Osama El-Temtamy b Assistant Professor, Accounting Faculty of Business, University of New Brunswick Saint John, E2L4L5, Canada

Transcription

1 Oil Price Plunge: Are Conventional and Islamic Banks Equally Vulnerable? Ghulame Rubbaniy a (Corresponding Author) Assistant Professor, Finance College of Business, Zayed University Abu Dhabi, UAE ghulame.rubbaniy@zu.ac.ae Osama El-Temtamy b Assistant Professor, Accounting Faculty of Business, University of New Brunswick Saint John, E2L4L5, Canada OTemtamy@unb.ca W. A. Khan c Professor of Finance Schroeder Family School of Business Administration, University of Evansville, 1800 Lincoln Avenue Evansville, IN 47722, USA wk3@evansville.edu Abstract In response to the recent debate on the potential vulnerability of the banking industry to oil price plunge, we investigate the effect of oil price plunge on credit and insolvency risks of the banking industry in the Gulf Cooperation Council (GCC) over a period from 2000 to 2014 with a particular focus on Islamic and conventional banks due to their different risk characteristics. Our findings show that falling oil prices significantly increase the credit risk for the overall banking industry in the GCC region and specifically for banks operating in Kuwait and UAE. However, the results become insignificant and remain invariable across different bank categories, (conventional, Islamic, and Islamic window only), even using different proxies. Our insolvency risk analysis does not provide any convincing evidence to support the hypothesis that oil price plunge increases the insolvency risk of the banking industry in the GCC region. Our findings based on recent time period offer some insights to the various stake holders in banking industry as well as regulators in the gulf region in terms of credit risk vulnerability and absence of insolvency risk. Keywords: Islamic Banks, Conventional Banks, Credit Risk, Insolvency Risk, Oil Price Plunge, Loan Loss Reserves, Withdrawal Risk JEL Codes: G21, G32, G33 We acknowledge and thankful to Zayed University, Abu Dhabi, UAE for the financial support to complete this project. a College of Business, Zayed University, Abu Dhabi, UAE b Faculty of Business, University of New Brunswick, Saint John, Canada. c Schroeder Family School of Business Administration, University of Evansville, Evansville, USA. Correspondence address: College of Business, Zayed University, PO box , Khalifa City, Abu Dhabi, UAE Phone: ; Ghulame.rubbaniy@zu.ac.ae 1

2 1. Introduction: In the backdrop of recent oil price plunge of 2015, analysts and policy makers are examining the consequences of the falling oil prices as well as possible solutions to cope with the problems faced by oil exporting economies such as Gulf Cooperation Council (GCC), Russia and Iran. More than 50% decline in oil price from its peak has not only affected the oil companies but also the banks operating in oil dependent economies such as Gulf Corporation Council (GCC). For instance, the credit and insolvency risks may have increased for the banks exposed to the sectors that are dependent on revenues from energy sector directly or indirectly in GCC countries. Banking sector plays a critical role in any globally integrated economy. Adverse conditions in this sector can quickly generate systemic pressures in other sectors of economy generating crisis as witnessed in the 2007/2008 global financial crisis. However, existing studies mostly investigate and report a significant influence of oil prices on stock market performance in both developing and developed countries (Alquist & Guénette, 2014; Basher & Sadorsky, 2006; Chen, 2014). The impact of declining oil prices on banking sector has largely been ignored in existing literature. We attempt to fill this gap by investigating the impact of oil price plunge on banks risks, particularly the credit and solvency risks, in the GCC countries. The oil fuels global economic growth and oil prices changes have a significant impact on world economies. Mussa (2000) for instance estimates that an increase of US$ 5/barrel in oil price reduces world economic growth by 0.3% in the following year. The recent oil price plunge is a matter of concern for GCC countries as it impacts the real and financial sectors. For example, in October 2014, IMF chief Christine Lagarde reported that a decline of $25/ barrel in oil prices would reduce GDP of GCC economies by 7%. Also continuing lower prices would increase the fiscal deficit of Gulf States. Gulf News (December 29, 2015) 1 has reported that prolonged lower oil prices would reduce government revenues and put high pressure on private sector institutions to finance investment projects. Significant decline in oil prices have an impact on credit portfolios of banks operating in the GCC region. These banks' exposures are concentrated in sectors that directly or indirectly depend on government spending which in turn is driven by the demand and revenue from the oil sector. Hence, the performance of banking sector in GCC region is tied to the oil prices 1 2

