The current issue and full text archive of this journal is available at

www.emeraldinsight.com/1086-7376.htm

SEF
31,1

106

The use of discretionary loan loss

provisions by Islamic banks and
conventional banks in the Middle
East region
A comparative study
Hakim Ben Othman and Hounaida Mersni
Accounting and Finance Department, Tunis Business School,
University of Tunis, Tunis, Tunisia and
LIGUE-ISCAE, University of Manouba, Tunis, Tunisia
Abstract
Purpose The purpose of this paper is to study earnings management practices of Islamic banks
and conventional banks in the Middle East region. First, the authors examine factors that may
influence Islamic banks managers use of discretion in reporting loan loss provisions (LLP). Second,
the authors investigate differences that may exist between Islamic banks and non-Islamic banks in
terms of discretionary loan loss provisions (DLLP) used to manipulate accounting earnings.
Design/methodology/approach This empirical study uses an unbalanced panel data of
21 Islamic banks, 18 conventional banks with Islamic windows and 33 conventional banks, from seven
Middle East countries during a period that ranges from 2000 to 2008. The authors use a two-stage
approach in order to examine factors that may influence the use of discretion by Islamic banks
managers.
Findings The empirical results reveal that Islamic banks use DLLP for both earnings and capital
management. External financing is also found to be a determinant of DLLP. Additional findings show
no significant differences among Islamic banks, conventional banks with Islamic windows and
conventional banks in using DLLP. These three groups of banks behave similarly in terms of
discretion based on DLLP.
Practical implications The findings are potentially useful for regulators, auditors and investors.
This study provides regulators with insights to strengthen their financial regulations in order to
improve accounting quality. In addition, it helps auditors when considering the provisioning policies
adopted by banks in order to detect specific manipulations of accounting earnings. The results may
also help investors to focus on the impact of managerial discretion on accounting earnings for
evaluation purposes.
Originality/value This study contributes to the literature on Islamic banking. On the one hand,
it extends prior research by examining the discretionary component of LLP, instead of being restricted
to total LLP. On the other hand, it compares the use of discretion among three groups of banks: full
Islamic banks, conventional banks with Islamic windows and full conventional banks.
Keywords Earnings management, Islamic banks, Conventional banks,
Conventional banks with Islamic windows, Discretionary loan loss provisions
Paper type Research paper

Studies in Economics and Finance

Vol. 31 No. 1, 2014
pp. 106-128
q Emerald Group Publishing Limited
1086-7376
DOI 10.1108/SEF-02-2013-0017

JEL classification G34, M40, M41, M42

1. Introduction
A wide literature has addressed the issue of earnings management in the banking
industry. It is shown that banks around the world are found to manage their earnings
and banks managers are motivated to minimize the earnings volatility over time.
Current literature provides evidence that loan loss provisions (LLP) are used by banks
as an instrument for long-term earnings management (Collins et al., 1995; Ismail and
Be Lay, 2002; Anandarajan et al., 2005, 2007; Taktak et al., 2010a). Most existing
studies have concentrated on conventional banks. However, little attention has been
given to earnings management in Islamic banks in the Middle East region.
Recent studies such as Zoubi and Al-Khazali (2007) and Taktak et al. (2010b) addressed
the issue of the use of LLP in Islamic banks. Zoubi and Al-Khazali (2007) argue that Islamic
banks use LLP to manage earnings while Taktak et al. (2010b) defend the idea that Islamic
banks do not use LLP to smooth their results. Although a few studies have provided some
evidence on the use of LLP by Islamic banks, the discretionary component of this variable
remained unexplored. Following Kanagaretnam et al. (2004) and Kwal et al. (2009), we
isolate, in this study, the discretionary component of loan loss provisions (DLLP) and we
use it as our dependent variable. The purpose of this paper is to examine factors that can
influence the Islamic banks managers use of DLLP. More specifically, we determine first,
whether managers use discretion to manage earnings and capital ratio in Islamic banks.
Second, we explore if there is any difference in the use of discretion between Islamic banks
and non-Islamic banks in the Middle East region.
Using panel data of 21 Islamic banks, 18 conventional banks with Islamic windows and
33 conventional banks pertaining to seven countries from the Middle East region
(i.e. Bahrain, Egypt, Jordan, Kuwait, Qatar, Saudi Arabia and UAE), we find that Islamic
banks use their discretion for earnings management. Indeed, we find that earnings before
tax and provisions ratio (EBTP) has influence on the discretion of Islamic banks proxied by
DLLP. In addition, results show a positive and significant relationship between capital
adequacy ratio (CAR) and DLLP, suggesting that managers in Islamic banks use discretion
to manage their capital ratio to enable banks to avoid violating minimum requirements.
Moreover, our findings reveal that Islamic banks and non-Islamic banks behave similarly in
terms of DLLP. In effect, there is no substantial difference in the use of DLLP by managers of
Islamic banks, conventional banks with Islamic windows and conventional banks.
Our study contributes to the earnings management literature in Islamic banking.
We extend prior work by focusing on the use of the DLLP rather than the total amount of
LLP. Furthermore, we compare the use of discretion between Islamic banks and their
non-Islamic counterparts. To our knowledge, this research is the first one that
distinguishes between conventional banks with Islamic windows and conventional banks.
Our findings are of interest to the interests of investors, standard setters and auditors.
This study helps investors to consider the impact of managerial discretion on
accounting earnings for evaluation purposes. Standard setters and regulators should be
able to provide adequate standards that regulate the estimation of LLP by banks. They
should require additional disclosure from banks in order to constrain discretionary
behavior of managers and to prevent from an aggressive earnings management.
Auditors have to focus more on the provisioning policies adopted by banks in order to
detect specific manipulations of accounting earnings.
The remainder of the paper is organized as follows. Section 2 presents the
specificities of Islamic banks. Section 3 provides backgrounds for the study and

Use of DLLP by
Islamic banks

107

SEF
31,1

108

develops hypotheses. Section 4 introduces the sample and describes our research
design. Section 5 presents descriptive statistics and reports results of our analysis based
on our DLLP panel-regression model. Section 6 concludes the paper.
2. Specificities of Islamic banks
2.1 Islamic banks principles
Islamic banks are based on the sharia law, which derives from the interpretation of the
Quran and Sunnah and from other secondary sources of Islamic law, such as Ijma
(unanimous agreement among sharia scholars about specific issues not envisaged by the
Quran or the Sunnah), Qiyas (the use of deduction by analogy) and Ijtihad
(personal reasoning) (Al-Gamal, 2006). The adoption of sharia leads Islamic banks to
focus more on ethical and moral values in their banking industry rather than credit value
(Hamdi and Zarai, 2012). The code of conduct of Islamic banks requires financing legal
activities which are sharia compliant. Certain commodities such as alcohol, illegal drugs
and illegal arm dealing are not allowed by the sharia law. Islamic banking, must avoid
also uncertainty (Gharar) and gambling (Mayser). Whence, any contract based on risky
or hazardous events and generated an easy profit, without breaking sweat, is not
allowed (Al-Gamal, 2006). The distinguishing feature of Islamic banks is the prohibition
of interest. According to Islamic law, money should not bread money, so both the
charging and the receiving of interest are strictly forbidden (Taktak et al., 2010b;
Taktak, 2011; Hamdi and Zarai, 2012). The profit sharing concept is used as an
alternative to the interest based banking. Consequently, Islamic banks become partners
and they share risk both with depositors and shareholders (Taktak et al., 2010b;
Farouk et al., 2012). The banning of interest and the activating of profit sharing principle
make the investment approach adopted by Islamic banks unique and different from
conventional banks (Taktak et al., 2010b; Hamdi and Zarai, 2012). Unlike conventional
banks, Islamic banks have only one kind of loan and that is Qard-el-hassan (literally
good loan). This type of loan is a profit-free loan for which the bank does not charge any
interest or additional amount over the money lent (Al-Gamal, 2006).
The main products offered by Islamic banks based on profit sharing principles are
Mudharaba and Musharaka contracts. The Mudharaba contract is a type of a
partnership, in which one partner provides the capital and the other provides expertise
and management. If the project generates profit, each partner gets a pre-arranged
percentage of profit. If there are losses, the capital provider bears all financial losses
while the entrepreneur supports only, the operating costs of their own efforts. The
Musharaka contract requires that the bank and the depositor establish a joint
commercial enterprise, in which all partners contribute capital as well as labor and
management and share profit and loss in a pre-arranged way (Al-Gamal, 2006). With
this respect, the loan portfolio of Islamic banks is diversified and it includes
Qard Hasan, Musharaka, Mudharaba and many others financing techniques that are
sharia compliant (Turen, 1996; Zoubi and Al-Khazali, 2007).
2.2 Provisioning practices of Islamic banks
Conventional banks and Islamic banks follow substantially different provisioning
practices (Taktak et al., 2010b; Hanif, 2011; Farouk et al., 2012). The Accounting and
Auditing Organization for Islamic Financial Institutions (AAOIFI) requires the use of
dynamic provisioning, which is a macro-prudential tool used to reduce the procyclicity of

