Performance differences between Islamic and conventional banking forms

This paper strives to recognize the possible performance differences between the two popular banking forms in the Gulf Cooperation Council (GCC) countries. Applying different methodologies on the data that span the period 2003–2015, this study documents significant differences with respect to the period, countries, and performance measures. Specifically, conventional banks in GCC countries outperform their Islamic counterparts in profitability. Also, bank specific factors such as liquidity, capital adequacy, bank size and growth all affect the profitability. In addition, GCC conventional and Islamic banks were isolated from the 2008 subprime crisis even though their profitability seems to be decayed differently over the period of the economic downturn. Turki Alshammari (Kuwait) BUSINESS PERSPECTIVES LLC “СPС “Business Perspectives” Hryhorii Skovoroda lane, 10, Sumy, 40022, Ukraine www.businessperspectives.org Performance differences


INTRODUCTION
The recent global financial crisis has attracted the attention to the Islamic banking as a different type of banking that mitigates the mismatch of short-term, on-sight demand deposits contracts with longterm uncertain loan contracts (with equity elements) (Cihak et al., 2010).According to the Islamic Financial Service Board (IFSB) report (May 2017), assets of Islamic banks totaled US$1.89 trillion in 2016 and grew 50% faster than the overall banking sector with an average annual growth of 17.6% from 2008 to 2012.Further, Islamic bank assets are expected to reach US$3.4 trillion by 2018 (Ernst & Young, 2013) and US$6.5 trillion by 2020 (IFSB, 2017).
Although the function of the intermediation process is the same, the business model of Islamic banks differs largely from their conventional counterparts.This difference may render diverse (and consistent) performance levels across time.Mainly, Islamic banks business model is based on the concept of reciprocal profit and loss sharing among related parties on both the liability and the asset side.In addition, financial transactions that involve interest payment and speculative trading are totally prohibited.Hence, the popular loan granting as a common financing form is prohibited.
For that, this paper strives to scan any differences in performance between two forms of banking, the Islamic banking form and the conventional banking form in a homogeneous environment, the Gulf Cooperation Council (GCC) countries.GCC countries include Kuwait, Saudi Arabia (KSA), Bahrain, Qatar, United Arab Emirates (UAE), and Oman.These countries share, to a high extent, an integrated economic environment as many economic policies are outstandingly coordinated before applied.In addition, GCC continues to be as the largest domicile for Islamic financial assets as it has experienced very recently a further increase in market share to 42.3% of the global Islamic financial assets.This paper shall contribute to the developing literature in the sense that it will shed light on the argumentative literature that relates to corporate performance trajectories across time and how different factors affect that performance, which completes a building block of the literature.The remainder of the paper is structured as follows: section one presents the literature review, where section two covers data and methodology of this research.The empirical results are contained in section three and final section presents the concluding remarks.

LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
Numerous researchers have conducted studies to analyze the financial performance determinants of conventional banks 1 .At the same time, researchers also examined performance of Islamic banks.Milhim and Istaiteyeh (2015) analyze the performance of conventional banks versus Islamic banks in Jordan during the period 2009-2013.As they employed the financial ratio analysis (profitability, liquidity, solvency and efficiency), they document significant differences in performance of Islamic and conventional banks.Specifically, they find that Islamic banks are less profitable, more liquid and less efficient as compared to conventional banks.One can glean from the literature review above a conjecture that there exists a difference between Islamic and conventional banks in many different business settings.This paper strives to explore any differences in financial performance employing a distinctive methodology and more recent data in order to detect any alleged differences in performance.The sample is unique, as it comprises country evidence, as well as regional evidence with respect to bank performance.Countries included in this study comprise the GCC countries where about 42% of all Islamic banks assets are domiciled.Moreover, GCC Islamic banks contribute about 70% of the growth rate in Islamic banking assets in 2014 (IFSB, 2017).Hence, the milieu of the two banking forms provides an inimitable and peerless setting by which one can analyze differences in bank performance during an extended period, before and after the 2008 subprime crisis, as the integrated environment minimizes the effect of any external shocks.Based upon that, the following hypotheses are of concern: H1: There are no differences between Islamic and conventional banks in terms of financial performance.
H2: Using different financial performance measures does not influence the obtained results.
H3: The obtained results are not sensitive to country selection.