3 which is likely to impact the credit and solvency risks of banks. In other words, falling oil prices can increase the bank borrowers difficulty to pay back their loans and interests to the banks. And a persistence downward trend in oil prices will likely produce defaults and high ratio of non-performing loans for banks in the Gulf region. Taking these considerations into account we intend to investigate the influence of falling oil prices on credit and insolvency risks of the banks in GCC region and whether this effect varies across Islamic and conventional banks as macro shocks may produce different results for Islamic versus conventional banks due to inherent differences in both sets of banks (as discussed below). Figure 1 shows that oil prices remained at lower levels until the year 1999 but increased to (US$ 138/barrel) in June Such an enormous price hike resulted in negative consequences for different economies and stock markets (Arouri & Rault, 2012; Fayyad & Daly, 2011; Hamilton, 2009; Jones & Kaul, 1996; Papapetrou, 2001). However, oil prices plummet to US$37 in December This decline might have brought good news for many oil importing countries but likely to have adverse impact on oil dependent countries particularly the GCC region whose main export is oil. A persistent oil price decrease can badly hurt oil-exporting economies resulting in lower revenues, falling foreign reserves, increasing budget deficits, unemployment surge and decreasing economic growth. Since oil exporting economies depend on performance of the domestic oil companies, falling oil prices reduce firms revenue and thus put them in a need to renegotiate the financing terms and conditions with their banks to continue their operations. In addition, their decreasing assets values and increasing default risks may prompt the underwriting banks to ask for additional collateral or increase the cost of financing in order to protect their financing. A tightening of financial conditions of these companies can be observed in Figure 1, which plots the crude oil prices against oil and gas sector stock indices of World, GCC, emerging markets, United States and China. Overall, a strong positive correlation between oil prices and oil and gas sector indices can be observed in Figure 1, where falling (increasing) oil prices are accompanied by plummeting (surging) oil and gas sector indices of World, GCC, emerging markets, United States and China. Insert Figure 1 about here 3

4 This historically strong correlation between oil prices and stock indices of oil and gas sector at international level may prompt a possible threat for financial institutions having high exposures in the energy sector (for instance worldwide outstanding loans of banks towards energy sector were around US$3 trillion in the year 2014), and the higher concentration of credit portfolios to the energy sector is likely to increase the credit portfolios risk of the banks in the GCC region as well as their net incomes are driven by developments in the energy market (Husain et al., 2015). Existing literature is scarce in documenting the relationship between oil prices and the risks faced by the banks particularly in a setting where banks loan portfolios are concentrated to the energy sector. In addition, almost nothing has been documented in the existing literature comparing how the oil prices plunge and banks risks vary across Islamic and conventional banks in GCC region. Most Islamic banks operate in the Islamic countries and face a unique set of risks. For instance, they neither have control over management of the projects they finance nor can they demand collateral from customers in case of profit and loss sharing (PLS) Mudarba accounts (Errico & Farahbaksh, 1998). Khan and Ahmed (2001) explain the withdrawal risk 2 associated with PLS investment account holders (IAH) in both conventional and Islamic banks having unique risk characteristics. This withdrawal risk pressurizes the Islamic banks to pay competitive premium returns to investment account holders irrespective of banks actual performance which in turn forces the shareholders to raise more equity capital (Obaidullah, 2005). Osoba (2003) suggest that although withdrawal risk disciplines the Islamic banks; however, some depositors show loyalty to these banks because of unavailability of other options or their preference for Islamic banks on religious grounds. Farooq and Zaheer (2015) show that Islamic banks can perform relatively better during panics as they are less prone to deposit withdrawals and less likely to cut lending during down market times. This loyalty factor helps mitigating the adverse impact of withdrawal risk and reduces the probability of bank defaults. However, complexity of Islamic loans, moral hazard 3 problems associated with PLS contracts, lack of collateral and limited penalties on defaults makes these banks vulnerable to credit risks (Elnahass, Izzeldin, & Abdelsalam, 2014; Gropp & Vesala, 2004; Hamza & Saadaoui, 2013). While these banks have long-term relationship with depositors making them 2 It is the risk of withdrawal of deposits from Islamic banks and then investing in conventional banks for higher returns. 3 Moral hazard problem is associated with manager s unnecessary risk taking at the cost of depositors, as the investors are ultimate risk bearers. 4

5 capable of bearing higher losses, they also face operational limitations (prohibition of interest, speculation or Gharar and gambling) simultaneously, which reduce their credit portfolio optimization compared to the conventional banks. Thus, they are more exposed to insolvency risks compared to conventional banks, as their credit is concentrated in specific sector and not well diversified (Abedifar, Molyneux, & Tarazi, 2013). Given that Islamic banks face different risks compared to the conventional banks, the impact of oil price plunge on banks riskiness across these two groups is likely to vary as well. Given that GCC region has a significant share of the oil market and possess a unique feature of concentrated banks loan portfolios to the energy sector, falling oil prices can destabilize the financial system. A persistent decline in oil prices may increase the credit and insolvency risks for the banks operating in GCC. Thus, insights about the relation between oil prices decline and banks vulnerability to these risks become more important in GCC region. Surprisingly however, existing academic studies have overlooked this area although there are some traces of the effect in policy documents. Financial stability report (2014) published by Central Bank of Oman 4, for example, documents that in their macro-financial stress testing analysis fall in oil price leads to a decline in real GDP, which in turn adversely affects default rates and capital adequacy ratio of the banking sector. Assuming that risk characteristic may vary across conventional and Islamic banks operating in GCC countries, we attempt to formulate the following research questions to fill the above gap: (1) Does the oil prices plunge increase both the credit and insolvency risk of the banks operating in the GCC countries? (2) Does this vulnerability due to the increase in credit and insolvency risks vary cross-sectionally across conventional and Islamic banks? We contribute to the banking and energy finance literature in several ways. We are the first to examine the effect of falling oil prices on the credit and insolvency risk of banks in the GCC region at both aggregate and country levels over the period from 2000 to 2014 the period which carries some remarkable oil price fluctuations. Second, existing financial market literature (Arouri & Rault, 2012; Effiong, 2014; Mussa, 2000) used crude oil prices, oil price shocks, oil price risk and higher oil prices as proxies of oil price shock. Contrary to the mainstream literature we use falling oil prices (negative sock) as a proxy to the oil price shock proposed by Hamilton (1983) in the study of crude oil and economic performance. Third, we also control for banks specialization as Islamic and conventional banks have different business 4 5

