The Impact of Global Financial Crisis on Islamic Banking System in Comparison to the Conventional Banking System
Chapter 3: Methodology
3.1 Research Design
This is a comparative study between Islamic banking and the conventional banking system within the period from 2005 to 2009, when the global financial crisis occurred. It involves comparative analysis of the performance of Islamic banks to that of the conventional banks. The choice of this research design is most appropriate because it helps explain the variation in the performance of the selected systems. Finally, in addition to revealing the systematic structure (invariance) for the cases under investigation, the research design allows generalization of the entire groups of Islamic and conventional banks.
3.2 Data Sampling and Collection
The data used in this study were collected from a variety of secondary sources and include panel data from 40 Islamic and 40 conventional banks for 2005-2009. This set of banks includes a synthesis of both listed and unlisted financial institutions. The overarching aim of this bank selection and subsequent data collection was to check the profitability of Islamic banking system compared to conventional banks both in listed and non-listed categories. Therefore, it was important to categorize banks as either Islamic or conventional for purpose of effective comparison, as recommended by Makkar and Singh (2013). Additionally, it was important to select banks operating from a particular geographic region as it is likely that they face the same operating environment (Munir, Perera and Baird, 2011). For this reason, the banks were selected from the Muslim territories where both Islamic and conventional banks are present. Banks from the Sultanate of Oman were excluded from the list of Islamic banks because full-fledged Islamic banks do not exist in this country (Jaffar and Manarvi, 2011).
3.3 Data Analysis
The primary subject of the bank performance evaluation is profitability (Tahir, 2007). There are various indicators of profitability, such as Operating Profit Ratio, Net Profit Ratio, Return on Share Capital, Return on Assets, and Return on Total Equity. These ratios play a critical role in the influence of profitability on share capital, total income, total equity, and so on (Jaffar and Manarvi, 2011). In addition to profitability, the equity-to-asset ratio plays a major role in assessing the performance of banks. Other indicators of performance include the proportion of customer deposits in relation to total liabilities and total equity. As these indicators could have differed between the Islamic financial institutions and the conventional ones, they were computed and a comparison made between the two groups of banks.
The second and most important step was to assess the changes of the performance indicators during the global financial crisis period. The study to illuminate on the performance during the said period was done through the calculation of the average annual growth rates between 2005 and 2009. The purpose of determining these ratios was to collect adequate data to enable the performance of CAMEL (Capital Adequacy, Asset Quality, Management Quality, Earning Ability and Liquidity) analysis, which is a standard for performance analysis of financial institutions (Rozzani and Rahman, 2013). The final part was to focus on the inferential statistics to determine whether there were significant differences between the islamic and conventional banks operating within the Gulf Cooperation Council (GCC) region.
The data collected was entered into the statistical software, and then a range of regression analyses were carried out to compare the performance of Islamic banks with respect to that of conventional banks using panel data. The panel data analysis was selected because it is useful in describing change over time (Semykina and Wooldridge, 2010). Additionally, panel data can be used to indicate performance trajectories and occurrence of events (Hsiao, 2014). As the study was looking for a superior estimate of the performance trends of banks during the global financial crisis, panel data becomes increasingly useful. In this study, the researcher modelled complex data to compare observations in banks, applied fixed parameters that varied with individual banks, which justified the application of the fixed effects model. Theis model can be explained using the standard linear model and regression theory (Greene, 2001).
According to George Box, All models are wrong, but some are useful (cited in Weinstein, 2012). Therefore, the characteristics of the data used for the models calibrated determined the type of inferences made. Consequently, the analysis in the present study was expected to provide a reasonable proximity between the assumptions of the regression model and the data. As such, an exploration of the important features of the data was essential to assess the proximity (Yang, Ward and Rundensteiner, 2003). Exploration of data involved summarizing them without reference to the model to choose the most appropriate model and to detect any unusual observations. For the purpose of this study, identification of such features in the modeling process was essential. In the regression analysis, the data exploration techniques included multiple time series plots, scatter plots and variable plots.
Once the data were found to be fit for analysis, the next focus was to conduct estimation and inference. The usual regression setup provided the estimator of variance parameter conveniently generated by statistical and econometric packages as analysis of variance (ANOVA) table (Spriensma, 2016).
3.3.1 Analytic techniques
This analysis involves estimating measures of central tendency such as average, median, standard deviation, etc. These tests compared the performance of Islamic banks to conventional banks in the Middle East and North Africa (MENA) and GCC region.
Analysis of variance (ANOVA)
This analysis was used to examine whether financial performance differed between the Islamic and conventional banks using the Capital Adequacy, Asset Quality, Management Quality, Earning Ability, and Liquidity variables.
Multiple Linear Regression (MLR)
The MLR model was used to examine the influence of each independent variable (the CAMEL variables) on the financial performance of the banks. The two banking systems were used as the dummy variables to determine the differences in the financial systems.
