Keywords

1 Introduction

Following the recent financial crisis, there is a widespread desire to create a stable financial system in general and to ensure the stability of the banking system in particular.

Having a stable and efficient banking system contributes to many economic benefits and reduces the possible negative consequences of increased capital mobility.

Thus, an adequate banking structure can attest to banks effectiveness and capacity enabling them to overcome the risks to which they are prone in order to avoid distress situations.

During recent years, concerns on competition evolution in the banking industry and the possible trade-offs between the efficiency and stability of the system, are often at the centre of banking sector policy debate. The ongoing crisis leads to re-consider the links of changes in the competitive environment of banks and the stability issues.

This issue raises important challenges for the MENA region where the banking system is the main conduit for economic activities. Moreover, the last decade was characterized by substantial reforms to restructure the banking system, increase the privatization process and foreign bank penetration.

Indeed, financial reforms adopted by the MENA countries involved abolition of control regulations which affect the behaviour of the banks (such as the abolition of controls on the debtor and credit rates, fixing of quotas) and ending regulations of structures which condition the organization of banking industry directly (functional separation, entries barriers). As a consequence, three radical changes of the banking structures in all their components took place: change of property structures by a movement of privatization, foreign banks penetration, and the intensity of competition (the deregulation made possible the appearance of new actors which created a new exacerbated competition).

So these changes of structure in all their dimensions raise many interrogations for the maintenance of financial stability especially that several banks encountered various difficulties to adapt to this new context and various followed strategies did not save many banks from significant situations of distress following a considerable rise in risks.

Thus, the aim of this paper is to study the impact of financial integration and banking competition changes in MENA countries on the likelihood of suffering a systemic banking crisis and to seek the optimal structure of the banking system.

The remainder of this paper is organized as follows: Sect. 10.2 presents related theoretical and empirical literature on the relationship between banking market concentration and financial stability. A detailed exposition of the empirical methodology is presented in Sect. 10.3. We report the results and a variety of robustness tests in Sect. 10.4. Some conclusions and policy implications are offered in the final section.

2 Literature Review

The relation between financial liberalization and banking stability is the subject of several theoretical and empirical studies that try mostly to identify the effect of liberalization policies on banking stability.

The literature addresses the macroeconomic and microeconomic approaches of banking failure. ResearchersFootnote 1 use mostly empirical approaches to study the macroeconomic sequence leading to banking instability. A clearly achieved result of these approaches is that the inappropriate institutional preconditions of the financial liberalization process constitute the main cause of crises and economic recessions.

The microeconomic aspect insists on banks behaviour as the origin of financial disturbances. In this respect, two major explanatory visions are opposed. The first is of neo-classic inspiration, where banking instability corresponds to exceptional episodes caused by external factors such as excess of public policies as sources of moral risk. According to this explanatory diagram and in spite of the frequency of banking crises in developing countries, the failure is not directly ascribed to the financial liberalization process, but to internal countries distortions. The second vision which is a post-Keynesian one supposes that the behaviour of banks and credit market imperfections are the main factors of dysfunctions.

This approach gives a central role to speculative behaviours resulting from the environment change created by financial liberalization.

Admittedly, there is no theoretical consensus on financial liberalization effect on banking stability. For the liberal dogma, the main problem is not due to the liberalization process, but rather in monitoring gaps and prudential regulation whose consequences are simply amplified by the financial opening.

However, critical approaches (post-Keynesian and neo-structuralist) while criticizing the liberal theoretical justifications, warn against the negative effects of financial liberalization policies and the failure of many reforms to preserve banking stability.

We can conclude that the majority of these contributions on the causes of banking instability did not address the strategic interaction between banks and various banking structures on banking stability.

Recently, many studies (Demirguç-Kunt and Detragiache 2002; Demirguç-Kunt et al. 2004; Claessens and Laeven 2004; Bikker 2004; Bikker and Spierdijk 2007) have focused their analysis on the evolution of competition intensity and the banking market structure characteristics based in the New Empirical Industrial Organization (NEIO). These models, which include those of Bresnahan (1982) and Panzar and Rosse (P-R) (1982, 1987), do not rely on explicit information about market structure in order to determine the level of competition and provides a very simple approach to test the market structure of an industry for competitiveness.