6 models and this effect needs to be controlled for robustness of results, as we believe riskiness level is not same across conventional and Islamic banks. Our findings show that falling oil prices significantly increase the credit risk of the banks at aggregate level in the Gulf region; and, the results are robust to different proxies of the credit risk. Contrary to our initial expectations, we do not find any empirical support for variation in the effects of oil prices decrease across different bank specialisations (for instance conventional banks and Islamic banks). Our country level analysis documents that the credit risk for Kuwait and UAE banking industry increases as oil prices plunge. Our insolvency risk analysis does not provide any convincing evidence to support the hypothesis that oil price plunge increases the insolvency risk of the banking industry in the Gulf region. These findings might also point to the possibility of very well diversified investments of banking industry in the Gulf region. The rest of this paper follows as: Section 2 explains the research design and methodology. Section 3 presents the empirical results and finally Section 4 concludes. 6

7 2. Research Design and Methodology: 2.1. The Data: The envisaged annual data for this study comes from three sources, i.e., Bankscope, Datastream and World Bank database. We extracted bank specific data from Bankscope, historical crude oil (WTI) prices in USD/barrel from DataStream and macro-economic variables from World Bank database. Our sample includes almost all Islamic banks, conventional banks, and commercial banks with Islamic windows operating in Gulf countries. The choice of the Gulf countries is motivated by apparent dependence of these countries on oil revenues, and thus any oil price shock is likely to affect the financial sector and economic activities within these countries. To capture the changes in economic and geo political conditions, we used a time span of , which is also marked with significant regional growth of Islamic and conventional financial institutions and high oil prices volatility. To cope with the currency risk, we converted all of our local currencies data into US dollars at the exchange rate of December 31, 2014, the last day of our sampled data. In order to make our results comparable with existing literature, we dropped insurance companies, investment and trust corporations, central banks, discontinued and merged banks, investment banks and duplicated banks from the data extracted from BankScope database. We also checked and corrected the specialization of banks by accessing the websites of relevant banks, as specialization of most of banks is not correct in the BankScope database. Our final sample includes data of 797 banks operating in GCC region divided into 320 commercial banks, 184 Islamic banks and 293 conventional banks with Islamic windows. Our total banks distribution includes data of 117 banks from Bahrain, 76 from Kuwait, 65 from Oman, 133 of Qatar, 157 banks operating in Saudi Arabia and 269 banks of United Arab Emirates. Table 1 (Panel A) shows the descriptive statistics of the variables used in this study. The average of our main explanatory variable, oil price negative shock (oil price decrease OPD), is -0.03% with a standard deviation of The average value of OPD clearly shows the dominant prevalence of negative values in OPD s computed observations, which is supported by a negative skewness value of -3.14; and a lower value of standard deviation indicates limited spread of the OPD around the mean value. Following Abedifar et al. (2013) credit risk is proxied by liquid liability ratio that has an average value of 5.85, a standard deviation of 4.74 and a positive skewness with a kurtosis value of For robustness purposes, we also use non-performing loans ratio of banks as proxy of credit risk suggested by Abedifar et al. (2013), 7

