BANKING SECTOR CONTRIBUTION TO ECONOMIC GROWTH IN ETHIOPIA: EMPIRICAL STUDY

BANKING SECTOR CONTRIBUTION TO ECONOMIC GROWTH IN ETHIOPIA: EMPIRICAL STUDY A.S.Kannan Dr. S. Sudalaimuthu Associate Professor of Management Studies, Associate Professor of Banking Technology, Sri Manakula Vinayagar Engineering College, Pondicherry University, Puducherry-605 107. Puducherry-605 014. Abstract In a financial-institutions-dominated-economy like Ethiopia, banks play a vital role in the economic development of the country. This study intends to identify banking sector contribution to the economic growth of the country. This descriptive study is fully based on secondary data obtained from the annual and other published reports of the National Bank of Ethiopia, and other commercial banks of the country. The study used 16 years data, i.e. from 1997-98 to 2012-13 both years inclusive. Based on the literature, the researcher used OLS regression. The paper presents two models, the first one with Gross Domestic Product (GDP) as the dependent variable, and the second with Gross Capital Formation (GCF). The first OLS model established that Investments, Advances, Deposits and Capital proved to be the right predictors of Gross Domestic Product. The second OLS model tested the predictability of Advances, Deposits and Capital on Gross Capital Formation in the country. Granger Causality tests on both the models established bi-directional relationship between (i) Deposits and GDP, and (ii) Deposits and GCF, and unidirectional relationship between DEP and NLA (in both models), indicating that Deposits creates Advances in the Ethiopian banking system, but not vice-versa. Key Words: Banking Sector, Economic Growth, Empirical Study, Ethiopia, OLS Regression. JEL Classification: C32, G21, O16, E51 (Journal of Economic Literature) 1.0 Introduction Banking institutions are the backbone of the financial system of a country. They function by mobilizing deposits (thereby inducing domestic and commercial savings), and lending loans (thus providing much needed capital funds to the manufacturing, industrial and service enterprises in the country). They facilitate payment system in the country by allowing the use of cheque system. They provide a lot of agency services to their customers, and general utility services to all thus establishing a vital channel in services provision in a country. Finance theorists identify every economy to be either financial-institutions-dominant 72

Million Birr or financial-markets-dominant. That way, Ethiopia is financial institutions dominant country. From a meagre 358 bank-branches all over Ethiopia in 2003-2004, the banking industry has outgrown to 2206 bank-branches [1] in 2013-2014, thus registering a growth rate of 616% over a decade which is phenomenal by any standards. Chart 1: Capital, Advances and Deposits of entire Banking industry of Ethiopia from 1994-95 to 2013-14 Ethiopian Banking Industry: Capital, Advances & Deposits (1995 to 2014) 300000 250000 200000 150000 100000 50000 DEP, 244630.1 ADV, 162232.8 CAP, 28116.6 0 Source: National Bank of Ethiopia Chart 1 shows the actual performance of Ethiopian Banking industry in terms of Capital & Reserves, Loans & Advances, and Deposits mobilized over a period of 19 years from 1994-95 to 2013-14. If one analyzes the data involved, one can find an average growth rate of 25% in Capital & Reserves, of 19.64% in Loans & Advances, and of 19.72% in Deposits mobilized over these 19 years. Ethiopia is reported to have registered a continual double digit growth over the recent decade. According to Deutsche Bank Research (2013), Ethiopia has been experiencing impressive economic growth: double-digit between 2004 and 2010, it has averaged 8.7% annually over the past five years, thanks to the expansion of agriculture and services. Ethiopia has thus been the fastest-growing economy in Sub-Saharan Africa [2]. The Ministry of Finance and Economic Development has announced that Ethiopia s Gross Domestic Product (GDP) has reached 1.5 trillion birr, and registered an estimated 10.3 percent economic growth last Ethiopian fiscal year (i.e. 2013-2014). Ethiopia Fiscal Year 2006 was the 11th consecutive year in which Ethiopia's economy grew on average by 10.9 percent. [3] The World Bank s latest African Pulse, the organization s bi-yearly analysis of Africa s economic status and subsequent prospects, has attributed this growth rate to the 73