Some of those studies (Barth et al. 2004; Beck et al. 2006a, b; Boyd et al. 2006; Schaeck et al. 2006) raise a controversial question in the context of the on-going financial crisis to examine if the banking stability is enhanced or weakened by the changes in banking structures (ownership and market structure) and the changes in the competition intensity.

The theoretical and empirical results reached show a considerable uncertainty concerning the relationship between competition and financial stability (Carletti 2010).Footnote 2

The synthesis of the theories shows a controversy between theories which advocate for concentration as guarantees of stability and those which consider it as a source of fragility.

Thus, in this respect two opposing views arise. The competition- fragility hypotheses (Keeley (1990), Allen and Gale (2000, 2004), Beck and Laeven (2008), Jimenez et al. (2007), Beck (2008), Berger et al. (2008)) where more concentrated and less competitive banking systems are more stable. Those theoretical and empirical studies suggest that a less concentrated banking system is more responsible for the financial crises. Indeed, banks with a high capitalization can better face unfavorable shocks. The market power and profits which result are regarded as guarantees of stability. This market power confers to banking institution revenue which gives an additional value constituting an additional capital, which decreases the excessive risk-taking improving the quality of loans. However the competition intensification induced by financial liberalization, diminished the dominant position of banks. In short, the partisans of “concentration-stability” predict that large banks can better diversify and face unfavourable shocks.

The increase in banks sizes can be regarded as a force and constitute a generating factor of profit. But it can also constitute a multiplying factor of risk according to the partisans of “concentration fragility”.

Whereas, for the new “competition-stability” view (Caminal and Matutes 2002; Boyd and de Nicola 2005; Boyd et al. 2006), increased competition may enhance bank stability, and may reduce significant implications for stressed banking systems in developing economies.

According to this view, several banks generally choose strategies of fusion to face competition. This process of concentration can cause the “too-big-to fail” policy as these banks will systematically be the subject of rescue in the event of failures. This intervention can exacerbate moral hazard and the risk taking of banks and thus increase the whole system’s vulnerability. Moreover, the emergence of the financial conglomerates generates also possible moral hazard by the complication of their monitoring.

In the following paper, our aim is to investigate how financial liberalization, banking structures change, MENA country’s institutions and business environment affect banking stability.

3 Methodology

3.1 The Data Set

Our sample represents panel data on commercial banks in the MENAFootnote 3 region for the period 1990–2009. The time interval under examination corresponds to an era characterized by substantial reforms to restructure MENA banking systems, increased banking market changes and foreign bank penetration.

Notes on variables and data sources are presented in Table 2 (Appendix). Table 3 reports descriptive statistics for the entire set of included variables (Bank specific variables, macroeconomic variables, institutional variables).

Bank level financial statements, structure and ownership are obtained from Bankscope database, different banks reports, and the financial structure Data set developed by Beck et al. (2008).

It‘s imperative to control for macroeconomic, regulatory and institutional factors that are likely to affect market structures, banking stability or both. The data used are retrieved from the World Development indicators (WDI) provided by the World Bank.

3.2 Empirical Models and Variables Selection

To test several hypotheses between liberalization, banking structure and financial stability, we follow a model similar to Berger et al. (2008), Demirgüç-Kunt and Enrica Detragiache (2010):

$$ zit=\alpha it+\sum {\beta k(Bit;Mit,Iit)+\varepsilon it} $$

The dependant variable is Z-score for country i at time t, Bit is the banking specific variables, Mit is the macroeconomic country specifications, It is the institutional variable and \( \varepsilon it \) an error term. \( \alpha it \) and \( \beta it \) denote the parameters to be estimated.

We estimate versions of panel data to test the following hypotheses:

  • H1: Financial liberalization and stability,

  • H2: Banking market structure and stability

  • H3: Banking market structure, governance and stability.

Three sets of factors might explain the stability of the banking system in a cross-country framework: namely bank industry factors, business environment and institutional variables. A set of control variables that capture macro-economic differences between countries is also used.