DATA AND METHODOLOGY
This paper uses year-end financial statement data that spans the period 2003-2015 and produced by the Kuwait Institute of Banking Studies financial database (KIBS).The extended 13 years of data renders more credibility in the results of the analysis and allows examining the influence of the 2008 subprime crisis on bank performance2 .
After some preliminary analysis, it was sought to focus on three (out of six) countries of the GCC countries, namely, Kingdom of Saudi Arabia (KSA), Kuwait and United Arab Emirates (UAE), for most of the analysis.Banking systems, assets level, regulations and activities are relatively massive in these three countries relative to those in the excluded countries.Besides, approximately 58.5% of the international participation banking assets are based on these three countries (IFSB, 2017).

EMPIRICAL RESULTS
The thrust of this paper is to analyze bank performance of two banking forms in the GCC countries.Table 2 presents performance differences between conventional and Islamic banks in the GCC countries in three different time intervals, name- The evidence in Table 3 suggests that the decrease in conventional banks' profitability was more severe than that of the Islamic banks, which seems to withstand the financial shock caused by the 2008 economic downturn, a result that is supported by Cihak et al. (2010).
In order to identify the sources of differences in the performance, Table 4   8.75%, a result that shows conventional banks to plow back more profit, hence, their growth potential is higher than that of the Islamic banks in the GCC countries.
Considering the other two periods, the 2008 and the 2013-2015, do not change the conclusions found earlier in terms of statistical significance.This indicates some consistency and reliability in the obtained results.
Because the significance of the influence of the sources of profitability on bank performance may not be consistent across different performance measures and across the banking form, Table 5 presents a comparison of the effect of profitability sources on both ROA and ROE for both Islamic and conventional banks.Panel (A) presents the pooled model, while panel (B) presents the Seemingly Unrelated Regression (SUR) (or Zellner method) 4 .
Remarkably, for the conventional banks, the pooled as well as the SUR methods exceedingly fit the data and all variables have a significant effect irrespective of the expected sign.Besides, the significance levels are almost the same when ROE replaces ROA.The results show that the higher the liquidity, the higher the performance of the conventional banks.Specifically, an increase by 1% in liquidity would result in an increase of ROA by about 11.5%.One would expect a negative sign as liquidity is well thought-out an idle asset, although, if it is relatively higher, it tends to reduce the operating risk of a bank.In addition, liquidity by itself scales the risk of insufficient reserves of cash in response to unexpected withdrawal requests; hence, it positively affects ROA.The effect of liquidity on profitability for Islamic banks is insignificant.One interpretation is that Islamic banks tend to link their credit to real pledged assets (i.e., the profit-loss system), which guarantees the credit in case a customer fails to service the debt.However, for Islamic banks, the effect of liquidity becomes significant when ROE replaces ROA.For this, the effect of leverage is considered, hence, more liquidity matters.This result supports the evidence established by Hussein et al. (2014) 4 Since we have cross sectional data, and as data are structured such that we have three countries, hence, three groups, and as the popular fixed and random effect models require the number of cross sections to be larger than the number of coefficients, so the Seemingly Unrelated Regression (SUR) method is used as a reliable alternative.
As for capital adequacy, one can note a positive coefficient value of both Equity to Assets and Deposit to Equity.The higher capital adequacy safeguards a bank from insolvency (and reduces bankruptcy costs) and as higher level of capital increases a bank lending capacity, which should lead to a higher profitability.For Islamic banks, the relation between capital adequacy and bank profitability is less confirmed, as only Deposits to Equity ratio is significant (considering ROA) and only Loan to Equity ratio is significant (considering ROE).This conclusion is in line with that of Hussein et al. ( 2014), but comes contrary to the evidence provided by Altamimi et al. (2015).
Bank efficiency, as proxied by the ratio of fixed assets to assets, is expected to have a positive effect on profitability.At the same time, the ratio by itself leads to the indication of a negative relation to profitability, given the financial operative feature of banks.Thus, the functional relationship between the operational efficiency and profitability is expected to be with a positive sign, a result which is illustrated by Table 5 for conventional banks.As for Islamic banks, the same relation is not supported, even after considering the ROE.These findings are supported by Hussein et al. (2014).Moreover, results in Table 5 show that the bank internal growth proxy has positive effect on bank's profitability for banking forms.This is vividly expected as the more funds plowed back in the bank, the more capacity to lend or to strengthen the fi-nancial stability of the bank, and the higher will be the profitability of the bank.
In order to obtain additional evidence on the impact of bank size, bank type, home country of the bank, and the impact of the 2008 subprime crisis on the GCC banks' profitability, a regression of bank profitability on several variables is carried out, and the results are given in Table 6.Kuwait is excluded from the home country dummy variable and Islamic banks are excluded from bank type.Therefore, "BT" is conventional banks profitability relative to Islamic banks, while Dksa and Duae are bank profitability in these countries relative to banks in Kuwait.Pooled and SUR regressions are both run to examine the robustness of the results.
The results reported in Table 6 are consistent across the two regression equations (pooled OLS and SUR).The OLS pooled regression indicates a positive and significant effect of bank size on ROA, which indicates the larger the bank, the higher its profitability (which is supported by Hussein et al. (2014) and Mollah et al. (2017)).This relation is statistically insignificant when employing the SUR regression model.Note: CTA = Cash to Assets; ETA = Equity to Assets; DTE = Deposit to Equity; LTE = Loan to Equity; FTA = Fixed Assets to Assets; IGR = Internal Growth Rate (the % of end of year retained income divided by the beginning of the year bank equity fund); ROA (ROE) = Return on Assets (Equity); LNTA = Log of Total Assets.Since we have a cross-sectional data, heteroscedasticity was found to be significant.And since it is of unknown form, it was corrected by using White's (1980) consistent covariance matrix.
Consistent with the previous results, conventional banks in the GCC countries outperform the Islamic banks.The coefficient of the OLS pooled model shows the conventional banks to outperform their Islamic counterparts by 1.96% (in ROA) and by 0.87% when considering the pooled regression method.Consistent with previous analysis, there seems to be no effect of the 2008 subprime crisis on the profitability of Islamic and conventional banks in the GCC countries, except the perceived significant shrinkage of the conventional banks' profitability, noted in Table 3.This supports the evidence by Almanaseer (2014)  For the country dummy variables, the results in Table 6 show that Kuwaiti banks have significantly higher ROA than that of Saudi banks.Also, Kuwaiti banks' ROA is not significantly different from that of Emirates banks (considering the OLS pooled model), but higher than that of Emirates banks by almost 1.3% (when considering the SUR model, which seems more reliable as we have cross section data).