17.87% 17.41% 18.37% 17.51% 15.86% 13.68% 14.58% 14.91% 13.80% 13.67% 11.69% 12.54% 13.23% 14.33% 15.14% 15.49% 15.00% increasing investment in natural resources and infrastructure, as well as strong household spending in recent years. Not only have the more obvious resource-rich countries, such as, Sierra Leone and the Democratic Republic of Congo, experienced high levels of growth, but countries improving their political stability, like Mali, as well as those not rich in resources, namely, Ethiopia and Rwanda have also been experiencing solid growth rates. [4] An analysis by The Economist finds that over the ten years to 2010, six of the world's ten fastest-growing economies were in sub-saharan Africa. [5] As observed in the Economist report above, Ethiopia is placed the fifth in GDP growth for the first decade of the third millennium, and is expected to be in the third position for the next five years period, being only behind Asian giants China and India. Chart 2: Proportion of Advances to GDP in Ethiopia (1998 to 2014) 20.00% 18.00% 16.00% 14.00% 12.00% 10.00% 8.00% 6.00% 4.00% 2.00% 0.00% Bank Advances/GDP Source: National Bank of Ethiopia 74

Bank Advances / Total Pvt. Credit As observed in chart 2, the commercial bank advances in Ethiopia accounted for an average of 15 percent of the GDP of the country in the 16 years period reported. The highest proportion of 17.87 percent is recorded in 1998-99 and the lowest of 11.69 percent was recorded exactly 10 years later, i.e. in 2008-09. Recent years indicate some kind of stability with 15 percent proportion in GDP. Chart 3: Proportion of Bank Advances in Total Private Credit in Ethiopia 140% 120% 100% 80% 60% 40% 20% 0% Proportion of Bank Advances in Total Private Credit (1998-99 to 2013-14) 105% AVERAGE Advances to Total Private Credit, 97% Source: National Bank of Ethiopia Chart 3 presented in the previous page evidences the greater proportion of commercial bank advances in total private credit in Ethiopia, thus proving the fact that Ethiopia s economy is financial-institutions dominant economy. If one observes the trend of the proportion of commercial bank advances in total private credit which is marginally declining, with the highest proportion of 123 percent being achieved in 2004-05 and the lowest of 77 percent in the latest fiscal year, i.e. 2013-14, with the average for the reported 16 years as 97 percent. The data presented above indicates that banks do contribute significantly to the total private credit in the country, and a reasonable proportion of bank advances to the GDP of Ethiopia. However, the question remains to be investigated whether banks significantly contribute to the growth of the economy of Ethiopia. The purpose of the paper is to empirically evaluate the contributions of Ethiopian Banking sector to the economic growth of the country. For this purpose, the paper takes into account the macro level data pertaining to 14 years from 1999-2000 to 2012-2013. 2.0 Review of Literature Koivu (2002) investigates the relationship between financial sector and economic growth by using empirical methods, data variables INT = Difference between lending and 75

deposit interest rates as percentage points. CREDIT = Ratio of bank credit to private sector to GDP. RI = Reform index. INF = Annual consumer price index as percentages. GDP growth = Real GDP growth rate. Fixed-effects panel model techniques have been used. [6] Poshakwale and Qian (2007) have empirically investigated the impact of financial reforms on assertiveness and growth efficiency of the banking industry, as well and find the long-term and short-term impact on economic growth in Egypt during the period of 1992 to 2007. The study suggests that the reforms have a positive and significant effect on assertiveness and growth efficiency in Egypt banking industry. It shows that government banks are generally well efficient than private banks and foreign banks are less aggressive than domestic banks. The total inefficiency of Egyptian government banks is approximate 30%. [7] The main objective of the study by Khulaifi et al (2009) was to examine empirically the functional relationships between bank's profit, from one hand and between the real sectors, such as, GDP, foreign interest rates, government revenues, government expenditures' and bank's equity effect on the other hand. Different statistical exercises including stability tests for structural stability of the model and causality experiments were conducted. Granger causality tests were performed on all variables, and bank equities. GDP Granger causes bank profit indirectly through bank equities. The statistical findings indicated that the predictability of these models is highly accurate and could be applied to similar countries or institutions. [8] Fadare (2010) empirically identifies the effect of banking sector reforms on economic growth in Nigeria by using the data 1999-2009. Variables used for the study are interest rate margins, parallel market premiums, total banking sector credit to the private sector, inflation rate, inflation rate lagged by one year, size of banking sector capital and cash reserve ratios. Results indicate that the relationship between economic growth and other exogenous variables of interest rate margins, parallel market premiums, total banking sector credit to the private sector, inflation rate and cash reserve ratio is negative and insignificant. [9] Mwenda and Mutoti (2011) have investigated the effects of market-based financial sector reforms on the competitiveness and efficiency of commercial banks, and economic growth, in Zambia. It was found that bank overall cost efficiencies, financial depth during stage two and stage three in which financial sector reforms have been discussed. They concluded that degree of economic freedom and rate of inflation had a significant impact on economic growth. Stage two policies and the inflation rate have negative effects; however, the remaining of the variables has positive results on economic growth. [10] Yazdani (2011) investigates the role and performance of private banks on the economic growth of Iran by using the variables economic growth, profitability, cash, and 76