banks provisions and earnings and thus their probability of default (Wezel et al., 2012).
The fundamental principle underpinning dynamic provisioning is that a forward-looking
provision, based on long-run expected annual losses, is made each year.
The AAOIFIs standard (11), which sets out the accounting for provisions and
reserves, defines provisions as setting aside certain amount from income as expenses to
revaluate receivables, financing and investment assets, when the probability of
uncollectible amounts or assets impairment occur. This standard advocates the
recognition of two types of provision: general and specific provisions. The first one
should be recorded in order to cover potential losses that results from unidentifiable
risks related to assets, and the second one is recognized when the asset is impaired, in
order to reduce its amount to its net realizable value. Therefore, the LLP is recorded to
better anticipate the credit risk of the bank. These characteristics in provisioning policy
adopted by Islamic banks are more sophisticated than conventional banks in that it does
not take into consideration the actual loss only, but it considers the expected future
losses (Salman, 2004; Taktak et al., 2010b; Taktak, 2011; Quttainah, 2011).
However, it is noteworthy that, the AAOIFIs standards are not applied by all Islamic
banks in all countries[1]. It is not easy for these banks to adopt unique accounting
standards into their practice because of the absence of a legal framework (Sarea, 2012).
In this regards, Islamic banks must comply with the accounting standards applicable in
the country and that can be different from the AAOIFI standards. Accordingly, the
provision practice may differ across Islamic banks and countries. Therefore, it becomes
necessary to examine the effect of the use of AAOIFI standards on DLLP.
3. Background and hypothesis development
A growing body of empirical research provides evidence that banks manage their earnings
(Shen and Chih, 2005; Cornett et al., 2009; Wang et al., 2012). It is noteworthy that earnings
management in banks is more problematic than in other firms. This is due to the
importance of banks to national, regional, and global economy. Banks have an important
role for the economic growth, the stability and the welfare of the countries (Quttainah, 2011;
Hamdi and Zarai, 2012). As such, earnings manipulation can have harmful implications to
the whole economy, as visualized by the last financial depression that originated in the
banking sector. During the last financial crisis, the collapse of banks made it clear that
information asymmetry problems between managers and shareholders are very severe
(Palia and Porter, 2007). Earnings management is considered as a constraint for investors
to predict banks future performance accurately using the current financial information.
This practice increases information dissemination problems between banks and investors
and reduces banking sector stability (Quttainah, 2011; Hamdi and Zarai, 2012).
This information dispersal originates from the agency theory, which suggests that
managers do not act in the best interest of the shareholders, they exhibit tendencies to
divert from their duties and to pursue strategies that meet their own goals, rather than
those of the owners (Jensen and Meckling, 1976; Fama, 1980). Theses agency
problems occur when shareholders lack the necessary power to monitor and control the
managers that have an opportunistic behavior (Macey and OHara, 2003).
Prior work offers evidence that accounting accruals are affected by agency issues
and asymmetric information (Bae et al., 2009; Cornet et al., 2009). Managers in the
banking industry have incentives to smooth earnings via LLPs considered as the most
important accruals in the banking sector. These incentives are basically related to

Use of DLLP by
Islamic banks

109

SEF
31,1

110

earning management (Greenwalt and Sinkey, 1988; Beatty et al., 1995) and capital
management (Moyer, 1990; Collins et al., 1995; Kim and Kross, 1998). Most of these
studies focused on conventional banks and their findings reveal some mixed evidence.
A large body of empirical studies provides evidence that banks engage in earnings
management throughout LLP (Ma, 1988; Collins et al., 1995; Greenwalt and Sinkey,
1988; Bhat, 1996; Lobo and Yang, 2001; Kanagaretnam et al., 2004; Anandarajan et al.,
2005, 2007; Kwal et al., 2009; Pinho and Martins, 2009; Taktak et al., 2010a). In the same
vein, Wetmore and Brick (1994, p. 299) showed that bank managers take large LLPs in
a good year so that extra reserves are available for bad years. Conversely, some other
studies found no relationship between LLP and earning management (Wetmore and
Brick, 1994; Beatty et al., 1995; Ahmed et al., 1998; Ismail et al., 2005). Table I provides
summary of these studies.
As shown previously, recent investigations have given rise to a vivid interest in
empirical research related to earnings management in conventional banks. However,
studies related to Islamic banks are limited and reported mixed results. We start our
evidence by referring to the paper of Ismail and Shahimi (2003) that proved the use of
LLP by Islamic Malaysian banks for the purpose of manipulating their earnings and
capital management during the period 1997-2001. Similarly, Zoubi and Al-Khazali (2007)
examined 55 conventional banks and ten Islamic banks in the Gulf Cooperation Council
(GCC) region during the period 2000-2003. They outlined that managers of Islamic and
conventional banks in the GCC region use total LLP to smooth their results. They
asserted that both Islamic and conventional banks in the GCC countries follow similar
income smoothing practices. Quttainah (2011) examined 11 countries pertaining to
Economic Research Forum (ERF) (Arab countries, Iran and Turkey) during the period
1994-2008. He used loss avoidance and abnormal LLP as proxies of earnings
management to compare earnings management behavior between Islamic banks and
conventional banks. Findings show that Islamic banks are less likely to engage in
earnings management compared to their non-Islamic counterparts.
In the same vein, Misman and Ahmed (2011) compared Islamic and conventional banks
in Malaysia. They proved that both Islamic and conventional banks use LLP for their
earnings and capital management purpose. However, findings reveal significant
differences between Malaysian Islamic and conventional banks when managing their LLP.
Conversely, Taktak et al. (2010b) provided contrary evidence. Using a sample of
66 Islamic banks in Muslim countries over the period 2001-2006, they investigated
income smoothing practices by Islamic banks and examined the use of total LLP for this
purpose. Based on Beidlemans and Eckels coefficients, they pointed out an extensive
use of income smoothing by Islamic banks. However, they did not find evidence on the
use of total LLP They showed that, contrary to conventional banks, Islamic banks do not
smooth income via total LLP. Moreover, Taktak (2011) examined the nature of
smoothing returns practices in Islamic banks. She aimed to assess if the smoothing
mechanism is natural or intentional. Using a sample of 79 Islamic banks from
19 countries during the period that ranges from 2001 to 2006, she found that a larger
number of Islamic banks engage in natural Income smoothing. Furthermore, results do
not provide evidence that Islamic banks resort only to natural income smoothing.
Empirical literature on Islamic banks provides conflicting predictions and reveals
mixed results about the use of LLP for earnings management purpose. However, contrary
to conventional banks, prior studies do not examine the use of managerial discretion.