The MLR model:
Return on Assets= a1 + b1(Capital Adequacy)+ b2(Asset Quality)+ b3(Management Quality)+ b4(Bank Earnings)+ b5(Liquidity Ratio)+ b6(GDP)+ b7(INF)+ e
Return on Equity= a1 + b1(Capital Adequacy)+ b2(Asset Quality)+ b3(Management Quality)+ b4(Earning Ability)+ b5(Liquidity Ratio)+ b6(GDP)+ b7(INF)+ e
Net Interest Margin= a1 + b1(Capital Adequacy)+ b2(Asset Quality)+ b3(Management Quality)+ b4(Earning Ability)+ b5(Liquidity Ratio)+ b6(GDP)+ b7(INF)+ e
Where: a is the intercept
b1 to b7 are coefficient parameters
GDP is the Gross Domestic Product (GDP) during the global financial crises period
INF is the Average Annual Inflation Rate global financial crises period
e is the Error term
Model adapted from Aspal and Nazneen (2014)
For the moderating effect of the bank type, the following extended MLR model was used:
Return on Assets= a1 + b1(Capital Adequacy*B)+ b2(Asset Quality*B)+ b3(Management Quality*B)+ b4(Bank Earnings*B)+ b5(Liquidity Ratio*B)+ b6(GDP*B)+ b7(INF*B)+ e
Return on Equity= a1 + b1(Capital Adequacy*B)+ b2(Asset Quality*B)+ b3(Management Quality*B)+ b4(Earning Ability*B)+ b5(Liquidity Ratio*B)+ b6(GDP*B)+ b7(INF*B)+ e
Net Interest Margin= a1 + b1(Capital Adequacy*B)+ b2(Asset Quality*B)+ b3(Management Quality*B)+ b4(Earning Ability*B)+ b5(Liquidity Ratio*B)+ b6(GDP*B)+ b7(INF)+ e
Where B is the bank type. Model adapted from Aspal and Nazneen (2014)
MLR model assumptions
The data variables used in this study must suit the proposed multiple linear regression models. Therefore, it was important to carry out diagnostic tests, such as normality and multicollinearity tests, for the basic assumptions. Also, heteroscedasticity test was necessary for testing assumptions about the variance of errors across observations in the regression analysis (Harvey, 1976; Grewal, Cote and Baumgartner, 2004; Ghasemi and Zahediasl, 2012).
These are mainly graphical methods that check the normal distribution of the model residuals. As the number of observations was high, the Kolmogorov-Smirnov test was performed to determine whether the samples originated from a population with a particular distribution. Besides, normal probability plots and histograms were generated to assess the normal distribution of data (Ghasemi and Zahediasl, 2012).
Multiple regression models have two or more predictors. When a model has two or more independent variables in a regression model, they are likely to be highly correlated, which causes multicollinearity. The CAMEL variables are often moderately or highly correlated, and the researcher has no control over the model predictors, which might increase the level of multicollinearity (Grewal, Cote and Baumgartner, 2004).
Heteroscedasticity may also occur as a result of the violation of other model assumptions. The Breusch-Pagan/Cook-Weisberg test was conducted to test heteroscedasticity in the variables. The test assesses whether the error variances are equal or the variance is as a result of a multiplier effect of one or more predictors (Harvey, 1976).
3.4 Summary of Variables
The Return on Assets (ROA) is calculated by dividing the banks net income by its total assets (Simpson and Kohers, 2002). Consequently, it is a measure of the ability of the management to use the available assets to generate profits. The Return on Equity is the Net Income to Total Equity ratio that measures the profitability of the shareholders capital (Barth, Beaver and Landsman, 1998). The Net Interest Margin is used to measure the interest income relative to the total bank assets. It is a measure that can be used to indicate the prudence of the banks loan investment. The CAMEL variables, in turn, influence these three variables. Merchant (2012) estimated Capital Adequacy using the equity to total assets ratio, which can be used to determine the bank’s capability to withstand shocks during times of financial risk. Asset quality, management quality, earnings quality, and liquidity have a direct influence on return on assets, return on equity, and the net interest margin.
Asset quality refers to the quality of the banks loan portfolio. The proper administration of the loan and other credit products is critical to the bank because it is the primary source of profits for the financial institutions. In this study, the asset quality was calculated by estimating the ratio of Loan Loss Reserves (LLR) to the Total Asset portfolio. LLR represents the estimated loss on loans that arises from defaults and nonpayment. It also evaluates the banks creditworthiness. In general, the ratio is an important indicator to the bank to devise ways of managing their loan portfolio to avoid incurring further losses through defaults and non-payments. During a period of financial crisis, banks may experience a higher LLR as economies suffer from the downturn. In such a scenario, it is likely that many borrowers will find a difficult time repaying the loan leading to defaulting or nonpayment. Therefore, banks are expected to have higher LLR during the global financial crises as compared to when the situation is normal. A lower LLR value is an indicator of a healthier financial situation of a bank than a higher value (Simpson and Kohers, 2002). Comparing the LLR values of the Islamic banks to those of conventional banks during the global financial crisis time revealed differential performances of the two types of banks.;
However, the Asset Quality could be a direct consequence of the bank management quality. The performance of banks depends on the type of management and the mechanisms put in place to manage their assets vis a vis their liabilities. The ability of the managers to acquire superior depositors and high-quality borrowers is a prerequisite to enhancing the asset quality and therefore the financial performance of the bank. In this study, this phenomenon was estimated through the calculation of the Loans to Deposit Ratio (LDR). This ratio is essential to highlight the proportion of the loans the bank issued through the total deposits. A high LDR is a positive score for the management. Therefore, the quality of the management has direct implications on the performance of the banks (Merchant, 2012).