investment. The part of financial sectors with relation to private banks and the impact on economic growth of the Islamic republic of Iran have been analysed. Spearman correlation test, Pearson Correlation test, David Watson test, independent t test, variance analysis F and linear regression chart have been used. Results obtained through analysis shows that all variables have a definite impact on the economic growth of Iran. [11] The study by Aurangzeb (2012) investigated the contributions of banking sector in economic growth of Pakistan. The data used in this study were collected from the period of 1981 to 2010 of 10 banks. Augmented Dickey Fuller (ADF) and Philip Perron unit root test, ordinary least square and granger causality test have been used. Unit root test confirms the stationary of all variables at first difference. Regression results indicate that deposits, investments, advances, profitability and interest earnings have significant positive impact on economic growth of Pakistan. The Granger-Causality test confirms the bidirectional causal relationship of deposits, advances and profitability with economic growth. On the other side, the study found unidirectional causal relationship of investments and interest earnings with economic growth runs from investments and interest earnings to economic growth. [12] The study to examine the impact of bank lending on economic growth in Nigeria was carried out by Mamman & Hashim [13]. In addition, the objective of this study was to examine the impact of bank lending on economic growth in Nigeria for the period 1987 to 2012. This study relied purely on secondary data and used multiple regression model. The study found out that bank lending accounted for about 82.6% variation in economic growth in Nigeria for the period under study. The study concluded that there is a statistically significant impact of bank lending on economic growth in Nigeria. The studies reviewed above established the nexus between banking sector of the country and the country s economic growth. In almost all these studies, GDP has been taken as the proxy for economic growth. As predictor variables, the studies mostly used some of the variables, such as, Advances (ADV), Investments (INV), Deposits (DEP), Interest Earnings (INE), Profitability (PRF), Gross Capital Formation (GCF), Total Bank Credit (TBC), Inflation (INF), and Capital and Reserves (CAR). Almost all studies used different forms of Regression, mainly focusing on Ordinary Least Squares (OLS) method. 3.0 Methodology This study intends to identify banking sector contribution to the economic growth of the country. This descriptive study obtains secondary data from the annual and other published reports of the National Bank of Ethiopia, and other commercial banks of the country. The study used 16 years data, i.e. from 1997-98 to 2012-13, both years inclusive. Based on the 77

understandings from the literature, the researcher decided to use OLS regression model with appropriate diagnostic tests for Classic Linear Regression Model (CLRM) as prescribed in authoritative texts. The paper presents four models, the first one with GDP as the dependent variable, the second with GCF, the third with Total Private Credit (TPC), and the final with Panel Data on GDP as DV. 3.1 OLS Model with GDP as DV While following the tradition of taking GDP as the dependent variable as a proxy for economic growth, this study used INV (Investments), NLA (Net Loans & Advances), DEP (Deposits), and CAR (Capital and Retained Earnings) as predictors in the primary model. The model is specified thus: Where Y = α + β 1 INV + β 2 NLA + β 3 DEP + β 4 CAR + Ut Y = Economic Growth (proxy by the GDP) α = constant β 1 to β 4 = regression coefficients INV = Investments of the Commercial Banks NLA = Net Loans & Advances of the Commercial Banks DEP = Deposits of the Commercial Banks CAR = Capital and Retained Earnings of the Commercial Banks U t = Unknown or the error term 3.2 OLS Model with GCF as DV This OLS regression model used GCF as Dependent Variable, and NLA, DEP and CAR as Independent Variables. The model is thus specified: Where Y = α + β 1 NLA + β 2 DEP + β 3 CAR + Ut Y = Gross Capital Formation β 1 to β 3 = regression coefficients NLA = Net Loans & Advances of the Commercial Banks DEP = Deposits of the Commercial Banks CAR = Capital and Retained Earnings of the Commercial Banks U t = Unknown or the error term 3.3 Description of the Explanatory Variables Various predictor variables used in all the above models are briefly described and their expected signs are presented in the following paragraphs. 78