8 which has an average value of 5.85, approximately similar to liquid liability ratio, but a higher standard deviation of 6.79 indicating more variation in this proxy of credit risk. We also use insolvency risk as an independent variable in a separate regression that shows a mean value % and median as 63.46% that suggests the presence of higher insolvency risk of banks in the Gulf region. The distribution is positively skewed with kurtosis of and standard deviation of pointing towards the differences in default risk probability on different bank specialization as some banks can be badly affected by falling oil prices compared to the others. We also control for bank specific characteristics such as bank size and market share, loans and total assets growth, cost inefficiency, capital asset ratio and loan earning asset ratio. Bank size has a mean value of 8.72% and a normal distribution with kurtosis of A low value of bank size s standard deviation (1.45) excludes the possibility of large fluctuations in bank size in our sample. Market share is positively skewed with a high kurtosis value of and capital asset ratio is on average 17.07% while, loan earning asset ratio is lower 0.63% with low standard deviation. Loans and total asset growth are on average 22.73% and 20.58% respectively. The mean value of cost inefficiency ratio of our sampled banks is 41.10% with a standard deviation of We also control for country specific factors such as GDP per capita that is on average 10.13%, positively skewed with little variation as the value of standard deviation (0.52) is low. Finally, GDP per capita growth with an average value of -0.50% shows negative effect of macro-economic on GDP growth of Gulf countries. We also divide our data into subsamples across the banks specialization - conventional, Islamic, and Islamic window. Table 1 (panel B) displays the descriptive statistics. For Islamic banks, liquid liability ratio (as proxy of credit risk) is 4.11% with a standard deviation of 4.44% and a positive skewness of 2.63% as compared to commercial banks which yields a higher liquid liability ratio of 5.57% and a standard deviation of 5.27%. We get almost similar values for commercial banks with Islamic windows. These figures suggest that due to unique characteristics of Islamic banks (e.g. interest free operations) their default loan ratio is less, as compared to conventional banks, because Islamic banks offer deposits on profit and loss basis and do not charge interest on loans. We also use non-performing loans ratio as proxy of credit risk. For Islamic banks, nonperforming loan ratio average is 4.80% which is less than both the commercial banks, with average of 6.68%, and commercial banks with Islamic windows average of 5.58%. These 8

9 statistics supports the view that non-performing loans are higher in commercial banks as these banks advance loans without transferring the losses to borrowers. For robustness purpose, we also estimate insolvency risk using z-score as proxy. For Islamic banks, the mean value is % and for commercial banks the mean value is %. This suggests the presence of higher insolvency risk of commercial banks pointing towards the differences in default risk probability for different specialization of banks. Insert Table 1 about here The pairwise correlation matrix with significance level is presented in Table 2 which does not indicate any major issue of multicollinearity among independent variables. Most variables exhibit small correlation values. Moreover, the mean variance inflation factor (VIF) is also less than 10, which rejects the existence of multicollinearity in independent variables Research Methodology: Insert Table 2 about here We begin our analysis by using panel regressions, i.e., fixed effects and random effects regression models, to investigate the effect of oil prices plunge on credit and insolvency risks. Although we used both fixed and random effects regressions, most of the results are estimated using fixed effects regression permitted by Haussmann Test. For instance, Haussmann Test suggests random effects model for our insolvency risk analysis but we also apply fixed effects method for the consistency of the estimations Hypothesis: Following literature, we use loan loss reserves to gross loan ratio as a proxy of the credit risk. As a robustness check, we also use impaired loans to gross loans ratio as indicator of the credit risk. The tight situation of oil prices in energy market reduces the savings of the governments and, different oil and gas exploration companies as these companies are unable to meet the costs of extraction of oil at lower prices. These negative consequences of lower oil prices transfer to the financial sector and increase the credit risk of the lending banks. Any reduction in oil prices 9

10 is likely to have an adverse impact on earnings of the banks through interest income, reduced savings and increased loan losses (Husain et al., 2015). Kinda, Mlachila, and Ouedraogo (2016) support the view that commodity price negative shocks and their magnitudes have adverse impact on the financial system. These negative shocks increase the non- performing loans, reduce profits of banks, and may cause potential banking crisis. The negative effects are particularly prominent in countries without proper governance mechanism and those lacking diversified export base. For instance, a prolonged fall in oil prices may increase the nonperforming loans of individual banks in GCC because of their exposures concentrated towards oil and gas sector. In addition, since information asymmetry and moral hazards problems have more likelihood to occur in Islamic banks compared to conventional banks, we expect a negative impact of oil prices plunge on credit risk of banks operating in the GCC region. Therefore, we expect Islamic banks to be more prone to credit risk compared to the conventional banks. A persistent fall in oil prices is likely to affect the revenue of both the oil companies and the governments operating in the GCC, which in turn is likely to decrease the loan recovery and increase the loan losses of the banks. The persistent decline in oil prices changed the outlook for hydrocarbon exporters. These low oil prices are a cause of an increase in interbank rates in GCC countries because of lower liquidity in banking system. Further, credit rating agencies lowered the ratings of oil exporting countries (e.g. Bahrain, Kazakhstan, Oman, and Saudi Arabia). Banks in oil exporting countries mainly finance the government deficits. For example, in 2015, Saudi Arabia withdrew $106.7 from Saudi Arabian Monetary Agency. On the other hand, commercial banks mainly finance the fiscal deficit in United Arab Emirates. This adverse outlook may increase the default risk of banks (Sommer & Sommer, 2016). Therefore, we expect a negative effect of oil price plunge on insolvency risk of the banks in the GCC; and, we expect this effect to vary across different specializations and countries Empirical Model: We use panel regressions with credit risk as dependent variable proxied by loan loss reserves to gross loan ratio. For robustness purposes, we also use impaired loans to gross loans as a proxy for the credit risk. To make our analysis comparable with the existing literature, we also use a few control variables. For instance, we control for Market Share, Bank Size, Capital Asset Ratio, Loan Earning Asset Ratio, Loan Growth, Asset Growth, Cost Inefficiency, GDP and GDP per Capita. 10

