ed.): International Financial Markets: Harmonization Versus Competition, Washington, DC, A EI Press. Mercieca, S., Schaeck K. and Wolfe, S. (2007) ‘Small European banks: benefits from diversification?’ Journal of Banking and Finance, Vol. 31, No. 7, pp.1975–1998. Meslier, C., Tacneng, R.C. and Tarazi, A. (2014) ‘Is bank income diversification beneficial? evidence from an emerging economy’, Journal of International Financial Markets, Institutions and Money, Vol. 31 No. 3, pp.97–126. Mian, A. (2003) Foreign, Private Domestic, and Government Banks: New Evidence from Emerging Markets, University of Chicago, Mimeo. Pennathur, A.K., Subrahmanyam, V. and Vishwasrao, S. (2012) ‘Income diversification and risk: Does ownership matter? An empirical examination of Indian banks’, Journal of Banking and Finance, Vol. 36, No. 8, pp.2203–2215. Rocha, R., Arvai, Z. and Farazi, S. (2011) Financial Access and Stability: A Road Map for the Middle East and North Africa, World Bank, Washington, DC. Sahoo, D. and Mishra, P. (2012) ‘Operational diversification and stability of financial performance in indian banking sector: a panel data investigation’, Research Journal of Finance and Accounting, Vol. 3, No. 3, pp.70–87. Sanya, S. and Wolfe, S. (2011a) ‘Can banks in emerging economies benefit from revenue diversification?’, Journal of Financial Services Research, Vol. 40, pp.79–101. Sanya, S. and Wolfe, S. (2011b) Ownership Structure, Revenue Diversification and Insolvency Risk in European Banks Amiable at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1787476. Shleifer, A. and Vishny, R. (1986) ‘Large shareholders and corporate control’, The Journal of Political Economy, Vol. 94, No. 3, pp.461–488. Stein, J.C. (2002) ‘Information production and capital allocation: decentralized versus hierarchical firms’, Journal of Finance, Vol. 57, No. 5, pp.1891–1921. Stiroh, K.J. (2006) ‘A portfolio view of banking with interest and noninterest activities’, Journal of Money Credit and Banking, Vol. 38, No. 5, pp.1351–1361. Stiroh, K.J. and Rumble, A. (2006) ‘The dark side of diversification: the case of US financial holding companies’, Journal of Banking and Finance, Vol. 30, No. 8, pp.2131–2161. Sullivan, R.J. and Spong (2007) ‘Manager wealth concentration, ownership structure, and risk in commercial banks, Journal of Financial Intermediation, Vol. 16, pp.229–248. Valverde, S. and Fernandez, F. (2007) ‘The determinants of bank margins in European banking’, Journal of Banking and Finance, Vol. 31, No. 7, pp.2043–2063. Zhou, K. (2014) ‘The effect of income diversification on bank risk: evidence from China’, Emerging Markets Finance and Trade, Vol. 50, p.201–213. 80 N. Lassoued and H. Sassi Notes 1 2 3 4 5 6 7 8 DeYoung and Roland (2001); Stiroh et al. (2006); Lepetit et al. 2008), Stiroh and Rumble (2006). The assets of the 3 largest banks to the total assets of the banking sector. The index indicates whether foreign banks are prohibited from entering through subsidiary acquisition and joint ventures. Lower values point to greater restrictions on foreign entries. According to Farazi et al. (2013), assets of foreign banks increased from 18% of total banking assets in 2001 to 20% in 2008. Like underwriting, brokering and securities trading and all aspects of the mutual fund industry in terms of insurance underwriting and selling in real estate investment, development and management. Our empirical approach is widely inspired from Sanya and Wolfe (2011b). like underwriting, brokering and dealing in securities… and all aspects of the mutual fund industry in insurance underwriting and selling in real estate investment, development and management. Lower HHI indices show increased diversification.