INV (INVESTMENTS): This explanatory variable refers to the investments made by commercial banks in Ethiopia in various options. In the recent period, this amount is predominantly comprised of the Investment in NBE Bills, with interest accruing at 3% per annum. This is expected to possess (-)ve sign, since it gives a less than market return. Since the rate of interest fetched by the investments is lower than the project rate of return, it is likely to be negatively contributing to the growth of the GDP. NLA (NET LOANS & ADVANCES): This predictor variable refers to the amount of loans and advances granted by commercial banks to various sectors of the economy, after being adjusted for provision as per norms. This variable is expected to be (+)ve sign, since the bank advances are likely to aid capital formation in the country, and thereby, enhance the scope for higher growth in country s GDP. DEP (DEPOSITS): This variable comprises of various types of deposits (viz., savings, demand and time) accepted by commercial banks of the country. This variable is expected to carry a (+)ve sign, as it is contributing to the loanable funds for banks. CAR (CAPITAL & RETAINED EARNINGS): This variable consists of paid-in capital, retained earnings, and profit & loss account balance for the commercial banks of Ethiopia. This is expected to be of (+)ve sign. 4.0 Results and Discussion This section of the paper deals with the results obtained by the use of different regression models in this study. It presents the descriptive statistics, the OLS regression model results summary, the fulfillment of diagnostic tests for the classical linear regression model (CLRM) in a brief way with appropriate explanations. Table 1: Descriptive Statistics (for Variables used in Model 1) GDP INV NLA DEP CAR Mean 247637.4 26080.02 32881.45 66279.76 7172.310 Median 117874.5 15672.00 17464.50 39228.50 3254.500 Maximum 864673.0 106025.8 115296.6 232950.4 23225.41 Minimum 55534.00 2209.000 8854.000 14528.00 1229.000 Std. Dev. 257453.2 29535.73 32563.61 64728.54 7030.476 Skewness 1.353575 1.661343 1.523196 1.463896 1.060704 Kurtosis 3.575443 4.758552 4.135916 4.037495 2.903019 Observations 16 16 16 16 16 Source: EViews output 79

Table 1 presents the descriptive statistics, i.e. Mean, Medium, Maximum, Minimum, Standard Deviation, Skewness, and Kurtosis values, for all the variables, viz., GDP, Investments (INV), NLA, Deposits (DEP), and CAR. Among the variables, Deposits show the highest Mean, Median and Standard Deviation values. 4.1 Summary results of the OLS Regression Model 1 Table 2: Regression Model 1 Dependent Variable Method Sample Observations Gross Domestic Product Least Squares 1998 to 2013 16 (years) Predictor Variables Table 3: Summary Results of the Model 1 Coefficient Standard Error t-statistic Probability (p-value) Level of Significance Constant -24270.48 6183.287-3.925174 0.0024 1 percent *** Investment (INV) -2.749438 1.147666-2.395677 0.0355 5 percent ** Net Loans & Advances (NLA) 5.964863 1.037964 5.746696 0.0001 1 percent*** Deposits (DEP) 1.520759 0.787382 1.931411 0.0796 10 percent * Capital & Retained Earnings (CAR) 6.508979 3.165811 2.056023 0.0643 10 percent * R-squared value Adjusted R-squared value F-statistic (probability) 0.998367 0.997774 1681.601 (0.000000) Durbin-Watson test value 2.005596 The OLS model used GDP as Dependent Variable, and INV, NLA, DEP and CAR as Predictor Variables. The model established that NLA is highly significant (at 1% level) in predicting the GDP with a co-efficient of 5.964863. It has positive sign as expected. INV is the next significant variable (at 5% level) in predicting GDP with a co-efficient of -2.749438. It is bearing the negative sign since the investments in Ethiopian commercial banks are mostly represented by NBE bills which bear a meagre 3% interest rate (obviously lower than the market return). Hence, this is negatively contributing to the GDP growth. Both DEP and CAR are significant (at 10% level) in predicting GDP with co-efficient of 1.520759 and 6.508979 respectively. The R-squared value for the model is 0.998367, and the adjusted R- squared value is 0.997774, indicating that the model fit is 99.7%. The Durbin-Watson test 80