11 Market Sharei,t is the natural log of market value of the assets of bank i to total assets of banking system in time t used as indicator to control for the market power of banks. The high market share may increase the riskiness of banks as with vast branch network big banks can transfer the risks to a mass level (Kick & Prieto, 2013). Bank Sizei,t is natural log of total assets of the bank i in time t. The major purpose of using this control variable is that large conventional and Islamic banks are well diversified and reduce their operating expenses due to their large size and many banks also merge to exploit economies of scale (Hughes, Mester, & Moon, 2001). Contrary to this, some authors believe that large banks are riskier because of too big to fail assumption as small banks with less liabilities are more efficient (Demirgüç-Kunt & Huizinga, 2013; Kane, 2010). Capital Asset Ratioi,t shows the equity capital to total assets ratio of bank i in time t, which is used to control for moral hazard problem and monitoring incentives of banks. Inclusion of the variable also allows us to account for variation of risk and equity capital as banks with more equity tend to bear more risk. Loan Earning Asset Ratioi,t is the share of net loans to earning assets of bank i in time t, which is used to control for stability of Islamic banks as they face more investment limitations compared to conventional banks in the system. Loan growthi,t displays the annual growth of loans of bank i in time t, and is used in credit risk equation to control for financial instability. It controls for the observation that when banks have private information of borrowers, asymmetries are reduced and result in relaxed screening, low profitability and reduction in interest rates on loans for instance during subprime mortgage crisis (Dell'Ariccia & Marquez, 2006). Asset growthi,t is annual growth of assets of bank i in time t. It is used to control for the growth strategy of banks as Islamic banks are growth oriented and like the customers to transfer their deposits from conventional to Islamic banks. The use of the control variable in the Equation 4 is motivated by the notion that it is more likely to affect insolvency risk. Cost Inefficiencyi,t displays the cost to income ratio of bank i in time t, and controls for other risks of the banks, i.e., moral hazard problem as badly managed banks benefit more from taking risks (Beck, Demirgüç-Kunt, & Merrouche, 2013; Bourkhis & Nabi, 2013). Finally we control for country level variables such as LNGDP Per Capitai,t that shows the natural log of per capita GDP in US$ of country i in time t, GDP Per Capita Growthi,t is 11

12 annual growth rate of GDP Per Capita of country i in time t. These variables are used to control for overall economic development and growth of population. Our panel regression models are extension of Abedifar et al. (2013) by incorporating oil price decrease as a main explanatory variable in this study. The oil price decrease proxy is proposed by Hamilton (1983) and computed as: log(op t ) if log(op t ) < 0, 0 otherwise (1) We code negative changes in oil prices as 1 for proxy of negative oil price shock because we want to check the effect of oil price decrease on riskiness of banks. Equation 2 has been developed to investigate the effect of decreasing oil prices on credit risk of the banks in GCC region. The model follows as: Credit Risk i,t = β 0 + β1 Oil Price decrease t + β2 Market Share i,t + β3capital Asset Ratio i,t + β4loan Earning Asset Ratio i,t + β5loan Growth i,t + β6lngdp Per Capita i,t + β7gdp Per Capita Growth i,t + α i + μ i,t (2) Here Credit Riski,t is proxied by either ratio of loan loss reserves to gross loans or impaired loans to gross loans of bank i in time t, Oil Price decreaset ( log(op t ) if log(op t ) < 0, 0 otherwise 1) shows the negative oil price changes at time t, Market Sharei,t is the natural log of market value of assets of bank i to total assets in the banking system in time t, Capital Asset Ratioi,t shows the equity capital to assets ratio of bank i in time t, Loan Earning Asset Ratioi,t is the share of net loans in earning assets of bank i in time t, Loan Growthi,t displays the annual growth of loans of bank i in time t, LNGDP Per Capitai,t shows the natural log of per capita GDP in US$ of country i in time t, GDP Per Capita Growthi,t is annual growth rate of GDP per capita of country i in time t, αi is fixed effect, µi,t is the error term, β0 is the intercept, and βi s represent the loadings of independent variables on credit risk. To investigate the effect of oil price decrease on insolvency risk of the individual banks, we use Z-score to measure the insolvency risk and capturing the stability of banks. Here Z-Score is calculated using accounting information of banks in the following equation: ZScore = E(ROA)+CAR SD(ROA) (3) 12

