literature by providing empirical evidence as to whether higher ownership concentration and. ٢٤ ... Contrary to the fi
IRTI Working Paper Series WP/2016/10
Are Islamic Banks Any Different in Financial Stability? An Empirical Investigation Dawood Ashraf, Mohamed Ramady, Khalid Albinali
24 Rajab 1437H | May 1, 2016
Islamic Economics and Finance Research Division
IRTI Working Paper 2016-10 Title: Are Islamic Banks Any Different in Financial Stability? An Empirical Investigation Author(s): Dawood Ashraf, Mohamed Ramady, Khalid Albinali
Abstract
This paper investigates the role that ownership structure and diversification of income plays in the financial stability of banks from the GCC region. We find evidence that suggests that higher concentration of ownership in any type of shareholding is associated with higher insolvency risk. However, this higher insolvency risk is not associated with any specific type of shareholders. Higher financial fragility is also associated with the size and whether the bank is an Islamic bank. Banks engaged in substantial fee-based activities are more financially stable as compared with banks that predominantly generate their incomes from traditional intermediation activities. Keywords: Emerging markets, GCC, banks, financial stability, ownership structure, income diversification. JEL Classification: G21, G28, G32
_____________________________________________
Suggested citation:
IRTI Working Paper Series has been created to quickly disseminate the findings of the work in progress and share ideas on the issues related to theoretical and practical development of Islamic economics and finance so as to encourage exchange of thoughts. The presentations of papers in this series may not be fully polished. The papers carry the names of the authors and should be accordingly cited. The views expressed in these papers are those of the authors and do not necessarily reflect the views of the Islamic Research and Training Institute or the Islamic Development Bank or those of the members of its Board of Executive Directors or its member countries.
Islamic Research and Training Institute 8111 King Khalid St. Al Nuzlah Al Yamania Dist., Jeddah 22332-2444 Kingdom of Saudi Arabia
Are Islamic Banks Any Different in Financial Stability? An Empirical Investigation
1 2 3 4 5 6
1.
Introduction
7 8
During the recent global financial crisis (GFC) the fragility of the financial system and elevated
9
risk-taking behavior of banks became more pronounced especially for banks located in the more
10
financially open and globally integrated economies. At the time when international banks were
11
facing a financial turmoil, banks from the GCC1 region exhibited a lower probability of default as
12
compared with most of their Western counterpart. The banking sector of the GCC region is roughly
13
10% of the US banking sector in terms of asset size2 with comparable income streams from non-
14
traditional fee-based activities3. The ownership of banks is most developed countries is widely
15
held. However, the ownership structure of banks in the GCC region is concentrated either through
16
significant government, institutional, or family-group membership. Aside from a dissimilar
17
ownership structure a sizeable proportion of banks in the GCC region are Islamic banks adhering
18
to Shari’ah principles for their investment and financing activities. These differences warrant an
19
investigation into whether there is any association between the financial stability of banks in the
20
GCC region with that of their ownership structure, the type of bank (Islamic or conventional) and
21
their respective income diversity. To the best of the authors’ knowledge, there is no empirical
22
study that provides an insight into the relationship between the ownership structure and financial
23
fragility of banks from the GCC region4. This paper attempts to address this shortcoming in the
24
literature by providing empirical evidence as to whether higher ownership concentration and
25
income diversification has any association with the financial stability of banks from the GCC
26
region.
27
1
GCC member countries include Bahrain, Kuwait, Oman, Qatar, United Arab Emirates and Saudi Arabia. FDIC reports total assets of all insured commercial and saving banks assets in the US, as at December 31, 2011 stands at USD 13.892 trillion versus our compilation of the total assets of GCC commercial, savings and Islamic banks assets of USD 1.423 trillion at the same time. 3 The fee-based activities of Western banks are not similar to that of banks in the GCC region. However, the motivation behind the fee-based activities is ‘income diversification’ and we focus on that part only in this paper. 4 Hassan et al. (2013) and Abraham (2013) provide some insight on the performance of banks with foreign ownership in the area but neither of these studies specifically examined the relationship between financial stability and the ownership structure of banks within the GCC region. 2
28
Using a panel data set extracted from the financial statements of 125 GCC banks from 2000
29
to 2011 matched with cross-section ownership data5, we asked whether cross-sectional variation
30
in ownership concentration is related to the subsequent financial stability of these financial
31
institutions. Specifically we evaluate the impact that ownership patterns, types of banks,
32
diversification of income along with other bank-specific and macroeconomic variables has on the
33
financial stability of banks within the GCC region. Z-score is used as a composite measure of
34
financial stability/risk among GCC banks. As a robustness check, we further test whether any
35
deviation from traditional lines of business (disintermediation) affected the financial stability of
36
banks in a simultaneous equation framework.
37 38
The key finding from the empirical estimations is that a higher concentration of ownership
39
in any type of shareholders (government, institutional, or family-group) affected the financial
40
stability of banks in the GCC region over the sample period. Banks with higher ownership
41
concentration in any type of shareholders exhibited a higher financial fragility and thus higher risk-
42
taking. Furthermore, banks larger in size and Islamic banks exhibited a lower risk aversion. We do
43
not find any statistically significant change in the financial stability indicator during the peak of
44
the GFC. Contrary to the findings of Demirgüç-Kunt and Huizinga (2010) we found that
45
engagement in non-intermediation fee-based activities improved the financial stability of banks in
46
the GCC region after controlling the endogeneity.
47 48
The findings of this paper should be of interest to both banking academics and policy
49
makers. The engagement in non-traditional fee based activities is usually considered as detrimental
50
to the financial stability of banks. However, careful diversification of income sources may result
51
in more financially stable banks. However, there is clearly a need for careful regulatory scrutiny
52
to ensure that these activities are not used in ways that jeopardize financial stability.
53 54
The remainder of the paper is organized as follows. Section 2 provides the literature review.
55
Section 3 presents an overview of the GCC financial sector. Section 4 describes the specification
56
of the empirical model. Section 5 describes the data sources and reports descriptive statistics.
5
Since there is very little variation in the ownership of banks within the GCC region, we used cross-sectional data.
27
57
Section 6 presents and interprets empirical estimations. Section 7 addresses endogeneity concerns
58
and presents the results of the robustness test. Section 8 summarizes and concludes the paper.
59 60
2.
The GCC Financial Sector
61 62
The GCC region has experienced a tremendous growth over the last decade due to high oil prices,
63
expanded oil production and expansionary fiscal policies (See Figure 1). The GCC region has one
64
of the world’s highest household savings ratios, which enabled the region to absorb potential
65
financial losses arising from the fallout of the global financial crisis during 2008-2009. Table 1
66
reports the GCC countries savings ratio, as measured by the nominal gross savings ratio as a
67
percentage of GDP for the period 2004-2012. It is evident from Table 1 that the GCC countries,
68
without exception, experienced a reduction in their nominal savings ratio in 2009 when inflation
69
rates rose sharply in the region following the onset of the global financial crisis in 2008. The crisis
70
did lead to a few non-banking financial institutions and family business groups to default.6
71
However, these defaults were isolated and did not have systemic consequences.
72 73
Regulatory authorities across the GCC region have put in place some advanced regulatory
74
practices, drawing upon international standards as well as domestic requirements to reduce
75
systemic financial risk (Naceur and Omran, 2011). During the financial crisis, regulatory
76
authorities of the region stepped in to provide a range of supportive policies to stabilize any fallout
77
to domestic financial markets. These measures are summarized in Table 2.
78 79
The GCC region is host to a large number of local, joint-venture, government and foreign
80
owned banks. Table 3 presents the ownership structure within GCC region with shareholding of
81
15 percent and above. It is evident from Table 3 that ownership among the GCC banks is
82
concentrated among large shareholders however, the pattern of ownership concentration is not
83
similar across all GCC countries. Government ownership holding is more prominent in the case of
84
Saudi Arabia, Qatar and the UAE, but less pronounced for Kuwait, Bahrain and Oman. The level
85
of local shareholding is more significant in Kuwait and Bahrain, while foreign shareholding is a
86
significant factor in the Saudi Arabian banking sector through the legacy of enforced
6
These include Saudi owned Al Sanei and Al Gosaibi groups, and Bahrain based Awal Bank and Arcapita, with the latter becoming the first Gulf institution to filing bankruptcy under Chapter 11 in 2011 and later emerge from it in 2013.
87
“Saudization” of the original foreign banks operating in the Kingdom 7. Foreign shareholding in
88
Bahraini institutions mostly reflects inter-GCC ownership.
89 90
On the surface, the shareholding structure with strong government influence/control and
91
individual/family management may lead to a higher risk-aversion among financial institution from
92
GCC countries. In the next section, we develop a model on the premise whether variations in
93
ownership structure has some influence on the risk-taking behavior of banks from the GCC region.
94 95 96 97
3.