value is 2.005596. It lies between d u and 4-d u (the no-evidence-of-auto-correlation region in the DW line). Hence, it can be concluded that this model is free from auto-correlation. 4.3 Equation derived from the model As explained in the paragraphs above, this model fulfills almost all the diagnostic tests prescribed for the Classical Linear Regression Model (CLRM), and thus considered robust. Thus, the model is expressed in the form of the following equation: Y = α + β 1 INV + β 2 NLA + β 3 DEP + β 4 CAR + Ut Accordingly, as found in the summary results presented above, the regression equation will be: GDP = f (INV, NLA, DEP, CAR) GDP = -24270.48-2.749438 INV + 5.964863 NLA + 1.520759 DEP + 6.508979 CAR + U t Since the adjusted R-square is 0.9977, it indicates that this model explains 99.77% of the explanatory variables. 4.4 Granger Causality Tests Granger causality is a statistical concept of causality that is based on prediction. According to Granger causality, if a signal X 1 "Granger-causes" (or "G-causes") a signal X 2, then past values of X 1 should contain information that helps predict X 2 above and beyond the information contained in past values of X 2 alone. (Anil Seth, 2007). [15] Table 4: Granger Causality Tests for Model 1 Pairwise Granger Causality Tests Date: 05/28/15 Time: 08:17 Sample: 1998 2013 Lags: 2 Null Hypothesis: Obs F-Statistic Prob. INV does not Granger Cause GDP 14 1.09759 0.3745 GDP does not Granger Cause INV 2.68766 0.1216 NLA does not Granger Cause GDP 14 0.96820 0.4161 GDP does not Granger Cause NLA 2.34053 0.1519 DEP does not Granger Cause GDP 14 5.96676 0.0224 GDP does not Granger Cause DEP 39.8229 3.E-05 DEP does not Granger Cause INV 14 15.5490 0.0012 INV does not Granger Cause DEP 2.06861 0.1823 DEP does not Granger Cause NLA 14 33.6169 7.E-05 81

NLA does not Granger Cause DEP 2.57140 0.1308 Source: EViews output As indicated by the results of Granger Causality tests, presented in table 4, there is a bi-directional causality relationship between Deposits (DEP) and Gross Domestic Product (GDP). There exists a unidirectional relationship between Deposits and Investments (DEP granger causing INV), and between Deposits and Net Loans & Advances (DEP granger causing NLA). There is no causality relationship between NLA and DEP. This is obvious from the economic scenario of the country. In Ethiopia, there is a strong association between Deposits and Net Loans & Advances, evidenced by the correlation co-efficient between them being 0.995249 (as per table 3) which indicates perfect positive correlation between the two. NLA does not granger cause DEP, because Ethiopia is more of Cash Economy, the funds loaned by commercial banks are not fully ploughed back to the banking system of the country. 4.6 OLS Regression Model 2 with GCF as DV The second model considers the Gross Capital Formation as a Dependent Variable, and NLA, DEP and CAR as Predictors. The results and analysis of the model are briefly presented in the following paragraphs. 4.6.1 Summary results of the OLS Regression Model 2 Table 5a: Regression Model 2 Dependent Variable Method Sample Observations Gross Capital Formation Least Squares 1998 to 2013 16 (years) Table 5b: Summary Results of the Model 2 Predictor Variables Coefficient Standard Error t-statistic Probability (p-value) Level of Significance Constant -19089.50 2627.898-7.264170 0.0000 1 percent *** Net Loans & Advances 1 percent *** (NLA) 3.255150 0.589536 5.521546 0.0001 Deposits (DEP) 0.285843 0.322443 0.886489 0.3928 Capital & Retained Earnings (CAR) -4.343231 1.529703-2.839264 0.0149 5 percent ** R-squared value Adjusted R-squared value F-statistic (probability) 0.995392 0.994241 1681.601 (0.000000) Durbin-Watson test value 1.857776 82