CONCLUSION
This paper analyzes the performance of both Islamic, as well as conventional banks in GCC countries.GCC economic environment is very integrated as many economic policies are highly coordinated among GCC countries.Hence, banks in these countries represent unique set of financial institutions that operate in special environment.The paper employs different statistical analysis tools such as mean test and different regression methods in order to corroborate the results and to indorse their credibility.
The paper predominantly documents vivacious differences between Islamic versus conventional banking form with respect to bank performance.The different regression models divulge the conventional banks to outperform their Islamic counterparts, irrespective of the period and regardless of the country as well as the performance measure.Furthermore, the direct determinants of the performance differ between the two banking forms.For example, as liquidity seems to affect the performance of conventional banking form, it does not appear to be a factor in the performance function of Islamic banks.Other Note: ROA = Return on Assets; Ln (TA) = natural logarithm of total assets; Dksa = 1 if bank is in Saudi Arabia and 0 otherwise; Duae = 1 if bank is in Emirates and 0 otherwise; D8 = 1 if year is 2008 and 0 otherwise; BT = 1 if bank is conventional and 0 otherwise.we have a cross-sectional data, heteroscedasticity was found to be significant.And since it is of unknown form, it was corrected by using White's (1980) consistent covariance matrix.
factors that seem to affect bank performance of GCC banks are bank efficiency, banks size and bank growth rate.The evidence in this analysis also shows that the 2008 subprime crisis seems to affect the performance of conventional banks but not the Islamic banks in GCC countries as the plodding reduction in banks performance was more overfed for conventional banks.
Note: CTA = Cash to Assets; ETA = Equity to Assets; DTE = Deposits to Equity; LTE = Loan to Equity; FTA = Fixed assets to Assets; IGR = Internal Growth Rate (the % of end of year retained income divided by the beginning of the year bank equity fund); ROA (ROE) = Return on Assets (Equity); Obs Con = Observations of conventional banks; Obs Islamic = Observations of Islamic banks; p-value = testing the null hypothesis that the mean difference is equal to zero.