3.2.1 The Dependent Variable: The Z-Score

The Z-score became a popular measure of banking solidity in several empirical studiesFootnote 4: This ratio represents the bank’s distance to default, and used as a proxy of banking soundness. This index combines the profitability and return volatility in only one measurement and is denoted as follows:

$$ Z=\frac{{\lambda +ROA}}{{\sigma ROA}} $$

Where \( \lambda \) is the equity to assets ratio (capital ratio)

ROA is the Return on Assets (Bank’s profitability)

\( \sigma ROA \) = The standard deviation of ROA (Volatility of earnings)

The Z-score increases with bank’s profitability and capitalization, and decreases with increasing volatility of earnings. Hence, the Z-score can measure the probability of a bank becoming insolvent when the value of assets becomes lower than the value of debt. Thus, a higher Z-score implies a lower probability of insolvency risk and vice versa.

3.2.2 Bank Specific Variables

  • Structure variables:

In this study, we focus our analysis on one main banking structure changes and its impact on stability. Various measurements of banking market structure and competition belong to two principal approaches: structural and nonstructural ones.Footnote 5

Structural approaches are based on the SCP paradigm, the efficiency hypothesis and oligopoly models. The bulk of studies opt for the k-bank concentration ratio (CRk) or the Herfindahl-Hirschmann index (HHI) as a measure of market concentration.

To investigate market structure of MENA banking, we use the most frequently applied measure of concentration: the 3-bank concentration ratio (CR3). It is calculated as the fraction of assets held by the three largest commercial banks.

Two others indicators are used in this study the H-statistic and the Lerner index to measure banking market power. These two measures are based on the nonstructural approach hypotheses.Footnote 6

The H-statistic frequently used in the “new empirical industrial organization literature” is designed to test the market structure of an industry for competiveness.

It is a variable that captures the competitiveness of the banking industry whereby H ≤ 0 indicates monopoly equilibrium; 0 < H < 1 indicates monopolistic competition and H = 1 indicates perfect competition.

Another variable used as a proxy for market power is the Lerner index. It ranges from 1 to 0, with higher numbers implying greater market power. The approach followed is similar to that of de Guevara et al. (2007) and Delis and Pagoulatos (2009) who defined the Lerner index as

$$ {L_{it }}=\frac{{({p_{it }}-m{c_{it }})}}{{{p_{it }}}} $$

Where p is the price of total assets computed as the ratio of total revenue to total assets; mc is the marginal cost of total assets. To calculate the Lerner index, we first estimate the following trans-log cost function:

$$\begin{array}{lllll}\ln C= {\alpha_0}+{\alpha_1}\ln Y+\frac{1}{2}{\alpha_2}\ln{Y^2}+\sum\limits_{k=1}^2 {{\beta_k}} {W_k}+\sum\limits_{h=1}^2 {{\mu_h}} \ln{E_h} \hfill \cr +\frac{1}{2}\sum\limits_{k=1}^2 {\sum\limits_{m=1}^2 {{\gamma_{km }}} } \ln{W_k}\ln{W_m}+\sum\limits_{k=1}^2 {{\rho_k}} \ln Y\ln{W_k}+\sum\limits_{h=1}^2 {{\varepsilon_h}} \ln Y\ln{E_h} \hfill \cr +\sum\limits_{k=1}^2 {\sum\limits_{h=1}^2 {{\lambda_{kh }}} } \ln{W_k}\ln{E_h}+\frac{1}{2}\sum\limits_{h=1}^2 {\sum\limits_{n=1}^2 {\psi \ln} } {E_h}\ln{E_n}+\ln{u_c} + \ln{\varepsilon_c} \end{array}$$

Where C denotes total cost, and Y is total assets. W is the vector of inputs (labor, funding, and other costs), and E is the vector which includes fixed assets, loan loss provisions, and equity capital. To obtain marginal cost, we differentiate with respect to Y as follows:

$$ m{c_{it }}=\frac{{\partial C}}{{\partial Y}}=[{a_1}+{\alpha_2}\ln Y+{\rho_1}\ln{W_1}+{\rho_2}\ln{W_2}+{\varepsilon_1}\ln{E_1}+{\varepsilon_2}\ln{E_2}]\frac{{{c_{it }}}}{Y} $$

  • Other specific variables:

Domestic credit provided by the banking sector (% of GDP) DCPS includes all credit to various sectors on a gross basis, with the exception of credit to the central government, which is net. This ratio is used to measure the growth of the banking system

Bank liquid reserves to bank assets ratio (%) is the ratio of domestic currency holdings and deposits with the monetary authorities xcwc to claims on other governments, nonfinancial public enterprises, the private sector, and other banking institutions. This ratio captures the banking system’s liquidity. In countries whose banking system is liquid, adverse macroeconomic conditions should be less likely to lead to banking and financial crisis. An increase of this ratio is positively correlated with stability.