98
Existing empirical literature on financial fragility of banks usually employs the Z-score as an
99
indicator of relative financial stability8. It is an indicator of the probability of failure of a bank and
100
measures the number of standard deviations a return would have to fall to deplete the sum of
101
income and equity9. Z-score has an advantage over other accounting based measures of risk such
102
as non-performing loans since it includes the return on both intermediation and fee-based activities
103
of the bank. We calculate Z-score as:
An Empirical Model for the Financial Fragility and Ownership Structure of GCC Banks
104
Z it
105
rit Eit σ it
(1)
106
where subscript i and t indicates bank and time respectively. rit , σit and Eit are the return on
107
assets, standard deviation of bank i’s return on assets, and the book value of equity to capital ratio.
108
A bank with a lower Z-score indicates higher probability of failure. It is widely argued in the
109
literature that the Z-score is highly skewed (Laeven and Levine, 2009; Schaeck et al., 2012)
110
therefore, we used its log transformation in all empirical estimations.
7
The process of Saudization involved wholly foreign owned banks operating in Saudi Arabia to sell 60 percent of their ownership to Saudi private sector founding members and the public in order to remain operating in the Kingdom. At the same time, the newly ‘Saudized’ banks continued to be managed by the foreign minority shareholder and were granted the right to set-up branches in the Kingdom and were offered attractive terms on their retained profits. This Saudization process ended in the mid 1980’s with the Saudization of all foreign banks completed (such as Arab Bank Ltd., Citibank, HSBC, ABN Amro, Banque Indochine, etc.). However, with the accession of Saudi Arabia to the WTO in 2005, wholly owned foreign banks were once again permitted to operate in the Kingdom (Ramady, 2010). 8 See for example Boyd and Runkle (1993), De Nicoló (2000), Stiroh (2004), Stiroh and Rumble (2006), Laeven and Levine (2009), Demirgüç-Kunt and Huizinga, (2010), Barrel et al. (2010), De Haan and Poghosyan (2012). 9 Boyd and Runkle (1993) argue that use of z-score is valid even when the return series is not normally distributed as long as return and standard deviation exist.
29
111
The model specification is as follows:
112
Z _ scoreit X it ui it
113
i = 1, . . ., N; t = 1, . . ., T
(2)
114 115
where Z_scoreit represent as a measure of financial stability used for the analysis as dependent
116
variable, i denoting banks and t denoting time. Xit contains a set of explanatory variables. ui and
117
it denotes the time-invariant unobservable individual bank specific effect and the remainder
118
disturbance respectively. The next section presents the list of covariates as candidates for Xit for
119
the estimation of Equation 2.
120 121
3.1
Determinant of Banks’ Stability Factors among GCC Countries
122 123
The literature on the risk-taking behavior of banks identifies several variables as candidates for
124
inclusion in X it . These include bank-specific and macroeconomic variables that are believed to
125
affect the risk-taking behavior of banks.
126 127
3.1.1 Bank Ownership Structure
128 129
The bank ownership structure is a complex matter and has various dimensions. The canonical
130
shareholders’ moral hazard problem suggests that shareholders of non-financial firms with
131
excessive debt in their capital structure have incentives to expropriate their lenders by accepting
132
excessive levels of risk (Jensen and Meckling, 1976; Amihud and Lev, 1981; Jensen and Murphy,
133
1990; and Coles et al., 2006). This moral hazard problem is further compounded in the case of
134
banks where depositors are least interested in monitoring the risk-taking behavior of banks.
135 136
The role of ownership structure on the risk-taking behavior of banks is well established in
137
the empirical literature at regional as well international level10. The findings of these studies are at
138
best mixed. The focus of earlier empirical studies on the relationship between risk-taking behavior
139
and ownership structure of banks was to isolate the impact of managerial ownership on the risk-
10
For detailed list of studies, please see Table A-1.
140
taking behavior of banks in a single country context11. Recently, emphasis has shifted towards
141
much broader geographic regions wherein researchers not only used samples from several different
142
countries but also included categories of shareholders other than managerial ownership 12.
143
Empirical literature reports a positive relationship between risk-taking and incremental change in
144
stockholding by bank management (Anderson and Fraser, 2000; Sullivan and Spong, 2007). On
145
the contrary many researchers have found that banks with a higher degree of managerial ownership
146
(management group as shareholders) are more risk averse as compared with widely held banks
147
(Saunders et al., 1990; Gorton and Rosen, 1995; Lee, 2002).
148 149
Our main variable RIGHTi, is the proportion of highest shareholding in the ownership
150
structure of bank i as per type of ownership: government, general public (widely held banks),
151
individual/family and institutional/corporate investors by the end of 2011. RIGHTi measures the
152
concentration by any type of shareholder in the ownership structure. We use the minimum five
153
percent ownership threshold such as one used by Laeven and Levine (2009). By focusing on
154
concentration of ownership, we expect to identify the incentives for risk-taking influenced by the
155
largest shareholders. We conjecture a negative association between the financial stability of banks
156
and the rising concentration in the ownership structure.
157 158
Corporate governance literature on non-financial firms suggests increasing the
159
proportionate ownership of management as a strategy for mitigating the shareholder-management
160
agency problem. However, for banking firms, it is debatable whether higher managerial ownership
161
provides any incentive for lower risk aversion due to a very low proportionate ownership stake as
162
compared with the depositors. This is also evident from the empirical findings confirming on one
163
hand, a negative association between the insider ownership and risk-taking (Saunders et al., 1990;
164
Demsetz et al., 1997; Lee, 2002) and on the other hand reporting a positive association between
165
managerial ownership and risk-taking (Gorton and Rosen, 1995; Anderson and Fraser, 2000). In
11
See for example Saunders et al. (1990), Gorton and Rosen (1995), Demsetz et al. (1997), Anderson and Fraser (2000) and Lee (2002) used a sample of US banks. Konishi and Yasuda (2004), Jia (2009), and Chou and Lin (2011) used sample of banks from Japan, China, and Taiwan. 12 See for example Micco et al (2007), Laeven and Levine (2009), Angkinand and Wihlborg (2010), Forssbaeck (2011) and Hossain et al. (2013) used a sample of global banks while authors who have studied the regional banks’ risk taking include Haw et al. (2010), Barry et al. (2011), Chalermchatvichien et al. (2014a) and Chalermchatvichien et al. (2014b) for Asian, European and East Asian regions. For a review of sample, period and their findings, please refer to Appendix A-1. For a more comprehensive review, specific to Asia Pacific region, please see Benson et al., 2014 and 2015.
31
166
line with the above argument, we conjecture that a shareholder sitting on the board of directors
167
may influence the strategic direction of the bank and hence the financial stability as well. We use
168
a dichotomous variable DIRi, that takes the value of unity if a shareholder with five percent or
169
more ownership stake has the representation on the board of directors or zero otherwise. The data
170
for DIRi is hand-collected from various sources including Bankscope, and bank annual filings with
171
stock exchanges and annual reports.
172 173
Aside from the ownership concentration and insider ownership, the stability of banks can
174
be affected by type of ownership. For example, banks owned by institutional investors are less risk
175
averse as compared with banks owned by individual/family or government due to their focus on
176
short-term returns (Berger et al., 2005; Barry et al., 2011; Chou and Lin, 2011; Hossain et al.,
177
2013). On the hand state-owned banks are less efficient as compared with widely-held or foreign
178
owned banks (Angkinand and Wihlborg, 2010; Jia, 2009; Casu and Molyneux, 2003; Yildirim and
179
Philippatos, 2007). To capture the impact of type of ownership on the financial stability of banks
180
in the GCC region, dummy variables: INDi, INSTi, GOVi and PUBi which take the value of one if
181
the majority shareholding of bank i is with an individual/family, institutional investors,
182
government or widely held respectively and zero otherwise.
183 184
3.1.2 Other Bank Specific Determinants of Risk-Taking
185 186
Among other bank-specific determinants for financial fragility, a bank’s income sources could
187
play a major role in the stability of banks. Income sources for banks have changed considerably
188
over the past couple of decades. Income diversification has been identified as one of the major
189
factors that may contribute to the fragility of banks whereby banks have become increasingly
190
dependent on fee based income and thus jeopardizing the financial stability of the whole financial
191
system (Ashraf and Goddard, 2012; Demirgüç-Kunt and Huizinga, 2010). Busch and Kick (2009)
192
found that fee income is more stable for commercial banks in Germany from 1995 to 2007 period.
193
Ashraf and Goddard (2012) using a dataset of the US commercial banks from 2001 to 2009 find
194
that banks facing downward pressure in growth of their loan portfolio are more likely to diversify
195
their income stream by engaging into more fee based activities such as dealing in derivatives.
196
Stiroh (2004) reports that US banks with greater reliance on noninterest income faced higher risk
197
and lower risk-adjusted returns using a quarterly data set from 1984 to 2001. While Demirgüç-
198
Kunt and Huizinga (2010) found similar results and suggest that higher fee income may have some
199
potential risk diversification benefit but at very low levels. NONIIit is the ratio of non-interest
200
income to gross income for banks i at time t. A positive relation with Z-score implies the
201
diversification benefits for banks. On the other hand, a negative and statistically significant
202
association of NONIIit with Z-score would imply a higher risk-taking due to greater reliance on
203
income from fee-based activities.