Authors

Results

Panel A: studies that examined association between LLP and earnings management in conventional
banks
Ma (1988)
Focusing on the income smoothing practices in US banks and using a
panel of 900 observations during the period 1980-1984, results prove
that US banks engaged in earnings management through provisioning
policy
Greenwalt and Sinkey (1988) Using a sample of 106 banks during the period 1976-1984, results
assert that US banks smooth their earnings using LLP. Thus, moneycenter banks are less likely to engage in income smoothing than
regional banking companies in the USA
Collins et al. (1995)
Considering the period 1971-1991, findings show that LLP is used as
an instrument for earning management while loan charge-off and
securities insurances are used for capital management
Bhat (1996)
Examining the income smoothing hypothesis for a sample of the 148
banks during the period 1981-1991, results reveal that US banks do not
use LLP for earning management
Lobo and Yang (2001)
Using a sample of US banks during the 1981-1996 period, and
analyzing the use of discretionary LLP, findings indicate strong
evidence for income smoothing, capital management and signaling
Ismail and Be Lay (2002)
Studying the case of 34 commercial banks in Malaysia across 19971999, and using a model of LLP which incorporates the sectorial effect
and the economic risk pertaining to those sectors, results outline that
conventional banks in Malaysia use LLP to manage their earnings
Kanagaretnam et al. (2004)
Based on 22,640 firm-year observations over the period 1992-2001,
findings provide evidence that US banks use DLLP to reduce earnings
volatility and to manage capital. Results also prove that bank
managers decisions to reduce earnings volatility are related to the
need for external financing and to securities gains and losses
Anandarajan et al. (2005)
Using a panel of 970 observations of depository institutions in Spain
during the 1986-1995 periods, empirical results assert the use of LLP
for capital and earnings management
Anandarajan et al. (2007)
Using a sample of 50 Australian commercial banks over the period
1991-2001, results prove that Australian banks use LLP for capital and
earnings management. However, there is no evidence on the use of LLP
for signaling future intentions of higher earnings to investors.
Further, listed banks are more likely to engage in earning management
than unlisted commercial banks
Kwal et al. (2009)
Based on a sample of 31 Japanese banks across the period 1996 to 1999,
findings indicate that DLLP are used extensively for earnings and
capital management. Results also show that DLLP are positively
related to the demand for external financing, realized securities gains
and prior years taxes
Pinho and Martins (2009)
Using a sample of 35 financial institutions operating in Portugal from
1990 to the end of 2000, the findings indicate that Portuguese banks
have a discretionary behaviour in setting up their provisions, and find
evidence of income-smoothing and capital management
Taktak et al. (2010a)
Using a sample of 278 commercial banks operating in OECD
countries. Results highlight that a large number of banks use
intentional smoothing results either by using LLP or by selling trading
securities
(continued)

Use of DLLP by
Islamic banks

111

Table I.
Summary of prior studies
that examined LLP

SEF
31,1

112

Table I.

Authors

Results

Panel B: studies that examined no association between LLP and earnings and capital management in
conventional banks
Ismail et al. (2005)
Based on a sample of 21 Malaysian banks during the period 1996-2002,
results indicate that banks in Malaysia do not use LLP to smooth
income
Wetmore and Brick (1994)
Considering 82 US banks across the period 1986-1990, results assess
that contrary to prior studies, there is no evidence of income smoothing
practices. Indeed, US banks do not use LLP for the earnings
management purpose
Beatty et al. (1995)
Considering the period 1987-1990, for a sample148 banks, findings
prove that LLP is not used as a tool to manage earnings in US banks
Ahmed et al. (1998)
Results indicate no evidence of earning management through LLP, it is
shown that LLP is used only for capital management

Unlike prior studies, that stick to evaluate total LLP, we follow Kanagaretnam et al. (2004)
and Kwal et al. (2009). We isolate the DLLP in order to investigate the factors that influence
managers use of discretion in estimating LLP.
A variety of studies provide evidence that managers use their discretion in
reporting LLP for both earnings and capital management purposes.
3.1 Earnings management
Empirical investigations about the relationship between DLLP and earnings
management asserted that managers are inclined to recognize provisions when
accounting earnings are high enough. Managers of banks with high earning variability
will have stronger incentives to smooth earnings through LLP (Lobo and Yang, 2001;
Pinho and Martins, 2009). Kwal et al. (2009) documented that managers, through loss
provisions, are able to shift earnings among periods to smooth income over time. The
EBTP is widely used in the literature to capture earnings management practices. Based
on the existing literature, we expect that managers use discretion to underestimate LLP
if the EBTP ratio is low, and overestimate LLP if the EBTP is high. Hence, the first
hypothesis:
H1. There is a positive relationship between DLLP and EBTP in Islamic banks.
3.2 Capital management
Following previous studies, the capital structure is measured by the CAR. Considering the
use of DLLP for capital management, prior literature supports either a positive and
negative relationship between DLLP or LLP and CAR. Some papers showed that banks
with low capital ratios are inclined to use their discretion and report low DLLP in order to
report higher capital and earnings (Kim and Kross, 1998; Ahmed et al., 1999). Other studies
suggested that well-capitalized banks are subject to a lower level of monitoring by
regulatory agencies (Kanagaretnam et al., 2004; Taktak et al., 2010b). Therefore, they are
able to use more discretion to boost earnings and capital. Consequently, we hypothesize
the following:
H2. There is a negative association between DLLP and CAR in Islamic banks.

The incentives of Islamic banks to manage accounting earnings via LLP are likely to
be influenced by bank-specific factors. We examine three factors that may affect the
earnings management behavior of Islamic banks managers.
3.3 External financing
Several studies considered external financing as an instrument to smooth reported
earnings. In fact, to attract external funds, a bank reports low LLP to reduce the
perceived risk and to increase reported income. Loan to deposit (LD) ratio is often used
as a proxy for external financing (Kanagaretnam et al., 2004; Zoubi and Al-Khazali,
2007). If LD ratio is high, this indicates that total loans are greater than deposits and,
therefore, banks need to attract more deposits from customers. For that, banks
managers have incentives to report low LLP. Thus, we expect that the degree of
earnings management through DLLP is negatively related to the demand for external
financing. This suggests the following hypothesis:
H3. There is a negative relationship between DLLP and LD ratio in Islamic banks.
3.4 Bank size
In the existing literature, it is often argued that bank size is considered as an important
factor that influences earnings management behavior. Following Zoubi and Al-Khazali
(2007), Taktak et al. (2010b) and Quttainah (2011) who are interested in Islamic
banking, we expect that the larger is the bank size, the higher is the DLLP. We state the
following hypothesis in the null form:
H4. There is a positive relationship between DLLP and bank size in Islamic banks.
3.5 Accounting standards
As stated above, the Islamic financial accounting standard (IFAS 11), which is related
to provisions and reserves, requires the use of dynamic provisioning. Subsequently,
Islamic banks are more inclined to set-up an allowance for loss provision to absorb any
future losses. Thus, the use of Islamic standards leaves little discretion to managers to
manipulate accounting earnings.
However, it is to be noted that, Islamic standards are not implemented in all
Middle East region. These standards (IFAS) are mandatory or recommended for a
limited number of countries. A dummy variable will be included in the model to control
for the use of Islamic accounting standard. Hence, we expect that managers are less
motivated to manage earnings through DLLP when IFASs are used by banks to
prepare their financial statements. More specifically we expect that the use of dynamic
provisioning reduces the DLLP. Our hypothesis is:
H5. There is a negative relationship between DLLP and AAOIFI.
4. Methodological approach
4.1 Sample selection
The sample selection process started with identifying Islamic banks and conventional
banks with Islamic windows in the Middle East region from the leading online business
intelligence platform focusing on the Middle East and North Africa, Zawiya. We found a
list of 200 Islamic banks from 12 countries. We intended first to include all these banks
and countries, but given that data are not available for all the banks, our sample covered