13 Here E(ROA) is the expected return on assets, CAR is the capital to assets ratio and SD(ROA) is termed as standard deviation of return on assets. Higher values of Z-Score depict that bank is more stable and has lower chances of its default and lower values of Z-score means higher chances of bank s default. Our insolvency risk panel regression is defined as: Insolvency Risk = β 0 + β1 Oil Price Decrease i.t + β2 Market Share i,t + β3bank Size i,t + β4loan Earning Asset Ratio i,t + β5asset Growth i,t + β6cost Inefficiency i,t + β7lngdp Per Capita i,t + β8gdp Per Capita Growth i,t + α i + μ i,t (4) Here Insolvency Riski,t is proxied by Z-score (computed above) of bank i in time t, Oil Price decreaset ( log(op t ) if log(op t ) < 0, 0 otherwise) shows the negative oil price changes at time t, Market Sharei,t is the natural log of market value of assets of bank i to total assets in the banking system in time t, Bank Sizei,t is natural log of total assets that shows the size of the bank i in time t, Loan Earning Asset Ratioi,t is the share of net loans in earning assets of bank i in time t, Asset growthi,t displays the annual growth of assets of bank i in time t, Cost Inefficiencyi,t display the cost to income ratio of bank i in time t, LNGDP Per Capitai,t shows the natural log of per capita GDP in US$ of country i in time t, GDP Per Capita Growthi,t is annual growth rate of GDP per capita of country i in time t, αi is individual effect, µi,t is the error term, β0 is the intercept, and βi s represent the loadings of independent variables on credit risk. 3. Results and Discussion: Table 3 displays the aggregate results of oil price decrease effect on credit risk of the banks using loan loss reserve ratio as proxy of credit risk. These aggregate results of different bank classes have been presented in Columns 2, 3, 4 and 5 respectively. These estimates are developed using fixed effects regression permitted by Hausman Test. In general, our results document a negative effect of oil price decrease on credit risk (liquid liability ratio) in all estimations; however, the effect is statistically significant with coefficient of only at aggregate level. A 1 percentage point decrease in oil price increases the credit risk by 1.9%. On individual level results are mostly insignificant, however, there economic significance changes across bank specializations. For instance, the insignificant coefficient of decreasing oil prices in case of Islamic banks is , which is much higher 13

14 compared to commercial banks , and conventional banks with Islamic windows (-0.296). On aggregate level the results support our hypothesis that oil prices plunge increases the credit risk of banks; however, our hypothesis does not find statistical support at the banks specialisation level. Our control variable for instance market share is significant, both economically (-71.77) and statistically, for Islamic banks at 10% significance level, which is consistent with (Kick & Prieto, 2014). Capital asset ratio has a negative and highly significant (-0.174) impact on credit risk at aggregate level; however, at banks specialisation level the coefficient (-0.302) is significant only for commercial banks, which is consistent with (Bourkhis & Nabi, 2013; Foos, Norden, & Weber, 2010). The effect of loan growth on credit risk is also highly significant at aggregate as well as each bank category. It means that higher loan growth positively contributes to credit risk of banks. Our country specific control variables, for instance, GDP per capita (LNGDPC) shows negative and significant effect on credit risk with coefficient of for Islamic banks. GDP per capita growth (GDPCG) has a statistically positive and significant impact but economically less meaningful with a coefficient of The most striking feature of the GDP per capita and GDP per capita growth variables is that their effect is significant only for either Islamic banks or banks with Islamic windows. This finding complements the results of existing literature (Foos et al., 2010; Mirzaei & Mirzaei, 2011). This visible difference can be explained by relatively higher market share of Islamic banks in GDP growth in the region. Insert Table 3 about here Table 4 shows the results of effect of oil price decrease on credit risk using non-performing loans to gross loans ratio as proxy of credit risk. Results show that oil price decrease has a negative influence on the credit risk at aggregate level; however, this effect is not statistically supported at the individual specialization levels. The statistically significant value of the coefficient of oil price decrease is , which can be interpreted as a 1 percentage point decrease in oil price increases the credit risk of banking sector by 3.087%. Using nonperforming loans to gross loans ratio most of the results do not qualitatively differ from the 14

15 results posted in Table 3, suggesting that our results are robust to different proxies of credit risk. Insert Table 4 about here Although countries in the Gulf region might significantly depend on oil revenue, the banks in this region might be diversifying their assets to minimize their portfolio risk due to heavy investment in the oil sector. In addition, risk management practices might also vary across Gulf countries due to development of the financial and legal system in these countries. This motivates us to conduct a country wide analysis of all the banks to investigate whether effect of falling prices varies significantly across banks in GCC countries. For this purpose, we run the fixed effects panel regressions and report the results of our six Gulf countries in Columns 2-7 of Table 5. Results show a significantly negative impact of oil price decrease on credit risk of the banks in Kuwait and UAE with coefficients of and respectively; however, the effect is negative but insignificant in case of banks in Bahrain, Qatar and Saudi Arabia with an exception of Oman, which has a positive and significant coefficient. Insert Table 5 about here We re-run the country wide analysis using non- performing loans ratio as a proxy of credit risk. Results reported in Table 6 are qualitatively consistent with results of Table 5, i.e., oil price decrease effect is still significantly negative in case of Kuwait and UAE; however, this effect is also statistically significant in case of Saudi Arabia. In general, using two proxies, our results show that aggregately oil price plunge increases the credit risk of the banks operating in the Gulf region; however, some countries for instance Kuwait and UAE are more prone to this risk compared to other countries. The different results for Kuwait and UAE may be due to ignoring the risk of sustained lower oil prices by these countries compared to for instance Saudi Arabia which cut their revenue estimations in an anticipation of oil price fall to overcome this risk. 5 The increased public debt in UAE might also be a reason of significantly increased riskiness