Table 1
Assets are denominated in US dollars from the data source.One can glean from the table that a number of banks have not changed.The book value of their assets have increased considerably though.While Emirates have the largest number of banks (45%), it is apparent that its banks assets are less than those of Saudi Arabia.In addition, in the three countries, conventional banks are the largest in terms of total assets relative to Islamic banks.In addition, mean differences are used to test any differences in ROA and ROE.In addition, pooled regressions are employed to test any differences in the sources of performance for the two banking forms.Moreover, dummy variables are used to test for the performance differences between the two banking forms.Specifically, as in Cornett et al.
presents summary statistics of the data used in the study for the years 2011 and 2015.The variables used to measure and evaluate bank performance are ROA and ROE, where ROA shows the management ability to use bank assets in generating profits, while ROE reflects the effectiveness of management to use shareholders equity, although it neglects the financial leverage.Hence, both indicators indicate different outlook of profitability.(2010), bank liquidity, operating efficiency, capital adequacy and growth indicator are all employed.

Table 1 .
Number of all banks, total assets (in US$), number of conventional banks and the percentage of total assets in conventional banks

Table 2 .
Performance differences between conventional and Islamic banks in the GCC countries

Period Conventional banks Islamic banks Obs (conv) Obs (Islamic) Sig (p-value) If only Ku, KSA, UAE
the 2008 subprime crisis, before and after the 2008 subprime crisis.In general, the table figures assert clearly the superiority of the conventional banks over the Islamic banks in profitability.The last column shows the results if only three countries of GCC are considered (i.e., Kuwait, Saudi Arabia, and Emirates).The results resemble those in the table in terms of the significance level (5% level) for the whole period.Before the 2008 subprime crisis though, the results of the mean difference test are insignificant, except for the 2005.The same is true when considering only the three countries group.Nevertheless, the absolute value of ROA and ROE shows the dominance of conventional banks over Islamic banks in terms of profitability.In 2008, the results are reversed.their conventional counterparts (although not significant) confirms Tlemsani et al.'s (2016) conclusions in that Islamic banks were less hit as they are better capitalized 3 .
Shah (2014)al.(2014)nvalues of ROA of the relative banks, while those between parentheses relate to ROE; p-values that test the hypothesis of "equal means' are reported for ROA of the relative banks, while those between parentheses relate to ROE; Ku represents Kuwait, KSA represents Kingdom of Saudi Arabia, and UAE represents Emirates; Obs stands for "number of observations"; p-value = testing the null hypothesis that the mean difference is equal to zero.ly 3The coefficient of variations (standard deviation divided by ROA and ROE) was calculated and was found to be unconditionally higher for Islamic banks at all years, which indicates riskier Islamic banks relative to their conventional counterparts (confirmed byHussein et al. (2014)andShah (2014).In addition, using the nonparametric Wilcoxon ranksum test produced qualitatively similar results to those obtained by employing the t-test.Hence, it is not reported for brevity purposes.than

Table 4 .
Mean differences of some financial ratios that relate to both conventional and Islamic banks