Use of DLLP by
Islamic banks

113

SEF
31,1

114

only 21 Islamic banks and 18 conventional banks with Islamic windows in

seven countries from the Middle East region: Bahrain, Egypt, Jordan, Kuwait, Qatar,
Saudi Arabia and UAE. Then we collected the list of conventional banks operating in
these countries, we selected 33 conventional banks.
Table II presents the sample selection. Data consists of annual year-end information
for Islamic and conventional banks in the Middle East during the period ranges from
2000 to 2008. We included all banks, for which complete data across this time period
was available. We have to note that, at the firm level, some data are missing for one or
several years. This constitutes an unbalanced panel study of the data sets including
519 total firm-year observations of which 129 firm-year observations represent Islamic
banks, 141 firm-year observations represent conventional banks providing Islamic
services (with Islamic windows) and 251 firm-year observations represent
conventional banks (with no Islamic windows).
4.2 Model
To examine the use of discretion by the managers of Islamic banks and compare it to
conventional banks in the Middle East region, we use the two-stage approach. At the
first stage, we use specific accruals to measure artificial earnings management in
Islamic banks. More specifically, we use a major accrual in the banking sector, LLP.
This proxy is divided into two components: discretionary and non-discretionary.
Whence, the basic model takes the form:
LLP Non-discretionary LLP Discretionary LLP
Following Zoubi and Al-Khazali (2007) and Taktak et al. (2010b), we take into
consideration the specificities of Islamic banks. These institutions operate under
Sharia law, and use techniques conform to sharia principles. The composition of their
portfolio loan is different from conventional banks. The loan portfolio of Islamic banks
includes four important types; Quard Hassan (loan), Musharaka, Murabaha and
Mudaraba investment. Consequently, we use for in our study the item loss provision
for loans, Murabaha, Musharaka, Mudaraba investment (LLPI) to estimate total loss
provisions for Islamic banks.

Table II.
Number of banks and
observations in the
sample by country

Country

(1)a

Bahrain
Egypt
Jordan
Kuwait
Quatar
Saoudi Arabia
UAE
Total
Seven countries

4
3
1
4
3
3
3
21

Number of banks
(2)b
1
0
0
1
5
6
5
18
72 banks

(3)c

(1)a

4
6
10
5
0
1
7
33

21
14
6
29
15
18
24
129

Number of observations
(2)b
(3)c
9
0
0
9
34
49
40
141
519 observation

34
43
83
34
0
9
48
251

Notes: aIslamic banks that are fully Sharia compliant; bconventional banks with Islamic windows;
c
conventional banks

The non-discretionary component of LLP represents the portion of total accruals

dictated by changes in bank business conditions. Because it cannot be directly
observed, it is estimated through variables reflecting the level of losses in the loan
portfolio. Similar to Kanagaretnam et al. (2004) and Kwal et al. (2009), NDLLP
component is estimated using a set of informational variables including the beginning
balance of non-performing loans, change in non-performing loans and change in total
loans. We expect to have a positive correlation between LLP and the independents
variables mentioned above. We expect that, if the beginning balance of non-performing
loans is high, banks will report a high level of loss provisions. On the other hand,
an increase in non-performing loans is likely to result in an increase in LLP and a
positive change in total loans increases the risk of uncollectible accounts. This involves
an increase in the amount of loss provisions. We estimate NDLLP using equation (1).
We used panel data over the period 2000-2008. Each observation of our sample has
two dimensions (firm, year). The estimation method of our model using panel
regression techniques is appropriate. Due to data availability, we used non-balanced
panel data techniques. Our sample consists of unbalanced panel data because each
variable is observed over varying time-period length:
LLPit b0 b1 NPLit21 b2 DNPLit b3 DTL 1it

Use of DLLP by
Islamic banks

115

where:
LLPit

total LLP for bank i at the year t, deflated by beginning loans.

NPLit2 1

the beginning balance of non-performing loan for bank i at the year t

deflated beginning loans.

DNPLit

change in the value of non-performing loan for bank i at the year t,

deflated by beginning loans.

DTL

change in the value of total loan, for bank i at the year t, deflated by
beginning loans.

The DLLP consists of the LLP prediction error; it is estimated through the residual
obtained from equation (1).
First of all, we estimate equation (1) for Islamic banks that are fully Sharia
compliant to obtain the estimates of b0, b1, b2 and b3. Table III reports the mean

Variables

Mean coefficient estimates for the model in equation (1)

Coefficient estimate

Intercept
Beginning balance NPL
DNPL
DTL
F-Fisher
R 2-overall

0.0169
20.1498
20.0001
0.0078
10.2300
39.03%

( p-value)
(0.005) * * *
(0.163)
(0.780)
(0.000) * * *
(0.0001) * * *

Notes: Significant at: *10, * *5 and * * *1 percent levels; LLPit b0 b1 NPLit21 b2 DNPLit
b3 DTL 1it

Table III.
Result of the regression
model from equation (1)

SEF
31,1

116

coefficient estimates for equation (1) based on panel estimation techniques. The
Hausman specification test is used to choose between the fixed or random effects model
for our sample of Islamic banks. The statistic of Hausman test is significant. As a
consequence, we validate the fixed effect. The explanatory power is relatively high,
with the mean R 2 overall 39.03 percent. As expected, we find a positive and significant
relationship at the level of 0.01 between LLP and total loan. However, the coefficients of
NPL and DNPL are not significant.
Then, using the estimated coefficients (b^ 0 , b^ 1 , b^ 2 , b^ 3 ) from equation (1), we evaluate
the non-discretionary component of LLP, NDLLP:
NDLLPit b^ 0 b^ 1 NPLit21 b^ 2 DNPLit b^ 3 DTLit

Finally, we obtain the discretionary component of LLP by calculating the difference

between total LLP and estimated non-discretionary LLP. Our basic estimation
equation becomes:
DLLPit LLPit 2 b^ 0 b^ 1 NPLit21 b^ 2 DNPLit b^ 3 DTLit

At the second stage, we use the discretionary LLP component as our dependent
variable. The independent variables in equation (4) below represent factors
hypothesized to influence DLLP. We examine whether, Islamic banks managers use
their discretion for both earnings and capital management:
DLLPIit b0 b1 EBTPit b2 CARit b3 LDit b4 Sizeit b5 AAOIFIit

11
X

bj countries 1it

j6

where:
DLLPIit

discretionary loss provisions for loans, Murabaha, Musharaka,

Mudaraba investment for bank i at the year t.

EBTPit

earning before taxes and provisions deflated by total assets for bank i at
the year t.

CARit

capital adequacy ratio for bank i at the year t, measured by average total
equity over average total assets.

LDit

loan to deposit for bank i at the year t.

Sizeit

bank size for bank i at the year t, expressed as natural log of asset.

AAOIFIit

dummy variable that takes 1 if the bank uses AAOIFIs accounting

standards, 0 otherwise.

S Country a set of country dummy variables controlling for specific differences

across countries.
We conduct our study first, on Islamic banks, then conventional banks with Islamic
windows and conventional banks. The procedure for determining the DLLP for

conventional banks which provide also Islamic services and conventional banks is
similar to that used before.
In order to determine whether observed differences between groups, the separate
samples are pooled and we include the variable type to compare the use of discretion
between:
.
Islamic banks vs conventional banks with Islamic windows.
.
Islamic banks vs conventional banks.
.
Islamic banks and conventional banks with Islamic windows vs conventional
banks.
The new model is presented as follows:
DLLPIit b0 b1 EBTPit b2 CARit b3 LDit b4 Sizeit
b5 AAOIFIit

11
X

bj countries Type 1it

j6

where:
DLLPIit

discretionary loss provision for loans, Murabaha, Musharaka,

Mudaraba investment for bank i at the year t.

EBTPit

earning before taxes and provisions deflated by total assets for bank i at
the year t.

CARit

capital adequacy ratio for bank i at the year t, measured by average total
equity over average total assets.