204 205
Since major population within the GCC region are Muslim. The GCC region hosts some
206
of the biggest Islamic banks in the world. The Islamic banks use risk-return sharing models and
207
hence may exhibit different behavior than conventional banks. To isolate the impact of the type of
208
bank as Islamic, we use a dummy variable (ISLi) that takes the value of one, if a bank is Islamic or
209
zero otherwise. We expect ISLi to be less risk averse as compared with conventional banks hence
210
expects a negative association with the Z-score.
211 212
The size of the bank is considered as an important risk-taking incentive. Larger banks tend
213
to engage higher risk activities due to the existence of implicit and explicit deposit insurance and
214
their role in the financial stability of the economy (too big to fail). On the other side, large banks
215
have larger pool of assets and thus are better able to diversify their loan portfolio (Diamond, 1984).
216
SIZEit is the natural logarithm of total assets of bank, i in year t. We anticipate a negative coefficient
217
of SIZEit with Z-score due to existence of too-big-to-fail phenomenon where larger banks assume
218
higher risk-taking due to their size.
219 220
Globally, bank supervisors have used risk-based capital requirements to improve the
221
stability of banks. However, the effectiveness of such regulatory stringency is disputed. Early
222
studies such as Sharpe (1978), Furlong and Keeley (1989) argue that increasing capital
223
requirements reduces the value of put option and thus reduce the excessive risk-taking. On the
224
other hand, stringent capital requirement may reduce the size of the risky portfolio but may
225
increase overall portfolio risk (Acharya, 2009; Ashraf, 2008; Kahane, 1977; Koehn and
226
Santomero, 1980). The relationship between the tighter capital regulations and risk-taking
227
incentives may be U-shaped, i.e. a bank may respond to capital regulations by reducing the risk
228
level but ultimately takes more risk (Calem and Rob, 1999). Furthermore, once a bank is in full
229
compliance with the capital requirements, changes in the minimum capital requirement have little
230
impact on risk-taking behavior however, if capital declines to the minimum level, banks are likely 33
231
to accept excessive levels of risk due to the shareholders’ moral hazard problem (Milne and
232
Whalley, 1998; Jokipii and Milne, 2009).
233 234
RWCit is bank i’s ratio of sum of risk-weighted assets to total assets. Banks with a higher
235
ratio of risk-weighted assets to total assets are more likely to maintain a buffer and manage their
236
portfolio risk when they have sufficient capital to comply with the capital requirement.
237
Accordingly, a positive coefficient is anticipated.
238 239
Among other notable bank specific factors that may contribute to the risk-taking of banks
240
is LLOSSit, the logistic transformed ratio of bank i’s loan loss reserve to gross loans. A positive
241
relationship of LLOSSit with Z-score would imply income smoothing practices of banks where
242
banks increase their loan loss reserves during years with abnormal profit and reduce during
243
economic downturns (Bouvatier et al., 2013; Bushman and Williams, 2012; Collins et al., 1995).
244
On the other hand, a negative relationship may indicate deteriorating quality of loan portfolios and
245
thus increases the insolvency risk for banks.
246 247
3.1.3 Macroeconomic Covariates Affecting the Risk-Taking Behavior among GCC Banks
248 249
Several researchers have examined the relationship between macroeconomic indicators and bank
250
performance (Berger et al., 2000; Daly et al., 2004; Albertazzi and Gambaracorta, 2009; Laeven
251
and Levine, 2009; Demirgüç-Kunt and Huizinga, 2010; Bushman and Williams, 2012). Albertazzi
252
and Gambaracorta (2009) suggest that GDP, as an indicator of business cycle, affects not only the
253
portfolio quality (increased loan losses) but also the intermediation activities (lower interest
254
income due to lower growth in loan portfolio). We use the per capita GDP income (GDPjt) as a
255
proxy for the business cycle fluctuations and overall economic condition in the local economy.
256
GDPjt is the natural log of the per capita GDP income of country j in year t. A positive relationship
257
of GDPjt with Z-score implies stability of banks during the economic expansion and otherwise
258
during the economic contractions as suggested by Albertazzi and Gambaracorta (2009).
259 260
Highly concentrated financial markets create opportunities for non-competitive behavior,
261
resulting in higher risk-taking (Boyd el al., 2006; Jiménez et al., 2013; Pilloff and Rhoades, 2002).
262
The intensity of competition in the local banking market may influence the risk-taking behavior of
263
banks. For market concentration we use the Hirchman Herfindahl Index (HHIjt), is the sum of the
264
squares of the market shares of all banks based on total assets of each bank existing in country j in
265
year t. More formally, HHI jt (TAit TAjt ) 2 where TAit is the total assets of bank i in year t and
266
i = 1, . . . , N. TAjt is the sum of total assets of all banks in country j at time t. We expect a negative
267
coefficient of HHIjt, with Z-score.
268 269
Regulatory restrictions on banking activities can impact the risk management behavior of
270
banks. We develop a composite index (CONTROLj) for each of the GCC countries based on
271
authors’ own assessment of six major regulatory factors that may impact the risk-taking behavior
272
of banks in GCC region. These factors include regulatory capital, central bank regulatory oversight
273
and support, regulatory controls – stress testing, corporate governance, international liquidity
274
benchmarking and international surveillance - capital control. A detailed description of assessment
275
is provided in Appendix A-II. CONTROLj measures the degree to which bank face regulatory
276
restriction on their activities relative to banks in other member GCC countries. A higher value on
277
the scale means tighter control on risk-taking. A positive coefficient for CONTROLj with Z-score
278
would suggest regulatory stringency helps improve the stability among GCC banks.
279 280
To control for the impact of global financial crisis, we use a dummy variable (GFC), which
281
takes the value of one for years 2008 and 2009, zero otherwise. Since larger banks were mostly
282
affected by the global financial crisis, an interaction term of GFC with SIZE is introduced to isolate
283
the impact of global financial crisis as per the size of a bank in GCC region. If banks are affected
284
by the events of GFC a negative coefficient is anticipated on GFC. We expect a negative
285
coefficient with GFC_SIZE since larger banks were mostly affected during the financial crisis.
286 287
4.
Data Sources and Descriptive Statistics
288 289
The sample includes annual data from year 2000 to 2011 (inclusive) on all commercial, savings,
290
and Islamic banks in the six GCC countries by the end of 2011 reporting in the Bankscope
291
database. The initial sample consists of 135 financial institutions with 45 Islamic banks. From this
292
sample we dropped all those financial institutions with gross loans showing a nil amount assuming
293
these banks are not engaged in traditional intermediation activities of deposit taking and lending.
294
Balance sheet, income statement and ownership structure data are taken from Bankscope while we
295
source macroeconomic variable data from the World Bank. Entry into and exit from the sample is 35
296
unrestricted to avoid sample selection bias. The final dataset consists of a sample of 125 banks
297
with a total of 996 year/bank observations.
298 299
Table 4 presents descriptive statistics for the overall sample and individual countries in the
300
sample. It is evident from Panel (1), the summary statistics on the overall sample that the largest
301
shareholder on average owns 71 percent of a bank in the GCC. On average 44 percent of a bank’s
302
shares are widely-held while corporate ownership constitutes more than 30 percent including both
303
local and foreign institutions, followed by government ownership about 21 percent in the overall
304
ownership structure. Banks in GCC region are well capitalized. Income from traditional
305
intermediation activities on average, contributes 58 percent while fee based activities contributes
306
about 42 percent on average to income during the sample period.
307 308
Inter-GCC-country comparison of descriptive statistics is provided from Panel (2) through
309
Panel 7 of Table 4. It is evident that shareholding is generally concentrated with specific type of
310
shareholders among all GCC countries. However, concentration patterns vary considerably. 58
311
percent shares of banking sector are widely-held in Qatar as compared with 35 percent in Bahrain.
312
Above 40 percent of banking sectors’ shares are concentrated with Institutional investors in
313
Bahrain and Oman. While government shareholding represents about 29 percent in case of Saudi
314
Arabia and UAE as compared with 7.5 percent of government shareholding in the overall banking
315
sector of Oman.
316 317
Interestingly, large disparity exists in relative riskiness and other bank specific variables
318
across GCC member countries. In terms of relative riskiness, Kuwait reports the lowest Z-score
319
and highest proportion of loan loss reserves with the highest proportion of income from non-
320
intermediation activities within GCC region. On the contrary, Qatar reports the highest Z-score
321
with 68 percent of income from traditional intermediation activities and relatively lower levels of
322
loan loss reserves. Findings from the descriptive statistics are in line with our hypotheses as set
323
out in the previous section. In summary, descriptive statistics suggest that banks in countries where
324
ownership is more concentered with corporate sector and accrue major portion of income from
325
non-traditional activities are less stable as compared with countries where banking sector is
326
concentrated and traditional intermediation activities contribute larger proportion of income for
327
banking sector.
328
The correlation matrix as reported in Table 5 echoed the findings from summary statistics
329
that relatively riskier banks (low Z-score) have negative and statistically significant association
330
with higher concentration in specific shareholdings, income from fee-based activities, and reserve
331
for loan losses. More stable banks are smaller in size with the main source of revenue from
332
traditional intermediation activities and domiciled in more concentrated markets.