Regarding the earnings quality, the importance of the operational aspects of the bank becomes clearer. The banks increase their profitability through increased production and highly controlled costs. The profitability arising from the minimization of costs and enhanced productivity is an indicator of the banks operational efficiency, and therefore the earnings quality. Earnings quality is estimated using the cost to income ratio, which is the cost incurred by the bank to generate more income (Simpson and Kohers, 2002). It can be hypothesized that the lower the cost to income ratio the more profitable the bank is. Therefore, more profitable banks should reveal a lower ratio. In the context of the present study, this ratio was applied to find out which set of banks had the least costs to income ratio between the Islamic banks and the conventional financial institutions.
Liquidity is another predictor of the financial performance of banks. All financial institutions are in the business of receiving money from depositors and lending the same to borrowers. Therefore, this variable is perhaps the most important to all banks. Liquidity tracks the movement of the bank’s finances, and the management can easily tell from the liquidity figures whether their balance sheet is healthy or they are headed towards bankruptcy. A critical assessment of the banks liquidity is the net loan to total assets ratio. The ratio is expected to be very low because banks are often mindful about their liquidity (Barth, Beaver and Landsman, 1998). Nevertheless, it was hypothesized that there were significant variances in the loan to total asset ratios between the Islamic banks and the traditional banking institutions.
In conclusion, the CAMEL framework was extremely useful in studying the various underlying factors that affect the financial performance of banks. Each aspect was analyzed in detail, and the synthesis of this information used to find statistical differences between the two chosen groups of banks.
Aspal, P.K. and Nazneen, A., 2014. An empirical analysis of capital adequacy in the Indian private sector banks. American Journal of Research Communication, 2(11), pp.28-42.
Barth, M.E., Beaver, W.H. and Landsman, W.R., 1998. Relative valuation roles of equity book value and net income as a function of financial health. Journal of Accounting and Economics, 25(1), pp.1-34.
Greene, W., 2001. Estimating econometric models with fixed effects. Department of Economics, Stern School of Business, New York University.
Grewal, R., Cote, J.A. and Baumgartner, H., 2004. Multicollinearity and measurement error in structural equation models: Implications for theory testing. Marketing Science, 23(4), pp.519-529.
Harvey, A.C., 1976. Estimating regression models with multiplicative heteroscedasticity. Econometrica: Journal of the Econometric Society, pp.461-465.
Hsiao, C., 2014. Analysis of panel data (No. 54). Cambridge university press.
Jaffar, M. and Manarvi, I., 2011. Performance comparison of Islamic and Conventional banks in Pakistan. Global Journal of Management and Business Research, 11(1).
Makkar, A. and Singh, S., 2013. Analysis of the financial performance of Indian commercial banks: A comparative study. Indian Journal of Finance, 7(5), pp.41-49.
Merchant, I.P. (2012). Empirical Study of Islamic Banks Versus Conventional Banks of GCC. Global Journal of Management and Business Research, 12 (20).
Munir, R., Perera, S. and Baird, K., 2011. An analytical framework to examine changes in performance measurement systems within the banking sector. Australasian Accounting, Business and Finance Journal, 5(1), pp.93-115.
Rozzani, N. and Rahman, R.A., 2013. Camels and performance evaluation of banks in Malaysia: conventional versus Islamic. Journal of Islamic Finance and Business Research, 2(1), pp.36-45.
Simpson, W.G. and Kohers, T., 2002. The link between corporate social and financial performance: Evidence from the banking industry. Journal of business ethics, 35(2), pp.97-109.
Spriensma, A., 2016. Descriptive Versus Inferential Statistics. Journal of Endourology, 30(1), pp.1-4.
Tahir, S., 2007. Islamic banking theory and practice: a survey and bibliography of the 1995-2005 literature. Journal of Economic Cooperation among Islamic Countries, 28(1), pp.1-72.
Weinstein, J.N., 2012. Drug discovery: Cell lines battle cancer. Nature, 483(7391), pp.544-545.
Yang, J., Ward, M.O. and Rundensteiner, E.A., 2003. Interactive hierarchical displays: a general framework for visualization and exploration of large multivariate data sets. Computers & Graphics, 27(2), pp.265-283.
Get access to
Guarantee No Hidden