LDit

loan to deposit for bank i at the year t.

Sizeit

bank size for bank i at the year t, expressed as natural log of asset.

Type

is a dummy variable taking 1, if the bank is Islamic and 0 otherwise.

AAOIFIit

dummy variable that takes 1 if the bank uses AAOIFIs accounting

standard, 0 otherwise.

S Country a set of country dummy variables controlling for specific differences

across countries.
5. Empirical results
5.1 Descriptive statistics
Tables IV and V present descriptive statistics for the dependent and independent
variables used in this study. On average, in Islamic banks, LLP and beginning
non-performing loans represent, respectively, 1.38 and 4.3 percent of beginning loans.
Our findings are consistent with those reported by Zoubi and Al-Khazali (2007),
Taktak et al. (2010b) and Quttainah (2011) who found that Islamic banks make a low
estimation of loss provisions and non-performing loan. The mean ratio of change in total
loan equals 60.81 percent, like Taktak et al. (2010b) and Quttainah (2011), results in
standard deviation indicating a large dispersion in the level of loans provided by
Islamic banks. The mean DLLP, measured as the residual value from equation (1),

Use of DLLP by
Islamic banks

117

Table IV.
Descriptive statistics for
all the variables included
in the regression models

LLP ratio
Change in total
loan ratio
Beginning nonperforming loan
ratio
Change in nonperforming loan
ratio
EBTP ratio
CAR
Loan to deposit
Size
DLLP
3.6165
0.0315

0.6081

0.0436

0.0043

21

2 0.0071

0.1279

35.9628

0.1312

0.0564

0.2048

0.0063

0.0737

0.0047

0.8616 2 1.0000

0.0081 2 0.0097

0.4128

8.7917

0.0465

0.0836 0.1252

0.0015

0.2021 0.9415 21.0000

1.0597

8.0073

0.2067

Conventional banks
SD
Min.
Max.
0.0116 0.0202 20.0228

Mean

6.21 102 07 355.2172 2 2,324.2000 1,753.7000 0.0062

0.0621 2 0.2629
0.4128 0.0111 0.1086 20.3358 1.0597
0.0348
0.03030
0.0017
0.1447 0.0355
0.0483
0.0092
0.2169 0.0238 0.0147 20.0720 0.1070
17.8890
13.3056
4.0960
51.9500 12.768
9.6230
8.0262 51.6000 11.5240 3.8248
2.7000 32.2000
156.1305
175.5431
26.9
995.3000 74.8510
21.6613 23.4000 134.5000 0.7398 0.2271
0.2514 1.3080
3.4374
0.4524
2.5064
4.4285 3.906075 0.4060
3.0850
4.4387 3.4933 0.6346
1.8739 5.2443
0.0000
0.0349
0.2979
0.0839 0.0000
0.0080 2 0.0172
0.0412 0.0000 0.0215 20.0493 0.2277

0.0221

0.0138

Conventional banks which Islamic

windows
Mean
SD
Min.
Max.

118

Variables

Islamic banks that are fully sharia compliant

Mean
SD
Min.
Max.

SEF
31,1

Variables
LLP ratio
Change in total loan
ratio
Beginning nonperforming loan
ratio
Change in nonperforming loan
ratio
EBTP ratio
CAR
Loan to deposit
Size
DLLP

Islamic banks vs
conventional banks with
Islamic windows
Difference in means

Islamic banks and

conventional banks with
Islamic windows vs
Islamic banks vs
conventional banks
conventional banks
Difference in means
Difference in means

0.0074 * * *

0.0022

0.4033

0.4060 *

0.1836

20.0128

2 0.0400 * *

20.0317 * *

20.0062
20.0006
5.1213 * * *
81.2794 * * *
20.4686 * * *
20.0083 * * *

2 0.0111
0.01103 * * *
6.4386 * * *
155.3910 * * *
2 0.0559
2 0.0121 * * *

20.0072
0.0113 * * *
3.7224 * * *
111.1420 * * *
0.1856 * * *
20.0074 * * *

20.0022

Note: Significant at: *10, * *5 and * * *1 percent levels

was zero, with a minimum of 3.49 percent and a maximum of 29.79 percent. The average of
EBTP to total assets is 3.48 percent with a maximum of 26 percent and a standard
deviation of 3.56 percent and is slightly higher to the mean of 2.23 and 2.29 percent
reported, respectively, by Zoubi and Al-Khazali (2007) and Taktak et al. (2010b). The mean
value of the natural log of total assets in Islamic banks is 3.43 percent and the standard
deviation 0.50 percent.
As reported in Table V, we use an ANOVA analysis, to compare variables among
the three different types of banks used in our sample. The evidence on LLP indicates
that LLP ratio for Islamic banks is higher than for conventional banks with Islamic
windows and is the same as for conventional banks. We conclude that Islamic banks
and conventional banks behave in the same manner in dealing with LLP.
Beginning non-performing loan ratio for Islamic banks is lower than those for
conventional banks providing Islamic services and conventional banks. This indicates
that full-fledged Islamic banks have less non-performing loans problem than the others
types of bank.
Change in total loans, EBTP and CAR of Islamic banks are significantly higher than
those of conventional banks. Islamic banks report statistically significantly lower
smaller DLLP than their non-Islamic counterparts. This indicates that Islamic banks
are less likely to use LLP to manage accounting earnings compared to conventional
banks. Referring to Taktak et al. (2010b), this finding could be explained by the
provisioning policy used by Islamic banks. Indeed, to determine provisions and
reserves, Islamic institutions use FAS 11 (Financial Accounting Standards (FAS)).
This standard requires an adequate level of provision and therefore, leaves little
discretion to Islamic banks to manage their earnings.
Table VI exhibits the correlation matrix for the variables in model (4). Pearson
correlation coefficients are reported on the left down side of the matrix while spearman

Use of DLLP by
Islamic banks

119

Table V.
Results of ANOVA
analysis

Table VI.
Correlation matrix for all
independent variables
0.3456
(0.0003) * * *
20.4634
(0.0000) * * *
0.1639
(0.0787) *
20.2921
(0.0015) * *
20.2376
(0.0102) *
20.1295
0.1660
0.1829
(0.0495) *
20.1501
0.1079
0.2118
(0.0225) *

0.3944
1.0000

20.3235
(0.0006) * * *
0.4084
(0.0000) * * *
20.0817
0.3792
20.1357
0.1428
20.1984
(0.0313) *
0.0113
0.9033
20.2294
(0.0125) *
20.0703
0.4493

1.0000

0.2706
0.6004

LD

Note: Significant at: *10, * *5 and * * *1 percent

Saudi
Arabia
UAE

Qatar

Kuwait

Jordan

Egypt

AAOIFI

Size

LD

1.0000
0.4294
(0.0000) * * *
0.0308
0.7515
0.0982
0.2779
0.2550
(0.0043) * *
20.1654
(0.0663) *
20.0826
0.3616
20.0889
0.3263
0.4792
(0.0000) * * *
20.0025
0.9780
20.1115
0.2178

CAR

20.0642
0.4785
20.1644
(0.0681) *
20.0561
0.5358
0.3438
(0.0001) * * *
20.0170
0.8514
0.2860
(0.0013) * *
20.1484
(0.0999) *

1.0000

20.2372

0.1235
20.3604

SIZE

20.3486
(0.0000) * * *
0.1624
(0.0259) *
20.1580
(0.0303) *
0.4810
(0.0000) * * *
20.3486
(0.0000) * * *
20.3486
(0.0000) * * *

1.0000

0.0154

0.2840

0.2789
0.0154

AAOIFI

20.0918
0.2101
20.1959
(0.0071) * *
20.1677
(0.0214) *
20.1677
(0.0214) *
20.1677
(0.0214) *