333 334
5.
Empirical Estimations
335 336
Before embarking on empirical estimation, it is pertinent to provide rationale for using the
337
empirical estimation method. A random effect model is preferred since it provides consistent
338
estimates despite little or no time variation among several independent variables. We use a panel
339
data model with random effects to correct for cross-sectional heteroscedasticity and panel specific
340
autocorrelation. For testing whether pooling of the data is warranted, we employ the Breusch and
341
Pagan (1980) Lagrange-multiplier test with the null hypothesis that variance of error term across
342
banks is zero that is Var (ui ) 0 . Wald test are also conducted to test the joint significance of
343
model parameters.
344 345
Table 6 reports the estimation results for various panel model specifications where Z-score
346
is a dependent variable. Model 1 reports the estimation results on the association of ownership
347
concentration of shareholding (RIGHTi) with the financial stability of banks while controlling for
348
the size and type of banks. Models 2 to Model 5 reports the estimation results after controlling
349
bank-specific factors, macroeconomic factors, impact of global financial crisis and dichotomous
350
variables for the type of ownership as discussed in the previous section.
351 352
The key finding from these estimation results is that concentration in the ownership
353
structure of a bank in the GCC region results in greater financial fragility. The proportionate
354
ownership of the largest ownership category is negative and statistically significant at 5 percent or
355
better in each of the models. These results are in line with the view that larger shareholders exhibit
356
a lower risk aversion and have greater incentives to enhance risk-taking at the expense of minority
357
shareholder (Laeven and Levine, 2009; Haw et al., 2010).
358 37
359
Beside proportionate ownership, the coefficients of SIZEit are negative and statistically
360
significant in all models, suggesting that banks larger in size have greater incentives to elevate
361
risk-taking due to the existence of too-big-to-fail phenomenon. Concurrent with our findings,
362
Demirgüç-Kunt and Huizinga (2010) report negative relationship between the size of banks and
363
Z-score after analyzing an international sample of 1134 listed banks from 101 countries over the
364
period from 1995 to 2007. De Nicoló (2000) also reports higher risk-taking by larger banks
365
utilizing a sample of 826 banks in 21 industrialized countries from 1988 to 1998. The covariates
366
on the dichotomous variable ISLi is also negative and significant for all models suggesting that
367
banks following Islamic intermediation model are less financially stable as compared with
368
conventional banks.
369 370
Among bank specific control variables, non-traditional fee-based income (NONIIit) is
371
insignificant suggesting that increasing income from fee based activities does not increase the
372
vulnerability of banks within the GCC region. In contrast with the empirical findings from the
373
study of Demirgüç-Kunt and Huizinga (2010), a higher proportion of income from fee-related
374
businesses did not reduce financial stability of banks within the GCC region. The covariates of
375
loan portfolio quality variable LLOSSit, are also insignificant for all models suggesting that banks
376
expecting higher losses in the future and make provisions for such losses are less fragile. It may
377
also indicate the existence of possible income smoothing practices whereby bank management
378
increases the loan loss provision during the economic expansion and reduce during the contraction.
379
Regulatory capital ratio (RWCit) is positive and significant for all models suggesting that banks
380
complying with regulatory capital requirements by maintaining higher capital ratio are more
381
financially stable (Ashraf 2008).
382 383
Next we control the macroeconomic variables in our regression and estimation results are
384
reported under Model 3 in Table 6. Economic growth as measured by per capita GDP (GDPjt) is
385
positive and significant suggesting that banking sector is less vulnerable for countries facing
386
economic expansion. Contrary to the expectations, the covariates of regulatory control
387
(CONTROLjt) variable are negative and significant in all models suggesting that tighter control
388
may increase the insolvency risk of GCC banks. Similarly, concentration measure (HHIjt) enters
389
in Model 3 as negative and significant suggesting that concentration within the GCC banking
390
increase the insolvency risk. As reported in Model 4, both GFC and its interaction with SIZE are
391
insignificant, suggesting that on average, banks in GCC regions remained unaffected during the
392
global financial crisis. Finally, we re-estimate the Model 4 after including dummy variable DIRi,
393
for individual sitting on the board of directors with shareholdings greater than 5 percent and four
394
dummy variables as per the type of ownership of the largest shareholders. This helps us to further
395
understand the influence of ownership structure on the risk-taking behavior of banks in GCC
396
region. The coefficients on type of ownership are insignificant in explaining the relationship
397
between type of ownership and insolvency risk.
398 399
In summary, empirical findings from the Z-score as a measure of insolvency risk for banks
400
in GCC region suggest that banks with concentrated ownership exhibit higher risk-taking. Among
401
bank specific factors type and size of the bank reduce the stability of banks among GCC countries.
402
Overall, GCC banking sector was unaffected by macroeconomic factors and remained resilient
403
during the financial crisis. However, tighter regulatory control and concentration increased the
404
fragility of banks within the GCC region.
405 406
6.
Robustness Check - Endogeneity Issues
407 408
In this study we have assumed non-traditional sources of revenue (NONIIit) for banks an exogenous
409
variable however, banks adjust their sources of income in response to bank’s risk and return
410
situation (Demirgüç-Kunt and Huizinga, 2010; Ashraf and Goddard, 2012). This simultaneous
411
determination of risk and sources of revenue could potentially cause the endogeneity problem.
412
Since stability of banks can be affected by return volatility emanating from non-traditional fee-
413
based activities consequently, participation in such activities can be altered according to the
414
perceived level of risk13. We address the endogeneity concern by using a two-stage least squares
415
(2SLS) model which allows the simultaneous determination of bank risk and sources of income
416
using the panel data14.
417 418
Banks may engage more in non-traditional fee based activities when there is a pressure on
419
traditional line of businesses or banks have the competitive advantage in offering such services.
420
(Ashraf and Goddard, 2012; Busch and Kick, 2009). We included four variables as excluded 13
Other potential instruments may include risk weighted capital. However, banks in GCC regions are well capitalized so we don’t perceive a situation where banks adjust their risk level to regulatory capital requirements as suggested by Laeven and Levine (2009), Keeley (1990) and Sharieves and Dahl (1992) among several others. 14 For a detailed discussion on 2SLS regression methodology, please see Baltagi (2005), pages 113-133
39
421
instruments in the first stage of 2SLS model. The ratio of interest income to revenue (INT2REVit),
422
the ratio of net-charge-offs to gross loans (NCAGit) and the ratio of total equity to total assets
423
(LEVGit) are included in the estimation with the assumption that banks facing pressure on
424
traditional line of business either by way of reduced portion of income from interest, higher loan
425
defaults or on their equity may engaged more in fee-based activities. FOREIGNi is the dummy
426
variable that takes the value of unity if a bank has a foreign financial institution as a shareholder.
427
The affiliation with a foreign bank increases the potential for cross-selling opportunities and thus
428
local bank engages more in non-traditional fee-based activities.
429 430
Table 7 reports the estimation results based on 2SLS regression model with Z-score as
431
dependent variable for three different models similar to models 2, 3, and 4 in Table 6. Before we
432
present our empirical estimation based on 2SLS model, it is pertinent to test whether instruments
433
are valid and correctly excluded from the estimated model. We conduct a Sargan-Hansen test of
434
overidentification with the joint null hypothesis that the instruments and exclusion restriction are
435
valid. Rejecting the null hypothesis implies that the instruments may not be valid. Sargan-Hansen
436
test statistics are reported in the last row of Table 9 for each model and indicate that endogeneity
437
cannot be rejected at the 5 percent significance level for any of the model and thus application of
438
2SLS regression is appropriate.
439 440
The empirical estimation using the 2SLS regression model generally confirms the previous
441
estimation results as reported in Table 6 after controlling for the endogeneity. Among bank-
442
specific factors, the association of Z-score with RIGHTit, and SIZEit is negative and significant
443
suggesting that higher ownership concentration and bigger size of bank increase the vulnerability
444
of banks in GCC region. While positive and significant coefficient of NONIIit suggests that
445
participation in fee-based activities provide not only the income diversification benefit but increase
446
their financial stability of banks within the GCC region. Both loan portfolio quality and risk
447
weighted capital are insignificant however, after controlling the endogeneity the coefficient of
448
RWCit is positive and significant suggesting the banks with higher risk based capital ratio are more
449
stable. In terms of macroeconomic variables, the empirical estimations are similar to the main
450
model except for GDPjt which is not significant. Coefficients on both CONTROLj and HHIjt are
451
negative and significant suggesting that stronger regulatory controls and concentrated banking
452
sector elevate financial fragility among GCC banks. Similar to Table 6, all other covariates
453
including differential ownership variables are insignificant.
454 455 456
7.