The OLS model used GCF as Dependent Variable, and NLA, DEP and CAR as Predictor Variables. The model established that NLA is highly significant (at 1% level) in predicting the GDP with a co-efficient of 3.255150. It has positive sign as expected. CAR is the next significant variable (at 5% level) in predicting GDP with a co-efficient of 4.343231. It is bearing the negative sign since the capital and earnings accumulation by banking industry does not contribute to the gross capital formation (mainly in the private sector) in the country. This is because the loanable capacity of the commercial banks is reduced by the reserves they maintain since: Deposits Actual Reserves Compulsory Investments (NBE Bills) = Loanable Funds. Hence, this is negatively contributing to the GDP growth. Though DEP is not significant, it has a positive co-efficient of 0.285843. The R-squared value for the model is 0.995392, and the adjusted R-squared value is 0.994241, indicating that the model fit is 99.4%. The Durbin-Watson test value is 1.857776. It lies between d u and 4-d u (the noevidence-of-auto-correlation region in the DW line). Hence, it can be concluded that this model is free from auto-correlation. Table 6: Results of Granger Causality Test for Model 2 Pairwise Granger Causality Tests Date: 05/28/15 Time: 08:44 Sample: 1998 2013 Lags: 2 Null Hypothesis: Obs F-Statistic Prob. NLA does not Granger Cause GCF 14 1.46237 0.2819 GCF does not Granger Cause NLA 0.19729 0.8244 DEP does not Granger Cause GCF 14 61.3444 6.E-06 GCF does not Granger Cause DEP 8.16451 0.0095 CAR does not Granger Cause GCF 14 1.77209 0.2244 GCF does not Granger Cause CAR 0.01865 0.9816 DEP does not Granger Cause NLA 14 33.6169 7.E-05 NLA does not Granger Cause DEP 2.57140 0.1308 Source: EViews output Table 6 presents the results of the Granger Causality test for model 2. Accordingly, there is a bi-directional relationship between DEP and GCF, thus indicating that DEP granger causes GCF, and GCF granger causes DEP. There is a unidirectional relationship between DEP and NLA, which is similar to the result obtained from model 1 earlier. This indicates 83

again that Deposits creates Advances in the Ethiopian banking system, but not vice-versa. Advances do not create Deposits as they are not fully ploughed back to the banking system, but largely remain as cash economy. 5.0 Conclusion to the study This study attempted to identify banking sector contribution to the economic growth of the country. This descriptive study is fully based on secondary data from official and authentic sources. Data obtained from the annual and other published reports of the National Bank of Ethiopia, and other commercial banks of the country are used for this purpose. The study used 16 years data, i.e. from 1997-98 to 2012-13 both years inclusive. Based on the understandings from the literature, the researcher decided to use OLS regression model. The first OLS model results indicate that Investments, Net Loans & Advances, Deposits and Capital & Reserves proved to be the right predictors of the Gross Domestic Product, thus proving the contribution of the banking sector to the economic growth of Ethiopia. The second OLS model tested the predictability of Net Loans & Advances, Deposits and Capital & Reserves on the Gross Capital Formation in the country. The model proved that NLA, DEP and CAR as right predictors of the Gross Capital Formation in Ethiopia thus proving the role of banking sector in economic growth through capital formation. Granger Causality tests on both the models established bi-directional relationship between (i) Deposits and GDP, and (ii) Deposits and GCF, and unidirectional relationship between DEP and NLA (in both models), indicating that Deposits creates Advances in the Ethiopian banking system, but not vice-versa. 5.1 Practical Implications of the study There have been many studies in the world in different countries establishing econometrically through various models the contribution of banking sector to the economic growth. But such studies are a bit rare in Ethiopia. In that count, this study assumes significance as it clearly established through proper diagnostic tests with a robust model that the Ethiopian banking sector does contribute significantly to the economic growth of the country not with one model but with two models taking Gross Domestic Product in the first, and Gross Capital Formation in the second as Dependent Variables. The study will be highly useful for different stakeholders including the policymakers and academic researchers. 5.2 Research Limitations and Scope for further research Inadequacy of the data is a major limitation for the study. The study took only 16 years data from 1997-98 to 2012-13. The reason for this restriction is Ethiopia opened up the banking industry for privatization only in the 1994-95, and till 1997-98, only 4 private banks were 84