1.0000

20.2799

20.1621

20.2610

20.2514
20.4141

Egypt

Kuwait

0.3943

20.0771

20.1259

0.4498

0.0374

0.1353

0.4633
0.1750

Qatar

20.0945

20.1544

20.3805

0.2815

20.3753

20.0138
20.1043

Saudi
Arabia

20.0842

20.1376

20.3390

20.3268

0.2876

20.0649
0.4665

UAE

20.1072
1.0000
20.2186
20.2681
20.2389
0.1429
20.0918 20.1959
1.0000
20.1711
20.1525
0.2101 (0.0071) * *
20.0918 20.1959
20.1677
1.0000
20.1870
0.2101 (0.0071) * * (0.0214) *
20.0918 20.1959
20.1677
20.1677
1.0000
0.2101 (0.0071) * * (0.0214) * (0.0214) *

1.0000 20.1207

20.0695 20.1973

0.1435 20.0573

20.1167

20.3190 20.0728

20.2024 20.0964
20.2917 20.1767

Jordan

120

EBTP
CAR

EBTP

SEF
31,1

correlations are presented on the right upside. Table VII presents correlations among
the independent dummy variables.
Results report a high correlation between CAR and LD. Therefore, we conduct two
separate specifications. In the first one (Panel A), we keep LD and we drop CAR and in
the second one (Panel B) CAR is kept and LD is dropped.
Tables VI and VII report correlations below 0.6. Therefore, correlations are not
sufficiently high to pose any serious multicollinearity problem according to the limit
set by Gujarati (1995).

Use of DLLP by
Islamic banks

121

5.2 Panel regression analysis

In this section, we present and discuss our empirical results concerning the factors that
my influence the use of discretion by Islamic banks managers in estimating LLP.
Table VIII (Panels A and B) reports the main results of the second stage regression
analysis.
For both specifications, our dependent variable is measured by DLLPI, which
corresponds to the discretionary component of LLP and represents the residual from
equation (1).
The variables of the two regression models are globally significant in Panels A
and B. The wald x 2-test statistic is significant at the 1 percent level in Panel A and at
10 percent level in Panel B. R 2 overall is 50.65 percent in Panel A and 14.63 percent in
Panel B. Therefore, the variables included in the two models provide a relatively good
explanation of DLLP.
The result of the regression in Panel A is in accordance with our expectations on the
earnings management hypothesis. Similarly to Zoubi and Al-Khazali (2007), the
coefficient of EBTP has the predicted sign and is significant at the 1 percent level. This
finding indicates that EBTP is an important factor that can influences the use of
direction by managers when reporting LLP. Unlike Taktak et al. (2010b), we found that
managers of Islamic banks use their discretion to manipulate accounting earnings
through LLP.
With respect to capital management hypothesis, results in Panel B are in contrast
with our expectations. Indeed, we find a positive and significant coefficient at the level
of 10 percent for CAR. This outcome confirms the mixed results reported in previous
studies as mentioned above in the hypothesis development section. Our findings
corroborate those obtained by Kim and Kross (1998) and Ahmed et al. (1999). This
suggests that Islamic banks with high equity incentives are more likely to report
discretionary LLP. Overall, the respect of banking regulatory requirements is the main

AAOIFI
Egypt
Jordan
Kuwait
Qatar
Saudi Arabia
UAE

AAOIFI

Egypt

Jordan

Kuwait

Qatar

1.0000
20.3448
0.1624
20.1580
0.4810
20.3486
20.3486

1.0000
20.0918
20.1959
20.1677
20.1677
20.1677

1.0000
20.1072
20.0918
20.0918
20.0918

1.0000
20.1959
20.1959
20.1959

1.0000
20.1677
20.1677

Saudi Arabia

1.0000
2 0.1677

UAE

1.0000

Table VII.
Spearman correlation
matrix for dummy
variables

SEF
31,1
Variables

122

Intercept
EBTP
CAR
LD
Size
AAOIFI
Egypt
Jordan
Kuwait
Qatar
Saudi Arabia
UAE
Wald x 2
R 2 overall (%)
R 2 between (%)

Mean coefficient estimates for the model in equation (4)

Panel A
Panel B
Coefficient estimate
( p-value)
Coefficient estimate
0.0257
1.0956

20.0008
20.0160
0.0103
0.0188
20.0089
0.0044
0.0082
20.0040
20.0129
69.8700
50.65
64.89

0.7040
(0.0000) * * *
(0.0560) *
0.2810
0.7380
0.7180
0.9880
0.9230
0.8380
0.9390
0.8110
(0.0000) * * *

( p-value)
0.1750
0.2400
(0.0560) *

0.6270
0.7710
0.2390
0.4580
0.5610
0.2170
0.6710
0.8950
(0.02690) *

20.0710
0.2191
0.0010

0.0061
0.0073
0.0463
0.0310
0.0192
0.0346
0.0169
0.0052
12.2500
14.63
43.96

Notes: Significant at: *10, * *5 and * * *1 percent levels:

Panel A : DLLPIit b0 b1 EBTPit b2 LDit b3 Sizeit b4 AAOIFTIit

10
X

bj countries 1it

j5

4
Panel B : DLLPIit b0 b1 EBTPit b2 CARit b3 Sizeit b4 AAOIFTIit

10
X

bj countries 1it

j5

Table VIII.
Result of the regression
model from equation (4)

where: DLLPIit discretionary loss provisions for loans, Murabaha, Musharaka, Mudaraba
investment for bank i at the year t; ETPit earning before taxes and provisions deflated by total assets
for bank i at the year t; CARit capital adequacy ratio for bank i at the year t, measured by average
total equity over average total assets; LDit loan to deposit for bank i at the year t; Sizeit bank size
for bank i at the year t, expressed as natural log of asset; AAOIFI dummy variable that takes 1 if the
bank uses AAOIFIs accounting standard, 0 otherwise; S Country a set of country dummy variables
controlling for specific differences across countries

concern of managers. Whence, they are motivated to use discretion in estimating LLP
in order to enhance earnings and therefore capital.
The analysis of the results related to our control variables shows that contrary to
our predictions the coefficient of size is not significant. Results in Panels A and B do
not support the findings of Zoubi and Al-Khazali (2007) and Taktak et al. (2010b).
Indeed, we found that the large-sized banks are not found to report higher DLLP.
As expected, the LD coefficient is negative and significant at 10 percent level. This
result shows that external financing need is a prominent factor affecting the estimation
of LLP by managers.
As regards the effect of Islamic accounting standard on the estimation of LLP by
managers, results provide evidence that Islamic banks managers behave in the same
way when reporting LLP. Therefore, the implementation of AAOIFI standard does not

have any influence on their behavior. Accordingly, the use of the dynamic provisions
policy does not reduce the discretion of managers. In fact, their behavior remains the
same.
To test for the effect of the type of the bank (Islamic banks, conventional banks with
Islamic windows and conventional banks) on DLLP, we include the variable type in the
regression model. Type is a dummy variable that equals 1 if the bank is Islamic,
0 otherwise. Thus, the equation to be estimated becomes:
DLLPIit b0 b1 EBTPit b2 CARit b3 LDit b4 Sizeit
b5 AAOIFIit

11
X

bj countries b12 Type 1it

j6

Equation (6) is used to compare the use of discretion between:

.
Islamic banks versus conventional banks providing Islamic services (Panel A).
.
Islamic banks versus conventional banks (Panel B).
.
Islamic banks and conventional banks providing Islamic services versus
conventional banks (Panel C) operating in the Middle East region.
Table IX shows the main results of our comparison study. It reports the regression
results pertaining to our three groups of banks.
The correlation coefficients show the existence of multicollinearity between CAR
and TL in Panel A and between LD and Type in both Panels B and C. As a
consequence, we run two separate specifications for each model. We found that results
remain the same. This finding shows that the correlation between the variables does
not affect the robustness of our model. Therefore, we report in Table IX, the results of
the three regressions including all variables. It is noteworthy that, the matrix
correlation and the separate specifications are note reported in our study.
On average, results suggest that type is not significant for all the groups. There is
no difference in the use of discretion by managers when reporting LLP between Islamic
banks and their non-Islamic counterparts. This finding is in accordance with the one
reported by Zoubi and Al-Khazali (2007), who found that Islamic and conventional
banks follow the same way of provisioning.
Additional results reported in Table IX outline that EBTP and CAR coefficients are
positive and significant at different level for our three specifications (Panel A-B-C). So,
banks in the Middle East region use DLLP for earnings and capital management
purpose. And all banks behave in the same way when dealing with DLLP.
The need for external financing is also found to be a determinant of DLLP. Results
show a negative and significant relationship between LD and DLLP.
6. Conclusion
The purpose of this paper is to examine factors that may influence Islamic banks
managers use of discretion in estimating LLP. Using a sample of 21 Islamic banks in
the Middle East region over the period 2000-2008, we find that, Islamic banks
managers use their discretion in reporting LLP for both earnings and capital
management purpose. Findings reveal a positive and significant relationship between
DLLP in the one side and EBTP and CAR in the other side. Our evidence of positive

Use of DLLP by
Islamic banks

123

SEF
31,1

124

Variables

Islamic banks vs
conventional banks with
Islamic banks vs
Islamic windows
conventional banks
Panel A
Panel B
Coefficient
( p-value) Coefficient ( p-value)

Intercept
EBTP ratio
CAR
LD
SIZE
Type
AAOIFI
Bahrain
Egypt
Jordan
Kuwait
Qatar
Saudi Arabia
UAE
Wald x 2
R 2 overall (%)
R 2 between (%)

2 0.0277
0.1179
0.0012
2 0.0009
0.0036534
2 0.0069
0.0188
0.0020
0.0271
2 0.0007
2 0.0023
0.0030

2 0.0052
40.5400
23.27
56.75

0.2460
(0.0850) *
(0.0000) * * *
(0.0000) * * *
0.4760
0.5180
0.1190
0.8970
0.1060
0.9760
0.8710
0.7980

0.6440
(0.0001) * *

2 0.0250
0.2662
0.0008
2 0.0008
0.0024
2 0.0034
0.0129
2 0.0005
0.0168

2 0.0002
0.0128
0.0134
2 0.0001
49.67
48.34
21.25

Islamic banks and

conventional banks with
Islamic windows VS
conventional banks
Panel C
Coefficient
( p-value)

0.1230
2 0.0206
(0.0010) * *
0.1467
(0.0030) * *
0.0005
(0.0000) * * * 2 0.0008
0.5850
0.0029
0.6690
2 60.91 102 06
0.2670
0.0126
0.9630
0.0034
0.0171
(0.0750) *

2 0.0007
0.9820
2 0.0011
0.4460
0.0070
0.7940
0.0010
0.9840

(0.0000) * * *
56.48
17.48
39.88

(0.0760) *
(0.0020) * *
(0.0010) * *
(0.0000) * * *
0.2870
0.9990
(0.0290) *
0.5750
(0.0060) * *
0.9010
0.8470
0.2780
0.8590

(0.0000) * * *

Notes: Significant at: *10, * *5 and * * *1 percent levels:

DLLPIit b0 b1 EBTPit b2 CARit b3 LDit b4 Sizeit
b5 AAOIFIit

11
X

bj countries Type 1it

j6

Table IX.
Results of the regression
model from equation (5)

where: DLLPIit discretionary loss provisions for loans, Murabaha, Musharaka, Mudaraba investment
for bank i at the year t; ETPit: earning before taxes and provisions deflated by total assets for bank i at
the year t; CARit capital adequacy ratio for bank i at the year t, measured by average total equity over
average total assets; LDit loan to deposit for bank i at the year t; Sizeit bank size for bank i at the year
t, expressed as natural log of asset; AAOIFI dummy variable that takes 1 if the bank uses AAOIFIs
accounting standard, 0 otherwise; S Country a set of country dummy variables controlling for specific
differences across countries; Type: is a dummy variable taking 1, if the bank is Islamic and 0 otherwise

relationship indicates that managers are found to use discretion to underestimate LLP
if the EBTP and CAR are low in order to enhance earnings and to comply with the
banking requirement regulation. Additional results show that external financing need
is an important factor that may influence the managerial discretion.
To test the effect of the type of banks on DLLP, we use a sample of 18 conventional
banks with Islamic windows and 33 conventional banks in the Middle East region. Our
findings indicate the lack of relationship between the type of banks and the use of
DLLP. This led us to conclude that there is no difference between Islamic banks,
conventional banks providing Islamic services and conventional banks in terms of
DLLP. Considering DLLP practice, all these banks behave similarly.
Overall, our finding can be useful for standard-setters, auditors and investors.
It helps standard setters to take appropriate measures and regulate the provisioning

policies of banks. Such measures can reduce the discretion of managers, and
therefore increase the information transparency and improve the accounting quality.
Auditors should be conscious of the existence of earnings management practices.
Thus, they should focus more on the policy of provision used by banks. Detecting
discretionary behavior in Islamic banks can also help investors in their decision
making.
One limitation of our paper is that our sample size is relatively small, given that
data are not available for all Islamic banks. Future research may extend the sample to
all Islamic institutions in the MENA region. Another avenue for future research could
be to investigate factors that potentially influence earnings-management policy in
Islamic banks.
Note
1. Countries where AAOIFI standards are either mandatory or recommended include: Bahrain,
Jordan, UAE, Saudi Arabia, Lebanon, Syria, Sudan and Malaysia www.islamicbanker.com/aaoifistandards
References
Ahmed, A.S., Takeda, C. and Thomas, S. (1998), Bank loan loss provisions: a reexamination of
capital management, earnings management and signaling effects, Journal of Accounting
and Economics, Vol. 28, pp. 1-25.
Ahmed, A.S., Takeda, C. and Thomas, S. (1999), Bank loan loss provisions: a reexamination of
capital management, earnings management and signaling effects, Journal of Accounting
and Economics, Vol. 28 No. 1, pp. 1-25.
Al-Gamal, M. (2006), Islamic Finance Law, Economics, and Practice, Cambridge University Press,
New York, NY.
Anandarajan, A., Hasan, I. and Lozano-Vivas, A. (2005), Loan loss provisions decisions:
an empirical analysis of the Spanish depository institutions, Journal of International
Accounting, Auditing and Taxation, Vol. 14 No. 1, pp. 55-77.
Anandarajan, A., Hasan, I. and McCarthy, C. (2007), Use of loan loss provisions for capital,
earnings management and signalling by Australian banks, Accounting and Finance,
Vol. 47, pp. 357-379.
Bae, G.S., Hamao, Y. and Kang, J.K. (2009), Bank monitoring incentives and borrower earnings
management: evidence from the Japanese banking crisis of 1993-2002, Working Paper
No. 279, Vol. 279, Center on Japanese Economy and Business, Columbia University,
New York, NY.
Beatty, A., Chamberlain, S. and Magliolo, J. (1995), Managing financial reports of commercial
banks: the influence of taxes, regulatory capital, and earnings, Journal of Accounting
Research, Vol. 33, pp. 231-262.
Bhat, V. (1996), Banks and income smoothing: an empirical analysis, Applied Financial
Economics, Vol. 6 No. 6, pp. 505-510.
Collins, J., Shackelford, D. and Wahlen, J. (1995), Bank differences in the coordination of
regulatory capital, earnings and taxes, Journal of Accounting Research, Vol. 33 No. 2,
pp. 263-292.
Cornett, M.M., McNutt, J.J. and Tehranian, H. (2009), Corporate governance and earnings
management at large US bank holding companies, Journal of Corporate Finance, Vol. 15,
pp. 412-430.