Interest rate spread (IRS) is the interest rate charged by banks on loans to prime customers minus the interest rate paid by commercial or similar banks for demand, time, or savings deposits.

In simple terms, the net interest spread is like a profit margin. The greater the spread, the more profitable the financial institution is likely to be and vice versa.

3.2.3 Control Variables

  • The global financial liberalization index: IGLF

In the last decades, we note the proliferation of empirical studies trying to detect the impact of financial liberalization process on financial systems, investment and growth. However, in spite of this growing number, the indicators of financial liberalization measurement are still very limited.Footnote 7 In fact, the majorities of researchers build their own liberalization chronology and concentrate on the abolition of restrictions on financial sector.

We pursue an approach inspired from Laeven et al. (2003), Demirguc-Kunt and Detragiache (1999) and Kaminsky and Schmukler (2003). We will try, in what follows, to build a global index of financial liberalization for MENA countries. It is a compound index of three fundamental aspects of the financial liberalization process: The domestic financial sector deregulation, the stock market liberalization and capital account deregulation. The total index of liberalization evaluates jointly with the capital account, the stock market and the domestic financial sector liberalization.

Our sample covers 13 MENA countries over the period 1980–2009. For each country, three forms are identified:

  • Completely liberalized: If the three sectors are perfectly liberalized;

  • Partially liberalized: At least a sector is partially liberalized;

  • Repressed: Total restrictions for all sectors.

This index is an inter-country average whose value varies between 1 and 3:

3 = Strong restrictions; 2 = Partial Liberalization; 1 = Total liberalization

To capture the capital account liberalization, we study the regulations evolution on the external financing of domestic financial institutions, the evolution of the exchange control, and controls of capital outflows. To measure the liberalization of the domestic financial system, we analyze regulations on debtor and credit interest rates, easing of banking legislation as regards of granting of credit, and deposits in different currencies. Finally, to detect stock markets liberalization, we study acquisition evolution and shares of domestic stock market by foreigners, capital repatriation, and interest and dividends.

We established these criteria after having gathered all the payments and have carefully studied the range of restrictions adopted throughout countries and years of several sources.Footnote 8

The information sources include arrangements of exchange rate (Publications of the IMF and the restrictions and economic developments, various issues). The data base of emerging markets is from the publication of IFC. We use also various reports/ratios of the World Bank, annual reports of the central banks, as well as of the research tasks with chronologies of the restrictions of the financial system.

  • Macroeconomic variables

Bank soundness is also affected by macroeconomic variables, as slow output growth, high and volatile inflation, rapid exchange rate depreciation, high real interest rates, and rapid credit expansion have been found to be associated with bank instability. We included several financial and macroeconomic variables that have been consistently identified in the literature as significant in the determination of banking crises.Footnote 9 The control variables included in all regressions are:

  • Growth rate: A decline of growth rate generally weakens the capacity of the domestic borrowers to ensure the refunding of their debt and thus increases the risk of insolvency; borrowers’ insolvency may be higher under decreasing economic performance which in turn deteriorates the banks’ asset quality. In addition, recessions generally advance the episodes of banking distress Laeven and Majoni (2003). Thus, we expect a positive coefficient sign if investment opportunities increase under economic booms with the growth rate variable expected to be positively correlated with banking stability.

  • Inflation: INF, the volatility inflation rate in general affects the solidity of the economy and the banking system in particular, through several transmission channels. This volatility accentuates the credit and market risks. Inflation is measured by the annual growth rate of the GDP implicit deflator showing the rate of price change in the economy as a whole. The GDP implicit deflator is the ratio of GDP in current local currency to GDP in constant ones.

  • M2/Reserves internationalFootnote 10: It is an indicator of confidence for investors in the interior financial system. This indicator informs about the capacity of the economy to resist the speculative pressures without correction of foreign exchange rate. Any increase in this ratio is a sign of vulnerability.

Thus, in robustness tests, we employ various combinations of these macroeconomic variables in alternative specifications. Macroeconomic variables are mainly retrieved from the world development indicators.