16 Insert Table 6 about here Table 7 reports the results of the impact of oil price decrease on the insolvency risk using random effects panel regressions. The choice of the Random effect regressions is granted by the Hausman test. Using Z-score as proxy of insolvency risk we regress insolvency risk on oil price decrease variable after controlling for all the other variables at an aggregate as well as individual bank specializations levels. Our results in Table 7 show that oil price decrease has a negative but statistically insignificant effect on the insolvency risk at aggregate as well as at the level of each specialization except the conventional banks with Islamic windows where the effect is statistically significant at 10% level. This finding suggests that banking system in Gulf region is not significantly prone to insolvency risk due to falling oil prices. However conventional banks with Islamic windows are exposed to insolvency risk to some extent. Thus contrary to our initial expectations banking system in the Gulf region is not significantly prone to the solvency risk arising from the oil prices plunge. We also added two new control variables namely firm size and asset growth in regression (Equation 4) of insolvency risk as larger banks are risky and more prone to default due to too big to fail assumption. The asset growth variable is used to control for the growth strategy in insolvency risk Equation 4 (Foos et al., 2010). Our results show that firm size has an insignificant effect on insolvency risk at aggregate as well as each specialization level except for the commercial banks, where the effect is significant at 5% level of significance. This might be due to large size of commercial banks compared to Islamic banks and conventional banks with Islamic windows which are of small size and have less likelihood of default due to size. We find the effect of asset growth significantly negative at aggregate as well as for Islamic banks; however, it is insignificant in other cases. Overall the insignificant effect of oil price decrease on solvency risk of the banks in the Gulf region can point to the banks ability to diversify their investments across different industry to survive in case of persistent oil prices decrease in the region. Insert Table 7 about here Table 8 shows the fixed effects panel regressions results of the influence of oil price decrease on the insolvency risk. We used both the fixed effects and the random effects regressions for 16

17 comparison purposes. Using Z-score as proxy of insolvency risk, we regress insolvency risk on oil price decrease variable after controlling for all the other variables. The reported results in Table 8 show that oil price decrease has a negative but statistically insignificant effect on the insolvency risk at aggregate as well as at the level of each specialization except the Islamic banks with commercial where the effect is positive and insignificant. Therefore, we find that although banks in GCC countries are facing strict regulations and some liquidity problems especially in UAE and Saudi Arabia where governments finance their fiscal deficits by debt financing from central and conventional banks, they are not significantly prone to insolvency risk. Insert Table 8 about here Since banks operating in different countries of the same region might face a different level of solvency risk and thus any decrease in oil prices might have different impact on the insolvency risk across GCC countries, we therefore investigate the effect of oil prices decrease on the insolvency risk of the banking sector across different Gulf countries. The results of our fixed effects panel regressions for different countries are presented in Table 9. Our country level analysis show that effect of oil prices decrease on insolvency risk is insignificant across all countries suggesting that banking sector in any of these countries is not significantly exposed to insolvency risk. Insert Table 9 about here For consistency purposes, we also run the random effects panel regressions and the results for different countries are presented in Table 10. Our country level analysis show that effect of oil prices decrease on insolvency risk is negative and insignificant in different countries except Kuwait where the effect is positive and significant at 10% level, which is contrary to our hypothesis as we expected a negative influence of the oil price plunge on insolvency risk of banks. Overall we do not find convincing evidence that oil prices decrease increases the insolvency risk of the aggregate banking sector or individual banks across GCC countries. The reason may be the shorter span of time of falling oil prices and a persistent decrease in oil prices might increase the default risk of banks. 17

18 We find that oil prices decrease significantly increase the credit risk of the banking sector in Gulf region at aggregate level; however, across countries only UAE and Kuwait are significantly prone to credit risk in case of oil prices plunge. This could be due to weak response from policy makers towards the implementation of tariff subsidies and budget cut policies that may not allow the financial institutions to adjust their portfolios in time to manage their credit risk. The results are particularly surprising for UAE, which has a diversified investment portfolio towards tourism, construction, and international investments. The surprizing result of UAE might be due to the reason that commercial banks mainly finance the fiscal deficits in United Arab Emirates (Sommer & Sommer, 2016). Insert Table 10 about here 18