added to the system, one in each year. Furthermore, the democratic government started its operations in 1994 only, and meaningful data for banking industry was available only from 1997-98 fiscal years. Future researches can be taken up for study with more years of data. Since the data for state-owned banks and of private banks were not really comparable in the early years of data, the panel model failed to be robust. Once again future research can successfully handle this issue. Another major problem faced in this study is the multicollinearity problem. The study took only banking related variables such as Bank Investments, Bank Deposits, Bank Advances, and Bank Capital as Predictor Variables, which are highly correlated to each other as the banking system in general moves in tandem with the deposits mobilization. This problem can be avoided in future researches by adding some of the macro economic variables and thus, overcome the multicollinearity issue. REFERENCES [1] National Bank of Ethiopia. Annual Reports. Available at www.nbe.gov.et accessed in March-April 2015. [2] Kannan Simhakutty & Sudalaimuthu Shanmugam. (2015). MSME Financing: Empirical comparison between select Emerging and Developing Economies. Paper presented in the National Seminar on Empowering Small and Medium Enterprises for Global Competitiveness at Pondicherry University, India, in Feb.2015. [3] Ministry of Foreign Affairs. (2014). Ethiopia s GDP reaches 1.5 trillion birr. News item available at http://www.mfa.gov.et/news/more.php?newsid=3627, accessed in April 2015. [4] KPMG Africa. (2014). Africa s Top 10 Fastest Growing Economies. Available at http://www.blog.kpmgafrica.com/africas-top-10-fastest-growing-economies/ accessed in April 2015. [5] The Economist Online. (2011). Daily Chart: Africa s impressive growth. Available at http://www.economist.com/blogs/dailychart/2011/01/daily_chart, accessed in April 2015. [6] Koviu, T. (2002, cited in Aurangzeb, Economics & Finance Review, Vol.2 (6), August 2012). Do efficient banking sectors accelerate economic growth in transition countries? [7] Poshakwale & Qian. (2007). Competitiveness and efficiency of the banking sector and economic growth in Egypt. Available at: http://www.uneca.org/sites/default/files/page_attachments/aec2009competitivenessandefficiency-of-thebankingsector.pdf, accessed in May 2015. 85

[8] Khulaifi, Sulaiti & Khatib. (2009). Banking Performance and Economic Growth in Qatar: An empirical investigation. Journal of Administrative Sciences and Economics, 199, Vol.10, pp.5-27. Available at http://qspace.qu.edu.qa/handle/10576/6927, accessed in May 2015. [9] Fadare, S.O. (2010). Recent banking sector reforms and economic growth in Nigeria. Middle Eastern Finance and Economics. Issue: 146-160. [10] Mwenda, Abraham & Mutoti, Noah. (2011). Financial Sector Reforms, Bank Performance and Economic Growth: Evidence from Zambia. African Development Review, 2011, Vol.23, Issue 1, pp.60-74. [11] Yazdani, Masoud. (2011). Role of performance of privately owned banks in economic growth of Iran. Australian Journal of Basic and Applied Sciences, Vol.5, Issue 11, pp.695-700, 2011. [12] Aurangzeb. (2012). Contributions of Banking Sector in Economic Growth: a case of Pakistan. Economics and Finance Review, Vol.2(6), pp.45-54, August 2012. Available online at http://www.businessjournalz.org/efr, accessed in March 2015. [13] Aliyu Mamman & Yusuf Alhaji Hashim.(2014). Impact of Bank Lending on Economic Growth in Nigeria. Research Journal of Finance and Accounting, Vol. 5, No.18, 2014, pp.174-182. [14] Brooks, Chris. (2008). Introductory Econometrics for Finance, 2 nd Edition, Cambridge University Press. [15] Anil Seth. (2007). Scholarpedia, 2(7):1667, Available at: doi:10.4249/scholarpedia.1667, accessed in May 2015. 86