4 Empirical Results

4.1 A Method

We rely on a fixed effect model. This one was preferred to the random one despite that the Hausman test didn’t succeed to allow us to choose between random and fixed effect model.Footnote 11 In fact, the latter is adapted to capture unobserved specific effects of countries, such as institutions, geographical characteristics, cultural norms, etc. Our regressions are estimated with a weighted least-squares procedure, employing a White correction for heteroskedasticity (cross-section weights).

We present main empirical results in Tables 10.1 and 10.2. Table 10.1 reports regressions results assessing the impact of financial liberalization sequencing on systemic stability as measured by the Z-score-technique. Regression specifications in Table 10.2 use different concentration and competition measures to seek the impact of banking structure changes on banking stability.

Table 10.1 Descriptive statistics
Full size table
Table 10.2 Financial liberalization sequencing, concentration, and banking stability.
Full size table

4.2 Main Findings

4.2.1 Financial Liberalization Sequencing, Concentration, and Banking Stability

Contrary to our anticipations, the empirical results indicate that financial liberalization measured by the global financial index is significantly positive at 1 %, so it has a positive impact on Z-score which induces more stability (Table 10.2). According to the majority of existing empirical studies, liberalization generally has a negative effect on banking stability.

In our case, this result can be explained by the non advanced stage and recent adoption of financial liberalization process in MENA region. The majority of countries did not carry out a total financial liberalization yet.

In regressions 1 and 2 of Table 10.1, we try to study the effect of sequencing on banking stability. For stock market liberalization, we didn’t find a significant result. Capital account liberalization enters the third regression with a negative and significant sign at 1 % level inducing more fragility for the banking system.

Hence, to guarantee a successful financial liberalization, MENA countries have to ensure banking solidity by an adequate prudential regulation able to limit the harmful effects of a total opening.

4.2.2 Banking Market Structure, Competition and Stability

As Table 10.3 reports, concentration (C3) enters regression (1) significantly negative at the 1 % level suggesting that an increase in banking market concentration has a negative impact on MENA banks financial soundness, which corresponds to the “concentration- fragility” view in the theoretical literature and generally confirms empirical findings by De Nicolo et al. (2004). In contrast, this result does not support theoretical arguments and earlier empirical findings (Beck et al. 2006) promoting the “concentration-stability view”. We additionally control for the robustness of our main findings by adopting other competition measures.

Table 10.3 Banking market structure, competition and stability
Full size table

In Specifications (2) we include the H-Statistic and in specification (3) the Lerner index.

The H-Statistic enters specification 2 positively and significantly at the 1 % level. The positive sign for the coefficient implies that the Z-score increases as the degree of competitive behaviour among financial institutions increases and therefore does not support the view that competitiveness gives rise to banking system vulnerabilities.

For Lerner index, we find that more monopolistic banking systems can induce to more fragility which confirms our first result.

4.2.3 Banking Market Structure, Governance and Stability

Finally, to test the impact of institutional variables on banking stability, we adopt the same methodology. We add to the baseline model the vector of institutional variables composed of the six indicators derived from World Governance Indicators compiled by Kaufmann and Kraay (2008). The empirical results are reported in Table 10.4.Footnote 12

Table 10.4 Banking market structure, governance and stability
Full size table

Examining the coefficients on the various institutional variables leads to a number of additional interesting results. The signs of all institutional variables are negative, but only one variable (political stability) is reported to affect positively the bank stability.

Our findings highlight also the importance of institutional environment in enhancing banking stability. Specifically, a sound regulatory quality and a better enforcement of rules of law, play an important role in reducing fragility in the MENA banking system. So we can conclude that stability in the banking system depends on legal and political institutions.

5 Conclusion

The purpose of this paper is to examine the impact of financial liberalization, bank structures changes, and institutional environment on MENA banking stability over 1990–2008. Empirical results indicate that financial liberalization has a positive impact on Z-score which induces more banking stability. Our analysis provides also empirical evidence that MENA banking market concentration has a negative impact on banks’ financial soundness as measured by the Z-score technique while controlling for macroeconomic, bank-specific, regulatory and institutional factors. Empirical results from panel estimations hold when employing alternative concentration and competition measures. Our findings are consistent with the “concentration-fragility view”. Finally, our findings highlight the importance of institutional environment in enhancing banking stability. Specifically, a better control of corruption, a sound regulatory quality, a better enforcement of rules of law, and a free voice and accountability play an important role in reducing fragility in the MENA countries.