19 4. Conclusion: Quite recently, the returns on the investments in the oil and energy sectors have been decreasing in Gulf region due to falling oil prices which in turn is adding to risks for the banks as they potential have higher exposures to the petroleum sector. We investigate the effect of oil prices plunge on the credit and insolvency risks in the Gulf banking industry with a particular focus on Islamic and conventional banks due to their different risk characteristics. Our findings show that falling oil prices significantly increase the credit risk of the banks at aggregate level in the Gulf region. Our results are robust to two proxies of the credit risk. However, we do not find any empirical support for the variation in the negative impact of oil prices decline on credit risk across different bank categories, for instance, conventional banks, Islamic banks and conventional banks with Islamic windows. At country level, the credit risk for Kuwait and UAE banking industry increases as oil prices fall. In these countries banks mainly finance the fiscal deficit that arises due to lower oil prices. These countries don t have excess oil reserves for generating revenues through increasing the oil supply in the market. Therefore, governments of these countries get loans from commercial banks to finance their deficit that increase the non-performing loans of the commercial banks (Sommer & Sommer, 2016). Results of our insolvency risk analysis do not provide any convincing evidence to support the hypothesis that oil price plunge increases the insolvency risk of the banking industry in the Gulf region. These findings might also point to the possibility of well diversified investments of banking industry in the Gulf region. Our results offer some insights to banks in gulf region with respect to the credit and solvency risks. Our findings offer added insights to Kuwait and UAE which are vulnerable to decline in oil prices. Also, absence of insolvency risk will be a welcoming finding for the various stake holders especially the regulators who are in constant search for finding more and more evidence about the sound functioning of the banking industry in particular after the global financial crisis of 2007/2008 and the observed volatility in the markets since then. 19

20 Appendix: Figure 1: Log Crude oil prices and DataStream oil and gas sector indices from

21 Table 1: Descriptive Statistics Panel A: Descriptive Statistics of All Banks Variable N Mean Median Skewness Kurtosis Std. Dev. Oil Price Decrease Liquid Liability Ratio Non-Performing Insolvency Risk Bank Size Market Share Capital Asset Ratio Loan Earning Assets Loan Growth Total Assets Growth Cost Inefficiency Ratio LN GDP Per Capita GDP Per Capita Panel B: Descriptive Statistics of Banks Regarding Bank Specialization Islamic Banks Commercial Banks Commercial Banks with Islamic Windows Variables Obs. Mean SD Skewness Kurtosis Obs. Mean SD Skewness Kurtosis Obs. Mean SD Skewness Kurtosis Oil Price Decrease Liquid Liability Ratio Non-Performing Loans Ratio Insolvency Risk Bank Size Market Share Capital Asset Ratio Loan Earning Assets Ratio Loan Growth Total Assets Growth Non-Interest Income Cost Inefficiency Ratio LN GDP Per Capita GDP Per Capita Growth Source: Self-construction from Bankscope and Datastream databases 21

22 Table 2: Pairwise Correlation Matrix Variables OPD FSIZE MKSHR CAR LEAR LGROW AGROW CINEF LNGDPC GDPCG OPD FSIZE * MKSHR * CAR * * * LEAR * * * * LGROW * * * AGROW * * * * * CINEF * * * * * * LNGDPC * * * * * * GDPCG * * * * * Pairwise correlation matrix in which * indicates significance at 5% level. In addition, our mean variance inflation factor (VIF) is less than 10 that also rejects the multicollinearity in independent variables. Source: Self-construction from Bankscope and Datastream databases 22

23 Table 3: Fixed effects Results - Oil Prices and Credit Risk of Banks (LLR) LLR All Banks Islamic Banks Commercial Banks Islamic Windows Banks OPD ** (0.740) (2.404) (1.455) (0.792) MKSHR * (34.47) (38.72) (66.62) (24.35) CAR *** * *** (0.0466) (0.0678) (0.103) (0.0875) LEAR *** ** ** ** (3.841) (4.824) (8.264) (5.427) LGROW *** ** *** *** ( ) ( ) (0.0108) (0.0104) LNGDPC * (1.794) (3.259) (3.358) (2.173) GDPCG ** 0.145** ** (0.0337) (0.0563) (0.0633) (0.0382) Constant ** (18.28) (35.26) (36.90) (19.00) Obs R Hausman-P Where, Credit Risk i,t represent the proxies of loan quality i.e. ratio of loan loss reserves to gross loans of bank i in time t, Oil Price decrease t log(op t ) if log(op t ) < 0, 0 otherwise and it shows then negative oil price changes at time t, Market Share i,t is the natural log of market share of bank i to total assets of banking system in time t, Capital Asset Ratio i,t shows the equity capital to assets ratio of bank i in time t, Loan Earning Asset Ratio i,t is the share of net loans in earning assets of bank i in time t, Loan growth i,t displays the annual growth of loans of bank i in time t, LNGDP Per Capita i,t shows the natural log of per capita GDP in US$ of country i at time t, GDP Per Capita Growth i,t is annual growth rate of GDP Per Capita of country i at time t, α i is fixed effect, µ i,t is the error term and β 0 is the intercept. R 2 shows the explanatory power, robust standard errors are in brackets and * shows significance at 10%, ** shows significance at 5%, *** shows significance at 1%. Source: Self-construction from Bankscope and World Bank databases 23 P a g e