Summary and Conclusion
457 458
During the recent global financial crisis, risk-taking behavior of financial institutions has been
459
criticized. The moral hazard problem associated with the ownership structure and engagement of
460
banks in relatively riskier non-traditional fee-based activities are blamed for the crisis (Demirgüç-
461
Kunt and Huizinga, 2010, Cyree et al, 2011). During the crisis, banking sector from the GCC
462
region not only remained resilient but also provided liquidity to the global financial markets. The
463
ownership of financial institutions from the GCC region is concentrated among specific type of
464
shareholders as compared with widely-held banks in most developed market with comparable
465
income from fee-based activities. This paper investigates the association between the ownership
466
concentration, and income diversification with the financial stability of banks by analyzing a
467
sample of 125 banks from GCC region for the period 2000-2011.
468 469
This study uses Z-score, a composite measure of financial stability, that takes into
470
consideration overall stability of financial institution without any regard to the source of risk. The
471
estimation results suggest that shareholder concentration do influence the stability of banks. Banks
472
with higher concentration of ownership tends to face higher insolvency risk; however higher
473
insolvency risk is not associated with any specific type of shareholders’ concentration. This can
474
be attributed to relatively stringent regulatory oversight and stringent approval process for
475
participation in relatively riskier products such as financial derivatives. Larger banks on average,
476
exhibits lower risk aversion however this does not reflect in relative riskiness among traditional
477
loan portfolios. We find those banks that generate more income from fee-based-activities are more
478
financially stable than banks relying of traditional intermediation activities. We did not find any
479
statistically significant change in the financial stability indicator during the peak of the GFC among
480
banks from the GCC region. In general, empirical findings are consistent after controlling the
481
endogeneity.
482 483
The income diversification and its positive impact on the financial stability of banks in the
484
region is a welcome sign. However, the relationship between the propensity of higher risk-taking
485
with concentrated ownership and more reliance on fee-based income of the GCC banks needs 41
486
careful regulatory oversight to ensure the stability of financial system. Entry of wholly owned
487
foreign banks in the GCC region are encouraging however, the impact of wholly owned foreign
488
banks on the risk-taking behavior warrants further investigation in future research.
489
490
References
491 492
Abraham, A., 2013. Foreign ownership and bank performance metrics in Saudi Arabia. International Journal of Islamic and Middle Eastern Finance and Management, 6(1), pp.43–50.
493 494
Acharya, V. V., 2009. A theory of systemic risk and design of prudential bank regulation. Journal of Financial Stability, 5(3), pp.224–255.
495 496
Albertazzi, U., Gambacorta, L., 2009. Bank profitability and the business cycle. Journal of Financial Stability, 5(4), pp.393–409.
497 498
Amihud, Y., Lev, B., 1981. Risk Reduction as a Managerial Motive for Conglomerate Mergers. The Bell Journal of Economics, 12(2), pp.605–617.
499 500
Anderson, R.C., Fraser, D.R., 2000. Corporate control , bank risk-taking , and the health of the banking industry. Journal of Banking and Finance, 24, pp.1383–1398.
501 502
Angkinand, A., Wihlborg, C., 2010. Deposit insurance coverage, ownership, and banks’ risktaking in emerging markets. Journal of International Money and Finance, 29(2), pp.252–274.
503 504
Ashraf, D., 2008. Capital and risk adjustment under risk-based capital regulations and diversification loss. Working Paper
505 506 507
Ashraf, D., Goddard, J., 2012. Derivatives in the wake of disintermediation: a simultaneous equations model of commercial and industrial lending and the use of derivatives by US banks. Int. J. Banking, Accounting and Finance, 4(3), pp.250–271.
508 509 510
Barrell, R.E., Davis, P., Fic, T., Karim, D. 2011. The Future of Financial Regulations–is There a Link from Bank Size to Risk. Available at http://niesr.ac.uk/sites/default/files/publications/dp367.pdf
511 512
Barry, T.A., Lepetit, L., Tarazi, A., 2011. Ownership structure and risk in publicly held and privately owned banks. Journal of Banking and Finance, 35(5), pp.1327–1340.
513 514 515
Berger, A., Clarke, G., Cull, R., 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, 29(8-9), pp.2179–2221.
516 517
Benson, K., Faff, R., Smith, T. (2014). Fifty years of finance research in the Asia Pacific Basin. Accounting & Finance, 54(2), 335-363
518 519
Benson, K., Clarkson, P. M., Smith, T., Tutticci, I. (2015). A review of accounting research in the Asia Pacific region. Australian Journal of Management, 40(1), 36-88.
520 521 522
Berger, A., Davies, S., Flannery, M., 2000. Comparing market and supervisory assessments of bank performance: who knows what when? Journal of Money, Credit and Banking, 32(3), pp.641– 667.
523 524 525
Bouvatier, V., Lepetit, L., Strobel, F., 2013. Bank income smoothing , ownership concentration and the regulatory environment. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2133345.
526 527 528
Boyd, J.H., De Nicoló, G., Jalal, A.M., 2006. Bank Risk-Taking and Competition Revisited : New Theory and New Evidence. Available at: http://www.imf.org/external/pubs/ft/wp/2006/wp06297.pdf. 43
529 530
Boyd, J.H., Runkle, D.E., 1993. Size and performance of banking firms: Testing the predictions of theory. Journal of Monetary Economics, 31, pp.47–67.
531 532
Bushman, R.M., Williams, C.D., 2012. Accounting discretion, loan loss provisioning, and discipline of Banks’ risk-taking. Journal of Accounting and Economics, 54(1), pp.1–18.
533 534
Calem, P., Rob, R., 1999. The impact of capital-based regulation on bank risk-taking. Journal of Financial Intermediation, 8, pp.317–352.
535 536
Casu, B., Molyneux, P., 2003. A comparative study of efficiency in European banking. Applied Economics, 35(17), pp.1865–1876.
537 538 539
Chalermchatvichien, P., Jumreornvong, S., Jiraporn, P. (2014b). Basel III, capital stability, risktaking, ownership: Evidence from Asia. Journal of Multinational Financial Management, 28, 28– 46.
540 541 542
Chalermchatvichien, P., Jumreornvong, S., Jiraporn, P., & Singh, M. (2014a). The Effect of Bank Ownership Concentration on Capital Adequacy, Liquidity, and Capital Stability. Journal of Financial Services Research, 45, 219–240.
543 544
Chou, S., Lin, F., 2011. Bank’s risk-taking and ownership structure - evidence for economics in transition stage. Applied Economics, 43(12), pp.1551–1564.
545 546
Coles, J., Daniel, N., Naveen, L., 2006. Managerial incentives and risk-taking. Journal of Financial Economics, 79(2), pp.431–468.
547 548
Collins, J.H., Shackelford, D.A., Wahlen, J.M., 1995. Bank Differences in the Coordination of Regulatory Capital, Earnings, and Taxes. Journal of Accounting Research, 33(2), pp.263–291.
549 550
Cyree, K.B., Huang, P., Lindley, J.T., 2011. The Economic Consequences of Banks’ Derivatives Use in Good Times and Bad Times. Journal of Financial Services Research, 41(3), pp.121–144.
551 552 553
Daly, M., Krainer, J., Lopez, J., 2004. Does regional economic performance affect bank conditions? New analysis of an old question. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=530782 [Accessed April 18, 2013].
554 555
De Haan, J., Poghosyan, T., 2012. Size and earnings volatility of US bank holding companies. Journal of Banking and Finance, 36(11), pp.3008–3016.
556 557
De Nicolo, G., 2000. Size, charter value and risk in banking: An international perspective. Available at: http://www.federalreserve.gov/pubs/IFDP/2000/689/ifdp689.pdf.
558 559
Demirgüç-Kunt, A., Huizinga, H., 2010. Bank activity and funding strategies: The impact on risk and returns. Journal of Financial Economics, 98(3), pp.626–650.
560 561 562
Demsetz, R.S., Saidenberg, M.R., Strahan, P.E., 1997. Agency Problems and Risk-taking At Banks. , p.Federal Reserve Bank of New York Staff Report No. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=943507.
563 564
Forssbæck, J., 2011. Ownership structure, market discipline, and banks’ risk-taking incentives under deposit insurance. Journal of Banking and Finance, 35(10), pp.2666–2678.
565 566
Furlong, F., Keeley, M., 1989. Capital regulation and bank risk-taking: A note. Journal of Banking and Finance, 13, pp.883–891.
567 568
Gorton, G., Rosen, R., 1995. Corporate control, portfolio choice, and the decline of banking. The Journal of Finance, 50(5), pp.1377–1420.
569 570 571
Hassan, M.K., Sanchez, B., Safa, M.F., 2013. Impact of financial liberalization and foreign bank entry on Islamic banking performance. International Journal of Islamic and Middle Eastern Finance and Management, 6(1), pp.7–42.
572 573
Haw, I.-M. et al., 2010. Concentrated control, institutions, and banking sector: An international study. Journal of Banking and Finance, 34(3), pp.485–497.
574 575
Hesse, H., Cihák, M., 2007. Cooperative banks and financial stability. Available at: http://www.imf.org/external/pubs/ft/wp/2007/wp0702.pdf [Accessed July 27, 2013].
576 577
Hossain, M., Jain, P.K., Mitra, S., 2013. State ownership and bank equity in the Asia-Pacific region. Pacific-Basin Finance Journal, 21(1), pp.914–931.
578 579 580
International Monetary Fund, 2012. Economic prospects and policy challenges for the GCC countries, Available at: http://www.imf.org/external/pubs/ft/dp/2012/mcd1012.pdf. [Accessed July 27, 2013].
581 582
Jensen, M., Meckling, W., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), pp.305–360.
583 584
Jia, C., 2009. The effect of ownership on the prudential behavior of banks – The case of China. Journal of Banking and Finance, 33(1), pp.77–87.
585 586
Jiménez, G., Lopez, J. a., Saurina, J., 2013. How does competition affect bank risk-taking? Journal of Financial Stability, 9(2), pp.185–195.
587 588
Jokipii, T., Milne, A., 2011. Bank capital buffer and risk adjustment decisions. Journal of Financial Stability, 7(3), pp.165–178.
589 590
Kahane, Y., 1977. Capital adequacy and the regulation of financial intermediaries. Journal of Banking and Finance, 1(1), pp.207–218.
591 592
Koehn, M., Santomero, A., 1980. Regulation of bank capital and portfolio risk. The Journal of Finance, 35(5), pp.1235–1244.
593 594
Konishi, M., Yasuda, Y., 2004. Factors affecting bank risk-taking: Evidence from Japan. Journal of Banking and Finance, 28(1), pp.215–232.
595 596
Laeven, L., Levine, R., 2009. Bank governance, regulation and risk-taking. Journal of Financial Economics, 93(2), pp.259–275.
597 598
Lee, S.W., 2002. Insider Ownership and Risk-taking Behaviour at Bank Holding Companies. Journal of Business Finance and Accounting, 29(7 & 8), pp.989–1005.
599 600
Micco, A., Panizza, U., Yañez, M., 2007. Bank ownership and performance. Does politics matter? Journal of Banking and Finance, 31(1), pp.219–241.
601 602
Milne, A., Whalley, a. E., 1999. Bank Capital and Risk-taking. SSRN Electronic Journal. Available at: http://www.ssrn.com/abstract=147550.
603 604 605
Morales, R., Ramady, M., Brennan, C., 2012. GCC Sovereign Wealth Funds: challenges, opportunities and issues arising from their growing presence on the global landscape, in Ramady, M.A. (ed) The GCC Economies: Stepping up to future Challenges, Springer, N.Y.
606 607
Murphy, K., Jensen, M., 1990. Performance pay and top management incentives. The Journal of Political Economy, 98(2), pp.225–264. 45
608 609
Naceur, S. Ben, Omran, M., 2011. The effects of bank regulations, competition, and financial reforms on banks’ performance. Emerging Markets Review, 12(1), pp.1–20.
610 611
Pilloff, S.J., Rhoades, S.A., 2002. Structure and Profitability in Banking Markets. Review of Industrial organization, 20, pp.81–98.
612 613
Ramady, M.A. (2010). The Saudi Arabian Economy: Policies, Achievements and Challenges, Springer. New York.
614 615
Saunders, A., Strock, E., Travlos, N.G., 1990. Ownership structure, deregulation, and bank risktaking. The Journal of Finance, 45(2), pp.643–654.
616 617
Schaeck, K., Cihak, M., Maechler, A., Stoltz, S., 2012. Who disciplines bank managers? Review of Finance, 16, pp. 197-243.
618 619
Sharpe, W., 1978. Bank capital adequacy, deposit insurance and security values. Journal of financial and quantitative analysis, 13(4), pp.701–708.
620 621
Stiroh, K., 2004. Diversification in banking: Is noninterest income the answer? Journal of Money, Credit and Banking, 36(5), pp.853–882.
622 623
Stiroh, K., Rumble, A., 2006. The dark side of diversification: The case of US financial holding companies. Journal of Banking and Finance, 30(8), pp.2131–2161.
624 625
Sullivan, R.J., Spong, K.R., 2007. Manager wealth concentration, ownership structure, and risk in commercial banks. Journal of Financial Intermediation, 16(2), pp.229–248.
626 627
World Bank, 2013. Global financial development database, http://data.worldbank.org/data-catalog/global-financial-development
628 629
Yeyati, E.L., Micco, A., 2007. Concentration and foreign penetration in Latin American banking sectors: Impact on competition and risk. Journal of Banking and Finance, 31(6), pp.1633–1647.
630 631
Yildirim, S.H., Philippatos, G.C., 2007. Efficiency of Banks: Recent Evidence from the Transition Economies of Europe, 1993–2000. The European Journal of Finance, 13(2), pp.123–143.
available
at
Figure 1: Comparison of aggregate GDP growth rate between the World and the GCC Region for the period 2000-2012 10 8 6 4 2 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 -2 -4 WORLD (AGG.)
Source: EIU Country Data
GCC (AGG.)
Table 1: GCC savings ratio 2004-2012 Country
2004
2005
2006
2007
2008
2009
2010
2011
2012
Saudi Arabia
34.6
47.2
46.6
45.9
50.7
31.7
38.6
46.9
45.3
Bahrain
27.7
35.4
32.6
36.0
32.7
22.8
24.1
25.0
25.0
Qatar
56.7
66.9
66.1
59.8
56.4
51.6
51.8
60.1
57.0
UAE
29.8
34.9
36.0
31.7
31.8
32.5
36.1
40.9
40.6
Oman
31.6
37.9
41.6
41.7
42.1
24.7
39.0
38.4
39.7
Kuwait
45.5
50.4
60.6
56.5
58.7
38.3
41.8
48.5
45.2
Source: GCC Central Banks, HSBC, 2012, p. 22
27
Table 2: GCC support mechanism during the crisis period: 2008-2010 Central bank liquidity support
Long-term Govt. deposits
Oman
Qatar
Country
Deposit guarantees*
Bahrain Kuwait
Saudi Arabia
UAE
Capital injections
Bank asset purchases
Monetary easing
Source: Ahmed, 2010, GCC Central Bank. *
Stock market purchases
includes expansion of retail deposit insurance and guarantee of wholesale liabilities
Table 3: GCC Financial Institutions Survey: Significant shareholding ownership holding (over 15%) Country
Public shareholding
Government shareholding
Individual shareholding
Institutional ownership Local
Foreign
Total
Saudi Arabia
3
5
3
4
4
13
Qatar
3
7
-
5
2
13
Kuwait
5
9
2
20
2
30
Oman
1
4
3
5
4
8
UAE
3
16
5
13
4
33
Bahrain
4
9
2
22
11
37
TOTAL
19
50
15
69
27
134
Source: Authors’ own compilation from Bankscope, local stock markets, and annual reports of individual banks Notes: The total number of financial institutions surveyed was 134. Some institutions had significant shareholding in more than one category.
29
Table 4: Descriptive statistics of dependent and explanatory variables
Pneumonic Bank variables Z_score
Overall sample (1) Standard Mean deviation
Saudi Arabia (2) Standard Mean deviation
Qatar (3) Standard Mean deviation
Kuwait (4) Standard Mean deviation
Bahrain (5) Standard Mean deviation
Mean
UAE (6) Standard deviation
Oman (7) Standard Mean deviation
8.020 2.393 1.527 71.099 14.824 42.356
14.456 4.401 1.006 17.813 1.674 27.625
9.004 0.727 1.216 62.695 16.529 28.307
12.636 0.982 0.830 10.038 0.878 13.417
13.664 0.732 0.873 74.088 15.035 31.539
26.932 0.754 1.228 17.846 1.453 16.917
3.898 4.573 1.905 74.970 14.501 56.668
6.057 5.109 1.000 16.287 1.682 32.952
5.093 4.074 1.779 77.790 14.166 49.582
7.555 6.705 0.983 18.918 1.830 32.217
10.235 1.460 1.363 67.505 14.800 39.013
16.448 2.877 0.928 17.996 1.515 22.539
9.650 0.738 1.667 65.032 14.533 30.141
13.451 0.976 0.585 18.160 0.953 13.961
RWCit 39.305 Macroeconomic variables GDPjt 10.219 CONTROLjt 0.608 Shareholdings variables DIRi 0.091 Public 43.791 Government 21.555 Individual/family 4.392 Institutions 29.947
26.648
25.949
19.633
28.974
20.235
53.807
27.155
40.915
26.178
39.887
26.253
26.124
22.471
0.576 0.256
9.480 0.752
0.278 0.000
10.906 0.011
0.388 0.000
10.501 0.815
0.408 0.000
9.788 0.632
0.149 0.000
10.628 0.696
0.140 0.000
9.538 0.116
0.379 0.000
0.288 29.161 29.218 13.014 33.121
0.109 47.459 25.775 7.009 19.757
0.312 20.458 26.500 17.335 22.667
0.000 58.936 18.048 2.273 20.743
0.000 28.704 14.669 9.904 35.777
0.038 51.254 14.141 2.029 31.300
0.192 31.299 27.358 7.902 30.805
0.093 35.509 22.321 1.751 40.419
0.290 31.688 34.044 6.239 38.618
0.138 37.211 29.917 7.542 25.195
0.346 26.070 32.105 18.021 31.289
0.146 44.701 7.527 5.224 42.548
0.356 25.884 8.828 7.697 29.193
( ROAit )
LLOSSit SHAREit SIZEit NONIIit
Notes: This table provides the descriptive statistics for all banks in (1) GCC region, (2) Saudi Arabia, (3) Qatar, (4) Kuwait, (5) Bahrain, (6) UAE, and (7) Oman. Mean and Standard deviation are time-series-cross-sectional mean and standard deviation for the period 2000-2011 for all variables except for Z-score, σ(RAOit) and Shareholding variables. Shareholding variables are cross-sectional mean and standard deviation over 2011 while Z-score, σ(RAOit) are time-series-cross-sectional over the period 2002-2011 due to use of 3 years’ rolling window for computing these variables. Definition of variables is provided in Appendix A-III. Public, government, individual/family and institutions represent the descriptive statistics based on proportionate ownership of each category and thus are different from dummy variables of the same as defined in Appendix A-III.
Table 5: Correlation Matrix Z-score SHAREit
-0.101
SHAREit
SIZEit
NONIIit
LLOSSit
RWCit
GDPjt
0.054*
-0.173***
NONIIit
-0.208***
0.102***
-0.321***
LLOSSit
-0.181
***
**
-0.348***
0.190***
RWCit
-0.140***
0.069***
-0.599***
0.304***
0.217***
GDPjt
0.006
0.055
-0.007
0.100***
-0.193***
0.089***
-0.134***
0.007
0.029
0.197***
0.152***
0.219***
HHIjt
0.112
***
0.071
*
GFC
-0.129***
0.037
0.078
0.008
-0.098***
GFC_SIZE
-0.123***
0.023
0.137***
-0.018
0.000
-0.130***
0.011
-0.071**
0.029
-0.100
***
-0.042
*
0.070
**
SIZEit
CONTROLjt
DIRi PUB
CONTROLjt
HHIjt
GFC
GFC_SIZE
DIRi
PUB
GOV
IND
***
0.061
-0.106
0.348
***
***
-0.095
***
-0.057 -0.122
-0.734***
-0.046
0.150***
0.017
-0.170***
-0.121***
-0.083***
0.149***
0.014
-0.167***
0.992***
0.065*
-0.044
-0.054*
0.027
-0.049
0.049
***
-0.057
-0.105
0.093 *
***
-0.037 -0.170***
0.063
*
***
-0.170
***
0.058
*
-0.089
***
0.089
**
0.001
0.124
-0.004
0.001
***
-0.022
-0.022
-0.021
**
-0.033
-0.015
-0.180***
-0.329***
GOV
0.024
-0.069
IND
0.004
-0.217***
0.089**
-0.111***
-0.015
-0.057*
0.020
0.027
-0.098***
-0.006
0.003
0.752***
-0.124***
-0.147***
INST
-0.050
0.146***
-0.293***
0.191***
0.023
0.078*
-0.069**
-0.018
-0.007
0.048
0.029
-0.115***
-0.529***
-0.529***
-0.151***
Notes: This table presents the pair-wise correlation matrix. Definition of variables is provided in Appendix A-III. PUB, GOV, IND and INST are based on percentage of ownership of general public, government, individual/family and institutions respectively not as dummy variables as defined in Appendix A-III in the capital structure of banks in the GCC region at end of year 2011. Superscript *, **, and *** indicate statistical significance at 10, 5 and 1 percent level respectively. .
31
Table 6: GLS random effect estimation results for bank risk-taking as measured by Z-Score Model 1
Model 2
Model 3
Model 4
Model 5
Z-score
Z-score
Z-score
Z-score
Z-score
-0.01964 (0.00613)*** -0.02108 (0.00385)*** -0.64175 (0.25082)**
-0.01950 (0.00569)** -0.01265 (0.00435)** -0.65448 (0.23255)** 0.00020 (0.00016) -0.00375 (0.00359) 0.00095 (0.00019)**
-0.01954 (0.00557)*** -0.02429 (0.00543)*** -0.63586 (0.22814)*** 0.00012 (0.00016) -0.00205 (0.00359) 0.00094 (0.00018)*** 0.03208 (0.01225)*** -1.32572 (0.40930)*** -0.19445 (0.09394)**
-0.01956 (0.00543)*** -0.02475 (0.00554)*** -0.63620 (0.22231)*** 0.00012 (0.00016) -0.00215 (0.00362) 0.00094 (0.00019)*** 0.03176 (0.01232)*** -1.32358 (0.39886)*** -0.18493 (0.09522)* 0.03222 (0.05429) -0.00182 (0.00351)
2.42387 (0.45875)*** 9.85*** 734.84***
2.25073 (0.42804)** 10.62*** 670.96***
2.91221 (0.49261)*** 14.50** 544.25***
2.92009 (0.48114)*** 17.35** 547.43***
Dependent Variable SHAREit SIZEit ISLi NONIIit LLOSSit RWCit GDPjt CONTROLj HHIjt GFC GFC_SIZE DIRi PUBi INSTi GOVi INDi Constant Wald χ2 LM statistics
-0.01730 (0.00593)*** -0.02469 (0.00554)*** -0.62940 (0.23459)*** 0.00012 (0.00016) -0.00213 (0.00361) 0.00094 (0.00019)*** 0.03158 (0.01232)** -1.36659 (0.41721)*** -0.18495 (0.09520)* 0.03208 (0.05427) -0.00181 (0.00351) -0.03025 (0.46040) 0.19857 (0.30186) 0.28240 (0.39053) 0.25004 (0.34098) 1.00737 (0.64867) 2.55678 (0.60266)*** 12.00 503.67***
Notes: This table presents regression estimates of Z-score as an overall measure for bank risk-taking on ownership concentration, income diversification, and several other bank specific and macroeconomic variables. Definition of each variable is provided in Appendix AIII. Wald χ2 test reports the test for the joint restriction of the model parameters. The LM statistics reports the Breusch-Pegan test results for random effects. Standard errors are reported in parentheses Superscript *, **, and *** indicate statistical significance at 10, 5 and 1 percent level respectively
Table 7: Robustness check: Two stage least square (2SLS) estimation results Model 1
Model 2
Model 3
Z-score
Z-score
Z-score
-0.01673 (0.00572)*** -0.00368 (0.00256) -0.23400 (0.23310) 0.00053 (0.00018)*** -0.00348 (0.00271) 0.00035 (0.00013)***
-0.01691 (0.00581)*** -0.00776 (0.00295)*** -0.24816 (0.23669) 0.00034 (0.00017)** -0.00211 (0.00260) 0.00037 (0.00012)*** 0.00581 (0.00568) -1.08321 (0.38214)*** -0.21471 (0.04156)*** 0.00904 (0.03175) -0.00084 (0.00198)
Dependent variable SHAREit
-0.01139 (0.00640)* SIZEit -0.00773 (0.00293)*** NONIIit -0.29391 (0.25282) ISLi 0.00034 (0.00017)** LLOSSit -0.00209 (0.00259) RWCit 0.00037 (0.00012)*** GDPjt 0.00574 (0.00566) CONTROLj -1.04492 (0.40285)*** HHIjt -0.21475 (0.04136)*** GFC 0.00910 (0.03160) GFC_SIZE -0.00084 (0.00197) DIRi 0.13711 (0.46045) INSTi 0.32468 (0.43537) INDi 0.96342 (0.59754) GOVi 0.48180 (0.33045) PUBi 0.55095 (0.30206)* Constant 1.92730 2.62037 1.75642 (0.40421)*** (0.47172)*** (0.62517)*** Hansen J statistic 7.092 4.219 3.677 (0.2139) (0.2387) (0.2985) Notes: This table presents the estimation results based on the 2SLS regression model of the association between Z-Score as a measure of overall risk-taking behavior of banks within GCC region and ownership concentration, income diversification, and several other bank specific and macroeconomic variables. Definition of each variable is provided in Appendix A-III. Hansen J-statistics reports test statistics of Sargan-Hensen test. p-value of Hansen J-statistics are reported in the last row within parenthesis. Excluded instruments in the first stage of G2SLS model include the ratio of interest income to revenue (INT2REVit), the ratio of net-charge-offs to gross loans (NCAGit) and the ratio of total equity to total assets (LEVGit) and a dummy variable FOREIGNi equal to one if a foreign bank is a shareholder. Standard errors are reported in parentheses. Superscript *, **, and *** indicate statistical significance at 10, 5 and 1 percent level respectively
33
Appendix A-I: Survey of empirical literature on association between the ownership structure and risk-taking behavior of banks. Study Saunders et al. (1990) Gorton and Rosen (1995)
Major interest Managerial ownership and risktaking. Portfolio risk and ownership structure
Sample
Ownership indicators
Risk /return indicators
38 listed US BHCs from 1978 to 1985.
Percentage of managerial ownership Percentage of managerial ownership and sum of the block shareholders >5% Percentage of managerial ownership and sum of the block shareholders >5%
Interest rate risk, total return risk, nonsystematic risk and market risk non-performing loan ( NPL)
Demsetz et al. (1997)
Risk-taking, franchise value, and ownership structure
US commercial banks from 1984 to 1990 350 listed US BHCs between 1991 and 1995 period.
Anderson and Fraser (2000)
Managerial shareholdings and risk-taking.
150 US banks between 1987 and 1994 period.
Percentage of managerial ownership
Total risk (standard deviation of daily stock returns), systematic risk and firm specific risk.
Lee (2002)
Managerial ownership and risktaking
65 listed US BHCs from 1987 to 1996
Percentage of managerial ownership
Konishi and Yasuda (2004)
Capital regulations, government official in the management and shareholders and risk-taking
48 regional banks listed on Tokyo Stock Exchange
Percentage of ownership data in terms of size
Total risk (Standard deviation of daily stock returns), systematic risk and firm specific risk based on the size of banks Total risk (standard deviation of daily stock returns), systematic risk and firm specific risk.
Sullivan and Spong (2007)
Managerial ownership, monitoring of large stockholders and risk-taking Ownership structure, governance and risk taking. Type of ownership and risk-taking (prudential behavior)
304 US Statechartered banks from 1985 to 1994
Bank manager’s ratio of investment in bank’s shares to total wealth
Standard deviation of operating return on equity
270 banks from 48 countries for the period 1996-2001.
Largest shareholder > 10% of outstanding shares, board representation, and managerial ownership State-owned or joint-equity banks.
Z-score, asset return volatility, equity return volatility
Laeven and Levine (2009)
Jia (2009)
A sample of 14 Chinese banks from 1985 to 2004 (annual data)
Standard deviation of weekly stock returns.
Bank excess reserve ratio, loan to total assets and deposit to loan ratio
Findings
GCC
Higher management ownership is associated with more risk aversion. Manger-shareholders may engage in higher risk-taking, when are likely to be entrenched. Higher risk-taking is associated with higher managerial ownership and lower franchise value. Managerial shareholding is positively associated with the total risk and firm specific risk during relatively less stringent regulatory regime. Larger widely-held banks with lower market return volatility and lower balance sheet risk exhibits higher risk-taking.
No
The relationship between the ownership and risk taking is non-linear. It decrease as the proportion of ownership increase in the beginning and later increase Bank risk taking is inversely related to the managerial/large ownership. Bank manager exhibit higher risk taking as their ownership of bank increases. Risk taking increase with the proportionate ownership of the largest shareholder.
No
State-owned banks lending practices are less prudent as compared with joint equity banks.
No
No
No
No
No
No
No
Angkinand and Wihlborg (2010)
ownership of banks, deposit insurance system, and risktaking
Aggregate country level banking system data from 52 countries for the period 1997- 2003.
Proportion of banks at country level as foreign or government owned banks with ownership threshold of >50%
Z-score, NPL to average capital, and standard deviation of NPL to average capital
Haw et el. (2010)
Bank performance, cost efficiency, risk, and block-shares Market discipline, ownership of banks and risk-taking
325 listed banks from 22 countries from 1990 to 1996 331 publicly traded banks from 47 countries during 1995-2005
Voting right of controlling owners
Probability of insolvency (Zscore), standard deviation of ROE, Z-score, NPL to equity capital
Barry et al. (2011)
Ownership structures and risktaking
Bankscope database: categories of shareholders.
Z-score, loan loss provision to net loans, volatility of return on assets and equity, bank portfolio risk and leverage risk
Chou and Lin (2011)
Type of ownership concentration and risk-taking
249 commercial banks from 16 Western European countries between 1999 and 2005. 37 listed banks on Taiwan Exchange from 2001 to 2006
Chalermchatv ichien et al. (2013)
Ownership structure on the risk-taking behavior of banks
68 banks from 11 East Asian countries from 2005 to 2009
Ownership concentration largest shareholder: private, financial holding corporation (FHC), FHC with government as major shareholder, and government. Ownership concentration – largest shareholder in the ownership structure
Hossain et al. (2013)
Impact of government ownership on the market valuation of banks Government ownership and bank risk-taking
Forssbæck (2011)
Reuter’s database: insider/stakeholder, institutions and mutual funds.
Banking systems with a larger state-ownership experience higher defaults, and lower stability as compared banking system with higher proportion of foreign ownership. Banks controlled by larger shareholders exhibit higher risktaking. Banks with larger shareholders exhibit lower risk-taking measured by the default risk. However Z-score as measure of risk is statistically insignificant. Concentration of ownership is not associated with the level of risk-taking
No
6-months and 3-months overdue loan rate, regulatory capital
Banks with foreign institutions as largest shareholders exhibit higher risk taking in terms of higher overdue loans and/or lower regulatory capital.
No
Capital adequacy ratio, current ratio, loan to deposit
Banks with larger shareholders are better capitalized and are more liquid. Concentrated ownership enhances the capital stability. Higher state ownership mitigate the investment risk among the common stock of banking industry.
no
International sample NPL to total assets, and risk of 2467 banks from weighted capital ratio 107 countries during the period from 1990 to 2009 Iannotta et al. European banks Bankscope ownership Issuers’ rating, Z-score Government owned banks have (2013) with total assets > database lower default risk but higher €10 billion from operating risk 2000 to 2009 Notes: The last column labeled as ‘GCC’ indicates if the sample of empirical study contains banking sector of GCC countries individually or as a region.
35
No
No
No
No
No
Appendix A-II: GCC Financial Markets Regulatory/ Supervisory Settings Oversight mechanism
Information for consideration 1.Banking Supervision: a) Basel II b) Basel III c) BCBS (Basel Committee on Banking Supervision)
Regulatory Oversight and Control
2. Equity participation by local banks 3. Bank Verification, examination & inspection 4. Mergers and Acquisitions 5. Related party approval exposures limits 6. Stress testing 7. Consumer lending limits 8. Liquidity reserve requirements 9. Asset classification provisioning and NPL’s 10. Senior management screening & approval 11. New product launch approval 12. Auditing and accounting requirement (IASB Int. Fin. Reporting Standards) 13. Restricting bank activities 14. Capital flows control
Central Bank Support
15. Foreign Exchange/ Swaps 16. Affiliated training center 17. Explicit CB guarantees (Too big to fail) moral hazard 18. Central Bank liquidity support
Stress Testing
19. Tools to measure systemic risk-VAR 20. Eliminate practices promoting excessive risk taking (compensation practices) 21. Capital surcharge for systemically important firms
Corporate Governance
22. Cross-border bank resolution regime 22. Unified Financial Markets Regulator 23. Corporate Governance Mena – OECD initiative on Governance and Investment for Development 24. Financial crime:
Oversight mechanism
Information for consideration a) FATF-MENA (Financial Action Task Force) b) OECD
International Liquidity Benchmarking
25. Securities markets (IOSCO- International Organizations of Securities Commission) 26. International Organizations: a) G20 b) IMF c) World Bank d) UNCTAD
Financial Sector Soundness Indicators
International Surveillance
e) BIS (Bank for International Settlement) 27. Official Surveillance: a) IMF-World Bank
FSSA(Fin. Sector Stability Assessment)
b) IOSCO – FSAP (Fin. Sector Stability Program) 28. Financial soundness Indicators a) Non-performing loans (% of gross loans) b) Provisioning Rate (% of non-performing loans)
Appendix A-III: Variable definition Variable Z-score
LLOSSit
( ROAit )
RIGHTi
SIZEit ISLi NONIIit RWCit GDPjt (HHIjt) CONTROLj
DIRi INDi INSTi GOVi PUBi GFC GFC_SIZE
Definition The log of the ratio of sum of the return on assets and the book value of equity to capital ratio to the standard deviation of the return on assets, and computed as moving average over the three most recent years. The logistic transformation of the ratio of bank i’s loan loss reserve to gross loans The ratio of net income to total assets computed by taking the most recent three years’ rolling standard deviation The proportion of highest shareholding in the ownership structure of bank i as per type of ownership: government, general public (widely held banks), individual/family and institutional/corporate investors by the end of 2011. The natural logarithm of total assets of bank, i in year t. A dummy variable equal to one if bank i is an Islamic bank, and zero otherwise. The ratio of non-interest income to gross income for banks i at time t. The ratio of sum of risk-weighted assets to total assets. The natural log of the per capita GDP income of country j in year t. The Hirchman Herfindahl Index and is the sum of the squares of the market shares of all banks based on total assets of each bank existing in country j in year t. Regulatory stringency index based on authors’ own assessment of six major regulatory factors: regulatory capital, central bank regulatory oversight and support, regulatory controls – stress testing, corporate governance, international liquidity benchmarking and international surveillance - capital control. A dummy variable equal to one if a shareholder with five percent or more ownership stake has the representation on the board of directors or zero otherwise. A dummy variable equal to one if majority shareholding of bank i is with individual/family, and zero otherwise. A dummy variable equal to one if majority shareholding of bank i is with institutional investors, and zero otherwise. A dummy variable equal to one if majority shareholding of bank i is with government, and zero otherwise. A dummy variable equal to one if majority shareholding of bank i is with widely held, and zero otherwise. A dummy variable equal to one if years 2008 and 2009, zero otherwise. An interaction term of GFC with SIZE
3