Use of DLLP by
Islamic banks

125

SEF
31,1

126

Fama, E.F. (1980), Agency problems and the theory of the firm, Journal of Political Economy,
Vol. 88, pp. 288-307.
Farouk, S., Hassan, K. and Clinch, G. (2012), Profit distribution management by
Islamic banks: an empirical investigation, The Review of Economics and Finance,
Vol. 52, pp. 333-347.
Greenwalt, M. and Sinkey, J. (1988), Bank loan loss provisions and the income smoothing
hypothesis: an empirical analysis, Journal of Financial Services Research, Vol. 1,
pp. 301-318.
Gujarati, D.N. (1995), Basic Econometrics, McGraw-Hill, New York, NY.
Hamdi, M.F. and Zarai, M.A. (2012), Earnings management to avoid earnings decreases and
losses: empirical evidence from Islamic banking industry, Research Journal of Finance
and Accounting, Vol. 3, pp. 88-107.
Hanif, M. (2011), Differences and similarities in Islamic and conventional banking,
International Journal of Business and Social Science, Vol. 2, pp. 166-175.
Ismail, A.G. and Be Lay, A.T. (2002), Bank loans portfolio composition and the disclosure of
loan loss provision: an empirical evidence of Malaysian banks, Asian Review of
Accounting, Vol. 10 No. 1, pp. 147-162.
Ismail, A.G. and Shahimi, S. (2003), Do risk-based capital requirements allocate financing and
cause a bigger loan loss provisions for Islamic banks?, Journal of Islamic Economics,
Vol. 4 No. 1, pp. 1-11.
Ismail, A.G., Shaharudin, R.S. and Samudhram, A.R. (2005), Do Malaysian banks manage
earnings through loan loss provisions?, Journal of Financial Reporting and Accounting,
Vol. 3 No. 1, pp. 41-47.
Jensen, M.C. and Meckling, W.H. (1976), Theory of the firm: managerial behavior, agency costs,
and ownership structure, Journal of Financial Economics, Vol. 3, pp. 305-360.
Kanagaretnam, K., Lobo, G.J. and Mathieu, R. (2004), Earnings management to reduce earnings
variability: evidence from bank loan loss provisions, Review of Accounting and Finance,
Vol. 3, pp. 128-148.
Kim, M. and Kross, W. (1998), The impact of the 1989 change in bank capital standards on loan
loss provisions and loan write-offs, Journal of Accounting and Economics, Vol. 25 No. 1,
pp. 69-100.
Kwal, W., Lee, Y.H. and Eldridge, S. (2009), Earnings management by Japanese bank managers:
using discretionary loan loss provisions, Review of Pacific Basin Financial Markets and
Policies, Vol. 12 No. 1, pp. 1-26.
Lobo, G.J. and Yang, D.H. (2001), Bank managers heterogeneous decisions on
discretionary loan loss provisions, Review of Quantitative Finance and Accounting,
Vol. 16, pp. 223-250.
Ma, C.K. (1988), Loan loss reserve and income smoothing: the experience in the
US banking industry, Journal of Business Finance & Accounting, Vol. 15 No. 4,
pp. 487-497.
Macey, J.R. and OHara, M. (2003), The corporate governance of banks, Economic Policy Review,
Vol. 9, pp. 91-107.
Misman, F.A. and Ahmed, W. (2011), Loan loss provisions: evidence from Malaysian Islamic
and conventional banks, International Review of Business Research Papers, Vol. 7 No. 4,
pp. 94-103.
Moyer, S.E. (1990), Capital adequacy ratio regulations and accounting choices in commercial
banks, Journal of Accounting and Economics, Vol. 13 No. 2, pp. 123-154.

Palia, D. and Porter, R. (2007), Agency theory in banking: an empirical analysis of

moral hazard and the agency costs of equity, Banks and Bank Systems, Vol. 2 No. 3,
pp. 142-156.
Pinho, P.S. and Martins, N.C. (2009), Determinants of Portuguese banks provisioning policies:
discretionary behaviour of generic and specific allowances, Journal of Money, Investment
and Banking, No. 10, pp. 43-56, special issue.
Quttainah, M.A. (2011), Do Islamic banks employ less earnings management?, working paper,
paper presented at ERF 17th Annual Conference, Turkey, March 20-22.
Salman, S.A. (2004), Islamic modes of finance and associated liquidity risks, working paper,
paper presented at Conference on Monetary Sector in Iran: Structure, Performance and
Challenging Issues Tehran, February.
Sarea, A.M. (2012), The level of compliance with AAOIFI accounting standards: evidence from
Bahrain, International Management Review, Vol. 8, pp. 27-32.
Shen, C. and Chih, H. (2005), Investor protection, prospect theory, and earnings management:
an international comparison of the banking industry, Journal of Banking & Finance,
Vol. 29, pp. 2675-2697.
Taktak, N. (2011), The nature of smoothing returns practices: the case of Islamic banks, Journal
of Islamic Accounting & Business Research, Vol. 2 No. 2, pp. 142-152.
Taktak, N., Shabou, R. and Dumontier, P. (2010a), Income smoothing practices: evidence from
banks operating in OECD countries, International Journal of Economics and Finance,
Vol. 2 No. 4, pp. 140-151.
Taktak, N., Zouari, S. and Boudrigua, A. (2010b), Do Islamic banks use loan loss provisions to
smooth their results?, Journal of Islamic Accounting & Business Research, Vol. 1 No. 2,
pp. 114-127.
Turen, S. (1996), Performance and risk analysis of the Islamic banks: the case of Bahrain Islamic
bank, Journal of Islamic Economics, Vol. 8, pp. 3-14.
Wang, C.Y., Chen, M.C., Lee, C.C. and Shyu, S.D. (2012), Income smoothing and earnings
management in the banking industry: evidence in Asian countries, International Research
Journal of Finance and Economics, No. 99, pp. 37-52.
Wetmore, J.L. and Brick, J.R. (1994), Loan-loss provisions of commercial
banks and adequate disclosure: a note, Journal of Economics and Business, Vol. 46,
pp. 299-305.
Wezel, T., Jorge, A., Lau, C. and Columba, F. (2012), Dynamic loan loss provisioning: simulations
on effectiveness and guide to implementation, IMF working paper, May.
Zoubi, T.A. and Al-Khazali, O. (2007), Empirical testing of the loss provisions of banks in the
GCC region, Managerial Finance, Vol. 33 No. 7, pp. 500-511.
About the authors
Dr Hakim Ben Othman is a Matre de Conferences Agrege of Accounting and Finance at Tunis
Business School, University of Tunis. He is a Senior Researcher and member of the
Interdisciplinary Laboratory in Management University-Enterprise (LIGUE) at ISCAE University
of Manouba. His main research interests include international accounting, corporate governance,
financial reporting, Islamic finance and accounting, voluntary disclosure, creative accounting and
earnings management, directors incentives and firm performance. He has published in
The International Journal of Accounting (Elsevier); Managerial Auditing Journal, Research
in Accounting in Emerging Economies, Journal of Accounting in Emerging Economies, Studies in
Economics and Finance (Emerald Publishing); Journal of Accounting Auditing and Performance
Evaluation, International Journal of Managerial and Financial Accounting (Inderscience) and other

Use of DLLP by
Islamic banks

127

SEF
31,1

128

academic journals. Dr Hakim Ben Othman is committed to several editorial activities. He has
reviewed articles for both local and international journals such as Management Research Review,
Journal of Islamic Accounting and Business Research (Emerald), International Journal of
Accounting and finance (Inderscience Publishers), The Journal of Credit Risk. He has been
an External Examiner for a large number of PhD students, both at the local universities and
abroad. Hakim Ben Othman is the corresponding author and can be contacted at: hakim.
bo@planet.tn
Hounaida Mersni is a Teaching Assistant at Tunis Business School, University of Tunis. She is
a PhD student at ISCAE, and member of LIGUE, University of Manouba. Her research interests are
in corporate governance, voluntary disclosure, earnings management and Islamic finance and
accounting.

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints