1. INTRODUCTION

Islamic finance was practicedpredominantly in the Muslim worldthroughout the middle Ages, fostering tradeand business activities with the developmentof credit. In Spain and the Mediterranean andBaltic states, Islamic merchants becameindispensable middlemen for tradingactivities. In fact, many concepts,techniques, and instruments of Islamic

finance were later adopted by Europeanfinanciers and businessmen (Alasrag, 2010).

According to Alasrag (2010), the term“Islamic financial system” is relatively new,appearing only in the mid-1980s. In fact, allthe earlier references to commercial ormercantile activities conforming to Islamicprinciples were made under the umbrella ofeither “interest free” or “Islamic” banking.However, describing the Islamic financialsystem simply as “interest-free” does not

ISLAMIC BANKING DURING THE FINANCIAL CRISIS OF 2007

Abdelkader Derbali

University of Sousse, Higher Institute of Management of Sousse, Sreet Abdlaaziz il Behi. Bp 763. 4000 Sousse, Tunisia

(Received 6 July 2014; accepted 8 September 2014)

Abstract

The purpose of this paper is to initially contribute to the literature linking the global financialcrisis and performance of Islamic banks. Thus, it is important for everyone's future to study thecurrent crisis in order to develop sustainable financial practices and in quest of a new business modelbased on sharing the profit and loss. This sustainable financial practice is based on non-interest-basedtransactions but profit and loss sharing, which should be in practice at the financial system. In thispaper, the performance of the Islamic banks in a period of crisis, were tested. The sample of 29Islamic banks from 7 countries and from a period of study of 7 years (2006-2012), was used.According to the empirical results, it was concluded that the Islamic banks are not affected by thefinancial crisis of 2007.

Keywords: financial crisis, islamic finance, conventional finance, performance, panel data

* Corresponding author: derbaliabdelkader@outlook.fr

S e r b i a n

J o u r n a l

o f

M a n a g e m e n t

Serbian Journal of Management 10 (1) (2015) 89 - 108

www.sjm06.com

DOI: 10.5937/sjm10-6402

provide a true picture of the system as awhole. Undoubtedly, prohibiting the receiptand payment of interest is the nucleus of thesystem, but it is supported by other principlesof Islamic doctrine advocating risk sharing,individuals’ rights and duties, propertyrights, and the sanctity of contracts (Iqbal &Abbas, 2007).

With most explosive growth incapitalization and estimated at nearly 1trillion USD funds management, the branchbegins to gain followers and renownedglobal scale: some authority as in Germany,British Prime Minister, the Japanesemonetary authorities and even the Frenchmonetary authorities begin to show interestin Islamic finance.

In our research, the question relative tostudy the situation of Islamic banksfollowing during of the financial crisis of2007 was answered.

This paper will focus on the answer to thisquestion. In addition, a literature review onthe study of the performance of the Islamicand the difference between Islamic banksand conventional banks in the secondsection, was developed. The third sectionwas devoted to present the researchmethodology and empirical model used inour paper. In the third section, empiricalresults were interpreted. Finally, weconcluded in the last section.

2. LITERATURE REVIEW: ISLAMIC

FINANCE IN CRISIS PERIOD

The study by Chia and Wang (2008) andDar and Presley (1999) showed how acyclical fluctuation come from theapplication of interest rates in theconventional banks, unlike the paradigm ofIslamic economics as it prohibits the

application of interest rates which in turncontribute less to the economy.

Kaminsky and Reinhart (1996) studiedthe relationship from the point of macro-economic perspective. They are choosing theinflation rate and the GDP rate in their study.

According to Demirguc-Kunt et al,(2006), Chapra (2000), El-Gamal (2000) andDemirguc-Kunt and Detragiache (1998), itwas conclude that the presence of economicinstability transactions based on interestcharged by commercial banks should bereplaced by the system of profit and lossapplied by Islamic banks.

According to Hasan and Dridi (2010), theonset of the global crisis of 2007 hasrenewed interest, specifically, the resilienceof the Islamic banking industry during crises.Their empirical work is based on the study ofthe relationship between Islamic finance andfinancial stability. The nature of Islamicfinance and their bases regulation protect theIslamic bank to the impact of the financialcrisis.

Aisyah (2009) and De Nicolò et al. (2006)studied the relationship between bank sizes,prime of risk and insolvency risk for asample of listed banks in 21 industrializedcountries. The results showed that banksoperating in the most developed countrieshave a higher risk of insolvency. The variousmethods used by researchers to study thestability of banks based approaches using theZ-score model. The Z-score has been widelyused in the empirical literature regarding theextent and determinants of the safety andsoundness of financial institutionsspecifically Islamic banks. Among thestudies using Z-score model in their analysisare Hasan and Dridi (2010), Karwowski(2009), Laeven and Levine (2009), Čihákand Hesse (2008), Demirguc-Kunt et al.(2006), Boyd and Nicolo (2006), Hesse and

90 A.Derbali / SJM 10 (1) (2015) 89 - 108

Čihák (2006) and Boyd and David (1993).There is a vast theoretical and empirical

literature on Islamic finance in general andspecifically Islamic banks.

The test of Čihák and Hesse (2010) wasmade to study the stability of Islamic bankscompared to conventional banks. Errico andFarahbaksh (1998) and Sole (2007)conducted a discussion on regulatory issuesrelated to Islamic finance. This general lackof academic work on Islamic finance is incontrast with the growing importance ofIslamic banks in many Muslim countries inAsia and Africa.

Comparing indicators of profitability,asset quality and stability of conventionaland Islamic banks, we see few significantdifferences between the two groups of banks.While it was find that Islamic banks are moreprofitable in the case of several countries.Therefore, conventional banks are moreprofitable than Islamic banks in the countriesthat contain the two types of banks (Beck etal., 2010).

The difference between the activities ofIslamic banks and conventional banks inIslamic countries can be translated in Table1.

3. DATA AND METHODOLOGY

This section was devoted to thepresentation of the sampling methodology,on the one hand, and the presentation of themodel used on the other.

3.1. Data

The objective of this research is to test ourhypothesis research focuses on determiningthe impact of the financial crisis on Islamicbanks.

In other words, it was tested whetherIslamic banks can be considered as a solutionto the failure of the conventional financialsystem.

Indeed, a sample of 29 Islamic banks(Appendix 1) spread over 7 countries(United Arab of Emirates (UAE), Bahrain,Jordan, Kuwait, Malaysia, Saudi Arabia,Turkey) during the period of 7 years (2006-2012), was chosen. Thus, these banks arechoosed in our study because they are themost banks in term of capitalization. It waschoosen to use the data from the 7 countriesin this paper because they are Islamiccountries and they are not affected directlyby the financial crisis of 2007. For the

91A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 1. Functioning of Islamic banks and Conventional banks

Source: Islamic banks and Conventional banks (annual reports)

context of crisis, it was choosen to study aperiod post and during financial crisis of2007.

In fact, our empirical test is to measure theposition of Islamic banks after the outbreakof the financial crisis. Therefore, twocategories of variables related to the specificcharacteristics of banks and the macro-economic indicators, were used.

3.2. Model

Islamic finance is one of the fastestgrowing segments in the global financialsector. Several factors have contributed tothe growth of Islamic finance, including (i)the high demand in many Islamic countriesfor products compliant with the Shariah, (ii)the progress made in the regulationframework the Islamic finance, (iii) theincreasing demand of conventional investorsincluding for the purpose of diversification,and (iv) the ability of the industry to developa number of financial instruments that meetmost of the needs of investors institutionaland individual. The size of Islamic bankingin the global industry was estimated at 820billion $ in 2008.

In our research, we will refer to the modeldeveloped by Hasan and Dridi (2010), inwhich they studied the impact of thefinancial crisis on conventional banks andIslamic banks. Their study focuses on agroup of 120 banks for a period of 3 years(2007-2009). They were used for the 4models performance measurement;profitability, changes in net assets, changesin appropriations and external ratings. Thesample is composed of 90 conventionalbanks and 30 Islamic banks.

However, the model developed by thetwo authors to test the impact of the financialcrisis on the performance of Islamic banking

model was used. Then, in our paper a samplecomposed only by Islamic banks and for aperiod of 7 years was used. Thus, it waschoosen to analyze the Islamic banks whichexisted in Islamic countries.

The model used is in the following form:

DependentVariableijt = f (Bankspecificsijt,Macrovariablesjt)

Three measures of the performance ofIslamic banks, were used, which arepresented as follows:

Profitabilityijt: is the relationship betweenprofits at time t and profits at the time (t-1)for bank i belong to country j.

CreditGrowthijt: is the ratio between loansat time t and credits at the time (t-1) for banki belong to country j.

AssetGrowthijt: is the ratio between theassets at time t and assets at time (t-1) forbank i belong to country j.

The explanatory variables used in themodels to be estimated are grouped into twocategories:

- The specific characteristics of banks:Investijt: is the ratio between the amount

of investment and total assets of bank ibelonging to country j at time t.

Leverageijt: is the ratio between theamount of capital and total assets of bank ibelonging to country j at time t.

ROAijt: is the ratio between the netincome and total assets of bank i belongingto country j at time t (ROA: Return onAssets).

- Macroeconomic variables:GDPjt: is the growth rate of GDP of

country j at time t.

92 A.Derbali / SJM 10 (1) (2015) 89 - 108

INFjt: is the inflation in country j at time trate.

UAEit: is a dummy variable for bank i attime t. It takes the value 1 if bank i belong tothe UAE and 0 if not.

Bahrainit: is a dummy variable of bank iat time t. It takes the value 1 if bank i belongto Bahrain and 0 if not.

Jordanit: is a dummy variable of bank i attime t. It takes the value 1 if the bank ibelongs Jordan and 0 if not.

Kuwaitit: is a dummy variable of bank i attime t. It takes the value 1 if bank i belongsto Kuwait and 0 if not.

Malaysiait: is a dummy variable of bank iat time t. It takes the value 1 if bank i belongsto Malaysia and 0 if not.

Saudiit: is a dummy variable of bank i attime t. It takes the value 1 if the bank ibelong to Saudi Arabia and 0 if not.

Turkeyit: is a dummy variable of bank i at

time t. It takes the value 1 if bank i belong toTurkey and 0 if not.

The estimated models are threefold andare presented in the following text:

Model 1:

Profitabilityijt = α0 + α1tInvestijt +α2tLeverageijt + α3tROAijt + α4tGDPjt +α5tINFjt + α6tUAEit + α7tBahrainit +α8tJordanit + α9tKuwaitit + α10tMalaysiait +

α11tSaudiit + α12tTurkeyit + εijt

Where: αpt: are the coefficients of the explanatory

variables (p = 1, ..., 12) and (t = 1, ..., 7).α0 : is a constant.i: the index for each bank (i = 1, ..., 29).j: the index for each country (j = 1, ..., 7)εijt: the error term.

Model 2:

CreditGrowthijt = β0 + β1tInvestijt +β2tLeverageijt + β3tROAijt + β4tGDPjt +β5tINFjt + β6tUAEit + β7tBahrainit +β8tJordanit + β9tKuwaitit + β10tMalaysiait +

β11tsaudiit + β12tTurkeyit + ωijt

Where:βqt : are the coefficients of the explanatory

variables (q = 1, ..., 12) and (t = 1, ..., 7).β0: is a constant.i: the relative index to each bank (i = 1, ...,

29).j: the index for each country (j = 1, ..., 7)ωijt: the error term.

Model 3:

AssetGrowthijt = λ0 + λ1tInvestijt +λ2tLeverageijt + λ3tROAijt + λ4tGDPjt +λ5tINFjt + λ6tUAEit + λ7tBahrainit +λ8tJordanit + λ9tKuwaitit + λ10tMalaysiait +

λ11tSaudiit + λ12tTurkeyit + ϕijt

Where:λst : are the coefficients of the explanatory

variables (s = 1, ..., 12) and (t = 1, ..., 7).λ0: is a constant.i: the relative index to each bank (i = 1, ...,

29).j: the index for each country (j = 1, ..., 7)ϕijt: the error term.

4. RESULTS

Throughout this section, the differentempirical results are presented. First, it wasdecided to present the various descriptivestatistics. Then, the correlation matrix was

93A.Derbali / SJM 10 (1) (2015) 89 - 108

presented. Finally, the different estimationresults of the three models were interpreted.

4.1. Descriptive statistics

Throughout this section it was tried toanalyze and interpret the different resultsobtained from the estimates made on thethree variables Profitability, CreditGrowthand AssetGrowth.

Therefore, the models used in our paperwill be estimated by a regression with PanelData. The choice of this type of regression isjustified by the presence of the twodimensions in the data used; the firstdimension is time (a period of 7 years) andthe second is individual (29 Islamic banksused in this study).

The Table 2 summarizes the descriptivestatistics for each variable used in theestimation of the three models.

The Profitability variable, whichexpresses the level of profit of each Islamicbank throughout the study period, can reacha maximum value of 11.33624, as itsminimum value is (-8.836468). The volatilityof the variable profitability, which ismeasured by the standard deviation, is1.325935.

Thus, the variable CreditGrowthmeasuring the growth rate of loans to clients(individual or company) has a maximumvalue of 96.76 and a minimum value of(-290.02). Level of risk is 27.14795.

In addition, the variable AssetGrowthmeasuring the rate of asset growth has amaximum value of 88.77 and a minimumvalue of (-16.8). Thus, the level of risk is16.2307.

Other statistics on other variables werepresented in Table 3.

In continuation of the analysis of theempirical resultsa test of correlation betweenthe variables used was conducted. Table 3summarizes the results. Furthermore, theresults show that all of Pearson correlationcoefficients do not exceed the tolerance limit(0.7), which does not cause problems withthe estimation of the three models.According to the correlation matrix, it wasremarked that some coefficients areinsignificant. This implies that thedependence is low between the variables forwhich the coefficient is insignificant andlow.

All banks used as the investigation sampleare presented in the Appendix 1. In addition,by observing the Appendix 2 it was

94 A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 2. Descriptive statistics

95A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 3. Correlation matrix

concluded that all Islamic banks in oursample have a positive level of profitability.Similarly, Appendix 3 and Appendix 4,which respectively represent the rate ofcredit growth and the rate of asset growth ofIslamic banks during the period of study(2006- 2012). In fact, the resources collectedfrom agents with financing capacity areadequate to meet the demands of creditagents to financing needs.

4.2. Estimation results

The results of the unit root, for themodeling procedure. are presented in Table4. The results of the estimation of thevariable Profitability, CreditGrowth andAssetGrowth are presented respectively inthe three tables 5, 6 and 7. These tablesinclude several specifications. Thedependent variables were estimated by usingthe explanatory variables on the specificcharacteristics of Islamic banks and variablesthat represent macro-economic indicators.

Note here that the panel structure ishomogeneous. In this case, the method ofordinary least squares, was applied, thatallows a better fit by minimizing the sum ofsquared residuals.

Therefore, three models in which threevariables were addopted were estimated:Profitability, CreditGrowth and AssetGrowthas depended variables. The estimation resultsof the three OLS models are presented in thetables 5, 6 and 7.

A test of the unit root panel data wasconducted. Thus, it was decided to use thetest of Levin Lin Chu. The null hypothesis ofthis test is H0: all series are non-stationaryand the alternative hypothesis is H1: allseries are stationary. The reject of the nullhypothesis is based on the value of the p-value. This value is compared to a threshold

of 10%. If the value of the p-value is lessthan 10%, then we reject H0 and if the valueof the p-value is greater than 10%, then wereject the alternative hypothesis H1.

In our case, it was noticed that the valuesof p-values on different variables are lessthan 10%. In this case, one rejects H0 andthereafter all these variables are stationary.

The dummy variables (Dummy) areconsidered stationary from the beginning.Subsequently, a test of stationary for thesevariables was not done.

The problem that arises when estimatingis the choice of the estimation method;estimating a fixed effects model or theestimation of a random effects model. Thus,approached this problem solution is theHausman test which allows you to choosebetween the estimation of a fixed effectsmodel and the estimation of a random effectsmodel.

In the model 1 (Profitability), it waschosen to use the random effects model as anestimation method in both estimations. Thischoice is justified by the probability of theHausman test which is equal to 0.5459 in thefirst estimate and 0.6269 in the secondestimate. Thus, this probability is greaterthan 10%, therefore, it was chosen therandom effects model.

Next, the tests of autocorrelation of thefirst order of each estimated model were

96 A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 4. Testing the unit root

conducted. This test is based on theinterpretation of the probability value (Prob>F). This value is compared to a threshold of5%. If the probability is less than 5%, rejectH0 therefore, that is to say it rejects thehypothesis of absence of the self-correlation

of the first order. In this case, we will correctthis problem in the presence ofautocorrelation.

In the model 1 (Profitability), theprobability value (Prob> F = 0.0002) is lessthan 5% in the two estimates of this model

97A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 5. Estimation of the Model 1

and it was correct this problem, which ispresented in Table 5. Within this framework,there is no a problem of correlation betweenthe explanatory variables and residuals.

It was also decided to conduct additionaltests to show the validity of the estimatedmodels and justify the significance of theestimates. It was used to test the correlationbetween the explanatory variables and theresiduals. This type of test is based on thevalue (Prob> chi2). If this probability is lessthan 5%, so it was accept H0 that verifies theabsence of correlation between the residualsand the explanatory variables. If thisprobability is greater than 5% in this casethere is a problem of correlation between theresiduals and the explanatory variables thatmust be corrected.

In both estimates of the model (1), theprobability values (Prob> chi2) are all lessthan 5%. So there are no problems ofcorrelation between the explanatoryvariables and residuals.

The test of significance of the model isbased on the probability of Fisher. It wasnoticed that all the probability value Fisher isless than 5% in all estimates first model. Sowe can deny that the estimated model 1(Profitability) is generally significant.

Thus, it was found that the coefficient ofdetermination R2 is equal to 0.7190 and0.6269 in the two estimates made, so themodel (1) is characterized by a good linearfit.

From Table 5, it was remarked that thereare two significant variables e.g., thevariable Leverage and ROA variable.

However, the variable Leverage isnegatively statistically significant at athreshold of 10% with a value of t-studentwho is (-1.94) in the first estimation and a10% threshold with a value of t-student isequal to (-1.66) in the second estimation. So

the variable, which measures the ratiobetween capital and total assets of Islamicbanks, negatively influences the dependentvariable Profitability, measures the rate ofgrowth in profitability of Islamic banks. Inthis case, the more this ratio increases as thelevel of profit of Islamic banks decreases.

The second variable ROA has a positiveimpact on the variable Profitability. TheROA ratio is statistically significant at the5% level with a t-student value which isequal to (2.31) in the first estimate and athreshold of 5% with a value of t-studentwhich is equal to (1.99) in the secondestimate. In this case, the increase in thelevel of economic efficiency has a positiveimpact on the profitability of Islamic banks.

Macroeconomic variables have a positiveimpact on the dependent variable, but it isinsignificant who justifies that volatility ofthe economic indicators did not affect theprofitability of Islamic banks.

In this case, and for the first model, it wasaccepted that the second hypothesis has noimpact of the 2007 financial crisis on Islamicbanks. Therefore, Islamic banks can beconsidered as a solution for conventionalfinancial model.

And note that the dummy variableTurkeyit was not retained in the twoestimates, which is a problem because of co-linearity with other variables.

The estimation of the model (2) ispresented in Table 6.

In the Model 2 (CreditGrowth), it waschosen to use the fixed effects model for thefirst estimation and the random effects modelfor the second estimation. This choice isjustified by the probability of the Hausmantest which is equal to 0.0330 in the firstestimation and 0.1438 in the secondestimate. Thus, the probability is less than10% in the first estimation and is more than

98 A.Derbali / SJM 10 (1) (2015) 89 - 108

10% in the second estimation. Then, it waschosen to use the fixed effects model in thefirst estimation and the random effects modelin the second estimation.

Next, it was decided to conduct tests ofautocorrelation of the first order of each

estimated model. This test is based on theinterpretation of the probability value (Prob>F). This value is compared to a threshold of5%. If the probability is less than 5%, rejectH0 therefore, that is to say it rejects thehypothesis of absence of the self-correlation

99A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 6. Estimation of the model 2

of the first order. In this case, it was correctthis problem in the presence ofautocorrelation.

In the Model 2 (CreditGrowth), theprobability value (Prob> F = 0.0000) is lessthan 5% in the two estimates of this modeland it was corrected this problem, which ispresented in Table 7. Within this framework,there is not a problem of correlation betweenthe explanatory variables and residuals.

It was also decided that additional testsshould be conducted to show the validity ofthe estimated models and justify thesignificance of the estimates. The correlationbetween the explanatory variables and theresiduals was tested. This type of test isbased on the value (Prob>chi2). If thisprobability is less than 5%, so it was acceptH0 that verifies the absence of correlationbetween the residuals and the explanatoryvariables. If this probability is greater than5% in this case there is a problem ofcorrelation between the residuals and theexplanatory variables that must be corrected.

In both estimates of the model (2), theprobability values (Prob>chi2) are all lessthan 5%. So there are not problems ofcorrelation between the explanatoryvariables and residuals.

The test of significance of the model isbased on the probability of Fisher. It wasnoticed that all the probability value Fisher isless than 5% in all estimates first model. Soit was deny that the estimated model 2(CreditGrowth) is generally significant.

Thus, it was found that the coefficient ofdetermination R2 is equal to 0.7639 and0.6236 respectively in estimation 1 and 2 inthe estimate, so the model (2) ischaracterized by a good linear fit.

From Table 6, it was remarked that thereis only one significant variable as Leverage.

However, the variable is statistically

significant if Leverage is positive to athreshold of 1% with a value of t-student thatis equal to (2.97) in the first estimation;while it is statistically significant to anegative threshold of 5% with a value of t-student which is equal to (-2.35) in thesecond estimate. So the variable, whichmeasures the ratio between capital and totalassets of Islamic banks, positively influencesthe level of loans to customers in the absenceof macro-economic indicators. While thisvariable have a negative impact on thedependent variable which measures the rateof growth of credit extended by Islamicbanks.

Macroeconomic variables have a negativeimpact on the dependent variable except forthe growth rate of GDP which has a positiveimpact, but it is not significant, that justifiesthe fact that the financial crisis of 2007 hasno impact on profitability of Islamic banks.In this case, and the second model, it waschosen to accept the first hypothesis of thepresence of impact of the financial crisis of2007 on Islamic banks. Therefore, Islamicbanks can be considered as a solution forconventional financial model under thecondition that the state of the economy isexpanding.

And note that the dummy variableTurkeyit was not retained in the twoestimates of a problem because of co-linearity with other variables.

The estimation of the model (3) ispresented in the Table 7.

For the model 3 (AssetGrowth), wechosed the fixed effects model for the firstestimation and the random effects model forthe second estimation. This choice isjustified by the probability of the Hausmantest which is equal to 0.0001 in the firstestimate and 0.2204 in the secondestimation. Thus, the probability is less than

100 A.Derbali / SJM 10 (1) (2015) 89 - 108

10% in the first estimation and is more than10% in the second estimation. Because ofthis; it was chosen to use the fixed effectsmodel in the first estimation and the random

effects model in the second estimation.Next, it was decided to conduct tests of

autocorrelation again.For model 3 (AssetGrowth), the

101A.Derbali / SJM 10 (1) (2015) 89 - 108

Table 7. Estimation of the model 3

probability value (Prob>F = 0.0001) is lessthan 5% in the two estimates of this modeland it was corrected for this problem, whichis presented in Table 9. Within thisframework, there is no problem ofcorrelation between the explanatoryvariables and residuals.

It was also decided to conduct additionaltests to show the validity of the estimatedmodels and justify the significance of theestimates. It was based on test of thecorrelation between the explanatoryvariables and the residuals, as described inprevious text.

In both estimates of the model (3), theprobability values (Prob> chi2) are all lessthan 5%. So there are no problems ofcorrelation between the explanatoryvariables and residuals.

The test of significance of the model wasbased on the probability of Fisher. It wasnoticed that all the probability values ofFisher are less than 5% in all estimates of thefirst model. So we can deny that theestimated model 3 (AssetGrowth) isgenerally significant.

Thus, it was found that the coefficient ofdetermination R2 is equal to 0.6707 and0.6611 in the two estimates made, so themodel (3) is characterized by a good linearfit.

From Table 7, it was showed that there arefour significant variables Leverage,Leverage variable, the dummy variable(Jordan) and the dummy variable (Kuwait).

Invest, as the first variable has a positiveimpact on the variable AssetGrowth. Thisvariable is statistically significant at the 10%level with a t-student value which is equal to(1.90) in the second estimate. In this case, theincrease in the value of investments ofIslamic banks can have a positive impact onthe growth rate of assets of Islamic banks.

However, if the investments are increasing,they allow Islamic banks to increase thevalue of their assets.

However, the variable Leverage isstatistically significant positively to athreshold of 1% with a value of t-student thatis equal to (3.99) in the first estimate and isstatistically significant at a threshold ofpositive 1% with a value of t student-whichis equal to (7.10) in the second estimate. Sothe variable, which measures the ratiobetween capital and total assets of Islamicbanks, positively influences the growth rateof assets. More this ratio is increasing thevalue of assets in turn increases.

The two dummy variables Jordan andKuwait have an impact on the growth rate ofassets in Islamic banks. In this case, thegrowth in the value of assets is determinedby the state of the economies of Jordan(positive impact) and Kuwait (negativeimpact).

Other macroeconomic variables have anegative impact on the dependent variable,but it is not significant to justifie that thefinancial crisis of 2007 has an impact on theprofitability of Islamic banks. In this case, ofthe third model, we accepted the firsthypothesis of the presence of impact of thefinancial crisis of 2007 on Islamic banks.Therefore, again, Islamic banks can beconsidered as a solution for conventionalfinancial model under the condition that thestate of the economy is expanding. Whilethis impact does not make sense since theprofitability of Islamic banks is stillincreasing.

It should be noted again that the dummyvariable Turkeyit was not retained in the twoestimates of a problem because of co-linearity with other variables.

102 A.Derbali / SJM 10 (1) (2015) 89 - 108

5. CONCLUSION

Islam is a comprehensive way of life,which strikes the balance between thespiritual and the material need of humanbeing. One of the important aspects in humanlife is the need for a comprehensive systemin order to govern the life and to ensure allthe needs are catered adequately includingthe material needs such as the financialmanagement. This aspect of life is closelyrelated to the fast growing industry in theworld nowadays, which is the Islamicfinancial services industry.

In this study, it was answered to afundamental question that leads us todetermine the impact of the financial crisis of2007 on the functioning of Islamic banks.

This paper is answering that question. Inaddition, a second section was developed, inwhich it was tried to present a literaturereview on the study of the performance ofthe Islamic banks and the difference betweenIslamic banks and conventional banks.

Third section was dedicated to present theresearch methodology and the models whichwill be estimated. For this section, theSTATA 12 software was used to get differentresults that allow us to respond all objectivesof our paper.

Following our study, which wasdeveloped on the situation of Islamic banksafter the outbreak of the financial crisis of2007, we can assume that Islamic banks arenot affected by the international turbulence.

References

Aisyah, A.R. (2009). Lending Structureand Market Risk Exposures: The MalaysianCase. Asian Academy of ManagementJournal, 14 (2), 1-20.

Alasrag, H. (2010). Global Financial crisisand Islamic finance. Working Papers No.22167. Online at http://mpra.ub.uni-muenchen.de/22167/.

Beck, T., Demirguc-Kunt, A., &Merrouche, O. (2010). Islamic vs

103A.Derbali / SJM 10 (1) (2015) 89 - 108

ИСЛАМСКО БАНКАРСТВО ТОКОМ

ФИНАНСИЈСКЕ КРИЗЕ ОД 2007

Abdelkader Derbali

Извод

Сврха овог рада је да иницијално допринесе литературном повезивању светске економскекризе и перформанси Исламских банака. Од глобалног је значаја проучавати тренутну светскукризу са циљем развоја одрживих финансијских пракси, у циљу потраге за новим пословниммоделима, заснованим на заједничком учешћу у профиту и губицима.По мишљењу аутора,одржива финансијска пракса се треба да заснива на трансакцијама без провизије и на применизаједничког удела у профиту и губицима у финансијском систему. У овом раду, тестиране суперформансе исламских банака у периоду кризе. Коришћен је узорак од 29 Исламских банакаиз 7 земаља у периоду од 7 година (2006-2012). На основу емпиријских резултата, закљученоје да Исламске банке нису погођене финансијском кризом.

Kључне речи: финансијска криза, исламско финансирање, конвенционално финансирање,перформансе, панел подаци

conventional banking business modelefficiency and stability. World bank WP, No.5446.

Boyd, J., & Nicolo, D. (2006). The Theoryof Bank Risk Taking Revisited. Journal ofFinance, 60, 1329-1343.

Boyd, J.H., & David, E.R. (1993). Sizeand Performance of Banking Firms. Journalof Monetary Economics, 31, 47-67.

Chapra, M.U. (2000). Why has IslamProhibited Interest: Rationale Behind theProhibition of interest. Review of IslamEconomic, 9, 5-20.

Chia, D.J., & Wang, Y. (2008). Risk andRegulation of Islamic Banks: A Perspectivefrom a NonIslamic Jurisdiction. PrudentialPolicy Department Monetary Authority ofSingapore. Paper published in “Theory andPractice of Islamic Finance,” MAS StaffPaper No. 49, December, 2008, Singapore.

Čihák, M., & Hesse, H. (2010). IslamicBanks and Financial Stability: An EmpiricalAnalysis. IMF Working Papers, No. 10/16.

Čihák, M., & Hesse, H. (2008). IslamicBanks and Financial Stability: An EmpiricalAnalysis. IMF (International MonetaryFund), Working Paper 08/16 (Washington:International Monetary Fund).

Dar, H.A., & R. Presley, J.R. (1999).Islamic Finance: A Western Perspective.International Journal of Islamic FinancialServices, 1 (1), 1-9.

De Nicolò, G., Laeven, L., & Ueda, K.(2006). Corporate Governance Quality:Trends and Real Effects. IMF WorkingPapers, No. 06/293.

Demirguc-Kunt, A., & Detragiache, E.(1998). The Determinants of Banking Crisesin Developing and Developed Countries. IMFStaff Papers, 45 (1), 81-109.

Demirguc-Kunt, A., Detragiache, E., &Thierry, T. (2006). Banking on the Principles:Compliance with Basel Core Principles and

Bank Soundness. Retrieved 1 January 2011from http://siteresources.worldbank.org/INTFR/Resources/Banking_on_the_PrinciplesCompliacewithBaselCorePrinciples.pdf

El-Gamal, M. (2000). An economicexplicaiton of the prohibition of riba inclassical Islamic jurisprudence. The ThirdHarvard University Forum on IslamicFinance. Harvard University Center forMiddle-Eastern Studies.

Errico, L., & Farahbakch, M. (1998).Islamic Banking: Issues in PrudentialRegulations and Supervision. IMF WorkingPapers, No. 98/30.

Hasan, M., & Dridi, J. (2010). The Effectsof the global crisis on Islamic andConventional banks: A comparative study.IMF WP 10/201.

Hesse, H., & Čihák, M. (2006).Cooperative Banks and Financial Stability.IMF Working Papers, No.07/1.

Iqbal, Z., & Abbas, M. (2007),Introduction to Islamic Finance: Theory andPractice, John Wiley and Sons (Asia),Singapore.

Kaminsky, G.L., & Reinhart, C.M. (1996).The Twin Crises: The Causes of Banking andBalance-of-Payments Problems. InternationalFinance Discussion Paper No. 544, Board ofGovernors of the Federal Reserve System.

Karwowski, E. (2009). Financial Stability:The Significance and Distinctiveness of LevyEconomics Institute Working PaperCollection Islamic Banking in Malaysia.Working Paper No. 555, University ofLondon.

Laeven, L., & Levine, R. (2009). BankGovernance, Regulation and Risk-Taking.Journal of Financial Economics, 93, 259-75.

Sole, J. (2007). Introducing Islamic Banksinto Conventional Banking Systems. IMFWorking Paper 07/175 (Washington:International Monetary Fund).

104 A.Derbali / SJM 10 (1) (2015) 89 - 108

105A.Derbali / SJM 10 (1) (2015) 89 - 108

APPENDIX 1.

List of banks used in sample

106 A.Derbali / SJM 10 (1) (2015) 89 - 108

APPENDIX 2.

The profitability of Islamic banks

107A.Derbali / SJM 10 (1) (2015) 89 - 108

APPENDIX 3.

Credit Growth in Islamic banks

A.Derbali / SJM 10 (1) (2015) 89 - 108108

APPENDIX 4.

Assets Growth in Islamic banks